ORIGINAL ARTICLE

Introduction

Diabetes mellitus (DM) is a chronic metabolic disorder characterized by persistent hyperglycemia which, in the case of type 2 diabetes mellitus (T2DM), is due to peripheral resistance to insulin actions <sup>1 ➤
1. Solis-Herrera C, Triplitt C, Cersosimo E, DeFronzo RA. Pathogenesis of Type 2 Diabetes Mellitus. MDText.com; 2021 [citado el 9 de julio de 2024]. Disponible en: https://www.ncbi.nlm.nih.gov/books/NBK279115/</sup> . It affects approximately 422 million people worldwide (62 million in the Americas), and this number is expected to increase by 200 million by the year 2040 <sup>2 ➤
2. Medina-Chávez JH, Vázquez-Parrodi M, Mendoza-Martínez P, Ríos-Mejía ED, de Anda-Garay JC, Balandrán-Duarte DA. Protocolo de Atención Integral: prevención, diagnóstico y tratamiento de diabetes mellitus 2. Rev Med Inst Mex Seguro Soc . 2022;60(Suppl 1):S4.</sup> . In Peru, it affects approximately four individuals per 100,000 inhabitants <sup>3 ➤
3. Carrillo-Larco RM, Bernabé-Ortiz A. Diabetes mellitus tipo 2 en Perú: una revisión sistemática sobre la prevalencia e incidencia en población general. Rev Peru Med Exp Salud Publica. 2019;36(1):26. Disponible en: 10.17843/rpmesp.2019.361.4027</sup> .

Systemic organ damage is attributed to the synergy between chronic hyperglycemia and metabolic abnormalities, primarily affecting the microvasculature in early stages, and later the macrovasculature, which increases the risk of cardiovascular disease by two to four times <sup>4 ➤
4. Faselis C, Katsimardou A, Imprialos K, Deligkaris P, Kallistratos M, Dimitriadis K. Microvascular complications of type 2 diabetes mellitus. Curr Vasc Pharmacol. 2020;18(2):117–24. doi: 10.2174/1570161117666190502103733</sup> . T2DM is a disease of multifactorial etiology, with environmental and lifestyle components manifesting in anthropometric characteristics detrimental to health, such as overweight or obesity, high blood pressure, and metabolic dysfunction <sup>5 ➤
5. Rydén L, Ferrannini G, Mellbin L. Risk factor reduction in type 2 diabetes demands a multifactorial approach. Eur J Prev Cardiol. 2019;26(2_suppl):81–91. doi: 10.1177/2047487319872015</sup> .Regarding overweight or obesity, the predominance of abdominal fat is a strong predictor of T2DM, as it promotes insulin resistance through complex metabolic mechanisms, such as the action of dipeptidyl peptidase-4 secreted by hepatocytes, which promotes inflammation of abdominal adipose tissue <sup>6 ➤
6. 6. Autor/es. Función de la DPP-4 y su rol en insulinoresistencia. Publicación. Año. doi o URL.</sup> . Moreover, excess abdominal fat acts as an endocrine organ, releasing inflammatory cytokines such as tumor necrosis factor alpha, reduced adiponectin, and non-esterified fatty acids, among others <sup>7 ➤
7. Korac A, Srdic-Galic B, Stancic A, Otasevic V, Korac B, Jankovic A. Adipokine signatures of subcutaneous and visceral abdominal fat in normal-weight and obese women with different metabolic profiles. Arch Med Sci. 2021;17(2):323–36. doi: 10.5114/aoms/92118</sup> . The association between abdominal obesity and T2DM has been widely established in international studies and meta-analyses, which suggest that reducing waist circumference lowers the risk of developing T2DM <sup>8 ➤
8. Freemantle N, Holmes J, Hockey A, Kumar S. How strong is the association between abdominal obesity and the incidence of type 2 diabetes?: abdominal obesity and type 2 diabetes. Int J Clin Pract. 2008;62(9):1391–6. doi: 10.1111/j.1742-1241.2008.01805.x</sup> . However, international study guidelines to assess this risk are primarily based on body mass index (BMI), and waist circumference is measured only in cases of overweight or obesity. This is a limitation, as BMI does not account for the distribution of body fat <sup>9 ➤
9. Goossens GH. The metabolic phenotype in obesity: fat mass, body fat distribution, and adipose tissue function. Obes Facts. 2017;10(3):207–15. doi: 10.1159/000471488</sup> . Furthermore, guidelines and recommendations on waist circumference for assessing T2DM risk are based on studies conducted in countries with different anthropometric characteristics than the Peruvian population. Therefore, the aim of this study was to determine a reference cutoff point for waist circumference associated with the risk of T2DM in Peruvian adults. The results will provide a screening tool with potential for improvement and application in rapid and large-scale assessments for estimating the risk of this endocrinopathy.

