From the New England Society for Vascular Surgery
Stent-graft versus open-surgical repair of the
thoracic aorta: Mid-term results
David H. Stone, MD, David C. Brewster, MD, Christopher J. Kwolek, MD, Glenn M. LaMuraglia, MD,
Mark F. Conrad, MD, Thomas K. Chung, MA, and Richard P. Cambria, MD, Boston, Mass
Objective: Pivotal and comparative trial data are emerging for stent graft (SG) vs open repair of the thoracic aorta. We
reviewed procedure-related perioperative morbidity, mortality, and mid-term outcomes in a contemporary series of
patients treated with SG of the thoracic aorta. The data were compared with those of a patient cohort concurrently treated
with open surgical repair confined to the descending aorta.
Methods: A review of patients undergoing SG procedures and open surgery of the thoracic aorta from January 1, 1996, to
November 30, 2005, was performed from a prospectively compiled database. Study end points included perioperative
complications, late survival, freedom from reinterventions, and graft-related complications. Multivariate methods were
used to assess variables potentially associated with study end points; late outcomes were compared with actuarial
Results: In 105 patients (mean age, 70 years; 66 male [62.9%]) SG repairs were done for 68 degenerative aneurysms
(64.7%), 12 penetrating ulcers (11.4%), 15 pseudoaneurysms (14.3%), 9 traumatic tears (8.6%), and 1 acute dissection
(0.9%). Mean follow-up was 22 months (range, 0 to 101 months). Eighty-nine (84.8%) SG patients were asymptomatic
at presentation and underwent elective repair, whereas 16 (15.2%) presented with acute conditions and underwent urgent
repair. Perioperative mortality was 7.6% (8/105), and actuarial survival at 48 months was 54% ? 7%. The perioperative
mortality rate among SG patients treated for degenerative pathology was 10.4% (8/77). Seven (6.7%) of 105 patients
experienced spinal cord ischemic complications, including 2 patients with transient paraparesis that resolved by the time
of discharge. Reinterventions were performed in 10.5% of patients (11/105), with freedom from reintervention
approaching 81% by 48 months. Over the same interval, 93 patients were treated with open-surgical repair for descending
thoracic aneurysm (anastomosis cephalad to the celiac axis). Perioperative mortality in the open cohort was 15.1%
(14/93; P ? .09 vs SG repair), and the 48-month actuarial survival was 64% ? 6%. The incidence of spinal cord ischemic
complications was 8.6% (8/93), including 4 patients with transient paraparesis (P ? .44 vs SG repair). Nine patients
(9.7%) required surgical reintervention during the follow-up period, with 48-month freedom from reintervention
approaching 79% (P ? .73 vs SG repair).
Conclusions: Operative mortality was halved with SG, with similar late survival for both cohorts. Reinterventions were
required at a nearly identical rate for open repair and SG, and both groups experienced similar rates of spinal cord
ischemic complications. (J Vasc Surg 2006;44:1188-97.)
The incidence of degenerative thoracic aortic aneu-
rysms has increased substantially in contemporary prac-
tice.1,2 Such increase in prevalence will likely demand an
increased need for thoracic aortic intervention.2-4 Histori-
cally, conventional open surgery has been the mainstay of
therapy; yet even in centers of excellence, operative mortal-
ity and paraplegia risks generally run in the 10% range.5-8
Although initial reports of stent graft repair (SG) in the
thoracic aorta created enthusiasm for a less invasive treat-
ment alternative,9,10 subsequent reports have since docu-
mented the safety and efficacy of SG of the thoracic
aorta.11-13 Pivotal and comparative (vs open repair) trial
data are emerging for the spectrum of thoracic aortic pa-
thology.14-16 With the recent advent of an Food and Drug
Administration–approved, commercially available device,
the treatment paradigm for a variety of thoracic aortic
pathologies has rapidly evolved toward stent-grafting strat-
egies, although late results and level 1 evidence in the form
of randomized clinical trials do not exist.17-19
Most reports detailing experience with TAA SG repair
lack direct comparison with patients treated with open
surgery; such data are largely limited to a single industry-
sponsored pivotal trial.11,14 With a study goal of providing
such comparative data, we reviewed a contemporary con-
currently treated series of patients managed with both SGs
and open surgical repair for thoracic aortic pathology.
