Phenotype-Based Stratification and Early Prediction of Staphylococcal Infective Endocarditis: Development of a...

Background: Early identification of Staphylococcus spp. infective endocarditis (IE) remains clinically challenging but essential for timely initiation of targeted antimicrobial therapy. We aimed to characterize pathogen-specific clinical phenotypes and to develop a pragmatic bedside prediction model for staphylococcal etiology. Methods: We conducted a retrospective cohort study including 112 patients diagnosed with definite IE. Demographic, clinical, echocardiographic, and laboratory data were analyzed. Independent predictors of staphylococcal etiology were identified using multivariable logistic regression, with internal validation by bootstrap resampling. A simplified clinical risk score was derived from regression coefficients. Platelet kinetics were evaluated as a potential complementary biomarker. Results: Patients with staphylococcal IE (n = 66) were younger and exhibited a distinct clinical profile characterized by a high prevalence of intravenous drug use, right-sided valve involvement, and a markedly elevated inflammatory response. Independent predictors included intravenous drug use (OR 7.1, 95% CI 2.45-20.6, p < 0.001), higher C-reactive protein levels (OR 1.08 per unit increase, p = 0.025), and lower oxygen saturation (OR 0.75 per 1% increase, p = 0.007). The model demonstrated good discrimination (AUC 0.82, 95% CI 0.74-0.90) and calibration. The simplified exploratory risk score stratified patients into low-, intermediate-, and high-risk groups, with observed probabilities of 27%, 82%, and 94%, respectively. Conclusions: Staphylococcal IE represents a distinct clinical phenotype. A simple three-variable model enables early bedside identification of high-probability cases and may support risk-adapted management decisions. Nevertheless, external validation is required before clinical implementation of a score.