METHODS

Study Design and Area
An analytical cross-sectional study was conducted using data from the 2022 Demographic and Family Health Survey (ENDES-22, by its Spanish acronym). ENDES is a population-based survey with a complex, probabilistic, two-stage, and independent sampling design <sup>10 ➤
10. Martina Chávez M, Amemiya Hoshi I, Suguimoto Watanabe SP, Arroyo Aguilar RS, Zeladita-Huaman JA, Castillo Parra H. Depresión en adultos mayores en el Perú: distribución geoespacial y factores asociados según ENDES 2018 - 2020. An Fac Med. 2022;83(3):180–7. doi: 10.15381/anales.v83i3.23375</sup> . The analyzed population consisted of adults aged 18 years or older who had complete measurements of weight, height, waist circumference, and blood pressure.

Population and Sample

The sample included the entire target population (N=28,407) with complete data within the 2022 ENDES database, derived from a total of 34,301 individuals (Table 1). No additional sampling procedures were conducted. Individuals with a prior diagnosis of T2DM were excluded from the analysis, as this condition causes changes in body weight <sup>11 ➤
11. Genuth SM, Palmer JP, Nathan DM. Classification and diagnosis of diabetes. 2018.</sup> and leads to macrovascular and microvascular complications that affect systemic blood pressure <sup>12 ➤
12. Jia G, Sowers JR. Hypertension in diabetes: an update of basic mechanisms and clinical disease. Hypertension. 2021;78(5):1197–205. doi: 10.1161/HYPERTENSIONAHA.121.17981</sup> . For external validation of the waist circumference cutoff points determined in ENDES-2022, the adult population from ENDES-2023 (N=30,782) was used. The selected population is shown in (Figure 1).

Figure 1

Selected and excluded population for the study.

Variables and Instruments

The variables included sex (female/male) and waist circumference (WC), considered a cardiovascular risk indicator. The WC scale varied by sex: in men, it was defined as normal if less than 95 cm, elevated risk between 95 and 101 cm, and high risk if greater than or equal to 102 cm. In women, it was considered normal if less than 82 cm, elevated risk between 82 and 87 cm, and high risk if greater than or equal to 88 cm <sup>13 ➤
13. 13. Brahim Q, Ahsan M. Measurement of visceral fat, abdominal circumference and waist-hip ratio to predict health risk in males and females. Pak J Biol Sci. 2019;22(4):168–73. doi: 10.3923/pjbs.2019.168.173</sup> . For the analysis, WC was dichotomized into two categories: normal and elevated (from 95 cm). According to the ENDES-2022 technical specifications, the measurement was taken with a tape measure placed between the costal margin and the iliac crest, perpendicular to the body’s longitudinal axis, with the person standing <sup>14 ➤
14. 14. Vega Abascal JB, Leyva Sicilia Y, Teruel Ginés R. La circunferencia abdominal. Su inestimable valor en la Atención Primaria de Salud. CCH Correo Cient Holguín. 2019;23(1):270–4.</sup> .

T2DM risk was determined using the diabetes risk questionnaire developed by Bang et al., validated by the American Diabetes Association (ADA) <sup>15 ➤
15. 15. Bang H. Development and validation of a patient self-assessment score for diabetes risk. Ann Intern Med. 2009;151(11):775. doi: 10.7326/0003-4819-151-11-200912010-00005</sup> . This instrument was chosen instead of the FINDRISC test because the available items in the database were more compatible with the Bang et al. test, as it does not include the frequency of fruit and vegetable consumption, an item missing in ENDES. The applied formulas were as follows <sup>16 ➤
16. Calculators:Diabetes Risk Self-Assessment [Internet]. MSD Manual Professional Edition. [citado el 9 de julio de 2024]. Disponible en: https://www.msdmanuals.com/professional/multimedia/clinical-calculator/diabetes-risk-self-assessment</sup> :

BMI SCORE = (weight / 2.205) / (height / 39.37)²
Total diabetes risk score = BMI score + age + sex + family history of DM + physical activity + hypertension (HTN)

Mean arterial pressure (MAP), calculated from the systolic blood pressure (SBP) and diastolic blood pressure (DBP) values obtained from the second measurement, was used as a reference to define the presence of HTN, following the formula <sup>17 ➤
17. Calculators:Mean Vascular Pressure (systemic or pulmonary) [Internet]. MSD Manual Professional Edition. [citado el 9 de julio de 2024]. Disponible en: https://www.msdmanuals.com/professional/multimedia/clinical-calculator/mean-vascular-pressure-systemic-or-pulmonary</sup> :

MAP = [(2 × DBP) + SBP] / 3.

