Outcomes of surgeon-modified fenestrated-branched endograft repair for acute aortic pathology

Abstract

Objective

Open surgical repair for acute aortic pathologies involving the visceral vessels is associated with morbidity and mortality rates of 40% to 70% and 30% to 60%, respectively. Due to these poor outcomes, the application of fenestrated/branched endovascular aortic repair (F/B-EVAR) has been expanded in this setting; however, durability remains unknown. The purpose of this analysis was to describe outcomes after F/B-EVAR for acute aortic disease.

Methods

A single center retrospective review of all F/B-EVARs for acute aortic disease was completed. Primary end points included mortality and reintervention-free survival. Secondary end points were patency and freedom from endoleak, as well as change in aneurysm diameter and estimated glomerular filtration rate. Life-tables were used to estimate end points, while mixed statistical models were used to determine aneurysm diameter change.

Results

Thirty-seven patients (mean age ± standard deviation, 67 ± 10 years; 75% male) underwent F/B-EVAR for acute aortic disease, and median follow-up time was 10.3 months (range, 0.5–31.4 months). Indications included thoracoabdominal aneurysm (65%; n = 24), pararenal aneurysm (17%; n = 6), postsurgical anastomotic pseudoaneurysm (8%; n = 3), dissection (5%; n = 2), and penetrating ulcer (5%; n = 2). Mean preoperative aneurysm diameter was 7.3 ±1.8 cm. All patients were American Society of Anesthesiologists class IV or IV-E, and 38% (n = 14) had history of aortic repair. There were 105 visceral vessels revascularized (celiac, 26; superior mesenteric artery, 29; renal, 50) and 24 (65%) patients underwent three- or four-vessel repair. Technical success was 92% (n = 34), with no intraoperative deaths and one conversion (3%). Median length of stay was 6 days (range, 2–60 days), and postoperative morbidity was 41% (n = 15; spinal cord ischemia, 14% [8% permanent]; pulmonary, 14%; renal, 14%; extremity ischemia, 8%; stroke, 5%; cardiac, 3%; bleeding, 3%) with 30-day mortality of 19% (n = 7; in-hospital, 8%; n = 3). Endoleak was detected at some point in follow-up in 27% (n = 10), and a majority were type II (n = 7). Six (16%) patients underwent reintervention, and no late conversions occurred. Postoperative imaging was available in 27 (73%), and one celiac fenestration lost patency at 12 months. One-year branch vessel patency and freedom from reintervention was 98% ± 6% and 70% ± 9%, respectively. Estimated 1- and 4-year survival were 70% ± 8% and 67% ± 8%, respectively. During follow-up, aortic diameter decreased 0.5 cm (95% confidence interval, 1.1–0.2; P = .1) while estimated glomerular filtration rate decreased by 2 mL/min/1.73 m².

Conclusions

F/B-EVAR can be performed to treat a variety of symptomatic and/or ruptured paravisceral aortic pathologies. Perioperative morbidity and mortality can be significant; however, it is less than literature-based outcomes of open repair. Short-term fenestrated/branched graft patency is excellent, but reintervention is frequent, highlighting the need for diligent follow-up. Patients surviving the initial hospitalization for F/B-EVAR of acute aortic disease can anticipate good long-term survival.

The global experience with fenestrated and branched endovascular aortic repair (F/B-EVAR) has increased rapidly in the past 10 years.^1–4 In the elective setting, excellent outcomes have been reported among highly selected patients.^2,5–7 However, unlike infrarenal EVAR, where use in urgent presentations has increased,^8,9 F/B-EVAR has not been widely adopted for acute disease due to significant limitations. In this scenario, obstacles to F/B-EVAR implementation include device availability, mode of presentation, timing of repair, and anatomic suitability, as well as institutional and surgeon expertise, among others. Despite these limitations, F/B-EVAR may be beneficial in this situation due to poor outcomes after open surgical repair of acute visceral segment aortic pathology, where morbidity and mortality rates can be 40% to 70% and 30% to 60%, respectively.^10–12

The development of F/B-EVAR programs and the emergence of aortic treatment centers,¹³ where acute aortic pathology may represent 30% to 50% of referrals,^13,14 will likely eventually lead to selective utilization of F/B-EVAR in the acute setting. To date, there are limited data on short- and mid-term outcomes after urgent or emergent F/B-EVAR for thoracoabdominal disease. Importantly, the implications of patient selection, procedural complexity, and periprocedural renal/neurologic morbidity on longer term outcomes such as endoleak, remediation risk, and survival after F/B-EVAR for acute presentations are unknown.

The purpose of this analysis is to describe our outcomes with F/B-EVAR for management of acute aortic disease.

METHODS

This study was approved by the University of Florida institutional review board, and the need for patient consent was waived (IRB#201400448).

