Temporal Trends and Outcomes of Abdominal Aortic Aneurysm Care in the United States

Abstract

Background:

Endovascular aortic aneurysm repair(EVAR) has had a dynamic impact on abdominal aortic aneurysm(AAA) care, often supplanting open AAA repair(OAR). Accordingly, U.S. AAA management is often highlighted by disparities in patient selection and guideline compliance. The purpose of this analysis was to define secular trends in AAA care.

Methods:

The Society for Vascular Surgery Vascular Quality Initiative(SVS-VQI) was queried for all EVARs and OARs(2011–2021). End-points included procedure utilization, change in mortality, patient risk profile, SVS-endorsed diameter compliance, off-label EVAR use, cross-clamp location, blood loss, in-hospital complications, and post-EVAR surveillance missingness. Linear regression was used without risk-adjustment for all endpoints except for mortality and complications for which logistic regression with risk adjustment was used.

Results:

66,609 EVARs(elective, 85%[n=55,805], non-elective, 15%[n=9,976]) and 13,818 OARs(elective, 70%[n=9,706], non-elective, 30%[n=4,081]) were analyzed. Elective EVAR:OAR ratios increased(0.2/year[95%CI 0.01–0.32]) while non-elective ratios were unchanged. Elective diameter threshold non-compliance decreased for OAR(24%→17%;p=.01) but not EVAR(mean-37%). Low-risk patients increasingly underwent elective repairs(EVAR, +0.4%/year[0.2–0.6];OAR, +0.6 points/year[0.2–1.0]). Off-label EVAR frequency was unchanged(mean-39%) but intraoperative complications decreased(0.5%/year[0.2–0.9]). OAR complexity increased reflecting greater suprarenal cross-clamp rates(0.4%/year[0.1–0.8]) and blood loss(33 mL/year[19–47]).

In-hospital complications decreased for elective(0.7%/year[0.4–0.9]) and non-elective EVAR(1.7%/year[1.1–2.3]) but not OAR(mean-42%). 30-day mortality was unchanged for both elective OAR(mean-4%) and EVAR(mean-1%). Among non-elective OARs, an increase in both 30-day(0.8%/year[0.1–1.5]) and 1-year mortality(0.8%/year[0.3–1.6]) was observed. Postoperative EVAR surveillance acquisition decreased(67%→49%) while 1-year mortality among patients without imaging was 4-fold greater(9.2% vs. with imaging-2.0%;OR 4.1[3.8–4.3];p<.0001).

Conclusions:

There has been an increase in EVAR and corresponding reduction in OAR across the U.S., despite established concerns surrounding guideline adherence, reintervention, follow-up, and cost. Although EVAR morbidity has declined, OAR complication rates remain unchanged and unexpectedly high. Opportunities remain for improving AAA care delivery, patient and procedure selection, guideline compliance, and surveillance.

Introduction

Abdominal aortic aneurysm(AAA) care remains associated with significant morbidity and mortality which affects ~1–2 million Americans and over 30–40 million people globally^{1, 2}. After the seminal description of a successful open AAA repair(OAR) of a ruptured aneurysm in 1951, operative paradigms and their respective outcomes have evolved³. Timely screening and prophylactic repair have become hallmarks of AAA care to avert the sobering consequences of rupture which remains associated with ~25%−35% perioperative mortality and complication rates exceeding 50%^{4, 5}. While advances in technology have improved AAA care, significant outcome disparities persist when comparing high and low volume centers^6–11. Indeed, efforts to mitigate ongoing variation have been a central focus of professional society clinical practice guidelines(CPGs) ^{12, 13}.

In this context, the evolution and widespread adoption of endovascular AAA repair(EVAR) has had a significant impact on both elective and non-elective aneurysm care across the United States. Nevertheless, there has been sustained variation in AAA management, noted by inconsistent patient selection, conflicting societal recommendations and center guideline compliance^{6, 14, 15}. While nuanced outcomes of operative treatment have remained a focus of investigation, the secular patterns of AAA care delivery remain less defined and have important practice implications.

Therefore, the objective of this study was to describe the longitudinal changes in operative AAA care, patient selection, and corresponding associated outcomes after EVAR and OAR using a large national quality registry. We hypothesized that there would be significant changes in patient selection, procedural complexity, and outcomes as a result of endoluminal strategies being increasingly utilized relative to OARs.

