A study was conducted to determine whether a direct relationship exists between beta-lactam and/or aminoglycoside activity measured in vitro and bacteriological outcome in acute pulmonary exacerbations of cystic fibrosis. Twenty-seven patients, aged between 6 months and 24 years (mean age 10 1/2 years), were included in the study and received 41 i.v. courses of a beta-lactam agent combined with an aminoglycoside. A total of 63 Pseudomonas aeruginosa strains were found in sputum taken on admission at densities exceeding 10(6) cfu/g of sputum. For each episode, the serum inhibitory quotient (SIQ) and the serum bactericidal quotient (SBQ) of the beta-lactam agent and of the aminoglycoside administered were determined for the Pseudomonas aeruginosa isolate(s). The SIQs and SBQs were calculated by dividing the average peak serum levels achievable in the patients by the minimal inhibitory concentrations and minimal bactericidal concentrations, respectively. The SIQs and SBQs were compared to bacteriological outcome. Bacteriological success was defined as a decrease of 2 log10 counts or more in the Pseudomonas aeruginosa density in sputum between days 0 and 7 of therapy. The SIQ and SBQ of beta-lactam agents were good predictors of bacteriological outcome: SIQs of < 1:16 were 100% predictive of failure (chi 2 28; p < 0.001) and of > or = 1:64 were 92.9% predictive of success (chi 2 35.68; p < 0.001); SBQs of < 1:8 were 100% predictive of failure (chi 2 42.78; p < 0.001) and of > or = 1:32 were 95.8% predictive of success (chi 2 31.5; p < 0.001). Aminoglycoside SIQs and SBQs were not predictive of outcome.