Objectives. This study examined the association between recreational physic
al activity and mortality in middle-aged and older women and the possibilit
y that physical activity serves as an important marker of health.
Methods. Analyses were conducted among participants in the Nurses' Health S
tudy. Levels of physical activity were assessed by questionnaire in 1980 an
d updated every 2 to 4 years.
Results. Levels of physical activity were inversely associated with mortali
ty risk; however, each activity level above the reference level had approxi
mately the same level of risk reduction (20%-30%). The inverse association
was stronger for cardiovascular deaths than for cancer deaths and was stron
gest for respiratory deaths. Women who died of noncardiovascular, noncancer
causes were more likely to have reported that poor health limited their ph
ysical activity than were women who died of other causes or who remained al
ive.
Conclusions. Part of the link between physical activity and mortality risk
is probably spurious and difficult to remove analytically; however, on the
basis of epidemiologic evidence, much of the health benefit of activity is
real.