A prospective, hospital-based, multicenter study was undertaken to identify the reasons for hospital admission, to describe antibiotic treatment before and during hospitalization, and to determine the outcome of community-acquired pneumonia (CAP). Data collected included prehospital management of CAP, Pneumonia Outcome Research Team (PORT) classification on admission, in-hospital antibiotic treatment, and predictors of death within 30 days. Among the 215 patients (mean age, 66.7 years; M:F ratio, 1.1) recruited, 24 (11.2%) were living in nursing homes. CAP had been diagnosed prior to admission in 55 (25.6%) patients. At admission, 75 (34.9%) patients had a low risk of death (PORT classification I-II). A pathogen was isolated for 55 (25.6%) patients, primarily Streptococcus pneumoniae (n=18), atypical agents (n=16), influenza virus (n=10), and respiratory syncytial virus (n=4). Amoxicillin (with or without clavulanate), cefotaxime, or ceftriaxone monotherapy was prescribed to 121 (56.3%) patients. Dual combination therapy was prescribed to patients at higher risk of death (PORT classification III-V; OR, 3.09). Mortality was 7%. Logistic-regression analysis identified nursing-home residency (OR, 8.36), serum creatinine > or =88 micromol/l (OR, 7.88), and Pneumonia Outcome Research Team classification (OR, 1.02) as independent predictors of death. CAP remains a serious disease for elderly persons living in nursing homes. This population should benefit from immunization with pneumococcal and influenza vaccines.