Predictors and outcomes of readmission for Clostridium difficile in a national sample of medicare beneficiaries

BACKGROUND: Rates of Clostridium difficile (CD) infections are increasing. Elderly patients may be at particular risk of recurrent CD infection. Little is known about the risk for CD readmission specifically in this age group. METHODS: A 5% random sample of Medicare data (2009-2011) was queried for patients surviving a hospitalization for CD by ICD-9 code. Demographic (age, sex, gender), clinical (Elixhauser index, gastrointestinal comorbidities), and hospitalization (length of stay, ICU admission) characteristics as well as exposure to antibiotics and interim non-CD hospitalization were compared for those with and without a readmission for CD. A multivariable survival analysis was used to determine predictors of readmission. RESULTS: Of 7,564 patients surviving a CD hospitalization, 8.5% were readmitted with CD in a median of 25 days (interquartile range (IQR) 14-57). In multivariable survival analyses, interim non-CD hospital exposure was the strongest predictor of CD readmission (hazard ration (HR) 3.75 95%, confidence interval (CI) 3.2-4.42). Oral and intravenous/intramuscular (IV/IM) antibiotic use, Elixhauser index, and CD as the primary diagnosis also increased the risk of CD readmission. Discharge to hospice, long-term care or a skilled nursing facility decreased the odds of CD readmission. CONCLUSION: Hospital exposure and antibiotic use put elderly patients at risk of CD readmission. Exposure to these factors should be minimized in the immediate post discharge period