Trimethoprim and trimethoprim-sulphamethoxazole (co-trimoxazole) are often prescribed in HIV patients treated with zidovudine. The pharmacokinetics of zidovudine, after a dose of 3 mg kg-1 by constant rate intravenous infusion over 1 h were evaluated in nine HIV patients in an open, randomized, three-phase crossover study, without and with trimethoprim (150 mg) and trimethoprim-sulphamethoxazole (160 and 800 mg). The metabolic clearance of zidovudine was not significantly influenced by trimethoprim-sulphamethoxazole and trimethoprim. However, the renal clearance of zidovudine was decreased by 58 and 48%, respectively, and that of its glucuronide by 27 and 20% (P < 0.05). The fraction of the dose excreted as the parent compound fell by 47 and 39% and the metabolic ratio by 48 and 43% (P < 0.05). This kinetic drug interaction, apparently due solely to trimethoprim, may only be clinically important when hepatic glucuronidation is also impaired by liver disease or inhibited by other drugs.