REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
20
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
21
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
22
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
23
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
24
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
25
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
26
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Variable Prediabetes n=221
[RIC]
Sin prediabetes
n=442 [RIC]
Total
n=663 [RIC] p-valor
IMC 29,46 [26,28-33,46] 25,40 [21,87-30,13] 27,02 [22,89-31,30] <0,001
HbA1c 5,90 [5,80-6,10] 5,30 [5,10-5,5] 5,50 [5,20-5,9] <0,001
Glicemia 109 [104-115] 93,3 [87,0-99,20] 99,20 [89,6-108] <0,001
ITG 8,86 [8,54-9,24] 8,44 [8,01-8,89] 8,59 [8,17-9,04] <0,001
Colesterol 194 [168-223] 181 [153-212] 184 [157-218] <0,001
Trigliceridos 131 [94-186] 98 [66-152] 107 [73-163] <0,001
LDL 113 [92-137] 101 [79-129] 103 [83-134] <0,001
HDL 50 [43-59] 53 [44-63] 52 [44-62] 0,030
Albúmina 4,1 [3,9-4,3] 4,3 [4-4,5] 4,2 [4-4,4] <0,001
AST 22 [17-29] 19 [15-26] 20 [16-27] <0,001
ALT 23 [20-28] 23 [20-27] 23 [20-28] 0,080
Las variables se expresan como medianas y rangos intercuartílicos (RIC). Las diferencias se evaluaron mediante la
prueba U de Mann-Whitney; IMC: índice de masa corporal; HbA1C: Hemoglobina glicosilada; ITG:Índice
triglicéridos-glucosa; LDL: lipoproteína de baja densidad; HDL: lipoproteína de alta densidad; AST: Aspartato
aminotransferasa; ALT: Alanina aminotransferasa.
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
27
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Variable
β
Wald
p-value
OR [95% CI]
Age
-0.007
0.244
0.622
0.99 [0.97 – 1.02]
Sex
-0.426
4,155
0.042
0.65 [0.43 – 0.98]
TGI
1,041
28,904
0.000
2.83 [1.94 – 4.14]
LDL
0.002
0.847
0.358
1.00 [1.00 – 1.01]
HDL
0.013
2,916
0.088
1.01 [1.00 – 1.03]
Albumin
-1,097
16,374
0.000
0.33 [0.20 – 0.57]
AST
0.002
0.033
0.856
1.00 [0.98 – 1.03]
ALT
0.008
0.253
0.615
1.01 [0.98 – 1.04]
OO
1,196
25,707
0.000
3.31 [2.08 – 5.25]
Constant
-6,250
7,438
0.006
BMI: Body mass index; HbA1c: Glycated hemoglobin; TGI: Triglyceride-glucose index; LDL: Low-density
lipoprotein; HDL: High-density lipoprotein; AST: Aspartate aminotransferase; ALT: Alanine aminotransferase;
OO: Overweight and obesity.
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
28
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Criterion VP FP FN VN S AND VPP VPN
TGI >=8.54 166 185 55 257 75.1% 58.1% 0.47 0.82
ALB <4.15 121 165 100 277 54.8% 62.7% 0.42 0.73
TGI+OO 146 128 75 314 66.1% 71.0% 0.53 0.81
TGI+ALB+OO 80 59 141 383 36.2% 86.7% 0.58 0.73
TGI+ALB+OO 89 79 132 363 40.3% 82.1% 0.53 0.73
VP: True positives; FP: False positives; FN: False negatives; TN: True negatives; S: Sensitivity; E: Specificity;
PPV: Positive predictive value; NPV: Negative predictive value; TGI: Triglyceride-glucose index; OO: Overweight
and obesity; ALB: Albumin.
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
29
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
30
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
31
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
32
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
33
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
34
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
35
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
Triglyceride-Glucose Index in the Prediction of Prediabetes
Índice triglicéridos-glucosa en la predicción de prediabetes
https://doi.org/10.37135/ee.04.25.02
Authors:
Jorman Francisco Choez Alava
1,2
- https://orcid.org/0000-0002-0073-3795
Marja Morales Baldeon
2
- https://orcid.org/0009-0000-3150-3290
Carmen Vanessa Vaca Vera
2,3
- https://orcid.org/0009-0001-8867-1276
Bertha Carolina Cruz Murillo
4
- https://orcid.org/0009-0001-9399-2939
Affiliation:
International University of La Rioja – Spain.
Surgical Clinical Center of the Ecuadorian Social Security Institute - Guayaquil, Ecuador.
Hemispheres University – Quito, Ecuador
University of Guayaquil – Guayaquil, Ecuador
Corresponding author: Jorman Francisco Choez Alava, International University of La Rioja, Rectorate, Av.
de la Paz, 93-103, 26006 Logroño, La Rioja, Spain, E-mail: jormanfrancisco.choez064@comunidadunir.net,
+593 967646036
Received: May, 19 2025 Accepted: November, 21 2025
ABSTRACT
Prediabetes is a metabolic disorder characterized by insulin resistance long before the diagnosis of type 2
diabetes mellitus (T2DM) and represents a key opportunity for intervention and prevention of T2DM. The
triglyceride-glucose index (TGI) has been identified as an accessible marker of insulin resistance with potential
diagnostic value. This study aimed to evaluate the diagnostic accuracy of the TGI in predicting prediabetic
status in nondiabetic adults. A case-control study was conducted using retrospective data from 663 nondiabetic
adults treated at an outpatient care center in Guayaquil between 2019 and 2023. 221 cases with Prediabetes
and 442 controls matched for age and sex were selected. Nonparametric tests, binary logistic regression, and
ROC curve analysis were applied. TGI was significantly associated with OR: 2.83 [95 % CI 1.94–4.14]. A
0.82. The combination of TGI with overweight/obesity and albumin levels <4.15 g/dL improved specificity
to 86.7 %. Low albumin and being overweight were also independently associated with an increased risk of
Prediabetes. The TGI demonstrated adequate diagnostic capacity in detecting Prediabetes, making it a valuable
and cost-effective marker for T2DM screening. Its combination with other variables improves diagnostic
accuracy, and future validations were planned to expand its clinical application.
