Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair

Abstract

Purpose:

Our objective was to investigate whether patients who receive anticoagulation therapy have different outcomes after endovascular aneurysm repair (EVAR) from those who do not.

Materials and Methods:

We conducted a systematic review of studies that compared outcomes of EVAR in patients who were on therapeutic anticoagulation vs those who were not. We developed and reported the review in accordance with the PRISMA guidelines with a registered protocol (CRD42022375894). The Ovid interface was used to search Medical Literature Analysis and Retrieval System Online (MEDLINE), Excerpta Medica Database (EMBASE), and Cochrane Central Register of Controlled Trials (CENTRAL) up to November 2022. The quality of studies was assessed with the Newcastle-Ottawa Scale (NOS) (maximum score=9), and the evidence was appraised with the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) framework. The hazard ratio (HR) and 95% confidence interval (CI) was the effect estimate in time-to-event meta-analyses, calculated using the inverse-variance statistical method and random-effects models.

Results:

Sixteen studies qualified for inclusion reporting a total of 35 739 individuals. Anticoagulated patients had a statistically significantly higher hazard of death (HR=1.93, 95% CI=1.03-3.63), endoleak (HR=2.13, 95% CI=1.55-2.93), reintervention (HR=1.79, 95% CI=1.27-2.52), and aneurysm sac expansion (HR=2.72, 95% CI=1.57-4.72) than patients not receiving anticoagulation therapy. The median score on the NOS was 7 (range=4-9). The certainty of evidence was very low for mortality and reintervention and low for endoleak and sac expansion.

Conclusions:

Anticoagulation is a poor prognostic factor after standard EVAR and should be considered in decision-making, consent processes, and surveillance strategies.

Clinical Impact

The number of individuals who take anticoagulation treatment has been rapidly increasing over the recent years. We aimed to investigate the effect of such treatment on outcomes after endovascular aneurysm repair (EVAR). Anticoagulated patients were found to have increased mortality, endoleak, and reintervention rates after EVAR compared to their non-anticoagulated counterparts. Anticoagulation therapy has a prognostic role in EVAR and should be considered in decision making and EVAR surveillance. Anticoagulated patients need to be informed of the higher failure rates of EVAR, and intensified surveillance strategies may need to be implemented in this patient cohort.

Keywords

aortic aneurysm endovascular aneurysm repair EVAR anticoagulation warfarin

Introduction

Endovascular aneurysm repair (EVAR) is currently the preferred treatment method for patients with abdominal aortic aneurysm (AAA) and the standard of care according to clinical practice guidelines, with open repair being reserved for patients with specific clinical characteristics and aortic anatomies.¹ Abdominal aortic aneurysm is predominantly a disease of the elderly; thus, EVAR, being a less invasive operative procedure, confers incremental benefits in this patient cohort, which are particularly important in elderly and frail patients with an increased prevalence of cardiovascular comorbidities.^2,3 Commonly, individuals with AAA take anticoagulation treatment for other clinical conditions, such as cardiac arrhythmia or thromboembolic events, which, in theory, can affect the results of EVAR. The number of individuals who take anticoagulation treatment has been rapidly increasing over the recent years, and there appears to be a shift from traditional anticoagulation agents, such as warfarin, to newer direct oral anticoagulants.^4,5 The effect of such treatment on clinical outcomes after EVAR has not yet been established. Increased rates of reintervention, rupture, and aneurysm-related mortality after EVAR have cast a shadow over its long-term effectiveness and durability compared with open repair, and the increasing use of anticoagulation treatment may account for such adverse events. Confirming an association between anticoagulation use and adverse events after EVAR has important implications in clinical decision-making and surveillance strategies, eg, enhanced follow-up may be needed for this cohort of patients. We aimed to investigate whether individuals who take anticoagulation treatment at the time of or after EVAR have worse clinical outcomes than those who are not on such treatment.

Materials and Methods

Review Registration and Reporting

The objectives and methodology of the review were prespecified in a protocol, which was registered in PROSPERO (CRD42022375894). The review was developed according to principles described in the Cochrane Handbook for Systematic Reviews of Interventions.⁶ Reporting of the review complied with the updated PRISMA 2020 guidelines.⁷ The PRISMA 2020 checklist was generated using a Shiny App available at https://prisma.shinyapps.io/checklist/ (available upon request) and the study flow diagram using a Shiny App available at https://www.eshackathon.org/software/PRISMA2020.html.

Eligibility Criteria

Types of studies

Eligible studies reported comparative outcomes of standard EVAR in patients taking therapeutic anticoagulation vs those not on such treatment. Studies of any design were eligible, eg, retrospective or prospective cohort studies, case-control studies, cross-sectional studies, and randomized clinical trials. There was no limit in the duration of follow-up post-EVAR that studies should report to be included in the review. Owing to funding constraints, only articles published in the English language were considered. No time constraints were applied, and the entire literature was searched from the date of the first publication on EVAR. Studies published in an abstract format, ie, no full text available is available, and those with an unpublished manuscript were excluded.

Types of participants

Eligible participants had an intact or ruptured infra-renal AAA treated with standard EVAR. Endovascular aneurysm repair could be performed with any commercially available bifurcated or aorto-uni-iliac device and any type of anesthesia, ie, local, regional, or general, using percutaneous access or surgical exposure of the femoral arteries. Patients treated with complex endovascular procedures—eg, fenestrated or branched EVAR—were excluded.

