An enzyme-linked immunosorbent assay (ELISA) was developed to measure total amoxicillin concentrations penetrating lung secretions, which were compared with "active" concentrations measured by conventional bioassay. An antibody was raised in rabbits to amoxicillin conjugated to bovine serum albumin and used in a competitive binding ELISA (sensitivity, 10 ng/ml; precision [coefficient of variation], 9%). The measurement of amoxicillin in lung secretions by using the ELISA method was verified by high-performance liquid chromatography. Amoxicillin concentrations were found to be similar in both whole sonicated sputum and sol-phase sputum obtained by ultracentrifugation following single oral doses of 3 g (4.6 mg/liter for sonicated and 4.7 mg/liter for sol-phase preparations) and 250 mg (0.23 mg/liter for both preparations). Eight patients with bronchiectasis received 500 mg of amoxicillin three times daily. On the second day of therapy (4 h after the morning dose), the mean concentration of amoxicillin in sputum was 0.88 mg/liter (standard error of the mean [SEM], 0.11) by ELISA and 0.40 mg/liter (SEM, 0.05) by bioassay, suggesting a significant degree of local inactivation. This difference between total and active amoxicillin levels was found to correlate significantly (r = 0.693; P < 0.05) with -lactamase levels (mean, 29.5 mU/ml; SEM, 9.4). A pharmacokinetic study on day 3 revealed maximum levels in secretions 2 to 4 h after dosing (mean, 1.36 mg/liter-, SEM, 0.26). At the end of successful therapy (day 14), total and active levels were lower (mean, 0.48 mg/liter; SEM, 0.11 [total]; mean, 0.21 mg/liter-, SEM, 0.06 [active]); this result was associated with a reduction in lung inflammation (decreased serum-derived albumin in the lung secretions). In conclusion, antibiotic penetration is partly dependent on the degree of lung inflammation. The differences observed in total and active levels of amoxicillin and the relationship to 13-lactamase activity in sputum suggest why higher doses of antibiotic may be required to produce a therapeutic response in some patients.Conventional doses of ,-lactam antibiotics often fail to elicit a clinical response in patients with chronic bronchial sepsis associated with the chronic lung disease bronchiectasis (2,10,13,20). Recent studies have shown that these patients require higher doses of oral P-lactam antibiotics or even direct airway delivery of the drug to clear their purulent secretions (3,4,5,10,29), suggesting that factors governing local antibiotic concentrations are critical in determining the clinical efficacy.In order for antibiotic therapy to be effective, the concentrations of the antibiotic in lung secretions should, theoretically, exceed the MIC or MBC for the colonizing bacteria (19, 23). Conventional doses of antibiotics used in the treatment of patients with chronic bronchial sepsis associated with bronchiectasis frequently fail to achieve these concentrations (2, 3). Effective antibiotic concentrations in lung secretions are dependent on several factors, inclu...