ABSTRACT
The aim of this study was to evaluate the association between indices of adiposity and 20‐year cumulative incidence of type 2 diabetes (T2D) among apparently healthy adults participating in the ATTICA cohort study (2002–2022). The present analysis included data from 2000 individuals free of atherosclerotic cardiovascular disease (ASCVD) and T2D at baseline (age 43 ± 13 years; 51% women). Sociodemographic, anthropometric, lifestyle, clinical, and biochemical parameters were collected at baseline and follow‐up examinations. Obesity‐related anthropometric indices, in terms of total and central adiposity, along with dysfunctional adipose tissue‐related indices, were measured. The 20‐year cumulative incidence of T2D was 26.3% (95% CI [24.4%, 28.3%]). All indices were independently associated with the 20‐year incidence of T2D, with an overall predictive accuracy of approximately 75%. With the exception of LAP (lipid accumulation product), all indices reflecting adipose tissue dysfunction demonstrated moderate discriminative accuracy for incident T2D (C‐statistic varied between 0.70 and 0.80) over the 20‐year follow‐up period, in line with obesity‐related anthropometric indices assessing total and central adiposity (C‐statistic 0.70–0.80). Obesity‐related anthropometric indices as well as dysfunctional adipose tissue‐related indices were moderately predictive of long‐term T2D onset, highlighting the complex role of increased body weight on glucose metabolism.
Keywords: adiposity indices, dysfunctional adipose tissue, obesity, type 2 diabetes
Key Points
Traditional anthropometric indices (BMI, WC, WHR and WHtR) are widely used to assess adiposity and predict type 2 diabetes (T2D).
Dysfunctional adipose tissue–related indices (VAI, LAP, DAI, CMI, WTI and TyG) better reflect adipose tissue dysfunction and metabolic status, integrating lipid metabolism and fat distribution.
In the 20‐year ATTICA cohort, both index categories showed comparable discrimination for incident T2D (C‐statistic = 0.70–0.80).
However, dysfunctional adiposity indices provided up to a 4.7% higher predictive value for residual T2D incidence, supporting their integration into clinical stratification models.
1. Introduction
Over the past 20 years a notable rise in the prevalence of type 2 diabetes (T2D) has been observed worldwide. This rise is largely attributed to urbanisation, ageing populations and unhealthy lifestyle factors, including the widespread adoption of high‐fat diets and decreased physical activity, which contribute to increasing obesity rates [1, 2]. Adiposity and insulin resistance have long been recognised as interrelated and well‐established determinants in the pathogenesis of T2D [3, 4]. Adiposity indices are strongly associated with insulin resistance and impaired glucose metabolism, both of which precede the clinical onset of T2D. In this context, the gap between 2‐h post‐load plasma glucose (2 h PG) and fasting blood glucose (FBG) has been proposed as an informative indicator of early dysglycemia, predictive of subsequent prediabetes and diabetes development [5]. Therefore, examining this glucose gap alongside adiposity indices may provide complementary insights into adiposity‐related metabolic disturbances and their contribution to diabetes incidence.
Traditional obesity‐related anthropometric indices such as body mass index (BMI), waist circumference (WC), waist‐to‐hip ratio (WHR) and waist‐to‐height ratio (WHtR) have been used to assess adiposity and its association with metabolic disorders, including ASCVD and T2D. Furthermore, these indices have been associated with diabetes‐related microvascular complications, particularly peripheral neuropathy [6].
However, recent research has increasingly focused on dysfunctional adipose tissue‐related indices, such as the visceral adiposity index (VAI), lipid accumulation product (LAP), dysfunctional adiposity index (DAI), cardiometabolic index (CMI), waist‐circumference‐triglyceride index (WTI) and triglyceride‐glucose (TyG)‐based indices, due to their strong association with insulin resistance and metabolic dysfunction. Adipose tissue dysfunction is characterised by impaired adipocyte expandability, ectopic fat accumulation, chronic low‐grade inflammation, altered adipokine secretion, insulin resistance and dysregulated lipid metabolism. As direct assessment of adipose tissue dysfunction is challenging in large‐scale epidemiological studies, these composite anthropometric and biochemical indices have been developed as surrogate markers of visceral adiposity and metabolic dysfunction. By integrating measures of body fat distribution with metabolic parameters, they may provide a more comprehensive assessment of cardiometabolic health and have been associated with an increased incidence of cardiometabolic diseases, including T2D [7, 8, 9, 10, 11, 12].
Τhe predictive ability of obesity‐related anthropometric indices—whether related to total or central adiposity—or dysfunctional adipose tissue‐related indices has not been thoroughly assessed in relation to the odds of developing T2D. The predictive power of different obesity‐related indices (total vs. central adiposity) and dysfunctional adipose tissue‐related indices, in relation to T2D incidence is a crucial area of research, because traditional markers like BMI may not fully capture metabolic status, whereas central obesity measures (i.e., WC and WHR) and metabolic dysfunction indicators (i.e., adipokines, ectopic fat) associated with lipid metabolism and fat distribution might provide better insight. Therefore, this study aimed to evaluate the association between adiposity indices and the cumulative incidence of T2D over a long‐term period (i.e., 20 years) in apparently healthy adults from the ATTICA cohort study. Assessing the predictive value of both obesity‐related anthropometric indices and dysfunctional adipose tissue‐related indices for T2D onset is essential for enhancing our understanding of their role in the early identification of individuals likely to develop T2D.
2. Materials and Methods
2.1. Study Design
The ATTICA study is a population‐based prospective cohort study initiated in 2001–2002 in the Attica region of Greece. The study aimed to investigate the distribution and longitudinal trajectories of various sociodemographic, anthropometric, lifestyle, clinical, biochemical and psychological parameters related to cardiovascular and metabolic diseases. In 2006, 2012 and 2022, the 5‐, 10‐ and 20‐year follow‐up examinations were performed. Detailed information regarding the study design, sampling procedures and methodological framework is available in previously published papers [13, 14].
2.2. Sample and Sampling Procedure
Of the 4056 individuals who were initially invited, 3042 agreed to participate in the study (participation rate of 75%); 1514 were males (mean age: 43 ± 13 years; range: 18–87 years) and 1528 were females (mean age: 43 ± 13 years; range: 18–89 years). Participation in the ATTICA study was subject to specific exclusion criteria, including a history of any cardiovascular disease, chronic viral infections, and residence in institutional settings. Moreover, for the present study, 212 participants who were diagnosed with T2D at baseline examination in 2002 or had a history of T2D prior to enrollment were also excluded. Trained health professionals—including cardiologists, nurses, general practitioners, and dietitians—conducted standardised assessments at participants' workplaces or residences, based on a stratified by age‐sex random sampling scheme of the Attiki region (see for details [13]).
