Abstract
Introduction
Abdominal obesity is the most common risk factor of pre-diabetes and diabetes. Currently, several types of indices are used for the determination of visceral fat-related abdominal obesity. To better understand the effect of the different adiposity indices, we sought to evaluate the association of different adiposity measurements, assessed using dual-energy X-ray absorptiometry (DXA), and pre-diabetes.
Research design and methods
This cross-sectional study included 1184 adults between 18 and 65 years who participated in the Kuwait Wellbeing Study. Anthropometry measurements included body mass index (BMI) and waist-to-hip ratio. Total body fat (TBF) mass, android fat mass, gynoid fat, and visceral adipose tissue (VAT) mass were measured using the Lunar iDXA. Pre-diabetes was defined as 5.7≤HbA1c%≤6.4. Adjusted prevalence ratios (aPRs) and 95% CIs were estimated. Area under the curve (AUC) was estimated for each adiposity measurement as predictor of pre-diabetes.
Results
A total of 585 (49.4%) males and 599 (50.6%) females were enrolled in the study. Increased BMI (aPR obese vs normal=1.59, 95% CI: 1.19 to 2.12), waist-to-hip ratio (aPR Q4 vs Q1=1.25, 0.96 to 1.61), TBF (aPR Q4 vs Q1=1.58, 1.20 to 2.07), android fat (aPR Q4 vs Q1=1.67, 1.27 to 2.20), gynoid fat (aPR Q4 vs Q1=1.48, 1.16 to 1.89), android-to-gynoid fat ratio (aPR Q4 vs Q1=1.70, 1.27 to 2.28), and VAT mass (aPR Q4 vs Q1=2.05, 1.49 to 2.82) were associated with elevated pre-diabetes prevalence. Gynoid fat was associated with pre-diabetes among males (aPR Q4 vs Q1=1.71, 1.22 to 2.41), but not among females (aPR Q4 vs Q1=1.27, 0.90 to 1.78). Moreover, in terms of AUC, VAT had the highest estimated AUC of 0.680, followed by android-to-gynoid fat ratio (AUC: 0.647) and android fat (AUC: 0.646).
Conclusions
Pre-diabetes prevalence increased as adiposity measurements increased, with VAT mass demonstrating the highest AUC for pre-diabetes.
Keywords: Adiposity; Fat; Diabetes Mellitus, Type 2; Epidemiology
WHAT IS ALREADY KNOWN ON THIS TOPIC.
Abdominal obesity indicators, such as body mass index, waist-to-hip ratio, abdominal fat, and visceral adipose tissue (VAT), have been linked to the risk of type 2 diabetes mellitus. Nonetheless, their association with pre-diabetes is less known.
WHAT THIS STUDY ADDS
The prevalence of pre-diabetes increased as the following body composition measures increased: body mass index, total body fat mass, android fat mass, gynoid fat mass, android-to-gynoid fat ratio, and VAT mass.
The aforementioned adiposity indices showed similar associations with pre-diabetes prevalence across males and females, except for gynoid fat mass, which showed an association with pre-diabetes only among males.
Moreover, receiver operating characteristics analysis showed that VAT was the best predictor of pre-diabetes in terms of the estimated area under the curve, followed by android-to-gynoid fat ratio and android fat mass.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Adiposity measurements assessed using dual-energy X-ray absorptiometry demonstrate improved prediction of pre-diabetes compared with the usual anthropometric measurements.
Introduction
Unhealthy lifestyle habits such as increased consumption of an energy-dense diet and physical inactivity have led to a sharp increase in obesity incidence over recent decades. The exponential increase in obesity is strongly associated with an increased prevalence of type 2 diabetes mellitus (T2DM)1 2 and pre-diabetes3 4 worldwide. According to the International Diabetes Federation, in 2017, approximately 425 million people worldwide are living with diabetes with the number expected to exceed 629 million by the year 2045.5 In the Middle East and North African region, nearly 38.7 million people, or 9.6% of adults aged 20–79 years, have diabetes.5 Among college students and adolescents in Kuwait, the estimated prevalence of pre-diabetes is 6.3% and 33%, respectively.6 7 In the adult population, the WHO STEPS Study estimated the prevalence of pre-diabetes to be 20%8 and diabetes to be 23.1%. In a recently published study, the prevalence of pre-diabetes was estimated at 31% and undiagnosed diabetes at 4.9% among adults in Kuwait.9
Abdominal obesity is strongly associated with several metabolic syndrome components, such as abnormal lipid profile,10 hypertension (HT),11 and T2DM.12 Abdominal obesity indicators include waist-to-hip ratio, abdominal fat, trunk fat, subcutaneous adipose tissue, and visceral adipose tissue (VAT), with VAT being the most vital link with pre-diabetes incidence.13 14 Studies have shown that VAT is strongly associated with insulin resistance,15 16 T2DM, and pre-diabetes.17 VAT cells release proteins that contribute to inflammation, atherosclerosis, dyslipidemia, and HT.18 Furthermore, visceral obesity decreases insulin action via increased delivery of free fatty acids in insulin-sensitive tissues.19 20 Therefore, VAT is a crucial predictor of pre-diabetes and is strongly associated with insulin resistance.15 16 21 Studies assessing associations of VAT with diabetes and pre-diabetes have been mainly conducted on Western and Chinese populations, and such studies on Middle Eastern populations are limited.
Dual-energy X-ray absorptiometry (DXA) is a non-invasive method to measure whole-body composition, trunk, abdominal fat mass, and lean and fat tissue distribution in populations, which helps define diabetes and cardiovascular risk.22 23 The recent introduction of an algorithm that can estimate VAT from the abdominal DXA scans has made it easier to collect visceral adiposity data in extensive population-based studies, which is more cost-effective than the gold standards, CT and MRI, where the cost of running the scans is substantially higher.22 23 In addition, recent validation studies showed that the DXA-derived VAT is highly correlated with the VAT estimated from CT and MRI.24
In Kuwait, overweight and obesity affect 74.5% and 84.3% of men and women aged ≥20 years, respectively.25 Since pre-diabetes is a high-risk state for developing T2DM, metabolic syndrome, and other complications, it is essential to ascertain parameters that can identify high-risk individuals and subgroups within a population. Therefore, in this population-based cross-sectional study, we sought to estimate the prevalence of pre-diabetes and assess associations between different adiposity indices measured by DXA, such as VAT, total body fat (TBF) mass, android fat, gynoid fat, and android-to-gynoid fat ratio, with the prevalence of pre-diabetes in a sample of Kuwaiti adults. Moreover, we have stratified our analysis by sex since males and females have different body fat distribution and scientific evidence suggests that sex dimorphism in the effect of adiposity on diabetes/metabolic syndrome risk exists.
Methods
Study participants
This population-based cross-sectional study enrolled 1240 adults between 18 and 65 years who participated in the Kuwait Wellbeing Study. The exclusion criteria included pregnancy, known diabetes, inability to walk unaided, psychosis, or terminal illness. All measurements on each individual were performed on the same day. In the current analysis, we excluded participants with undiagnosed diabetes (ie, no history of doctor-diagnosed diabetes and measured HbA1c (glycated hemoglobin) ≥6.5% (48 mmol/mol); n=56). Hence, of the total enrolled study sample, 95.5% (1184 of 1240) satisfied our inclusion criteria.