Keywords: Triglycerides, Blood Glucose, Diabetes Mellitus, Prediabetic State, Insulin Resistance.
RESUMEN
La prediabetes es un estado de alteración metabólica caracterizado por la resistencia a la insulina mucho antes
del diagnóstico de diabetes mellitus tipo 2 (T2DM) y representa una oportunidad clave para la intervención
y prevención hacia T2DM. El índice triglicéridos-glucosa (ITG) se ha identificado como un marcador accesible
de resistencia a la insulina, con valor diagnóstico potencial en este contexto. El objetivo de este estudio fue
evaluar la precisión diagnóstica del ITG en la predicción del estado prediabético en adultos no diabéticos. Se
realizó un estudio de casos y controles con datos retrospectivos de 663 adultos no diabéticos atendidos entre
2019 y 2023 en un centro de atención ambulatoria de Guayaquil. Se seleccionaron 221 casos con prediabetes
y 442 controles emparejados por edad y sexo. Se aplicaron pruebas no paramétricas, regresión logística binaria
y análisis de curvas ROC. El ITG se asoció significativamente OR: 2,83 [IC95 % 1.94 – 4.14]. Un punto de
0,82. La combinación de ITG con sobrepeso/obesidad y albúmina <4,15 g/dL mejoró la especificidad hasta
86,7 %. La albúmina baja y el sobrepeso también se asociaron independientemente con mayor riesgo de
prediabetes. El ITG mostró adecuada capacidad diagnóstica en la detección de prediabetes, por lo que
representa un marcador útil y económico para el tamizaje de T2DM. Su combinación con otras variables
mejora la precisión diagnóstica, además de futuras validaciones a fin de ampliar la aplicación clínica.
Palabras clave: triglicéridos, glucemia, diabetes mellitus, estado prediabético, resistencia a la insulina.
INTRODUCTION
Metabolic syndrome is a well-known clinical entity characterized by the presence of specific factors that
predispose individuals to developing cardiovascular disease and type 2 diabetes mellitus (T2DM).
(1–3)
Globally,
diabetes is the eighth leading cause of death.
(4)
In Ecuador, the prevalence of diabetes is estimated at 10% in
adults over 50 years of age, making it the second leading cause of death in 2022 and 2023.
(5)
These figures
are alarming, due to the rapid increase in the incidence of diabetes,
(6,7)
but mainly because its diagnosis is
becoming less exclusive to older people, and at the same time, society is rapidly adopting sedentary lifestyles
in young people.
(8,9)
According to reports from a study conducted in 146 countries on adolescents between 11
and 17 years of age, the global trend of insufficient physical activity up to 2019 was 80 %, and it is 86.5 %
in Ecuador.
(10)
Regarding the pathophysiological basis of type 2 diabetes mellitus (T2DM), it is known to be a metabolic
disorder that initially involves insulin resistance and pancreatic beta-cell dysfunction.
(11,12)
This leads to a
transition between normal glucose metabolism and T2DM, a condition known as Prediabetes. The prediabetic
state is defined as an intermediate condition between normal glucose metabolism and type 2 diabetes
mellitus (T2DM), characterized by blood glucose levels higher than usual but not yet meeting the diagnostic
criteria for diabetes. Current criteria consider blood glucose levels between 100 and 125 mg/dL as Prediabetes
and a level greater than or equal to 126 mg/dL as diabetes.
(13)
Over the years, there has been a considerable
increase in the prevalence of diabetes mellitus;
(9,14)
however, early diagnosis using current diagnostic criteria
and measures to treat the disease do not appear to be significantly impacting the decline of this epidemic.
(14,15)
Estimating insulin resistance is helpful for predicting type 2 diabetes mellitus (T2DM); however, precise
measurement of blood insulin levels is not readily available to the entire population, especially in low-income
countries.
(16)
Therefore, other options have been proposed, such as determining the triglyceride-glucose
index (TGI) for assessing metabolic status and insulin resistance,
(17–19)
which has demonstrated equal or greater
quantification value. The triglyceride-glucose index is defined as the negative logarithm of the product of
glucose and triglyceride values divided by two, represented by the following formula: I<sub>n</sub>
[Triglycerides [mg/dl] × glucose [mg/dl]/2).
(20)
Research over the last decade has demonstrated the usefulness of the TGI in estimating metabolic status and
insulin resistance
(20–26)
, interpreted as a sign of the initial deterioration of metabolic status that precedes the
development of T2DM. In the Mexican population, the TGI has been shown to assess insulin resistance
accurately.
(19)
Systematic reviews have evaluated cutoff points; however, it is considered that further studies
are still needed in this regard.
