Zidovudine is widely prescribed for the treatment of human immunodeficiency virus (HIV) infection. Trimethoprim and dapsone are commonly used in the management of Pneumocystis carinii pneumonia in HIV-infected patients. To examine the pharmacokinetic interactions among these drugs, eight HIV-infected patients (26 to 43 years old) with a mean CD4 count of 524.4 +/- 405.7 cells per mm3 received zidovudine (200 mg), trimethoprim (200 mg), and dapsone (100 mg) as single agents and in two- and three-drug combinations. Blood and urine samples were collected at a specified time and analyzed for zidovudine, zidovudine-glucuronide, trimethoprim, dapsone, and monoacetyl-dapsone concentrations under single-dose and steady-state conditions. Zidovudine did not influence the pharmacokinetic disposition of dapsone or trimethoprim. Dapsone had no effect on the pharmacokinetic disposition of zidovudine. Trimethoprim significantly decreased the renal clearance of zidovudine by 58% (5.0 +/- 1.8 versus 2.1 +/- 0.5 ml/min/kg of body weight [P < 0.05]). There was a concurrent 54% decrease in the mean urinary recovery of zidovudine (11.7 +/- 3.5 versus 5.4 +/- 3.0 [P < 0.05]), and the metabolic ratio was decreased by 78% (0.32 +/- 0.4 versus 0.07 +/- 0.05 [P < 0.05]). The mean area under the concentration-time curve from 0 to 6 h of the zidovudine-glucuronide/ zidovudine ratio was unchanged. We conclude that zidovudine, trimethoprim, and dapsone can be coadministered to patients with AIDS without significant pharmacokinetic interaction. However, in AIDS patients with liver impairment and impaired glucuronidation, doses of zidovudine may need to be decreased.