2.3. Follow‐Up Examination
In 2022, the 20‐year follow‐up examination of the participants was performed. Complete data for assessing the 20‐year cumulative incidence of T2D were available from 2000 participants who were free of T2D at baseline examination (71% participation rate); of them, 974 were males (mean age 43 ± 13 years, 18–89 years) and 1026 were females (mean age 42 ± 13 years, 18–89 years). T2D was defined according to the criteria established by the American Diabetes Association, that is, fasting plasma glucose level > 125 mg/dL and/or the administration of antidiabetic medications [15]. No significant differences were observed regarding the age‐sex distribution between the baseline sample and the 20‐year follow‐up sample (p > 0.10).
2.4. Bioethics
The ATTICA study was carried out in compliance with the ethical principles set forth in the Declaration of Helsinki. The study protocol was approved by the Ethics Committee of the First Cardiology Department of the National and Kapodistrian University of Athens (#017/01.05.2001) and by the Ethics Committee of Harokopio University (#38/29.03.2022).
2.5. Measurements
Various sociodemographic, lifestyle, psychological, anthropometric, clinical and biochemical characteristics were measured at baseline examination through standardised procedures, as it can be seen in the methodology paper of the ATTICA study [13]. Briefly, hypertension was defined as an average of three measurements of systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg, or the current use of antihypertensive agents [16], and hypercholesterolemia was defined as a total cholesterol level ≥ 200 mg/dL and/or the use of lipid‐lowering therapy [17]. Biochemical analyses included the measurement of high‐density lipoprotein cholesterol (HDL‐C), triglycerides (TGs) and fasting glucose, determined using standardised laboratory methods. HDL‐C, TGs and glucose concentrations were measured via an enzymatic chromatographic technique using a Technicon automatic analyser RA‐1000 (Dade Behring, Marburg, Germany). Dietary intake was evaluated using a validated semi‐quantitative food‐frequency questionnaire, with habitual food consumption recorded in servings per day or week [18]. Adherence to the Mediterranean diet (MedDiet) was assessed using the MedDietScore [19]. Physical activity levels were determined using the International Physical Activity Questionnaire (IPAQ), validated for the Greek population, which assessed frequency (times per week), duration (minutes per session), and intensity (calories expended per session) on a weekly basis [20]. Smoking status was classified as ‘current smoking’ for participants who reported smoking at least one cigarette per day or who had ceased smoking within the previous 12 months.
Regarding the main exposure factors of interest for the present analyses, in total 22 adiposity indices were measured at baseline examination or calculated through other measurements and categorised into groups according to the reflection of total, central adiposity, and dysfunctional adipose tissue‐related indices (see Table S1). These indices included body weight (kg), height (m), waist and hip circumferences (cm), which were obtained following established protocols; height was recorded to the nearest 0.5 cm using a stadiometer, with measurements taken at the end of a normal expiration. Participants stood upright with their head, upper back, buttocks, calves, and heels aligned against the stadiometer, ensuring proper positioning of the head in the Frankfort horizontal plane. Body weight was measured to the nearest 100 g using a digital scale placed on a firm, level surface. Participants stood barefoot, wearing light clothing and were positioned centrally on the scale without external support. Waist circumference (WC) was assessed to the nearest 0.1 cm at the midpoint between the lowest rib and the upper border of the iliac crest, with measurements taken at the end of normal expiration. Hip circumference was recorded to the nearest 0.1 cm at the widest point of the buttocks. Both measurements were performed using a non‐stretchable tape, ensuring it remained parallel to the floor.
BMI was computed as weight (kg) divided by height squared (m2), with overweight and obesity defined as BMI values of 25–29.9 kg/m2 and ≥ 30 kg/m2, respectively [21]. In addition, body surface area of Dubois (BSAD) and body surface area of Mosteller (BSAM) were calculated based on established equations and classified as indicators of total adiposity because of the incorporation of weight and height solely. Body adiposity index (BAI), a body shape index (ABSI), body roundness index (BRI), weight‐adjusted‐waist index (WWI), conicity index (ConI), relative fat mass (RFM), waist‐corrected body mass index (WCBMI) and abdominal volume index (AVI) were also calculated and categorised as indices of central adiposity together with WC, WHR and WHtR.
Indices incorporating lipid measurements, that isVAI, LAP, DAI, WTI, CMI, TyG × BMI, TyG × WC and TyG × WHtR were calculated and classified as dysfunctional adipose tissue‐related indices. VAI combines WC, BMI, TGs and HDL‐cholesterol and has been proposed as a surrogate marker of visceral adipose tissue dysfunction [7]. LAP incorporates WC and TG concentrations and reflects excessive lipid accumulation [22]. CMI combines the TG‐to‐HDL cholesterol ratio with waist‐to‐height ratio, integrating information on dyslipidemia and central adiposity [9]. The TyG index has been proposed as a surrogate marker of insulin resistance and metabolic dysfunction. The TyG‐derived indices (TyG‐BMI, TyG‐WC and TyG‐WHtR) combine the TyG index with anthropometric measures and have been proposed as markers of insulin resistance and cardiometabolic status [23, 24, 25]. DAI and the WTI have also been developed to capture adipose tissue dysfunction through the combined assessment of adiposity and lipid‐related parameters [11, 26]. To account for body weight changes, body weight, height and waist/hip circumferences were measured in subsequent follow‐up examinations. All formulas used for the calculations of various indices of adiposity with the relevant references are presented in Table S1.
2.6. Statistical Analysis
Categorical variables were expressed as relative frequencies (%), continuous variables with normal distributions as mean values with corresponding standard deviations (SDs) and continuous variables with skewed distributions as medians with corresponding interquartile ranges (IQRs). The association between categorical variables was assessed using the Pearson chi‐square test. Associations between normally distributed continuous variables and the 20‐year cumulative incidence of T2D were examined using the independent samples t‐test. The normality of continuous variables was assessed through probability–probability (P–P) plots, and homogeneity of variances was tested using Levene's test. The primary endpoint, the 20‐year cumulative incidence of T2D cases, was calculated as the ratio of incident cases to the total number of participants at follow‐up. The associations between adiposity indices and the study endpoint were evaluated using multivariable logistic regression analysis, with odds ratios (OR) serving as proxies for relative risks (RRs) and corresponding 95% confidence intervals (95% CI) reported. Multi‐adjusted logistic regression models were constructed for each adiposity index individually and in combination, adjusting for several covariates, including blood glucose levels, age, sex, smoking status at baseline, level of adherence to the MedDiet through the MedDietScore, baseline physical activity level, family history of T2D, hypercholesterolemia and hypertension. Moreover, changes in BMI or waist/hip circumferences during the first 10 years of follow‐up were considered as covariates in all analyses. Logistic regression analysis was preferred over survival models due to the unavailability of exact T2D onset times for all cases. The discriminatory ability of the estimated models was assessed using the C‐statistic, where values greater than 0.70 were considered indicative of acceptable discrimination. To evaluate which of the tested adiposity indices provided additional predictive information for incident T2D beyond established predictors, a core model including FBG levels, age, sex and family history of T2D was constructed. Based on this model, each adiposity index was evaluated against the likelihood of observing false negative cases of the core model. Missing data were not imputed, and the models ran only with complete data cases. All statistical analyses were performed using STATA software (version 17, MP & Associates, Sparta, Greece), and statistical significance was set at a two‐sided p value of < 0.05.