Study cohort

A retrospective review of all patients undergoing urgent/emergent endovascular treatment for visceral segment disease from 2010 to 2014 was performed. Elective F/B-EVAR cases were excluded (n = 104), and patients undergoing urgent/emergent repair (n = 37) were further reviewed. Urgent patients were identified by symptomatic presentation including abdominal, flank, and/or back pain that was not attributable to nonaortic pathology and who underwent repair within 24 hours of admission. Emergent presentation was defined by radiographic and/or hemodynamic evidence of rupture.

Demographics, comorbidities, operative variables (including adjuncts as defined by Society for Vascular Surgery reporting standards^15,16), and postoperative outcomes were abstracted from a prospectively maintained F/B-EVAR registry. Comorbidities and complications were defined using reporting guidelines.^15–17 Pararenal aortic disease was categorized (eg, juxtarenal, suprarenal) using suggested reporting standards¹⁸ and/or the Crawford classification when applicable.¹⁶ Technical success was defined as successful deployment into the intended aortic segment(s), with revascularization of all visceral branch vessels, and no evidence of type I or III endoleak on completion arteriogram.

Clinical practice and patient selection

Our current practice is characterized by a group of board-certified Vascular and Cardiothoracic surgeons who work synergistically to manage a broad range of complex aortic diseases. This long-standing relationship was formalized under the auspices of the University of Florida Aorta Center in 2012 (https://ufhealth.org/uf-health-aorta-center/). All subjects were initially evaluated for feasibility of repair by multiple faculty; however, subsequent consensus judgment determined them to be prohibitive anatomic and/or medical high risk for development of profound morbidity or death with open aortic reconstruction.

Factors that defined anatomic high risk were acute complicated dissection with visceral segment false lumen aneurysm, visceral patch pseudoaneurysm, visceral degeneration proximal to prior infrarenal repair, and mycotic involvement of the paravisceral aorta. Medical high risk was defined as patients deemed to be unable to tolerate aortic clamping or open thoracotomy based on various combinations of advanced medical comorbidities (Supplementary Table I, online only). Comorbidity grading and severity was determined in a manner previously reported (≥8 considered high medical risk).^17,19

Due to the requisite time to complete preoperative planning, and provide patient/family counseling, subjects remaining hemodynamically unstable after initial presentation were not offered F/B-EVAR. Most commonly, emergent patients were taken directly to the operating room and underwent preanesthetic stabilization and invasive monitoring line placement. Patients presenting urgently were initially admitted and stabilized in the surgical intensive care unit. No patients died while awaiting F/B-EVAR for an acute aortic presentation once deemed to be eligible. The ‘off-label’ nature of surgeon-modified F/B-EVAR was disclosed in all cases, and written informed consent was obtained from the patient (or health care surrogate).

Perioperative details

All procedures were performed by a single surgeon (A.B.) with assistance from other attending surgeons or fellowship trainees. We have previously described in detail our principles of preoperative stabilization,²⁰ spinal drain use,^21,22 F/B-EVAR planning using centerline reconstruction (Aquarius 3D; Tera Recon, Foster City, Calif), and device modification and implantation,^20,23 as well as use of intraoperative three-dimensional fusion computed tomography (CT) to facilitate visceral vessel cannulation.²³ An example of perioperative planning and device implantation technique is demonstrated in Fig 1, which highlights a patient who presented with a ruptured paravisceral aortic aneurysm after previous infrarenal endovascular aneurysm repair (EVAR).

Fig. 1.

Example of a patient transferred from an outside hospital with a hemodynamically stable contained rupture of a thoracoabdominal aneurysm after a previous endovascular aneurysm repair (EVAR). A–C, Preoperative imaging of a ruptured Crawford extent III thoracoabdominal aortic aneurysm (TAAA) above a previous Medtronic Talent EVAR (A; red arrow). This patient had a chronically occluded celiac and left renal artery, necessitating a two-vessel fenestrated repair with inclusion of the superior mesenteric artery (SMA) and right renal artery (RRA). The contained rupture was adjacent to the left renal artery (B/C; yellow arrow), and the RRA was nearly occluded and appeared dissected at the origin (C; white arrow). D–F, The repair and postoperative three-dimensional reconstruction. D, The nearly occluded RRA, with successful revascularization through the graft fenestration (E). F, A 6-month postoperative computed tomography (CT) that demonstrated no endoleaks and continued perfusion of the RRA and SMA.

Postoperative recovery occurred in a dedicated cardiovascular surgical intensive care unit. Unless a pre-existing allergy or contraindication existed, patients were given dual antiplatelet therapy postoperatively consisting of clopidogrel (75 mg/day) and aspirin (81 mg/day) for 6 weeks, followed by aspirin (81 mg/day) indefinitely. Patients were typically initiated on postoperative statin therapy if not already on the medication. In-hospital CT angiography was typically performed prior to discharge and then at 1, 6, and 12 months, and then annually in the absence of significant renal dysfunction (estimated glomerular filtration rate [eGFR] <30 mL/min/1.73 m²). In those cases, surveillance consisted of noncontrasted CT with concurrent visceral/renal artery stent duplex ultrasound (DUS).