PATIENTS AND METHODS
During the interval from January 1996 to November
2005, 105 patients underwent SG repair of the thoracic
aorta. Open surgical repair for degenerative (dissections
excluded) descending thoracic aneurysms was done in 93
patients, of which 60% involved the entire descending
aorta. Inclusion criteria included a distal aortic suture line
cephalad to the celiac axis.
From the Division of Vascular and Endovascular Surgery and the Depart-
ment of Surgery, Massachusetts General Hospital, Harvard Medical
Supported in part by the Harold and June Geneen Vascular Research Fund.
Competition of interest: none.
Presented in part at the Thirty-second Annual Meeting of the New England
Society of Vascular Surgery, Stowe, Vt, September 2005.
Reprint requests: Richard P. Cambria, MD, Division of Vascular and Endo-
vascular Surgery, Massachusetts General Hospital, 15 Parkman St, WAC
458, Boston, MA 02114 (e-mail: firstname.lastname@example.org).
Copyright © 2006 by The Society for Vascular Surgery.
cially manufactured devices or custom-made SGs. The 20
custom-made devices consisted of an endoskeleton of self-
expanding Gianturco Z stents (W. A. Cook, Inc, Bloom-
ington, Ind) covered with an ironed woven Dacron graft
(Cooley Veri-soft; Meadox Medicals, Inc, Oakland, NJ).
These “first-generation” constructs were last used in 2004.
All patients underwent perioperative evaluation with fine-
cut (3-mm) computed tomography (CT) imaging of both
the thoracic and abdominal aortic segments. Three-dimen-
sional reconstructions were obtained to aid in device sizing
for SG patients (Medical Metrx Solutions, West Lebanon,
NH). Diagnostic angiography was also used selectively in
the early stages of our experience.
Inclusion criteria generally involved 2-cm segments at
proximal and distal fixation sites with vessel diameters not
exceeding 37 mm in maximum diameter. More proximal
fixation was achieved with associated exclusion of the left
subclavian artery or left carotid artery.
Twenty-six patients (24%) were treated as part of the
the Treatment of Thoracic Aortic Aneurysms (VALOR)
trial, which included subgroups of both high-risk and com-
passionate device use patients (n ? 16). Criteria for this
group included high risk for open surgery (Society for
Vascular Surgery (SVS) score of ?3), or considered a
nonsurgical candidate not associated with SVS scoring, or a
patient with traumatic thoracic aortic injury who was oth-
erwise stable, or a combination of these. TAG devices
(W.L. Gore & Associates, Flagstaff, Ariz) were used to treat
22 patients (21%) in the context of phase II clinical trials,
whereas 29 patients (28%) received Gore commercially
available devices. Eight patients were treated with Cook
Zenith Tx2 devices as part of the Zenith Tx2 thoracic
surgical repair were not applied over the study interval. In
general, custom-made or compassionate use devices were
applied in patients unfit for open repair; alternatively, such
patients were entered into the VALOR high-risk arm,
wherein enrollment ceased in 2002. Patients with anatomy
amenable for SG repair were routinely offered participation
in one of 4 industry-sponsored FDA–approved clinical
trials with a variety of clinical exclusion criteria. Aneurysms
were repaired in the context of symptoms, rapid enlarge-
ment, or approximate 6 cm size. Penetrating ulcers were
repaired in context of an acute aortic syndrome, threatened
rupture, or size ? 6 cm.
imaging social security database follow-up was performed
in each patient group. Comorbidities, including chronic
obstructive pulmonary disease (COPD) with associated
clinical impairment (forced expiratory volume in 1 second
[FEV1] ?50% predicted), chronic renal insufficiency (se-
rum creatinine ?1.5 mg/dL), hypertension, diabetes mel-
litus, and tobacco use were assessed. Procedural outcome
complications including cardiac (myocardial infarction),
renal failure (?30% serum creatinine), pulmonary, and
spinal cord ischemic complications were prospectively re-
corded. End points included periprocedural mortality
(within 30 days or in-hospital death during the same ad-
mission), actuarial mid-term survival, freedom from rein-
tervention, and endoleak.
Statistical analysis of comparative demographic and
clinical variables, as well as study end points between the
two cohorts, was performed with ?2testing Fisher’s exact
test, Student’s t test, and Mann-Whitney tests. Survival
analysis was performed by using Kaplan-Meier life tables
with Mantel-Cox log-rank univariate analysis to identify
differences between groups.
