Anthropometric indices as predictors for diabetes risk: epidemiological findings and their implications

It is widely known that obesity confers an increased risk of insulin resistance and Type 2 diabetes mellitus, but uncertainty remains as to the exact nature of the relationship of body fat proportion and regional fat distribution with diabetes mellitus or prediabetic condition. Obesity is a morbid phenotype of excess body fat resulting from an excess energy balance. The direct measurement of body fat proportion is difficult and BMI (kg/m²) has long been used in clinical practice, as well as in epidemiological research. Internationally, a BMI of 25.0–29.9 and 30.0 kg/m² or greater are classified as overweight and obese, respectively. Fat accumulates subcutaneously and intra-abdominally, and the distribution of fat mass differs by sex. Men tend to have fat mass in the upper body, predominantly in intra-abdominal fat tissue, whereas women are more likely to have fat mass in the lower body, such as hips, buttocks and thighs. For the characteristic distribution in men, the android pattern is referred to as central, abdominal or visceral adiposity. Waist-to-hip circumference ratio (WHR) is a typical measure of central obesity. Cross-sectional and prospective studies in the 1980s consistently showed that WHR was related to an increased risk of Type 2 diabetes mellitus more strongly than (and independently of) BMI in the USA and Europe [1,2]. A relatively stronger association of WHR with Type 2 diabetes mellitus, as determined by a 75-g oral glucose-tolerance test, compared with BMI, was also confirmed in a cross-sectional study of middle-aged Japanese men [3]. Unlike BMI, however, operational cutoff points for WHR have not been established.

More than 10 years ago, Lean and colleagues proposed the use of waist circumference (WC) to detect individuals at a high risk of obesity-related diseases owing to being overweight and central fat distribution [4]. In conjunction with popularity of the metabolic syndrome, WC seems to have gained much credit as a measure of abdominal obesity. The metabolic syndrome is a cluster of metabolic abnormalities (e.g., glucose intolerance, hypertension and dyslipidemia) associated with increased risks of Type 2 diabetes mellitus and cardiovascular disease, and abdominal obesity as measured by WC is a primary component of the metabolic syndrome [5]. The International Diabetes Federation has defined different cutoffs for WC as a criterion of central obesity for different ethnicities and sexes (i.e., Europeans: 94 cm in men and 80 cm in women; South Asians and Chinese: 90 cm in men and 80 cm in women; and Japanese: 85 cm in men and 90 cm in women) [6]. The appropriateness of these cutoffs is one issue to be answered, and another concern has been directed to the question of whether WC is superior to WHR or BMI in predicting the risk of Type 2 diabetes mellitus.

A limited literature has provided somewhat inconsistent findings regarding the latter question. In a large prospective study of male health professionals aged 40–75 years in the USA [7], WC and WHR were related to an increased risk of Type 2 diabetes mellitus independently of BMI, but WC was shown to be a better predictor of diabetes risk than WHR. BMI itself was associated with increased risk of diabetes, more strongly than WHR, and WC and BMI were similar with respect to the prediction of diabetes risk in the receiver-operating characteristic (ROC) analysis. In this study, however, anthropometric measurements were self-reported and Pearson’s correlation coefficients between self-reported and measured values were much smaller for WHR (r = 0.69) than for WC (r = 0.95). Thus, direct comparison between WC and WHR is obviously difficult because inaccurate measurements always weaken the true association. On the other hand, in a prospective study of German men and women aged 35–74 years in which anthropometric indices were measured directly [8], both BMI and WC were associated with increased risk of Type 2 diabetes mellitus with almost the same magnitude in strength. Furthermore, additive effects of BMI and WC or BMI and WHR were observed. These prospective studies suggest that WC is a better predictor of Type 2 diatbetes risk than WHR.

The relative importance of the adiposity-related anthropometric indices has also been investigated in cross-sectional studies on the basis of prevalent diabetes mellitus or homeostasis model assessment of insulin resistance (HOMA-IR). The ability to identify prevalent diabetes mellitus, defined as fasting plasma glucose of at least 126 mg/dl and/or hemoglobin A1c of at least 6.5% or treatment for diabetes mellitus, was shown to be the highest for WHR in Japanese men and women with an average age of 45 years [9]. Values of the AUC in the ROC analysis for WHR, WC and BMI were 0.76, 0.66 and 0.58 in men, respectively, and the corresponding values in women were 0.84, 0.77 and 0.73, respectively. Similar findings were reported across different ethnicities in the National Health and Nutrition Examination Survey in the USA, although the difference between WHR and WC was less evident [10]. The AUC in relation to elevated HOMA-IR were almost the same for WC (0.63 in men and 0.66 in women), BMI (0.62 in men and 0.66 in women) and WHR (0.61 in men and 0.65 in women) in another middle-aged Japanese population [11]. Similarly, both WC and BMI were strongly, positively related to HOMA-IR in the Self Defense Forces Health Study in Japan, in which WHR was not included in the analysis [12].

Findings from cross-sectional studies demonstrating that WHR is related to diabetes mellitus more strongly than or at least equally to WC are in disagreement with the prospective findings that WHR was least strongly related to diabetes risk. The discrepancy may be ascribed to weaknesses of cross-sectional studies, such as reverse causality and attrition bias. It is, however, unlikely that the presence of diabetes mellitus affected WHR alone. Diagnostic bias is more likely to have occurred in the reported prospective studies because the majority of nondiabetic subjects probably did not receive laboratory screening for diabetes mellitus. On the other hand, the similarity in the associations of WHR, WC and BMI with HOMA-IR is notable. In the earlier study showing a stronger association between WHR and Type 2 diabetes [3], BMI was as strongly associated with impaired glucose tolerance as WHR. Overall adiposity may be linked to low-to-moderate levels of insulin resistance, and abdominal adiposity may be more closely involved in the advance of abnormalities in glucose metabolism. Abdominal obesity results in enhanced portal flow of free fatty acids and inflammatory cytokines into the liver from omental adipocytes, which further deteriorates glucose metabolism, probably through oxidative damage of pancreatic islet cells and disturbed hepatic function [5].

In conclusion, although WC and WHR are regarded as a measure of abdominal obesity, they have shown different profiles as predictors of diabetes risk in prospective and cross-sectional studies, causing difficulty in interpretation. Similar associations of WHR, WC and BMI with HOMA-IR or a prediabetic condition support a hypothesis that abdominal adiposity is more closely linked to the enhanced deterioration of glucose metabolism.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.