Absorption and disposition characteristics of carprofen were compared in the dog after intravenous, oral, and rectal administrations. Rectal formulations included an aqueous solution and a suppository. Single doses of 100 mg carprofen were given in a cross-over design and plasma concentrations of unchanged drug were determined by HPLC. Plasma concentration-time profiles could be adequately described after intravenous and extravascular administrations by tri-and biexponential functions, respectively. After intravenous applications the basic disposition parameters could be determined: mean ( f S.D.) elimination half-life was about 11.7 (f3.1) h; volume of distribution ranged from 0.12 to 0.22 I kg-' and total plasma clearance was about 0.017 f 0.003 1 h kg-'. After oral dosing, carprofen was rapidly absorbed (time of maximum concentration: 0.83 f 0.61 h) and comparison with the intravenous solution indicated complete bioavailability. After rectal administration, the rate of absorption evaluated through t,,, and calculation of mean absorption times was always slower than after oral dosing. Relative bioavailabilities of the drug from the suppository were about 20 per cent lower than from rectal solutions. No significant difference in rates of absorption of carprofen from rectal solution and suppository was seen; this allowed the conclusion that drug release from the semi-solid dosage form was not the rate-limiting step in carprofen absorption from the suppository. From the present study, it is concluded that rectal administration of carprofen offers an alternative to the oral route of drug intake.