(27)
The TGI has become an essential predictor of prediabetic status and its progression or regression toward
normoglycemia or diabetes. Several studies have found that TGI can serve as a surrogate marker for insulin
resistance, as it has shown a non-linear relationship with glucose status conversion, with an inflection point at
a TGI value of 8.88. Beyond this value, the probability of returning to normoglycemia decreases significantly
in individuals with Prediabetes.
(28)
Furthermore, combining TGI with body mass index (BMI) improves the
predictive accuracy of prediabetes recovery or progression, with specific thresholds identified for predicting
recovery and progression.
(29)
The predictive capacity of TGI is further supported by its significant correlation
with markers of insulin resistance and its superior predictive ability compared to other indices, particularly
in women and obese individuals.
(30,31)
Furthermore, the TGI has been validated as a reliable predictor of
prediabetes risk in several populations, including middle-aged and older adults, with a demonstrated
non-linear relationship between TGI values and diabetes risk.
(32,33)
In most cases, the time of diabetes diagnosis does not represent a point at which the progression of the underlying
metabolic disorder can be reversed.
(34,35)
Therefore, the need arises to predict diabetes at its earliest stages,
that is, at the first signs of insulin resistance, even when fasting glucose levels fluctuate between Prediabetes
and normal.
(36)
Thus, it is essential to investigate tools that allow us to know the metabolic state before
reaching the point of no return that type 2 diabetes and the prediabetic state represent. Considering this
background and the evidence on estimating insulin resistance from TGI, we hypothesize that it is possible
to predict the diagnosis of Prediabetes from the TGI estimate. The objective of this research is to evaluate
the diagnostic accuracy of the TGI in predicting the prediabetic state.
MATERIALS AND METHODS.
A case-control design is presented to evaluate the diagnostic accuracy of the TIG in predicting Prediabetes
in nondiabetic adult patients treated at the outpatient service of the Surgical Clinical Center of Northern
Guayaquil, Ecuador, between 2019 and 2023, as part of the Ecuadorian Social Security Institute (IESS).
Population and sample
The population consists of 41,713 adult patients who attended CCQANT-IESS for outpatient follow-up for
causes other than diabetes during the period from January 2019 to December 2023.
The minimum sample size was estimated using Epi Info™ StatCalc software, assuming a population of
41,713 patients, an expected prevalence of 50 %, a 99 % confidence level, and a 5 % margin of error, resul-
ting in a minimum of 653 participants.
To form the sample, 9096 clinical records with data on HbA1c, lipid profile, and glucose levels were identified.
Those individuals who met the criteria for Prediabetes (ADA 2024)
(13)
(fasting glucose between 100 and 125
mg/dL, HbA1c between 5.7 % and 6.4 %, and compatible symptoms recorded in the medical history) were
then identified. 829 records with Prediabetes were identified, from which 221 prediabetes cases were randomly
selected, and from the remaining 442 controls, matched by age and sex, were randomly selected at a ratio of
2 controls per case to improve statistical power, according to the literature.
(37)
Inclusion and exclusion criteria
Nondiabetic patients were included based on laboratory test records of HbA1c, fasting glucose, lipid profile
(Total Cholesterol, High and Low Density Lipoproteins (HDL and LDL), triglycerides), and body mass
index (BMI).
Patients under 18 years of age were excluded, as were those with a prior diagnosis of metabolic diseases or
endocrinopathies (type 1 diabetes mellitus, uncontrolled thyroid disorders, Cushing's syndrome, or other
hormonal dysfunctions); documented history of cardiovascular disease (myocardial infarction or heart failure);
advanced chronic renal failure; liver cirrhosis; pregnancy; and those with incomplete clinical records for the
study variables. The exclusion of these clinical conditions was considered to control for confounding bias.
Variables
Quantitative variables include age (measured in years), body mass index (BMI), fasting glucose, triglycerides,
HDL, LDL, total cholesterol (all in milligrams per deciliter), and HbA1c (in grams per deciliter). Qualitative
variables include sex and prediabetes diagnosis. BMI is classified as an ordinal qualitative variable, with
ranges defined by the WHO.
(38)
Data collection
After obtaining authorization from the center for data collection, a database from the Laboratory Department
containing 41713 laboratory records of nondiabetic adult patients (2019–2023) was retrospectively
reviewed. Of these, 9096 had records of HbA1c, lipid profile, and glucose levels. Following the initial
selection of cases and controls, the medical records were individually reviewed to verify compliance with the
inclusion and exclusion criteria. In cases where a patient had a documented exclusion condition, they were
removed from the sample and replaced with another randomly selected patient who met the corresponding
age and sex criteria to control for selection bias. Relevant clinical, anthropometric, and biochemical data
were extracted from the electronic records for analysis. To control for confounding bias, clinical conditions
associated with hyperglycemia were excluded, and multivariate models were used in the analysis. To minimize
selection bias, only complete laboratory records were included as study variables.
Statistical analysis
After collecting and compiling a database of the study population in Microsoft Excel, the data were
exported to IBM SPSS Statistics 27. The normality of the quantitative variables was assessed using the
Kolmogorov-Smirnov test. Since most variables did not follow a normal distribution, nonparametric tests
were used for inferential analysis.
Quantitative variables were reported as medians and interquartile ranges (IQRs), and qualitative variables
were reported as absolute frequencies and percentages. The Mann-Whitney U test was used to compare
continuous variables between the groups with and without Prediabetes. Subsequently, a binary logistic
regression analysis was performed to identify independent predictors of Prediabetes. Initially, all study variables
were included, excluding those with clinical or statistical collinearity with TGI (glucose and triglycerides) and
glycated hemoglobin (HbA1c) due to their diagnostic overlap with the outcome. Total cholesterol was omitted
due to overlap with LDL and HDL cholesterol fractions. A second model was evaluated, adjusting for body
and Snell, Nagelkerke). Model results are reported as odds ratios (OR) with 95 % confidence intervals.