Types of prognostic factor

The prognostic factor of interest is therapeutic anticoagulation at the time of EVAR or at any time during the post-EVAR follow-up. From here on, therapeutic anticoagulation is referred to as anticoagulation. Any type of anticoagulant agent or anticoagulation regime was considered including unfractionated heparin, low molecular weight heparin, pentasaccharide, vitamin K antagonist, direct thrombin inhibitor, or factor Xa inhibitor. Any indication for therapeutic dose anticoagulation was considered including, but not limited to, atrial fibrillation, mechanical cardiac valve, or venous thromboembolic disease. Comparator to therapeutic dose anticoagulation was no antithrombotic treatment, any antiplatelet regime, a combination of antiplatelet and nontherapeutic dose anticoagulation, or nontherapeutic dose anticoagulation alone.

Types of outcomes

Studies should report at least 1 primary or secondary outcome to be eligible for inclusion. Primary outcomes are all-cause mortality, AAA-related mortality, and AAA rupture. Secondary outcomes are endoleak, reintervention, and sac expansion.

Information Sources and Search Strategy

The literature search strategy was developed by the senior review author, who has experience in outreach, knowledge, and evidence search, with support from library services at our institution. Access to health care databases was via online sources of institutional library services.

MEDLINE (Medical Literature Analysis and Retrieval System Online) and EMBASE (Excerpta Medica Database) were searched using the Ovid interface. CENTRAL (Cochrane Central Register of Controlled Trials) was also searched for eligible studies. The PICO (patient, intervention, comparison, outcome) framework was used to develop the search strategies. A combination of controlled vocabulary (subject headings) and free-text terms was used to search electronic bibliographic databases. Subject headings/thesaurus trees, search operators, and search limits in each of the above databases were adapted accordingly. Search syntaxes in each of the databases are presented in Appendix 1. Electronic searches were last run on November 30, 2022. A second-level search was conducted by manually interrogating the bibliographic list of articles that qualified for inclusion in this review and previous systematic review reports on the same or a similar topic.

Study Selection and Data Collection Process

Two review authors screened each record and each report retrieved. The reviewers worked independently at each stage of the screening process. Disagreements between screeners were resolved by discussions, with the senior review author acting as an arbitrator, if necessary.

Collected data from individual studies were prespecified during the development of the review protocol. Additional relevant data identified during the data collection process were extracted and entered into a Microsoft Excel spreadsheet. Data were extracted from the main text, figures, and tables of the original publications. Data extracted from published Kaplan-Meier curves were digitalized using an open-source software (http://plotdigitizer.sourceforge.net). Two review authors extracted data from selected studies independently. Once checked for accuracy and consistency, data were subsequently transferred to the statistical computer program. Only published material was considered, and no study investigators were contacted to obtain or confirm relevant information.

Data items were grouped as follows: (1) study level data: first author, journal, year of publication, study period, country where the study was conducted, type of anticoagulant agent, and length of follow-up; (2) individual study population data: age, sex, maximum AAA diameter, number of patent lumbar arteries, and inferior mesenteric artery patency; (3) data pertaining to risk of bias assessment; and (4) outcome data, as outlined in section “Eligibility criteria.”

Study Risk of Bias Assessment and Evidence Appraisal

The quality of nonrandomized studies was assessed with the Newcastle-Ottawa Scale (NOS).⁸ The tool uses a star system, with a maximum score of 9, by which studies are judged on 3 broad methodological areas: selection of the study groups, comparability of the groups, and ascertainment of the exposure or outcome of interest for case-control and cohort studies, respectively. Studies achieving less than 7 stars were arbitrarily deemed of lower methodological quality for the purposes of sensitivity analyses. The assessment was done by 2 review authors independently, and the results were then cross-checked for consistency.

The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) framework was used to assess the certainty of the body of evidence. The GRADE approach specifies 4 levels of evidence for a given outcome, namely, high, moderate, low, and very low. A summary of findings table was generated using an online platform (https://gdt.gradepro.org/app/).⁹ The GRADE assessment was performed for the primary (time-to-event) analyses.

Effect Measures and Synthesis Methods

Primary analyses were time-to-event data meta-analyses for primary and secondary outcomes using the hazard ratio (HR) and 95% confidence interval (CI) as the effect estimate. A mixture of direct (eg, from reported HR with CI) and indirect methods (eg. from survival curves incorporating numbers at risk) was used to calculate individual study log HR and its standard error (SE). An open-source software/spreadsheet was used to facilitate the estimation of HR from published summary statistics or data extracted from Kaplan-Meier curves.^10,11

Secondary analyses used dichotomous data for the primary and secondary outcomes. Data synthesis of binary data was performed using the odds ratio (OR) and 95% CI as the summary statistic. Numbers of events and total numbers of patients in each group were extracted from individual studies.

Data were then inputted into the Review Manager (RevMan) computer program (Version 5.4, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2020). Effect estimates for binary outcomes were calculated using the Mantel-Haenszel statistical method. Time-to-event data meta-analyses were conducted using the inverse-variance method. A forest plot was generated for graphical presentation of meta-analysis for each outcome.

Important between-study heterogeneity was anticipated, eg, due to different anticoagulant agents, and therefore, the random-effects method for meta-analysis was used.¹² The extent and impact of between-study heterogeneity were assessed by inspecting the forest plots and by calculating the τ² and the I² statistics, respectively. Inconsistency was quantified and interpreted as previously described.¹³ To explain potential sources of heterogeneity, subgroup analyses for type of anticoagulation treatment—eg, vitamin K and newer direct oral anticoagulants—were planned but not performed due to insufficient data.

Sensitivity analyses were conducted to explore the robustness of meta-analyses by excluding studies that were deemed to be of high risk of bias (score on NOS <7). Furthermore, the analyses were repeated after removing 1 study at a time to examine the impact of each study on the overall meta-analysis.

To assess risk of bias due to missing results in a synthesis arising from reporting biases, the effect by the inverse of its SE was plotted for each study. The possibility of publication bias was assessed visually evaluating the symmetry of the funnel plot. Such reporting bias assessments would be conducted for outcomes reported by at least 10 studies.