Anthropometric measurements
Trained nurses performed anthropometry measurements. Weight was measured using a calibrated scale (TANITA model BC-418 MA, Tokyo, Japan) and recorded to the nearest 0.1 kg. Height was measured by a wall-mounted stadiometer (SECA model 240, Birmingham, UK) and recorded to the nearest 0.1 cm. Both height and weight were measured without shoes and in light clothing in a standardized manner. Body mass index (BMI) was calculated as the weight in kilograms divided by the height in meters squared (kg/m2). Waist and hip circumferences were measured with a D-loop measurement tape at the midpoint between the lowest rib margin and the iliac crest and the widest level over the greater trochanters, respectively, with both measurements recorded to the nearest 0.1 cm.
DXA measurements
Total body imaging was acquired using the Lunar iDXA (GE Healthcare, Bedford, UK). Estimates of TBF mass, android fat mass, gynoid fat mass, and VAT content (mass and volume) were derived using the iDXA enCORE software (V.14.10.022; GE Healthcare). The software estimates the VAT content within the android region; the software automatically places a quadrilateral box, that is, the region outlined by the iliac crest, with a superior height equivalent to 20% of the distance from the top of the iliac crest to the base of the skull.
Daily quality assurance and quality controls were carried out during the study period before using the equipment according to standard procedures supplied by the manufacturer. The volunteers were scanned using standard imaging and positioning protocols by trained operators.26
Blood test and pre-diabetes ascertainment
Participants were asked to fast overnight for at least 10 hours. Fasting blood samples were collected upon arrival at the research unit, and HbA1c was determined using the VarianT M device (BioRad, Hercules, California, USA). Among individuals with no history of DM diagnosis and/or no history of using DM pharmacological treatments, pre-diabetes was defined as HbA1c levels of 5.7–6.4% (39–47 mmol/mol), according to the American Diabetes Association 2015 guidelines.27 Participants with HbA1c levels ≥6.5% (48 mmol/mol) were excluded from the analysis.
Statistical analysis
All statistical analyses were conducted using SAS V.9.4 (SAS Institute). The statistical significance level was set to α=0.05 for all association analyses. Descriptive analyses were conducted to estimate frequencies and proportions of categorical variables and medians and 5th and 95th percentiles of continuous variables. Prevalence of pre-diabetes, in the total population and stratified by sex, was estimated along with their binomial 95% CI. The Wilcoxon rank-sum test was used to compare the medians of quantitative variables across two groups. Correlations between body fat composition measures were tested for males and females separately using Spearman correlation coefficients.
Associations between anthropometric and DXA-derived measures of body composition with pre-diabetes status were evaluated in the total sample and stratified by sex. BMI (kg/m2) was categorized as follows: underweight/normal weight (<25.0), overweight (25.0–29.9), and obese (≥30.0). On the other hand, waist-to-hip ratio, TBF mass (kg), android fat mass (kg), gynoid fat mass (kg), VAT (kg), and android-to-gynoid fat ratio were categorized into quartiles, with the first quartile set as the reference group. To account for sex differences, quartiles of body fat composition measures were determined separately for males and females. Two analytical approaches were applied to assess the associations between the measures of body composition and pre-diabetes prevalence: the variables were treated as (1) categorical and (2) quantitative to infer the linear trends. To determine if sex is an effect modifier, statistical interactions on multiplicative scale were evaluated by including product terms (sex×BMI/waist-to-hip ratio/TBF/AF/GF/AF:GF ratio/VAT) in the regression models. Given that statistical power to detect higher-order terms is usually limited in epidemiological studies,28 interaction term p value (Pinteraction) <0.2 was considered as a ‘possible’ statistical suggestion for interaction (effect modification). Adjusted prevalence ratios (aPRs) and their 95% CIs were estimated by applying a modified Poisson regression with robust variance estimation using the GENMOD procedure in SAS V.9.4.29 In the analysis evaluating the associations in the total study sample (ie, including males and females), PRs were adjusted for sex and age. In the sex-stratified analysis, PRs were adjusted for age.
Moreover, logistic regression models were fitted to assess the discrimination ability of the different measures of body composition to accurately identify subjects with and without pre-diabetes. Receiver operating characteristics (ROC) curves, plotting sensitivity (true positives) by 1–specificity (false positives) at different cut-off points, were used to assess the discriminatory performance of models with different predictors. The area under the ROC curve (AUC) was estimated, providing a quantitative summary measure of the discriminatory performance of the fitted models. A model with an AUC of 0.5 has no discrimination value, whereas a model with an AUC of 1.0 has perfect discrimination.30 AUC and their 95% CIs were estimated for each assessed body composition measure. In addition, to test differences between the estimated AUC values, the estimated AUCs of the different measures of body fat were compared with the estimated AUC of BMI (reference measure) using the roccontrast statement in the LOGISTIC procedure in SAS V.9.4.
Results
The general characteristics of the study participants are shown in table 1. In regard to the total analytical sample (n=1184), 49.4% (n=585) were males and 50.6% (n=599) were females. The median age of the study participants was 44.0 years (5th, 95th percentile: 23.5, 60.0 years). The prevalence of pre-diabetes was estimated to be 32.6% (355 of 1088; 95% CI: 29.9% to 35.5%) in the total study sample, with males having higher pre-diabetes prevalence (35.6%, 187 of 525; 95% CI: 31.5% to 40.0%) compared with females (29.8%, 168 of 563; 95% CI: 26.1% to 33.8%; p=0.042).
Table 1.
