Predicting readmission stroke type among blacks and whites in California.

Background: It is generally assumed that initial and recurrent strokes are of the same type, but data from South London, United Kingdom; Lausanne, Switzerland; and other studies suggest this may not be true for certain subtypes. In these studies, however, the number of recurrent strokes observed during the follow-up period was small, thereby limiting the ability of these studies to provide reliable estimates of stroke type concordance.

Methods: Using a large, diverse, inpatient database, this study sought to: (1) estimate the relative agreement (Cohen's kappa) between initial and recurrent stroke types for blacks and whites; and (2) develop a risk-adjusted logistic model for readmission stroke type, with the initial admission stroke type as the main predictor and race, other sociodemographic variables, and clinical and hospital characteristics as potential covariates.

Results: Stroke type concordance was similar for blacks (kappa = 0.77, 95% confidence interval [CI] = 0.71-0.83) and whites (kappa = 0.77, 95% CI = 0.74-0.79). In the adjusted logistic regression models, the initial admission stroke type strongly predicted the readmission stroke type (subarachnoid hemorrhage: odds ratio [OR] = 738.31, 95% CI = 422.58-1289.93; intracerebral hemorrhage [ICH]: OR = 80.86, 95% CI = 61.57-106.19; ischemic: OR = 125.81, 95% CI = 96.12-164.67). Other patient factors, but not race, also predicted readmission stroke type (e.g., younger age increased the odds of having an subarachnoid hemorrhage readmission; atrial fibrillation increased the odds of having an ICH readmission; older age, diabetes mellitus, and heart failure increased the odds of having an ischemic stroke readmission).

Conclusion: This study showed that the initial stroke type and other factors were independently associated with the readmission stroke type and that patterns of stroke type concordance were similar for blacks and whites. These results may help to identify patients in high-risk subgroups who are more likely to have a recurrent hemorrhagic stroke, which could inform patient treatment decisions. For example, patients with atrial fibrillation may be at greater risk for having an ICH readmission because of the adverse effects of anticoagulant therapy, antiplatelet treatment, or both, which should be investigated further.