From the Society for Vascular Surgery
`
`Late graft explants in endovascular aneurysm repair
`
`Eric J. Turney, MD, Sean P. Steenberge, MS, Sean P. Lyden, MD, Matthew J. Eagleton, MD,
`Sunita D. Srivastava, MD, Timur P. Sarac, MD, Rebecca L. Kelso, MD, and Daniel G. Clair, MD,
`Cleveland, Ohio
`
`Objective: With more than a decade of use of endovascular aneurysm repair (EVAR), we expect to see a rise in the number
`of failing endografts. We review a single-center experience with EVAR explants to identify patterns of presentation and
`understand operative outcomes that may alter clinical management.
`Methods: A retrospective analysis of EVARs requiring late explants, >1 month after implant, was performed. Patient
`demographics, type of graft, duration of implant, reason for removal, operative technique, length of stay, complications,
`and in-hospital and late mortality were reviewed.
`Results: During 1999 to 2012, 100 patients (91% men) required EVAR explant, of which 61 were placed at another
`institution. The average age was 75 years (range, 50-93 years). The median length of time since implantation was
`41 months (range, 1-144 months). Explanted grafts included 25 AneuRx (Medtronic, Minneapolis, Minn), 25 Excluder
`(W. L. Gore & Associates, Flagstaff, Ariz), 17 Zenith (Cook Medical, Bloomington, Ind), 15 Talent (Medtronic), 10
`Ancure (Guidant, Indianapolis, Ind), 4 Powerlink (Endologix, Irvine, Calif), 1 Endurant (Medtronic), 1 Quantum LP
`(Cordis, Miami Lakes, Fla), 1 Aorta Uni Iliac Rupture Graft (Cook Medical, Bloomington, Ind), and 1 homemade tube
`graft. Overall 30-day mortality was 17%, with an elective case mortality of 9.9%, nonelective case mortality of 37%, and
`56% mortality for ruptures. Endoleak was the most common indication for explant, with one or more endoleaks present in
`82% (type I, 40%; II, 30%; III, 22%; endotension, 6%; multiple, 16%). Other reasons for explant included infection (13%),
`acute thrombosis (4%), and claudication (1%). In the first 12 months, 23 patients required explants, with type I endoleak
`(48%) and infection (35%) the most frequent indication. Conversely, 22 patients required explants after 5 years, with type
`I (36%) and type III (32%) endoleak responsible for most indications.
`Conclusions: The rate of EVAR late explants has increased during the past decade at our institution. Survival is higher
`when the explant is done electively compared with emergent repair. Difficulty in obtaining a seal at the initial EVAR often
`leads to failure #1 year, whereas progression of aneurysmal disease is the primary reason for failure >5 years. (J Vasc Surg
`2014;59:886-93.)
`
`Endovascular aneurysm repair (EVAR) has increased in
`use during the past decade and now has become the
`primary mode of treatment of infrarenal abdominal aortic
`aneurysms (AAAs) in private and academic practices.1 In
`addition, clinicians have continued to push the limits of
`the application of the technology, attempting to address
`more challenging anatomic conditions outside the instruc-
`tions for use (IFU).2,3 The long-term success of EVAR
`requires surveillance imaging to identify problems that
`develop, most of which can usually be solved with
`
`From the Department of Vascular Surgery, Cleveland Clinic Lerner College
`of Medicine at Case Western Reserve University.
`Author conflict of interest: S.L. is a Consultant for Medtronic, Cook, and
`Covidien; a Board Member of VIVA Physicians; and an advisory board
`member of Covidien, Medtronic, Trivascular, and Ariste Medical. ME
`is a Consultant for Cook and Bolton Medical. TS is the founder of Peri-
`tec; DC is a Consultant for Cordis, Covidien, Endologix, Vessix, and
`Gore; and is on the Advisory Board for Medtronic and Boston Scientific.
`Presented as an abstract at the 2013 Vascular Annual Meeting of the Society
`for Vascular Surgery, San Francisco, Calif, May 30-June 1, 2013.
`Reprint requests: Sean P. Lyden, MD, Department of Vascular Surgery,
`Cleveland Clinic Lerner College of Medicine at Case Western Reserve
`University, Cleveland Clinic Foundation, Desk H32, 9500 Euclid Ave,
`Cleveland, OH 44195 (e-mail: lydens@ccf.org).