Results

Literature Search Results and Study Characteristics

Database searches retrieved 1769 reports. Sixteen studies^14
–29 qualified for inclusion in quantitative syntheses. All reported comparative data on patients who were on anticoagulation therapy at the time of EVAR or during the post-EVAR follow-up vs a control group of patients who were not taking anticoagulation treatment at the time of or after the EVAR. The PRISMA flow diagram is presented in Figure 1.

Figure 1.

Literature flow diagram generated using a Shiny App available at https://www.eshackathon.org/software/PRISMA2020.html.

Fourteen studies^{14
–24,26,28,29} were observational (11 retrospective and 3 prospective) and another 2²⁵,²⁷ reported administrative databases (from the Society for Vascular Surgery Vascular Quality Initiative). The 2 studies that used the Vascular Quality Initiative databases had overlapping populations with study periods from 2003 to 2017 and 2003 and 2019 but reported different outcomes and therefore were considered as reports of the same study for the purposes of the analysis. The studies were published between 2002 and 2022, and their recruitment period spanned from 1997 to 2019.

The studies reported a total of 35 739 patients, of whom 3180 were on anticoagulation therapy and the remaining 32 559 were not. Most studies reported use of warfarin. One study also included patients receiving low molecular weight heparin and another 2 included patients receiving direct oral anticoagulants. The mean/median follow-up ranged across the studies from 1.1 to 6.2 years. The individual study characteristics are summarized in Table 1. Most patients included in the analysis were male in their seventies. Baseline characteristics of patients are summarized in Table 2.

Table 1.

Characteristics of Studies Reporting Comparisons of Patients With Versus Without Anticoagulation Treatment at the Time of or After Endovascular Aneurysm Repair.

Studies	Country	Design	Anticoagulant agent	Recruitment period	Number of patients (anticoagulation/control)	Follow-up
Fairman et al¹⁴	The United States	Prospective	Warfarin	NR	36/196	Mean 18 months
Biebl et al¹⁵	The United States	Retrospective	Warfarin	1999-2003	21/161	Mean 16.3±12.6 months
Bobadilla et al¹⁶	The United States	Prospective	Warfarin	2000-2007	24/103	Mean 2.14 years
Abullarange et al¹⁷	The United States	Retrospective	Warfarin	1999-2007	70/525	Median 34.8 months (range=6.4-121.2)
Johnson et al¹⁸	The United States	Retrospective	Warfarin	2003-2011	68/295	Mean 29 months
De Rango et al¹⁹	Italy	Retrospective	VKA+heparin	1997-2011	103/1306	Mean 64.3±45.2 months
Wild et al²⁰	The United Kingdom	Retrospective	Warfarin	2006-2011	45/362	Median 48.5 months (range=14-85.2)
Lal et al²⁹	The United States	Retrospective	Warfarin	2002-2008	42/397	Mean 6.2±2.4 years
Seike et al²³	Japan	Retrospective	Warfarin	2007-2013	29/180	Mean 37±12 months
Marcos et al²²	Spain	Prospective	NR	2003-2011	9/78	Mean 41.5 months
Kumar et al²¹	Australia	Retrospective	Warfarin	2009-2013	68/625	NR
Castelli et al²⁴	Argentina	Retrospective	Warfarin/dabigatran	2009-2014	33/308	Median 16 months (range=3-33)
Flohr et al²⁵	The United States	Administrative database VQI	Warfarin + DOAC	2013-2019	2303/27 480	12 months
D’Oria et al²⁶	Italy	Retrospective	Warfarin	2010-2013	12/76	Median 4.3 years (IQR=1.4-7.2)
Kong et al²⁷	The United States	Administrative database VQI	Warfarin + DOAC	2003-2017	301/301	Mean 1.12±0.63 years
Morisaki et al²⁸	Japan	Retrospective	NR	2007-2019	16/166	Median 18 months

±indicates standard deviation.

Abbreviations: DOAC, direct oral anticoagulant; IQR, interquartile range; NR, not reported; VKA, vitamin K antagonist; VQI, Vascular Quality Initiative.

Table 2.

Patient Characteristics of Studies Reporting Comparisons of Patients With Versus Without Anticoagulation Treatment at the Time of or After Endovascular Aneurysm Repair.