Definitions

Survival was calculated as time from index operation to the last known date alive, and deaths were confirmed using the Social Security Death Masterfile. Timing and/or need for reintervention was at the senior author’s discretion. Type I and III endoleaks were considered significant and warranted repair unless the patient’s clinical status or anatomic factors limited remedial options. Although no formal protocol for reintervention existed, additional findings that usually prompted reintervention included: (1) persistent type II endoleak (≥6 months) associated with ≥5 mm of aneurysm diameter growth; (2) CT and/or DUS evidence of impending branch failure (renal stent graft peak systolic velocity ≥225 cm/s with eGFR decline ≥ 25%; superior mesenteric artery [SMA] stent graft peak systolic velocity ≥325 cm/s and/or symptoms); or (3) endograft main body, fenestrated/branched vessel, and/or iliac limb deformation resulting in clinical symptoms, ischemia, and/or impression of imminent loss of graft integrity.

Endoleak type was defined using reporting guidelines.^15,16 Aortic diameter evaluation was completed by comparison of the preoperative and most recent postoperative CT scan using centerline reconstruction (Aquarius 3D). The eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration formula.²⁴ Acute kidney injury/failure were based on the Risk, Injury, Failure, Loss, and End-stage kidney disease (RIFLE) criteria.²⁵ Spinal cord ischemia (SCI) has been consistently defined at our institution^22,26 as any new lower extremity motor deficit not explained by any intracranial process or peripheral nerve dysfunction (eg, epidural hematoma, stroke, peripheral neuropathy, or neuropraxia) and may range from mild paraparesis to frank paralysis. Neurology consultation was obtained only in equivocal cases.

End points and statistics

The primary end points of the study were mortality and reintervention-free survival. Secondary end points included freedom from endoleak and branch vessel patency, along with aortic diameter and eGFR change over time. The R statistical package (V.3.0.2; Vienna, Austria) was used to calculate descriptive statistics and make group comparisons. The Fisher exact test was used to compare groups on categorical variables, and the Kruskal-Wallis test to compare groups on ordered variables. The Kaplan-Meier method was applied to estimate freedom from end points. Linear mixed models with random intercepts and slopes for each patient were used to determine aortic diameter change with time. Significance was determined with a P value < .05.

RESULTS

Study population

From 2010 to 2014, there were a total of 1013 emergency room (n = 200) or hospital-to-hospital (n = 813) patient referrals for aortic disease in any anatomic location. Of this group, 530 patients underwent operations for thoracic (n = 313), thoracoabdominal (n = 104), and abdominal (n = 113) aortic disease (Fig 2, A). During the same time period, F/B-EVAR was completed in 141 patients. The annual elective and urgent/emergent open and EVAR volumes for pararenal and thoracoabdominal aortic disease are further highlighted in Fig 2, B.

Fig. 2.

Description of inpatient aortic referral volumes and annual elective, urgent, and emergent operative volumes for complex aortic disease at the University of Florida. A, The total number of inpatient aortic referrals that encompasses all potential patients with acute visceral aortic disease is depicted in this figure. Approximately 40% of cases were found to be nonurgent/emergent and subsequently managed in the outpatient setting. The remaining patients underwent open and/or endovascular operations involving the thoracic, thoracoabdominal, and abdominal aorta. B, The elective and ruptured pararenal and thoracoabdominal open and endovascular surgical volumes are highlighted in the graph. Notably, despite adoption of fenestrated/branched endovascular aortic repair (F/B-EVAR) for elective and acute visceral aortic disease in selected patients, no decrease in open operative volumes is noted, supporting the assertion that judicious application of the technology occurred during this time period. AAA, Abdominal aortic aneurysm. FEVAR, fenestrated endovascular aneurysm repair; OR, operating room; TAAA, thoracoabdominal aortic aneurysm.

F/B-EVAR for acute aortic disease was performed in 37 patients with mean age ± standard deviation of 67 ± 10 years (75% male), and median follow-up time was 8.3 months (range, 0.5–31.4 months). All patients were American Society of Anesthesiologists class IV or IV-E, and 14 (38%) had history of prior open aortic repair or EVAR. The demographics and comorbidities of these subjects are detailed in Table I. Additional characteristics on presentation including symptom status, aortic morphology, and indications are in Table II.

Table I.

Demographics and comorbidities of patients undergoing urgent/emergent fenestrated/branched endovascular aortic repair (F/B-EVAR)

Demographics	Rupture (n = 10)	Symptomatic (n = 27)
Age, years	71 ± 13	70 ± 8
Male	80	73
Body mass index	28 ± 7	27 ± 6
Prior aortic repair	30	41
ASA IV/V(E) status	90	96
Comorbidities
Hypertension	80	93
Renal insufficiency (eGFR <50 mL/min/1.73 m2)	60	25
Coronary artery disease	20	37
Chronic obstructive pulmonary disease	30	26
Diabetes mellitus	30	19
Congestive heart failure	10	15
SVS comorbidity scorea	9 (7–12)	8 (6–12)

ASA, American Society of Anesthesiologists; eGFR, estimated glomular filtration rate; SD, standard deviation; SVS, Society for Vascular Surgery.

