Résumé
Background Single-sampling strategy (SSS) for blood cultures (BCs) has not been evaluated in infective endocarditis (IE). We assessed the diagnostic performance of SSS versus conventional multisampling strategy (MSS) in IE diagnosis.Methods Patients suspected of IE were prospectively enrolled in 8 tertiary-care centers. Five BC bottle pairs were sampled from each patient. Pairs 1, 2, and 3 were sampled simultaneously, followed by 2 additional separate pairs (4 and 5) sampled more than 1 hour later. Pairs 1, 2, and 3 emulated SSS and pairs 1, 4, and 5 emulated MSS. The sensitivity and specificity of the major microbiologic criterion of the 2015 European Society of Cardiology IE diagnostic criteria, based on the SSS and MSS BC results, were calculated using the endocarditis team's diagnosis as the gold standard.Results An IE was diagnosed in 101 (39.4%) of the 256 patients enrolled. Sensitivity rates of SSS and MSS were 50.5% (95% CI: 40.7-60.2) and 45.5% (95% CI: 35.8-55.3), respectively (P = .063), whereas specificity rates were 94.8% (95% CI: 91.4-98.3) and 95.5% (95% CI: 92.2-98.8), respectively (P = 1). In IE patients, SSS compared to MSS accurately upgraded the diagnosis from possible to definite IE in 1 patient and downgraded it in none.Conclusions Using SSS to define the major microbiologic criterion was as sensitive and specific as using MSS for diagnosing IE. Using SSS instead of MSS BC results did not lead to erroneous changes in diagnostic class according to the 2015 European Society of Cardiology criteria. Consequently, SSS may be regarded as standard practice for IE diagnosis.
A single sampling strategy for performing blood cultures is at least as effective as the conventional multiple sampling strategy in diagnosing and classifying infective endocarditis