The diagnostic accuracy of the TGI and other parameters was evaluated using receiver operating characteristic
(ROC) curves, and the area under the curve (AUC) was calculated. Optimal cutoff points were identified, and
sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated
for each criterion. In addition, combinations of variables (TGI, albumin, overweight/obesity) were analyzed
to determine if they improved the diagnostic performance of TGI alone. A p-value < 0.05 was considered
statistically significant.
Ethical considerations
This study received institutional authorization from the CCQANT-IESS for data collection and a
confidentiality agreement from the principal investigator. The protocol was evaluated by the Master's
Thesis Research Committee of the International University of La Rioja (UNIR) [2023_2643], which
issued a favorable opinion in May 2023. Data were obtained from anonymized clinical records without
requiring additional informed consent, as the retrospective design implies minimal risk. The research
was conducted in compliance with the principles of the Declaration of Helsinki, current Ecuadorian
legislation, and the Organic Law on the Protection of Personal Data, ensuring confidentiality and
responsible data handling.
RESULTS
A total of 663 patients were analyzed, comprising 221 (33.3 %) in the prediabetes case group and 442 (66.7 %)
in the control group. The patient population consisted of 54.8 % males and 45.2 % females. The glucose
tolerance index (TGI) distribution showed values close to normal (skewness of -0.080 and kurtosis of 0.534).
However, the Kolmogorov-Smirnov test indicated that all quantitative variables were non-normal, except for
age (p = 0.037), which justified the use of nonparametric tests for comparisons. The median age was 52 years
[IQR 47–57], with no significant differences between the two groups due to age- and sex-matching. Regarding
body mass index (BMI), the case group had higher values than the patients without Prediabetes (Table 1).
Regarding biochemical parameters, patients with Prediabetes had significantly higher fasting glucose,
HbA1c, triglycerides, total cholesterol, LDL, TGI, and AST levels than controls (p < 0.001 for all variables).
On the other hand, the prediabetes group showed significantly lower HDL (p = 0.03) and albumin (p < 0.001)
levels, whereas no statistically significant differences were observed in ALT levels (Table 1).
Table 1. Comparison of BMI and biochemical parameters between patients with and without Prediabetes.
A binary logistic regression analysis was performed to identify factors associated with a prediabetes diagnosis.
In the first model, the study variables were included, excluding blood glucose and triglycerides due to
collinearity with the glucose tolerance test (GTT), HbA1c due to collinearity with the dependent variable,
and total cholesterol due to the simultaneous inclusion of its HDL and LDL fractions. The model showed
of adequate fit (not shown in the table).
Subsequently, a second model was fitted incorporating the dichotomous variable BMI. This model showed
considering its sensitivity to the sample size, and its interpretation should be made in conjunction with other
In this second model, the TGI index was significantly associated with a diagnosis of Prediabetes (OR: 2.831;
95% CI: 1.937–4.137; p < 0.001), indicating that for every unit increase in the TGI, the odds of having
Prediabetes increased by 2.83. Significant associations were also observed with albumin (OR: 0.334 [95 %
CI: 0.196–0.568] p < 0.001), showing a protective effect, and with overweight/obesity status (OR: 3.307
[95% CI: 2.083–5.251] p < 0.001), which tripled the risk of Prediabetes. Female sex was also associated with
a lower risk (OR: 0.653 [95 % CI: 0.434–0.984] p = 0.042). The remaining variables, including age, LDL,
HDL, AST, and ALT, did not show statistically significant associations (Table 2).
Table 2. Multivariate association between clinical variables and the diagnosis of Prediabetes using binary
logistic regression.
Diagnostic accuracy of the triglyceride-glucose index
The diagnostic ability of the TGI to predict prediabetic status was evaluated using ROC curve analysis (Figure
1A).
(75.1%; 95 % CI: 69.0–80.4) and specificity (58.1 %; 95% CI: 53.5–62.7), a positive predictive value (PPV) of
0.47, and a negative predictive value (NPV) of 0.82 (Table 3). The area under the curve (AUC) was 0.691 (95 %)
CI: 0.65–0.73; p < 0.001), indicating moderate diagnostic accuracy.
Since albumin was one of the significant variables in the multivariate analysis, its diagnostic performance
was evaluated using an additional ROC curve (Figure 1B), finding an AUC of 0.635 (95 % CI: 0.59–0.68;
p <0.001) and an optimal cutoff point at <4.15 g/dL, with a sensitivity of 54.8 %, specificity of 62.7 %, PPV
of 0.42 and NPV of 0.73.
Subsequently, combinations of the TGI with other clinical variables were analyzed to assess whether
its diagnostic performance was improved. Combining the TGI with overweight or obesity (OO) increased
specificity to 71.0 % and maintained an acceptable sensitivity of 66.1 % (PPV: 0.53; NPV: 0.81).
increase in specificity to 86.7 %, although sensitivity decreased to 36.2 %. A second alternative combination
(Table 3).
Figure 1. ROC curves for the prediction of Prediabetes using A) the triglyceride-glucose index (TGI); B)
serum albumin.
