A Retrospective Analysis of Risk Factors for Severe Clostridium Difficile Infection In Acute Leukemia and Aggressive Lymphoma.

Abstract Abstract 3797 Introduction and Goals. Clostridium difficile infection (CDI) is increasing in frequency and severity. Risk factors for severity of illness and treatment recommendations for patients with concurrent aggressive hematologic malignancies have not been investigated, despite data showing that this population experiences a high rate of CDI. The goal of this study was to assess mortality associated from CDI as well as risk factors for severe CDI in a population of patients with acute hematologic malignancies. Methods. Ninety-three cases of CDI among 78 patients with active acute leukemia or aggressive lymphoma at the University of Virginia (UVa) Health System from 2004–2009 were studied retrospectively. Rates of all-cause mortality, attributable mortality secondary to CDI (defined as CDI with at least one of the following: active diarrhea, toxic megacolon, or pseudomembranes on colonoscopy at time of death), and a composite outcome (defined as at least one of the following: attributable mortality, ICU stay from CDI, pseudomembranous colitis, toxic megacolon, or colectomy) were assessed as adverse outcomes. The following potential risk factors for severe CDI were investigated, adjusting for age: onset and duration of neutropenia, exposure to high dose cytarabine, number of antimicrobials given, exposure to fluoroquinolone prophylaxis for neutropenic fever, and a designation of severe CDI as previously described (at least two of the following: white blood count >15,000 cells/mL, age >60 years, albumin <2.5 g/dL, temperature >38.3° C, or one of the following: ICU stay or pseudomembranous colitis). Results. All-cause mortality was 21.7%. CDI-attributable mortality was 14.1%. The rate of the composite outcome was 20.3%. All-cause mortality was higher in patients with leukemia and lymphoma than for all other patients hospitalized at this institution with CDI in 2008 (12.5% all-cause mortality, 76/610, p=0.015, 95% CI [13.2, 32.6]). Duration of neutropenic episode >14 days and neutropenia at diagnosis of CDI were associated with increased risk of attributable mortality (OR 3.12, 95% CI [0.82, 11.92], p=0.09; OR 3.39, 95% CI [0.57, 20.05], p=0.11). Cumulative duration of neutropenia >21 days was not associated with increased risk in adverse outcomes. Designation of severe CDI predicted an increase in all-cause mortality (OR 6.7, 95% CI [2.1, 21.5], p <0.001), attributable mortality (OR 11.5, 95% CI [2.3, 57.5], p <0.001), and the composite outcome, (OR 19.3, 95% CI [5.5, 74.7], p <0.001). Fluoroquinolone prophylaxis for neutropenic fever, high dose cytarabine, induction chemotherapy, or receipt of more than 2 concurrent antimicrobials were not associated with increased risk of all-cause mortality, attributable mortality or the composite outcome. Discussion. Patients with acute hematologic malignancies and concurrent CDI were at increased risk of all-cause and CDI-attributable mortality compared with similar data from UVa of all hospitalized patients. Prolonged neutropenia at time of CDI diagnosis was associated with an increased risk of CDI-attributable mortality, although this result was not statistically significant. Established criteria for diagnosing severe CDI were helpful at predicting poor outcomes in this population. Seven of the 10 neutropenic patients who died in this study did not receive oral vancomycin. Neutropenic patients with CDI may benefit from initial treatment with oral vancomycin instead of metronidazole, especially if other criteria for severe illness are present. Future research may assess effectiveness of oral vancomycin as initial treatment for CDI in neutropenic patients. Disclosures: No relevant conflicts of interest to declare.