Objective: To assess prospectively the risk of future stroke associated with baseline concentration of lipoprotein(a), abbreviated Lp(a).
Design: Nested case-control study using baseline plasma samples.
Setting: Men in the Physicians' Health Study.
Participants: A cohort of 14,916 male physicians with no prior history of stroke, transient ischemic attack, or myocardial infarction provided plasma samples at baseline and were followed prospectively for 7.5 years. Samples from 198 physicians who subsequently developed stroke (155 thromboembolic, 35 hemorrhagic, eight indeterminate) were analyzed for Lp(a) concentration together with paired controls, matched for age and smoking habit.
Main outcome measure: Fatal and nonfatal stroke.
Results: Median Lp(a) concentration (8.88 mg/dL [0.23 mmol/L] vs 8.55 mg/dL [0.22 mmol/L]), P = .69) and overall distributions of Lp(a) (P = .54) were similar at baseline in men who did and did not develop future stroke. In analyses controlling for age, smoking status, blood pressure, obesity, and the presence of diabetes, the relative risks (RRs) associated with baseline Lp(a) concentration exceeding the 25th, 50th, 75th, 90th, and 95th percentiles of the control distribution were 1.26, 0.99, 1.06, 0.90, and 1.03 (all P values nonsignificant). There was likewise no association in analyses limited to thromboembolic events. For example, among subjects with baseline Lp(a) values exceeding the 95th percentile of the control distribution, the RR of future thromboembolic stroke was 1.01 (P = .9). No evidence of association between Lp(a) and stroke risk was found in analyses limited to individuals with hypercholesterolemia.
Conclusions: Among nearly 15,000 predominantly white, healthy, middle-aged men followed in the Physicians' Health Study for a period of 7.5 years, we found no evidence of association between baseline plasma concentration of Lp(a) and future risk of total or thromboembolic stroke.