Percutaneous occlusion of patent ductus arteriosus with the double umbrella Rashkind prosthesis is an alternative to surgery. The authors report the results of a series of 40 patients aged 16 to 203 months, who underwent transcatheter occlusion of patent ductus arteriosus between January 1989 and March 1993. These children were followed up regularly, clinically and by echocardiography to determine the presence or absence of a residual shunt. Immediately after implantation, angiography showed a residual shunt in 75% of cases. However, the shunt progressively regressed during follow-up. Three types of shunt were described using colour Doppler based on the diameter of the transductal jet at its origin and its extension in the pulmonary artery: type I, a wide jet (> 3 mm), type II, or minimal shunt, with a jet diameter of 2 to 3 mm and an extension to the pulmonary valve (flame-like), and finally the type III, or punctiform shunt with a jet diameter < 2 mm and extension in the main pulmonary artery < 1.5 cm. During follow-up, a transition between the 3 types was observed in 37% of cases. Kaplan-Meier actuarial analysis showed a residual shunt rate of 33% at one year and 19% at 2 years after implantation. The search for predictive factors of residual shunt showed that very young patients at the time of catheterisation was the only indicator of a residual shunt (p = 0.03). Occlusion of a patent ductus arteriosus with a Rashkind device is a reliable method which gives good results. The essential problem remains that of a residual shunt but this disappears in 3/4 of patients during follow-up. When the shunt persists, a 1 to 2 year interval is advised before considering implantation of a second occluder.