Patient characteristics. During the study interval,
thoracic endografts were deployed in 105 patients; their
mean age was 70 years, and 65 (61.9%) were men. Pathol-
ogies treated included 68 degenerative aneurysms (64.7%),
12 penetrating ulcers (11.4%), 15 pseudoaneurysms
(14.3%), 9 traumatic tears (8.6%), and 1 acute dissection
(0.9%). Devices used included 51 TAG (48.5%), 26 Talent
By comparison, 93 patients (51 [54.8%] male; mean age,
71 years) underwent open-surgical repair of thoracic aortic
aneurysms, as defined previously.
No difference was noted between groups regarding
age, sex, hypertension, baseline renal function, COPD, or
diabetes. A significant difference was noted in tobacco use
among patients in the endograft cohort: 75.2% (79/105)
shown in the Table. Of significance, 27.6% (29/105) of
stent graft patients were deemed unfit for conventional
open-surgical repair secondary to significant comorbidities,
including severe congestive heart failure, myocardial infarc-
tion, severe COPD (forced expiratory volume in 1 second
30%-50% predicted), or renal insufficiency (serum creati-
nine ?2.2 mg/dL). Among patients with degenerative
pathology, 29.9% (23/77) were deemed unfit for open
Perioperative mortality. Perioperative mortality was
7.6% (8/105) in the endograft cohort. By comparison,
perioperative mortality in the open-surgical group was
15.1% (14/93; P ? .09). Etiologies for perioperative mor-
failed attempt at deployment with subsequent withdrawal
of support (n ? 1). There was one intraoperative death,
which involved faulty device deployment with subsequent
conversion to unsuccessful open repair. Among patients
who underwent open-surgical repair, myocardial infarction
(n ? 5), stroke (n ? 6), aortoenteric fistula (n ? 1), sepsis
(n ? 1), and pancreatitis (n ? 1) accounted for periproce-
Subgroup analysis of perioperative mortality rates
among SG patients with degenerative pathology (degener-
ative aneurysms and penetrating ulcers) was 10.4% com-
pared with 15.1% in the open-surgical controls (P ? .37).
Furthermore, when ruptures were excluded from analysis,
JOURNAL OF VASCULAR SURGERY
Volume 44, Number 6
Stone et al 1189
perioperative mortality rates were 8.1% in the SG group vs
9.6% in the open surgical cohort (P ? .74).
Procedure-related data. Technical success of implan-
tation, defined as device deployment at the intended seg-
ment/seal zone, was achieved in 100 (95.2%) of 105
to the intended fixation site because of either tortuosity or
vascular injury. Faulty or incomplete device deployment
occurred in two additional patients. A total of 185 devices
were deployed in 105 subjects; 46.7% (n ? 49) of patients
received a single device, 30.5% (n ? 32) received 2 devices,
21.9% (n ? 23) received 3 device pieces, and 1% (n ? 1)
received 5 pieces at the initial procedure. The mean aortic
length covered was 28.2 cm.
Arterial access was obtained via femoral cutdown in
76.2% (80/105) of patients, and 23.8% (n ? 25) of pa-
tients required arterial conduit access for device introduc-
tion. Among this subgroup, 17.1% (n ? 18) required iliac
access via retroperitoneal cutdown, and 6.7% (n ? 7)
of SGs were placed intentionally covering the left subcla-
vian artery origin, and 11.4% (n ? 12) underwent carotid
subclavian bypass (57% of those with left subclavian cover-
age). Four patients (4%) required carotid-to-carotid bypass
to enable SG seal proximal to the left common carotid
origin. Management of the left subclavian artery varied by
surgeon and clinical presentation. Ileofemoral access com-
plications, defined as any injury incurred to the femoral or
iliac arteries during the procedure, irrespective of access
type, occurred in 17.2% (n ? 14) of patients in the stent-
graft cohort. These patients required lower midline lapa-
rotomy or retroperitoneal repair of the access vessels.
