A pharmacokinetic study was performed in 13 adult patients with acute nonlymphoblastic leukemia to compare two formulations of 4'-(9-acridinylamino)-methanesulphone-m-ansidide (AMSA): the original formulation, AMSA-NCL, and a water-soluble lyophilized formulation, AMSA-lactate (Bristol Myers, Syracuse, N.Y. USA). Initially, the patients received either AMSA-NCL or AMSA-lactate, 75-90 mg/m2 daily, for 3-7 days as a 1-h infusion. Eight patients subsequently crossed over to receive the other formulation. Plasma samples for drug determination were collected during the first 3 days. A new method for determination of AMSA is described. Acidified plasma samples containing an internal standard were extracted with hexane, then made alkaline, whereafter, AMSA was extracted with ethylacetate. Extracts were reconstituted in absolute ethanol and analyzed by high-pressure liquid chromatography (HPLC) using a reverse-phase C-18 column and UV detection at 254 nm. There were no clear differences in clinical effects and toxicity between the two formulations. Patients with the highest total area under the drug concentration-versus-time curves (AUCs) for plasma concentrations versus time had significantly lower nadir for white blood cell count, suggesting a relation between plasma levels and bone marrow toxicity for AMSA. The pharmacokinetics showed a biphasic elimination for both formulations. The mean terminal elimination half-life of AMSA-NCL and AMSA-lactate was 7.1 and 6.3 h, respectively, and the mean volume of distribution was 105 and 99 L/m2, respectively. No significant differences in the pharmacokinetics comparing days 1 and 3 were seen.(ABSTRACT TRUNCATED AT 250 WORDS)