Studies	Group	Number of patients	Age	Number of male sex	Aneurysm diameter	IMA patent	Number of patent lumbar arteries
Fairman et al¹⁴	Anticoagulation	36	NR	NR	NR	NR	NR
Fairman et al¹⁴	Control	196	NR	NR	NR	NR	NR
Biebl et al¹⁵	Anticoagulation	21	Mean 75.0±5.3	NR	Mean 56.9±6.8	NR	NR
Biebl et al¹⁵	Control	161	Mean 75.4±7.4	NR	Mean 57.4±8.9	NR	NR
Bobadilla et al¹⁶	Anticoagulation	24	Mean 72.3±8.5	22	Mean 6.34±1.2	NR	NR
Bobadilla et al¹⁶	Control	103	Mean 74.1±8.3	91	Mean 5.77±1.2	NR	NR
Abullarange et al¹⁷	Anticoagulation	70	NR	NR	NR	NR	NR
Abullarange et al¹⁷	Control	525	NR	NR	NR	NR	NR
Johnson et al¹⁸	Anticoagulation	68	Median 74.5	59	Mean 57.5	NR	NR
Johnson et al¹⁸	Control	295	Median 73	244	Mean 56	NR	NR
De Rango et al¹⁹	Anticoagulation	103	Mean 73.57±7.2	91	Mean 56.43±9.8	NR	NR
De Rango et al¹⁹	Control	1306	Mean 72.86±7.7	1196	Mean 54.65±9.7	NR	NR
Wild et al²⁰	Anticoagulation	45	Median 74.7 (range=60.4-91.7)	36	NR	NR	NR
Wild et al²⁰	Control	362	Median 74.6 (range=54.8-90.6) Median 74.9 (range=39.9-91.1)^a	331	NR	NR	NR
Lal et al²⁹	Anticoagulation	42	NR	NR	NR	NR	NR
Lal et al²⁹	Control	397	NR	NR	NR	NR	NR
Seike et al²³	Anticoagulation	29	Mean 78±9	23	Mean 49±7	69%	Mean 3.7±1.9
Seike et al²³	Control	180	Mean 78±7	136	Mean 48±6	68%	Mean 3.9±1.9
Marcos et al²²	Anticoagulation	9	Mean 61.3±3	NR	NR	NR	NR
Marcos et al²²	Control	78	Mean 56.8±2 ASA (n=50) Mean 52.8±4 Clop (n=16) Mean 59.4±4 No therapy (n=12)	NR	NR	NR	NR
Kumar et al²¹	Anticoagulation	68	NR	NR	NR	NR	NR
Kumar et al²¹	Control	625	NR	NR	NR	NR	NR
Castelli et al²⁴	Anticoagulation	33	Mean 75.8±7.9^b	31	Median 60 (IQR=50-68)	NR	NR
Castelli et al²⁴	Control	308	Mean 72.4±7.7^b	280	Median 57 (IQR=52-63)	NR	NR
Flohr et al²⁵	Anticoagulation	2303	NR	NR	NR	NR	NR
Flohr et al²⁵	Control	27 480	NR	NR	NR	NR	NR
D’Oria et al²⁶	Anticoagulation	12	NR	NR	NR	NR	NR
D’Oria et al²⁶	Control	76	NR	NR	NR	NR	NR
Kong et al²⁷	Anticoagulation	301	Mean 74.1 (SEM 0.4)	263	NR	NR	NR
Kong et al²⁷	Control	301	Mean 74.4 (SEM 0.5)	262	NR	NR	NR
Morisaki et al²⁸	Anticoagulation	16	Mean 76.2±9.0	14	Mean 49.1±4.3	50%	Mean 4.63±1.02
Morisaki et al²⁸	Control	166	Mean 74.7±7.5	148	Mean 49.5±6.1	50%	Mean 4.16±1.47

±indicates standard deviation.

Abbreviations: IMA, inferior mesenteric artery; IQR, interquartile range; NR, not reported; SEM, standard error of mean.

n=292/n=70 for patients on antiplatelet therapy/no antithrombotic treatment.

Statistically significant differences between anticoagulation and control groups.

Results of the Risk of Bias Assessment and Evidence Appraisal

The median score on the NOS was 7 (range=4-9). In general, issues were identified with comparability on the basis of the design and/or analysis and with ascertainment of exposure and assessment of outcome for case-control and cohort studies, respectively. The results of the NOS assessment are presented in Supplementary Tables 1 and 2.

The results of the GRADE assessment are presented in a summary of evidence table (Table 3). The level of evidence was very low for most time-to-event outcomes, with the exception of freedom from endoleak and freedom from sac expansion, for which the certainty of evidence was judged to be low. Downgrading of the evidence was mainly due to high risk of bias of studies, inconsistency resulting from between-study heterogeneity, and imprecision.

Table 3.

Summary of Findings Presenting the Certainty of Evidence Assessed With the Grading of Recommendations, Assessment, Development, and Evaluation Method.

Quality assessment							Number of patients		Effect (95% CI)	Quality	Importance
Number of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Anticoagulation	No anticoagulation	Effect (95% CI)	Quality	Importance
Overall mortality
4	Observational studies	Serious^a	Serious^b	No serious indirectness	No serious imprecision	None	2495	29 572	HR 1.93 (1.03-3.63)	⊕OOO Very low	Critical
Endoleak
6	Observational studies	Serious^a	No serious inconsistency	No serious indirectness	No serious imprecision	Strong association^c	482	2190	HR 2.13 (1.55-2.93)	⊕⊕OO Low	Important
Reintervention
4	Observational studies	Serious^a	No serious inconsistency	No serious indirectness	No serious imprecision	None	493	2063	HR 1.79 (1.27-2.52)	⊕OOO Very low	Important
Sac expansion
2	Observational studies	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious^d	Strong association^c	45	346	HR 2.72 (1.57-4.72)	⊕⊕OO Low	Important

Abbreviations: CI, confidence interval; HR, hazard ratio.

Information derived from studies with high risk of bias.

Significant heterogeneity between studies.

HR>2.

Small sample size.

Results of Synthesis

Forest plots for time-to-event (primary analyses) and binary data (secondary analyses) meta-analyses are presented in Figure 2 and Supplementary Figure 1, respectively. The methods of calculation of HR and information on time-to-event outcome data are presented in Supplementary Table 3.

Figure 2.

Forest plots for primary and secondary post-endovascular aneurysm repair outcomes for the comparison of anticoagulation versus no anticoagulation. The solid squares denote the hazard ratios, the horizontal lines represent the 95% confidence intervals, and the diamonds denote the pooled hazard ratios. (A) Overall mortality. (B) Endoleak. (C) Reintervention. (D) Aneurysm sac expansion. (E) Reintervention-free survival. SE, standard error; IV, inverse variance; CI, confidence interval.

Primary outcomes

Overall mortality

Time-to-event data on overall mortality were reported in 4 studies^{15,19,21,25,27} that included a total of 32 067 patients (2495 on anticoagulation and 29 572 controls). Patients on anticoagulation had a statistically significantly higher hazard of death at any time after the EVAR than patients not receiving anticoagulation therapy (HR=1.93, 95% CI=1.03-3.63, p=0.04). The between-study heterogeneity was considerable (p=0.0001, I²=88%). Similarly, meta-analysis of 3 studies^20,22,23 reporting binary data on 1705 patients (141 on anticoagulation and 1564 controls) found statistically significantly higher odds of mortality in patients receiving anticoagulation treatment (OR=2.64, 95% CI=1.41-4.69, p=0.002; heterogeneity: p=0.92, I²=0%).