3. Results
3.1. Cumulative T2D Incidence at 20‐Year Follow‐Up
During the 2002–2022 period, 526 new cases of T2D were recorded out of the 2000 participants, resulting in a cumulative incidence rate of 26.3% (21.4% for females and 31.4% for males). Details regarding factors associated with the development of T2D have been presented in previous analyses and reports of the study [13, 14].
3.2. Association of Indices of Overall Adiposity and the 20‐Year Cumulative T2D Incidence
In Table 1, various adiposity indices, in terms of obesity‐related anthropometric indices (either total or central adiposity), as well as the dysfunctional adipose tissue‐related indices, were presented, both in total and stratified according to the 20‐year cumulative incidence of T2D. All markers were significantly increased among subjects who developed T2D during the 20‐year follow‐up as compared to those who did not, except for LAP.
TABLE 1.
Obesity‐related anthropometric and dysfunctional adipose tissue‐related indices according to the 20‐year cumulative incidence of T2D.
| Total sample (N = 2000) | T2D or not T2D during 20‐year FU | p | ||
|---|---|---|---|---|
| Yes (N = 526) | No (N = 1474) | |||
| Obesity‐related anthropometric indices | ||||
| Total adiposity | ||||
| BMI (kg/m2), mean (SD) | 26 (4.4) | 27 (4.6) | 25.6 (4.3) | < 0.001 |
| BMI trajectories (2002–2022) | ||||
| Overweight/obesity‐overweight/obesity | 48.4 | 60.0 | 44.9 | < 0.001 |
| Normal weight status‐overweight/obesity | 3.4 | 2.2 | 3.8 | |
| Overweight/obesity‐normal weight status | 1.6 | 2.4 | 1.4 | |
| Normal weight status‐normal weight status | 46.6 | 35.4 | 49.9 | |
| BSAD, mean (SD) | 1.9 (0.2) | 1.9 (0.2) | 1.8 (0.2) | < 0.001 |
| BSAM, mean (SD) | 1.9 (0.2) | 1.9 (0.2) | 1.9 (0.2) | < 0.001 |
| Central adiposity | ||||
| BAI, mean (SD) | 29.2 (5.9) | 29.9 (6.0) | 29.0 (5.9) | 0.003 |
| ABSI, mean (SD) | 0.08 (0.01) | 0.08 (0.01) | 0.08 (0.01) | < 0.001 |
| BRI, mean (SD) | 4.0 (1.8) | 4.4 (1.6) | 3.8 (1.8) | < 0.001 |
| WWI, mean (SD) | 10.3 (1.1) | 10.6 (1.0) | 10.2 (1.2) | < 0.001 |
| ConI, mean (SD) | 1.2 (1.1) | 1.3 (1.1) | 1.2 (1.1) | < 0.001 |
| WC (cm), mean (SD) | 89.1 (15.1) | 93.1 (13.9) | 87.7 (15.3) | < 0.001 |
| WHR, mean (SD) | 0.9 (0.1) | 0.9 (0.1) | 0.8 (0.1) | < 0.001 |
| WHtR, mean (SD) | 0.5 (0.8) | 0.6 (0.1) | 0.5 (0.1) | < 0.001 |
| RFM, mean (SD) | 31.1 (8.6) | 31.7 (6.5) | 30.9 (9.2) | 0.059 |
| WCBMI, mean (SD) | 23.6 (7.7) | 25.6 (8.0) | 23.0 (7.5) | < 0.001 |
| AVI, mean (SD) | 16.6 (5.8) | 17.9 (5.2) | 16.1 (5.9) | < 0.001 |
| Dysfunctional adipose tissue‐related indices | ||||
| VAI, median (IQR) | 1.4 (1.3) | 1.5 (1.6) | 1.3 (1.2) | < 0.001 |
| LAP, median (IQR) | 30.1 (39.6) | 32.9 (41.1) | 29.3 (39.2) | 0.318 |
| DAI, mean (SD) | 1.1 (1.2) | 1.4 (2.0) | 1.0 (0.8) | < 0.001 |
| WTI, mean (SD) | 116.9 (95.6) | 142.4 (135.1) | 108.8 (77.4) | < 0.001 |
| CMI, mean (SD) | 0.6 (0.8) | 0.8 (1.3) | 0.6 (0.5) | < 0.001 |
| TyG × BMI, mean (SD) | 217.8 (45.6) | 235.6 (45.6) | 212.2 (43.8) | < 0.001 |
| TyG × WC, mean (SD) | 746.5 (156.1) | 811.4 (146.6) | 725.9 (153.4) | < 0.001 |
| TyG × WHtR, mean (SD) | 4.4 (0.9) | 4.8 (0.8) | 4.3 (0.9) | 0.047 |
Note: p value was based on the independent samples t‐test (normally distributed continuous characteristics) and on the Mann–Whitney U test (non‐normally distributed continuous characteristics).
Abbreviations: ABSI, a body shape index; AVI, abdominal volume index; BAI, body adiposity index; BMI, body mass index; BRI, body roundness index; BSAD, body surface area of Dubois; BSAM, body surface area of Mosteller; ConI, conicity index; CMI, cardiometabolic index; DAI, dysfunctional adiposity index; FU, follow‐up; IQR, interquartile range; LAP, lipid accumulation product; RFM, relative fat mass; SD, standard deviation; T2D, type 2 diabetes; TyG, triglyceride‐glucose index; TyG × BMI, triglyceride‐glucose index × body mass index; TyG × WC, triglyceride‐glucose index × waist circumference; TyG × WHtR, triglyceride‐glucose index × waist circumference‐to‐height ratio; VAI, visceral adiposity index; WC, waist circumference; WCBMI, waist‐corrected body mass index; WHR, waist circumference‐to‐hip ratio; WHtR, waist circumference‐to‐height ratio; WTI, waist circumference‐triglyceride index; WWI, weight‐adjusted‐waist index.
