Title of article:
Component distributions of body mass index defining moderate and extreme overweight in Danish women and men
Authors: Price RA, Sorensen TI, Stunkard AJ.
Journal: Am J Epidemiol, Jul 1989;130(1):193-201
Abstract
Skewness in the distribution of body mass index in a population of Danish men (n = 1,589) and women (n = 1,988) adopted in Copenhagen between 1923 and 1947 can be explained by mixture of three component distributions, after removing age and sex effects. Even when residual skewness was allowed for in component distributions, similar results were obtained. The upper component distributions corresponded to extreme overweight in both women and men. The results indicate that while there are differences in the means and ranges of body mass index in men and women, the component distributions are similar. Both genetic and environmental factors can produce such component distributions. The authors speculate that the underlying predisposition to obesity is distributed similarly in women and men. Finally, on the basis of the component distributions, the authors propose thresholds of body mass index which allow for 85-98 per cent separation of moderate and extreme overweight groups in men and women.
Comments and Key points
This article reported on the Danish population, and suggested suitable cut-off thresholds to define Overweight and Obesity. (Actually, the article labelled these as "Moderate overweight" and "Extreme overweight"). Anyway, here are the thresholds suggested.
| Overweight | Sensitivity | Specificity | |
|---|---|---|---|
| Men |
26.1
|
77.5% | 84.7% |
|
27.2
|
62.4% | 93.5% | |
| Women |
25.0
|
90.0% | 92.4% |
|
26.9
|
63.2% | 98.8% | |
| Obese BMI | |||
| Men |
31.9
|
92.9% | 98.7% |
|
31.8
|
93.5% | 98.6% | |
| Women |
33.1
|
97.6% | 99.7% |
|
31.4
|
99.6% | 98.7% |
The article gives two choices for each threshold, and shows the sensitivity and specificity that would result. I should point out that the numbers for sensitivity and specificity are high, much higher than typically seen in other studies, which suggests that there is a problem with this study somehow.
But at least, this article is an example of an opinion about BMI thresholds, that differs from the usual international standard values of 25 and 30.
The mean age in this study was 41.5 yrs for men, and 42.7 yrs for women. The study didn't measure body fat percentages. Instead, it analyzed distributions of Body Mass Index in the population, and adjusted the thresholds, to treat men and women equally on a statistical basis. So even if you don't agree with the exact numbers for the thresholds, this article gives some guidance to the suitable gap between mens and women's values ( when sensitivity and specificity are reasonably equivalent).
Review & comments by Steven B. Halls, MD, Last changed
23-June, 2008, Copyright.
Return to Body Mass Index articles index.