Data are presented as mean ± standard deviation or as percentages.

^aData are presented as median (range); Society for Vascular Surgery Reporting Standards.^21,23

Table II.

Patient presentation

Feature	Rupture (n = 10)	Symptomatic (n = 27)
Preoperative aortic diameter, cm	6.8 ± 2.1	7.5 ± 1.7
Pathologic indication
TAAAa
Extent II	0	26
Extent III	20	26
Extent IV	30	22
Suprarenal aneurysm	20	4
Juxtarenal aneurysm	10	7
Postsurgical anastomotic pseudoaneurysmb	0	11
Penetrating ulcer with rupture	20	0
Dissection with false lumen aneurysm (Debakey IIIb)	0	7

TAAA, Thoracoabdominal aortic aneurysm.

Data are presented as mean ± standard deviation or percentage.

^aCrawford classification.

^bOne patient had a mycotic paravisceral pseudoaneurysm after aortic reconstruction.

Operative details

The mean arrival systolic blood pressure was 139 ± 27 mm Hg (median, 135 mm Hg; interquartile range, 122–156 mm Hg). Similarly, mean lowest intraoperative systolic blood pressure was 85 ± 13 mm Hg (median, 86 mm Hg; interquartile range, 76–90 mm Hg), and no patients required intraoperative aortic balloon occlusion. No difference in arrival or lowest intra-operative systolic blood pressure was present between ruptured and symptomatic patients (P = .4). Specific details regarding perioperative vasoactive medications, spinal drain use, device specifications, and operative characteristics are tabulated in Table III. There were 105 visceral vessels revascularized (celiac, 26; SMA, 29; renal, 50), and 24 (65%) patients underwent three- or four-vessel repair. Technical success was 92% (n = 34), with no intraoperative deaths. The three technical failures were due to abandonment of one renal artery (n = 2) and one intraoperative conversion to a hybrid debranching procedure due to graft maldeployment (3%). However, all patients left the operating room with a repaired aneurysm (no type I or III endoleak with an endograft deployed at the intended aortic segment; n = 36; hybrid repair; n = I). Intraoperative endoleak was detected in eight (27%) patients (type I, 0; type II, 7; type III, repaired intraoperatively with a bridging graft, 2; type IV, 1).

Table III.

Operative details and graft configurations

Feature	N = 37
Arrived to OR on vasopressor infusion	14%
Arrived to OR on vasodilator infusion	14%
Received intraoperative vasopressor medicationa	86%
Left OR on vasopressor infusion	16%
Preoperative spinal drain	41%
Operative time, minutes	309 ± 154
Fluoroscopy time, minutes	78 ± 42
Contrast volume (visipaque), mL	100 ± 39
Estimated blood loss, mL	501 ± 544
Fenestrated/branch vessel detailsb
1 or 2 vessel	35%
3 or 4 vessel	65%
Fenestration	82
Directional branch	23
Permanent diameter reducing tie	5
Graft scallop	3
Stent ring excision	1
Aortic stent graft types
Cook TX2	27
Cook TX2 + Endologix main body	7
Cook TX2 + Zenith Flex	1
Zenith Flex	1

OR, Operating room; SD, standard deviation.

Data are presented as mean ± standard deviation or number.

^aVasoactive medication: can be a continuous infusion or single aliquot or both at some point during the conduct of the operation.

^bA total of 109 targeted vessels; 105 revascularized, one patient underwent intraoperative conversion during attempted four-vessel fenestrated endovascular aneurysm repair; two renal arteries abandoned due to cannulation difficulty; data are presented as percentages.

Perioperative results

There were seven patients who died before 30 days (19%), with three in-hospital deaths (8%). Median length of stay was 6 days (range, 2–60 days), and any postoperative morbidity occurred in 15 (41%) cases. The list of postoperative complications is displayed in Table IV. Notably, 19% (n = 7) of patients suffered a neurologic complication. SCI accounted for five of those events, while two patients (5%) had a stroke. Four of the five SCI patients had preoperative spinal drains, and three had permanent SCI. One patient developed postoperative SCI without a preoperative spinal drain, but subsequent drain placement led to prompt neurologic recovery.

Table IV.

Outcomes of urgent/emergent fenestrated endovascular aneurysm repair (FEVAR)

Feature	N = 37
In-hospital death	8
30-day mortality	19
Median length of stay (range), days	6 (2–60)
Complications
Any postoperative morbidity	40
Neurologic	19
Stroke	5
Any SCI	14
Permanent SCI	8
Pulmonary	14
Renal	14
Extremity ischemia	8
Bleeding	3
Cardiac	3
Wound	3

SCI, Spinal cord ischemia.

Data are presented as median (range) or as percentages.

Factors associated with experiencing a postoperative complication are outlined in Table V. There were no differences in demographics, comorbidities, aortic pathology, or procedural urgency. However, patients developing postoperative complications were more likely to have Zone 2/3 coverage, longer fluoroscopy/procedure times, and greater blood loss.

