Data are reported from three step-wise pharmacokinetic studies in 43 patients who received acyclovir. Dosage regimens began at 0.5 mg/kg administered as a single dose intravenously and were increased to 15 mg/kg per dose given three times daily for five days. All patients evaluated were immunocompromised by underlying disease or received cytolytic and/or cytotoxic therapy. Patients with virologically confirmed herpes simplex or zoster infections were assessed in the multiday, multidose pharmacokinetic trial. Postinfusion plasma concentrations of acyclovir declined in a biphasic manner such that the plasma-concentration time data were fitted by a two-compartment, open-model with zero-order input. The drug's half-life showed little variation with a mean of 3.16 +/- 0.20 hours. In both single-dose and multiple-dose studies there was dose proportionality with peak plasma levels and area under the plasma concentration-time curve indicating dose-independent pharmacokinetics. The kidney was the principal route of drug clearance with a mean recovery of 60 +/- 12 percent. Renal clearance exceeded creatinine clearance indicating renal tubular secretion of drug. No significant clinical or laboratory evidence of toxicity appeared. These studies provide a foundation for the evaluation of acyclovir in controlled trials.