Heart. 2026 Sep 18:heartjnl-2026-328064. doi: 10.1136/heartjnl-2026-328064. Online ahead of print.
ABSTRACT
BACKGROUND: A major challenge in managing left-sided infective endocarditis (LSIE) is identifying patients who will benefit from surgery including those with prohibitive operative risk. However, no widely validated models are currently available to estimate 30-day mortality early after the initial diagnostic evaluation in the broader LSIE population including patients managed without surgery. We aimed to develop and externally validate the European Infective Endocarditis (EURO-ENDO) risk score for predicting 30-day mortality in patients with LSIE.
METHODS: The prospective, multicentre European Society of Cardiology (ESC)-EURObservational Research Programme EURO-ENDO registry included 2171 patients with LSIE diagnosed according to ESC 2015 criteria. Clinical, biological, microbiological and imaging variables were analysed and a multivariable logistic regression model was developed. Internal validation was performed using bootstrap resampling. External validation was conducted in an independent temporal cohort of 377 consecutive patients with LSIE from two tertiary centres in Barcelona, Spain, recruited between 2019 and 2024.
RESULTS: Among 2140 registry patients included in the final analysis, 254 (11.9%) died within 30 days. Eleven variables were independently associated with 30-day mortality and included in the EURO-ENDO prediction model: previous stroke or transient ischaemic attack, previous cardiac surgery, heart failure or cardiogenic shock, creatinine >2 mg/dL, vegetations >10 mm, severe regurgitation, abscess, negative blood cultures, Staphylococcus aureus infection, absence of surgical indication and surgical indication without intention to operate. The model showed good discrimination and calibration (area under the receiver operating characteristic (AUROC) 0.763; Hosmer-Lemeshow p=0.825; Brier score 0.09). In the external validation cohort, 48/377 patients (12.7%) died; discrimination remained excellent (AUROC 0.799; 95% CI 0.730 to 0.868) with good calibration (slope 1.017; expected-to-observed ratio 1.124).
CONCLUSIONS: The EURO-ENDO risk score, incorporating 11 variables available during the initial diagnostic evaluation, accurately predicts individualised 30-day mortality in patients with LSIE and showed consistent performance in both internal and external validation cohorts. This tool may support early risk stratification, multidisciplinary endocarditis team discussions and clinical decision-making (https://scoreei.web.app/score).
PMID:42760106 | DOI:10.1136/heartjnl-2026-328064