Background: Stroke is a major complication of infective endocarditis (IE). The features and predictors of ischemic and hemorrhagic stroke subtypes are unknown.
Methods and Results: In this retrospective study (1980-2011) we used original Duke criteria to diagnose IE in 1149 episodes (1081 patients; 81% definite, 80% native valve). Neuro complications occurred in 28% including infarcts (17.6%), ICH (4.6%), meningitis (4%), brain/spinal abscess (8%), and seizures (2%). Ischemic stroke was the isolated presenting symptom in 11%; its most frequent presentations were total/partial anterior-circulation syndrome (47%) and embolic encephalopathy (21%). On DWI, infarcts were disseminated small (45%) or multi-sized (31%), single (16%), or territorial (9%). Disseminated infarcts were associated with poor outcome (p=0.026). Post-ischemic hemorrhage occurred in 6.4%. Primary ICH was evident on admission in 55% including 36 parenchymal (92% lobar), 13 SAH (69% convexal) and 4 SDH. As compared to the non-stroke group (n=885), the ischemic (n=202) and hemorrhagic stroke (n=53) groups had higher rates of complications (e.g., meningitis, seizures, sepsis, renal failure), longer LOS, higher mortality, and lower rates of home discharge (all p<0.05). Both stroke groups had higher rates of S. viridans and Group B strep infections and more valvular vegetations; the ischemic stroke group had more intracardiac thrombi, and the hemorrhagic stroke group had more mechanical valve infection (all p<0.05). Heparin use was associated with a higher rate of new infarcts (12% vs. 5%, p=0.027) but not ICH (2.4% vs. 1.4%, p=0.6). Overall mortality (14.6%) was higher in both stroke groups, however both groups had similar neurological mortality. Mitral valve involvement, sepsis, and strep infections were independent predictors of both ischemic and hemorrhagic stroke. Age, renal failure, heart failure, sepsis, and Staph aureus endocarditis were independent predictors of non-home discharge. Over three decades the LOS decreased, ischemic stroke rates increased, but there was no change in ICH or mortality rates.
Conclusion: Endocarditis-related ischemic and hemorrhagic stroke subtypes have overlapping risk factors, predictors, clinical phenotypes, and effects on outcome.