RMP-7, a bradykinin analog, temporarily increases the permeability of the blood-brain tumor barrier to chemotherapy drugs like carboplatin. We conducted a randomized, controlled trial of carboplatin and RMP-7 versus carboplatin and placebo in patients with recurrent malignant glioma. The primary outcome measure was time to tumor progression (TTP). Adults with recurrent glioblastoma multiforme or anaplastic glioma were randomized in a 1:1 ratio to receive carboplatin and either RMP-7 or placebo. Radiation therapy had failed in all patients, and they may have received prior chemotherapy. Carboplatin (dosed to achieve an area under the curve of 5 mg/ml x time for patients who had received prior chemotherapy, or 7 mg/ml x time for those who had not) was given intravenously every 4 weeks, followed by intravenous infusion of either RMP-7 or placebo (300 ng/kg). TTP, tumor response, neuropsychological assessments, functional independence, and quality of life assessments were analyzed every 4 weeks. There were 122 patients enrolled, 62 in the RMP-7 and carboplatin group and 60 in the placebo and carboplatin group. Median TTP was 9.7 weeks (95% CI, 8.3-12.6 weeks) for the RMP-7 and carboplatin group and 8.0 weeks (95% CI, 7.4-12.6 weeks) for the placebo and carboplatin group. Median survival times were 26.9 weeks (95% CI, 21.3-37.6 weeks) for the RMP-7 group and 19.9 weeks (95% CI, 15.0-31.3 weeks) for the placebo group. No differences were noted for time to worsening of neuropsychological assessments, functional independence, or quality of life assessments. The use of RMP-7 had no effect on the pharmacokinetics or toxicity of carboplatin. At the dose and schedule used in this trial, RMP-7 did not improve the efficacy of carboplatin. Recent preclinical pharmacokinetic modeling of RMP-7 suggests that higher doses of RMP-7 may be required to increase carboplatin delivery to tumor.