ED Prevalence of C. Diff Infection
Abrahamian, Fredrick M., et al. “Clostridium difficile Infection Among US Emergency Department Patients With Diarrhea and No Vomiting.” Annals of Emergency Medicine (2017).
C. Diff infection has increasingly become a source of morbidity and mortality, especially for patients recently hospitalized or treated with antibiotics. Diarrhea, however, is a relatively common chief complaint, and emergency departments require more data to guide testing and treatment. This study narrowed its focus to patients with diarrhea but not vomiting, in order to identify a subgroup more like to have C. Diff rather than other common causes of gastroenteritis such as norovirus (which typically also cause vomiting), and attempted to quantify prevalence and risk factors for C. Diff.
10 US Emergency Departments prospectively enrolled cases of diarrhea (at least 3 loose stools within 24 hours) without vomiting, who were all tested with stool culture and toxin assays for C. Diff (both needing to be positive to confirm a diagnosis of C. Diff).
Of the 422 patients enrolled, 43 (10%) tested positive for C. Diff infection. Statistically significant risk factors were recent antibiotic use (26% of patients overall, 56% in the C. Diff-positive group), recent hospital admission (27% overall, 44% in the C. Diff-group), and prior C. Diff infection (3% overall, 10% in the C. Diff-group). 40% of patients had no major risk factors, with a 7% prevalence among these patients. Many of the patients (40%) reported greater than 10 stools within 24 hours, but this was not predictive of C. Diff infection, and nor were laboratory findings or other comorbidities. There was a wide range from 2% to 30% of prevalence among the hospital sites, suggestive of some enrollment bias (our hospital was one of the sites, and had an overall prevalence of 6%).
Diarrhea without vomiting may prompt providers to consider, test for, or empirically treat for C. Diff diarrhea, especially in the presence of recent admission or antibiotics. However, overall prevalence varies geographically so clinicians should use local data to avoid harms associated with overtesting and overtreatment.