Objectives: This study was undertaken to determine the effect of the preoperative diameter of abdominal aortic aneurysms on the midterm outcome after endovascular abdominal aneurysm repair (EVAR).
Method: The data for 4392 patients who had undergone EVAR were analyzed. Patients were enrolled over 6 years to June 2002 in the EUROSTAR database. Outcomes were compared between three groups defined by the preoperative diameter of the aneurysm: group A (n = 1962), 4.0 to 5.4 cm; group B (n = 1528), 5.5 to 6.4 cm; and group C (n = 902), 6.5 cm or larger. Patient characteristics, details of aortoiliac anatomy, operative procedures, old or current device generation, and postoperative complications in the three patient groups were compared. Outcome events included aneurysm-related death, unrelated death, conversion, and post-EVAR rupture of the aneurysm. Life table analysis and log-rank tests were used to compare outcome in the three study groups. Multivariate Cox models were used to determine whether baseline and follow-up variables were independently associated with adverse outcome events.
Results: Patients in group C were significantly older than patients in groups A and B (73 years vs 70 and 72 years, respectively; P =.003 - P <.0001 for different group comparisons), and more frequently were at higher operative risk (American Society of Anesthesiologists classification >or=3; 63% vs 48% and 54%; P =.0002-P <.0001). Device-related (type I) endoleaks were more frequently observed at early postoperative arteriography in group C compared with groups A and B (9.9% vs 3.7% and 6.8%; P =.01-P <.0001). Postoperatively systemic complications were more frequently present in group C (17.4% vs 12.0% in group A and 12.6% in group B; P <.0001 and.001). The first-month mortality was approximately twice as high in group C compared with the other groups combined (4.1% vs 2.1%; P <.0001). Late rupture was most frequent in group C. Follow-up results at midterm were less favorable in groups C and B compared with group A (freedom from rupture, 90%, 98%, and 98% at 4 years in groups C, B, and A, respectively; P <.0001 for group C vs groups A and B). Aneurysm-related death was highest in group C (88% freedom at 4 years, compared with 95% in group B and 97% in A; P =.001 and P <.0001, respectively; group B vs A, P =.004). The annual rate of aneurysm-related death in group C was 1% in the first 3 years, but accelerated to 8.0% in the fourth year. Incidence of unrelated death also was higher in groups C and B than in group A (76% and 82% freedom at 4 years vs 87%; P <.0001 for both comparisons). Ratio of aneurysm-related to unrelated death was 23%, 21%, and 50% in groups A, B, and C, respectively. Cox models demonstrated that the correlation between large aneurysms (group C) and all assessed outcome events was independent and highly significant. Older generation devices had an independent association with aneurysm-related and unrelated deaths (P =.02 and P =.04, respectively). However, this correlation was less strong than large aneurysm diameter (P =.0001 and P =.0009, respectively).
Conclusions: The midterm outcome of large aneurysms after EVAR was associated with increased rates of aneurysm-related death, unrelated death, and rupture. Reports of EVAR should stratify their outcomes according to the diameter of the aneurysm. Large aneurysms need a more rigorous post-EVAR surveillance schedule than do smaller aneurysms. In small aneurysms EVAR was associated with excellent outcome. This finding may justify reappraisal of currently accepted management strategies.