OBJECTIVE Diethylenetriamine/nitric oxide (DETA/NO) has been shown to be an effective treatment for delayed posthemorrhagic vasospasm when released abluminally from ethylene-vinyl acetate copolymer (EVAc). However, the observed mortality associated with this drug warrants further investigation. To establish a maximum tolerable dose, this study evaluated the toxicity of DETA/NO released from EVAc in a dose-escalation series in cynomolgus monkeys (Macaca fascicularis). METHODS DETA/NO was incorporated into EVAc at a 20:80 dry weight ratio (DETA/NO:EVAc). A total of 13 animals underwent a right frontotemporal craniotomy for placement of a single polymer delivering no drug (n = 3), 0.5 +/- 0.1 mg/kg (n = 3), 0.9 +/- 0.1 mg/kg (n = 3), 1.9 +/- 0.2 mg/kg (n = 3), or a 3.2 mg/kg dose (n = 1) into the subarachnoid space. RESULTS The animal receiving the highest dose of DETA/NO (3.2 mg/kg) died 46 hours after surgery. The remaining animals survived for the planned duration of the study. One animal in the group receiving the 1.9 mg/kg dose experienced a seizure 25 hours after surgery and remained lethargic for 2 days before making a complete recovery. The remaining animals exhibited no adverse behavioral effects. Histopathological examination of brain tissue revealed hemorrhagic and ischemic changes at doses above 0.9 mg/kg. No evidence of vascular wall pathology or infection was observed in any animal. CONCLUSIONS The greatest amount of DETA/NO safely delivered from EVAc copolymer to the subarachnoid space of the cynomolgus monkey is approximately 1.0 mg/kg. These findings show that continuous intracisternal delivery of DETA/NO is a safe and potentially effective strategy for prophylactic treatment of delayed cerebral vasospasm.