Whole-genome sequencing of clinical Clostridioides difficile isolates reveals molecular epidemiology and discrepancies with conventional laboratory diagnostic testing.

SUMMARY Background The high clinical burden of Clostridioides difficile infections merits rapid and sensitive identification of affected individuals. However, effective diagnosis remains challenging. Current best practice guidelines recommend molecular and/or direct toxin detection-based screening for symptomatic individuals, but previous work has called into question the concordance and performance of extant clinical assays. Aim To better correlate the genomic and phenotypic properties of clinical C. difficile isolates with laboratory testing outcomes in both C. difficile-infected patients and asymptomatic carriers. Methods Whole-genome sequencing of clinical C. difficile isolates collected from an inpatient population was performed at a single healthcare institution, enabling examination of their molecular epidemiology and toxigenic gene content. Genomic findings were compared with clinical testing outcomes, identifying multiple diagnostic discrepancies. Findings Toxigenic culture, considered a ‘reference standard’, provided perfect sensitivity and specificity in predicting toxigenic gene content, whereas reduced performance was observed for Simplexa C. difficile Direct Assay (100% specificity, 88% sensitivity), Gene Xpert CD/Epi Assay (86% specificity, 83% sensitivity), and Quick Check Complete Tox A/B (100% specificity, 30% sensitivity). Genomic analysis additionally revealed variability in toxin gene sequences among C. difficile strains, phylogenomic equivalency between isolates from affected patients and carriers, and patient carriage with uncommon environmentally derived C. difficile lineages, as well as presenting opportunities for tracing pathogen transmission events. Conclusion These results highlight the variable performance of clinical stool-based testing approaches as well as the potential diagnostic utility of whole-genome sequencing as an alternative to conventional testing algorithms.