Although mivacurium is eliminated by plasma cholinesterase, previous investigations have revealed either no relationship or limited correlation between mivacurium infusion rates (IRs) and plasma cholinesterase activity. Assuming that such a relationship should exist, we used a novel approach to better demonstrate the relationship in humans. Fourteen isoflurane-anesthetized adults underwent standard neuromuscular monitoring. Mivacurium was then infused at 1.0 mu g . kg(-1). min(-1) until twitch tension stabilized. The LR was then adjusted, using the Hill equation, to produce approximately steady state 50% (n = 14) or 90% (n = 13) twitch depression. Using these values for IR and steady-state twitch depression, the IRs expected to produce 50% and 90% twitch depression (IR(50) and IR(90), respectively) were estimated by nonlinear regression. Both IR(50) (r(2) = 0.51, P < 0.005) and IR(90) (r(2) = 0.48, P < 0.01) were related to plasma cholinesterase activity; the coefficient of the Hill equation did not vary with plasma cholinesterase activity. We conclude that mivacurium IRs are, as expected, influenced by the activity of the enzyme responsible for its elimination.