Twenty years ago an expert panel at the NIH created a furor among obesity researchers by suggesting that the BMI cutoff point for a person to be considered overweight be lowered from 27 to 25. That guidance was accepted, and Americans' girths continued to rise. Now, an editorial in JAMA has suggested that we should lower that cut point even more — but for only one group — postmenopausal women.
We, and plenty of others, have pointed out the weakness of using BMI to indicate fatness, whether overweight, obese or extremely obese. In a nutshell, BMI can't distinguish body fat from muscle, and thus could label someone with a lot of muscle as obese when he or she isn't. The measure is, however, simple to obtain and useful for gaining information about changes in populations' levels of fatness, and so will not go out of use in the near future.
The authors of the JAMA editorial acknowledge such issues, but point out that even if a woman's BMI remains constant as she ages, her body composition usually does not. Especially in the postmenopausal years, a woman's complement of muscle and bone (lean body mass) tends to decrease, while the proportion of fat increases. And that fat likely deposits around her middle — perhaps as subcutaneous fat, and/or in deeper depots as visceral fat. In addition, people lose height as they age — which could also affect the accuracy of using BMI to assess fatness.
This is particularly well supported by a recent study published in the journal Menopause. Dr. Hailey R. Banack from the State University of New York at Buffalo and colleagues. They used dual-energy x-ray absorptiometry (DXA) scans [the gold standard] to actually measure the amount and distribution of body fat in 1,329 postmenopausal women aged 59 to 83 years. These investigators noted that there is currently no consensus on what actual percentage of body fat (BF%) should be considered defining of obesity. Thus they compared the BMIs of their participants to the BF% equal to 35, 38, and 40 percents.
They found that with a BF% cutoff of 40, they was the greatest proportion of women (82 percent) correctly classified as obese. Further analysis indicated that if BF% 40 is used as the cutpoint, then for such women the true standard for an obese BMI would be 27, substantially below the current BMI of 30. For a BF% of 35, that BMI would be 24.85, and for BF% 38 it would be 26.49.
The authors stated:
The results of the present study indicate that a substantial proportion of older women who are truly obese are misclassified as nonobese by BMI, regardless of the BF% used as the reference standard.
They also noted that while BMI isn’t a perfect index of obesity, the fact that as the BF% increased, so did the BMI to be used, and that this concordance supports the utility of BMI to some extent.
What does this all mean for clinical practice at present? It could mean that additional care must be taken with this population of older women to assess the extent to which accretion of additional body fate over the years has been seen mostly in the visceral depots. Perhaps, while taking into account the information on BF% and lower cut points for BMI, one can also consider indirect indices of visceral obesity such as waist circumference.