No abstract

OBJECTIVE To assess nosocomial infection (NI) as a risk factor for death and to estimate the population-attributable risk of death from NI. DESIGN. A prospective cohort study of patients with and without NI.SETTING. Nimes University Hospital, Nimes, France.PATIENTS. Patients were recruited from May 7, 2001, to January 10, 2003. Patients in acute care and long-term care units who had NI were enrolled, and patients without NI were randomly selected and matched with patients with NI for age, sex, type of care (acute care vs. long-term care) and length of stay in hospital at study inclusion. OUTCOME MEASURES. Vital status within 60 days after study inclusion was assessed. We used conditional logistic regression to estimate the relative death risk from NI after adjusting for comorbidities, severity of the underlying disease, and all other confounding factors. The adjusted population-attributable risk was assessed using the Mantel-Haenszel method.RESULTS. We recruited 1,914 patients with NI and 5,172 patients without NI. The median age of the patients with NI was 73 years; 1,045 (54.6%) were female. NI was associated with death within 60 days (adjusted odds ratio, 1.7 [95% confidence interval {CI}, 1.4-;2.2]; P<.001). The adjusted population-attributable risk of death for all sites of infection was 1.7% (95% CI, 1.4-2.1). If we consider the NI incidence to be 3%-6% in French hospitals, the population-attributable risk of death from NI would range from 2.1% (95% CI, 1.7%-2.5%) to 4.0% (95% CI, 3.3%-4.9%).CONCLUSION. In this study, NI appeared to have a significant impact on mortality. Multicenter studies will be needed to confirm these results.