A prospective cohort study on hospital mortality due to Clostridium difficile infection

Abstract

Purpose: Although an increase in burden of disease has frequently been reported for Clostridium difficile infection (CDI), specific data on the effect of CDI on a patient's risk of death or overall hospital mortality are scarce. Therefore, we performed a prospective cohort study to analyse the effect of CDI on the risk of pre-discharge all-cause death in all inpatients with CDI compared to all inpatients without CDI during 2009 in a single hospital.

Methods: Clostridium difficile infection was defined as by the European Society of Clinical Microbiology and Infectious Diseases. Data were collected from the medical charts of CDI patients and from the hospital discharge data of non-CDI and CDI patients. The effect measures of CDI used to compute the risk of pre-discharge all-cause death were risk ratio, attributable risk, mortality fraction (%) and population attributable risk percentage. Co-morbidity was categorized using the Charlson co-morbidity score in which a value of ≤2 was defined as low co-morbidity and that of >2 as moderate/severe co-morbidity. A stratified analysis and a Poisson regression model were applied to adjust for the effects of the risk factors sex, age and severity of co-morbidity.

Results: A total of 185 hospitalized patients with CDI were compared to 38,644 other hospitalized patients without CDI admitted between 1 January 2009 and 31 December 2009. The mean age of the CDI and non-CDI patients was 74.3 (range 72.3-76.4) and 51.9 (range 51.6-52.1) years, respectively. Of the 185 CDI, 136 (73.5%) and 49 (26.5%) were categorized with low and high co-morbidity, respectively, versus 32,107 (83.4%) and 6,352 (16.5%), respectively, in non-CDI patients. Overall, 24 of the 185 CDI patients (13%) versus 1,021 of the 38,459 non-CDI patients (2.7%) died during their hospital stay, resulting in a relative risk of pre-discharge death of 4.89 [95% confidence interval (CI) 3.35-7.13] for CDI patients, a CDI attributable risk of death of 10.3 per 100 patients and a CDI attributable fraction of 79.5 % (95% CI 70.1-86 %). After adjustment for age, sex and co-morbidity the relative risk of pre-discharge death was 2.74 (95% CI 1.82-4.10; p < 0.0001) for patients with CDI, and the proportion of hospital deaths due to CDI was 1.72 (95% CI 1.22-2.05).

Conclusion: The results of this study lead to the conclusion that hospitalized patients with CDI are--independent of age, sex and co-morbidity severity--2.74-fold more likely to die during their hospital stay than all other hospitalized patients. The eradication of CDI in the hospital could have prevented 1.72% of in-hospital deaths in our study population during the 1 year of the study.