A multicentre, randomized study was performed to compare the clinical and bacteriological efficacy of 500 mg ceftazidime i.v. t.d.s. with 1,000 mg ceftazidime i.v. t.d.s. for treatment of hospitalised, non-compromised patients with gram-negative infections. The study was conducted in ten hospitals in The Netherlands. Hospitalised patients with a suspected gram-negative lower respiratory tract infection, complicated urinary tract infection or septicaemia were included. Excluded were patients with neutropenia, limited life expectancy, or severe renal insufficiency as well as those on antibiotics in the 48 h prior to entry. Ceftazidime was administered via an intravenous infusion every 8 h. For patients with moderately impaired renal function the frequency was reduced to 12 h. Treatment was continued for as long as clinically indicated. Clinical response (cure, improvement or failure) and bacteriological response (elimination, persistence or non-evaluable) were assessed primarily by the investigator. Final assessments were made by a panel of experts without prior knowledge. In total 127 patients were randomized, 64 patients to the 500 mg group and 63 to the 1,000 mg group; 47 patients were excluded from evaluation, usually due to an incorrect diagnosis prior to randomization. Ultimately 37 patients of the 500 mg group and 43 patients of the 1,000 mg group were available for evaluation. Between these two groups of evaluable patients there were no significant differences in baseline characteristics, types of infection, isolated bacterial pathogens or treatment characteristics.(ABSTRACT TRUNCATED AT 250 WORDS)