`The editors and reviewers of this article have no relevant financial relationships
`to disclose per the JVS policy that requires reviewers to decline review of any
`manuscript for which they may have a conflict of interest.
`0741-5214/$36.00
`Copyright Ó 2014 by the Society for Vascular Surgery.
`http://dx.doi.org/10.1016/j.jvs.2013.10.079
`
`886
`
`secondary endovascular interventions. With the increasing
`number of patients treated, the number presenting without
`further options for endovascular salvage will likely produce
`an increasing population needing open conversion.
`A large review has found that the overall incidence of
`late conversion is 1.9% (range, 0.4%-22%)4 and that
`delayed rupture risk is 1% per year.5 EVAR can fail for
`many reasons, including graft infection, rupture, migration,
`component
`separation,
`thrombosis, device fracture or
`failure, and aneurysm growth with or without endoleak.
`Other authors have documented the frequency of endovas-
`cular option use before open conversion.6,7 We reviewed
`our institution’s growing experience for late conversion
`of failed EVAR to better understand current results, predic-
`tors of outcome, and reasons for failure.
`
`METHODS
`Patients were identified according to the Current
`Procedural Terminology codes (American Medical Associ-
`ation, Chicago, Ill) for open AAA repair, and International
`Classification of Diseases, 9th revision coding for AAA
`at the Cleveland Clinic Foundation between January 1,
`1999, and December 31, 2012. We reviewed all operative
`reports to identify EVAR explants. After Institutional
`Review Board approval, a Research Electronic Data
`Capture database was generated. Patients were only
`included if the endograft was removed >30 days after
`implantation. We also included patients who underwent
`
`TMT 2113
`Medtronic v. TMT
`IPR2021-01532
`
`

`

`JOURNAL OF VASCULAR SURGERY
`Volume 59, Number 4
`
`Turney et al 887
`
`partial endograft removal. Patients who underwent open
`surgical attempts to repair endoleaks or increase graft fixa-
`tion without graft were excluded.
`Patient demographics, EVAR type, operative factors,
`and complications were recorded. Demographic informa-
`tion included sex, duration of implant, age, and comorbid-
`ities such as coronary artery disease, chronic obstructive
`pulmonary disease, and chronic renal failure (glomerular
`filtration rate <60 mL/min/1.73 m2). Graft information
`included reason for explant, partial or full excision, facility
`where implanted, and need for elective, urgent, or emer-
`gent explant. Reasons for explant included endoleak by
`type,
`infection, migration, aneurysm enlargement, and
`rupture. Society for Vascular Surgery reporting standards
`were used to define migration, aneurysm enlargement,
`and endoleak type. Operative factors included repair type,
`surgical approach, clamp position, duration of clamp, and
`creatinine levels preoperatively and postoperatively and
`the peak level during the hospital stay.
`We identified patients from our previous analysis in
`20098 and compared outcomes for patients who under-
`went repair after that point. We compared modes of failure
`for conversions #1 year, between 1 and 5 years, and
`>5 years.
`Outcomes analyzed were 30-day mortality, arrhythmia,
`myocardial
`infarction, pneumonia, prolonged ventilation,
`temporary hemodialysis, permanent hemodialysis, length of
`stay, and long-term survival status. Statistical analysis was per-
`formed using R software (The R Foundation for Statistical
`Computing, http://www.r-project.org/foundation/) and
`Excel (Microsoft Corp, Redmond, Wash). Analysis of associ-
`ations with 30-day mortality was performed using Pearson c2
`tests and Fisher exact tests for categoric factors and Wilcoxon
`rank sum tests for continuous measures. Kaplan-Meier esti-
`mates and Cox proportional hazard models were used to
`compare long-term survival. Owing to the small number
`of 30-day and long-term mortality events, multivariable
`modeling was not performed for either outcome. Statistical
`significance was identified for a P value <.05.
`
`RESULTS
`From 1999 to 2012, our institution performed 1881
`EVARs and 2075 open infrarenal aortic repairs. During
`this interval, 100 patients required late graft explant, of
`which 61 were placed at another institution. Six acute graft
`explants were excluded from analysis. Multiple types of
`EVAR devices have been explanted (Table I). We have
`noted an increased frequency of EVAR explant, performing
`28 in the last 2 years (Fig 1). Suprarenal fixation was
`present in 37% of stent grafts removed.
