Automated reading of RAST results using laboratory automation and ReadABrowser(TM) software ; Évaluation des performances d’un protocole automatisé de détermination rapide de la sensibilité aux antibiotiques (DRSA) à partir d’hémocultures positives

Mémoire de Diplôme d'Etudes Spécialisées (DES) tenant lieu de thèse d'exercice. ; Introduction: Bloodstream infections (BSI) are the leading cause of sepsis, which is associated with an important morbi-mortality. Its prognosis is correlated with its rapid identification and introduction of appropriate antimicrobial therapy. The gold standard for BSI diagnostic is blood culture. Recently, EUCAST proposed the Rapid Antibiotic Susceptibility Testing (RAST) methodology from positive blood culture with zone diameter breakpoints defined after short incubation times (4h, 6h and 8h) for the rapid management of bloodstream infections. However, this protocol requires precise reading times (plus or minus 5 minutes) making it difficult to implement in a manual routine workflow. Solution can be brought by BD KIESTRATM laboratory automation and ReadACompact™ incubators. It allows automated reading of AST plates at precise time points and zone diameter measurement on the digital pictures of AST disk diffusion plates using ReadABrowser™ software. Methods: We evaluated the performances of automated RAST compared to standard AST on all overnight positive blood cultures from July 2020 to February 2021. Species was identified by MALDI-TOF mass spectrometry performed on the pellet of 1 mL of positive blood culture using an in-house protocol. If the species identified was one of the eight species for which RAST breakpoints are defined, we performed manual and BD InoqulA™ automated plating of Mueller-Hinton agar medium and added antibiotic panels following RAST methodology. An automated picture of the plate was taken after 4h, 6h and 8h of incubation and inhibition diameters were adjusted manually using ReadABrowser™ software. If the inhibition zone was readable, the strains were categorized as S/ATU/R according to EUCAST RAST breakpoints (v2.1). Results: Readability was 94.8%, 98.2% and 99.9% for E. coli (65 strains) with automated plating and 70.5%, 94.9% and 99.4% for S. aureus (39 strains) after 4h, 6h and 8h respectively using ...