Multi‐adjusted models were unveiled to evaluate the association between various adiposity indices and the 20‐year cumulative incidence of T2D, to control for potential residual confounding (Table 2). In the nested (unadjusted) models, all adiposity indices, by the exception of LAP, were significantly associated with the 20‐year cumulative incidence of T2D. After considering adjusting variables, that is, age, sex, smoking at baseline, MedDietScore, physical activity at baseline, family history of T2D, hypercholesterolemia and hypertension, all the studied adiposity indices appeared to retain their aggravating effect, except for ABSI, WWI, ConI, WHR and LAP. Among the studied indices, those corresponding to body surface area, that is, BSAD and BSAM, exhibited the strongest association with 20‐year cumulative incidence of T2D.
TABLE 2.
Multi‐adjusted logistic regression models and the 20‐year cumulative incidence of T2D.
| Adiposity index | Model A: Nested model | Model B (naïve): Model A, plus age and sex | Model C: Model B plus smoking, MedDietScore and physical activity | Model D: Fully adjusted model (Model C, plus family history of T2D, hypercholesterolemia, and hypertension) |
|---|---|---|---|---|
| Total adiposity | ||||
| BMI (per 1‐unit increment) | 1.07 (1.05, 1.09) | 1.05 (1.03, 1.08) | 1.06 (1.03, 1.08) | 1.11 (1.07, 1.15) |
| BSAD (per 1‐unit increment) | 3.13 (1.96, 4.99) | 2.24 (1.20, 4.19) | 2.34 (1.24, 4.41) | 5.52 (2.10, 14.83) |
| BSAM (per 1‐unit increment) | 3.18 (2.04, 4.95) | 2.30 (1.30, 4.08) | 2.42 (1.35, 4.35) | 5.90 (2.39, 14.70) |
| Central adiposity | ||||
| BAI (per 1‐unit increment) | 1.03 (1.01, 1.05) | 1.04 (1.02, 1.06) | 1.05 (1.02, 1.07) | 1.07 (1.04, 1.11) |
| ABSI (per 0.1‐unit increment) | 16.63 (4.44, 62.23) | 3.26 (0.85, 1.58) | 3.24 (0.84, 12.54) | 1.86 (0.21, 16.10) |
| BRI (per 1‐unit increment) | 1.19 (1.12, 1.26) | 1.11 (1.05, 1.19) | 1.12 (1.05, 1.20) | 1.22 (1.09, 1.36) |
| WWI (per 1‐unit increment) | 1.28 (1.16, 1.41) | 1.14 (1.03, 1.26) | 1.14 (1.03, 1.26) | 1.17 (0.98, 1.40) |
| ConI (per 1‐unit increment) | 9.23 (4.08, 20.90) | 3.13 (1.32, 7.43) | 3.19 (1.33, 7.63) | 3.74 (0.86, 16.17) |
| WC (per 1‐unit increment) | 1.02 (1.02, 1.03) | 1.02 (1.01, 1.02) | 1.02 (1.01, 1.02) | 1.03 (1.01, 1.04) |
| WHR (per 0.1‐unit increment) | 1.33 (1.21, 1.46) | 1.14 (1.01, 1.28) | 1.14 (1.02, 1.28) | 1.16 (0.96, 1.41) |
| WHtR (per 1‐unit increment) | 1.53 (1.35, 1.74) | 1.33 (1.16, 1.54) | 1.40 (1.18, 1.58) | 1.65 (1.30, 2.10) |
| RFM (per 1‐unit increment) | 1.01 (1.01, 1.02) | 1.05 (1.03, 1.07) | 1.05 (1.03, 1.08) | 1.09 (1.05, 1.13) |
| WCBMI (per 1‐unit increment) | 1.04 (1.03, 1.06) | 1.03 (1.01, 1.04) | 1.03 (1.02, 1.05) | 1.06 (1.01, 1.03) |
| AVI (per 1‐unit increment) | 1.05 (1.03, 1.07) | 1.03 (1.01, 1.05) | 1.03 (1.01, 1.05) | 1.05 (1.02, 1.09) |
| Dysfunctional adipose tissue | ||||
| VAI (per 1‐unit increment) | 1.19 (1.11, 1.29) | 1.13 (1.04, 1.21) | 1.14 (1.05, 1.23) | 1.21 (1.07, 1.36) |
| LAP (per 1‐unit increment) | 1.01 (0.99, 1.01) | 1.01 (0.99, 1.01) | 1.01 (0.99, 1.01) | 0.99 (0.99, 1.00) |
| DAI (per 1‐unit increment) | 1.33 (1.18, 1.51) | 1.19 (1.06, 1.35) | 1.21 (1.07, 1.37) | 1.32 (1.09, 1.60) |
| WTI (per 1‐unit increment) | 1.01 (1.01, 1.02) | 1.01 (1.01, 1.02) | 1.01 (1.01, 1.02) | 1.01 (1.01, 1.02) |
| CMI (per 1‐unit increment) | 1.67 (1.38, 2.02) | 1.35 (1.11, 1.65) | 1.40 (1.13, 1.70) | 1.65 (1.20, 2.26) |
| TyG × BMI (per 1‐unit increment) | 1.01 (1.01, 1.02) | 1.01 (1.01, 1.02) | 1.01 (1.01, 1.02) | 1.02 (1.01, 1.02) |
| TyG × WC (per 1‐unit increment) | 1.01 (1.01, 1.02) | 1.01 (1.01, 1.02) | 1.01 (1.01, 1.02) | 1.01 (1.01, 1.02) |
| TyG × WHtR (per 1‐unit increment) | 1.90 (1.65, 2.18) | 1.74 (1.49, 2.03) | 1.84 (1.56, 2.17) | 2.28 (1.75, 2.98) |
Note: RRs and their corresponding 95% CIs were obtained through logistic regression analysis as proxy of Odds Ratios. Fully adjusted model included: glucose levels, age (years), sex (men vs. women), smoking status at baseline (yes vs. no), adherence level to the Mediterranean Diet at baseline (MedDietScore), physical activity at baseline (yes/no), family history of T2D, hypercholesterolemia and hypertension. Level of adherence to the Mediterranean Diet was evaluated through the MedDietScore.
Abbreviations: ABSI, a body shape index; AVI, abdominal volume index; BAI, body adiposity index; BMI, body mass index; BRI, body roundness index; BSAD, body surface area of Du Bois; BSAM, body surface area of Mosteller; CI, confidence interval; ConI, conicity index; CMI, cardiometabolic index; DAI, dysfunctional adiposity index; LAP, lipid accumulation product; RFM, relative fat mass; RR, relative risk; T2D, type 2 diabetes; TyG, triglyceride‐glucose index; TyG × BMI, triglyceride‐glucose index × body mass index; TyG × WC, triglyceride‐glucose index × waist circumference; TyG × WHtR, triglyceride‐glucose index × waist circumference‐to‐height ratio; VAI, visceral adiposity index; WC, waist circumference; WCBMI, waist‐corrected body mass index; WHR, waist circumference‐to‐hip ratio; WHtR, waist circumference‐to‐height ratio; WTI, waist circumference‐triglyceride index; WWI, weight‐adjusted‐waist index.