`Cohort characteristics. Patients were an average age
`of 75 years (range, 50-93 years) and 91 were male. Coro-
`nary artery disease was present in 66%, chronic obstructive
`pulmonary disease in 23%, and chronic kidney disease in
`24%, with two or more present in 28%. The average preop-
`erative serum creatinine was 1.3 mg/dL (range, 0.5-
`4.68 mg/dL). We did not find a difference in patient
`demographics compared with the patient population
`
`Table I. Graft type
`
`Graft type
`
`AneuRxa
`Excluderb
`Zenithc
`Talenta
`Ancured
`Powerlinke
`Enduranta
`Quantum LPf
`Aorta Uni Iliac Rupture Graftg
`Homemade
`
`aMedtronic, Minneapolis, Minn.
`bW. L. Gore & Associates, Flagstaff, Ariz.
`cCook Medical, Bloomington, Ind.
`dGuidant, Indianapolis, Ind.
`eEndologix, Irvine, Calif.
`fCordis, Miami Lakes, Fla.
`gCook Medical, Bloomington, Ind.
`
`No.
`
`25
`25
`16
`15
`10
`4
`1
`1
`2
`1
`
`reported in 2009. The only difference we identified in
`the EVARs explanted since our original report is that we
`have converted a higher proportion of patients whose orig-
`inal EVAR was done at another institution, 15 patients
`(38%) and 46 patients (77%), respectively (P < .001).
`Grafts were excised after a median of 41 months
`(range, 1-144 months), with 71 explants performed elec-
`tively. Overall, endoleak was present in 82% of patients,
`some with multiple types. Type I endoleak was the most
`frequently occurring, present in 40% of total patients, fol-
`lowed by type II in 30%, and type III in 22%. Finally, endo-
`tension was present
`in 6%. Also contributing to the
`indications for explantation were graft migration in 15
`patients and claudication in one. The claudication patient
`who required explant was before 2009 and is detailed in
`our prior report.8
`Limb thrombosis in six patients contributed to the
`indication for explant. One had claudication and one had
`bilateral limb thrombosis with acute limb ischemia, and
`both were previously described.8 Two patients had prox-
`imal migration causing limb occlusion leading to the choice
`to explant. One had multiple limb occlusions treated
`several times with thrombolysis before deciding to remove
`the device. The last patient with limb thrombosis had prox-
`imal migration and developed a symptomatic aneurysm.
`Urgent or emergent repairs were done in 29 patients.
`Ten patients were treated urgently with seven staged
`reconstructions for infected stent grafts. Repairs were
`emergent in 19 patients, including infection in 3, rupture
`in 9, acute limb ischemia in 2, and a symptomatic nonrup-
`tured aneurysm in 5. The five patients with symptomatic
`aneurysm treated emergently had endoleaks contributing
`to the problem, comprising type I in 2, type II in 2, type
`III in 1, and type V in 1. One patient had only type II
`endoleak and a symptomatic expanding aneurysm.
`Two patients were originally treated at Cleveland
`Clinic for ruptured aneurysm in a physician-sponsored
`investigational device exemption trial with an aortouniiliac
`
`

`

`888 Turney et al
`
`JOURNAL OF VASCULAR SURGERY
`April 2014
`
`Fig 1. Number of explants per year.
`
`Fig 2. Time to explant.
`
`two-piece device. Both patients were explanted for recur-
`rent rupture.
`Infected grafts were treated with several methods. One
`patient was treated with excision and femoral vein recon-
`struction, eight were treated with axillobifemoral bypass
`and removal (one simultaneous), and four were treated
`with excision and replacement with rifampin-soaked poly-
`ester grafts.
`Twenty-three patients required explant <12 months
`after EVAR, with an additional 22 performed $5 years
`after placement (Fig 2). The primary indication for graft
`removal by time period is detailed in Table II.
`Operative characteristics. The operative approach
`was determined by surgeon preference and clinical factors
`such as the presence of suprarenal fixation, extent of prox-
`imal and distal repair required, and the reason for failure.