Methods

Data, Materials and Code Disclosure Statement.

The data that support the findings of this study are available from the corresponding author upon reasonable request. However, since these data are owned by the Society for Vascular Surgery Patient Safety Organization (also known as the Vascular Quality Initiative), data requests would require approval through this entity.

Data source.

Data for this study was derived from the Society for Vascular Surgery Vascular Quality Initiative(SVS-VQI) OAR and EVAR registries. The SVS-VQI registries contain demographic, clinical, procedural and outcomes data from more than one million vascular procedures performed across the United States. Participating centers prospectively enter data which undergoes auditing to ensure consecutive case capture. Over the study period, 370 centers contributed data to the analysis, which included procedures performed from January 2011-December 2021. Out of hospital survival is captured by the long-term follow-up modules, as well as through established linkage with the social security death index. Further details and information about the specific registries, as well as definitions for the variables are available upon request at http://www.vascularqualityinitiative.org.

The institutional review board at the University of Florida approved this study and deemed it to be exempt from a need for informed consent(IRB#1723–2021).

Cohort derivation.

All elective and non-elective AAA repairs in both the SVS-VQI OAR and EVAR modules were reviewed. No specific a priori inclusion or exclusion criteria were broadly applied; however, due to the various end-points analyzed and necessary anatomical or longitudinal follow-up information needed to assess them, serial adjustments were made to the study populations to account for these requirements. The rate of variable missingness and how this was addressed for each end-point is described in the Supplemental Table. Moreover, we are unable to quantify the impact of the COVID19 pandemic on surveillance missingness in the VQI EVAR registry. Accordingly, remain unable to methodologically account for potential confounding effects of the pandemic on practice patterns and perioperative outcomes among patients who were non-compliant with care and thus not captured in the registry.

Definitions.

For the purposes of this analysis, elective procedures included patients presenting with asymptomatic, intact AAA. Non-elective presentations included either symptomatic intact or ruptured AAA. Patient preoperative risk was determined using the previously published and validated SVS-VQI 1-year mortality risk calculator¹⁶. Major in-hospital complications included stroke, death, myocardial infarction, congestive heart failure, arrythmia, respiratory failure, new hemodialysis requirement, leg ischemia, intestinal ischemia, and return to the operating room for bleeding. Complications are defined by the SVS-VQI and details are available on-line at http://www.vascularqualityinitiative.org.

Center volume and SVS CPG endorsed diameter threshold compliance was determined using the following criteria: ‘low-volume’ hospitals were centers that recorded <10 OARs/year(which could include a composite of elective and non-elective infrarenal AAA operations) while centers performing ≥ 10 OARs/year were deemed to be ‘high-volume’. This threshold reflected current SVS CPGs¹². Similarly, when determining whether an AAA repair was compliant with diameter standards, only elective procedures that excluded concurrent iliac aneurysms were assessed. More specifically, elective AAA procedures recorded in the OAR or EVAR registries that did not involve an iliac aneurysm(e.g., iliac diameter <20mm) were reviewed. Moreover, if the preoperative diameter was <55mm for a man or <50mm for a women, then the procedure was considered to be ‘non-compliant’. Notably, the open and endovascular AAA registries do not document other indications for repair such as saccular morphology or rapid growth rates.

‘Off-label’ EVAR designation included only elective procedures that treated anatomy outside instructions for use(IFU) of commercially available devices. An elective EVAR procedure was categorized as being ‘off-label’ if it included any of the following: proximal landing zone diameter>32mm, proximal landing zone length<10mm, iliac diameter <6mm, and/or infrarenal neck angle >60 degrees. Due to registry variable updates in 2014 and data missingness, 58% of elective EVAR procedures(260 centers; N=24,532/42,543) had information available for this component of the analysis.

‘Complicated’ EVAR was defined by any use of intraoperative adjunct(s) or documented complication such as iliac conduit, femoral-femoral bypass(due to reported association of aorto-uni-iliac stent-graft implants with inferior long-term outcomes)¹⁷, access vessel injury, unplanned aortic/iliac device use, distal/lower extremity embolectomy/thrombectomy, unintentional renal coverage, and/or conversion to open repair. As a proxy for OAR procedure-related complexity, we examined trends in supra-renal cross-clamp placement and intraoperative estimated blood loss(EBL) due to their known associations with outcomes^{18, 19}.

