OBJECTIVE This study aimed to evaluate the symptomatic response and side effects of ventriculolumbar perfusion (VLP) methotrexate chemotherapy with a low perfusion rate in patients with leptomeningeal metastasis. METHODS Patients in a single-arm, two-stage phase II trial based on Simon's minimax design received VLP with a reduced (15 cc/h) perfusion rate with the purpose of decreasing constitutional side effects such as nausea/vomiting, insomnia, and confusion. The primary outcome was control of increased intracranial pressure (ICP). The secondary outcome was an occurrence of side effects. The results were compared with those of a previous trial of VLP with a 20-cc/h perfusion rate. RESULTS Total 90 patients were enrolled. Out of 65 patients with increased ICP, 32 achieved normalized ICP after VLP chemotherapy (bias-adjusted response rate = 51%). The incidence of moderate-to-severe nausea/vomiting was reduced to 46% from 64% in the previous study, and that of sleep disturbance was increased to 13% from 9%, but both failed to reach statistical significance. The incidence of moderate-to-severe confusion was significantly reduced to 12% from 23% in the previous study (p = 0.04). Median overall survival was better among patients with controlled ICP than among those who remained with increased ICP (193 days vs. 94 days, p = 0.013). CONCLUSIONS Compared with a higher perfusion rate, the low perfusion rate failed to provide non-inferior ICP control or improved side effects, except for confusion. The relationship between VLP perfusion rate and ICP control needs to be evaluated in future trials adjusting for bias from uncompleted protocol due to poor general condition.