The accuracy of the evaluated adiposity indices was found approximately 0.75 (C‐statistic). Almost all dysfunctional adipose tissue‐related indices exhibited moderate discriminative accuracy (i.e., C‐statistic varied between 0.70 and 0.80), for the development of T2D over the 20‐year follow‐up period (Figure 1). In line with the previous finding, obesity‐related anthropometric indices, assessing total and central adiposity exhibited also moderate discriminative accuracy (C‐statistic 0.70–0.80) (Figure 1).
FIGURE 1.
Accuracy (C‐statistic index) of adiposity indices for predicting incidence of T2D during a 20‐year follow‐up period. Values between 0.5 and 0.6 are considered as very poor discrimination ability, between 0.6 and 0.7 are considered as poor, between 0.7 and 0.8 as acceptable and above 0.8 as good/very good discrimination.
Furthermore, the additional predictive performance provided by the tested adiposity indices for incident T2D ranged from 0% for obesity‐related anthropometric indices to 4.7% for dysfunctional adipose tissue‐related indices. Among the latter, TyG × WHtR and TyG × WC demonstrated the highest discriminative ability.
4. Discussion
The present study examined the association between a variety of adiposity indices, in terms of obesity‐related anthropometric indices and dysfunctional adipose tissue‐related indices, and the 20‐year cumulative incidence of T2D among apparently healthy adults participating in the ATTICA cohort study (2002–2022). The assessment of excess adiposity using traditional anthropometric indices—encompassing measures of total and central adiposity—or indices indicative of adipose tissue dysfunction can improve the identification and stratification of future T2D cases, with comparable predictive accuracy at both the individual and population levels.
According to the findings of a previous analysis in the same cohort, obesity indices were strongly associated with the long‐term incidence of cardiovascular disease, highlighting the critical role of excess adiposity in its pathophysiology. Furthermore, incorporating an obesity index into a cardiovascular disease prediction model significantly improved its predictive accuracy and reclassification ability, emphasising the value of obesity indices in refining the assessment and identification of individuals likely to develop cardiovascular disease in the general population [27]. Regarding diabetes, obesity has long been associated with increased odds of developing T2D. Nonetheless, which obesity index most accurately correlates with the incidence of T2D continues to be a subject of considerable debate. Several studies have demonstrated that BMI is unable to differentiate between lean mass (muscle) and adipose tissue, particularly visceral fat. Thus, it does not provide an accurate reflection of abdominal fat distribution. This limitation has led to increasing criticism of BMI as a reliable measure of obesity, as it primarily assesses body weight without accounting for variations in body composition, such as fat mass versus lean mass [28]. Moreover, the use of BMI for obesity classification has significant limitations, particularly among individuals with moderate or high‐intensity physical activity. BMI does not distinguish between lean mass and fat mass, potentially leading to misclassification of overweight or obese [29]. According to a review related to the definition of obesity, BMI should be used only as a surrogate indicator of health status at the population level, primarily for epidemiological research or screening purposes, rather than as a standalone measure of individual health. Excess adiposity should be confirmed either through direct body fat assessment, where available, or by incorporating at least one additional anthropometric measure—such as WC, WHR or WHtR—alongside BMI, using validated methods and cutoff points appropriate for age, sex, and ethnicity. However, in individuals with a BMI ≥ 40 kg/m2, excess adiposity can reasonably be assumed without further verification. Additionally, individuals identified as having excess adiposity, with or without associated organ or tissue dysfunction, should be further assessed for clinical obesity [30].
Evidence suggests that abnormal fat accumulation and the transition from healthy to dysfunctional adipose tissue are more strongly correlated with an increased odds of developing T2D, dyslipidemia, hypertension and metabolic syndrome. According to findings from a meta‐analysis, metabolically unhealthy normal weight, a phenotype of obesity characterised by normal BMI but increased body fat percentage, was strongly associated with an increased incidence of cardiometabolic disorders, underscoring the limitations of BMI as a sole measure of obesity [31]. In light of these considerations, more comprehensive measures of obesity, including both fat quantity and distribution, in combination with the metabolic dysfunction associated with lipid metabolism and fat distribution, have emerged, demonstrating a possible association with T2D incidence [8, 10]. According to the findings of our study, BMI was strongly and independently associated with the 20‐year cumulative incidence of T2D, demonstrating moderate predictive accuracy.
BSAD and BSAM, have been used to estimate body surface area. Body size exerts a significant influence on the results of a standardised oral glucose tolerance test (OGTT). Smaller individuals tend to exhibit higher susceptibility to glucose intolerance and are more frequently diagnosed with impaired glucose metabolism, compared to individuals with larger body sizes, potentially leading to diagnostic disparities [32]. However, after adjusting for central adiposity, higher level of body surface area was inversely associated with the 2‐h plasma glucose levels after an OGTT, underlying the role of visceral fat and its strong association with insulin resistance. In our study, increased values of both BSAD and BSAM, were independently associated with an increased odds of developing T2D, during a long‐time frame, that is, 20 years, after adjusting for comorbidities and parameters of lifestyle, underlying the possible aggravating role of body surface, in terms of weight and height, on glucose metabolism. The discriminative capacity of both BSAD and BSAM was similar to the accuracy of BMI, concerning the cumulative 20‐year T2D incidence.
Although adiposity is a well‐established factor for the development of T2D, central adiposity, as reflected by the accumulation of visceral fat, contributes to the development of T2D through multiple mechanisms, including insulin resistance, increased inflammation, adipokine imbalance, ectopic fat deposition, and beta‐cell dysfunction. The excess visceral fat plays a pivotal role in disrupting metabolic homeostasis, and addressing central obesity is crucial for the prevention and management of T2D [33, 34, 35]. According to findings from a dose–response meta‐analysis of cohort studies, increased values of central adiposity were strongly and linearly associated with T2D incidence, independently of overall adiposity [33]. In line with the previous findings, obesity defined by BMI and markers of central adiposity, in terms of WC and WHR, were strongly associated with the incidence of T2D [36], while WCBMI, which incorporates WC and BMI, RFM, and AVI, demonstrated the best discrimination and accuracy compared to other traditional indices (i.e., BMI, WC and WHR) [37, 38]. According to the findings of our study, the majority of obesity‐related anthropometric indices of central adiposity were also positively associated with 20‐year T2D incidence, but with moderate discriminative capacity and comparable to the accuracy of indices of total adiposity.