Aneurysm-related mortality

One study²⁴ reported comparative data on aneurysm-related mortality with no statistically significant difference between patients with and without anticoagulation (3.03% in anticoagulated patients and 2.59% in controls, p=0.44). No data suitable for meta-analysis were available.

Aneurysm rupture

No time-to-event data on aneurysm rupture were available for meta-analysis. Dichotomous data on AAA rupture could be extracted from 3 studies^16,18,23 reporting a total of 699 patients (121 on anticoagulation and 578 controls). No statistically significant difference was found between patients who received anticoagulation therapy and those who did not (OR=2.25, 95% CI=0.37-13.53, p=0.38; heterogeneity: p=0.43, I²=0%).

Secondary outcomes

Endoleak

Time-to-event data on freedom from endoleak were reported by 7 studies^{15,19,23,25,27,28} with a total of 2672 patients (482 on anticoagulation and 2190 controls). Patients on anticoagulation had a statistically significantly higher hazard of developing endoleak than patients not receiving anticoagulation therapy (HR=2.13, 95% CI=1.55-2.93, p<0.00001) with an unimportant between-study statistical heterogeneity (p=0.22, I²=29%). Similarly, meta-analysis of 6 studies^{16,18,20,25
–27,29} reporting binary data on 2026 patients (492 on anticoagulation and 1534 controls) found statistically significantly higher odds of endoleak at any time in patients receiving anticoagulation treatment (OR=2.64, 95% CI=1.41-4.69, p=0.002; heterogeneity: p=0.92, I²=0%).

Binary data on early (detected within 30 days after the procedure) or perioperative endoleak were reported by 5 studies^{14,15,19,20,23} with a total of 2439 patients (234 on anticoagulation and 2205 controls). Patients on anticoagulation treatment had statistically significantly higher odds of early/perioperative endoleak (OR=1.55, 95% CI=1.10-2.19, p=0.01; heterogeneity: p=0.86, I²=0%).

Reintervention

Time-to-event data on freedom from reintervention were reported by 4 studies^{15,18,20,25,27} that included a total of 2556 patients (493 on anticoagulation and 2063 controls). Patients receiving anticoagulation therapy had a statistically significantly higher hazard of secondary intervention than patients not on anticoagulation therapy (HR=1.79, 95% CI=1.27-2.52, p=0.0008) with an unimportant between-study statistical heterogeneity (p=0.39, I²=1%).

Similarly, meta-analysis of 6 studies^{15,16,18,20,23,25,27} reporting binary data on 31 071 patients (2490 on anticoagulation and 28 581 controls) found statistically significantly higher odds of reintervention in patients receiving anticoagulation treatment (OR=2.06, 95% CI=1.08-3.93, p=0.03; heterogeneity: p=0.004, I²=72%). The type of reintervention was reported in 4 studies^15
–17,24 with the primary cause being endoleak followed by limb-related complications, such as stenosis or occlusion. The vast majority of secondary procedures were performed by endovascular means.

Sac expansion

Time-to-event data on sac expansion were reported by 2 studies^23,28 that included a total of 391 patients (45 on anticoagulation and 346 controls). Patients on anticoagulation had a statistically significantly higher hazard of developing sac expansion than those not receiving anticoagulation (HR=2.72, 95% CI=1.57-4.72, p=0.0004) with no statistical evidence of between-study heterogeneity (p=0.58, I²=0%).

Similarly, meta-analysis of 8 studies^{15,16,18,20,22,23,25,27} reporting binary data on 31 158 patients (2499 on anticoagulation and 28 659 controls) found statistically significantly higher odds of sac expansion in patients receiving anticoagulation treatment (OR=2.20, 95% CI=1.11-4.33, p=0.02; heterogeneity: p=0.03, I²=57%).

Additional outcomes

Persistent type II endoleak

Binary data on persistent type II endoleak were reported by 7 studies^{14
–17,23,25,28} including a total of 31 310 patients (2499 on anticoagulation and 28 811 controls). Definitions of “persistent” are presented in Supplementary Table 4. Patients on anticoagulation had statistically significantly higher odds of persistent endoleak (OR=2.63, 95% CI=2.35-2.95, p=0.00001; heterogeneity: p=0.46, I²=0%).

Spontaneous resolution of type II endoleak

Binary data on spontaneous resolution of type II endoleak were reported by 5 studies^{14
–16,19,25,27} including a total of 306 patients with type II endoleak (104 receiving anticoagulation and 202 controls). Patients on anticoagulation had lower odds of spontaneous resolution of type II endoleak without statistical significance (OR=0.50, 95% CI=0.23-1.09, p=0.08; heterogeneity: p=0.28, I²=22%).

Reintervention-free survival

Data on reintervention-free survival were reported by 2 studies^18,24 with a total of 704 patients (101 on anticoagulation and 603 controls). No statistically significant difference was found between the groups (HR=1.04, 95% CI=0.62-1.75, p=0.88; heterogeneity: p=0.67, I²=0%).

Sensitivity and subgroup analyses

Sensitivity analyses excluding studies of low methodological quality (<7 stars on the NOS scale) did not affect the direction of effect for any of the outcomes, but the statistical significance was lost for overall survival (HR=2.13, 95% CI=0.87-5.19, p=0.10; heterogeneity: I²=94%, p<0.0001). Exclusion of 1 study at a time did not affect the direction of effect estimate for any of the outcomes, but the statistical significance was lost for overall survival, when the study of Biebl et al¹⁵ was excluded (HR=1.78, 95% CI=0.83-3.78, p=0.14, heterogeneity: I²=92%, p<0.00001), and freedom from reintervention, when the study De Rango et al¹⁹ was excluded (HR=1.31, 95% CI=0.78-2.19, p=0.30; heterogeneity: I²=0%, p=0.80). Two studies included patients with ruptured AAA (5.6% of the total cohort in the study of Flohr et al²⁵ and 12% in the study of Johnson et al¹⁸). Exclusion of these studies in a posthoc sensitivity analysis found no change in the significance or direction of effect estimate. We were not able to do subgroup analysis for the type of anticoagulation treatment due to paucity of data.

