Endovascular surgery for abdominal aortic aneurysms (AAA) has been performed with different commercially available devices since more than 7 years. The technique of implantation even for ruptured aneurysm is feasible. Attention is now focused on complications and mid-term results. We present a single center experience over the last 7 years using nine different endografts. 653 patients with AAAs were treated endovascularly (158 tubes, 495 bifurcated) between August 1994 and September 2001. The total mortality rate in the follow-up period (median 37 months) was 6.1 % with a 30-day mortality of 0.76 % and a procedure-related lethality of 1.2 %. Tube endografts showed the highest endoleakage rate (37.3 %) due to secondary shortening, migration and material defects. Endoleakage was diagnosed in 18.2 % of unibody designed stentgrafts versus 28.3 % in modular grafts. Regarding proximal Type I endoleaks suprarenal fixation with uncovered stents seems to be of no further advantage in proximal graft fixation. The conversion rate was 6.6 % (n = 43), aneurysm rupture following endovascular repair occurred in 3 of these patients (0.45 %). Mortality in emergency conversion was 30 % compared to no deaths in elective cases. There was a surprisingly high rate of inadvertent renal infarctions (9 %) in follow-up CT scans, two times higher in suprarenal fixation (18 %) than in infrarenal position (7.4 %). In conclusion, we observed best results in unibody bifurcated stentgrafts, late failure of first generation endografts (Mintec(R)) and decreasing conversion rates over the last years (learning curve!). The clinical significance of a side branch endoleak and related maximal AAA diameter remains uncertain. The ideal graft has yet to be found, but improvements in design and implantation techniques evolve rapidly. Good results will be obtained in the future by patient selection and restrictive indications.