Background: Acute brain embolization (ABE) in left-sided infective endocarditis has significant implications for clinical decision making. The true incidence of ABE, including subclinical brain embolization, is unknown.
Methods and results: We prospectively studied 56 patients with definite left-sided infective endocarditis. Patients were examined by a study neurologist, and those without contraindication had magnetic resonance imaging of the brain. Patients without clinical evidence of acute stroke but with magnetic resonance imaging evidence of ABE were considered to have subclinical brain embolization. Clinical stroke was present in 14 of 56 patients (25%). Among 40 patients undergoing magnetic resonance imaging, the incidence rates of subclinical brain embolization and any ABE were 48% and 80%, respectively. ABE was present in 18 of 19 patients (95%) with Staphylococcus aureus infection. At 3 months, mortality was similar among patients with clinical stroke and subclinical brain embolization (62% versus 53%; P=NS) and was higher among patients with any ABE than among those without ABE (56% versus 12%; P=0.046). Valvular surgery was performed in 25 patients (45%), including 16 with ABE, at a median of 4 days. No patient suffered a postoperative neurological complication. Surgery was independently associated with a lower risk of mortality at 3 months (odds ratio, 0.1; 95% confidence interval, 0.03 to 0.6; P=0.008).
Conclusions: Magnetic resonance imaging detected subclinical brain embolization in a substantial number of patients with left-sided infective endocarditis, suggesting that the incidence of ABE may be significantly higher than reports based on clinical and computed tomography findings have indicated. Brain magnetic resonance imaging may play a role in the complex decision about surgical intervention in infective endocarditis.