Discussion

Meta-analysis of observational data showed that anticoagulation therapy has an important effect on clinical outcomes after standard EVAR. Primary outcomes were overall and aneurysm-related mortality and aneurysm rupture. Individuals receiving anticoagulation treatment at the time of EVAR or during the post-EVAR follow-up were found to have higher overall mortality than those not receiving such treatment. There is a paucity of data on aneurysm-related mortality, most probably due to difficulties with data capture in retrospective study designs, with only 1 study reporting pertinent information with no significant difference between groups. Similarly, limited data were identified for aneurysm rupture, the rate of which was not significantly different between groups, but the 3 studies included in this meta-analysis had a very low event rate, and therefore, the absence of difference may merely be the result of a type II error. Secondary outcomes were endoleak, reintervention, and aneurysm sac expansion. Endoleak was more common in anticoagulated patients, as was persistent type II endoleak, and patients on anticoagulation therapy were less likely to have spontaneous resolution of a type II endoleak, albeit without statistical significance. Patients receiving anticoagulation therapy were more likely to undergo secondary intervention and develop aneurysm sac expansion after the index EVAR.

The reduced survival of anticoagulated patients cannot be fully explained without considering confounding factors. Hazard estimates were calculated from survival curve data, and the exposed and nonexposed individuals were not matched in the design nor confounders were adjusted for in the analysis. Individual patient data would give further insight into potential causal associations of anticoagulation therapy and mortality. It may well be that patients receiving anticoagulation therapy may have more cardiac comorbidities putting them at higher risk of cardiovascular complications and mortality during the post-EVAR follow-up.

Although most studies failed to report detailed information on the type of endoleak, the higher rate of endoleak in anticoagulated patients is likely due to aortic side branches failing to thrombose, resulting in type II endoleak, often persisting despite secondary interventions. Of note, persistent type II endoleaks have been reported to correlate with sac expansion, secondary interventions, surgical conversion, and aneurysm rupture.^30,31 As opposed to early type II endoleaks, which resolve in up to 75% of cases, late endoleaks often persist, with only a third resolving spontaneously and more than half requiring reintervention in case of sac expansion or rupture.³²

Individuals receiving anticoagulation therapy were twice as likely to require secondary intervention. Although lack of data granularity does not allow making inferences about indications and types of reinterventions, we know that such interventions are commonly performed to treat type II endoleaks which cause the aneurysm sac to expand. Such procedures are technically complex, often with suboptimal results, and require a high level of expertise and experience. A systematic review found little evidence demonstrating the efficacy of secondary interventions for such endoleaks.³³ Another systematic review found a technical failure rate as high as 28.5%.³⁴ The reduced overall survival in anticoagulated patients might be due to higher rates of persistent type II endoleak and reintervention, as previously shown.^30,31 Unfortunately, the lack of data on aneurysm-related mortality does not allow reaching conclusions on potential causal associations between anticoagulation therapy and mortality. Similarly, a potential causal association of anticoagulation treatment and AAA rupture post-EVAR has to be investigated. The role of anticoagulation in patients with unrepaired AAA also remains controversial with some reports indicating a positive effect and some others demonstrating no effect.^35,36

With the considerably larger number of studies (16 vs 5) and patients (35 739 vs 1499) than the previous meta-analysis on this topic published in 2014, our work adds to the evidence base with additional important information and implications for clinical practice.³⁵ We conducted additional analyses to those of the previous report, including time-to-event meta-analyses for overall survival, freedom from endoleak, freedom from reintervention, freedom from sac expansion, and endoleak-free survival. We investigated additional outcomes, such as perioperative/early endoleak, persistent type II endoleak, and spontaneous resolution of type II endoleak.

Based on our findings, anticoagulation treatment needs to be considered in decision-making processes. Furthermore, patients need to be informed of the higher failure rates of EVAR compared with their nonanticoagulated counterparts, and intensified surveillance strategies may need to be implemented in this patient cohort. Identifying anticoagulation as a prognostic factor after EVAR has additional clinical and research implications. Risk prediction modeling in EVAR should be based on clinical as well as anatomical and procedural parameters, and anticoagulation therapy should be factored into such models.³⁶ Risk stratification based on a multifactorial approach will inform decision-making, optimize consenting processes, and help develop risk-informed surveillance strategies.^37,38 Anticoagulation should be considered alongside other parameters that have been described as risk factors for developing type II endoleak, such as number and size of patent lumbar arteries, patency, and size of the inferior mesenteric artery. In the presence of such factors, concomitant use of anticoagulants may have an additive effect, putting patients at risk for developing adverse events.³⁹

Our findings should be viewed and interpreted in the context of limitations. Meta-analyses are limited by the retrospective observational study design. The low event rate for several of the investigated outcomes suggests that such events may be under-reported and not representative of real-world practice. Most studies were judged to be of low methodological quality. As a result, the level of evidence was downgraded, increasing the uncertainty about the findings and not allowing safe clinical translation and transfer to clinical practices. The level of evidence was also downgraded due to heterogeneity in patient characteristics and anticoagulation regimes, as well as due to inconsistency in outcome reporting and definition, and was judged to be very low or low for all outcomes. Critical and important outcomes, such as aneurysm-related mortality and aneurysm rupture, were not reported or reported by a limited number of studies, resulting in imprecision and limiting the power of statistical analyses. Paucity of data did not allow subgroup analyses for type of anticoagulation agent; thus, effects of different anticoagulation agents or regimes on post-EVAR outcomes were not investigated. Most studies examined patients on warfarin. With the increasing use of the newer direct oral anticoagulant agents, the study cohorts included in our meta-analysis may not be representative of the overall anticoagulated population, resulting in diminished external validity of our findings. Furthermore, lack of data did not allow subgroup analyses of patients receiving a combination of antiplatelet and anticoagulation therapy vs those on anticoagulation therapy alone.