The use of MAP was justified because it more adequately represents tissue perfusion throughout the cardiac cycle <sup>18 ➤
18. DeMers D, Wachs D. Physiology, Mean Arterial Pressure. En: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2025 [citado el 9 de julio de 2024]. Disponible en: http://www.ncbi.nlm.nih.gov/books/NBK538226/</sup> . The total score was constructed by assigning values to the variables: age (<40 years = 0; 40–49 = 1; 50–59 = 2; ≥60 = 3), sex (female = 0; male = 1), physical activity (subtracts 1 point), hypertension (adds 1 point), family history of DM (adds 1 point), and BMI score (continuous value). T2DM risk was dichotomized: scores ≥4 indicated the presence of risk; scores <4 indicated the absence of risk.

Procedures

Data were grouped according to defined categories, and the proposed WC cutoff points were applied to the ENDES-2022 population. These cutoff points were then validated on the ENDES-2023 population to assess their diagnostic accuracy for estimating T2DM risk according to the Bang et al. questionnaire <sup>16 ➤
16. Calculators:Diabetes Risk Self-Assessment [Internet]. MSD Manual Professional Edition. [citado el 9 de julio de 2024]. Disponible en: https://www.msdmanuals.com/professional/multimedia/clinical-calculator/diabetes-risk-self-assessment</sup> . Additionally, significant differences between the findings in both databases were evaluated.

Statistical Analysis

Descriptive analysis was conducted using frequencies and percentages. To evaluate the strength of association between WC and T2DM risk, Cramer's V coefficient was used. Cohen's Kappa test was employed to estimate the agreement between WC categories (normal/elevated) and T2DM risk (present/absent). Prevalence ratios (PRs) with their respective 95% confidence intervals (95% CI) were calculated to determine the risk magnitude.

A receiver operating characteristic (ROC) curve analysis was applied to establish the WC cutoff point with the highest sensitivity and specificity. The area under the curve (AUC) was interpreted as follows: 0.6–0.7 (low diagnostic ability), 0.7–0.8 (acceptable), 0.8–0.9 (excellent), and >0.9 (outstanding) <sup>19 ➤
19. Corbacioglu Ş, Aksel G. Receiver operating characteristic curve analysis in diagnostic accuracy studies: a guide to interpreting the area under the curve value. Turk J Emerg Med. 2023;23(4):195. doi: 10.4103/tjem.tjem_182_23</sup> .

Sensitivity, specificity, and positive and negative predictive values were calculated using an online tool (http://araw.mede.uic.edu/cgi-bin/testcalc.pl). Results with a p-value <0.05 were considered statistically significant.

RESULTADOS

In women, the AUC was 0.688 (95% CI: 0.678–0.699; p-value < 0.001), with a waist circumference cutoff point associated with the risk of T2DM set at 91 cm (Figure 2A). In men, the AUC was 0.821 (95% CI: 0.814–0.828; p-value < 0.001), with a waist circumference cutoff point for the risk of T2DM set at 88 cm (Figure 2B).

Figure 2

ROC Curve for waist circumference cutoff point adjusted to T2DM risk according to ENDES 2022.

A bivariate comparison of frequencies was conducted between adults with waist circumference and T2DM risk according to the Bang et al. questionnaire <sup>16 ➤
16. 16. Calculators:Diabetes Risk Self-Assessment [Internet]. MSD Manual Professional Edition. [citado el 9 de julio de 2024]. Disponible en: https://www.msdmanuals.com/professional/multimedia/clinical-calculator/diabetes-risk-self-assessment</sup> . In men, on the left side of the table, using the international cutoff point of 95 cm, 63.3% of adults at risk of T2DM had elevated waist circumference. On the right side, after applying the ROC curve analysis and establishing a new cutoff point of 88 cm, the frequency of T2DM risk was 82.3%. In women, with the international cutoff point of 82 cm, the frequency was 87.5%, while with the new cutoff point of 91 cm, it was 71.5% (Table 2).

