Sex Differences in Ischemic Stroke Readmission Rates and Subsequent Outcomes After Coronary Artery Bypass Graft Surgery

Background and Purpose: Prior studies examining sex-related risk of readmission for ischemic stroke (IS) after coronary artery bypass grafting (CABG) did not adjust for preoperative comorbidities and used small study samples that were single-center or otherwise poorly generalizable. We assessed risk of readmission for IS after CABG for females compared to males in a nationwide sample. Methods: The 2013 Nationwide Readmissions Database contains data on 49% of all U.S. hospitalizations. We used population weighting to determine national estimates. Using all follow-up data up to 1 year after discharge from CABG hospitalization, we estimated Kaplan-Meier cumulative risk of IS, stratified by sex, using the log-rank test for significance. We created Cox proportional hazard models to calculate hazard ratios (HR) and 95% confidence intervals (CI) for IS readmission, with sex as the main independent variable. We ran unadjusted models and models adjusted for age, vascular risk factors, estimated severity of illness and risk of mortality, hospital characteristics, and income quartile of patient?s zip code. Results: An estimated 53,270 females and 147,396 males survived index CABG admission in 2013. There was a consistently elevated cumulative risk of readmission for IS after CABG for females versus males (log-rank p-value = 0.0014). In the unadjusted Cox model, the HR of IS in females vs. males was 1.35 (95% CI 1.12-1.62, p = 0.0015). The elevated risk for females remained after adjusting for severity of illness (1.30 [1.08-1.56], p = 0.0056) and risk of mortality (1.28 [1.07-1.54], p = 0.0086). This elevated risk persisted after adjusting for multiple vascular risk factors, hospital characteristics, and income quartile of patient?s zip code (1.23 [1.02-1.48], p = 0.03). Conclusions: We found a 23% increased risk of readmission for IS up to 1 year after CABG for females compared to males in a fully adjusted model utilizing a large, contemporary, nationwide database. Further research would clarify mechanisms of this increased risk among women.