And now a new Canadian study points to the possibility that use of BMI may be responsible for the oft-cited observation that some folks with higher levels seem to be healthier than those in the normal range — which is the obesity paradox.
The study authors, led by Dr. Raj Padwal from the Department of General Internal Medicine, University of Alberta, Edmonton in Canada, utilized a database of people aged 40 and up who had had DXA (dual-energy x-ray absorptiometry) imaging to determine bone mineral density. This imaging also provided a measure of the percentage of body fat, and the investigators compared the all-cause mortality risk predicted by BMI (kilograms/meter2) with that predicted by percent body fat to see which would be most accurate.
The study included about 49,500 women aged 64 years with an average BMI of 27 kg/m2and mean body fat of 32 percent. Data from about 5,000 men aged 66 years, with BMI of 27 and body fat of 30 percent. The women were followed for nearly seven years, during which time about 5,000 of them died. The men's followup lasted about five years, and during that time about 1,000 of them died.
In women, low BMIs (less than normal) and high percent body fat were associated with a significantly increased risk of dying. For men with low BMIs, again there was a significantly increased risk of dying, and the men with the greatest percent of body fat were again at a higher risk of dying. A high percent of body fat seemed to be more deleterious in men than women. It was associated with a 19 percent increased risk for women, and a 59 percent increased risk for men.
The investigators noted some limitations to their study. For example, the fact that participants were referred for DXA testing might limit the generalizabilty of their results. Also, they were not able to ascertain the prevalence of smoking or physical activity and control for those factors in their evaluation.
However, for this large cohort of Canadians it was clear that when BMIs were below normal they were significantly associated with an increased risk of death, but high BMIs were not. On the other hand, a high percent of body fat was independently associated with an increased risk of death. These results suggest that the imprecise indication of body fat provided by Body Mass Index may well be the root of the supposed obesity paradox.
Such results are important because BMI has been used, for example by insurance companies assessing risks of individuals, and suggest that this can be a misleading index of true obesity. However, it would be impractical to suggest that DXA imaging be used widely to determine body fat. Perhaps a new index that combines BMI with waist circumference (a measure that indicates abdominal adiposity — the most dangerous kind) could be established and validated, to provide a more accurate assessment of obesity risk.