BackgroundWhile early appropriate antibiotic therapy is a proven means of limiting the progression of infections, especially bacteremia, empirical antibiotic therapy in sepsis is ineffective up to 30%. The aim of this study was to compare early blood culture testing protocols in terms of their ability to shorten the delay between blood sampling and appropriate antibiotic therapy.MethodsIn this french observational study, we compared three blood culture testing protocols. Positive blood cultures were tested using either GenMark ePlex panels (multiplex PCR period), a combination of MRSA/SA PCR, β-Lacta and oxidase tests (multitest period), or conventional identification and susceptibility tests only (reference period). Conventional identification and susceptibility tests were performed in parallel for all samples, as the gold standard.ResultsAmong the 270 patients with positive blood cultures included, early and conventional results were in good agreement, especially for the multitest period. The delay between a blood culture positivity and initial results was 3.8 (2.9–6.9) h in the multiplex PCR period, 2.6 (1.3–4.5) h in the multitest period and 3.7 (1.8–8.2) h in the reference period (p<0.01). Antibiotic therapy was initiated or adjusted in 68 patients based on early analysis results. The proportion of patients receiving appropriate antibiotic therapy within 48 h of blood sampling was higher in the multiplex PCR and multitest periods, (respectively 90% and 88%) than in the reference period (71%).ConclusionThese results suggest rapid bacterial identification and antibiotic resistance tests are feasible, efficient and can expedite appropriate antibiotic therapy.