Thirty-seven children and adults aged 0.2-82 years were treated intravenously with 150 mg of trimethoprim (TMP) and 750 mg of sulfamethoxazole (SMZ)/m2 every 8 hr, usually for known or suspected pneumocystis pneumonia; when necessary dosage was adjusted to maintain peak TMP levels of 5-10 micrograms/ml. On day 2 of treatment, mean peak levels of TMP-SMZ were 7.02 and 148 micrograms/ml, respectively, and mean half-lives were 9.6 and 10.7 hr, respectively. All age groups achieved similar peak levels of TMP-SMZ, although dosages per weight were higher in children than in adults. Peak increments (peak levels minus levels before infusion) were higher and more reliable after iv than after oral dosage (P less than 0.001). The half-lives of TMP and SMZ increased with age (r = +0.73 and +0.39, respectively) and were correlated directly with the level of serum creatinine (r = +0.85 and +0.39, respectively). Serum concentrations of N4-acetyl-SMZ, the major hepatic metabolite of SMZ, increased in proportion to concentrations of creatinine in serum (r = +0.92; P less than 0.001). Adverse effects included fluid overload due to the large dilution volume and thrombocytopenia, which was associated with higher serum TMP levels and longer treatment as compared with nonthrombocytopenic patients. A loading dose of 250 mg of TMP and 1,250 mg of SMZ/m2 is recommended, followed by maintenance doses of 150 mg of TMP and 750 mg of SMZ/m2 every 8 hr for children aged 10 years or younger and every 12 hr for adults with normal renal function. In renal failure the dosage interval (hr) should be increased to 12 times the serum creatinine level (mg/dl) (maximum, 48 hr). Serum concentrations of TMP and perhaps of N4-acetyl-SMZ should be monitored in patients with severe renal failure.