End-points.

The primary end-points of the study included the change in procedure utilization(e.g., center ratio of EVAR:OAR), as well as 30-day and 1-year mortality outcomes over time. Secondary end-points included temporal variation in SVS CPG diameter treatment threshold compliance, preoperative patient risk, elective ‘off-label’ and ‘complicated’ EVAR use, estimated blood loss and cross-clamp location during elective OAR, in-hospital complications, as well as adherence to postoperative EVAR imaging surveillance.

Statistics.

Median annual center case volumes were determined by counting the number of combined elective and non-elective procedures entered into the VQI by each center per year. A single average annual case volume was determined for each center that had entered cases into the VQI for at least 3 consecutive years. Because centers can enter or leave VQI at any time during a calendar year, three consecutive years were required so that the first and last year could be excluded from the calculation of the annual average. Patients were assigned risk scores using the 1-year mortality model from Neal et al.¹⁶. This model scores patients on a 0–40 point scale by risk of mortality according to preoperative values of hemoglobin, creatinine, COPD, age, BMI, maximum AAA diameter, admission and transfer status, ejection fraction as well as statin and betablocker exposure. After scoring, patients were classified as low-risk if below the median risk score for all patients or high-risk if above the median. Annual rates of post-operative imaging were determined using the EVAR and OAR long-term follow-up registries.

Since the goal of the analysis was to establish if any changes occurred in outcomes of interest over time, no risk-adjustment was performed for most end-points. This was deliberate since if changes were detected it could be linked to either changes in patient population, changes in VQI center participation, or a trend toward better/worse quality of care(or some combination of these factors). However, specific end-points such as mortality and complications are presented with both crude and risk-adjusted analyses. This was done to determine if observed changes in mortality and in-hospital complications were changing as a result of variation in the patient population over time or if it was a result of hospital-associated changes(e.g., trend toward better/worse quality of care delivery and/or changes in VQI center population). For mortality and complications endpoints, logistic regression was used to estimate the effect of time when adjusting for patient factors. For all other analyses, linear regression was used to summarize observed changes in rates or means over time. All analyses were performed using the R statistical software package(V4.2.2, the R-Foundation for Statistical Computing).

Results

AAA center volume.

Over the 10-year study interval, temporal case accrual, as well as change in median center volume is depicted in Figure 1. By 2021, over 80,000 AAA procedures were recorded with 81% being EVAR(N=66,609: elective-85%[n=55,805], non-elective-15%[n=9,976]) and 19% being OAR(N=13,818: elective-70%[n=9,706], non-elective-30%[n=4,081]). A total of 370 centers contributed data to either registry during the study period while 247 participated in both the EVAR and OAR modules. Notably, 49 hospitals provided data in both registries for all 10-years. When examining the change in the median annual center volume over time, a significant decrease in both EVAR and OAR was evident. Specifically, in 2011, median EVAR volume was 24 procedures/year which decreased to 20/year by 2021(p=.02). Similarly, a reduction in median annual OAR volume was also observed(2011, 8/year vs. 2021, 4/year;p<.0001)(Figure 1B). Further, for elective OAR, the proportion of centers performing the SVS endorsed volume threshold of ≥10 OARs/year decreased over time(43%→22%;p<.0001).

Figure 1.

Change in Abdominal Aortic Aneurysm (AAA) Procedures and Center Volume Over Time, Stratified by Endovascular Aortic Repair (EVAR) and Open AAA Repair (OAAA) A. AAA Repair Procedures By Year B. Median AAA Repair Volume By Year

EVAR and OAR utilization.

The aggregated median EVAR:OAR ratio for all repairs increased over time but was driven by higher utilization among elective procedures(all repairs, EVAR:OAR ratio +0.2/year;p=.0004; elective EVAR:OAR ratio +0.2/year;p=.04 while non-elective EVAR:OAR was not significantly changed, +0.02/year;p=.3). These results were consistent when looking among centers contributing to both registries during the entire study period(N=49), as well as across all centers over time. The proportion of elective repairs that were performed for AAA diameters below the SVS endorsed diameter thresholds did not change for EVAR(mean-37%: 2011–37% vs. 2021–33%;p=.09) while a significant reduction was noted among OARs(mean-22%: 2011–24% vs. 2021–17%;p=.0002)(Figure 2).