Characteristics, anthropometry, and dual-energy X-ray absorptiometry measurements of the Kuwait Wellbeing Study sample
| Total (n=1184) |
Males (n=585) |
Females (n=599) |
P value* | |
| Age (years), % (n) | ||||
| Median (5th, 95th percentile) | 44.0 (23.5, 60.0) | 42.0 (25.0, 60.0) | 45.0 (23.0, 61.0) | <0.001 |
| ≤30 | 16.5 (195) | 19.1 (112) | 13.9 (83) | |
| 31–40 | 25.9 (307) | 29.7 (174) | 22.2 (133) | |
| 41–50 | 30.1 (356) | 27.9 (163) | 32.2 (193) | |
| ≥51 | 27.5 (326) | 23.3 (136) | 31.7 (190) | |
| Pre-diabetes, % (n) | ||||
| Yes (5.7≤HbA1c%≤6.4) | 32.6 (355) | 35.6 (187) | 29.8 (168) | 0.042 |
| Missing, (n) | (96) | (60) | (36) | |
| BMI (kg/m2), % (n) | ||||
| Median (5th, 95th percentile) | 28.2 (21.3, 39.6) | 28.2 (21.9, 39.3) | 28.2 (20.9, 39.9) | 0.698 |
| Normal† (<25.0) | 20.1 (238) | 18.4 (108) | 21.7 (130) | |
| Overweight (25.0–29.9) | 44.4 (526) | 47.2 (276) | 41.7 (250) | |
| Obese (≥30.0) | 35.5 (420) | 34.4 (201) | 36.6 (219) | |
| Waist-to-hip ratio, median (5th, 95th percentile) | ||||
| Overall | 0.89 (0.75, 1.02) | 0.93 (0.82, 1.04) | 0.84 (0.73, 0.96) | <0.001 |
| Quartile 1 | 0.78 (0.71, 0.89) | 0.86 (0.77, 0.89) | 0.76 (0.71, 0.79) | |
| Quartile 2 | 0.86 (0.79, 0.93) | 0.92 (0.90, 0.93) | 0.81 (0.79, 0.83) | |
| Quartile 3 | 0.91 (0.84, 0.96) | 0.95 (0.93, 0.97) | 0.86 (0.84, 0.88) | |
| Quartile 4 | 0.98 (0.89, 1.08) | 1.01 (0.97, 1.11) | 0.92 (0.89, 0.99) | |
| Total body fat mass (kg), median (5th, 95th percentile) | ||||
| Overall | 30.2 (16.4, 50.4) | 28.4 (13.4, 50.4) | 31.6 (17.8, 50.5) | <0.001 |
| Quartile 1 | 20.6 (11.3, 25.4) | 19.3 (9.0, 22.9) | 21.9 (13.8, 25.9) | |
| Quartile 2 | 27.1 (23.5, 31.1) | 25.3 (23.4, 28.3) | 29.1 (26.7, 31.3) | |
| Quartile 3 | 33.1 (28.9, 37.6) | 31.6 (28.6, 35.1) | 34.5 (31.8, 38.0) | |
| Quartile 4 | 43.5 (36.6, 59.8) | 41.1 (36.1, 59.7) | 44.4 (38.9, 60.0) | |
| Android fat (kg), median (5th, 95th percentile) | ||||
| Overall | 2.57 (0.91, 5.06) | 2.82 (0.89, 5.62) | 2.37 (0.93, 4.43) | <0.001 |
| Quartile 1 | 1.39 (0.52, 1.96) | 1.46 (0.45, 1.99) | 1.36 (0.60, 1.75) | |
| Quartile 2 | 2.22 (1.87, 2.74) | 2.42 (2.06, 2.77) | 2.08 (1.82, 2.35) | |
| Quartile 3 | 2.95 (2.42, 3.59) | 3.23 (2.86, 3.64) | 2.66 (2.40, 3.04) | |
| Quartile 4 | 4.09 (3.22, 6.34) | 4.55 (3.78, 6.68) | 3.73 (3.18, 5.66) | |
| Gynoid fat (kg), median (5th, 95th percentile) | ||||
| Overall | 4.99 (2.75, 8.53) | 4.32 (2.19, 7.69) | 5.67 (3.40, 7.69) | <0.001 |
| Quartile 1 | 3.38 (1.72, 4.53) | 3.04 (1.40, 3.52) | 4.01 (2.75, 4.59) | |
| Quartile 2 | 4.31 (3.61, 5.59) | 3.89 (3.59, 4.28) | 5.30 (4.73, 5.64) | |
| Quartile 3 | 5.42 (4.39, 6.61) | 4.80 (4.35, 5.33) | 6.20 (5.70, 6.65) | |
| Quartile 4 | 7.26 (5.62, 9.91) | 6.39 (5.52, 9.87) | 7.84 (6.81, 9.95) | |
| Android-to-gynoid fat ratio, median (5th, 95th percentile) | ||||
| Overall | 0.50 (0.28, 0.85) | 0.63 (0.35, 0.92) | 0.41 (0.24, 0.64) | <0.001 |
| Quartile 1 | 0.33 (0.21, 0.51) | 0.41 (0.28, 0.52) | 0.30 (0.18, 0.34) | |
| Quartile 2 | 0.52 (0.35, 0.62) | 0.58 (0.53, 0.62) | 0.38 (0.34, 0.41) | |
| Quartile 3 | 0.63 (0.42, 0.73) | 0.68 (0.63, 0.73) | 0.45 (0.41, 0.48) | |
| Quartile 4 | 0.75 (0.50, 1.02) | 0.83 (0.75, 1.11) | 0.55 (0.49, 0.77) | |
| Visceral adipose tissue (kg), median (5th, 95th percentile) | ||||
| Overall | 0.92 (0.15, 2.47) | 1.34 (0.19, 3.03) | 0.67 (0.11, 1.73) | <0.001 |
| Quartile 1 | 0.29 (0.06, 0.71) | 0.41 (0.08, 0.74) | 0.24 (0.04, 0.39) | |
| Quartile 2 | 0.77 (0.42, 1.29) | 1.06 (0.81, 1.32) | 0.52 (0.41, 0.66) | |
| Quartile 3 | 1.34 (0.71, 1.87) | 1.59 (1.37, 1.90) | 0.83 (0.69, 1.01) | |
| Quartile 4 | 2.00 (1.09, 3.43) | 2.30 (1.98, 3.94) | 1.35 (1.05, 2.26) | |
*P values calculated using Wilcoxon rank-sum tests to compare median values of the respective variable between males and females.
†Participants in the underweight groups (BMI <18.5) were analyzed with the normal group due to a small proportion (1.3%, n=16) being underweight.
BMI, body mass index; HbA1c, glycated hemoglobin.
The median BMI value of the total study sample was estimated to be 28.2 (5th, 95th percentile: 21.3, 39.6) kg/m2, with no difference between males and females. Based on BMI classification, of the total study sample, 44.4% (n=526) were overweight and 35.5% (n=420) were obese (table 1). The median waist-to-hip ratio, android fat mass, android-to-gynoid fat ratio, and VAT were higher among males than females (p<0.001, table 1). Females presented a higher TBF mass and gynoid fat mass than males (p<0.001, table 1).
Results of the correlation analyses between body fat composition measures are presented in online supplemental table 1 (males) and online supplemental table 2 (females). Overall, the magnitude and direction of correlations were similar in males and females. Although gynoid fat showed similar direction and magnitude of correlation with other body fat composition measures among males and females, an exception is the correlation between gynoid fat and waist-to-hip ratio, which showed a positive correlation in males (r=0.35, p<0.001; online supplemental table 1) and no correlation in females (r=−0.05, p=0.194; online supplemental table 2).
bmjdrc-2023-003412supp001.pdf (19.7KB, pdf)
Associations between measures of body composition with pre-diabetes in the total study sample and stratified by sex are presented in figure 1 and table 2. An increasing pattern in the prevalence of pre-diabetes across BMI categories was demonstrated, with pre-diabetes being more prevalent among individuals in the obese category compared with those in the underweight/normal weight group (aPR=1.59, 95% CI: 1.19 to 2.12; figure 1A). Although the waist-to-hip ratio was positively associated with pre-diabetes, this association did not gain statistical significance (figure 1B). Increasing patterns in the prevalence of pre-diabetes were seen across quartiles of TBF mass (figure 1C), android fat mass (figure 1D), gynoid fat mass (figure 1E), android-to-gynoid fat ratio (figure 1F), and VAT (figure 1G). For instance, pre-diabetes prevalence was increased among participants in the highest quartile of VAT compared with those in the lowest quartile (aPRQ4 vs Q1=2.05, 95% CI: 1.49 to 2.82; figure 1G).
Figure 1.
