Clinical utility of the DENOVA score for predicting infective endocarditis in Enterococcus faecalis bacteraemia: external validation with decision curve analysis

Objectives: The DENOVA score was developed to support risk stratification for infective endocarditis (IE) in patients with Enterococcus faecalis bacteremia (EfB) and to guide echocardiographic evaluation; external validation remains, however, limited. The objective of this study was to externally validate the DENOVA score in a large prospective international cohort of patients with EfB. Methods: Prospective multicentre cohort study (2019-2024) conducted across 23 centres in six countries (Italy, Spain, Israel, Brazil, Switzerland, and Romania), including adult patients with monomicrobial EfB who underwent at least one echocardiographic evaluation as part of the study protocol. Definite IE was defined according to the 2023 Duke criteria. The DENOVA score was externally validated; discrimination was assessed by the area under the receiver operating characteristic curve, calibration by observed versus predicted risk plots, and clinical utility by decision curve analysis at the predefined threshold (DENOVA ≥3). Performance of the NOVA score was evaluated as a secondary analysis. Results: Among 543 patients, 125 (23.0%) were diagnosed with IE. When evaluated as a continuous variable, the DENOVA score showed an area under the receiver operating characteristic curve of 0.871 (95% CI 0.835-0.906). At the predefined threshold (DENOVA ≥3), sensitivity was 79.2% (95% CI 71.0-85.9) and specificity was 83.0% (95% CI 79.1-86.5). Decision curve analysis showed that DENOVA ≥3 was associated with positive net clinical benefit across most threshold probabilities, with a reduction in unnecessary transoesophageal echocardiographies. Conclusions: The study demonstrated that DENOVA supports risk stratification for IE in patients with EfB, informing clinical decision-making and use of transoesophageal echocardiography.