Amrinone-associated thrombocytopenia is thought to result from nonimmune-mediated peripheral platelet destruction. Platelet destruction may be a concentration-dependent toxic effect of amrinone or its principal metabolite N-acetylamrinone. Eighteen children receiving amrinone after heart surgery were prospectively evaluated to correlate the pharmacokinetics of amrinone and N-acetylamrinone with thrombocytopenia. Amrinone and N-acetylamrinone plasma concentrations were determined by HPLC during loading, infusion, and terminal elimination, with concurrent monitoring of platelet counts. Thrombocytopenia developed in eight patients (platelet count, 66 +/- 17 x 10(9) platelets/L [mean +/- SD]). Peak and steady-state amrinone plasma concentration, amrinone total dose, duration of amrinone exposure, and amrinone area under curve (AUC) were similar between patients with and without thrombocytopenia. N-Acetylamrinone peak concentration, steady-state concentration, N-acetylamrinone AUC, and ratio of N-acetylamrinone to amrinone were greater in patients with thrombocytopenia. This association suggests that N-acetylamrinone, and not amrinone, may be the mediator of thrombocytopenia in children receiving amrinone.