Frequently, instructions for use (IFU) parameters published by device manufacturers are not adhered to in order to increase patient candidacy for endovascular aneurysm repair (EVAR). In a multicentre study of patients undergoing elective EVAR for abdominal aortic aneurysm, we examined whether IFU non-adherence was related to a worse mid-term outcomes. The primary outcome in our study was a composite endpoint of graft-related adverse events and included aneurysm-related death, sac growth, endoleak (except type II), endoleak intervention, limb occlusion, device migration or rupture.
Our study population included 461 patients with an IFU non-adherence rate of 43.8%. Patients with IFU non-adherence were less likely to be men (90% vs 78.7%) and more likely to have peripheral vascular disease (7.3% vs 12.4%). Iliac and neck diameter non-adherence (21.4% and 15.2%, respectively) occurred the most frequently. Graft-related adverse events occurred at a rate of 12.8% during the study period with endoleak and sac growth occurring the most frequently (8.5% and 8.1%, respectively). On survival analysis, freedom from graft-related adverse events was significantly higher in patients with IFU adherence compared to those with non-adherence (p=0.031). When considering neck IFU non-adherence in isolation, freedom from graft-related adverse events remained significantly higher in IFU adherent patients compared to non-adherent (p=0.003). On adjusted Cox regression, IFU non-adherence, age ≥80, peripheral vascular disease, and use of the Talent graft (Medtronic) were significant predictors of graft-related adverse events in both any IFU non-adherence and neck-only IFU non-adherence models.
IFU non-adherence in the literature ranges from 38–53%, which is similar to our own results.2–6 The relationship that exists between IFU non-adherence and negative outcome is not clear. Many studies have found that IFU non-adherence appears to be associated with lower mid-term survival and that freedom from graft-related adverse events was significantly lower in the IFU non-adherent group.2,3 Neck anatomy that falls outside of IFU parameters have also been found to be associated with a higher rate of type I endoleak but showed no significant difference in survival or graft patency.4 Several studies have found no difference in mortality, aneurysm-related mortality, sac enlargement, need for reintervention, or endoleaks in grafts placed within and outside of the IFU.5,6 However, there are several limitations in the published literature. Many of these studies employed solitary outcomes with low event rates as their primary outcome instead of using a composite endpoint as we did in this study. Consequently, these studies did not attain power and any signal of association between IFU non-adherence and outcomes may have been lost. As well, some studies used generic IFU parameters rather than device-specific parameters, which may not accurately predict graft performance. Finally, many of the survival analyses have rapid loss of follow-up (patients at risk), which decreases the robustness of the analysis.
Since its inception nearly 25 years ago by Juan Parodi, commercially available endovascular grafts have undergone multiple design modifications to overcome constraints imposed by complex anatomy. Lower profile delivery systems are now available as small as 14F providing EVAR access vessel suitability for 90% of men and 70% of women.7,8 Uncovered stents, suprarenal fixation, radial force from self-expanding stents, and increased columnar strength have been some additions commonly found in modern grafts to maximise neck apposition and reduce migration. More recently, sealing rings filled with radio-opaque polymer have allowed for reliable sealing of short neck (≥7mm) aneurysms, however, the long-term results of these devices are still unknown. An FDA-approved device for short necks <10mm has had some success however type I endoleak rates still remain higher than conventional neck lengths and long-term results are not yet known.
Chimney and snorkel adjunct offers an alternative to complex neck anatomy that falls outside the IFU for standard available endografts but are associated with high rates of gutter leaks9 that are exceedingly complex to treat. Caution should be used before considering these procedures since they require advanced endovascular skills to safely perform and both the endografts and the covered and uncovered stents employed to create the snorkel or chimneys are being deployed outside their respective IFUs. There is some favourable published short-term outcome data from experienced centres related to these procedures, although the durability of chimneys and snorkels is currently not known.
More complex solutions such as custom branched EVAR and fenestrated EVAR remain promising solutions to complex neck aneurysms. As they become more accessible to US centres, many are touting these custom grafts as the solution to complex anatomy. However, the four- to six-week custom order waiting period, the requirement of advanced endovascular skills, excessive coverage of the normal aorta, considerable procedural times and potential inadvertent branch vessel occlusion associated with custom grafts creates impracticality and makes them less appropriate for small-volume centres.
Custom grafts have been available in Canada for approximately 15 years and are commonly used in most academic centres. Through the time period of our study, custom grafts were available as an alternative for complex anatomy cases at the study centres, one of which also offers a complex aortic endovascular programme and fellowship. And yet, a relatively high rate of IFU non-adherence was still found in this series. It is likely that because of the impracticality of applying custom techniques to a high rate of abdominal aortic aneurysms that are outside of IFU parameters and for the other above-listed issues, that surgeons continue to elect to push the limits of device capabilities for standard endografts. It is our belief that this will also hold true in the US market and fail to solve the complex anatomy issue. Finally, let us not forget how abdominal aortic aneurysm repair began; open repair remains a viable and reliable solution for many of these cases.
Our study would indicate that offering EVAR repair in patients whose anatomy is not compliant with IFU parameters must be done with caution. Widespread aggressive application of EVAR in non-IFU-adherent patients must be limited and other endovascular and open options must be considered. In some co-morbid patients this may still remain the lowest-risk option for abdominal aortic aneurysm repair. Other patients may be willing to accept the higher risk of graft-related adverse events in order to avoid a laparotomy or complex open repair. In these instances, a patient must be properly advised and consented for these risks. At the very least, if patients are treated using standard EVAR devices outside IFU, these patients warrant very rigorous follow-up imaging. Further device development and creative solutions in EVAR technology may someday yield a universal device for complex anatomy, but for now we must use all the options available to provide the best care for patients.
- JC Parodi, JC Palmaz, HD Barone. Ann Vasc Surg 1991;5:491–9.
- TA Abbruzzese, CJ Kwolek , DC Brewster, et al. J Vasc Surg 2008;48(1):19–28.
- KP Donas, G Torsello, K Weiss, et al. J Vasc Surg 2015;62(4):848–54.
- AF AbuRahma, M Yacoub, AY Mousa, et al. J Am Coll Surg 2016;222(4):579–89.
- J Walker, LY Tucker, P Goodney, et al. J Vasc Surg 2015;61(5):1151–9.
- WE Beckerman, RO Tadros, PL Faries, et al. J Vasc Surg 2016;64(1):63–74.
- MP Sweet, MF Fillinger, TM Morrison, D Abel. J Vasc Surg 2011;54:931–7.
- M Mehta, FE Valdes, T Nolte, et al. J Vasc Surg 2012;55:906–13.
- A Wilson, S Zhou, P Bachoo, AL Tambyraja. Br J Surg 2013;100:1557e64.
Christine R Herman is at Dalhousie University, Halifax, Canada, and Oren K Steinmetz is at McGill University, Montreal, Canada.