Ectopic fat deposition in non‐adipose tissues has been associated with several cardiometabolic diseases, that is, T2D, and cardiovascular diseases, through the mechanism of insulin resistance in liver and skeletal muscles leading to hyperglycemia, and the disrupted insulin secretion from pancreas [3]. Although BMI remains a useful screening tool for identifying individuals with an unfavourable cardiometabolic profile, it has limited specificity in detecting impaired adipose tissue function. Furthermore, obesity‐related anthropometric indices of central adiposity cannot distinguish visceral and subcutaneous adipose tissue, which is important to evaluate atherosclerotic burden in dysmetabolic patients [39]. Dysfunctional adipose tissue‐related indices by integrating body fat distribution with lipid metabolism provide a more comprehensive assessment of metabolic health beyond traditional measures. A recent large cohort study revealed that TyG demonstrated the best predictive ability for T2D incidence, among middle‐aged and elderly males and females, compared to traditional anthropometric indices (i.e., BMI, WC, WHtR, ABSI and BRI); TyG's correlation indices (i.e., TyG‐BMI, TyG‐WC and TyG‐WHtR), and LAP, outperformed BMI, WC and WHtR, in predicting T2D [40]. Furthermore, DAI has been associated with early cardiometabolic abnormalities based on adipocytes morpho‐functional abnormalities, independently of adiposity, and was associated with an increased odds of developing diabetes, hypertension, non‐alcoholic liver disease, and subclinical atherosclerosis [11]. According to findings from a large cohort study, high CMI was a detrimental factor for new‐onset T2D, while significant associations were observed between changes in CMI status and the development of T2D [41]. CMI has been significantly non‐linearly correlated with insulin resistance in both males and females, and its accuracy was the largest, compared to traditional indices of adiposity and lipid profile, that is, BMI, WC, WHtR, VAI, DAI and TGs/HDL‐C [9]. In line with the previous findings, in our study the predictive ability of dysfunctional adipose tissue‐related indices, except for LAP, demonstrated moderate predictive ability for developing T2D, during the 20‐year period, comparable to traditional obesity‐related anthropometric indices of total and central adiposity, suggesting the aggravating role of excess body weight on both glucose homeostasis, and insulin sensitivity, as well.
Importantly, dysfunctional adiposity‐related indices demonstrated superior predictive performance compared with traditional anthropometric measures. While BMI and WC remain simple and widely used indicators of adiposity, indices incorporating metabolic parameters, such as TGs and glucose, provide additional discriminatory ability for incident T2D. These findings suggest that the integration of anthropometric and metabolic information may better capture the complex pathophysiological processes underlying adipose tissue dysfunction and diabetes development.
The similar C‐statistics observed across the evaluated dysfunctional adiposity‐related indices may be explained by the substantial overlap in their constituent components, including WC, TGs, glucose levels, BMI and other lipid‐related parameters. As these indices capture closely related aspects of adipose tissue dysfunction, insulin resistance, and metabolic dysregulation, they likely reflect overlapping pathophysiological pathways, resulting in comparable discriminatory performance. Nevertheless, dysfunctional adiposity‐related indices consistently demonstrated slightly better predictive performance than traditional anthropometric measures, suggesting that the incorporation of metabolic parameters may provide additional information for the identification of future T2D cases.
4.1. Strengths and Limitations
Τhe ATTICA study is a large‐scale prospective observational study with 20‐year follow‐up period, making it one of the few long‐term epidemiological studies of T2D worldwide. However, certain limitations should be acknowledged when interpreting the findings. Despite adjustments for multiple covariates, residual confounding may still be present due to the observational study design, potentially influencing the associations observed. The exact date of diabetes onset was not available, precluding the calculation of incidence rates and time‐to‐event analyses. Calculation of adiposity indices was conducted only at baseline, limiting insights into longitudinal changes and their potential role in disease development. The study did not utilise body composition analysers to assess fat distribution and muscle mass, limiting the ability to distinguish between sarcopenic obesity and metabolically unhealthy obesity phenotypes. Approximately 25% of participants were lost to follow‐up, introducing potential selection bias. Despite these limitations, the ATTICA study remains a valuable contribution to the epidemiology of diabetes, providing important insights into the long‐term impact of obesity‐related indices on diabetes incidence.
5. Conclusions
Through the evaluation of anthropometric indices representing total and central adiposity, together with markers of metabolic dysfunction associated with lipid metabolism and adipose tissue distribution, excess body weight emerged as a key determinant of future T2D development, independent of glucose levels. These findings underscore the importance of incorporating adiposity measurements into prevention strategies, focusing on the early identification and management of excess body weight to reduce the burden of T2D.
Author Contributions
Ioanna Kechagia: conceptualisation, data curation and writing. Sofia‐Panagiota Giannakopoulou: visualisation, review and editing. Fotios Barkas, Mary Yannakoulia, Evangelos Liberopoulos, Petros P. Sfikakis, Christos Pitsavos, and Demosthenes Panagiotakos: design, methodology, review and editing. All authors have read and agreed to the published version of the manuscript.
Funding
The ATTICA Study was funded by Hellenic Cardiological Society in 2002 and Hellenic Atherosclerosis Society in 2004 and 2015.
Ethics Statement
The ATTICA study was conducted in accordance with the Declaration of Helsinki (1989) of the World Medical Association and was approved by the Institutional Ethics Committee of Athens Medical School (#017/1.5.2001), and the Ethics Committee of the Harokopio University (#38/29.03.2022).
Consent
Informed consent was obtained from all subjects involved in the study. All participants were informed about the aims and procedures and agreed to participate providing signed written consent.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: Adiposity indices equations.
