Two dosages of clarithromycin were compared for treatment of disseminated Mycobacterium avium disease of AIDS patients: high-dose (HD): 1,000 mg or 750 mg b.i.d. according to body weight, and low-dose (LD): 1,000 mg or 750 mg q.d. Patients with high probability of M. avium positive blood culture on day 0 received a 42-day clarithromycin treatment with HD (n = 27) or LD (n = 28) at random after stratification according to body weight. Assessment procedures, including quantitative blood cultures, were performed at days 14, 28 and 42. Forty-five patients were eligible for clinical and 28 for bacteriological evaluation. Bacteriological success was observed in 12 HD and 11 LD patients, partial success in one HD and two LD and failure in none of the HD and two LD (p = 0.33). Between days 0 and 42, log decreases in CFU counts/ml were (mean +/- SD) 3.13 +/- 0.82 (HD) and 2.67 +/- 1.8 (LD) (p = 0.38). Fever and night sweats significantly improved similarly in both groups; no change in spleen and liver size was observed on CT scans. Eight patients died during the study but no death was reported as drug related. Sixteen patients (HD = 6, LD = 10) discontinued the treatment because of side effects. A trend towards improved bacteriological effectiveness and reduced tolerance was observed in the HD group but the difference was not significant. With a power of 0.70, no dose effect was demonstrated between the two tested dosages. A daily dose of 1,000 mg clarithromycin was tested in drug combinations to treat disseminated M. avium infection in AIDS patients.