Blood culture-negative infective endocarditis: a worse outcome? Results from a large multicentre retrospective Spanish cohort study

Background To assess the impact of blood cultures negative infective endocarditis (BCNIE) on in-hospital mortality. Methods Prospective multicentre study with retrospective analysis of a Spanish cohort including adult patients with definite IE. Cardiac implantable devices infection were excluded. Comparisons between blood cultures positive and BCNIE groups were performed to analyse in-hospital mortality. Results 1001 cases were included of which 83 (8.3%) had BCNIE. Alternative microbiological diagnosis was achieved for 39 (47%) out 83 cases. The most frequent identifications were: Coxiella burnetii (11; 28.2%), Tropheryma whipplei (4; 10.3%), Streptococcus gallolyticus (4;10.3%) and Staphylococcus epidermidis (3; 7.7%). Surgery was performed more frequently in BCNIE group (57.8 vs. 36.9%, p < .001). All-cause in-hospital mortality rate was 26.7% without statistical difference between compared groups. BCNIE was not associated to worse mortality rate in Cox regression model (aHR = 1.37, 95% CI 0.90-2.07, p = .14). Absence of microbiological diagnosis was also not associated to worse in-hospital prognosis (aHR = 1.62, 95% CI 0.99-2.64, p = .06). Conclusions In our cohort, BCNIE was not associated to greater in-hospital mortality based in multivariate Cox regression models. The variables most frequently associated with mortality were indicated but not performed surgery (aHR = 2.48, 95% CI 1.73-3.56, p < .001), septic shock (aHR = 2.24, 95% CI 1.68-2.99, p < .001), age over 65 years (aHR = 1.88, 95% CI 1.40-2.52, p < .001) and complicated endocarditis (aHR = 1.79, 95% CI 1.36-2.37, p < .001).