Acknowledgements
The authors would like to thank the ATTICA study group of investigators: Evrydiki Kravvariti, Evangelia Damigou, Elpiniki Vlachopoulou, Christina Vafia, Dimitris Dalmyras, Konstantina Kyrili, Petros Spyridonas Adamidis, Georgia Anastasiou, Amalia Despoina Koutsogianni, Evangelinos Michelis, Asimina Loukina, Giorgos Metzantonakis, Manolis Kambaxis, Kyriakos Dimitriadis, Ioannis Andrikou, Amalia Sofianidi, Natalia Sinou, Aikaterini Skandali, Christina Sousouni, for their assistance on the 20‐year follow‐up, as well as Ekavi N. Georgousopoulou, Natassa Katinioti, Labros Papadimitriou, Konstantina Masoura, Spiros Vellas, Yannis Lentzas, Manolis Kambaxis, Konstantina Palliou, Vassiliki Metaxa, Agathi Ntzouvani, Dimitris Mpougatsas, Nikolaos Skourlis, Christina Papanikolaou, Georgia‐Maria Kouli, Aimilia Christou, Adella Zana, Maria Ntertimani, Aikaterini Kalogeropoulou, Evangelia Pitaraki, Alexandros Laskaris, Mihail Hatzigeorgiou and Athanasios Grekas, Efi Tsetsekou, Carmen Vassiliadou, George Dedoussis, Marina Toutouza‐Giotsa, Konstantina Tselika and Sia Poulopoulou and Maria Toutouza for their assistance in the initial and follow‐up evaluations. The publication of this article in OA mode was financially supported by HEAL‐Link.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Sun H., Saeedi P., Karuranga S., et al., “IDF Diabetes Atlas: Global, Regional and Country‐Level Diabetes Prevalence Estimates for 2021 and Projections for 2045,” Diabetes Research and Clinical Practice 183 (2022): 109119, 10.1016/j.diabres.2021.109119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Ong K. L., Stafford L. K., McLaughlin S. A., et al., “Global, Regional, and National Burden of Diabetes From 1990 to 2021, With Projections of Prevalence to 2050: A Systematic Analysis for the Global Burden of Disease Study 2021,” Lancet 402, no. 10397 (2023): 203–234, 10.1016/S0140-6736(23)01301-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Galicia‐Garcia U., Benito‐Vicente A., Jebari S., et al., “Pathophysiology of Type 2 Diabetes Mellitus,” International Journal of Molecular Sciences 21, no. 17 (2020): 6275, 10.3390/ijms21176275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Klein S., Gastaldelli A., Yki‐Järvinen H., and Scherer P. E., “Why Does Obesity Cause Diabetes?,” Cell Metabolism 34, no. 1 (2022): 11–20, 10.1016/j.cmet.2021.12.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Liu W., Liu Z., Ding C., Li J., and Jiang H., “Associations of the Gap Between 2‐Hour Post‐Load Plasma Glucose and Fasting Blood Glucose With All‐Cause or Cardiovascular Mortality in US Normoglycemic Adults,” Biological Research for Nursing 27, no. 3 (2025): 391–399, 10.1177/10998004251316688. [DOI] [PubMed] [Google Scholar]
- 6. Fadlilah S., Amelia V. L., Tuppal C. P., et al., “Significant Impacts of the Body‐Mass Index, Blood Pressure, Blood Glucose, and Ankle‐Brachial Index on Peripheral Neuropathy Risk in Indonesian With Type 2 Diabetes: A Cross‐Sectional Study,” Biological Research for Nursing 27, no. 4 (2025): 558–569, 10.1177/10998004251336795. [DOI] [PubMed] [Google Scholar]
- 7. Shen F., Guo C., Zhang D., Liu Y., and Zhang P., “Visceral Adiposity Index as a Predictor of Type 2 Diabetes Mellitus Risk: A Systematic Review and Dose–Response Meta‐Analysis,” Nutrition, Metabolism and Cardiovascular Diseases 34, no. 4 (2024): 811–822, 10.1016/j.numecd.2023.04.009. [DOI] [PubMed] [Google Scholar]
- 8. Sadeghi E., Khodadadiyan A., Hosseini S. A., et al., “Novel Anthropometric Indices for Predicting Type 2 Diabetes Mellitus,” BMC Public Health 24, no. 1 (2024): 1033, 10.1186/s12889-024-18541-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Wu L. and Xu J., “Relationship Between Cardiometabolic Index and Insulin Resistance in Patients With Type 2 Diabetes,” Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy 17 (2024): 305–315, 10.2147/DMSO.S449374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Deng H., Hu P., Li H., et al., “Novel Lipid Indicators and the Risk of Type 2 Diabetes Mellitus Among Chinese Hypertensive Patients: Findings From the Guangzhou Heart Study,” Cardiovascular Diabetology 21, no. 1 (2022): 212, 10.1186/s12933-022-01660-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Reyes‐Barrera J., Sainz‐Escárrega V. H., Medina‐Urritia A. X., et al., “Dysfunctional Adiposity Index as a Marker of Adipose Tissue Morpho‐Functional Abnormalities and Metabolic Disorders in Apparently Healthy Subjects,” Adipocytes 10, no. 1 (2021): 142–152, 10.1080/21623945.2021.1893452. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Anto E. O., Frimpong J., Boadu W. I. O., et al., “Cardiometabolic Syndrome Among General Adult Population in Ghana: The Role of Lipid Accumulation Product, Waist Circumference‐Triglyceride Index, and Triglyceride‐Glucose Index as Surrogate Indicators,” Health Science Reports 6, no. 7 (2023): e1419, 10.1002/hsr2.1419. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Kechagia I., Tsiampalis T., Damigou E., et al., “Long‐Term Adherence to the Mediterranean Diet Reduces 20‐Year Diabetes Incidence: The ATTICA Cohort Study (2002‐2022),” Metabolites 14, no. 4 (2024): 182, 10.3390/metabo14040182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Kechagia I., Barkas F., Liberopoulos E., et al., “Association Between Simple, Combined Lipid Markers and 20‐Year Cumulative Incidence of Type 2 Diabetes: The ATTICA Cohort Study (2002‐2022),” Lipids in Health and Disease 23, no. 1 (2024): 413, 10.1186/s12944-024-02383-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus , “Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus,” Diabetes Care 20, no. 7 (1997): 1183–1197, 10.2337/diacare.20.7.1183. [DOI] [PubMed] [Google Scholar]
- 16. Chobanian A. V., Bakris G. L., Black H. R., et al., “Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure,” Hypertension 42, no. 6 (2003): 1206–1252, 10.1161/01.HYP.0000107251.49515.c2. [DOI] [PubMed] [Google Scholar]
- 17. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults . “Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III),” JAMA 285, no. 19 (2001): 2486–2497, 10.1001/jama.285.19.2486. [DOI] [PubMed] [Google Scholar]
- 18. Katsouyanni K., Rimm E. B., Gnardellis C., Trichopoulos D., Polychronopoulos E., and Trichopoulou A., “Reproducibility and Relative Validity of an Extensive Semi‐Quantitative Food Frequency Questionnaire Using Dietary Records and Biochemical Markers Among Greek Schoolteachers,” International Journal of Epidemiology 26, no. 1 (1997): S118–S127, 10.1093/ije/26.suppl_1.s118. [DOI] [PubMed] [Google Scholar]
