Currently, there are no drugs supported by sufficient evidence of efficacy for cerebral vasospasm in patients with subarachnoid hemorrhage, despite abundant evidence of anti-vasospasm drugs at an experimental level. We have developed a drug-delivery system using a vasodilating drug that can be implanted intracranially at the time of surgery for aneurysm clipping, without systemic side effects or side effects associated with long-term intrathecal drug administration through indwelling catheters. We started our project on 1994 for making slowly-releasing drug-delivery system in vitro, because cerebral vasospasm occurs 4-14 days following subarachnoid hemorrhage. A rod-shaped pellet (1 mm in diameter, 10 mm in length, and containing 1 mg of nicardipine) for animal study was prepared by heat compression. Release curve from the pellets was adjusted similar to the time course of cerebral vasospasm by changing the combination of molecular weight, lactic acid ratio of copoly (lactic/glycolic acid) and nicardipine. We presented the efficacy and safety of this drug delivery system using both canine double hemorrhage and clot placement model. The mean concentration of nicardipine in the clots was 1.5x10-4 mol/L on Day 7 and 5.1x10-6 mol/L on Day 14. This drug delivery system can prevent vasospasm significantly in dogs, while maintaining an appropriate concentration of nicardipine in the clot adjacent to the arteries, since maximal relaxation is achieved by 10-6 mol/L of nicardipine. Since October 1999, nicardipine pellets (NPs) (2 mm in diameter, 10 mm in length, and containing 4 mg of nicardipine) have been used to prevent vasospasm in patients with SAH. A frontotemporal craniotomy and a midline frontal craniotomy were performed for aneurysms in the internal carotid artery (ICA), middle cerebral artery (MCA), basilar artery, anterior communicating artery, and distal anterior cerebral artery (ACA). NPs were placed in the cistern of the ICA, the MCA, and/or the ACA, where thick clots existed, and, therefore, vasospasm related to delayed ischemic neurological deficits (DIND) was highly probable. The number of pellets and the location of the placement depended on the amount and site of the subarachnoid clot in the preoperative CT scans, the operative field, and the craniotomy. Cerebral vasospasm was assessed by DIND, cerebral angiography performed on Days 7 to 12 and CT scan. Vasospasm was completely prevented in the arteries in cisterns with thick clots, where vasospasm was highly expected, by placing NPs adjacent to the arteries during surgery. In the first 100 patients treated with NPs, the ratio of DIND, severe angiographical vasospasm and cerebral infarctions were 7%, 11%, and 5%, respectively. No complications were experienced. Thirty-two patients with severe SAH and undergoing aneurysm clipping were included into the single center, randomized, double-blind trial in Germany. The incidence of angiographic vasospasm in proximal vessel segments was significantly reduced after implantation of NPs (73% control versus 7% NPs). Implantation of NPs also improves clinical outcome of SAH patients. Implantation of NPs reduces the incidence of cerebral vasospasm and delayed ischemic deficits and improves clinical outcome after subarachnoid hemorrhage.