Predictors of hospital-onset Clostridioides difficile infection in children with antibiotic-associated diarrhea.

• Significant predictors of Clostridioides difficile infection (CDI) included presence of abdominal pain, exposure to meropenem and clindamycin, longer duration of hospital stay and antibiotic treatment. • The new scoring system showed good discrimination, with an AUROC of 0.88. • The optimal cut-off point for the score was 6, with a sensitivity and specificity of 89% and 74%, respectively. This study aimed to determine the predictors of hospital-onset Clostridioides difficile infection (CDI) in pediatric patients with antibiotic-associated diarrhea (AAD) and to develop a predictive scoring system to identify at-risk patients. This retrospective case-control study included patients aged ≥2-18 years with AAD who underwent C. difficile polymerase chain reaction testing >3 days after hospital admission. Patients with hospital-onset CDI were selected as cases and matched with the control patients without CDI. Univariate and multivariate logistic regressions were used to determine predictors of CDI and to construct a prediction score for the outcomes of interest. Sixty-five patients with hospital-onset CDI and 130 controls were enrolled. Independent predictors for CDI identified and combined into the prediction score included abdominal pain (adjusted odds ratio [95% confidence interval]: 7.940 [3.254-19.374]), hospitalization for ≥14 days before the onset of diarrhea (3.441 [1.034-11.454]), antibiotic use for ≥10 days before the onset of diarrhea (6.775 [1.882-24.388]), receipt of meropenem (4.001 [1.098-14.577]) and clindamycin (14.842 [4.496-49.000]). The area under the receiver operating characteristic curve for this score was 0.883. The presented scoring system can be easily applied by clinicians at the bedside to decide which patients with AAD are likely to have CDI. [ABSTRACT FROM AUTHOR]