Derivation of clinical predictive factors (CHIEF) for first recurrent Clostridioides difficile infection

Abstract

Current models for predicting Clostridioides difficile infection (CDI) recurrence rates have a limited capacity to account for important risk factors. This study developed a clinical prediction rule for CDI recurrence. This retrospective cohort study evaluated 209 patients with CDI at a university hospital in Japan. Logistic regression and receiver operating characteristic curve analyses were performed to identify potential predictors (age, sex, underlying diseases, antibiotic use, acid suppressants, immunosuppressants, CDI history) of CDI recurrence. Forty-five patients developed recurrent CDI. Univariate analyses identified several significant recurrence predictors (enteral feeding, inflammatory bowel diseases [IBD], community-onset CDI, severe CDI). Enteral feeding (odds ratio: 3.87, 95% confidence interval: 1.75-8.56) and IBD (odds ratio: 7.08, 95% confidence interval: 1.28-39.06) were significant factors in the multivariate analysis. The CHIEF predictive scoring system was developed using 5 relevant variables (carbapenem use, hematologic malignancy, IBD, enteral feeding, fluoroquinolone use); the area under the receiver operating characteristic curve for the CHIEF score was 0.70. The CHIEF score incorporates useful, clinically available factors and could help identify patients at risk of recurrent CDI. These findings contribute to the understanding of risk factors associated with CDI recurrence and provide support for the development of prevention strategies.