Clinicians are frequently faced with patients in whom the radiographic resolution of community-acquired pneumonia seems delayed. Previous studies of radiographic resolution of the disease have yielded conflicting results. We prospectively assessed the radiographic resolution of pneumonia in 81 non-immuno-compromised patients, presenting to the emergency room and ambulatory clinics of a large university hospital, who met clinical and radiographic criteria for pneumonia. Serial chest radiographs were obtained every 2 wk for an initial period of 8 wk, and then every 4 wk until 24 wk had passed, or until all radiographic abnormalities had cleared. Forty-one of the 81 patients (50.6%) demonstrated complete clearance after 2 wk. Fifty of the 75 patients (66.7%) followed to 4 wk demonstrated complete clearance. The rate of clearance was inversely correlated with age (p < 0.001) and involvement of single versus multiple lobes (p < 0.0001) (log-rank test). Clearance was faster in those patients treated as outpatients (3.8 wk versus 9.1 wk, p = 0.03) and in patients who were nonsmokers (4.5 wk versus 8.4 wk, p = 0.05) (log-rank test). Multivariate regression analysis demonstrated that only age (relative risk for clearance, +0.79 per decade) and single versus multiple lobes involved (relative risk for clearance, 0.55 for more than one lobe) had independent predictive value (Cox proportional hazards regression model). The radiographic resolution of pneumonia occurs more rapidly in younger patients and in those with only a single lobe involved.(ABSTRACT TRUNCATED AT 250 WORDS)