One hundred ten patients were randomized to receive one of the following antibiotic combinations: aztreonam + clindamycin, tobramycin + clindamycin, or amikacin + mezlocillin for the treatment of lower respiratory tract infections (LRTI) caused by gram-negative bacilli. Of the 68 patients who received aztreonam + clindamycin, 60 were clinically evaluable and 50 were bacteriologically evaluable. Of the 60 clinically evaluable patients, 54 were cured and 5 were treatment failures or died during the study period. Of the 50 bacteriologically evaluable patients, 46 were cured and 3 failed to respond to therapy. Of the 26 clinically evaluable patients in the tobramycin + clindamycin group, 22 were cured and 4 either failed to respond or died during the study period. Of 18 bacteriologically evaluable patients in this group, 16 were cured and 2 failed to respond. In the amikacin + mezlocillin group, 14 of the 15 clinically and bacteriologically evaluable patients were cured, and 1 failed to respond. The most commonly isolated pathogens were Klebsiella pneumoniae, Escherichia coli, and Pseudomonas aeruginosa. The very few adverse drug reactions that were seen were transient and comparable in all three groups except for renal function parameters, which deteriorated in 6-8% of patients receiving the aminoglycoside combination. All three antibiotic combinations were similar in effectiveness and safety.