`Our overall operative approach has not changed since our
`prior report. The approach was retroperitoneal in 56% of
`patients and transabdominal in 44%. Supraceliac clamping
`was required in 54 patients, suprarenal in 33, and infrarenal
`
`in 13, with an average clamp time of 35 minutes. Clamps
`were generally placed above the entire endograft, with
`few exceptions, when distal issues led to the failure and
`proximal fixation was not compromised. Aortobiiliac repair
`was performed in 65 patients and was the most common
`type of reconstruction required. This was followed by
`tube graft in 18 patients, aortobifemora1 in 8, axillobife-
`moral in 7, aortouniiliac in 1, and no repair in 1. Nine
`patients required revascularization of visceral or renal
`vessels. Stent grafts were partially excised in 22 patients,
`with 17 of those leaving distal limb components and four
`proximal components and one patient with both.
`Average length of stay was 15 days (range, 1-56 days),
`with in hospital complications including arrhythmia in 25
`patients, myocardial infarction in 6, prolonged ventilator
`support in 26, pneumonia in 7, and hemodialysis in 9, of
`which one was temporary. Complications rates by time
`cohort are detailed in Table III and did not change. Renal
`failure requiring dialysis developed in six patients, all of
`whom had a supraceliac clamp, with an average clamp
`
`

`

`JOURNAL OF VASCULAR SURGERY
`Volume 59, Number 4
`
`Turney et al 889
`
`Table II. Indication for explants by duration (patients
`may have more than one)
`
`Table III. Outcomes
`
`Reason for explant
`
`Type I endoleak
`Initial seal
`Migration
`Disease progression
`Type II
`Type III
`Type V
`Infection
`Occlusion
`Claudication
`Unknown
`Mortality
`
`1 year,
`No. (%)
`
`11 (38)
`7
`3
`1
`1(3)
`1(3)
`0
`8 (28)
`3 (3)
`3 (10)
`0
`4 (14)
`
`1-5 years,
`No. (%)
`
`$5 years,
`No. (%)
`
`19 (35)
`5
`4
`10
`11 (20)
`13 (24)
`4 (7)
`4 (7)
`1 (2)
`0
`1 (2)a
`11 (20)
`
`8 (36)
`1
`1
`6
`2 (9)
`7 (32)
`2 (9)
`1 (5)
`2 (9)
`0
`0
`2 (9)
`
`aPatient presented with free rupture, and the reason for rupture was not
`identified.
`
`time of 61 minutes compared with the overall average of
`35 minutes. Overall 30-day mortality was 17%, with 9.9%
`mortality in elective cases and 34% in nonelective cases.
`Mortality rate by indication is described in Table IV.
`Univariate predictors of 30-day mortality. The 30-
`day mortality was predicted by multiple factors that
`reached statistical significance. Renal or visceral revascu-
`larization was required in nine of the 100 patients, with
`a 30-day mortality of 44% (P ¼ .036). Conversely, the
`mortality rate was 13% in those who did not require renal
`or visceral revascularization. The second factor to reach
`statistical significance was clamp position in the supra-
`visceral position, which was required in 54 patients, with
`a mortality rate of 24% (P ¼ .001). Suprarenal clamping
`was required in 33 patients, with no deaths, and infrarenal
`clamping was performed in 13 patients, but the associated
`30% mortality was not statistically significant. The pres-
`ence of infection, which occurred in 13 patients, also
`reached statistical significance with respect to mortality at
`38% (P ¼ .038). The 19 patients who were repaired
`emergently and the nine who presented with rupture had
`a statistically significant worse 30-day mortality of 37%
`(P ¼ .015) and 55% (P ¼ .006), respectively (Table V).
`The 71 patients who underwent elective procedures had
`a statistically significant lower 30-day mortality at 9.9%
`(P ¼ .006).
`Finally, the 30-day mortality can be predicted by peak
`creatinine and postoperative creatinine levels, which were
`statistically significantly elevated (2.5 mg/dL; P < .001;
`1.9, mg/dL; P < .001) compared with those who did
`not die #30 days (1.4 and 1.06 mg/dL, respectively). A
`number of
`factors did not predict 30-day mortality,
`including age, sex, presence of coronary artery disease,
`chronic obstructive pulmonary disease, chronic renal
`failure, presence of suprarenal fixation, operative approach,
`clamp time, partial excision, and aneurysm enlargement.
