Body Mass History May Weigh In on Mortality Risk

By Ruth Kava — Jan 13, 2016
Although we know that obesity is associated with an increased risk of numerous ills, it hasn't been clear that it's also linked to an increased risk of death. A new study suggests that the way the BMI data has been examined may account for that dissonance, and that body weight history may also weigh in on mortality risk.
shutterstock_329042336 Courtesy of Shutterstock

It's widely accepted that obesity especially extreme obesity carries with it a heightened risk of a number of comorbid conditions. For example, Type 2 diabetes, liver disease and some types of cancer have reliably been linked to obesity. But one problem for which the data are somewhat confusing is an increased risk of death mortality.

Some studies of national surveys suggest that increased Body Mass Index might even be protective in the later years. Why might this be?

Drs. Andrew Stokes from the Boston University School of Public Health, and Samuel H. Preston from the University of Pennsylvania, point out that most surveys use a participant's body weight, or BMI, at the time the survey information is gathered, and thus ignore the weight history of the individual. This means that individuals who have always been normal weight are assessed the same way as those who have been obese, but who are normal weight at the time of the survey. These authors point out that "[t]his distinction would be important if individuals who were formerly obese were at higher risk than never-obese individuals."

To determine if this might be the case, the investigators examined the association betwen excess weight and mortality by setting up four different models of the relationship between BMI and mortality:

  • Model 1: Examined only the relationship between BMI at the time of the survey and mortality risk
  • Model 2: Examined only the relationship between lifetime maximum BMI and mortality risk
  • Model 3: Again used survey BMI, but differentiated between people who were in a particular weight class (i.e., underweight, normal weight, overweight, obese, extremely obese) and had never been heavier, and people who were in the same survey BMI but had been heavier at one time
  • Model 4: Again used maximum BMI, but differentiated between people who had reached a particular weight class and stayed in that class, compared to those who had moved to a lower weight at the time of the survey

Drs. Stokes and Preston used data from the 1988-1994 and 1999-2004 National Health and Nutrition Examination (NHANES) surveys, which include in-person interviews, weighing and measuring. They linked that information to death records from the National Death Index through 2011, to determine risk of death. Their data were obtained only from never-smokers aged 50-74 at the time of the survey, and a total of about 6,300 people were included in their analyses.

When they calculated the hazard ratios for deaths from any cause according to Models 1 and 2 described above, compared to individuals at normal weight, those in increasing BMI categories had increasing risks of death. These risks ranged from 10 to 72 percent in Model 1, and from 19 to 250 percent in Model 2.

Models 3 and 4 also provided interesting data. For Model 3, when the survey weight was the maximum weight, the increased risk ranged from 12 to 218 percent. But if people had been at a higher weight and lost weight prior to the survey, the risk ranged from 58 to 280 percent.

Model 4 also indicated that mortality rose with the highest weight achieved, and that those who then lost weight from their maximum had a greater risk of mortality (318 percent) than those who just stayed at that maximum (2.19 percent).

Along with these estimates of mortality risk, the investigators also examined the prevalence of both Type 2 diabetes and cardiovascular disease in people who had lost weight versus those who had remained at their maximum BMI. Although the prevalence was similar for those whose maximum BMI was either in the normal or overweight categories, for those whose BMIs were in the obese (BMI 30-34.9) or extremely obese (BMI >34.9) ranges, the prevalence of both diseases increased.

In summary, the authors found that risk of both illness and death rose with both the maximum BMI attained, and rose even more for those who then lost weight. They acknowledged that a weakness of their study was the reliance on self-reported maximum weight data, and suggested that future research should be done using contemporaneous data on measured weights over individuals' lifespan.

"Our results suggest the burden of overweight and obesity on mortality is likely substantially larger than commonly appreciated," the authors stated. "If correct, this may have serious implications for the future of life expectancy in the United States. Although the prevalence of obesity may level off or even decline, the history of rapidly rising obesity in the last [three] decades cannot be readily erased."