To assess the reproducibility of penile arterial colour duplex ultrasonography (CDU) in patients with impotence. Fifty-eight patients underwent CDU of their penile arteries on two occasions 2 weeks apart. Group A (n = 34) received 60 mg of papaverine intracavernosally before CDU and group B (n = 24) received a combination of papaverine, prostaglandin E1 and phentolamine. The variables measured during CDU were peak systolic velocity (PSV) and the end diastolic velocity (EDV), with the systolic rise-time (SRT) also recorded in group B. In both groups of patients the reproducibility between the first and second recordings was poor for all variables. The mean variation in PSV was 31.5% and 40.2%, and in EDV was 44.9% and 50.1%, for groups A and B, respectively. For the SRT, measured only in group B, the mean variation was 25.7%. These results show that CDU is a poorly reproducible technique for assessing vascular status in patients with impotence. Therefore, it is not reliable in the long-term follow-up of patients and single measurements should be interpreted with caution.