Table V.

Univariate comparison between patients with or without complication after fenestrated endovascular aneurysm repair (FEVAR) for acute aortic disease

Feature	No complication (n = 22; 59%)	Complication (n = 15; 41%)	P value
Gender
Male	17 (77)	10 (71)	.7
Indication
Dissection	0 (0)	2 (13)	.2
Juxtarenal aneurysm	2 (9)	1 (7)
Penetrating ulcer	1 (5)	1 (7)
Postsurgical pseudoaneurysm	3 (14)	0 (0)
Suprarenal aneurysm	3 (14)	0 (0)
TAAA	13 (59)	11 (73)
Urgency
Emergent/rupture	6 (27)	4 (27)	1
Urgent/symptomatic	16 (73)	11 (73)
ASA
IV(E)	22 (100)	15 (100)	1
Hypertension	19 (86)	14 (93)	.6
Coronary artery disease	7 (32)	5 (33.3)	1
Congestive heart failure	2 (9)	3 (20.0)	.4
Arrhythmia	3 (14)	1 (7)	.6
Chronic lung disease	5 (23)	5 (33)	.7
Current smoker	8 (36)	10 (67)	.1
Diabetes mellitus	5 (23)	3 (20)	1
Dyslipidemia	7 (32)	3 (20)	.5
Cerebrovascular disease	4 (18)	0 (0)	.1
Peripheral artery disease	2 (9)	2 (13)	1
Preoperative dialysis	1 (5)	0 (0)	1
Preoperative eGFR	60 [54.5, 60] (8, 60)	55.5 [34, 60] (19, 60)	.1
Prior aortic repair	9 (41)	5 (33)	.8
Preoperative aortic diameter ± SD, mm	73 ± 18	72 ± 17	.6
Proximal coverage zone
2 or 3	4 (18)	9 (60)	.05
4–7	18 (82)	6 (40)
Number of fenestrations
1 or 2	9 (41)	4 (27)	.9
3 or 4	13 (59)	11 (73)
Adjunctive procedure	7 (32)	7 (47)	.5
Intraoperative complication	1 (5)	2 (13)	.6
Fluoroscopy time, minutesa	56 (42–79) [32–165]	90 (66–125) [30–200]	.05
Contrast volume, mLa	90 (73–125) [20–155]	110 (75–138) [20–200]	.3
Procedure time, minutesa	240 (190–360) [134–512]	360 (242–440) [150–900]	.05
Estimated blood loss, mLa	200 (200–300) [100–800]	300 (300–450) [150–5000]	.04

ASA, American Society of Anesthesiologists; eGFR, estimated glomerular filtration rate; SD, standard deviation; TAAA, thoracoabdominal aortic aneurysm.

Other data are presented as number (%) unless otherwise indicated.

^aData are presented as median (range) [interquartile range].

Outcomes

There were seven perioperative deaths, and five additional patients died due to nonaortic-related causes after hospital discharge within 50.9 months of operation (maximum follow-up time). The estimated 1- and 4-year survival after F/B-EVAR for acute aortic disease was 70% ± 8% and 67% ± 8%, respectively (Fig 3). Patients who were discharged and survived beyond 30 days were more likely to be on a preoperative angiotensin-converting enzyme medication (or angiotensin receptor blocker; P = .01) but less likely to have a history of congestive heart failure (P = .04). No other significant differences were found between surviving patients and those who died (Table VI).

Fig. 3.

This Kaplan-Meier curve demonstrates the estimated midterm survival after urgent/emergent fenestrated/branched endovascular repair (F/B-EVAR). All reported intervals are less than 10% standard error of the mean. FEVAR, Fenestrated endovascular aneurysm repair.

Table VI.

Univariate comparison between patients who did or did not survive after fenestrated endovascular aneurysm repair (FEVAR) for acute aortic disease

