Another Ding for the Body Mass Index

A study published by JAMA about the relationship between body mass index (BMI) and mortality puts another ding in the already dubious practice of using BMI by itself as a measure of health. And along the way, the research was twisted into some pretty bad headlines.

The study examined the relationship between body mass and all-cause mortality in three cohorts of individuals in Denmark over a period of three decades. The objective was to determine the BMI associated with the lowest mortality rate in each of the three decades studied. The study found that the BMI with the lowest mortality increased from 23.7 in the 1976-78 cohort to 27.0 in the 2003-2013 cohort.

Shifting BMI Mortality CurvesNow in thinking about these results, it’s important to remember that mortality follows a U-shaped curve as body mass goes up. At a very low BMI, mortality is high and it steadily goes down until it reaches the bottom of the curve. Then the health risks of excess adiposity start adding up as BMI goes up from there. So mortality rates go up. In these cohorts from Denmark, it seems that the mortality curves have shifted right along with rising BMI in the population.

Why is this happening? This study can’t answer that question. But the study’s authors speculate that perhaps improved treatment for complications of obesity might account for some of the shifts in these curves. In other words, people living longer with conditions like heart disease and diabetes might account, at least in part, for these shifts.

What is clear from this study is that simplistic definitions of obesity based on fixed BMI thresholds don’t work. Obesity is a disease of excess adiposity and BMI is nothing more than a crude screening tool. Equating BMI with obesity itself is what gave us headlines that totally missed the mark of this study. “Obesity may not shorten your life” is an example of one of the worst. Even NPR offered a headline suggesting the possibility that “a few extra pounds help you cheat death.” The authors were clear and correct on this point:

The question of causality cannot be addressed in an observational study, i.e., whether the causal association between BMI and all-cause mortality has changed over time.

Health reporters should be just as clear.

Click here for the study and here for a recent review and meta-analysis on BMI and mortality. Click here for the story from NPR.

Ding, photograph © jaya / flickr

Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.


May 16, 2016

2 Responses to “Another Ding for the Body Mass Index”

  1. May 16, 2016 at 10:37 am, Stephen Phillips said:

    Whats old is new

    As early as 1987 Paul Ernsberger and Paul Haskew had their paper published by The Journal of Weight Regulation …(RETHINKING OBESITY).
    To summarize some of their finding they wrote. “Much medical literature has documented elevated risk factors in heavy people, but these risk factors fail to translated into mortality rates. In fact many studies show that maximum longevity is associated with above average weight.”

    Old science and new science seem in agreement with my Grandmother who said. ” You have to be thin enough to be healthy and fat enough to be happy.”

    Stephen Phillips
    American Association of Bariatric Counselors

  2. May 16, 2016 at 10:26 pm, Allen Browne said:

    Poor BMI. It’s easy, it’s useful, but it is what it is.