`Type II endoleaks were associated with statistically signifi-
`cant improvement in 30-day mortality, present in 30% of
`explanted patients and a mortality of 3% (P ¼ .025).
`
`Variable
`
`Total
`
`30-day mortality
`Complication
`Cardiac
`Pulmonary
`Renal
`Any
`
`17
`
`29
`26
`9
`44
`
`Initial cohort,
`No. (%)
`
`New cohort,
`No. (%)
`
`9 (23)
`
`8 (13)
`
`13 (33)
`10 (25)
`5 (12.5)
`19 (47.5)
`
`16 (27)
`16 (27)
`4 (6.67)
`25 (41.67)
`
`P
`
`.36
`
`.69
`.99
`.48
`.71
`
`Table IV. Mortality rates by indication
`
`Variable
`
`Total, No.
`
`Mortality, No. (%)
`
`Elective
`Nonelective
`Emergent
`Infected grafta
`Aortoenteric fistulaa
`Rupture
`
`71
`29
`19
`13
`5
`9
`
`7 (9.90)
`10 (34)
`7 (37)
`5 (38)
`3 (60)
`5 (56)
`
`aSome of the infected grafts and patients with aortoenteric fistula were done
`urgently with a staged reconstruction, whereas others were unstable and
`done emergently at single setting.
`
`Univariate predictors of long-term survival. Overall
`survival after open conversion was evaluated with a Cox
`proportional hazard model and with Kaplan-Meier esti-
`mates. This analysis demonstrated that the presence of
`chronic obstructive pulmonary disease, visceral revascular-
`ization, infection, emergent repair, rupture, and elevated
`creatinine conveyed an increased risk for death. Conversely
`an elective repair, presence of
`type II endoleak, and
`aneurysm enlargement predicted a decrease risk of death
`(Table VI).
`A comparison with patients from our previous series in
`2009 found no statistically significant difference in demo-
`graphics, procedural details, or outcomes. The indications
`for explant in the initial cohort included type I endoleaks
`in 16 (40%), type II in 8 (20%), type III in 11 (28%),
`and infection in 4 (10%). Indications in the new cohort
`were type I endoleak in 24 (40%), type II in 22 (37%),
`type III in 11 (18%), and infection in 9 (15%). There
`was no statistically significant difference in these factors
`as well.
`
`DISCUSSION
`The need for open conversion of failing endografts has
`accelerated during the last several years at our institution.
`This is most likely due to the growing number of patients
`having had EVAR, because this is the most common way
`AAAs are treated in the United States. The initial repair
`in most of our patients (n ¼ 61) was done at another
`facility, and many of the patients underwent attempted
`endovascular salvage before open conversion. The details
`of those interventions before referral for conversion were
`unavailable for this analysis. Pre-EVAR implant imaging
`was not available for most patients, making it difficult to
`
`

`

`890 Turney et al
`
`JOURNAL OF VASCULAR SURGERY
`April 2014
`
`Table V. Predictors of 30-day mortality
`
`Table VI. Factors predictive of overall survival
`
`Survival
`
`Predictor
`
`Variable
`
`With variable
`(n ¼ 17)
`
`Without variable
`(n ¼ 83)
`
`Median age (range), years
`Male sex
`Coronary artery disease
`COPD
`Chronic renal failure
`Suprarenal fixation
`Visceral revascularization
`Supraceliac clamp
`Emergent
`Infection
`Rupture
`
`77 (55-85)
`78
`54
`19
`20
`32
`2
`41
`12
`8
`4
`
`75 (50-93)
`5
`29
`64
`63
`51
`81
`42
`71
`75
`79
`
`P
`
`.58
`.67
`.99
`.99
`.99
`.67
`.036
`.001
`.015
`.038
`.006
`
`Negative predictors
`COPD
`Visceral revascularization
`Infection
`Emergent
`Rupture
`Creatinine
`Preoperative
`Postoperative
`Peak
`Positive predictors
`Type II endoleak
`Elective operation
`Aneurysm enlargement
`
`HR
`
`2.19
`4.02
`3.27
`4.07
`5.44
`
`2.1
`2.02
`1.37
`
`0.36
`0.2
`0.37
`
`P
`
`.038
`.003
`.002
`<.001
`<.001
`
`.004
`<.001
`<.001
`
`.035
`<.001
`.005
`
`COPD, Chronic obstructive pulmonary disease.