Feature	Alive at discharge and survived >30 days (n = 30; 81%)	In-hospital or 30-day death (n = 7; 19%)	P value
Gender
Male	22 (73)	5 (83)	1
Indication
Dissection	1 (3)	1 (14)	.4
Juxtarenal aneurysm	3 (10)	0 (0)
Penetrating ulcer	1 (3)	1 (14)
Postsurgical pseudoaneurysm	3 (10)	0 (0)
Suprarenal aneurysm	3 (10)	0 (0)
TAAA	19 (64)	5 (72)
Urgency
Emergent/rupture	8 (27)	2 (29)	1
Urgent/symptomatic	22 (73)	5 (71)
ASA
IV (E)	30 (100)	7 (100)	1
Hypertension	27 (90)	6 (86)	1
Coronary artery disease	8 (27)	4 (57)	.2
Congestive heart failure	2 (7)	3 (43)	.04
Arrhythmia	3 (10)	1 (14)	1
Chronic lung disease	9 (30)	1 (14)	.6
Current smoker	14 (47)	4 (57)	.7
Diabetes mellitus	8 (27)	0 (0)	.3
Dyslipidemia	9 (30)	1 (14)	.6
Cerebrovascular disease	3 (10)	1 (14)	1
Peripheral arterial disease	2 (7)	2 (29)	.2
Preoperative hemodialysis	1 (3)	0 (0)	1
Preoperative eGFR	60 (44–60) [8–60]	59 (49–60) [19–60]	.5
Preoperative ASA ± warfarin	18 (60)	3 (43)	.4
Preoperative statin	9 (30)	2 (29)	1
Preoperative ARB/ACEI	17 (57)	0 (0)	.01
Preoperative β-blocker	19 (63)	2 (29)	.2
Prior aortic repair	13 (43)	1 (14)	.2
Preop aortic diameter ± SD, mm	73 ± 17	71 ± 23	.7
Proximal coverage zone
2 or 3	9 (30)	4 (57)	.7
4–7	21 (70)	3 (43)
Number of fenestrations
1 or 2	11 (37)	2 (29)	.5
3 or 4	19 (63)	5 (71)
Adjunctive procedure	11 (37)	3 (43)	1
Intraoperative complication	3 (10)	0 (0)	.7
Fluoroscopy time, minutesa	68 (43–107) [18–200]	66 (49–75) [30–126]	.8
Contrast volume, mLa	100 (76–125) [20–200]	75 (62–108) [20–150]	.3
Procedure time, minutesa	280 (196–383) [134–900]	240 (200–410) [150–472]	.9
Estimated blood loss, mLa	300 (200–300) [100–5000]	300 (250–350) [150–1000]	.6

ACEI, Angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; ASA, American Society of Anesthesiologists; eGFR, estimated glomerular filtration; SD, standard deviation; TAAA, thoracoabdominal aortic aneurysm.

Data are presented as number (%) unless otherwise indicated.

^aData are presented as median (range) [interquartile range].

The median clinical/radiographic follow-up time was 10.3 months (range, 0.5–31.4 months). Six (16%) patients underwent reintervention, and no late conversions occurred. The 1-year freedom from reintervention was 70% ± 9% (Fig 4). Reinterventions after hospital discharge included renal/visceral artery stent placement for endoleak (renal, n = 1; SMA, n = 1; celiac, n = 1), branch vessel stent extension for impending loss of seal (renal, n = 1), iliac angioplasty with stent extension for impending limb failure (n = 1), and proximal thoracic EVAR for a ruptured aortic dissection (n = 1 at 21.3 months).

Fig. 4.

The freedom from reintervention after nonelective fenestrated/branched endovascular aortic repair (F/B-EVAR) is highlighted in the figure. Notably, six patients underwent remediation, and all were managed with endovascular techniques. All reported intervals are less than 10% standard error of the mean.

At least one postoperative CT was available in 68% (n = 25 with CT angiography after hospital discharge) of patients. Endoleak was seen at some point in follow-up in 10 (27%), and the majority were type II (n = 7; type III, n = 3). One-year freedom from endoleak after discharge was 68% ± 9% (Fig 5). Presence of intraoperative endoleak was not associated with a higher likelihood of secondary intervention after discharge (P = .6).

Fig. 5.

There were 10 patients (27%) who survived initial hospitalization with an outpatient contrasted postoperative computed tomography (CT) scan during follow-up with evidence of endoleak. If patients were noted to have an intraoperative endoleak with graft implantation, this did not lead to higher likelihood of endoleak after hospitalization. Seventy percent of the posthospital discharge endoleaks were type II. The remaining endoleaks (type III; n = 3) all underwent successful endovascular remediation. All reported intervals are less than 10% standard error of the mean.

Posthospital discharge CT and/or DUS was available in 27 (73%), and one celiac fenestration lost patency at 12 months. One-year branch vessel patency was 98% ± 6% (Fig 6).

Fig. 6.

The primary patency of all fenestrated/branched endografts in the series is displayed. Excellent 12-month patency is reported; however, one celiac fenestration was documented to have occluded at 12 months postoperatively. All reported intervals are less than 10% standard error of the mean.

During follow-up, aortic diameter decreased on average by 0.5 cm (95% confidence interval, 1.1–0.2; P = .1; Fig 7). The rate of diameter change among patients with or without an intraoperative endoleak was not different, but those with intraoperative endoleak were estimated to have an aortic diameter that is, on average, across all time points, 16.0 mm larger than those without endoleak (95% confidence interval, 3–29; P = .02).

Fig. 7.

The overall trajectory of aortic diameter remodeling is demonstrated in this figure. Because repeated measures over time in the same patient are not independent, simple linear modeling is not appropriate. Mixed statistical models were applied to better understand the behavior of the aorta after nonelective fenestrated/branched endovascular aortic repair (F/B-EVAR). Notably, aortic diameter stabilization and/or regression were observed in the majority of patients with available postoperative imaging. CI, Confidence interval; FEVAR, fenestrated endovascular aneurysm repair.

