Early vs Late Surgery for Patients With Endocarditis and Neurological Injury: A Systematic Review and Meta-analysis

Surgical timing in infective endocarditis (IE) with preoperative neurological events remains controversial. The relevant society guidelines are each on the basis of a small number of observational studies. This meta-analysis was designed to search the available literature broadly and assess the weight of available evidence as comprehensively as possible.

We searched MEDLINE and EMBASE to April 2018 for studies that compared mortality or neurological exacerbation in early vs late surgery for IE complicated by neurological events. Random effects meta-analysis was performed.

Twenty-seven observational studies (25 unadjusted, n = 879; 2 adjusted, n = 451) met inclusion criteria. Using early and late thresholds defined in each study (7 or 14 days), early surgery in ischemic or hemorrhagic stroke was associated with elevated perioperative mortality vs late surgery (pooled relative risk [RR], 1.74; 95% confidence interval, 1.34-2.25; P < 0.0001; I² = 0%) and greater neurological exacerbation (RR, 2.09; 95% confidence interval, 1.32-3.32; P = 0.002; I² = 33%). In subgroup analysis, for ischemic stroke, early surgery before 7 vs before 14 days exhibited similar perioperative mortality and neurological exacerbation. For hemorrhagic stroke, performing surgery before 21 vs before 28 days showed trends toward perioperative mortality (RR, 1.77 vs 0.63, interaction P = 0.14) and neurological (RR, 2.02 vs RR, 0.44; interaction P = 0.11) exacerbation. There was no difference in long-term mortality but reporting was sparse. Early surgery was often performed for clinical deterioration, negatively biasing outcomes.

Available observational data support delaying surgery by 7-14 days if possible in IE complicated by ischemic stroke and > 21 days in hemorrhagic stroke to decrease perioperative mortality and neurological exacerbation rates. Randomized trials are needed for definitive guidance.