This phase I safety study aimed to identify the optimal dose of the vascular disrupting agent 5,6-dimethylxanthenone-4-acetic acid (DMXAA) for combination studies. Using a crossover design, 15 patients with refractory tumors were allocated randomly to receive six sequential doses of DMXAA (300, 600, 1,200, 1,800, 2,400, and 3,000 mg m(-2)), each given once-weekly as a 20-minute i.v. infusion. The drug was generally well tolerated. Transient, moderate increases in the heart rate-corrected cardiac QT interval occurred at the two highest doses. DMXAA produced transient dose-dependent increases in blood pressure. Transient, dose-related visual disturbances occurred at the two highest doses. No significant changes in K(trans) and k(ep) were observed but V(e), a secondary dynamic contrast-enhanced magnetic resonance imaging variable, increased significantly after giving DMXAA. At 1,200 mg m(-2), the Cmax and the area under the concentration-time curve over 24 hours for total and free DMXAA plasma concentrations were 315 +/- 25.8 microg/mL, 29 +/- 6.4 microg/mL x d, 8.0 +/- 1.77 microg/mL, and 0.43 +/- 0.07 microg/mL x d, respectively. Plasma levels of the vascular damage biomarker 5-hydroxyindoleacetic acid increased in the 4 hours after treatment in a dose-dependent fashion up to 1,200 mg m(-2), with a plateau thereafter. Doses in the range of 1,200 mg m(-2) have been selected for further studies (phase II combination studies with taxanes and platins are under way) because this dose produced no significant effect on heart rate-corrected cardiac QT interval, produced near maximum levels of 5-hydroxyindoleacetic acid, achieved DMXAA plasma concentrations within the preclinical therapeutic range, and was well tolerated.