The purpose of this study was to assess the accuracy of multi-parameter measurements with color-coded duplex sonography (CCDS) for the diagnosis of venous leakage in patients with erectile dysfunction. Sixty patients with repeated unsatisfactory reactions after intracavernous injection of vasoactive substances underwent CCDS. Following intracavernous injection of prostaglandin E1, peak systolic velocity (PSV), enddiastolic velocity (EDV), time averaged velocity (TAV), resistance index (RI), and pulsatility index (PI) were measured in the cavernous arteries over 30 min (one measurement per minute). The results were compared with independent measurements based on dynamic pharmaco-cavernosometry/cavernosography (DPCC). Dynamic pharmaco-cavernosometry/cavernosography revealed venous leakage in 33 patients. Of 48 patients with normal PSV ( > 25 cm/s), 25 had veno-occlusive dysfunction and the remainder presented normal venous function. No statistically significant differences between these groups were found in EDV, RI, and PI measurements. In contrast, differences in TAV were significant between patients with (mean 9.4 +/- 4.6 cm/s) and without venous leakage (mean 5.5 +/- 2.2 cm/s; p = 0.001). Analysis of relative frequencies revealed a broad overlap of EDV, TAV, RI, and PI measurements between both groups. Sensitivities and specificities determined from receiver-operating-characteristic curves were > 80 % and > 50% for a TAV threshold of 5 cm/s, and an RI threshold of 1.0. Measurements of EDV, TAV, RI, and PI in patients with repeated unsatisfactory reactions on intracavernous prostaglandin injection are poor predictors of venous leakage and should not replace DPCC in the investigation of vasculogenic impotence.