RESUMO
INTRODUCTION: The performance of MALDI-TOF MS combined with analysis platform for identification of toxin-producing Clostridiodes difficile is yet to be known. METHODS: Between August 2018 and September 2020, 61 isolates from stool specimens of patients with C. difficile-associated diarrhea were analyzed using the MALDI Biotyper system. A C. difficile toxin-producer detection model was developed using ClinProTools. The model was validated using 28 known strains that differed from the isolates used to develop the model. RESULTS: The sensitivity and specificity of the Genetic Algorithm (GA) model using isolates grown on Brucella with hemin and vitamin K (BHK) agar plates were 91.7% and 44.4%, respectively. When isolates grown on cycloserine-cefoxitin mannitol agar were analyzed by the model, sensitivity and specificity were 6.3% and 100%, respectively. The GA model using BHK medium showed the highest discriminatory performance in detection of toxin-producing C. difficile. However, a discrepancy in detection of toxin-producing C. difficile was observed in the results generated when the model was being developed and when the model was validated which suggests that incubation conditions may have affected the results. CONCLUSION: MALDI-TOF analysis using ClinProTools has a potential to be a cost-effective tool for rapid diagnosis and contribute to antimicrobial stewardship by differentiating toxin-producing C. difficile from non-producers.