This prospective randomized trial was performed to compare the effectiveness of intra-arterial ACNU with intravenous ACNU in newly diagnosed patients with supratentorial glioblastoma. The primary end points were overall survival and progression-free survival. Within 3 weeks after surgery, patients were randomly assigned to receive either intravenous or intra-arterial ACNU (80 mg/m2) once every 6 weeks concomitant with radiotherapy. Intraarterial ACNU was administered for the first 3 courses followed by intravenous administration. Eighty-four patients were enrolled onto this study and among them 82 patients who passed eligibility criteria were analyzed. Patients characteristics were not different significantly between 2 treatment arms. Median survival and progression-free survival time was 59 and 24 weeks, respectively for intra-arterial arm and 56 and 45 weeks, respectively for intravenous arm. There was no significant difference respectively between two treatment arms. Among the prognostic variables including age, Karnofsky performance status, extent of surgery and treatment arm, Cox's proportional hazards model showed that age was the only significant factor for both survival and progression-free survival (P = 0.003 and 0.016, respectively). With regard to toxicity, there was no significant difference between two treatment arms. Leukoencephalopathy was not observed in intra-arterial arm. In conclusion, intra-arterial ACNU when administered by the method in this study does not increase the survival and progression-free survival of newly diagnosed patients with glioblastoma over that afforded by intravenous ACNU.