Prediction of poor outcome in Clostridioides difficile infection: A multicentre external validation of the toxin B...

Classification of patients according to their risk of poor outcomes in Clostridioides difficile infection (CDI) would enable implementation of costly new treatment options in a subset of patients at higher risk of poor outcome. In a previous study, we found that low toxin B amplification cycle thresholds (C_t) were independently associated with poor outcome CDI. Our objective was to perform a multicentre external validation of a PCR-toxin B C_t as a marker of poor outcome CDI. We carried out a multicentre study (14 hospitals) in which the characteristics and outcome of patients with CDI were evaluated. A subanalysis of the results of the amplification curve of real-time PCR gene toxin B (XpertTM C. difficile) was performed. A total of 223 patients were included. The median age was 73.0 years, 50.2% were female, and the median Charlson index was 3.0. The comparison of poor outcome and non-poor outcome CDI episodes revealed, respectively, the following results: median age (years), 77.0 vs 72.0 (p = 0.009); patients from nursing homes, 24.4% vs 10.8% (p = 0.039); median leukocytes (cells/μl), 10,740.0 vs 8795.0 (p = 0.026); and median PCR-toxin B C_t, 23.3 vs 25.4 (p = 0.004). Multivariate analysis showed that a PCR-toxin B C_t cut-off <23.5 was significantly and independently associated with poor outcome CDI (p = 0.002; OR, 3.371; 95%CI, 1.565-7.264). This variable correctly classified 68.5% of patients. The use of this microbiological marker could facilitate early selection of patients who are at higher risk of poor outcome and are more likely to benefit from newer and more costly therapeutic options.