For patients surviving hospital discharge with available laboratory data (n = 27), eGFR decreased on average by 2 mL/min/1.73 m² during the follow-up interval. When patients were stratified into those with or without preoperative renal dysfunction (eGFR <60 mL/min/1.73 m²), there was a significant difference in the trajectory of eGFR change with increasing follow-up time (P < .0001). Specifically, patients with pre-existing renal insufficiency experienced a more precipitous deterioration in renal function at 6 and 12 months postoperatively (Fig 8).

Fig. 8.

The posthospital discharged renal outcomes are significantly different for patients with pre-existing renal insufficiency. Specifically, patients with a preoperative estimated glomerular filtration rate (eGFR) < 60 mL/min/1.73 m² who survived initial hospitalization after urgent/emergent fenestrated/branched endovascular aortic repair (F/B-EVAR) had significantly worse renal function in short-term follow-up. These findings have significant implications on long-term survival, as well as surveillance and reintervention protocols. FEVAR, Fenestrated endovascular aneurysm repair; preop, preoperative.

DISCUSSION

This report represents the largest series of F/B-EVAR used to treat acute aortic pathology in the existing literature, and demonstrates that acceptable outcomes can be achieved with properly selected patients. Despite the procedural complexity, technical success was high. Postoperative morbidity and early mortality was significant; however, trends in aneurysm diameter stabilization and endoleak risk after discharge appear acceptable. Additionally, renal outcomes in early follow-up may be significantly different for patients with preoperative renal dysfunction. Reintervention risk was modest, but was consistent with the existing literature for elective F/B-EVAR,²⁷ and remediation was accomplished with endovascular techniques. Most importantly, this analysis demonstrates that good long-term survival can be anticipated after successful F/B-EVAR for acute aortic disease.

Early morbidity and mortality events appeared to be predominantly influenced by patient covariates, as well as anatomic and procedural complexity (Tables V and VI). Patients with a history of congestive heart failure appeared to have increased risk of 30-day death. The relatively protective effect of angiotensin-converting enzyme/angiotensin receptor blocker medication(s) was unexpected but may represent a spurious association related to a small sample size or may reflect a subset of patients with fundamentally different physiologic risk. Although the complication rate is significant, when placed in the context of elective F/B-EVAR, these results seem acceptable. Elective F/B-EVAR has an overall morbidity rate of 24% to 43%,^5,28–30 which is similar to results in this analysis. The most prevalent complications in the elective setting are neurologic (2%–15%), renal (1%–10%), and pulmonary (10%–15%) impairment.

The overall neurologic complication rate in our series was 19% and was primarily driven by development of SCI (any SCI, 14%; permanent, 8%), thus a heightened awareness for this complication is warranted. The elevated risk compared with elective F/B-EVAR is likely related to multiple factors, including: (1) elective F/B-EVAR reports predominantly focus on outcomes of juxtarenal/suprarenal aortic disease; (2) we use a comprehensive SCI definition that potentially has high sensitivity but low specificity^22,26; and (3) 65% of the operations were three or four-vessel repairs for extent II or III thoracoabdominal aortic aneurysm (TAAA) disease requiring extensive aortic coverage, a known risk factor for development of SCI.²⁶

We are aggressive about spinal drain use in the elective setting, but the acute clinical presentation of this cohort makes routine cerebrospinal fluid drains impractical. We also have a policy of staging long-segment coverage for endovascular thoracoabdominal aneurysm repair to decrease SCI risk³¹; however, this cannot be done safely in the acute setting. Indeed, these results motivated us to evolve our SCI protocols,²² which now resemble interventions for open TAAA repair.³² In addition to cerebrospinal fluid drain placement, these adjuncts include permissive hypertension (goal mean arterial pressure ≥90), mannitol prior to and after endograft deployment, a naloxone drip (1 μg/kg/h) prior to deployment and for 48 hours afterward, hemoglobin maintenance ≥9 mg/dL (≥10 mg/dL if SCI occurs), and a bolus dose of steroids at the beginning of the procedure (methylprednisolone, 30 mg/kg).

Another important outcome measure that merits specific discussion is renal morbidity. A systematic review by Katsargyris and colleagues⁵ documented a 9.8% postoperative renal complication rate after F/B-EVAR similar to this analysis. The impact of this complication is highlighted by reports demonstrating significantly elevated early and late mortality rates for patients experiencing postoperative acute kidney injury after endovascular³³ and open aneurysm repair.³⁴ Moreover, F/B-EVAR has a known association with long-term renal function decline in 25% to 33% of subjects,^2,35 which is related to multiple factors including comorbidities, alteration in target vessel perfusion, atheroemboli, and repeated contrast exposure. To reduce this risk, we now routinely employ three-dimensional fusion CT technology²³ to minimize intraoperative contrast exposure in all F/B-EVAR cases and modify surveillance in renal dysfunction patients (eg, noncontrast CT with branch vessel DUS).

