This study investigated the pharmacokinetics and activity of gallium nitrate in non-small cell lung cancer when 700 mg/m2 was given as a 30-min infusion with prehydration every 2 weeks. Gallium was measured in plasma and urine using flameless atomic absorption spectrophotometry, and pharmacokinetics of total and ultrafilterable gallium were calculated. Twenty-five patients with non-small cell lung cancer received 1-12 (median 2) courses of gallium nitrate every 2 weeks. Of 21 patients evaluable for response, 1 partial response was recorded, 4 patients had stable disease. and 16 had progressed. The most serious toxicities were renal impairment and optic neuritis. Hypocalcaemia was recorded in 3 patients. The mean C(max) was 15.2 +/- 3.1 microg/ ml (range 9.5-21.2). Most gallium remained ultrafilterable for the first 10 h, after which plasma protein binding increased, and at 48 h only 11% was present as ultrafilterable gallium. The elimination profiles of both total and ultrafilterable gallium were biphasic, and the distribution phase consisted of ultrafilterable gallium, with a distribution half-life of 1.4 h. Total gallium plateaued at 1.9 microg/ml at between 8 and 12 h, and the estimated elimination half-life was 63 h. The elimination half-life of ultrafilterable gallium was 16.5 h. Inter- and intra-patient variability in pharmacokinetics was minimal. A mean of 50 +/- 14% of the gallium dose was excreted in the urine within 48 h. A short infusion of gallium nitrate achieving high peak plasma concentrations results in little efficacy in non-small cell lung cancer.