Healthy jogger effect might explain differences in mortality

EDITOR—In their cohort study of mortality in Danish men, Schnohr et al compared the mortality in 96 men who reported that they were joggers at two examinations in 1976–8 and 1981–3 with that in 4562 men who were non-joggers at both examinations.1 The authors followed up the cohort from 1976–8 to 1998 and found that the joggers' relative risk of death was 0.39 (95% confidence interval 0.19 to 0.73).

The participants rather than the authors defined jogging. The paper does not include information on how or how much the men exercised at the time of the examinations or in between and after the examinations. It is not clear whether the men agreed on a common definition of jogging.

At the time of the two examinations, in 1976–8 and 1981-3, 217 (4.7%) and 202 (4.3%) respectively of the 4658 men, aged 20-79, reported that they were joggers. These numbers are surprisingly low. According to a study carried out in 1994 by the Danish National Institute for Public Health, 59.5% of men aged ≥16 exercised or considered themselves to be physically active. Even among men aged ≥67, 55.4% exercised or considered themselves to be physically active.2

No data on mortality are presented in the paper. Because of the small number of joggers the relative risk of death is presumably sensitive to small differences in mortality even though the difference is significant.

The authors point out that their study is observational and that it controls for age but not for differences in lifestyle. A healthy jogger effect rather than jogging in itself may explain the difference in mortality.

Benefits are seen in previously sedentary individuals who start exercising regularly

EDITOR—The Copenhagen city heart study confirms the association between regular exercise and a reduced risk of premature death from certain illnesses. 1 2 But the observed reduction in the risk of death from all causes only in those men who reported jogging at both of two examinations carried out five years apart conflicts with results reported elsewhere. 3 4 We believe that this arises from pooling the results of two patterns of behaviour.

To combine the results from men who went from being active to inactive with those from men who went from being inactive to active (jogging at one examination) implies that the direction of change between active and inactive is immaterial and that the health state between those behaving in these different ways does not differ. This is unlikely. Stopping regular exercise may indicate declining health and an increased risk of death, whereas starting an exercise habit may be a sign of increased wellbeing and fitness. Evidence from both the United States and Britain supports this view. 3 4

Observing changes in activity levels at three time points between 1962 and 1977, Paffenbarger et al reported that the relative risk of death from all causes in participants who at the first time point were engaged in moderately vigorous sports activity but who stopped exercising was slightly higher (1.15 (95% confidence interval 0.73 to 1.71)) than that in those who reported no exercise at all (1.0).3 Those who changed in the opposite direction, from being inactive to being active, had a relative risk similar to that of those who maintained a continuous exercise habit throughout the study period (0.77 and 0.71 respectively).

A similar pattern of health benefit in the direction from inactive to active was also found in the British regional heart study.4 This study reported that, relative to the risk in those who remained inactive, the fully adjusted risk of death from all causes in those who changed from an inactive lifestyle to participating in occasional light exercise was 0.55 (0.36 to 0.84), which was slightly lower than that for continuous exercisers (0.58).

This study suggests that the benefits of a reduced risk of premature death from all causes in those previously sedentary individuals who start exercising regularly are probably as large as those in continuous exercisers; they are unlikely to be the same as those in previous exercisers who gave up exercise.

Authors' reply

EDITOR—In 1976, when the Copenhagen city heart study was started, there was still some concern over whether jogging might be harmful to health. No other study had investigated this question, and we decided to ask each participant “Are you a jogger?,” without further quantification of jogging. The percentage of male joggers aged 20–79 was 4.7% in the first examination (1976–8) and had increased to 8.3% in 1991-4. In addition to jogging status, information on physical activity during work and leisure time was collected, both classified into four levels. Our findings agreed with the physical activity reported in the Danish health and morbidity survey 1994.

Follow up is based on the Danish national population register, and data on mortality are therefore available for nearly all participants. Because our paper was published as a short report, with only one table or figure, however, we were unable to present age specific mortality in addition to the table. As the figure indicates here, the crude survival function for the four groups of joggers shows a remarkably clear pattern over age. The positive effect of jogging is thus unlikely to be an artefact resulting from small sample sizes.

Estimated age specific survival function, using Kaplan-Meier estimator with left truncation and right censoring for four groups of joggers

• Download figure
• Open in new tab
• Download powerpoint

The purpose of conducting the Cox regression analysis was to separate the effect of jogging from the effect of potential confounding variables. One should always be concerned about a possible healthy jogger effect as a partial explanation of the difference in mortality. Previous studies have shown that among its beneficial effects physical exercise (including jogging) reduces blood pressure, being overweight, and the risk of non-insulin dependent diabetes mellitus; raises high density lipoprotein cholesterol concentrations; and decreases platelet aggregation.