Author + information
- Lars Bo Andersen, Dr Med Sci∗ ()
- ↵∗Department of Sport Science and Clinical Biomechanics, University of Southern Denmark, Campusvej 55, Odense 5230, Denmark
The recently published paper by Schnohr et al. (1) on jogging and long-term mortality concluded that there was a U-shaped association between all-cause mortality and dose of jogging, and that low-intensity jogging was more beneficial than high-intensity jogging. However, this conclusion cannot be drawn from the data presented.
Schnohr et al. claim to have analyzed the quantity, frequency, and pace of jogging. However, only frequency was assessed, and instead of quantity and pace, they assessed time spent jogging and perceived intensity, respectively. A fit person may cover a much longer distance during the same time as an unfit person and perceive the same absolute intensity as light instead of strenuous. This is a serious problem, because cardiorespiratory fitness is strongly related to mortality and to the analyzed exposures. This confounding was not taken into account, and adjustment for fitness could have resulted in the opposite conclusion.
None of the conclusions are justified by the statistical analysis. In the analysis of the 3 exposures—quantity, frequency, and pace of jogging—the confidence intervals of the subgroups overlap, which means there are no differences between the groups. Most of the subgroups include only 1 to 5 cases, and the conclusions based on this few cases cannot be justified. An analysis in which all of the jogging groups are merged could make statistical sense, and it could conclude that there is a substantial benefit to jogging, but not how fast, how long, or how frequently jogging should take place.
The investigators concluded that there is a U-shape between the dose of jogging and mortality. The definition of what is light, moderate, and strenuous jogging is problematic, because the study did not assess these exposures, as previously mentioned. The strenuous group experienced only 2 deaths, and to draw conclusions of a U-shape on the basis of 2 cases cannot be justified. The investigators do not describe how the 413 joggers in the reference group were selected from among the 3,950 nonjoggers, but it is obvious that an age difference of approximately 20 years between the reference group and all subgroups of joggers points to a severe selection bias, which makes a comparison difficult.
Finally, it is surprising that the investigators did not discuss in detail why 2 recently published studies that analyzed walking (2) and cycling (3) from the same data came up with the opposite conclusion in relation to intensity and duration of physical activity.
Please note: Dr. Anderson has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2015 American College of Cardiology Foundation