The effects of the NMDA antagonist dextromethorphan (DM) on levodopa-associated dyskinesias and motor fluctuations were studied in patients with advanced Parkinson's disease. During initial open-label dose escalation, 6 of 18 patients reported a beneficial effect at their individually determined optimal DM dose (range, 60-120 mg/day). The 12 remaining patients either experienced reversible side effects, particularly mild drowsiness, or decreased levodopa efficacy, and were therefore excluded from the study. The six responders entered the double-blind, placebo-controlled, crossover study with two 2-week arms separated by 1 week wash-out. On the last day of each arm, motor ratings were performed every 20 minutes for 8 consecutive hours. In addition, motor complications and Activities of Daily Living (ADL) were assessed using the Unified Parkinson's Disease Rating Scale (UPDRS) and patient diaries. With DM, dyskinesias improved by 25% according to physician's ratings and by 40% according to UPDRS interviews, without compromising the anti-Parkinson effect of levodopa. Motor fluctuations and ADL scores also improved significantly. Although the narrow therapeutic index of DM limits its clinical usefulness, these findings support the view that drugs acting to inhibit glutamatergic transmission at the NMDA receptor can ameliorate levodopa-associated motor complications.