Conclusions

Anticoagulation may be a poor prognostic factor after standard EVAR. Patients receiving anticoagulation treatment may have worse clinical outcomes than those not receiving such treatment, as indicated by the increased mortality, endoleak, and reintervention rates. Our results have important clinical implications in decision-making and patient consenting. Intensified post-EVAR follow-up should be considered in this patient cohort. Anticoagulation should be included in research on risk prediction building and modeling. The effects of direct oral anticoagulants compared with other anticoagulant agents should also be the focus of future research.

Supplemental Material

sj-docx-1-jet-10.1177_15266028231214761 – Supplemental material for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair

Supplemental material, sj-docx-1-jet-10.1177_15266028231214761 for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair by Nikolaos Kontopodis, Nikolaos Galanakis, Christos V. Ioannou and George A. Antoniou in Journal of Endovascular Therapy

Supplemental Material

sj-docx-2-jet-10.1177_15266028231214761 – Supplemental material for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair

Supplemental material, sj-docx-2-jet-10.1177_15266028231214761 for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair by Nikolaos Kontopodis, Nikolaos Galanakis, Christos V. Ioannou and George A. Antoniou in Journal of Endovascular Therapy

Supplemental Material

sj-docx-3-jet-10.1177_15266028231214761 – Supplemental material for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair

Supplemental material, sj-docx-3-jet-10.1177_15266028231214761 for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair by Nikolaos Kontopodis, Nikolaos Galanakis, Christos V. Ioannou and George A. Antoniou in Journal of Endovascular Therapy

Supplemental Material

sj-docx-4-jet-10.1177_15266028231214761 – Supplemental material for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair

Supplemental material, sj-docx-4-jet-10.1177_15266028231214761 for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair by Nikolaos Kontopodis, Nikolaos Galanakis, Christos V. Ioannou and George A. Antoniou in Journal of Endovascular Therapy

Supplemental Material

sj-docx-5-jet-10.1177_15266028231214761 – Supplemental material for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair

Supplemental material, sj-docx-5-jet-10.1177_15266028231214761 for Systematic Review and Meta-Analysis of the Effect of Anticoagulation on Outcomes After Endovascular Aneurysm Repair by Nikolaos Kontopodis, Nikolaos Galanakis, Christos V. Ioannou and George A. Antoniou in Journal of Endovascular Therapy

Footnotes

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: G.A.A. has received fees from GE Healthcare for providing educational support. The rest of the review authors have no competing interests to disclose or any relationships that could be perceived as conflict of interest for the conduct of this review.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Nikolaos Galanakis

George A. Antoniou

Availability of Data,Code,and Other Material

A list of studies that were excluded from this review with reasons, template data collection forms, data extracted from included studies, data used for all analyses, analytical code, and the PRISMA 2020 checklist can be obtained by the corresponding author upon request. Such data are not publicly available.

Supplemental Material

Supplemental material for this article is available online.

References

Wanhainen

Verzini

Van Herzeele

, et al. Editor’s choice—European Society for Vascular Surgery (ESVS) 2019 clinical practice guidelines on the management of abdominal aorto-iliac artery aneurysms. Eur J Vasc Endovasc Surg. 2019;57(1):8–93.

Antoniou

Torella

Editor’schoice—endovascularvs. Open repair for abdominal aortic aneurysm: systematic review and meta-analysis of updated peri-operative and long term data of randomised controlled trials. Eur J Vasc Endovasc Surg. 2020;59(3):385–397.

Chen

Yeh

Ulloa

, et al. Frailty among veterans undergoing abdominal aortic aneurysm repair. Ann Vasc Surg. 2023;92:18–23.

Hutchens

Hung

Briffa

, et al. Antithrombotic therapy in atrial fibrillation management in Western Australia: temporal trends and evidence-treatment gaps. Heart Lung Circ. 2021;30(7):955–962.

Kotalczyk

Gue

Potpara

, et al. Current trends in the use of anticoagulant pharmacotherapy in the United Kingdom are changes on the horizon. Expert Opin Pharmacother. 2021;22(8):1061–1070.

Higgins

JPT

Thomas

Chandler

, et al. Cochrane Handbook for Systematic Reviews of Interventions (Version 6.0). London, England: Cochrane; 2019.

Antoniou

Mani

Enhancing the reporting of systematic reviews and meta-analyses in vascular surgery: PRISMA 2020. Eur J Vasc Endovasc Surg. 2021;62(4):664–666.

The Ottawa Hospital Research Institute. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Date unknown. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed April 1, 2023.

Antoniou

GA.

The GRADE approach to appraising the evidence or how to increase the credibility of your research. Am J Surg. 2020;220(2):290–293.

10.

Tierney

Stewart

Ghersi

, et al. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials. 2007;8:16.

11.

Antoniou

Smith

CT.

A guide on meta-analysis of time-to-event outcomes using aggregate data in vascular and endovascular surgery. J Vasc Surg. 2020;71(3):1002–1005.

12.

DerSimonian

Laird

Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188.

13.

Sterne

JAC

Sutton

Ioannidis

JPA

, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ. 2011;343:d4002.

14.

