Clinical, immunological and microbiological predictors of poor outcome in Clostridium difficile infection

BackgroundClostridium difficile infection (CDI) causes increased morbidity and mortality. Clinical data cannot clearly predict poor CDI outcome. Data on the value of microbiological predictors is scarce.ObjectiveTo identify early predictors of poor outcome of CDI.MethodsWe prospectively included patients with CDI aged >2 years. Clinical, immunological (Toxin B IgG/Ig A and Toxin A IgG/Ig A), microbiological factors (bacterial load, toxin quantification, sporulation, germination, and metronidazole susceptibility) were evaluated to identify early independent predictors of poor outcome.ResultsWe identified 204 cases of CDI; outcome was poor in 22.1%. Advanced age, presence of comorbidities, leukocytosis and high toxigenic C. difficile load were independently associated with poor outcome. We could not demonstrate this correlation for antitoxin antibodies.ConclusionWe identified high bacterial load as a microbiological predictor of poor outcome. We propose this factor to be included in combined clinical and microbiological prediction rules of poor outcome in CDI.