21 Background 22 Laboratory diagnosis of Clostridium difficile infection (CDI) remains unsettled, despite updated 23 guidelines. We investigated the potential utility of quantitative data from a nucleic acid amplification 24 test (NAAT) for C. difficile toxin gene (tg) for patient management. 25 Methods 26 Using data from the largest ever C. difficile diagnostic study (8853 diarrhoeal samples from 7335 27 patients), we determined the predicative value of C. difficile tgNAAT (Cepheid Xpert C.diff) low cycle 28 threshold (CT) value for patient toxin positive status, CDI severity, mortality and CDI recurrence.
29Reference methods for CDI diagnosis were cytotoxicity assay (CTA) and cytotoxigenic culture (CTC).30 Results
31Of 1281 tgNAAT positive faecal samples, 713 and 917 were CTA and CTC positive, respectively. The 32 median tgNAAT CT for patients who died was 25.5 vs 27.5 for survivors (p = 0.021); for toxin-33 positivity was 24.9 vs 31.6 for toxin-negative samples (p<0.001) and for patients with a recurrence 34 episode was 25.6 vs 27.3 for those who did not have a recurrent episode (p = 0.111). Following 35 optimal cut-off determination, low CT was defined as ≤25 and was significantly associated with a 36 toxin-positive result (P<0.001, positive predictive value 83.9%), presence of PCR-ribotype 027 37 (P=0.025), and mortality (P=0.032). Recurrence was not associated with low CT (p 0.111).
Conclusions 39Low tgNAAT CT could indicate CTA positive patients, have more severe infection, increased risk of 40 mortality and possibly recurrence. Although, the limited specificity of tgNAAT means it cannot be 41 used as a standalone test, it could augment a more timely diagnosis, and optimise management of 42 these at-risk patients.