Figure 2.

Elective Abdominal Aortic Aneurysm Repairs Below Society for Vascular Surgery (SVS) Diameter Threshold, Stratified by Endovascular Aortic Repair (EVAR) and Open AAA Repair (OAAA)

Elective AAA patient selection and procedure complexity.

When patient risk factors were examined to determine if physiological complexity changed, significant differences were evident. Specifically, a larger proportion of subjects deemed to be low-risk(e.g., <5% preoperatively predicted 1-year mortality risk) underwent either elective EVAR or OAR. For example, 53% of EVAR patients in 2011 were classified as high-risk(e.g., ~10–22% preoperatively predicted 1-year mortality risk) but in 2021 this was 48%(low-risk +0.4 percentage points/year;p=.0008). A similar trend was evident with the elective OAR patients(low-risk +0.6 points/year;p=.009).

Among elective EVAR procedures with available anatomic information (260 centers; N=24,532/42,543 [58%]), ‘off-label’ repair incidence did not change over time and varied between 39%−41%. However, it is important to note that the rate of missing data used to assess IFU adherence has significantly increased since 2017 secondary to new center subscriptions to the VQI registry. In contrast, the frequency of complicated infrarenal EVAR cases decreased significantly(2011–50% vs. 2021–45%; −0.5 percentage points/year;p=.008). This reduction in complicated EVAR was predominantly driven by reduction in intraoperative adjunct utilization, as well as iliac vessel injury(adjunct: 2011, 21% vs. 2021, 16%, p=.007; iliac artery injury: 2011, 2.9% vs. 2016, 1.3%, p=.004). During the same period, elective OAR complexity, as determined by suprarenal cross-clamp use and EBL, appeared to increase(elective OAR procedures using suprarenal clamps increased 0.4 percentage points/year-p=.02; EBL increased 33mL/year;p=.0005).

Outcomes.

For elective and non-elective OAR, the overall incidence of any in-hospital complication did not change and was approximately 35% and 60%, respectively. By comparison, a significant reduction in elective and non-elective EVAR in-hospital complication rates was observed. In 2011, 10% of elective EVAR procedures had a recorded in-hospital complication, which subsequently declined to 4% in 2021(−0.7 percentage points/year;p=.0001). Similarly, non-elective EVAR procedure morbidity decreased 1.7 percentage points/year(2011–33% vs. 2021–19%;p=.0002)(Figure 3).

Figure 3.

Change in Any In-hospital Complication Rates after Abdominal Aortic Aneurysm (AAA) Repair, Stratified by Endovascular Aortic Repair (EVAR) and Open AAA Repair (OAAA) A. Elective Procedures Postoperative Complications By Year B. Non-elective Procedures Postoperative Complications By Year

The 30-day mortality for both elective EVAR and OAR did not change significantly during the study period(EVAR, ~1%;OAR, ~4%). When examining non-elective cases, no change occurred in the EVAR cohort(~14%); however, an increase in 30-day mortality was detected among OAR patients(2011–23% vs. 2021–29%;+0.8 percentage points/year;p=.03)(Figure 4). The analogous risk-adjusted mortality and in-hospital complication outcomes are featured in Table I.

Figure 4.

Change in Abdominal Aortic Aneurysm 30-day Mortality, Stratified by Endovascular Aortic Repair (EVAR) and Open AAA Repair (OAAA) A. Elective Procedures 30-day Mortality By Year B. Non-elective Procedures 30-day Mortality By Year

Table I.

Model-estimated effects on surgical year, unadjusted and adjusted for patient risk score (Odds Ratio {OR] per year, 95%CI, p-value) from mixed-effects logistic regression with random factor for center

