In an open, multicenter randomized clinical trial on patients with acute pneumonia, 5-day therapy using a new oral macrolide antibiotic, azithromycin (Az), was compared with standard 10-day therapy using another macrolide, josamycin (J). Eighty-nine patients were treated (46 Az, 43 J); both groups were comparable, except in terms of the sex ratios. The causative agent was determined in 31 cases; streptococcal pneumonia was the most common diagnosis (48%). Cure rates were 93% for J and 80% for Az (p greater than 0.30), as judged by a fall in body temperature and an improvement in clinical condition and in radiographic findings. Among 6 failures, there were 4 cases of empyema (1 J, 3 Az). Of 6 deaths, 2 were pneumonia-dependent (1 J and 1 Az, Haemophilus and Streptococcus pneumoniae, respectively). We conclude that 5-day Az is as effective as 10-day J and that such short-course therapy is an advantage in the treatment of pneumonia caused by macrolide susceptible pathogens.