Objective: Individuals with high levels of depressive symptoms have an incr
eased risk of many illnesses, including stroke. Measures of depressive symp
toms include questions about the presence of negative affect, such as sadne
ss, as well as the absence of positive affect, such as happiness and optimi
sm. We assessed whether positive or negative affect, or both, predicted ris
k of stroke. Methods: Data were from a 6-year prospective cohort study of a
population-based sample of 2478 older whites and blacks from five counties
in North Carolina who reported no history of stroke at the baseline interv
iew. Baseline, in-person interviews were conducted to gather information on
sociodemographic, psychosocial, and health-related characteristics of subj
ects. Thereafter interviews were conducted annually for 6 years. Results: I
ncreasing scores on the modified version of the Center for Epidemiological
Studies Depression Scale (CES-D) were significantly associated with stroke
incidence for the overall sample (relative risk [RR] = 1.04 for each one-po
int increase, 95% confidence interval [CI] = 1.01-1.09) over the 6-year fol
low-up period after adjusting for sociodemographic characteristics, blood p
ressure, body mass index, smoking status, and selected chronic diseases. Po
sitive affect score demonstrated a strong inverse association with stroke i
ncidence (RR = 0,74, 95% CI = 0.62-0.88), Conclusions: Increasing scores on
the modified CES-D are related to an increased risk of stroke, whereas hig
h levels of positive affect seem to protect against stroke in older adults.