Clinical and echocardiographic predictors of embolism in infective endocarditis: systematic review and meta-analysis

Abstract Objectives Prediction of embolic events (EEs) in infective endocarditis (IE) could inform clinical decisions, such as surgical timing. We conducted a systematic review to more precisely define associations between risk factors and EEs. Methods We searched two bibliographic databases (1994–2018) for observational studies that reported EEs in IE patients and considered clinical, microbiological or echocardiographic risk factors. Studies that did not use Duke criteria or only investigated a subset of IE patients were excluded. Study quality was assessed using the Newcastle–Ottawa scale. A pooled risk ratio (RR) for each risk factor was estimated using random-effects models; statistical heterogeneity was estimated using I 2 . Results Of 3862 unique citations, 47 cohort studies (11 215 IE cases) were included; 54 risk factors were analysed in at least two studies, with nine studies reporting other individual factors. Most studies were of high methodological quality. Major predictors of EEs were intravenous drug use (RR 1.69, 95% CI 1.32–2.17; I 2  = 46%), Staphylococcus aureus infection (RR 1.64, 95% CI 1.45–1.86, I 2  = 32%), mitral valve vegetation (RR 1.24, 95% CI 1.11–1.37, I 2  = 30%), and vegetation size >10 mm (RR 1.87, 95% CI 1.57–2.21, I 2  = 48%). EE risk was also higher with human immunodeficiency virus, chronic liver disease, elevated C-reactive protein, Staphylococcus spp. infection, vegetation presence, and multiple, mobile or prosthetic mechanical valve vegetation, and lower with Streptococcus spp. infection. Most findings were unchanged in sensitivity analyses that removed studies with pulmonary EEs from the outcome. Conclusions Given the serious consequences of embolism, surgical evaluation may be considered in patients with these risk factors.