Back

Optimized Surveillance Testing for Toxigenic Clostridioides difficile to Scale Prevention Programs Across Clinical Settings

Cersosimo, L.; Correa, N.; Delaney, M. L.; Dellostritto, L.; Misialek, M.; Baker, M.; Klompas, M.; Bry, L.

2026-07-23 infectious diseases
10.64898/2026.07.22.26358596 medRxiv
Show abstract

Clostridioides difficile is the leading cause of healthcare acquired infections (HAIs). Healthcare systems lack scalable strategies to identify patients colonized with toxigenic strains to inform prevention programs. We developed a scalable program using rectal swabs, collected for vancomycin resistance Enterococci surveillance, with ex vivo amplification of C. difficile in Cdiff Banana Broth (BB). Confirmatory testing of BB-positive tubes used the rapid C. difficile QuickCheck assay with Cepheid's Xpert(R) C. difficile/Epi (Xpert) to confirm toxin gene carriage for QuickCheck results that were C. difficile glutamate dehydrogenase-positive (GDH) and antigenic toxin-negative. BB with confirmation demonstrated 92% specificity and 100% sensitivity, to a limit of <5 colony forming units (CFU)/swab, versus published values of 100/swab by chromID(R) C. difficile agar and 460/swab by direct Xpert swab testing. The two-step confirmation detected toxigenic C. difficile in cases missed by these other methods. Prospective testing of 797 VRE swabs from 538 ICU and inpatient oncology patients seen at academic and community hospitals demonstrated 11% positivity for toxigenic and 4% for non-toxigenic C. difficile. Among swabs, 32% of toxigenic C. difficile were resulted at 24 hours and 68% at 48 hours. Addition of 17% glycerol to BB-positive aliquots, for -80C storage, supported 100% retrieval of C. difficile >6 months later. BB with confirmation provides a more sensitive method to detect patient colonization with toxigenic C. difficile and, by using VRE surveillance swabs, does not require additional patient samples. The simplified approach can be deployed across healthcare settings to support C. difficile-prevention programs.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.