Objective- Increased Lp(a) [lipoprotein(a)] is associated with coronary heart disease risk, but links with stroke are less consistent. Blacks have higher Lp(a) levels and stroke incidence than whites but have been underrepresented in studies. We hypothesized that Lp(a) is a risk factor for ischemic stroke and that risk differs by race. Approach and Results- REGARDS (Reasons for Geographic and Racial Differences in Stroke) recruited 30 239 black and white US adults aged ≥45 in 2003-2007 to study regional and racial differences in stroke mortality. We measured baseline Lp(a) by immunonephelometric assay in 572 cases of incident ischemic stroke and a 967-person cohort random sample. The hazard ratio of stroke by baseline Lp(a) was calculated using Cox proportional hazards models, stratified by race. Lp(a) was modeled in sex- and race-specific quartiles, given known differences in distributions by race and sex. Interactions were tested by including interaction terms in the proportional hazards models, with P<0.10 considered statistically significant. After adjustment for age, sex, and stroke risk factors, being in the fourth versus the first Lp(a) quartile was weakly associated with ischemic stroke overall, hazard ratio, 1.45 (95% CI, 0.96-2.19). In blacks, the hazard ratio was 1.96 (95% CI, 1.10-3.46), whereas in whites HR was 1.14 (95% CI, 0.64-2.04); P interaction=0.12. Lp(a) was lower in men than women, but associations with stroke in men and women were similar. Conclusions- We confirm that Lp(a) is a risk factor for ischemic stroke. Further research is needed to confirm the role of racial differences of the Lp(a) risk multiplier in ischemic stroke.