Objective To evaluate our policy of managing priapism for the success rate of the treatments, potency afterward, complications, and the risk factors responsible for erectile dysfunction in these patients. Patients and methods The study included 50 patients (mean age 37.1 years, range 22±66) with a diagnosis of priapism (1981±1999). Their records were reviewed; 35 patients were available for a long-term evaluation. Factors assessed were the duration of priapism, history of previous recurrent attacks, possible underlying causes (e.g. haematological disorders, medications or trauma), relation to sexual stimulation, pain, and any attempt at previous management. A complete blood screen and blood gases were assessed in corporal aspirates. Duplex ultrasonography was used in all impotent patients at their follow-up. Early and late complications were reviewed, and patients asked about their erectile function before priapism, and any recurrence. Results The median (range) duration of priapism was 48 (6±240) h; almost half the patients presented >48 h after the onset of priapism. Sixteen patients (32%) reported a history of previous recurrent attacks, of whom seven had a history of previous treatments. The main cause of priapism was idiopathic or intracavernosal injection with papaverine. All patients were initially treated by corporal blood aspiration and injection with ephedrine; if this failed or if the priapism was prolonged (>48 h) various shunts were used. The hospital stay was signi®cantly shorter among patients with papaverine-induced or brief priapism. In the long-term follow-up of 35 patients (mean 66.4 months, range 3±220) only 15 (43%) reported preserved erectile function, and this was more likely in patients with brief priapism (<48 h). Eight patients (23%) reported subsequent recurrent attacks of priapism; all were managed successfully as they presented shortly after their onset. Penile ®brosis was detected in 20 patients (57%), and was signi®cantly more common in those with prolonged priapism (>48 h) or from causes other than papaverine. The 20 impotent men evaluated by Doppler ultrasonography had severe echo-dense penile ®brosis and high end-diastolic velocities suggesting veno-occlusive incompetence in all except two. In ®ve men with shunts cavernosography showed extensive venous leakage irrespective of site of the shunt. MRI in ®ve patients with penile ®brosis showed heterogeneous areas of low signal intensity, corresponding with haemosiderin deposition and ®brosis. On univariate analysis the ®nal result of management (complete detumescence or not), the duration of priapism and the presence of penile ®brosis signi®cantly in¯uenced erectile function. On multivariate logistic regression only the ®rst remained signi®cant. Conclusions Low-¯ow priapism for >48 h, failure to maintain complete detumescence after management, and marked penile ®brosis during the follow-up are the most signi®cant risk factors responsible for erectile dysfunction, with failure to achieve complete detumescence the most detrimental.