Supplementary Material for: Magnetic Resonance Imaging-Based versus Computed Tomography-Based Thrombolysis in Acute Ischemic Stroke: Comparison of Safety and Efficacy within a Cohort Study

Background: In acute ischemic stroke, brain imaging is mandatory in the decision whether to perform intravenous thrombolysis with recombinant tissue plasminogen activator. The most widespread used imaging modality to exclude intracranial hemorrhage is plain computed tomography (CT). However, there is an ongoing debate whether the information provided by magnetic resonance imaging (MRI) could improve the selection of patients for thrombolysis. We investigated whether the choice of imaging modality (MRI vs. CT) affects therapy safety and the patients' outcome. Methods: Analyses are based on data from a prospective, single-center observational study that included all patients with acute ischemic stroke who received intravenous thrombolysis within 4.5 h. Stroke severity was assessed by the National Institutes of Health Stroke Scale. Safety was assessed by rates of symptomatic intracranial hemorrhage (SICH), brain edema with mass effect and 7-day mortality. Outcome was assessed at 3 months as mortality and proportion of independent patients (modified Rankin Scale score between 0 and 2). Results: We analyzed 345 patients of whom 141 received multimodal MRI and 204 received plain CT prior to treatment. Groups did not differ significantly in terms of age, neurological deficit, rate of elevated glucose level or rate of very high blood pressure. However, patients with CT-based thrombolysis had significantly higher rates of cardiac comorbidities (coronary artery disease, heart failure). In the MRI group, we observed a lower rate of 7-day mortality (1 vs. 10%; p = 0.001), a lower rate of SICH (1 vs. 6%; p = 0.010) and a nonsignificantly lower rate of brain edema with mass effect (2 vs. 6%; n.s.). In multivariable analysis, 7-day mortality was independently associated with MRI-based thrombolysis, even if cardiac comorbidities were taken into account. For mortality at 3 months, there was a nonsignificant difference in favor of the MRI group (16 vs. 23%; n.s.). In multivariable analyses, mortality at 3 months was independently associated with older age, higher stroke severity, brain edema with mass effect, SICH, pneumonia and coronary artery disease. Neither mortality nor independent outcome was influenced by initial imaging modality. Conclusions: Thrombolysis based on multimodal MRI is associated with reduced rates of SICH and early death. Our results suggest that these complications affect survival principally in the acute phase after thrombolysis. However, nonneurological and especially cardiac comorbidities also influence survival after stroke and are underrepresented in stroke patients undergoing MRI. Selection bias has to be considered.