Fairman

Carpenter

Baum

, et al. Potential impact of therapeutic warfarin treatment on type II endoleaks and sac shrinkage rates on midterm follow-up examination. J Vasc Surg. 2002;35(4):679–685.

15.

Biebl

Hakaim

Oldenburg

, et al. Does chronic oral anticoagulation with warfarin affect durability of endovascular aortic aneurysm exclusion in a midterm follow-up? J Endovasc Ther. 2005;12(1):58–65.

16.

Bobadilla

Hoch

Leverson

, et al. The effect of warfarin therapy on endoleak development after endovascular aneurysm repair (EVAR) of the abdominal aorta. J Vasc Surg. 2010;52(2):267–271.

17.

Abularrage

Crawford

Conrad

, et al. Preoperative variables predict persistent type 2 endoleak after endovascular aneurysm repair. J Vasc Surg. 2010;52(1):19–24.

18.

Johnson

Chiang

Eldrup-Jorgensen

, et al. Effect of chronic oral anticoagulation with warfarin on the durability and outcomes of endovascular aortic aneurysm repair. J Vasc Surg. 2013;58(2):319–323.

19.

De Rango

Verzini

Parlani

, et al. Safety of chronic anticoagulation therapy after endovascular abdominal aneurysm repair (EVAR). Eur J Vasc Endovasc Surg. 2014;47(3):296–303.

20.

Wild

Dattani

Stather

, et al. Effect of anticoagulation and antiplatelet therapy on incidence of endoleaks and sac size expansions after endovascular aneurysm repair. Ann Vasc Surg. 2014;28(3):554–559.

21.

Kumar

Cowled

Boult

, et al. Type II endoleak after endovascular aneurysm repair: natural history and treatment outcomes. Ann Vasc Surg. 2017;44:94–102.

22.

Álvarez Marcos

Llaneza Coto

Franco Meijide

, et al. Effect of antiplatelet therapy on aneurysmal sac expansion associated with type II endoleaks after endovascular aneurysm repair. J Vasc Surg. 2017;66(2):396–403.

23.

Seike

Tanaka

Fukuda

, et al. Influence of warfarin therapy on the occurrence of postoperative endoleaks and aneurysm sac enlargement after endovascular abdominal aortic aneurysm repair. Interact Cardiovasc Thorac Surg. 2017;24:615–618.

24.

Castelli

Deluca

Guillermo

, et al. Oral anticoagulation in aortic endovascular reinterventions. Rev Argent Cardiol. 2019;87:15–18.

25.

Flohr

Snow

Aziz

The fate of endoleaks after endovascular aneurysm repair and the impact of oral anticoagulation on their persistence. J Vasc Surg. 2021;74(4):1183–1192.e5.

26.

D’Oria

Di Girolamo

Calvagna

, et al. Remodeling of abdominal aortic aneurysm sac following endovascular aortic repair: association with clinical, surgical, and genetic factors. Cardiovasc Pathol. 2022;58:107405.

27.

Kong

Balceniuk

Mix

, et al. Long-term anticoagulation is associated with type II endoleaks and failure of sac regression after endovascular aneurysm repair. J Vasc Surg. 2022;76(2):437–444.e2.

28.

Morisaki

Matsubara

Furuyama

, et al. Effects of antithrombotic therapy on abdominal aortic aneurysm sac size after endovascular repair in patients with favorable neck anatomy. J Vasc Interv Radiol. 2022;33(2):113–119.

29.

Lal

Zhou

, et al; OVER Veterans Affairs Cooperative Study Group. Predictors and outcomes of endoleaks in the veterans affairs open versus endovascular repair (OVER) trial of abdominal aortic aneurysms. J Vasc Surg. 2015;62:1394–1404.

30.

Buck

Herrmann

, et al; Vascular Study Group of New England.Risk factors and consequences of persistent type II endoleaks. J Vasc Surg. 2016;63:895–901.

31.

Jones

Atkins

Brewster

, et al. Persistent type 2 endoleak after endovascular repair of abdominal aortic aneurysm is associated with adverse late outcomes. J Vasc Surg. 2007;46(1):1–8.

32.

Pineda

Calligaro

Tyagi

, et al. Late type II endoleaks after endovascular aneurysm repair require intervention more frequently than early type II endoleaks. J Vasc Surg. 2018;67(2):449–452.

33.

Ultee

KHJ

Büttner

Huurman

, et al. Editor’s choice—systematic review and meta-analysis of the outcome of treatment for type II endoleak following endovascular aneurysm repair. Eur J Vasc Endovasc Surg. 2018;56(6):794–807.

34.

Sidloff

Stather

Choke

, et al. Type II endoleak after endovascular aneurysm repair. Br J Surg. 2013;100:1262–1270.

35.

Lazarides

Georgiadis

Charalampidis

, et al. Impact of long-term warfarin treatment on EVAR durability: a meta-analysis. J Endovasc Ther. 2014;21(1):148–153.

36.

Antoniou

GA.

Risk stratification and risk specific surveillance for endovascular aneurysm repair. Eur J Vasc Endovasc Surg. 2020;60:950–951.

37.

Ul-Mulk

Antoniou

GA.

Prognostic prediction models for endovascular abdominal aortic aneurysm repair: protocol for a scoping review. BMJ Open. 2022;12:e061420.

38.

Antoniou

Schermerhorn

Forbes

, et al. Risk factors, risk stratification and risk-specific surveillance strategies after endovascular aneurysm repair: study protocol for a Delphi study by the International RIsk Stratification in EVAR (IRIS-EVAR) working group. BMJ Open. 2022;12:e055803.

39.

Lalys

Durrmann

Duménil

, et al. Systematic review and meta-analysis of preoperative risk factors of type II endoleaks after endovascular aneurysm repair. Ann Vasc Surg. 2017;41:284–293.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB

0.01 MB

0.02 MB

0.01 MB