`
`COPD, Chronic obstructive pulmonary disease; HR, hazard ratio.
`
`ascertain the proximal neck morphology. We did not have
`difficulty assessing angulation, neck diameters, and neck
`features after implant, but the effect of the device on these
`measurements may lead to potential errors in assumption
`of the reason for failure.
`Conversion was elective in most patients in our series,
`with endoleak and aneurysm enlargement as the primary
`indication. Compared with our previous series, the rate of
`device migration was
`lower, perhaps a reflection of
`improved device technology. Some data suggest that aneu-
`rysm growth of >8 mm, or absolute size >7 cm, defines an
`increased risk for rupture, although rupture can still occur
`in stable or shrinking aneurysms.9
`It is accepted type I and III endoleaks should be
`aggressively treated, but the threshold for explantation
`with type II or no endoleaks visualized remains to be estab-
`lished.4,10-12 We have used >5 mm growth without an
`identifiable and treatable cause as indication for explant.
`We found varied indications for conversion relative to
`the time period they presented after EVAR. The most
`common indication in patients presenting #1 year of
`EVAR for explant was type Ia endoleak or graft infection.
`The type Ia endoleaks in these patients were typically in
`disadvantaged necks that were short, angulated, or conical
`and typically outside of most device IFU. The willingness
`to push the envelope with endograft placement in chal-
`lenging anatomy and less than ideal adherence to the
`device IFU, as touted by many authors, may also play an
`important role in earlier failures.2,13-15 The increased use
`of EVAR for ruptured aneurysms, when durability of repair
`is often a secondary consideration, has also been a factor
`contributing to an increased need for late conversion in
`other series.16,17 We have seen a reduction of conversions
`in our own patients. Although we are not sure of the
`reason, we have limited our use of commercial devices
`outside of the IFU. This approach has been influenced
`by the increased referral for EVAR failure and the avail-
`ability of branch and fenestrated devices to treat this
`anatomy in a physician-sponsored investigational device
`exemption trial at our institution. This contrasts with other
`
`recent series where most conversions were implanted at the
`same institution.7
`Patients who required explant >5 years most commonly
`presented with type Ia or type III endoleaks. The type Ia
`endoleaks were typically from aneurysmal degeneration of
`the proximal sealing zone, and the type III endoleaks
`were typically from material failure. The average time to
`open conversion is variable across case series, but compared
`with our previous experience, there has been an overall
`increase, with six patients being more than a decade since
`EVAR. The increased type III endoleaks in the devices pre-
`senting for conversion >5 years as well as the failure of
`a patient >10 years after the initial implant underscores
`the fact that surveillance should not have an end point.
`Most patients in our series who required explant had an
`endoleak, with type II endoleaks present in 30%. A type II
`endoleak was never the cause of aneurysm rupture in
`a patient in our series, although this phenomenon has
`been reported.18 In one patient with a symptomatic
`enlarging aneurysm, we only found a type II endoleak.
`We and others have previously noted that the continuous
`pursuit of endoleaks with multiple secondary interventions
`reaches a point of futility where definitive open repair
`becomes prudent.8,19,20 We only recommend treatment
`of type II endoleaks with conversion when endovascular
`methods of treatment fail and aneurysm sac enlargement
`continues.19 In several patients in our series, type II endo-
`leaks were present with type I and III endoleaks, which
`were the primary indication for conversion.
`Emergent repair was most frequently required for
`patients with infection or rupture. Morbidity and mortality
`rates were similar to several recent studies, and our data
`echo the significantly better outcomes when performed
`electively.21 Nonelective repair
`resulted in an overall
`mortality of 37%, and 55% specifically for rupture.7,11 Graft
`fixation can have a direct effect on clamp position and time
`as well as the potential need for aortic endarterectomy. We
`frequently choose to clamp above the entire device. A supra-
`celiac clamp position had a negative effect on outcomes,
`likely due to the effect of visceral and renal ischemia. We,
`
`

`

`JOURNAL OF VASCULAR SURGERY
`Volume 59, Number 4
`
`Turney et al 891
`
`interestingly, did not see an influence of suprarenal fixation
`on outcome, suggesting that this can be done safely in most
`patients even though suprarenal fixation devices are
`commonly covered with neointima and this is disrupted
`and removed by aortic endarterectomy in many instances.