In men, both in the group evaluated with the international cutoff of 95 cm and in the group where the new cutoff of 88 cm determined by the ROC curve was applied, the association strength was relatively strong (V=0.519 and V=0.489, respectively), with moderate agreement (K=0.507; K=0.496). Sensitivity was 64% and 82%; specificity was 87% and 61%; positive predictive value (PPV) was 82% and 67%; and negative predictive value (NPV) was 72% and 78%. The prevalence ratio (PR) was 2.88 (95% CI: 2.77–2.99) and 3.07 (95% CI: 2.90–3.24), respectively. In women, with a cutoff based on international parameters (82 cm) and with the new cutoff of 91 cm, the association strength was V=0.115 and V=0.240; agreement was K=0.060 and K=0.203. Sensitivity was 88% and 71%; specificity was 23% and 57%; PPV was 25% and 33%; and NPV was 86% and 87%. The PR was 1.87 (95% CI: 1.71–2.05) and 2.56 (95% CI: 2.40–2.73), respectively (Table 3).

The cutoff points established in ENDES-2022 were evaluated by applying them to a different population corresponding to ENDES-2023. A bivariate table was used to compare the frequency of adults with elevated waist circumference and T2DM risk according to the Bang et al. <sup>16 ➤
16. Calculators:Diabetes Risk Self-Assessment [Internet]. MSD Manual Professional Edition. [citado el 9 de julio de 2024]. Disponible en: https://www.msdmanuals.com/professional/multimedia/clinical-calculator/diabetes-risk-self-assessment</sup> . questionnaire. In men, on the left side of the table, using the international cutoff point of 95 cm, 57.90% of adults at risk of T2DM had elevated waist circumference, while on the right side, with the new cutoff point of 88 cm, the frequency was 79.60%. In women, with the international cutoff point of 82 cm, the frequency was 85.90%, while with the new cutoff point of 91 cm, it was 67.80% (Table 4).

When applying the cutoff points to the ENDES-2023 population, it was observed that in men, both with the international cutoff (95 cm) and the new cutoff derived from the ROC curve (88 cm), the association strength was relatively strong (V=0.457 and V=0.424), with moderate agreement (K=0.444; K=0.403). Sensitivity was 60% and 81%; specificity was 86% and 61%; PPV was 77% and 64%; and NPV was 72% and 78%. PR was 2.62 (95% CI: 2.52–2.71) and 2.89 (95% CI: 2.74–3.01), respectively. In women, with a cutoff based on international parameters (82 cm) and with the new cutoff of 91 cm, the association strength was V=0.121 and V=0.206; agreement was K=0.066 and K=0.173.

Sensitivity was 86% and 68%; specificity was 27% and 57%; PPV was 24% and 31%; and NPV was 87% and 86%. PR was 1.90 (95% CI: 1.75–2.07) and 2.24 (95% CI: 2.10–2.38), respectively (Table 5).

DISCUSSION

It was observed that the waist circumference cutoff points after the ROC curve had greater sensitivity in detecting adults at risk of T2DM compared to the use of reference points based on international studies. This aligns with the anthropometric characteristics of the Peruvian population, which has a tendency toward overweight and obesity, primarily in men of all ages, and in women except for older adults, being strongly related to socioeconomic status <sup>20 ➤
20. Tarqui-Mamani C, Alvarez-Dongo D, Espinoza-Oriundo PL, Sanchez-Abanto JR. Análisis de la tendencia del sobrepeso y obesidad en la población peruana. Rev Esp Nutr Humana Diet. 2017;21(2):137–47. doi: 10.14306/renhyd.21.2.312</sup> , Additionally, the average heights of 1.66 in men and 1.54 in women are influenced by genetic and nutritional factors from childhood <sup>21 ➤
21. 21. Tarqui-Mamani CB, Alvarez-Dongo D, Espinoza-Oriundo PL. Análisis de la tendencia de la talla en niños y adolescentes peruanos; 2007 - 2013. Rev Esp Nutr Humana Diet. 2018;22(1):64–71. doi: 10.14306/renhyd.22.1.452</sup> . Therefore, it is suggested that anthropometric parameters should not be based on international ranges but rather adjusted to the characteristics of the Peruvian population, given the biological and sociocultural heterogeneity that may influence the underestimation of populations at cardio-metabolic risk <sup>24 ➤
24. 24. Rodríguez-Carrillo P, Aguirre-Tostado PI, Macías-Cervantes M, et al. Novel adiposity and biochemical–anthropometric indices to identify cardiometabolic risk and metabolic syndrome in Mexican adults. Healthcare. 2021;9(11):1561. doi: 10.3390/healthcare9111561</sup> .