Associations between general and abdominal adiposity indicators with pre-diabetes prevalence defined according to HbA1c in the total study sample and stratified by sex. (A) Body mass index is categorized by normal, overweight, and obese status, whereas (B) waist-to-hip ratio, (C) total body fat mass, (D) android fat, (E) gynoid fat, (F) android-to-gynoid fat ratio, and (G) visceral fat are ranked by quartiles. P values for linear trends are presented. Interaction term (sex×body fat measure) p values (Pinteraction) are presented to assess sex-related differences in the evaluated associations. Adjusted prevalence ratios (aPRs) along with their 95% CIs are presented. In the analysis evaluating the associations in the total study sample (ie, including men and women), PRs were adjusted for sex and age. In the sex-stratified analysis, PRs were adjusted for age. HbA1c, glycated hemoglobin.
Table 2.
Associations of measures of general adiposity and abdominal adiposity with pre-diabetes prevalence defined according to HbA1c in the total study sample and stratified by sex
| Total sample | Males | Females | ||||
| Pre-diabetes, % (n/total) |
aPR* (95% CI) | Pre-diabetes, % (n/total) |
aPR† (95% CI) | Pre-diabetes, % (n/total) |
aPR† (95% CI) | |
| Body mass index | ||||||
| Normal† | 19.5 (42/215) | 1.00 (ref) | 20.4 (19/93) | 1.00 (ref) | 18.9 (23/122) | 1.00 (ref) |
| Overweight | 32.0 (158/494) | 1.33 (0.99, 1.78) | 36.7 (94/256) | 1.57 (1.02, 2.42) | 26.9 (64/238) | 1.10 (0.74, 1.66) |
| Obese | 41.5 (152/366) | 1.59 (1.19, 2.12) | 42.4 (73/172) | 1.75 (1.13, 2.70) | 40.7 (79/194) | 1.45 (0.99, 2.14) |
| Ptrend | – | 0.019 | – | 0.033 | – | 0.042 |
| Waist:hip ratio | ||||||
| Quartile 1 | 23.5 (65/277) | 1.00 (ref) | 25.4 (34/134) | 1.00 (ref) | 21.7 (31/143) | 1.00 (ref) |
| Quartile 2 | 27.5 (76/276) | 0.96 (0.73, 1.26) | 29.9 (41/137) | 0.99 (0.68, 1.45) | 25.2 (35/139) | 0.92 (0.62, 1.36) |
| Quartile 3 | 37.8 (101/267) | 1.20 (0.93, 1.55) | 42.0 (55/131) | 1.31 (0.91, 1.87) | 33.8 (46/136) | 1.09 (0.76, 1.58) |
| Quartile 4 | 43.5 (110/253) | 1.25 (0.96, 1.61) | 47.5 (56/118) | 1.24 (0.85, 1.80) | 40.0 (54/135) | 1.24 (0.87, 1.77) |
| Ptrend | – | 0.021 | – | 0.098 | – | 0.087 |
| Total body fat mass | ||||||
| Quartile 1 | 19.4 (54/279) | 1.00 (ref) | 20.2 (27/134) | 1.00 (ref) | 18.6 (27/145) | 1.00 (ref) |
| Quartile 2 | 23.1 (93/281) | 1.36 (1.02, 1.81) | 39.0 (53/136) | 1.59 (1.07, 2.36) | 27.6 (40/145) | 1.13 (0.75, 1.72) |
| Quartile 3 | 38.6 (103/267) | 1.60 (1.22, 2.11) | 39.7 (52/131) | 1.64 (1.11, 2.43) | 37.5 (51/136) | 1.57 (1.07, 2.29) |
| Quartile 4 | 41.1 (102/248) | 1.58 (1.20, 2.07) | 45.0 (54/120) | 1.82 (1.23, 2.69) | 37.5 (48/128) | 1.35 (0.92, 1.98) |
| Ptrend | – | <0.001 | – | 0.003 | – | 0.042 |
| Android fat | ||||||
| Quartile 1 | 18.8 (53/282) | 1.00 (ref) | 21.3 (29/136) | 1.00 (ref) | 16.4 (24/146) | 1.00 (ref) |
| Quartile 2 | 30.2 (83/275) | 1.29 (0.96, 1.73) | 35.3 (47/133) | 1.35 (0.91, 1.99) | 25.4 (36/142) | 1.22 (0.78, 1.90) |
| Quartile 3 | 38.8 (106/273) | 1.55 (1.17, 2.05) | 41.0 (55/134) | 1.45 (0.99, 2.14) | 36.7 (51/139) | 1.67 (1.11, 2.51) |
| Quartile 4 | 44.9 (110/245) | 1.67 (1.27, 2.20) | 46.6 (55/118) | 1.67 (1.13, 2.41) | 43.3 (55/127) | 1.69 (1.13, 2.53) |
| Ptrend | – | <0.001 | – | 0.008 | – | 0.002 |
| Gynoid fat | ||||||
| Quartile 1 | 24.3 (66/272) | 1.00 (ref) | 25.2 (33/131) | 1.00 (ref) | 23.4 (33/141) | 1.00 (ref) |
| Quartile 2 | 35.5 (99/279) | 1.37 (1.07, 1.76) | 38.2 (52/136) | 1.46 (1.03, 2.07) | 32.9 (47/143) | 1.28 (0.90, 1.83) |
| Quartile 3 | 31.2 (83/266) | 1.17 (0.90, 1.53) | 34.9 (45/129) | 1.36 (0.94, 1.96) | 27.7 (38/137) | 0.99 (0.68, 1.46) |
| Quartile 4 | 40.3 (104/258) | 1.48 (1.16, 1.89) | 44.8 (56/125) | 1.71 (1.22, 2.41) | 36.1 (48/133) | 1.27 (0.90, 1.78) |
| Ptrend | – | 0.008 | – | 0.004 | – | 0.425 |
| Android-to-gynoid fat ratio | ||||||
| Quartile 1 | 18.4 (51/277) | 1.00 (ref) | 20.2 (27/134) | 1.00 (ref) | 16.8 (24/143) | 1.00 (ref) |
| Quartile 2 | 28.5 (80/281) | 1.26 (0.93, 1.70) | 33.8 (47/139) | 1.39 (0.93, 2.09) | 23.2 (33/142) | 1.10 (0.70, 1.74) |
| Quartile 3 | 36.5 (99/271) | 1.45 (1.08, 1.95) | 41.5 (54/130) | 1.53 (1.02, 2.30) | 31.9 (45/141) | 1.36 (0.88, 2.08) |
| Quartile 4 | 49.6 (122/246) | 1.70 (1.27, 2.28) | 49.2 (58/118) | 1.60 (1.06, 2.41) | 50.0 (64/128) | 1.79 (1.19, 2.70) |
| Ptrend | – | <0.001 | – | 0.025 | – | 0.001 |
| Visceral adipose tissue | ||||||
| Quartile 1 | 15.8 (44/279) | 1.00 (ref) | 16.9 (23/136) | 1.00 (ref) | 14.7 (21/143) | 1.00 (ref) |
| Quartile 2 | 24.7 (69/279) | 1.23 (0.87, 1.73) | 28.4 (38/134) | 1.41 (0.89, 2.24) | 21.4 (31/145) | 1.05 (0.64, 1.74) |
| Quartile 3 | 41.6 (114/274) | 1.83 (1.33, 2.51) | 44.5 (61/137) | 1.98 (1.27, 3.08) | 38.7 (53/137) | 1.68 (1.07, 2.63) |
| Quartile 4 | 51.4 (125/243) | 2.05 (1.49, 2.82) | 56.1 (64/114) | 2.29 (1.46, 3.59) | 47.3 (61/129) | 1.81 (1.15, 2.84) |
| Ptrend | – | <0.001 | – | <0.001 | – | <0.001 |
*Adjusted for age and sex.
