Seventy-two hospitalized patients with pneumonia or bacteremia were randomly allocated to receive ceftriaxone 2 g once daily i.v. or ceftazidime 2 g twice a day i.v. At the end of the study 60 patients were evaluable, 31 in the ceftazidime group and 29 in the ceftriaxone group. Thirty-four patients (ceftazidime = 15, ceftriaxone = 19) yielded one or more pathogens, of which 64% were gram-negative bacilli. Clinical cure or improvement was observed in 90% of patients in both groups. All 3 cases of bacteremia were cured. Three patients in each group failed to respond to the administered drug. Eradication of the pathogen(s) was observed in 82% of the ceftazidime group and in 86% of the ceftriaxone group. Two episodes of superinfection due to Pseudomonas aeruginosa were recorded in the ceftriaxone group, while Candida spp. was isolated from the sputum in 2 patients in the ceftazidime group. Three strains of P. aeruginosa (2 in the ceftazidime group, 1 in the ceftriaxone group) persisted despite the treatment. No side effects were seen except for skin rash in 2 patients receiving ceftazidime. Compliance was good in both groups, particularly with the once daily administration of ceftriaxone. Overall ceftriaxone and ceftazidime appear to be equally effective in the treatment of nosocomial pneumonia, with the exception of P. aeruginosa infection.