The results of our study, which show higher sensitivity in the cutoff points based on local ROC curves to detect the risk of T2DM in Peruvian adults, align with recent research highlighting the importance of adjusting these cutoff points to the characteristics of each population. Bello-Chavolla et al. <sup>25 ➤
25. 25. Bello-Chavolla S, Antonio-Villa NE, et al. Metabolic score for visceral fat (METS-VF): a novel estimator of visceral fat area in relation to cardio-metabolic risk. PLoS One. 2020;15(10):e0241841. doi: 10.1371/journal.pone.0241841</sup> and Lopez-Lopez et al. <sup>26 ➤
26. 26. Lopez-Lopez JA, Gonzalez F, et al. Waist circumference cutoff points to identify major cardiovascular risk events in Colombian population. PLoS One. 2023. doi: 10.1371/journal.pone.0241841</sup> noted that the use of local indices, such as waist circumference, offers superior performance compared to international criteria, which may be less sensitive in specific contexts. This approach emphasizes the need to adapt anthropometric parameters to the unique characteristics of the Peruvian population to improve diagnostic accuracy for cardio-metabolic risk.

However, the adjustment of the cutoff point through the ROC curve resulted in favorable changes for higher sensitivity only in men, decreasing in women, where the association between WC and T2DM risk was very low and poorly concordant before and after the ROC curve, while in men, the association and concordance were moderate in both cases. This can be attributed to the fact that body fat distribution differs by sex, where white adipose tissue (lipid storage) in men is predominantly abdominal (visceral fat) due to androgens promoting fat accumulation in the splenic and visceral region, whereas, in women, white adipose tissue is mainly stored in the hips and thighs (subcutaneous fat) <sup>22 ➤
22. Esteve Ràfols M. Tejido adiposo: heterogeneidad celular y diversidad funcional. Endocrinol Nutr. 2014;61(2):100–12. doi: 10.1016/j.endonu.2013.03.011</sup> , influenced by estrogens, to which they are more exposed and are more sensitive in alpha and beta receptors at the nuclear level, affecting adipose tissue signaling, storage, and distribution <sup>23 ➤
23. Lizcano F. Roles of estrogens, estrogen-like compounds, and endocrine disruptors in adipocytes. Front Endocrinol (Lausanne). 2022;13:921504. doi: 10.3389/fendo.2022.921504</sup> . PTherefore, using waist circumference measurement to determine the risk of T2DM, while feasible for both sexes, performs better as a screening tool in men, and may be used in women but considering the research of other anthropometric parameters such as the waist-to-hip ratio in women.

The use of the cutoff point of 88 cm in men and 91 cm in women was employed in another population group (ENDES-2023), observing similar results, so the use of these new reference points has remained consistent in the Peruvian population over both years, confirming that the 91 cm cutoff in women is not suitable as it loses sensitivity and specificity, so it is recommended that for T2DM risk screening, the new cutoff of 88 cm be used in men and the previous reference value of 82 cm in women as a screening method for T2DM risk at the national level in Peru.

The limitations of the study included the possible bias in waist circumference measurement since measurements were taken from over 30,000 adults, and before processing and analyzing the data, inconsistent and poorly recorded data, such as 999 cm heights or 1 kg weights, were removed. Another limitation was the fact that basal glucose or hemoglobin A1c measurements were not used, which is logistically and economically unfeasible for the territorial extension covered by ENDES. However, the quantification of diabetes risk through the Bang et al. <sup>16 ➤
16. Calculators:Diabetes Risk Self-Assessment [Internet]. MSD Manual Professional Edition. [citado el 9 de julio de 2024]. Disponible en: https://www.msdmanuals.com/professional/multimedia/clinical-calculator/diabetes-risk-self-assessment</sup> questionnaire provides an opportunity to measure T2DM risk based on more easily accessible measurement characteristics, being a highly reproducible tool in different populations. Additionally, the new cutoff points are not fixed; they need to be adjusted in future studies, which could also represent an opportunity to develop anthropometric measurement tools that are updatable based on the health, socio-economic, and demographic development of the population. This approach will allow for a more accurate assessment of risk, contributing to more effective public health strategies adapted to the changing needs of the population.

CONCLUSION

he waist circumference reference point for T2DM risk in men was 88 cm, and in women, it was 82 cm. Waist circumference measurement for screening is an effective tool for men, but less efficient for women, who may require instruments based on measurements related to the gynoid fat distribution. The advantages of waist circumference measurement, which only requires a tape measure, compared to BMI formulas or risk questionnaires, make it a useful tool for T2DM risk screening in large epidemiological studies, with the possibility of individualized clinical use in men, while being less recommended for individualized studies in women.