- 19. Panagiotakos D. B., Milias G. A., Pitsavos C., and Stefanadis C., “MedDietScore: A Computer Program That Evaluates the Adherence to the Mediterranean Dietary Pattern and Its Relation to Cardiovascular Disease Risk,” Computer Methods and Programs in Biomedicine 83, no. 1 (2006): 73–77, 10.1016/j.cmpb.2006.05.003. [DOI] [PubMed] [Google Scholar]
- 20. Papathanasiou G., Georgoudis G., Papandreou M., et al., “Reliability Measures of the Short International Physical Activity Questionnaire (IPAQ) in Greek Young Adults,” Hellenic Journal of Cardiology 50, no. 4 (2009): 283–294. [PubMed] [Google Scholar]
- 21. Kushner R. F., “Clinical Assessment and Management of Adult Obesity,” Circulation 126, no. 24 (2012): 2870–2877, 10.1161/CIRCULATIONAHA.111.075424. [DOI] [PubMed] [Google Scholar]
- 22. Department of Internal Medicine, Dr. Cipto Mangunkusumo National General Hospital, Faculty of Medicine University of Indonesia , Ayundini G., Astrella C., et al., “A Systematic Review on the Association Between Lipid Accumulation Product Index and Type 2 Diabetes Mellitus,” Journal of the ASEAN Federation of Endocrine Societies 34, no. 1 (2019): 16–20, 10.15605/jafes.034.01.04. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Song B., Zhao X., Yao T., et al., “Triglyceride Glucose‐Body Mass Index and Risk of Incident Type 2 Diabetes Mellitus in Japanese People With Normal Glycemic Level: A Population‐Based Longitudinal Cohort Study,” Frontiers in Endocrinology 13 (2022): 907973, 10.3389/fendo.2022.907973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Zheng S., Shi S., Ren X., et al., “Triglyceride Glucose‐Waist Circumference, a Novel and Effective Predictor of Diabetes in First‐Degree Relatives of Type 2 Diabetes Patients: Cross‐Sectional and Prospective Cohort Study,” Journal of Translational Medicine 14, no. 1 (2016): 260, 10.1186/s12967-016-1020-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Xuan W., Liu D., Zhong J., Luo H., and Zhang X., “Impacts of Triglyceride Glucose‐Waist to Height Ratio on Diabetes Incidence: A Secondary Analysis of A Population‐Based Longitudinal Data,” Frontiers in Endocrinology 13 (2022): 949831, 10.3389/fendo.2022.949831. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Chen D., Liang Z., Sun H., et al., “Association Between Hypertriglyceridemic‐Waist Phenotype and Risk of Type 2 Diabetes Mellitus in Middle‐Aged and Older Chinese Population: A Longitudinal Cohort Study,” International Journal of Environmental Research and Public Health 18, no. 18 (2021): 9618, 10.3390/ijerph18189618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Giannakopoulou S. P., Barkas F., Chrysohoou C., et al., “Comparative Analysis of Obesity Indices in Discrimination and Reclassification of Cardiovascular Disease Risk: The ATTICA Study (2002‐2022),” European Journal of Internal Medicine 134 (2025): 96–103, 10.1016/j.ejim.2025.02.007. [DOI] [PubMed] [Google Scholar]
- 28. Khan I., Chong M., Le A., et al., “Surrogate Adiposity Markers and Mortality,” JAMA Network Open 6, no. 9 (2023): e2334836, 10.1001/jamanetworkopen.2023.34836. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Sweatt K., Garvey W. T., and Martins C., “Strengths and Limitations of BMI in the Diagnosis of Obesity: What Is the Path Forward?,” Current Obesity Reports 13, no. 3 (2024): 584–595, 10.1007/s13679-024-00580-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Rubino F., Cummings D. E., Eckel R. H., et al., “Definition and Diagnostic Criteria of Clinical Obesity,” Lancet Diabetes & Endocrinology 13, no. 3 (2025): 221–262, 10.1016/S2213-8587(24)00316-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Mohammadian Khonsari N., Khashayar P., Shahrestanaki E., et al., “Normal Weight Obesity and Cardiometabolic Risk Factors: A Systematic Review and Meta‐Analysis,” Frontiers in Endocrinology 13 (2022): 857930, 10.3389/fendo.2022.857930. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Palmu S., Rehunen S., Kautiainen H., Eriksson J. G., and Korhonen P. E., “Body Surface Area and Glucose Tolerance—The Smaller the Person, the Greater the 2‐Hour Plasma Glucose,” Diabetes Research and Clinical Practice 157 (2019): 107877, 10.1016/j.diabres.2019.107877. [DOI] [PubMed] [Google Scholar]
- 33. Jayedi A., Soltani S., Motlagh S. Z. T., et al., “Anthropometric and Adiposity Indicators and Risk of Type 2 Diabetes: Systematic Review and Dose‐Response Meta‐Analysis of Cohort Studies,” BMJ 376 (2022): e067516, 10.1136/bmj-2021-067516. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Ruze R., Liu T., Zou X., et al., “Obesity and Type 2 Diabetes Mellitus: Connections in Epidemiology, Pathogenesis, and Treatments,” Frontiers in Endocrinology 14 (2023): 1161521, 10.3389/fendo.2023.1161521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Wondmkun Y. T., “Obesity, Insulin Resistance, and Type 2 Diabetes: Associations and Therapeutic Implications,” Diabetes, Metabolic Syndrome and Obesity 13 (2020): 3611–3616, 10.2147/DMSO.S275898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Issaka A., Cameron A. J., Paradies Y., et al., “Associations Between Obesity Indices and Both Type 2 Diabetes and Impaired Fasting Glucose Among West African Adults: Results From WHO STEPS Surveys,” Nutrition, Metabolism, and Cardiovascular Diseases 31, no. 9 (2021): 2652–2660, 10.1016/j.numecd.2021.05.028. [DOI] [PubMed] [Google Scholar]
- 37. Suthahar N., Wang K., Zwartkruis V. W., et al., “Associations of Relative Fat Mass, a New Index of Adiposity, With Type‐2 Diabetes in the General Population,” European Journal of Internal Medicine 109 (2023): 73–78, 10.1016/j.ejim.2022.12.024. [DOI] [PubMed] [Google Scholar]
- 38. Li X., Zhao D., and Wang H., “Association Between Weight‐Adjusted Waist Index and Risk of Diabetes Mellitus Type 2 in United States Adults and the Predictive Value of Obesity Indicators,” BMC Public Health 24, no. 1 (2024): 2025, 10.1186/s12889-024-19576-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Ferreira J., Cunha P., Carneiro A., et al., “Is Obesity a Risk Factor for Carotid Atherosclerotic Disease?—Opportunistic Review,” Journal of Cardiovascular Development and Disease 9, no. 5 (2022): 162, 10.3390/jcdd9050162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Zhang X., Wang Y., Li Y., et al., “Optimal Obesity‐ and Lipid‐Related Indices for Predicting Type 2 Diabetes in Middle‐Aged and Elderly Chinese,” Scientific Reports 14, no. 1 (2024): 10901, 10.1038/s41598-024-61592-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Qiu Y., Yi Q., Li S., et al., “Transition of Cardiometabolic Status and the Risk of Type 2 Diabetes Mellitus Among Middle‐Aged and Older Chinese: A National Cohort Study,” Journal of Diabetes Investigation 13, no. 8 (2022): 1426–1437, 10.1111/jdi.13805. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Adiposity indices equations.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.