Accuracy of two Sepsityper MALDI-TOF MS methods for bacterial identification in bloodstream infections in dogs, foals, and calves using Bayesian latent class model.

Abstract

Accurate diagnosis of bloodstream infections is crucial for survival and antimicrobial de-escalation in veterinary medicine. Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry enables faster identification of bacteria in BSIs. This study aimed to compare diagnostic accuracy of two Sepsitypermethods (Rapid Sepsityperand SepsityperExtraction) with conventional culture method for bacterial identification in clinical samples. Mean time-to-positivity and frequency of bacteria in BSIs were also determined. This diagnostic test study used bloodstream infection samples from 385 critically ill animals (121 dogs, 119 foals, and 145 calves) admitted to the Faculty of Veterinary Medicine, Ghent (October 2021-February 2024). Accuracy was compared using Bayesian latent class model with priors for sensitivity (99.9%) and specificity (96.0%) based on literature, and a prevalence of 26.0%. Conventional culture method identified 173 bacteria with(19.1%,33/173),spp. (12.1%,21/173) andspp. (8.1%,14/173) being most common. Sensitivity of Rapid Sepsityper, SepsityperExtraction, and conventional culture method was 62.1%, 86.1%, and 97.4%, respectively. Specificity was 94.3%, 90.4% and 92.3%, and accuracy was 85.8%, 89.3%, and 93.6%, respectively. Mean time-to-positivity and ±standard deviation for blood cultures flagging positive was 21h25min ±17.8h. Rapid Sepsityperidentified bacteria in approximately 30min, while SepsityperExtraction method required around 50min, and conventional culture method needed 12-48h. Altogether, SepsityperExtraction shows promise given the sensitivity and results were delivered more rapidly than conventional culture. Enhancing diagnostic workflow, resulting in a better prognosis, reduced hospital stays, and lower healthcare costs due to more rational use of (critically important) antimicrobials.