Table 3. Diagnostic accuracy of the triglyceride-glucose index (TGI) alone and combined with albumin and
overweight/obesity for the detection of Prediabetes
DISCUSSION
type 2 diabetes mellitus (T2DM). These findings are consistent with previous studies by Zhang and Zeng in
a cross-sectional analysis of more than 25,000 US adults using NHANES data, which found a non-linear
relationship between TGI and the prevalence of Prediabetes and diabetes, observing a progressive increase
in risk starting from an TGI > 8.00 in men and > 9.00 in women.
(39)
This behavior suggests that the risk threshold
for TGI may vary according to population characteristics, justifying the need for local studies such as the
present one.
In a prospective cohort study in China,
(31)
reported that a one-standard-deviation increase in TGI was
associated with a 1.38-fold increased risk of Prediabetes. Furthermore, they found that the TGI had better
diagnostic performance than other non-insulin-based markers, such as the triglyceride/HDL ratio or obesity,
with an AUC of 0.60,
(31)
a value comparable to that observed in this study.
In this study, the specificity of the TGI (58.1 %) implies that a considerable proportion of individuals without
Prediabetes could be initially classified as at risk, resulting in false positives. In clinical practice, this does
not invalidate its usefulness, as these individuals can benefit from follow-up and preventive guidance.
as an initial screening tool. Its value lies in facilitating the early detection of individuals at risk of Prediabetes,
even at the cost of a proportion of false positives. In this sense, the TGI should not be considered a definitive
diagnostic marker, but rather a complement to other tests or clinical criteria, especially in primary care
settings or environments with limited resources, where access to more complex methods may be restricted.
A key finding of the study was the identification of a significant relationship between low albumin levels and
Prediabetes, even after multivariate adjustment. This finding may differ from other studies, which indicate
increased albumin levels in patients with insulin resistance
(39,40)
, even though elevated albumin is not explicitly
linked to the development of type 2 diabetes mellitus (T2DM).
(40)
This association could be explained by
variations in liver albumin production under conditions of insulin resistance due to hepatic stimulation.
(41)
When analyzing diagnostic combinations, it was observed that incorporating SO into the TGI criterion
increased specificity to 71.0 %. This improvement was even more pronounced when combining TGI, OO,
and albumin, achieving a specificity of 86.7 %, which coincides with that reported by Chen et al., who
demonstrated that a TGI greater than 8.88 significantly decreases the probability of regression to normoglycemia,
especially in patients with a high BMI.
(28)
In the multivariate analysis, the TGI maintained a significant association with the diagnosis of Prediabetes,
positioning it as an independent predictor. This finding is consistent with a preliminary study reporting that
TGI has diagnostic capacity comparable to HbA1c,
(42)
but with the advantage of being a more accessible
method in resource-limited settings.
Additionally, it has been shown that the TGI not only predicts the onset of Prediabetes but is also associated
with cardiovascular complications. Another study demonstrated that an elevated TGI is associated with a
higher risk of cardiovascular disease in individuals under 65 years of age with Prediabetes or diabetes,
(43)
reinforcing its effectiveness as a prognostic marker and not just a diagnostic one. These results demonstrate
the TGI's functionality as a screening tool in adult populations at metabolic risk. The non-linear relationship
with regression to normoglycemia observed in longitudinal studies
(28)
suggests the importance of low TGI
levels, even in the early stages of dysglycemia, which could prevent progression to overt diabetes.
Limitations
Despite efforts to control for bias, limitations inherent to the study design were identified, including potential
recording errors or underestimation of relevant, undocumented clinical variables —such as family history of
diabetes, physical activity level, dietary habits, and inflammatory markers—leading to uncontrolled
confounding. Furthermore, the multivariate model showed marginal fit in the statistical analysis, and a third
model proved unfeasible. This suggests that the regression results require further refinement and validation.
Another limitation is that the observed moderate specificity carries a risk of false positives, which limits its
use as a standalone diagnostic tool. Therefore, the identified cutoff point should be interpreted with caution,
as it may require initial adaptation across populations with varying genetic, epidemiological, or lifestyle
profiles. Multicenter, longitudinal studies are needed to confirm the external validity of these findings.
In addition, limitations were identified, including periods of unreported results due to a lack of reagents at
the institution, as well as the absence of screenings based on insulin measurements or oral glucose tolerance
tests.
However, the study provides evidence on the usefulness of the TGI as an accessible marker for detecting
Prediabetes.
CONCLUSIONS
The TGI showed moderate discriminative capacity to predict prediabetic status in nondiabetic adults, with a
Serum albumin < 4.15 g/dL was associated with a higher risk of Prediabetes. The combination of TGI with
tool for early detection of dysglycemia, especially in resource-limited settings where insulin- or HbA1c-ba-
sed testing is unavailable. Prospective validation of these results in other populations is recommended to
strengthen their clinical applicability.
Financing: This research was self-funded by the authors
Acknowledgments: The authors express their gratitude to the health institution for its logistical support in
carrying out this study.
Conflicts of interest: The authors declare that they have no conflicts of interest related to this study.
Contribution statement:
Author 1: study design, statistical analysis, and initial writing, general supervision, and funding.
Author 2: collection and validation of clinical data.
Author 3: Collection of laboratory data and support in statistical analysis.
Author 4: discussion, review, and formatting adjustments of the final manuscript.
BIBLIOGRAPHIC REFERENCES
1. Lahsen M. Rodolfo. Sindrome metabólico y diabetes. Revista Médica Clínica Las Condes. 2014;
[cited May 15, 2025]; 25(1): 47–52. Available at: https://www.sciencedirect.com/science/article/pii/
S0716864014700100?via%3Dihub DOI: https://doi.org/10.1016/S0716-8640(14)70010-0.
