Oral contraceptive use and incident stroke in women with sickle cell disease.

OBJECTIVE: Oral contraceptives increase the risk of ischemic stroke among women. However, the effect is not studied in the stroke prone sickle cell disease (SCD) population. We want to determine the rate of incident stroke among women with SCD enrolled in a large cohort with longitudinal follow-up.

STUDY DESIGN: A total of 1257 women aged ≥15 and <45years, enrolled in Phase 1 of Cooperative Study of SCD, underwent a baseline examination for assessment of demographics, prior medical history, laboratory assessments, and clinical data. The effects of active oral contraceptive use at baseline interview on incident ischemic and hemorrhagic strokes were assessed after adjusting for potential confounders using Cox Proportional Hazards analysis.

RESULTS: A total of 178 (14.2%) of 1257 women with SCD reported use of oral contraceptives. The age adjusted annual incidence of incident stroke was four folds higher among women who reported active oral contraceptive use than those who did not report use (1.6/100 person-years versus 0.4/ 100 person-years, p=0.03). After adjusting for age, cigarette smoking status, history of exchange transfusions, alcohol use, body weight, systolic blood pressure, and heart rate, oral contraceptive use was not significantly associated with rate of ischemic stroke (hazards ratio [HR], 3.6 95% confidence interval [CI] 0.8-16.5, p=0.09) or hemorrhagic stroke (HR, 1.2 95% CI 0.2-5.7, p=0.8).

CONCLUSIONS: The four fold higher risk of incident stroke associated with use of oral contraceptives in women with SCD can be mitigated by controlling other cardiovascular risk factors.

IMPLICATIONS: Our results are expected to increase the awareness, among both medical practitioners and patients, regarding the four fold higher risk of incident stroke associated with use of oral contraceptives in women with SCD. Our results also identify the confounding effect of other cardiovascular risk factors such as cigarette smoking on the observed relationship and thus identify potential targets for prevention.