Genomic and clinical determinants of extraintestinal Clostridium perfringens infections in immunocompromised patients
Menif, B.; Wirth, S. E.; Wroblewski, D.; Connors, J.; Correa, N.; Delaney, M. L.; Bry, L.
Show abstract
BackgroundClostridium perfringens can cause life-threatening extraintestinal infections in immunocompromised patients, an area in which we have little information regarding strain factors that impact patient risks and outcomes. MethodsWe conducted genomic-epidemiologic analyses on C. perfringens isolates from 70 patients seen at Brigham and Womens Hospital over 2021-2024. Genomic analyses evaluated strain profiles within a broader context of 2,321 C. perfringens genomes from foodborne, veterinary, clinical, and environmental sources to identify factors associated with invasive infections. ResultsOf 70 patients with C. perfringens infections (mean age 67.6 years), 32 had invasive infections, of which two-thirds had active malignancies, and more than half were immunocompromised. Patients with invasive infections had a significantly higher 90-day mortality of 43.8% vs. 18.4% (p=0.035) and a higher median Charlson Comorbidity Index (6 vs. 3; p=0.003). Notably, no patient isolates were clonal, verifying the absence of hospital-based transmission. Patient isolates showed increased carriage of hyaluronidases (nagHIJKL), sialidases (nanIJ), and perfringolysin O (pfoA). Genomic-epidemiologic analyses identified a new independent association between the NagL hyaluronidase (OR 3.90, 95% CI 1.14 - 16.24) in highly morbid invasive infections. ConclusionWe present a comprehensive genomic analysis of C. perfringens and of strains infecting immunocompromised patients, including epidemiologic associations of the hyluronidase NagL, NanIJ sialidase, and perfringolysin O in highly morbid invasive infections. These genes provide potential markers to identify high morbidity strains that can infect these populations and to further elucidate their role in invasive infections.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.