In a prospective, randomized, controlled, three-arm study, the pharmacokinetics of hydroxyurea administered as an antiviral agent in patients infected with human immunodeficiency virus type 1 (HIV-1) were evaluated. The three arms of the study consisted of azidothymidine (AZT) 250 mg twice daily, hydroxyurea 500 mg twice daily, or a combination of the two. Nine patients receiving hydroxyurea in monotherapy (n = 4) or in combination with AZT (n = 5) agreed to undergo multiple venipunctures for pharmacokinetic analysis. Sample collection was performed at steady-state conditions and serum concentration-time data for hydroxyurea were fitted using a one-compartment model. Mean (+/- standard deviation) peak concentration (Cmax) was 0.135 +/- 0.06 mmol/L and mean trough level (Cmin) was 0.0085 +/- 0.003 mmol/L. Mean concentration at steady state was 0.045 +/- 0.006 mmol/L. Apparent clearance (Cl/F) was 0.18 +/- 0.005 L/hr/kg, and half-life (t1/2) was 2.5 +/- 0.5 hours. Hydroxyurea given orally to patients infected with HIV-1 was well absorbed from the gastrointestinal tract, with a tmax of 0.85 to 0.96 hours after ingestion. Serum levels of hydroxyurea ranged from 0.01 to 0.13 mmol/L. These values are similar to the concentrations (between 0.01 and 0.1 mmol/L) demonstrated to inhibit HIV-1 in vitro. Our data show that hydroxyurea given at a dosage of 500 mg twice daily is sufficient to yield serum concentrations potentially useful for in vivo inhibition of HIV-1.