Outcome, OR(95%CI), p-value	Elective EVAR	Non-elective EVAR	Elective Open AAA	Non-elective Open AAA
30-day mortality (unadjusted)	1.0(.96,1.03), .7	1.02(.99,1.05), .09	1.02(.98,1.1), .3	1.1 (1.02,1.1), .003
30-day mortality (adjusted)	1.0(.96,1.04), .9	1.03(1.0,1.1), .04	1.03(.98,1.1), .2	1.1(1.02,1.1), .0006
In-hospital complications (unadjusted)	0.88(.87,.89), <.0001	0.93(.91,.94), <.01	0.99(.97,1.00), .2	0.99(.97,1.01), .4
In-hospital complications (adjusted)	0.89(.87,.89), <.0001	0.93(.91,.95), <.01	0.99(.98,1.0), .4	0.99 (.97,1.02), .5
1-year mortality (unadjusted)	1.0(.99,1.02), .3	1.0(.98,1.03), .33	1.0(.98,1.04), .7	1.0(1.01,1.1), .01
1-year mortality (adjusted)	1.0(.997,1.03), .1	1.0(.99,1.04), .17	1.0(.98,1.1), .5	1.1(1.02,1.1), .002

EVAR, Endovascular AAA Repair; AAA, Abdominal Aortic Aneurysm Repair

One year mortality and EVAR surveillance.

The overall 1-year mortality for elective OAR and EVAR procedures did not change over time and was between ~6%−8%(Table I). In comparison, non-elective procedures demonstrated a disparate trend with no change observed among non-elective EVAR cases; however, an increase in 1-year mortality was detected after non-elective OAR(+0.8 percentage points/year;p=.04). There has been a significant decline in 1-year compliance with surveillance imaging after the index EVAR. For example, in 2015, approximately 65%−70% of patients had recorded surveillance imaging of any kind after the first 6 postoperative months; however, by 2020, this decreased to less than 50%. Among patients who survived 1-year the 2-year mortality rate was 4-fold higher among those not imaged(no imaging- 9.2% vs. received imaging-2.0%; 3-year mortality, no imaging-14.7% vs. received imaging-4.7%;p<.0001)(Supplemental Figure).

Discussion

This study is among the first to concurrently catalogue contemporary trends in procedure utilization, guideline compliance, patient selection, as well as perioperative outcomes after both elective and non-elective AAA repair. Specifically, we determined that EVAR continues to supplant OAR in the elective setting though not for non-elective cases. These national findings were associated with an increasing percentage of lower volume centers performing OARs and would seemingly refute any notion that there has been a trend of increased AAA care regionalization. Interestingly, lower risk patients are increasingly undergoing elective EVAR and OAR with ~35–40% of the endovascular procedures performed for small AAA(vs. ~20% OAR). Notably, while procedural complexity is decreasing among EVARs, the OAR technical intricacy appears to be increasing over time. Overall complication and mortality outcomes for elective EVAR and OAR remain stable but non-elective OAR 30-day and 1-year mortality is increasing. Moreover, it appears that compliance with post-EVAR surveillance has decreased over time. The underlying causes for this phenomena and related sequelae needs further study.

The findings herein identify several important aspects of AAA care provision that represent opportunities for targeted quality improvement efforts that are designed to better align both academic and community-based practice with SVS endorsed AAA CPGs¹². The fact that 40% of elective EVAR procedures are performed in patients with aneurysm diameters below current societally endorsed treatment thresholds is notable^{9, 15}. This rate is 2–4 fold higher than any other nation and deserves more dedicated study to understand what is driving this practice, as it remains unsupported by Level 1 evidence and is not cost-effective^{9, 15, 20}. Additional registry and large data analyses are necessary to continue to define and facilitate patient and procedure selection which may bolster initiatives to optimize practice guideline compliance which has been shown to improve outcomes¹⁴. Importantly, the current study also provides a comprehensive accounting of expected results for both elective and non-elective repair which informs stakeholders while providing minimum standards to lower volume centers that continue to deliver AAA care for benchmarking to higher volume hospitals.

The observation that EVAR adoption has not plateaued for elective indications has several important consequences given the growing concerns surrounding vascular surgery training^21–23. It is currently estimated that the average graduate will perform <10 OARs during their vascular residency or fellowship^{23, 24}. Not surprisingly, the impact that diminished OAR experience will have on AAA services is amplified once a trainee transitions to practice when they must perform these cases independently²⁵. The reflexive response to this finding is that AAA care should be regionalized to higher volume centers, due to known associations with improved outcomes after OAR^{10, 11, 26–28}, which the authors generally support. However, declining annual OAR volume trends among VQI centers in this analysis does not provide a signal that this is in fact occurring in a homogeneous fashion nationally. Even if the primary driver for this observation is growth of the OAR registry to include more low volume centers, this still underscores the fact that >75% of VQI centers do not meet current volume standards¹⁵. It would appear that despite SVS endorsed recommendations¹² and over 40 years of evidence in support of AAA regionalization²⁹ that novel strategies including financial incentives may be necessary before this can be fully implemented in U.S. practice^{8, 30}.

