Predicting Clostridium difficile toxin in hospitalized patients with antibiotic-associated diarrhea.

Objective: Clostridium difficile infection is implicated in 20%-30% of cases of antibiotic-associated diarrhea. Studying hospitalized patients who received antibiotic therapy and developed diarrhea, our objective was to compare the clinical characteristics of patients who developed C. difficile-associated diarrhea (CDAD) with those of patients with a negative result of a stool assay for C. difficile toxin.

Methods: A prospective study was done with a cohort of 217 hospitalized patients who had received antibiotics and developed diarrhea. Patients with CDAD were defined as patients who had diarrhea and a positive result for C. difficile toxin A/B by an enzyme immunoassay of stool. The variables that yielded a significant difference on univariate analysis between patients with a positive assay result and patients with a negative assay result were entered into a logistic regression model for prediction of C. difficile toxin.Setting. A 900-bed tertiary care medical center.

Results: Of 217 patients, 52 (24%) had a positive result of assay for C. difficile toxin A/B in their stool. The logistic regression model included impaired functional capacity, watery diarrhea, use of a proton pump inhibitor, use of a histamine receptor blocker, leukocytosis, and hypoalbuminemia. The area under the receiver operating characteristic curve for the model as a predictor of a positive result for the stool toxin assay was 0.896 (95% confidence interval, 0.661-1.000; P<.001), with 95% specificity and 68% sensitivity.

Conclusions: Our results may help clinicians to predict the risk of CDAD in hospitalized patients with antibiotic-associated diarrhea, to guide careful, specific empirical therapy, and to direct early attention to infection control issues.