†Adjusted for age.
aPR, adjusted prevalence ratio; HbA1c, glycated hemoglobin.
The measures mentioned above of body composition were associated with pre-diabetes similarly in males and females, with the exception of gynoid fat (Pinteraction=0.164, figure 1E), android-to-gynoid fat ratio (Pinteraction=0.034, figure 1F), and VAT (Pinteraction=0.094, figure 1G), which showed possible sex-related differences. In sex-stratified analysis, increased gynoid fat mass showed an association with pre-diabetes prevalence only among males (aPRQ4 vs Q1=1.71, 95% CI: 1.22 to 2.41), but not among females (aPRQ4 vs Q1=1.27, 95% CI: 0.90 to 1.78; figure 1E). Moreover, given the statistically significant ‘sex×android-to-gynoid fat ratio’ interaction term (Pinteraction=0.034, figure 1F), the association between android-to-gynoid fat ratio and pre-diabetes is stronger among females than males. In terms of VAT, the association between VAT and pre-diabetes may be stronger in males than in females (sex×VAT: Pinteraction=0.094, figure 1G).
In the total study sample, the ROC analysis predicting pre-diabetes showed that VAT measurement had the highest estimated AUC of 0.680 (95% CI: 0.646 to 0.714, table 3), followed by the android-to-gynoid fat ratio (AUC=0.647, 95% CI: 0.613 to 0.682) and android fat (AUC=0.646, 95% CI: 0.612 to 0.680). Compared with an estimated BMI-AUC of 0.611, android fat (difference in AUC: 0.035, p=0.001) and VAT (difference in AUC: 0.069, p<0.001) were better predictors of pre-diabetes. The gynoid fat compared with BMI had lower estimated AUC for predicting pre-diabetes (difference in AUC: −0.066, p<0.001). In predicting pre-diabetes, VAT AUC compared with android fat AUC was higher (difference in AUC: 0.034, 95% CI: 0.013 to 0.055, p=0.001). Moreover, VAT AUC compared with android-to-gynoid fat ratio AUC was higher (difference in AUC: 0.033, 95% CI: 0.013 to 0.052, p=0.002). There was no difference in the AUC of android-to-gynoid fat ratio and android fat (difference in AUC: 0.001, 95% CI: −0.027 to 0.029, p=0.914).
Table 3.
Estimates of area under the receiver operating characteristics curve (AUC) of body fat composition measurements for predicting pre-diabetes in the total study sample and stratified by sex
| AUC (95% CI) | Difference in AUC (95% CI) | P value | |
| Total sample | |||
| Body mass index | 0.611 (0.576, 0.646) | 0.00 (reference) | – |
| Waist-to-hip ratio | 0.606 (0.570, 0.642) | −0.005 (−0.049, 0.039) | 0.832 |
| Total body fat mass | 0.603 (0.568, 0.638) | −0.008 (−0.026, 0.010) | 0.400 |
| Android fat | 0.646 (0.612, 0.680) | 0.035 (0.014, 0.055) | 0.001 |
| Gynoid fat | 0.545 (0.509, 0.581) | −0.066 (−0.094, −0.038) | <0.001 |
| Android-to-gynoid fat ratio | 0.647 (0.613, 0.682) | 0.036 (−0.003, 0.075) | 0.069 |
| Visceral adipose tissue | 0.680 (0.646, 0.714) | 0.069 (0.037, 0.100) | <0.001 |
| Males | |||
| Body mass index | 0.592 (0.542, 0.642) | 0.00 (reference) | – |
| Waist-to-hip ratio | 0.620 (0.570, 0.670) | 0.028 (−0.020, 0.077) | 0.253 |
| Total body fat mass | 0.610 (0.561, 0.659) | 0.018 (−0.009, 0.046) | 0.187 |
| Android fat | 0.632 (0.584, 0.681) | 0.041 (0.013, 0.069) | 0.004 |
| Gynoid fat | 0.573 (0.523, 0.624) | −0.018 (−0.054, 0.018) | 0.319 |
| Android-to-gynoid fat ratio | 0.640 (0.591, 0.689) | 0.048 (−0.002, 0.098) | 0.059 |
| Visceral adipose tissue | 0.694 (0.646, 0.741) | 0.102 (0.063, 0.141) | <0.001 |
| Females | |||
| Body mass index | 0.628 (0.579, 0.678) | 0.00 (reference) | – |
| Waist-to-hip ratio | 0.597 (0.545, 0.648) | −0.032 (−0.098, 0.035) | 0.353 |
| Total body fat mass | 0.616 (0.566, 0.666) | −0.012 (−0.034, 0.010) | 0.289 |
| Android fat | 0.651 (0.602, 0.699) | 0.023 (−0.005, 0.050) | 0.108 |
| Gynoid fat | 0.561 (0.509, 0.613) | −0.068 (−0.104, −0.032) | <0.001 |
| Android-to-gynoid fat ratio | 0.674 (0.624, 0.723) | 0.046 (−0.007, 0.098) | 0.087 |
| Visceral adipose tissue | 0.684 (0.636, 0.731) | 0.056 (0.015, 0.096) | 0.007 |
Among both males and females, VAT, android-to-gynoid fat ratio, and android fat were the best predictors of pre-diabetes. Among males, VAT (difference in AUC: 0.102, p<0.001) and android fat (difference in AUC: 0.041, p=0.004; table 3) had higher estimated AUC than BMI. Among females, compared with BMI, VAT had higher AUC (difference in AUC: 0.056, p=0.007) and gynoid fat had lower AUC (difference in AUC: −0.068, p<0.001; table 3).
Discussion
In the present study, we evaluated associations between anthropometric and DXA-derived measures of body composition with pre-diabetes in a sample of Kuwaiti adults. Body composition measures that showed associations with pre-diabetes included BMI, TBF mass, android fat mass, gynoid fat mass, android-to-gynoid fat ratio, and VAT. The aforementioned adiposity indices showed similar associations with pre-diabetes prevalence across males and females, except for gynoid fat, android-to-gynoid fat ratio, and VAT. Moreover, ROC analysis showed that VAT was the best predictor of pre-diabetes in terms of the estimated AUC, followed by android-to-gynoid fat ratio and android fat mass compared with BMI.
