The assay of saliva samples provides a valuable alternative to the use of blood samples for therapeutic drug monitoring (TDM), at least for certain categories of patients. To determine the feasibility of using saliva sampling for the TDM of rufinamide, we compared rufinamide concentrations in paired samples of saliva and plasma collected from 26 patients with epilepsy at steady state. Within-patient relationships between plasma rufinamide concentrations and dose, and the influence of comedication were also investigated. Assay results in the two tested fluids showed a good correlation (r2 = .78, P < .0001), but concentrations in saliva were moderately lower than those in plasma (mean saliva to plasma ratio = 0.7 ± 0.2). In eight patients evaluated at three different dose levels, plasma rufinamide concentrations increased linearly with increasing dose. Patients receiving valproic acid comedication had higher dose-normalized plasma rufinamide levels than patients comedicated with drugs devoid of strong enzyme-inducing or enzyme-inhibiting activity. Overall, these findings indicate that use of saliva represents a feasible option for the application of TDM in patients treated with rufinamide. Because rufinamide concentrations are lower in saliva than in plasma, a correction factor is needed if measurements made in saliva are used as a surrogate for plasma concentrations.