Endovascular versus open repair of abdominal aortic aneurysms: A population-based evaluation of outcomes and resource utilization in Ontario

Abstract

Objective. Two large randomized trials that compared elective EndoVascular Aneurysm Repair (EVAR) with open repair for non-ruptured abdominal aortic aneurysms (AAA) have demonstrated similar long-term mortality rates but increased costs associated with EVAR. Despite these data, the use of EVAR continues to increase in North America. There are currently very limited population-based adjusted data looking at long-term outcomes and resource utilization. Methods. All patients who underwent elective AAA repair between April 2002 and March 2007 in Ontario were identified using data from hospital discharge abstracts. ICD-10-CA and Canadian Classification of health Interventions (CCI) codes were used in a validated algorithm to identify patients who underwent either EVAR or open repair of non-ruptured AAAs. Pre-operative co-morbidities were measured using the Charlson co-morbidity index. Risk stratification into quintiles was performed using propensity score analysis. Results. Overall, 6461 patients underwent treatment of non-ruptured AAAs (N: EVAR 888; open 5573). Patients undergoing EVAR were older and had more comorbidities. The adjusted 30-day mortality was significantly lower in the EVAR group (adjusted OR= 0.34 [0.20-0.59]). The adjusted all-cause long-term mortality was similar in both groups (OR= 0.95 [0.81-1.05]). After adjustment for significant confounders, rates of imaging studies and both urgent and vascular readmissions were statistically higher in the EVAR group. However, the EVAR group had significantly shorter length of stay for the index hospitalization, all subsequent hospitalizations, and the intensive care unit. Discharge to a nursing home or other chronic care facility after the index procedure was also lower in the EVAR group (OR= 0.55 [0.41-0.0.74]). The durability of the repair of EVAR vs. open techniques as indicated by the rate of repeat interventions following the index procedure for EVAR (OR= 1.3 [0.98-1.75]) did not reach statistical significance. Conclusion. After adjusting for pre-operative risk factors, there was no difference in long-term mortality between EVAR and open repair in Ontario. The significantly lower 30-day mortality rate in EVAR patients was not sustained over longer-term follow-up. Although the utilization of imaging studies and hospitalizations was significantly higher in the EVAR group, patients undergoing open repair spent more days in hospital (including readmissions), more time in ICU, and were more likely discharged to a chronic care facility.