Outpatient versus inpatient-only antibiotic therapy in left-sided infective endocarditis: long-term clinical and neurologic outcomes in a real-world cohort

Abstract Prolonged inpatient antibiotic therapy for infective endocarditis (IE) strains inpatient resources and may increase the risk of hospitalization-related complications. Outpatient continuation strategies are increasingly used, but long-term outcomes—particularly neurologic endpoints—in broad real-world populations remain incompletely characterized. [1] Purpose To compare long-term clinical and neurologic outcomes between an outpatient continuation strategy after discharge and inpatient-only antibiotic therapy among consecutive adults with left-sided IE who survived to discharge. Methods We performed a retrospective cohort study of consecutive adults with left-sided IE treated at a tertiary center (2013–2024) who survived to discharge and received either outpatient continuation of antibiotics (OAT) or inpatient-only therapy. The primary endpoint was a composite of all-cause mortality or unfavorable neurologic outcome through 8 years (secondary: 1 year). Confounding was addressed by multivariable Cox regression and propensity-score matching, with complete-case sensitivity analyses. Kaplan–Meier and Cox models were used for time-to-event analyses. An exploratory non-inferiority compatibility check used a margin informed by the POET trial. [2] Results Among 391 patients (65±13 years; 26% female), 79 (20.2%) received OAT and 312 (79.8%) inpatient-only therapy. Baseline differences supported non-random treatment allocation. In the full cohort, OAT was not associated with the 8-year primary endpoint (unadjusted HR 1.22, 95% CI 0.87–1.71; Figure 1A) and remained non-significant after adjustment (HR 0.86, 95% CI 0.60–1.24). Propensity-score matching yielded 79 pairs (n=158) with improved balance; effect estimates again showed no clear association (unadjusted HR 0.94, 95% CI 0.64–1.37; adjusted HR 0.68, 95% CI 0.43–1.07), consistent with Kaplan–Meier curves (Figure 1B) and sensitivity analyses (Figure 2). In the exploratory non-inferiority assessment, the adjusted matched estimate met the non-inferiority criterion because the upper 95% CI bound (1.07) remained below the pre-specified margin. Secondary 1-year analyses were concordant. Conclusions In this real-world consecutive cohort of left-sided IE, outpatient continuation strategies were associated with no detectable difference in long-term composite mortality/neurologic outcomes compared with inpatient-only therapy across complementary adjustment and matching approaches. Prospective studies in broad, unselected patient populations are warranted to refine patient selection and characterize long-term neurologic outcomes.