Mortality and prognostic factors in infective endocarditis: A Colombian multicenter cohort

Objectives This study aimed to estimate in-hospital mortality and identify prognostic factors in infective endocarditis (IE) in a Colombian multicenter cohort. Methods This was a retrospective cohort study of 308 consecutive patients with definitive IE (modified Duke criteria) across four Colombian hospitals (2007-2017). Multivariate logistic regression identified independent mortality predictors. The primary outcome was in-hospital death. Results In-hospital mortality was 32.7% (101 of 308), with substantial across-center heterogeneity (9.6-45.9%), contrasting with European (EURO-ENDO 17.1%) and Latin American (25.1%) registries. Independent predictors of death in multivariable logistic regression were renal replacement therapy (adjusted odds ratio [aOR] 9.19, 95% confidence interval [CI] 4.61-18.34), perivalvular abscess (aOR 6.42, 95% CI 2.69-15.35), central nervous system embolism (aOR 3.07, 95% CI 1.63-5.80), older age (aOR 1.03 per year, 95% CI 1.01-1.05), and Staphylococcus aureus etiology (aOR 1.87, 95% CI 0.99-3.51). Surgery was independently protective (aOR 0.34, 95% CI 0.18-0.63), an estimate that persisted after propensity score adjustment. Of 199 patients with surgical indication, 78.9% underwent surgery; non-operated cases had prohibitive risk profiles. Model discrimination was good (area under receiver operating characteristic curve 0.842, 95% CI 0.791-0.894). Conclusion Renal replacement therapy and perivalvular abscess are the strongest independent predictors of in-hospital mortality in this Colombian multicenter IE cohort, whereas cardiac surgery is protective but frequently precluded by prohibitive risk profiles. The 32.7% mortality reflects the high-severity spectrum of a tertiary referral population and is not generalizable to lower-complexity settings. Early identification of patients at risk for dialysis-requiring acute kidney injury and timely multidisciplinary surgical evaluation represent priority clinical intervention targets. In this Colombian multicenter cohort, in-hospital mortality was 32.7%, with substantial across-center heterogeneity. Renal replacement therapy and perivalvular abscess emerged as the strongest independent mortality predictors, identifying patients with limited physiological reserve and advanced infection. Surgery was strongly and consistently protective, including after propensity score adjustment, although unmeasured confounding by indication cannot be fully excluded. Whether multidisciplinary endocarditis team implementation can translate this prognostic information into mortality reduction in Colombian tertiary settings is a hypothesis that should be tested prospectively.