Review Article
Indices of abdominal obesity are better discriminators of cardiovascular risk factors than BMI: a meta-analysis

https://doi.org/10.1016/j.jclinepi.2007.08.012Get rights and content

Abstract

Objective

To determine which simple index of overweight and obesity is the best discriminator of cardiovascular risk factors.

Study Design and Setting

This is a meta-analysis of published literature. MEDLINE was searched. Studies that used receiver-operating characteristics (ROC) curve analysis and published area under the ROC curves (AUC) for overweight and obesity indices with hypertension, type-2 diabetes, and/or dyslipidemia were included. The AUC for each of the four indices, with each risk factor, was pooled using a random-effects model; male and female data were analyzed separately.

Results

Ten studies met the inclusion criteria. Body mass index (BMI) was the poorest discriminator for cardiovascular risk factors. Waist-to-height ratio (WHtR) was the best discriminator for hypertension, diabetes, and dyslipidemia in both sexes; its pooled AUC (95% confidence intervals) ranged from 0.67 (0.64, 0.69) to 0.73 (0.70, 0.75) and from 0.68 (0.63, 0.72) to 0.76 (0.70, 0.81) in males and females, respectively.

Conclusion

Statistical evidence supports the superiority of measures of centralized obesity, especially WHtR, over BMI, for detecting cardiovascular risk factors in both men and women.

Introduction

In recent years, there has been increasing speculation over which measure of overweight and obesity is best able to discriminate those individuals who are at increased cardiovascular risk. Body mass index (BMI) is used by the World Health Organization to define severity of overweight and obesity across populations [1]. But increasingly, measures of central adiposity, namely waist circumference (WC) and waist-to-hip ratio (WHR), have been adopted as more accurate predictors of obesity-related cardiovascular risk and have replaced BMI in several definitions for clinical diagnosis of metabolic syndrome [2].

The ability of BMI, WC, WHR, and waist-to-height ratio (WHtR) to discriminate major cardiovascular risk factors, namely hypertension, type-2 diabetes, and dyslipidemia, has largely been based upon receiver-operating characteristic (ROC) curve analysis [3], [4], [5], [6]. Reported differences in the discriminatory capability between different indices of obesity with cardiovascular risk factors are small, and despite many studies not formally conducting any statistical comparison, claims have been made to suggest the superiority of some measures of obesity over others in the discrimination of cardiovascular risk [7], [8], [9]. To date, there has been no systematic attempt to compare how well these different measures of obesity perform at discriminating cardiovascular risk factors across diverse populations. If significant, and clinically relevant, differences were shown to exist between these measures, it would provide a strong rationale for the universal adoption of a single measure for defining obesity.

A meta-analysis was conducted to determine which of the four simple indices of overweight and obesity (BMI, WC, WHR, or WHtR) is the best discriminator of hypertension, type-2 diabetes, and dyslipidemia, separately for men and women. We also sought to evaluate whether the combination of BMI with measures of abdominal obesity would increase the discriminatory capability of the model.

Section snippets

Identification of studies

We searched MEDLINE for relevant studies that were published from 1966 to December 1, 2006, using a combination of keywords: “receiver operating characteristic curve,” “anthropometry,” “diabetes,” “hypertension,” “dyslipidaemia,” “obesity,” “body mass index,” “waist circumference,” “waist-hip ratio,” and “waist-height ratio.” References from relevant studies were scanned visually to identify other relevant studies.

Inclusion criteria

We included studies on adults (aged ≥18 years) that used ROC analysis to compare

Results

A total of 25 studies that used ROC curve analysis to study at least one of the three cardiovascular risk factors were identified. Fifteen studies were excluded due to lack of data on at least one of the four indices (Fig. 1). Hence, 10 studies met the inclusion criteria; of those, nine were cross-sectional studies, and one was longitudinal (Table 1). Studies were conducted between 1990 and 2004, in nine countries. Six countries were from Asia, and one of each from the Caribbean, Europe, and

Discussion

This meta-analysis, including data on more than 88,000 individuals from diverse populations, supports previous claims that measures of central obesity, in particular, the WHtR [7], [8], [9], [18], [19], provide a superior tool for discriminating obesity-related cardiovascular risk compared with BMI. However, the observed differences in the discriminatory capability between BMI with each of the individual measures of central obesity were observed to be small, and in general, statistically

Acknowledgments

We thank Gary T.C. Ko, Cuong Q. Tran, and Harald J. Schneider for providing additional information from their studies. C.M.Y. Lee is supported by the National Health and Medical Research Council (Australia) Public Health Postgraduate Scholarship. R.R. Huxley is supported by a University of Sydney SESQUI Postdoctoral Fellowship.

References (29)

  • W.Y. Lin et al.

    Optimal cut-off values for obesity: using simple anthropometric indices to predict cardiovascular risk factors in Taiwan

    Int J Obes

    (2002)
  • Y.H. Pua et al.

    Anthropometric indices as screening tools for cardiovascular risk factors in Singaporean women

    Asia Pac J Clin Nutr

    (2005)
  • M. Woodward

    Epidemiology: Study design and data analysis

    (2005)
  • G.T.C. Ko et al.

    Prediction of hypertension, diabetes, dyslipidaemia or albuminuria using simple anthropometric indexes in Hong Kong Chinese

    Int J Obes

    (1999)
  • Cited by (0)

    View full text