Frequency, predictive factors, and consequences of stent-graft kink following endovascular AAA repair

To determine from the EUROSTAR registry the incidence of stent-graft kink, predictive factors for kinking, and the consequences of this complication on graft patency and aneurysm exclusion.

From January 1994 to June 2002, 4613 patients who underwent endovascular aneurysm repair were registered in the EUROSTAR registry. Presence of kink was determined according to the information available on the follow-up Case Record Form. The population was divided into those with and without stent-graft kink. Patient characteristics, morphological aneurysm features, team experience, type of device, period of implantation, and outcome were compared between the groups by univariate analysis. Significant factors were subsequently submitted to a multivariate Cox proportional hazard analysis.

One hundred seventy (3.7%) patients were reported as having a kink of the stent-graft during a mean follow-up of 21 months (range 1-72). Gender, neck angulation, team experience, period of implantation, ASA classification, and device type were independent predictors of kink. The presence of a kink was significantly associated with type I endoleaks (proximal and distal), type III endoleaks (midgraft), graft stenosis, graft limb thrombosis, graft migration, and conversion to open repair. No relationship was found between a decrease in the aneurysm diameter and the occurrence of stent-graft kink.

Kinks of stent-grafts were infrequent events in the intermediate term. Patients most at risk were women with angulated aortic necks treated by a minimally experienced team. Kinks are potentially damaging events because they may lead to delayed device-related endoleaks, graft stenosis, thrombosis, and conversion to open repair. They appear more closely associated to graft migration than to aneurysm diameter reduction.