Moreover, concepts such as aortic centers of excellence^{31, 32}, as well as generation of focused aortic training fellowships have been discussed among educators and societal leaders but the operationalization of these ideas and its scalability to address a national workforce problem have yet to be clarified^{30, 33}. The hospital surgical workforce will be increasingly comprised of new graduates, as well as established surgeons who remain exposed to diminished OAR case volumes over time. Accordingly, healthcare leaders may ultimately contemplate changes to credentialing paradigms and on-boarding strategies³⁰. This reality may catalyze requirements for instituting more formal clinical mentoring programs, as well as iterative competency and quality-based assessments that should include standards endorsed by professional societies^{30, 33, 34}. Indeed, this highlights how the promise of EVAR to deliver improved outcomes has several unanticipated sequelae in real-world practice. These issues are compounded by fiscal healthcare stewardship and the emergence of longer-term risks such as sac expansion, reintervention, loss to follow-up/imaging surveillance, as well as late rupture. These findings underscore the need for a more prudent and balanced approach to the continued adoption and utilization of endoluminal strategies^{20, 35–37}. Nowhere is this more evident than in the complex aneurysm space where a growing number of surgeons are positioned to ably perform branched/fenestrated EVARs though have diminished experience with the corollary open repair techniques.

In this analysis, we identified that elective EVAR performed outside the IFU criteria, occurred in roughly 40% of cases. This finding is concordant with other studies³ and highlights the importance of registry data since these patients are not included in randomized controlled trials. Interestingly, there is general agreement that treatment of AAA outside the IFU is associated with sac enlargement and reintervention during follow-up but the association is less evident with all-cause or aorta-related mortality^{38, 39}. While non-IFU EVAR utilization did not change over time; the frequency of more complex elective EVAR procedures did decrease significantly. This observation would support the narrative that lower risk patients with smaller AAA and more straightforward anatomy were being selected for elective EVAR treatment which not surprisingly was associated with improved short-term outcomes. Concordantly, there was a reciprocal increase in the operative complexity of OARs, which was consistent with prior studies^{40, 41}. In addition, it is important to highlight that ~40% of elective EVAR procedures were performed in patients with AAA diameters below currently endorsed diameter treatment thresholds. Although alternative indications such as saccular morphology or rapid expansion may account for a percentage of real-world ‘off guideline’ repairs, these numbers have been historically documented to be a dramatic outlier (e.g., 2–6 fold higher rate in US) relative to other international registries⁶ and thus would fail to account for our observation.

Interestingly, although centers were performing fewer OARs during the study period, that were more frequently associated with suprarenal cross-clamps and higher EBL, it did not translate into measurable decline in overall elective outcomes. However, higher non-elective OAR mortality was detected. Although speculative, it would appear that centers and providers have maintained enough infrastructure and institutional memory surrounding elective OAR to offset the impact of lower volumes with increasing procedural complexity, for the time being. Notably, it remains to be determined if this system-based practice resiliency can continue to stem the unrelenting pressures imposed by unfettered EVAR utilization. Nevertheless, the increasing non-elective mortality signal with OARs may herald the erosion of the unique teams, clinical decision-making, technical expertise, and postoperative care pathways that are needed to achieve optimal outcomes for ruptured AAA management.

One of the primary goals of this study was to generate a contemporary perspective on AAA care provision in the U.S. by describing EVAR and OAR utilization patterns within the context of patient selection and outcomes over time. This was in part to foster a larger discussion about how to define quality as it pertains to either EVAR or OAR. Historical benchmarks used for inter-hospital comparisons for OAR predominantly focus on perioperative complications and all-cause mortality but these parameters are less informative about EVAR quality given the significantly lower morbidity and mortality signals in the early postoperative period. Therefore, based on the temporal trends reported in this study, there appears to be a need to establish consensus about how to best define and subsequently measure AAA quality. This may include consideration of endorsing novel procedure and/or center-specific composite quality metrics that align with SVS CPGs^{12, 14, 42}.

