Because 4-aminopyridine (AP) improves residual deficits in some multiple sclerosis (MS) patients but has a narrow toxic-to-therapeutic margin, we compared the safety and efficacy of two target peak serum concentration ranges (low: 30 to 59 ng/ml and high: 60 to 100 ng/ml). We enrolled eight MS patients with temperature-sensitive visual and motor deficits in a randomized, placebo-controlled, double-blind, crossover trial of short-term oral AP treatment. We randomized patients to a sequence of three treatments on three separate days: placebo, low serum concentration, and high serum concentration. We determined dosing to achieve the desired steady-state peak serum concentration ranges from a test dose and population pharmacokinetic parameters using bayesian estimation. Contrast sensitivity, standard neurologic examination, ratings of videotaped neurologic examinations, and quantitative strength assessment all improved with treatment, but flicker fusion frequency, visual evoked response latencies, and Expanded Disability Status Scale scores did not. All patients experienced side effects during the high-serum-concentration arm. A grand mal seizure occurred at a serum AP level of 104 ng/ml, and an acute confusional episode occurred at 114 ng/ml. AP treatment produced improvements in residual deficits in MS patients, but the occurrence of significant toxicity suggests that AP serum levels should be monitored and peak levels above 100 ng/ml should be avoided. Concentration-control methodology may be useful in testing putative treatments for other neurologic diseases.