2. Civeira-Murillo F, Pérez-Ruiz MR, Baila-Rueda L. Síndrome metabólico: concepto, epidemiología,
etiopatogenia y complicaciones. Medicine. 2013; [cited May 15, 2025];11(40): 2402–2409. Available
at: https://www.sciencedirect.com/science/article/abs/pii/S0304541213706371?via%3Dihub DOI:
https://doi.org/10.1016/S0304-5412(13)70637-1.
3. Puchulu FM. Definition, Diagnosis and Classification of Diabetes Mellitus. In: Cohen Sabban EN,
Puchulu FM, Cusi K (eds) Dermatology and Diabetes. Cham: Springer International Publishing;
2018; [cited May 15, 2025]; p. 7–18. Available at: https://link.springer.com/chapter/10.1007/978-3-
319-72475-1_2 DOI: https://doi.org/10.1007/978-3-319-72475-1_2.
4. Organización Mundial de la Salud (OMS). Las diez causas principales de defunción. [Cited Apr 9,
2025] Available at: https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death.
5. Instituto Nacional de Estadística y Censos. Registro Estadístico de Defunciones Generales. Instituto
nacional de estadística y censos (INEC), 2024. Available at: https://www.ecuadorencifras.gob.ec/de
funciones-generales/.
6. Zavala Calahorrano AM, Fernández E. Diabetes mellitus tipo 2 en el Ecuador: revisión epidemiológica.
Mediciencias UTA. 2018; [cited May 18, 2025]; 2(4): 3–3. Available at: https://revistas.uta.edu.ec/e
revista/index.php/medi/article/view/1219 DOI: https://doi.org/10.31243/mdc.uta.v2i4.132.2018.
7. Flores JXD, Morán EEM, Gaytán ÁMM, Martinez JLT. La diabetes mellitus y diabetes gestacional,
en adolescente, en el mundo y en el Ecuador, manejo, prevención, tratamiento y mortalidad. Recimundo.
2023; [cited May 18, 2025]; 7(2): 33–48. Available at: https://dialnet.unirioja.es/servlet/articulo?codigo=
9006260 DOI: https://doi.org/10.26820/recimundo/7.(2).jun.2023.33-48.
8. García Matamoros WF. Sedentarismo en niños y adolescentes: Factor de riesgo en aumento. Recimundo.
2019; [cited May 20, 2025]; 3(1): 1602–1624. Available at: https://www.recimundo.com/index.php/
es/article/view/449 DOI: https://doi.org/10.26820/recimundo/3.(1).enero.2019.1602-1624.
9. Vasconcellos MB, Matta IEA da MP, Santana DD, Veiga GV da. Mudanças na obesidade, comportamento
sedentário e inatividade física, entre 2010 e 2017, em adolescentes. Journal of Physical Education.
2021; [cited May 20, 2025]; 32(1): e-3280. Available at: https://www.scielo.br/j/jpe/a/B89Pgbwcc98
z4cQ3WvTNSvt/?format=html&lang=pt DOI: https://doi.org/10.4025/jphyseduc.v32i1.3280.
10. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among
adolescents: a pooled analysis of 298 population-based surveys with 1·6 million participants. The
Lancet Child & Adolescent Health. 2020; [cited May 23, 2025]; 4(1): 23–35. Available at:
https://www.thelancet.com/journals/lanchi/article/PIIS2352-4642(19)30323-2/fulltext DOI:
https://doi.org/ 10.1016/S2352-4642(19)30323-2.
11. Galicia-Garcia U, Benito-Vicente A, Jebari S, Larrea-Sebal A, Siddiqi H, Uribe KB, et al. Pathophysiology
of Type 2 Diabetes Mellitus. International Journal of Molecular Sciences. 2020; [cited May 23, 2024];
21(17): 6275. Available at: https://www.proquest.com/docview/2440180571?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.3390/ijms21176275.
12.
Metabolism Journal. 2021; [cited May 26, 2025]; 46(1): 15–37. Available at: https://synapse.koreamed.
org/articles/1160546 DOI: https://doi.org/10.4093/dmj.2021.0280.
13. American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of
Diabetes: Standards of Care in Diabetes—2024. Diabetes Care. 2023; [cited May 30, 2025];
47(Supplement_1): S20–S42. Available at: https://diabetesjournals.org/care/article/47/Supplement_
1/S20/153954/2-Diagnosis-and-Classification-of-Diabetes DOI: https://doi.org/10.2337/dc24-S002.
14. Pan American Health Organization (PAHO/WHO). Prevalence of diabetes and diabetes treatment
coverage. Available at: https://www.paho.org/en/enlace/prevalence-diabetes-and-diabetes-treatment-
coverage.
15. Zhou B, Rayner AW, Gregg EW, Sheffer KE, Carrillo-Larco RM, Bennett JE, et al. Worldwide trends
in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-
representative studies with 141 million participants. The Lancet. 2024; [cited May 30, 2025]; 404(10467):
2077–2093. Available at: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)02317-1/
fulltext? DOI: https://doi.org/10.1016/S0140-6736(24)02317-1.
16. Park PH, Pastakia SD. Access to Hemoglobin A1c in Rural Africa: A Difficult Reality with Severe
Consequences. Journal of Diabetes Research. 2018; [cited Jun 06, 2025]; 2018(1): 6093595. Available
at: https://onlinelibrary.wiley.com/doi/full/10.1155/2018/6093595 DOI: https://doi.org/10.1155/2018/
6093595.