An alternative cause of renal (or mesenteric) morbidity is loss of target vessels during or after F/B-EVAR procedures. During follow-up, no renal or SMA vessels were lost; however, one asymptomatic celiac occlusion occurred at 12 months. The corresponding 1-year patency (98% ± 6%) in this analysis is consistent with published reports.^2,7 Notably, two renal arteries (intention to treat, n = 54) were abandoned during endograft implantation due to inability to catheterize the vessels. Both vessels were extremely tortuous, short, and had an orificial stenosis. However, neither fenestration required coverage because they approximated the aortic wall and did not perfuse the aneurysm sac. This resulted in a 3.7% rate of kidney loss in this study, which compares favorably with the 1.5% to 3.0% rate in other series.^5,7

The gold standard for comparison of elective infrarenal and juxtarenal aortic repair outcomes has been open surgical reconstruction. However, in the setting of acute thoracic and/or thoracoabdominal pathology, this paradigm has been challenged due to the morbidity of open repair.^11,36 Emergency open surgical repair of thoracoabdominal aneurysms results in a crude postoperative mortality rate exceeding 50% while morbidity approaches 70%, which has not improved in contemporary practice.^11,12 These dismal results have led to calls for development of new repair techniques to address this challenging clinical scenario.

Some authors advocate use of hybrid repair (eg, visceral debranching with thoracic EVAR) for acute visceral aortic pathology. However, the 30-day mortality rates are 25% to 40% with corresponding morbidity rates of 30% to 60%.^37,38 Parallel stents (“chimney, snorkel, or sandwich” stents) represent an alternative strategy, with encouraging results after elective pararenal aneurysm repair, but there may be an increased risk of major adverse events in follow-up.^5,39 The current parallel stent literature pertaining to symptomatic or ruptured suprarenal and TAAA disease remains anecdotal and limited to case reports, so outcomes are poorly understood. These sobering results further support a role for F/B-EVAR in selected patients with acute aortic disease.

The rationale for this analysis was related to a pilot program we developed at our institution examining feasibility and safety of attempting F/B-EVAR for acute aortic disease.²⁰ The original series focused on the technical descriptions and 30-day morbidity and mortality outcomes. In contrast, this analysis sought to describe out-of-hospital survival, reintervention, endoleak, renal morbidity, and aortic remodeling events. As F/B-EVAR adoption occurred in our elective practice, we expanded use of the technology in more urgent and emergent situations, similar to other groups.⁴⁰ However, appropriateness of F/B-EVAR for acute pathology remains undefined due to questions regarding patient selection, device availability, variable local institutional/surgeon expertise, and limited reports chronicling outcomes. Many of these concerns provide justification for concentration of these complex aortic cases to centers of excellence.

Over the course of the study interval, we did not have an investigational device exemption for use of surgeon-modified F/B-EVAR. This technology was selectively applied for compassionate treatment of high-risk patients with complex aortic disease who otherwise would not have access to a manufactured fenestrated/branched device, and/or could not wait for device customization, or where ‘off-the-shelf’ designs were not suitable. We share the opinion with others⁴⁰ that this population defines the subset of patients where niche utilization of surgeon-modified F/B-EVAR may be justified.

Our current practice is to attempt F/B-EVAR in patients with symptomatic and/or contained ruptured visceral aortic pathology if they are considered to be at excessive risk for morbidity and mortality with open repair, and are stable enough to allow graft planning, device modification, and implantation. All cases are discussed in detail with the patient/family, and we fully disclose the ‘off-label’ nature of the procedure and unknown long-term durability. Despite these concerns, the experience gained with surgeon-modified F/B-EVAR facilitated our application and subsequent Food and Drug Administration approval of a physician-sponsored investigational device exemption that currently allows use of both surgeon-modified and custom-designed fenestrated-branched endografts to treat aortic disease (www.clinicaltrials.gov identifier: NCT02043691).

There are important limitations to this analysis. The single-center, single-surgeon experience compounded by small patient numbers with short follow-up leaves several questions unanswered about the role of F/B-EVAR and treatment of acute aortic disease. The potential for type II error is significant and calls for greater patient numbers with longer term follow-up to further define the role of this technology in these challenging patients. Notably, no open surgical repair cohort was presented for comparison, as this was a highly selected patient population given the need for hemodynamic stability and anatomy amenable to F/B-EVAR.

Importantly, a learning curve effect can exist with complex operations, which may impact outcomes and practice patterns. However, no difference in patient risk profile or outcome was discovered during the study interval (Supplementary Table II, online only). We continued to judiciously apply F/B-EVAR over the course of the study as evidenced by the increasing open and endovascular volumes for both elective and emergent indications (Fig 2). Despite these limitations, we feel that these results provide evidence to support utilization of F/B-EVAR in the management of acute visceral aortic disease, particularly in centers specializing in the care of complex aortic disease.

CONCLUSIONS

F/B-EVAR can be performed to treat a variety of symptomatic and/or ruptured paravisceral aortic pathologies. Perioperative morbidity and mortality can be significant; however, it is less than historical outcomes of open surgical repair. Short-term branch patency is excellent, but reintervention is common, highlighting the need for diligent follow-up. Patients surviving the initial hospitalization for F/B-EVAR of acute aortic disease can anticipate good long-term survival.