`We would like to emphasize that well-incorporated
`suprarenal stents may be difficult to remove, and consider-
`ation should be to leave those in place. Several new devices
`are designed with more substantial active suprarenal fixa-
`tion, among them the Endurant (Medtronic, Santa Rosa,
`Calif) and Ovation Prime (Trivascular, Santa Rosa, Calif),
`compared with prior devices. These new-generation
`devices may lead to worse outcomes during conversion if
`longer clamp times are needed or damage occurs to the
`visceral segment. The clamping of devices is possible, and
`partial removal may be an important way to reduce compli-
`cations and mortality when removing endografts.8,22 We
`only left the proximal portion of the device in five patients
`and have only done so when well incorporated and clearly
`not associated with the failure mode of the device.
`We found no significant outcome differences in our data
`in this report compared with our initial report of 41 patients
`in 2009.8 Although not significant, we did note a higher
`mortality in elective cases in the current series (3.3% vs
`9.9%). When patient and procedural details were evaluated,
`no significant difference between the two groups could be
`found that might explain the reason for this finding.
`Clearly, with increased experience in conversion, we
`are more willing to offer this operation to patients without
`alternatives. We have the ability to offer custom fenestrated
`devices as a solution for failing EVAR devices; thus, not all
`patients with failed EVAR have open conversion at our
`institution. Some anatomic and device factors prevent
`treatment with custom fenestrated devices. A few of our
`patients without fenestrated or branch options who were
`considered a prohibitive risk for open conversion were
`not offered explant surgery. The disparity in outcomes after
`emergent explantation, specifically for rupture, highlights
`the importance of continued long-term surveillance as
`well as when to proceed with open repair.
`Multiple endoleaks are often encountered during graft
`removal, with type I often the most common indication
`and often implicated in rupture.16,21 Infectious etiologies,
`with an incidence of 0.62%, represent a significant popula-
`tion of patients who require graft removal and often result
`in poorer outcomes.23-25 Patients can present with a spec-
`trum of pathology, from smoldering low-grade infection to
`florid sepsis or exsanguination and, consequently, variable
`urgency for graft removal. For patients who are unfit for
`open repair, some series have suggested percutaneous
`drainage and suppressive antibiotics as a short-term option,
`but late failure has occurred.25-27 We successfully used
`percutaneous drainage with antibiotic
`instillation as
`a temporizing measure to control sepsis in one patient in
`this series. This allowed the patient to be converted
`without dying when better medically optimized.
`The decision to remove all or part of the endograft
`remains
`largely
`contested.
`Potential
`for
`delayed
`
`complications from the remnant endograft, such as migra-
`tion and endoleak remains, particularly if all or a large frag-
`ment of the graft remains. Owing to these risks, continued
`surveillance imaging should be performed, which may not
`be met with great patient compliance. However, well-
`incorporated distal and proximal components can be diffi-
`cult to remove, often leading to longer clamp times and
`more significant injury to the vessel wall. In addition,
`surrounding venous structures can be damaged when there
`is a large amount of inflammation. Some recent series have
`suggested decreased morbidity and mortality when the
`entire endograft is left intact, although bleeding complica-
`tions were present postoperatively.22,28
`The long-term risk for recurrence has not been estab-
`lished, but one series noted freedom from secondary endo-
`leaks at an average of 42 months of follow-up, whereas
`another noted the need for further procedures after partial
`removal.7,22 We still believe that optimally all graft material
`should be removed if safe and technically feasible and
`definitive repair is performed. When endograft explant is
`performed for infection, all of the graft should be removed
`to minimize persistent infection, and the best configuration
`of reconstruction remains undetermined.24,29 The follow-
`up for patients with remnant endograft is a fusion of the
`surveillance plan we use for endograft and open repairs,
`with timing in part tailored to the patient.
