The Eurostar Registry was established in 1996 to collate information, from centres across Europe, on the outcome from endovascular grafting of aortic aneurysms. At the end of the first year of the project, data on 430 patients had been entered onto the database. In 420 patients (97.7%), the endografts were deployed without major complications. The 30-day mortality rate was low at 3.4% and deaths were confined mostly to 'high risk' patients with major co-morbidity. Endoleaks, which were present on discharge from hospital in 15.7% of patients, were associated with a significantly increased incidence of continued expansion of the aneurysm sac postoperatively (P < 0.01). Thus the early results confirmed the feasibility and low complication rate of endovascular repair of aneurysms, but a higher than expected incidence of endoleaks. At 2 years, 895 patients had been registered. The rate of early endoleaks remained significantly unchanged but another 18% of patients had developed new endoleaks during the first year of follow-up. Six delayed ruptures had been reported, 3 fatal. There were indications that 'self sealed' endoleaks continued to pressurise the aneurysm sac. Severe distortion of the grafts with kinking and other structural changes associated with postoperative longitudinal shrinking of the aneurysm sac was observed in 69% of patients at 1 year. Clinical complications associated with these changes included late endoleak and graft limb occlusion. Early unrealistic optimism about endovascular aneurysm repair has been replaced with a more realistic understanding of its benefits and limitations as a result of the Eurostar project and other registries. Randomised studies are now required to establish the most appropriate role for this approach, alongside established therapies.