The pharmacokinetics of ceftriaxone (CRO) after subcutaneous and intravenous administration was studied in 10 healthy volunteers (5 males, 5 females, age 22-43 years, body weight 64.3 +/- 9.5 kg). Each of them received 2.0 g CRO i.v., and then 0.5 g i.v. and 0.5 g CRO s.c. in a randomized cross-over design. Subcutaneous administration of ceftriaxone was tolerable in combination with lidocaine. Serum and urine concentrations were determined by high performance liquid chromatography and for comparison with a bioassay. Mean serum concentrations were high after intravenous administration: 258 +/- 40 mg/1 (0 min, 2 g i.v.) resp. 84 +/- 40 mg/1 (0 min, 0.5 g i.v.). They declined to 11.6 +/- 4.2 mg/1 (2 g) resp. 6.5 +/- 2.2 mg/1 (0.5 g i.v.) within 24 h. Following subcutaneous application peak serum concentrations of 37.1 +/- 5.6 mg/1 were found after 138 +/- 49 min and a mean serum concentration of 6.6 +/- 1.6 mg/1 after 24 h. Concentrations of free ceftriaxone, determined by ultrafiltration, were 9.2 +/- 2.7% of total concentrations from 25 to 200 mg/1. Cumulated urine recoveries over a period of 24 h were: 51.2 +/- 8.9% (2 g i.v.), 47.1 +/- 7.9% (0.5 g i.v.) and 39.7 +/- 9.5% (0.5 g s.c.). There was no evidence for the presence of a microbiologically active metabolite in urine. Comparison of pharmacokinetic parameters for the 0.5 g dose did not show relevant differences between intravenous and subcutaneous administration (using an open two-compartment model): t beta 1/2 (min) 514 +/- 104 (s.c.), 592 +/- 133 (i.v.), VD,Area (1) 11.9 +/- 3.8 (s.c.), 13.3 +/- 3.8 (i.v.) and AUCtot (mg X h/1) 515 +/- 106 (s.c.) and 549 +/- 125 (i.v.). Bioavailability of the subcutaneous application was 0.96 +/- 0.26. For the 2 g i.v. dose the known nondosis-dependent kinetics was observed. Subcutaneous administration of ceftriaxone appears to be a possible alternative to the intravenous route in selected clinical situations or cases.