Abstract

Although previous studies revealed the association of physical activity with mortality rate, it is unclear whether there is a linear trend between physical activity and mortality rate. In this study, the association of physical activity with the risk of all-cause mortality was analysed using Cox's proportional hazard model for a cohort of 14,204 healthy Korean men aged 40-59 years followed up for 4 years(Jan. 1993 - Dec. 1996). Physical activity and other life style were surveyed by a postal questionnaire in December 1992. Total of 14,204 subjects were grouped into quartiles by physical activity. Using death certificate data, 123 deaths were identified. The second most active quartile had a lowest mortality rate with relative risk of 0.44(95% C.I.: 0.23-0.84) compared with most sedentary quartile, showing a J-shape pattern of physical activity-mortality curve. By examining the difference in proportion of cause of the death between most active quartile and the other quartiles, there was no significant difference of proportional mortality from cardiovascular deaths, cerebrovascular deaths or deaths from trauma. The covariates were stratified into two group between which the trend of RR was compared to test the effect modification. There was no remarkable effect modification by alcohol intake, smoking, body mass index, calorie consumption, percent fat consumption. In conclusion, moderate activity was found to have more protective effect on all-cause mortality than vigorous activity and that the J-shape pattern of physical activity-mortality curve was not due to the difference of mortality pattern or effect modification by alcohol intake, smoking, body mass index, calorie consumption and percent fat consumption.

• Keywords: physical activityall-cause mortalitycohort study