BMI has been extensively assessed in population-based studies as an indicator of general adiposity due to its ease of ascertainment and as a predictor of pre-diabetes status.31 32 In the current study, we observed association between BMI categories and pre-diabetes. However, as the BMI categories increased from overweight (aPR=1.33 CI: 0.99 to 1.78) to obese (aPR=1.59, 95% CI: 1.19 to 2.12), a higher prevalence of pre-diabetes was observed. In addition, the estimated average BMI in the total population was 28.2 kg/m2, and obesity prevalence was similar for males (42.4%) and females (40.7%). Another study on the Kuwaiti population had a median BMI of 28.4 kg/m2 among males and 29.1 kg/m2 among females, respectively.33 However, pre-diabetes prevalence was 20.9% for overweight and 22% for obese, compared with 32% for overweight and 41.5% of obese were pre-diabetic in the current study. In the United Arab Emirates (UAE), the prevalence of pre-diabetes among UAE nationals was 16% for both males and females; however, 33% were overweight and 49% obese.34 Since BMI is influenced by multiple body compositions, including fat mass, lean tissue, and bone mass, BMI presents certain shortcomings when predicting diabetes and pre-diabetes status.35
On the other hand, studies have shown that the waist-to-hip ratio is more strongly associated with the development of pre-diabetes and diabetes than BMI36 since the waist-to-hip ratio measures abdominal obesity, which has been proven to associate with metabolic syndromes. In a Spanish cohort, waist-to-hip ratio presented a stronger association with pre-diabetes than BMI.11 In the current study, the waist-to-hip ratio did not demonstrate a strong association with pre-diabetes prevalence in both males and females (total sample: aPRQ4 vs Q1=1.25, 95% CI: 0.96 to 1.61). These results contradict other studies that show the waist-to-hip ratio has a stronger association with pre-diabetes than BMI.
In recent years, DXA has been used as an alternative to standard epidemiological estimation techniques of obesity indices such as BMI and waist-to-hip ratio. However, waist-to-hip ratio addresses abdominal fat, including any muscle mass in the waist region. On the other hand, DXA measures regional adiposity more accurately than conventional anthropometry; in addition, the benefit of DXA is the dissection of fat compartments, including TBF, android fat, gynoid fat, and more recently, VAT.22 37 Therefore, it is advantageous in understanding the possible causes of pre-diabetes concerning the different fat compartments. In addition, compared with waist-to-hip ratio and BMI, studies have shown that DXA demonstrated more accurate cardiovascular risk estimates concerning total and regional adiposity.24 38 39 TBF is highly associated with diabetes and pre-diabetes.40 In our results (figure 1), we observed a more apparent association between TBF and a higher prevalence of pre-diabetes than observed with waist-to-hip ratio but not BMI. Compared with BMI, TBF measures the TBF and does not consider muscle and bone mass in the calculation.41
In comparison with general obesity indices (BMI and TBF), we used the ability of DXA to differentiate between various fat depots to measure android fat, gynoid fat, android-to-gynoid fat ratio, and VAT to assess which fat compartment is a better predictor of pre-diabetes. Android fat, which is measured in the abdominal region and can be described as interabdominal fat VAT,42 is generally accepted as an important risk factor for insulin resistance43; whereas gynoid fat (lower body adiposity) may lower that risk.44 In our study, android fat mass showed strong associations with pre-diabetes prevalence, and the association was similar in males and females. Interestingly, gynoid fat mass was positively associated with pre-diabetes prevalence among males but not females. The android fat PRs showed a similar association with pre-diabetes along with the four quartiles in comparison with TBF. As for gynoid fat, in general, there was a weaker association with and lower prediction (ie, AUC) of pre-diabetes in comparison with android fat, which was corroborated in a study where they found the DXA-measured android fat was highly associated with cardiovascular disease and T2DM, and showed a stronger correlation with impaired fasting glucose than gynoid fat.39 Nonetheless, in males, the gynoid fat showed similar associations with pre-diabetes to android fat, whereas in females, the association was weaker, indicating that gynoid fat is not an important risk factor for pre-diabetes in females.39 45
The presence of an association between gynoid fat and pre-diabetes in males, and the absence of such an association among females could be due to the differential correlation between gynoid fat and wait-to-hip ratio measure in males (r=0.35, p<0.001; online supplemental table 1) and females (r=−0.05, p=0.194; online supplemental table 2). Accordingly, we re-evaluated the association between gynoid fat and pre-diabetes while adjusting for the effect of waist-to-hip ratio in males and females (data not shown). Adjusting for the effect of waist-to-hip ratio did not attenuate (confound) the observed association between gynoid fat and pre-diabetes among males, nor did it induce an association among females. Hence, the observed positive association between gynoid fat and pre-diabetes in males and the lack of such an association in females is not confounded by waist-to-hip ratio.
Android-to-gynoid fat ratio has demonstrated association with an increased risk of metabolic syndrome in healthy adults.46 In addition, the android-to-gynoid ratio showed a relationship with increased risk of cardiovascular disease.39 Furthermore, from the National Health and Nutrition Examination Survey (NHANES) data, Okosun et al demonstrated that cardiometabolic dysregulation showed a stronger association with android-to-gynoid fat ratio than android fat and BMI in both males and females.47 In the current study, this is further corroborated, whereby the PRs for android-to-gynoid ratios showed a strong association with HbA1c.
Increasing proof suggests that the distribution of fat composition is critical in the development of insulin resistance.48 Studies have shown that high VAT levels increase lipolytic activity and release free fatty acids and adipokines into the portal circulation leading to hepatic insulin resistance.47 49 Therefore, it further highlights the physiological effects high VAT content has on the manifestation of metabolic syndromes. Our results showed that VAT measurements, on average, were higher in both males (1.34 kg) and females (0.67 kg) in comparison with a European cohort study of 0.542±0.451 kg and 0.258±226 kg for males and females, respectively. In addition, at the higher quartiles, Q3 (aPR=1.83 CI: 1.33 to 2.51) and Q4 (aPR=2.05 CI: 1.49 to 2.82), VAT showed strong associations with pre-diabetes. This is further verified in a cross-sectional study, whereby DXA-derived VAT showed a higher OR (10th percentile) for developing pre-diabetes and T2DM.17 Moreover, our AUC analysis showed that VAT is the best predictor of pre-diabetes in both males and females.
Our study had several strengths: a well-characterized study sample, a standardized protocol, and a large sample size. In addition, measuring different adiposity indices is a further strength of our study, which allowed us to investigate associations of different fat depots with pre-diabetes prevalence. Information bias is inevitable in epidemiological studies; however, using objective tests to ascertain pre-diabetes and adiposity indices in the current report helped minimize the effect of misclassification. Moreover, the average BMI of participants in our study (males: 28.2 kg/m2; females: 29.2 kg/m2) is similar to a prior study conducted among Kuwaiti adults (males: 28.4 kg/m2; females: 29.1 kg/m2; n=3589).33 This further indicates that our study sample is representative of the Kuwaiti population in terms of BMI, which is a main risk factor for diabetes and pre-diabetes. A limitation of the current analysis is that we did not exclude subjects with anemia, hemoglobinopathies, and other clinical conditions, which may potentially interfere with HbA1c measurements. Moreover, confounding in observational studies should not be overlooked. We did not observe associations between cigarette smoking and physical activity (ascertained as walking frequency) with body fat composition measures and pre-diabetes (data not shown). Accordingly, smoking and physical activity were not considered as possible confounders in the current analysis. Nonetheless, the assessment of physical activity was based on a single question (ie, how many times do you walk for pleasure (not as a means of transport)?), which might not capture the true physical activity state/nature. Also, information about alcohol use was not obtained in this study, which is a further limitation. It is essential to indicate that our associations represent concurrent (cross-sectional) associations and should not be interpreted as causal associations.