17. Unger G, Benozzi SF, Perruzza F, Pennacchiotti GL. Índice triglicéridos y glucosa: Un indicador útil
de insulinorresistencia. Endocrinologia y Nutricion. 2014; [cited Jun 06, 2025]; 61(10): 533–540.
Available at: https://www.sciencedirect.com/science/article/abs/pii/S2173509314001846 DOI:
https://doi.org/10.1016/j.endonu.2014.06.009.
18. Navarro-González D. El Índice Trigliceridos-Glucosa como predictor de Diabetes tipo 2 y su relación
con el Estado Metabólico y la Obesidad. [Tesis doctoral] [Pamplona, España]: Universidad de Navarra;
2016. Available at: https://dialnet.unirioja.es/servlet/tesis?codigo=246480.
19. Campos Muñiz C, León-García PE, Serrato Diaz A, Hernández-Pérez E. Diabetes mellitus prediction
based on the triglyceride and glucose index. Medicina Clinica. 2022; [cited Jun 06, 2025]; 160(6):
231–236. Available at: https://www.sciencedirect.com/science/article/abs/pii/S2387020623000761
DOI: https://doi.org/10.1016/j.medcli.2022.07.003.
20. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala
MG, Hernández-González SO, et al. The Product of Triglycerides and Glucose, a Simple Measure of
Insulin Sensitivity. Comparison with the Euglycemic-Hyperinsulinemic Clamp. The Journal of Clinical
Endocrinology & Metabolism. 2010; [cited Jun 06, 2025]; 95(7): 3347–3351. Available at:
https://academic.oup.com/jcem/article-abstract/95/7/3347/2596446 DOI: https://doi.org/10.1210/jc.
2010-0288.
21. Calcaterra V, Montalbano C, De Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, et al. Triglyceride
Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population.
Journal of clinical research in pediatric endocrinology. 2022; [cited July 06, 2025]; 0–0. Available at:
https://europepmc.org/article/med/31088046 DOI: https://doi.org/10.4274/jcrpe.galenos.2019.2019.0024.
22. Babic N, Valjevac A, Zaciragic A, Avdagic N, Zukic S, Hasic S. The Triglyceride/HDL Ratio and
Triglyceride Glucose Index as Predictors of Glycemic Control in Patients with Diabetes Mellitus
Type 2. Medical archives (Sarajevo, Bosnia and Herzegovina). 2019; [cited July 06, 2025]; 73(3):
163–168. Available at: https://www.proquest.com/docview/2287390456?pq-origsite=gscholar&
fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/10.5455/medarh.2019.
73.163-168.
23. Muhammad IF, Bao X, Nilsson PM, Zaigham S. Triglyceride-glucose (TyG) index is a predictor of
arterial stiffness, incidence of diabetes, cardiovascular disease, and all-cause and cardiovascular
mortality: A longitudinal two-cohort analysis. Frontiers in Cardiovascular Medicine. 2023; [cited
July 06, 2025]; 9. Available at: https://doi.org/10.3389/fcvm.2022.1035105.
24.
Insulin Resistance and Predicting Diabetic Kidney Disease. Medical Science Monitor. 2023; [cited
July 06, 2025]; 29: 0–0. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10337482/ DOI:
https://doi.org/10.12659/MSM.939482.
25. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off Values of Triglycerides - Glucose
Index for Metabolic Syndrome Associated with Type 2 Diabetes Mellitus. Baghdad Science Journal.
2022; [cited July 08, 2025];19(2): 0340–0340. Available at: https://bsj.uobaghdad.edu.iq/home/vol19/
iss2/7/ DOI: https://doi.org/10.21123/bsj.2022.19.2.0340.
26.
biomarker for predicting type 2 diabetes mellitus in children and adolescents. Endocrine Journal.
2022; [cited July 08, 2025]; 69(5): 559–565. Available at: https://www.jstage.jst.go.jp/article/endocrj/
69/5/69_EJ21-0560/_html/-char/en DOI: https://doi.org/10.1507/endocrj.EJ21-0560.
27. Sánchez García A, Rodríguez Gutiérrez R, Mancillas Adame L, González Nava V, Díaz González
Colmenero A, Solis RC, et al. Diagnostic Accuracy of the Triglyceride and Glucose Index for Insulin
Resistance: A Systematic Review. International Journal of Endocrinology. 2020; [cited July 08, 2025];
2020. Available at: https://onlinelibrary.wiley.com/doi/full/10.1155/2020/4678526 DOI: https://doi.org/
10.1155/2020/4678526.
28. Chen X, Liu D, He W, Hu H, Wang W. Predictive performance of triglyceride glucose index (TyG
index) to identify glucose status conversion: a 5-year longitudinal cohort study in Chinese pre-diabetes
people. Journal of Translational Medicine. 2023; [cited July 08, 2025]; 21(1): 624. Available at:
https://www.proquest.com/docview/2865458165?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12967-023-04402-1.
29.
index combined with body mass index in predicting recovery from prediabetic state to normal fasting
glucose: a cohort analysis based on a Chinese physical examination population. Lipids in Health and
Disease. 2024; [cited July 08, 2025]; 23(1): 71. Available at: https://link.springer.com/article/10.1186/
s12944-024-02060-w DOI: https://doi.org/10.1186/s12944-024-02060-w.