The central tenets that govern the provision of high quality AAA care include targeted screening and education, judicious repair and surveillance. The impact of loss to follow-up and lack of post-EVAR imaging is well-described³⁶ but this analysis determined that more recent negative trends overlap with the COVID-19 pandemic. Due to the abrupt disruption of surveillance programs and access to care that occurred during the last 3 years, it remains to be determined what the longitudinal influence will be on timely detection of endograft-related complications during this interval⁴³. These concerns are magnified by the dominant position that EVAR has in U.S. practice, and further highlights another opportunity for policy generation and quality improvement initiatives that incentivizes more balanced application of EVAR.

This study has several limitations. First, concerns about generalizability of these findings to non-VQI centers has been frequently cited and may reflect subtle inclusion disparities in the registry when compared to alternative data sources; however, the temporal growth of the registries now accounts for over 25% of infrarenal AAA care in the U.S.⁴⁴ so observations from this analysis are likely to be a reasonable proxy for national trends. We concede that OAR center volume estimates do not account for the contribution of complex repairs(e.g., suprarenal/thoracoabdominal aneurysms) or non-aneurysm related indications(e.g., aorto-iliac occlusive disease/aorto-femoral reconstructions), which are relevant to phenotyping institutional volume and supporting trainee education^{23, 45}. Moreover, we have deliberately omitted the inclusion of fenestrated/branched EVAR patients which we believe constitutes a profoundly anatomically heterogeneous cohort which may confound comparative analysis. Nevertheless, we acknowledge that there may be a subset of patients in the FEVAR cohort in whom this analysis would be conceptually applicable for inclusion. Similarly, we do not have information on surgeon experience or volume that has known associations with AAA outcomes^{27, 45}. The definitions that were employed for ‘off-label’ or ‘complicated’ EVAR could easily be scrutinized; however, we used conservative anatomical parameters that have been previously published and/or are known to predict longer term adverse events^{3, 39}. Lastly, the current analysis only focused on short-term outcomes so we cannot comment on the changes in longer term outcomes that may have occurred during the study interval.

Conclusion

In conclusion, there has been a sustained increase in elective EVAR utilization across the U.S. over time, despite documented concerns surrounding reintervention, cost, and corresponding reduction in OAR volume. Compliance with SVS endorsed diameter treatment thresholds for elective EVAR continues to lag and needs further study to understand the drivers for this practice variation, since it does not occur at similar rates in other international healthcare systems. Short-term morbidity associated with EVAR has improved but OAR complication rates remain high. Moreover, the implications of offering elective EVAR to younger patients with fewer comorbidities along with high rates of loss to follow-up within the first 1–2 years postoperatively will need continued examination over time especially since this remains associated with increased mortality. The reduction in exposure to OAR during training and thereafter in independent practice may have increasing negative implications on AAA management which requires novel approaches to early career clinical mentorship, maintenance of credentialing, and patient-centered quality improvement initiatives. Based on these real-world observations, there appears to be ample opportunity for quality improvement in AAA care delivery, patient and procedure selection, guideline compliance, and post-EVAR surveillance.

What is Known:

• Abdominal aortic aneurysm (AAA) operative outcomes persistently vary when comparing high and low volume centers in contemporary practice.

• While the widespread adoption of endovascular aneurysm repair (EVAR) has transformed current AAA care there is sustained variation in patient selection, and guideline compliance across the United States.

What the Study Adds:

• In this real world national registry analysis of current AAA repair in the U.S. there has been a sustained increase in EVAR utilization despite concerns regarding reintervention rates, procedure associated costs, and consequent reductions in open AAA repair.

• Current trends in compliance with recommended size thresholds for AAA repair remain poor in the U.S. and do not occur at similar rates in alternative international healthcare systems.

• The consequent reduction in open AAA repair rates have real world implications for the future of aneurysm care in the U.S. given the known limitations of EVAR and offers insight into opportunities for dedicated quality improvement in AAA care.

Non-standard Abbreviations and Acronyms(in order of appearance):

AAA

abdominal aortic aneurysm

EVAR

endovascular abdominal aortic aneurysm repair

OAR

open abdominal aortic aneurysm repair

SVS-VQI

Society for Vascular Surgery Vascular Quality Initiative

IRB

Institutional Review Board

CPG

clinical practice guideline

IFU

instructions for use

COPD

chronic obstructive pulmonary disease

BMI

body mass index

EBL

estimated blood loss