BACKGROUND:Intrathecal N-methyl-d-aspartate antagonists have antihyperalgesic efficacy. The authors examined toxicity in a canine model of chronic lumbar intrathecal infusion. METHODS:Dogs (10-16 kg) were prepared with lumbar intrathecal catheters connected to vest-mounted pumps (100 microl/h). In phase 1, stepwise incrementations in infusion concentration were performed at 48- to 72-h intervals to determine an infusion dose with minimal but detectable behavioral effects. In phase 2, the dose/concentration defined in phase 1 was infused for 28 days. Behavioral function during infusion and histopathology at sacrifice was assessed. Drugs examined were 2-amino-5-phosphono-valorate (AP5), MK801, memantine, amitriptyline, S-methadone, and saline. RESULTS:In the phase 1 dose ranging, the minimum effect doses for the several agents were as follows: AP5, 1 mg/day; amitriptyline, 1 mg/day; ketamine, 10 mg/day; MK801, 1 mg/day; and memantine, 4 mg/day. In phase 2, infusion of these doses typically resulted in mild hind limb weakness by 3-5 days after initiation of infusion, which progressed over the 28-day infusion interval. In a limited number of animals, a similar effect was observed with S-methadone. Histopathologically, vehicle-infused animals displayed a minor local catheter reaction. With the drug treatments, a gradient of increasing pathology from cervical to lumbar segments was noted. Pathology ranged from local demyelination to necrotizing lesions of spinal parenchyma near the catheter tip. All drugs given at their respective doses produced pathology scores significantly worse than saline controls. CONCLUSIONS:These drugs given for 28 days at acutely tolerable doses lead to spinal pathology. These data suggest a reevaluation of the use of these agents in chronic spinal delivery.