The efficacy and tolerance of clarithromycin (250 mg twice daily) were compared with those of roxithromycin (150 mg twice daily) in an open, multicentre trial of 77 inpatients with community-acquired pneumonia. Sixty-five patients were clinically evaluable (34, clarithromycin; 31 roxithromycin). Efficacy was comparable between treatment groups: 26 of 34 patients (76%) treated with clarithromycin were clinically cured, including four with atypical pneumonia. In the roxithromycin group 25 of 31 patients (81%) were clinically cured and one was improved. Cough, appearance of sputum, and fever improved in most patients in both treatment groups. Chest X-rays after treatment showed resolution or improvement in 76% of patients who received clarithromycin and 87% of those who received roxithromycin. The clinical evaluation of the response generally agreed with the bacteriological response. Among patients who were bacteriologically evaluable for four target organisms (Streptococcus pneumoniae, Haemophilus influenzae, H. parainfluenzae, and Branhamella catarrhalis) the pathogen was eradicated in four of seven (57%) in the clarithromycin-treated group and in five of six (83%) in the roxithromycin-treated group. Adverse events were reported in more patients who received roxithromycin (21.6%) than in those who received clarithromycin (12.5%) although the incidences were not statistically significantly different. The majority of adverse events were transient increases in serum alanine aminotransferase, serum aspartate aminotransferase, and alkaline phosphatase. Clarithromycin was shown to be effective and well-tolerated; the clinical efficacy and safety of clarithromycin and roxithromycin were comparable.