Stent graft migration after endovascular aneurysm repair: importance of proximal fixation.

OBJECTIVE We reviewed the incidence of stent-graft migration after endovascular aneurysm repair in a prospective multicenter trial and identified factors that may predispose to such migration. METHODS All patients who received treatment during the course of the multicenter AneuRx clinical trial were reviewed for evidence of stent-graft migration over 5 years, from 1996 to 2001. Post-deployment distance from the renal arteries to the proximal end of the stent graft and the proximal fixation length (length of the infrarenal neck covered by the stent graft) were determined in patients for whom pre-procedure and post-procedure computed tomography scans were measured in an independent core laboratory. RESULTS Stent-graft migration was reported in 94 of 1119 patients, with mean time after device implantation of 30 +/- 11 months. Freedom from migration was 98.6% at 1 year, 93.4% at 2 years, and 81.2% at 3 years (Kaplan-Meier method). Subset (n = 387) analysis revealed that initial device deployment was lower in 47 patients with migration, as evidenced by a greater renal artery to stent-graft distance (1.1 +/- 0.7 cm), compared with 340 patients without migration (0.8 +/- 0.6 cm; P =.006) on post-implantation computed tomography scan. Proximal fixation length was shorter in patients with migration (1.6 +/- 1.4 cm) compared with patients without migration (2.3 +/- 1.4 cm; P =.005). There was significant variation in migration rate among clinical sites (P <.001), ranging from 0% to 30% (median, 8%), with a greater than twofold difference in migration rate between the lowest quartile (6%) and the highest quartile (15%) clinical sites. Univariate and multivariate analysis revealed that renal artery to stent-graft distance (P =.001) and proximal fixation length (P =.005) were significant predictors of migration, and that each millimeter increase in distance below the renal arteries increased risk for subsequent migration by 5.8% and each millimeter increase in proximal fixation length decreased risk for migration by 2.5%. Pre-implantation aortic neck length, neck diameter, degree of device oversizing, correct versus incorrect oversizing, device type (stiff vs flexible), placement of proximal extender cuffs at the original procedure, and post-procedure endoleak were not significant predictors of migration. Migration was treated with placement of extender modules in 23 patients and surgical conversion in 7 patients; 64 patients (68%) with migration have required no treatment. CONCLUSIONS Stent-graft migration among patients treated in the AneuRx clinical trial appears to be largely related to low initial deployment of the device, below the renal arteries, and short proximal fixation length. Significant variation in migration rate among clinical sites highlights the importance of the technical aspects of stent-graft deployment. Advances in intraoperative imaging and deployment techniques that have been made since completion of the clinical trial facilitate precision of device placement below the renal arteries and should increase proximal fixation length. Whether this, together with increased iliac fixation length, will result in lower risk for migration remains to be determined in long-term follow-up studies.