REE 20(1) Riobamba ene. - abr. 2026
cc
BY NC ND
36
ISSN-impreso 1390-7581
ISSN-digital 2661-6742
30. Suleiman RR, Salih SF, Abdullah BI, Ibrahim IH, Saeed ZA. Triglyceride Glucose Index, its Modified
Indices, and Triglyceride HDL-C Ratio as Predictor Markers of Insulin Resistance in Prediabetic
Individuals. Medical Journal of Babylon. 2023; [cited July 08, 2025]; 20(2): 268. Available at:
https://journals.lww.com/mjby/fulltext/2023/20020/triglyceride_glucose_index,_its_modified_
indices,.10.aspx DOI: https://doi.org/10.4103/MJBL.MJBL_269_22.
31.
predicts incidence of Prediabetes: a prospective cohort study in China. Lipids in Health and Disease.
2020; [cited July 08, 2025]; 19: 226. Available at: https://www.proquest.com/docview/2451933152?
pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI: https://doi.org/
10.1186/s12944-020-01401-9.
32.
glucose index and prediabetes risk among Chinese adults: a secondary retrospective cohort study.
European Journal of Medical Research. 2024; [cited July 08, 2025]; 29(1): 529. Available at:
https://link.springer.com/content/pdf/10.1186/s40001-024-02121-x.pdf DOI: https://doi.org/10.1186/
s40001-024-02121-x.
33.
general population aged 45 years and older: a national prospective cohort study. BMC Endocrine
Disorders. 2025; [cited July 08, 2025]; 25(1): 25. Available at: https://www.proquest.com/docview/
3165519755?pq-origsite=gscholar&fromopenview=true&sourcetype=Scholarly%20Journals DOI:
https://doi.org/10.1186/s12902-025-01848-w.
34. Rothberg A, Lean M, Laferrère B. Remission of type 2 diabetes: always more questions, but enough
answers for action. Diabetologia. 2024; [cited July 08, 2025]; 67(4): 602–610. Available at:
https://www.proquest.com/docview/2933267153?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1007/s00125-023-06069-1.
35. Lima LMTR. Insulin resistance underlying type 2 diabetes. The Lancet Diabetes & Endocrinology.
2019; [cited July 08, 2025]; 7(6): 424. Available at: https://www.thelancet.com/journals/landia/article/
PIIS2213-8587(19)30147-0/fulltext DOI: https://doi.org/10.1016/S2213-8587(19)30147-0.
36. Zhang J, Zhang Z, Zhang K, Ge X, Sun R, Zhai X. Early detection of type 2 diabetes risk: limitations
of current diagnostic criteria. Frontiers in Endocrinology. 2023; [cited July 08, 2025]; 14. Available
at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2023.1260623/full
DOI: https://doi.org/10.3389/fendo.2023.1260623.
37. Katki HA, Berndt SI, Machiela MJ, Stewart DR, Garcia-Closas M, Kim J, et al. Increase in power by
obtaining 10 or more controls per case when type-1 error is small in large-scale association studies.
BMC Medical Research Methodology. 2023; [cited July 08, 2025]; 23(1): 153. Available at:
https://www.proquest.com/docview/2838780104?pq-origsite=gscholar&fromopenview=true&
sourcetype=Scholarly%20Journals DOI: https://doi.org/10.1186/s12874-023-01973-x.
38. Crews, L. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary123. The American Journal of Clinical Nutrition. 1998; [cited
July 15, 2025]; 68(4): 899–917. [cited July 08, 2025]; Available at: https://publishingimages.s3.
amazonaws.com/eZineImages/PracticePerfect/804/Clinical-guidelines-on-the-identification.pdf
DOI: https://doi.org/10.1093/ajcn/68.4.899.
39. Zhang L, Zeng L. Non-linear association of triglyceride-glucose index with prevalence of prediabetes
and diabetes: a cross-sectional study. Frontiers in Endocrinology. 2023; [cited July 15, 2025]; 14:
1295641. Available at: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.
2023.1295641/full DOI: https://doi.org/10.3389/fendo.2023.1295641.
40.
Insulin Resistance, and Incident Diabetes in Nondiabetic Subjects. Endocrinology and Metabolism.
2013; [cited July 15, 2025]; 28(1): 26–32. Available at: https://e-enm.org/journal/view.php?doi=
10.3803/EnM.2013.28.1.26 DOI: https://doi.org/10.3803/EnM.2013.28.1.26.
41. Kim S, Kang S. Serum Albumin Levels: A Simple Answer to a Complex Problem? Are We on the
Right Track of Assessing Metabolic Syndrome? Endocrinology and Metabolism. 2013; [cited July
15, 2025]; 28(1): 17–19. Available at: https://e-enm.org/journal/view.php?doi=10.3803/EnM.2013.
28.1.17 DOI: https://doi.org/10.3803/EnM.2013.28.1.17.
42. Darshan An V, Rajput R, Meena, Mohini, Garg R, Saini S. Comparison of triglyceride glucose index
and HbA1C as a marker of prediabetes – A preliminary study. Diabetes & Metabolic Syndrome:
Clinical Research & Reviews. 2022; [cited July 15, 2025]; 16(9): 102605. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S1871402122002193 DOI: https://doi.org/10.1016/
j.dsx.2022.102605.
43.
diabetes mellitus: a systemic review and meta-analysis. 2021; [cited July 15, 2025]; Available at:
https://ec.bioscientifica.com/view/journals/ec/10/11/EC-21-0234.xml DOI: https://doi.org/10.1530/
EC-21-0234
