Accuracy of BMI for Diagnosing ObesityWednesday, August 3, 2011
While I am on a brief holiday in Berlin, I thought I’d rerun a few earlier posts that discuss the issue of measuring obesity and how such measures may or may not be helpful in obesity management – as many readers may not have seen these posts before, comments are very much appreciated.
The following was first posted on July 30, 2008
Body mass index (BMI) is currently widely recommended and used as the best measure of obesity both in population and clinical studies. It dates back to the Belgian statistician Adolphe Quételet, who between 1830 and 1850 described this index as a way to characterize the level of adiposity in sedentary adults.
But how accurate is this index really to identify individuals with excess body fat?
This question was recently addressed by Abel Romero-Corral and colleagues from the Mayo Clinic, MN, USA, who analysed the relationship between BMI and body fat percent (BF%) as measured by bioelectrical impedence in 13,601 subjects (age 20-79.9 years; 49% men) from the Third National Health and Nutrition Examination Survey (Int J Obesity).
In this study, the authors defined obesity based on the World Health Organization (WHO) reference standard for obesity of BF%>25% in men and >35% in women.
BMI-defined obesity (>=30) was present in 19% of men and 25% of women, while BF%-defined obesity was present in 44% of men and 52% of women.
A BMI>=30 had a high specificity (men=95%, women=99%), but a poor sensitivity (men=36%, women=49%) to detect BF%-defined obesity. This means that while the BMI definition does identify the vast majority of men and women who have increased body fat, it also misses a significant number of individuals who have high percent body fat and would be considered obese by the BF% definition.
The diagnostic performance of BMI diminished as age increased and in the intermediate range of BMI (25-29.9), BMI failed to discriminate between BF% and lean mass in both sexes.
The authors conclude that accuracy of BMI in diagnosing obesity is limited, particularly for individuals in the intermediate BMI ranges, in men and in the elderly. Thus, the currently recommended BMI cutoff of >=30 kg for obesity has good specificity but misses more than half the people with excess fat.
The scary part of these results of course is in the fact that based on actual BF% the prevalence of obesity in this population doubled! On the other hand, we know that %body fat or body composition alone is not a particularly reliable measure of health.
I prefer to continue using my operational clinical definition of obesity: the presence of excess body fat that threatens or affects your health.
Given the wide variation in the inter-individual susceptibility to develop adiposity-related health problems, the diagnosis of obesity and the question of whether or not reducing the proportion of body fat will indeed benefit your health will always remain a matter of clinical judgement.