&NA; We determined the pharmacokinetics and duration of action of a bolus dose of doxacurium (15 μg/kg) in 27 patients anesthetized with isoflurane and nitrous oxide. Nine patients had normal renal and liver functions and were undergoing a variety of surgical procedures, nine were undergoing cadaveric kidney transplantation because of end‐stage renal disease, and nine were undergoing cadaveric liver transplantation because of end‐stage hepatocellular disease. Plasma concentrations of doxacurium were measured for 6 h after administration using a sensitive and specific capillary gas chromatographic assay. Plasma concentration versus time data were analyzed by a noncompartmental method based on statistical moments. Neuromuscular blockade was assessed by measuring the electromyographic evoked response of the adductor pollicis muscle to train‐of‐four stimulation of the ulnar nerve. The degree of neuromuscular blockade after doxacurium administration was described as the percent of control of the first train‐of‐four response. The pharmacokinetic variables were (normal vs hepatic failure vs renal failure, respectively): volume of distribution at steady state (220 ± 110 vs 290 ± 60 vs 270 ± 130 mL/kg [mean ± SD]), plasma clearance (2.7 ± 1.6 vs 2.3 ± 0.4 vs 1.2 ± 0.7 mL·kg−1·min−1), mean residence time (95.2 ± 57 vs 129.4 ± 30 vs 270 ± 210 min), and elimination half‐life (99 ± 54 vs 115 ± 31 vs 221 ± 156 min). Plasma clearance and mean residence time differed significantly between patients with renal failure and control patients. There was no significant difference in the onset times or in clinical effective duration, although the clinical duration tended to be longer and more variable in the patients with renal failure. This unpredictable response and the possibility of prolonged blockade should be borne in mind if doxacurium is to be used in patients with renal failure.