The pharmacokinetic properties of the lactate-lowering drug dichloroacetate were investigated in 111 adult patients with lactic acidosis who were randomized to receive dichloroacetate as part of a placebo-controlled clinical trial. The clinical symptoms and etiology of lactic acidosis varied markedly among patients. Dichloroacetate, at a dose of 50 mg per kilogram of body weight, was administered in a 30-minute intravenous infusion into a peripheral vein. A second dose, identical to the first, was administered 2 hours after beginning the first infusion. Plasma levels of dichloroacetate were determined from blood samples collected periodically up to 288 hours after administration and the data were subjected to pharmacokinetic modeling. The pharmacokinetic properties of dichloroacetate in these acutely ill patients were complex and differed markedly from those in healthy volunteers, whose data fitted a one-compartment pharmacokinetic model. In contrast, the data from patients fitted one-, two-, or three-compartment pharmacokinetic models or even none of these, depending on the individual. Drug clearance in plasma tended to decrease as the number of compartments required to fit the data increased or as the number of drug treatments increased.