In our study sample of Kuwaiti adults, adiposity measurements using DXA showed strong associations with pre-diabetes. VAT and android-to-gynoid fat ratio were strongly associated with pre-diabetes prevalence in this study. Adiposity measurements using DXA demonstrated improved prediction in terms of AUC analysis for pre-diabetes relative to BMI. In general, we did not identify any pronounced differences in the association between adiposity indices and pre-diabetes in males and females. Nevertheless, increased gynoid fat in males was associated with higher pre-diabetes prevalence, whereas, in females, no association was identified. Moreover, we observed a potential sex-related effect modification of the association between android-to-gynoid fat ratio and VAT with pre-diabetes.
Footnotes
Contributors: AM contributed to conceptualization and design of the study, collected the data, contributed to data analysis and interpretation, and drafted the manuscript. AZ contributed to conceptualization and design of the study, contributed to data analysis and interpretation, and critically reviewed and revised the manuscript for important intellectual content. TM is the principal investigator of the Kuwait Wellbeing Study and responsible for the study design, planning, management, and conduct. TM obtained the funding, contributed to data interpretation, and critically reviewed and revised the manuscript for important intellectual content. All authors have reviewed, revised, and approved the final manuscript. AM is the guarantor for the study, has access to the data and accepts full responsibility for integrity of the data and the accuracy of the data analyses, and controlled the decision to publish.
Funding: This research was supported by a donation from His Highness Shiekh Nasser Al-Mohammad Al-Sabah, funding the Dasman Diabetes Institute (RA-2010-001).
Disclaimer: The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: None declared.
Provenance and peer review: Not commissioned; externally peer reviewed.
Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Data availability statement
Data are available upon reasonable request. Data are available on reasonable request from the corresponding author.
Ethics statements
Patient consent for publication
Not required.
Ethics approval
This study involves human participants. The protocol of the present study was approved by the Ethical Review Board of the Dasman Diabetes Institute, Safat, Kuwait (RA-01-2010). Written informed consent was obtained from study participants.
References
- 1.Kahn SE, Hull RL, Utzschneider KM. Mechanisms linking obesity to insulin resistance and type 2 diabetes. Nature 2006;444:840–6. 10.1038/nature05482 [DOI] [PubMed] [Google Scholar]
- 2.Ziyab AH, Mohammad A, Maclean E, et al. Diabetes: A fast evolving epidemic. Kuwait Med J 2015;47:291–301. [Google Scholar]
- 3.Ramírez-Vélez R, Pérez-Sousa MÁ, González-Ruíz K, et al. Obesity- and lipid-related parameters in the identification of older adults with a high risk of Prediabetes according to the American diabetes Association: an analysis of the 2015 health, well-being, and aging study. Nutrients 2019;11:2654. 10.3390/nu11112654 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Liu L, Gao B, Wang J, et al. Joint Association of body mass index and central obesity with cardiovascular events and all-cause mortality in Prediabetic population: A prospective cohort study. Obes Res Clin Pract 2019;13:453–61. 10.1016/j.orcp.2019.08.004 [DOI] [PubMed] [Google Scholar]
- 5.Federation ID . IDF diabetes Atlas. 2017: 1–150. [Google Scholar]
- 6.Ben Haider NY, Ziyab AH. Prevalence of Prediabetes and its association with obesity among college students in Kuwait: A cross-sectional study. Diabetes Res Clin Pract 2016;119:71–4. 10.1016/j.diabres.2016.07.001 [DOI] [PubMed] [Google Scholar]
- 7.Almari M, Alsaedi S, Mohammad A, et al. Associations of Adiposity and parental diabetes with Prediabetes among adolescents in Kuwait: A cross-sectional study. Pediatr Diabetes 2018;19:1362–9. 10.1111/pedi.12780 [DOI] [PubMed] [Google Scholar]
- 8.Alkandari A, Longenecker JC, Barengo NC, et al. The prevalence of pre-diabetes and diabetes in the Kuwaiti adult population in 2014. Diabetes Res Clin Pract 2018;144:213–23. 10.1016/j.diabres.2018.08.016 [DOI] [PubMed] [Google Scholar]
- 9.Mohammad A, Ziyab AH, Mohammad T. Prevalence of Prediabetes and Undiagnosed diabetes among Kuwaiti adults: A cross-sectional study. Diabetes Metab Syndr Obes 2021;14:2167–76. 10.2147/DMSO.S296848 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Björnson E, Adiels M, Taskinen M-R, et al. Kinetics of plasma triglycerides in abdominal obesity. Curr Opin Lipidol 2017;28:11–8. 10.1097/MOL.0000000000000375 [DOI] [PubMed] [Google Scholar]
- 11.Sangrós FJ, Torrecilla J, Giráldez-García C, et al. Association of general and abdominal obesity with hypertension, Dyslipidemia and Prediabetes in the PREDAPS study. Rev Esp Cardiol (Engl Ed) 2018;71:170–7. 10.1016/j.rec.2017.04.035 [DOI] [PubMed] [Google Scholar]
- 12.Caspard H, Jabbour S, Hammar N, et al. Recent trends in the prevalence of type 2 diabetes and the association with abdominal obesity lead to growing health disparities in the USA: an analysis of the NHANES surveys from 1999 to 2014. Diabetes Obes Metab 2018;20:667–71. 10.1111/dom.13143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mainous AG, Tanner RJ, Jo A, et al. Prevalence of Prediabetes and abdominal obesity among healthy-weight adults: 18-year trend. The Annals of Family Medicine 2016;14:304–10. 10.1370/afm.1946 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Liu PJ, Ma F, Lou HP, et al. Visceral Adiposity index is associated with pre-diabetes and type 2 diabetes mellitus in Chinese adults aged 20-50. Ann Nutr Metab 2016;68:235–43. 10.1159/000446121 [DOI] [PubMed] [Google Scholar]
- 15.Storz C, Heber SD, Rospleszcz S, et al. The role of visceral and subcutaneous Adipose tissue measurements and their ratio by magnetic resonance imaging in subjects with Prediabetes, diabetes and healthy controls from a general population without cardiovascular disease. Br J Radiol 2018;91:20170808. 10.1259/bjr.20170808 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Gu D, Ding Y, Zhao Y, et al. Visceral Adiposity index was a useful Predictor of Prediabetes. Exp Clin Endocrinol Diabetes 2018;126:596–603. 10.1055/s-0043-120440 [DOI] [PubMed] [Google Scholar]
- 17.Jung SH, Ha KH, Kim DJ. Visceral fat mass has stronger associations with diabetes and Prediabetes than other Anthropometric obesity indicators among Korean adults. Yonsei Med J 2016;57:674–80. 10.3349/ymj.2016.57.3.674 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wu H, Ballantyne CM. Metabolic inflammation and insulin resistance in obesity. Circ Res 2020;126:1549–64. 10.1161/CIRCRESAHA.119.315896 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Moreno-Navarrete JM, Serino M, Blasco-Baque V, et al. Gut Microbiota interacts with markers of Adipose tissue Browning, insulin action and plasma acetate in morbid obesity. Mol Nutr Food Res 2018;62:1700721. 10.1002/mnfr.201700721 [DOI] [PubMed] [Google Scholar]
- 20.Hocking S, Samocha-Bonet D, Milner K-L, et al. Adiposity and insulin resistance in humans: the role of the different tissue and cellular lipid depots. Endocr Rev 2013;34:463–500. 10.1210/er.2012-1041 [DOI] [PubMed] [Google Scholar]
