Active Voice: Can Exercise Be Bad For You?

By Stephen M. Roth, Ph.D., FACSM

Viewpoints presented in SMB commentaries reflect opinions of the authors and do not necessarily reflect positions or policies of ACSM.

Stephen M. Roth, Ph.D., FACSM, is associate professor and associate chair in the Department of Kinesiology at the University of Maryland. His research is focused broadly on the interaction of DNA and exercise, including studies examining the genetic influences on exercise-related traits and also the role of chronic exercise on DNA structure. He is a co-author of the regularly published “Advances in Exercise Genomics” articles in Medicine & Science in Sports & Exercise® and Associate Editor-in-Chief of Exercise and Sport Sciences Reviews.

The media jumped all over a scientific article published a few months ago by Bouchard and colleagues in PLoS ONE that described adverse metabolic responses to aerobic exercise training in some subjects. As can be expected in an article having both the words “adverse” and “exercise” in the title, the natural reaction from the media was to question whether exercise was bad for some people. That is an overly simple interpretation to draw and not the most complete or thoughtful conclusion that should be made from the work.

For those not familiar with the article, the authors performed a retrospective analysis of cohorts from six well-designed and high-quality exercise training intervention studies, each of which examined the change in typical disease risk factors such as fasting insulin, HDL cholesterol, and systolic blood pressure. The authors performed an analysis that minimized the role of typical measurement error in the results, and then defined an adverse response as one that met specific statistical criteria in an adverse direction. For example, a 10 mm Hg or greater increase in SBP after training was categorized as an adverse response, as was a 24 pmol/L or higher increase in fasting insulin. The number of adverse responders was then calculated for each risk factor across the six cohorts. Perhaps surprising to many people was that a significant fraction (>20%) of individuals across each of the six cohorts experienced at least one adverse response for one of the measured risk factors, and a small number exhibited an adverse response in two risk factors.

So, how do we put this information in context? Is exercise “bad” for these people? Should these results change our thoughts on exercise as preventive or therapeutic medicine? No – absolutely not! The benefits of exercise have been clearly documented across all of these risk factors for decades, and across these six cohorts every subject enjoyed benefits in at least one and often in all four risk factors. But this study does shed light on the other end of the exercise adaptation spectrum, one that we often choose to ignore. The results clearly show that, in some cases for a specific risk factor, an individual response can be detrimental for that person. We don’t know why these adverse responses occur or whether they are predictable, but understanding these adverse responses is worth pursuing. If we are able to predict adverse responses to exercise for a particular trait, then we can use that information to provide the best (personalized) exercise prescription.

Some readers will argue about these findings and certainly there are limitations that can be considered. But this issue of adverse responses to exercise has been raised before and though we might disagree on the prevalence of such cases, that they occur should no longer be a question. While those of us affiliated with the American College of Sports Medicine may be reluctant to support findings that might encourage people not to exercise (which is the wrong conclusion to draw from the recent Bouchard et al. study), we cannot discount these data or try to hide them from view. Instead, we must seek to understand the significance of these adverse responses and determine the possible public health relevance.