Evaluation of empirical treatment for blood culture-negative endocarditis

International audienceObjectives: Much progress has been made in understanding the main causes of blood culture-negative endocarditis (BCNE). Few studies concerning BCNE treatment (due to previous antibiotics used or fastidious pathogens) are available. We performed this study to evaluate the effectiveness of our therapeutic protocol in BCNE, based on compliance with the protocol, outcome and 1 year mortality. Patients and methods: We collected prospectively and analysed retrospectively cases of BCNE between 2002 and 2014, using a simplified and standardized protocol developed by our multidisciplinary team. We apply two kinds of protocols to treat BCNE, which include only four intravenous antimicrobial agents: amoxicillin, vancomycin, gentamicin and amphotericin B. Results: We had 177 patients with definite BCNE. There were 154 (87.0%) patients treated with both appropriate antimicrobial agents and appropriate duration of treatment. We analysed the causes of inappropriate treatment in 13 (7.3%) cases and inappropriate duration in 10 (5.6%) cases. The treatment changes were justified in all cases except one of discharge against medical advice. The fatality rate was 5.1% (nine cases) and all deaths occurred in the group of patients who were treated with appropriate treatment; however, four deaths were not attributable to empirical treatment failure. Concerning the other deaths, the lack of surgical management, in association with empirical treatment, could explain our protocol's failure, such as poorly tolerated surgery. Conclusions: Our protocol is efficient and our mortality rate was low, compared with the literature review. This may result from a strategy that uses a sampling procedure and a standardized protocol at the same time