Background: Obesity can be divided into "general" and "central." Since abnormal glucose and lipid metabolism are more strongly associated with central obesity, it may not be adequate to use a general measure, such as a weight-for-height index, to assess for obesity. An index of central obesity, such as the waist-to-hip ratio, might be more appropriate.
Methods: Nurses measured height and weight for the body mass index (BMI = kilograms of mass divided by the square of the height in meters) and girths for the waist-to-hip ratio (WHR) in 414 patients aged 45 years and over. Patients completed an obesity-related questionnaire.
Results: Fifty-seven percent of patients had an elevated BMI. Fifty percent of men (95% confidence interval [CI], 46 to 55) and 78% of women (95% CI, 75 to 80) had central obesity based on elevated WHRs. Using an elevated WHR as the standard for central obesity, elevated WHR as the standard for central obesity, elevated BMI had a positive predictive value of only 64% and a negative predictive value of 68% in men. For women, the corresponding positive and negative predictive values were 84% and 31%, respectively.
Conclusions: The data indicate that the practice of using only scales to identify "overweight" patients should be reevaluated since doing so will miss patients at risk. In primary care patients, particularly those 50 years of age and over, weight-for-height indices such as the BMI result in underdiagnosis of central obesity.