A case-control study of Clostridioides difficile symptomatic infections in a pediatric cancer hospital

Abstract

Objective: The aim of this study was to analyze and identify documented infections and possible risk factors for Clostridioides difficile infections in children with cancer. Methods: This is a retrospective case-control study, carried out in a pediatric cancer hospital, covering the years 2016–2019. Matching was performed by age and underlying disease, and for each case, the number of controls varied from 1 to 3. Logistic regression models were used to assess risk factors. Results: We analyzed 63 cases of documented infection by C. difficile and 125 controls. Diarrhea was present in all cases, accompanied by fever higher than 38°C in 52.4% of the patients. Mortality was similar among cases (n=4; 6.3%) and controls (n=6; 4.8%; p=0.7). In all, 71% of patients in the case group and 53% in the control group received broad-spectrum antibiotics prior to the infection. For previous use of vancomycin, the Odds Ratio for C. difficile infection was 5.4 (95% confidence interval [95%CI] 2.3–12.5); for meropenem, 4.41 (95%CI 2.1–9.2); and for cefepime, 2.6 (95%CI 1.3–5.1). For the antineoplastic agents, the Odds Ratio for carboplatin was 2.7 (95%CI 1.2–6.2), melphalan 9.04 (95%CI 1.9–42.3), busulfan 16.7 (95%CI 2.1–134.9), and asparaginase 8.97 (95%CI 1.9–42.9). Conclusions: C. difficile symptomatic infection in children with cancer was associated with previous hospitalization and the use of common antibiotics in cancer patients, such as vancomycin, meropenem, and cefepime, in the last 3 months. Chemotherapy drugs, such as carboplatin, melphalan, busulfan, and asparaginase, were also risk factors.