Background. Our aim was to study how combinations of three unhealthy b
ehaviors (smoking, physical inactivity, and use of dairy fat) and an i
ndex describing their number were associated with the risk of cardiova
scular mortality in a society showing remarkable improvement in health
behaviors. Methods. The material consisted of random samples of annua
l adult health behavior surveys from 1978 to 1991; included were 8,869
men and 10,105 women ages 45 to 64 years. The mortality follow-up cov
ered Years 1978 to 1993. The study period was divided into four phases
on the basis of number of deaths and timing of health behavior change
s. Poisson multivariate models were used to determine the risk of card
iovascular mortality by all combinations of unhealthy behaviors and th
e index. In the models age, education, chronic morbidity, and body mas
s index were adjusted for. Results. Each unhealthy behavior was found
to be a risk factor for cardiovascular mortality when the whole study
period was examined. Among men, daily smoking was a significant predic
tor of cardiovascular mortality in the first three phases, among women
in the first and third phase. Among men physical inactivity became si
gnificant only in the last phase, among women in all except the last p
hase. The combinations of physical inactivity with use of dairy fat or
with smoking were more significant for women than for men. The combin
ation of smoking and use of dairy fat was significant only for men. Th
e association of the index with cardiovascular mortality was slightly
stronger among women than among men. Conclusions. The predictive value
s of unhealthy behaviors on cardiovascular mortality changed by period
depending on their frequency. The combinations of unhealthy behaviors
are to be preferred over the index in description of lifestyle determ
inants of cardiovascular mortality. (C) 1998 Academic Press.