- 21.Liu L, Feng J, Zhang G, et al. Visceral Adipose tissue is more strongly associated with insulin resistance than subcutaneous Adipose tissue in Chinese subjects with pre-diabetes. Curr Med Res Opin 2018;34:123–9. 10.1080/03007995.2017.1364226 [DOI] [PubMed] [Google Scholar]
- 22.Kaul S, Rothney MP, Peters DM, et al. Dual-energy X-ray Absorptiometry for Quantification of visceral fat. Obesity 2012;20:1313–8. 10.1038/oby.2011.393 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Sasai H, Brychta RJ, Wood RP, et al. Does visceral fat estimated by dual-energy X-ray Absorptiometry independently predict Cardiometabolic risks in adults J Diabetes Sci Technol 2015;9:917–24. 10.1177/1932296815577424 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Mohammad A, De Lucia Rolfe E, Sleigh A, et al. Validity of visceral Adiposity estimates from DXA against MRI in Kuwaiti men and women. Nutr Diabetes 2017;7:e238. 10.1038/nutd.2016.38 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Collaboration GO . Global, regional and national prevalence of overweight and obesity in children and adults 1980-2013: A systematic analysis. Lancet 2014;384:766–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Rolfe EDL, Sleigh A, Finucane FM, et al. Ultrasound measurements of visceral and subcutaneous abdominal thickness to predict abdominal Adiposity among older men and women. Obesity 2010;18:625–31. 10.1038/oby.2009.309 [DOI] [PubMed] [Google Scholar]
- 27.American Diabetes A. 2) classification and diagnosis of diabetes. Diabetes Care 2015;38(Supplement_1):S8–16. 10.2337/dc15-S005 [DOI] [PubMed] [Google Scholar]
- 28.Greenland S. Interactions in epidemiology: relevance, identification, and estimation. Epidemiology 2009;20:14–7. 10.1097/EDE.0b013e318193e7b5 [DOI] [PubMed] [Google Scholar]
- 29.Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004;159:702–6. 10.1093/aje/kwh090 [DOI] [PubMed] [Google Scholar]
- 30.Bewick V, Cheek L, Ball J. Statistics review 13: receiver operating characteristic curves. Crit Care 2004;8:508–12. 10.1186/cc3000 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hu H, Kawasaki Y, Kuwahara K, et al. Trajectories of body mass index and waist circumference before the onset of diabetes among people with Prediabetes. Clin Nutr 2020;39:2881–8. 10.1016/j.clnu.2019.12.023 [DOI] [PubMed] [Google Scholar]
- 32.Rahmanian K, Shojaei M, Sotoodeh Jahromi A, et al. The association between pre-diabetes with body mass index and marital status in an Iranian urban population. Glob J Health Sci 2015;8:95–101. 10.5539/gjhs.v8n4p95 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Weiderpass E, Botteri E, Longenecker JC, et al. n.d. The prevalence of overweight and obesity in an adult Kuwaiti population in 2014. Front Endocrinol;10. 10.3389/fendo.2019.00449 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hamoudi R, Saheb Sharif-Askari N, Saheb Sharif-Askari F, et al. Prediabetes and diabetes prevalence and risk factors comparison between ethnic groups in the United Arab Emirates. Sci Rep 2019;9:17437. 10.1038/s41598-019-53505-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Clifton EAD, Day FR, De Lucia Rolfe E, et al. Associations between body mass index-related genetic variants and adult body composition: the Fenland cohort study. Int J Obes (Lond) 2017;41:613–9. 10.1038/ijo.2017.11 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Jayawardana R, Ranasinghe P, Sheriff MHR, et al. Waist to height ratio: a better Anthropometric marker of diabetes and Cardio-metabolic risks in South Asian adults. Diabetes Res Clin Pract 2013;99:292–9. 10.1016/j.diabres.2012.12.013 [DOI] [PubMed] [Google Scholar]
- 37.Direk K, Cecelja M, Astle W, et al. The relationship between DXA-based and Anthropometric measures of visceral fat and morbidity in women. BMC Cardiovasc Disord 2013;13:25. 10.1186/1471-2261-13-25 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Achamrah N, Colange G, Delay J, et al. Comparison of body composition assessment by DXA and BIA according to the body mass index: A retrospective study on 3655 measures. PLoS One 2018;13:e0200465. 10.1371/journal.pone.0200465 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Vasan SK, Osmond C, Canoy D, et al. Comparison of regional fat measurements by dual-energy X-ray Absorptiometry and conventional Anthropometry and their association with markers of diabetes and cardiovascular disease risk. Int J Obes (Lond) 2018;42:850–7. 10.1038/ijo.2017.289 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Solanki JD, Makwana AH, Mehta HB, et al. Body composition in type 2 diabetes: change in quality and not just quantity that matters. Int J Prev Med 2015;6:122. 10.4103/2008-7802.172376 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Liu P-Y, Ilich JZ, Brummel-Smith K, et al. New insight into fat, muscle and bone relationship in women: determining the threshold at which body fat assumes negative relationship with bone mineral density. Int J Prev Med 2014;5:1452–63. [PMC free article] [PubMed] [Google Scholar]
- 42.Kang SM, Yoon JW, Ahn HY, et al. Android fat Depot is more closely associated with metabolic syndrome than abdominal visceral fat in elderly people. PLoS One 2011;6:e27694. 10.1371/journal.pone.0027694 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Ormazabal V, Nair S, Elfeky O, et al. Association between insulin resistance and the development of cardiovascular disease. Cardiovasc Diabetol 2018;17:122. 10.1186/s12933-018-0762-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Choi SI, Chung D, Lim JS, et al. Relationship between regional body fat distribution and diabetes mellitus: 2008 to 2010 Korean national health and nutrition examination surveys. Diabetes Metab J 2017;41:51–9. 10.4093/dmj.2017.41.1.51 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Wiklund P, Toss F, Weinehall L, et al. Abdominal and Gynoid fat mass are associated with cardiovascular risk factors in men and women. J Clin Endocrinol Metab 2008;93:4360–6. 10.1210/jc.2008-0804 [DOI] [PubMed] [Google Scholar]
- 46.Samsell L, Regier M, Walton C, et al. Importance of Android/Gynoid fat ratio in predicting metabolic and cardiovascular disease risk in normal weight as well as overweight and obese children. J Obes 2014;2014:846578. 10.1155/2014/846578 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Okosun IS, Seale JP, Lyn R. Commingling effect of Gynoid and Android fat patterns on Cardiometabolic dysregulation in normal weight American adults. Nutr Diabetes 2015;5:e155. 10.1038/nutd.2015.5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Patel P, Abate N. Body fat distribution and insulin resistance. Nutrients 2013;5:2019–27. 10.3390/nu5062019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Kojta I, Chacińska M, Błachnio-Zabielska A. Bioactive lipids, and Adipose tissue inflammation in insulin resistance. Nutrients 2020;12:1305. 10.3390/nu12051305 [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
bmjdrc-2023-003412supp001.pdf (19.7KB, pdf)
Data Availability Statement
Data are available upon reasonable request. Data are available on reasonable request from the corresponding author.