`
`CONCLUSIONS
`The incidence of endograft failure continues to rise,
`and the need for open conversion for salvage will grow
`more prominent. Explant results are improved when per-
`formed electively and are associated with mortality rates
`that may exceed those found for elective primary repair.
`Conversion for ruptured EVAR continues to have a very
`high risk of death and complications. Type I and III endo-
`leaks are most often associated with late rupture, and
`surveillance is necessary to identify these issues early.
`Endograft failure #1 year is most commonly due to failure
`of the proximal seal at the initial procedure and under-
`scores the need to understand device limitations when
`used outside the IFU. Late failures are most commonly
`due to aneurysmal degeneration of the seal zones and
`device material failure and can occur even after a decade
`of successful EVAR repair.
`
`AUTHOR CONTRIBUTIONS
`Conception and design: ET, SPS, SL
`Analysis and interpretation: ET, SPS, SL
`Data collection: ET, SPS
`Writing the article: ET, SS, SL
`Critical revision of the article: ET, SL, ME, SDS, TS, RK,
`DC
`Final approval of the article: ET, SPS, SL, ME, SDS, TS,
`RK, DC
`Statistical analysis: SL, ET, SPS
`Obtained funding: SL
`Overall responsibility: SL
`
`

`

`892 Turney et al
`
`REFERENCES
`
`1. Traul D, Street D, Faught W, Eaton M, Castillo J, Brawner J, et al.
`Endoluminal stent-graft placement for repair of abdominal aortic
`aneurysms in the community setting. J Endovasc Ther 2008;15:
`688-94.
`2. Schanzer A, Greenberg RK, Hevelone N, Robinson WP, Eslami MH,
`Goldberg RJ, et al. Predictors of abdominal aortic aneurysm sac
`enlargement after endovascular repair. Circulation 2011;123:2848-55.
`3. Abbruzzese TA, Kwolek CJ, Brewster DC, Chung TK, Kang J,
`Conrad MF, et al. Outcomes following endovascular abdominal aortic
`aneurysm repair (EVAR): an anatomic and device-specific analysis.
`J Vasc Surg 2008;48:19-28.
`I, Mylonas S, Giannakopoulos TG,
`4. Moulakakis KG, Dalainas
`Avgerinos ED, Liapis CD. Conversion to open repair after endograft-
`ing for abdominal aortic aneurysm: a review of causes,
`incidence,
`results, and surgical techniques of reconstruction. J Endovasc Ther
`2010;17:694-702.
`5. Harris PL, Vallabhaneni SR, Desgranges P, Becquemin JP, van
`Marrewijk C, Laheij RJ. Incidence and risk factors of late rupture,
`conversion, and death after endovascular repair of infrarenal aortic
`aneurysms: the EUROSTAR experience. European Collaborators on
`Stent/graft techniques for aortic aneurysm repair. J Vasc Surg 2000;32:
`739-49.
`6. Sampram ES, Karafa MT, Mascha EJ, Clair DG, Greenberg RK,
`Lyden SP, et al. Nature,
`frequency, and predictors of secondary
`procedures after endovascular repair of abdominal aortic aneurysm.
`J Vasc Surg 2003;37:930-7.
`7. Chaar CI, Eid R, Park T, Rhee RY, Abu-Hamad G, Tzeng E, et al.
`Delayed open conversions after endovascular abdominal aortic aneu-
`rysm repair. J Vasc Surg 2012;55:1562-1569.e1.
`8. Kelso RL, Lyden SP, Butler B, Greenberg RK, Eagleton MJ, Clair DG.
`Late conversion of aortic stent grafts. J Vasc Surg 2009;49:589-95.
`9. Koole D, Moll FL, Buth J, Hobo R, Zandvoort HJ, Bots ML, et al.
`Annual rupture risk of abdominal aortic aneurysm enlargement without
`detectable endoleak after endovascular abdominal aortic repair. J Vasc
`Surg 2011;54:1614-22.
`10. Phade SV, Keldahl ML, Morasch MD, Rodriguez HE, Pearce WH,
`Kibbe MR, et al. Late abdominal aortic endograft explants: indications
`and outcomes. Surgery 2011;150:788-95.
`JB, Sternbach Y,
`11. Mehta M, Paty PS, Roddy SP, Taggert
`Kreienberg PB, et al. Treatment options for delayed AAA rupture
`following end