Open-surgical patients underwent open thoracotomy/
thoracoabdominal incisions with clamp placement at or
distal to the left subclavian artery and proximal to the celiac
axis. Atrial-femoral bypass was used in 30 (32.3%) of the 93
patients; no cases were done with hypothermic circulatory
Spinal cord ischemia. The incidence of spinal cord
ischemic complications (any degree of deficit) was 6.7%
(7/105) in the SG group and 8.6% (8/93) in the open-
surgery group (P ? .44). The incidence of SCI complica-
tions in patients treated for degenerative pathology exclud-
ing ruptures was 10.4% in the SG cohort versus 7.2% in the
open group (P ? 0.49), paraplegia in the endograft cohort
was 4.8% (5/105) and was 4.5% (4/93) in the open-
surgical group (P ? .84). Spinal cord–protective adjuncts
were not routinely used in the endograft group but were
implemented on a case-by-case basis, typically in patients
with a history of abdominal aortic surgery or in anticipation
of extensive descending aortic SG coverage.21 In the open
surgery cohort, 80.6% (75/93) had adjunctive use of cere-
brospinal fluid drains, and 81.7% (76/93) had epidural
cooling, as previously described,21 vs close to 8% (8/105)
in the SG group. Among the five cases of paraplegia in the
endograft group, all were paraplegic with neurologic defi-
cits noted immediately after surgery or on postoperative
day 1. Thirty-three patients in the endograft group had
previous AAA. Among the seven SG patients with SCI
complications, 3 (42%) had previous AAA. Among paraple-
gic patients, two had previous AAA repair, and one of the
patients with paraparesis had prior AAA repair (odds ratio,
2.04; 95% confidence interval, 0.64-6.5; P ? .23). Among
the paraplegic patients in the open-surgical group, all had
spinal cord infarcts. Four patients in this group had tran-
sient lower extremity weakness. Two patients had early-
onset left lower extremity weakness that resolved within 2
days. Two additional patients had bilateral lower extremity
motor deficits that were resolved by discharge.
Stroke. Ten patients (9.5%) in the endovascular cohort
sustained an intraoperative stroke and stroke occurred in 7
(7.5%) of the 93 open surgery patients (P ? .62). Subgroup
analysis of degenerative pathology, excluding rupture, re-
vealed a 11.9% incidence of stroke in the SG group vs 6.0% in
the open surgery group (P ? .2). Among patients in the
the arch with seal zones proximal to the left subclavian artery.
One patient sustained a stroke secondary to device deploy-
ment proximal to the left carotid artery necessitating urgent
carotid-carotid bypass. Postoperatively, the patient was found
to have a large right middle cerebral artery distribution cere-
brovascular accident (CVA) and subsequently died. A cere-
brovascular accident occurred in a second patient after bare
metal stent deployment across the left carotid artery that
manifested as left hand weakness and in a third patient after
to aortic tortuosity.
stay among patients undergoing SG procedures was 3.65
days (range, 0-59 days) and was 9.26 days (range, 1-60
Table I. Demographic and clinical features among patient cohorts.
Endo Cohort (n?77)
Endo Cohort (n?74)
Degenerative excl. Rupt.
Open Cohort (n?93)
Open Cohort (n?83)
Degenerative excl. Rupt.
75.57 ? 7.76 [47-91]
75.61 ? 7.11 [47-89]
70.8 ? 9.8 (31-89)
70.2 ? 9.8 (31-89)
JOURNAL OF VASCULAR SURGERY
1190 Stone et al
days) in the open-surgical group (P ? .001). Mean total
hospital length of stay was 11.0 days in the SG group vs
18.76 days in the open-surgical group (P ? .001).
Late survival. Late survival for the SG and open sur-
gical cohorts is depicted in Figs 1, 2 and 3. Kaplan-Meier
48-month survival was 54% ? 7% in the SG cohort vs 64%
? 6% in the open-surgical group (P ? .91). Subgroup
survival analysis for SG patients with degenerative aneu-
rysms, excluding patients with traumatic tears, pseudoan-
eurysms, and acute dissection, revealed a 54% ? 7% 48-
month survival vs 64% ? 6% in the open surgical group (P
? .60). Furthermore, when the 30% of patients deemed
unfit for open surgery were excluded, 48 month survival
was 60 ? 8% in the SG group versus 64 ? 6% in the open
group (P ? 0.47).
Aneurysm-related mortality. The incidence of aneu-
rysm-related death was 11.4% (12/105) among patients
treated with SG versus 15.1% (14/93) in the open-surgical
group (P ? 0.45). Freedom from aneurysm-related mor-
tality in the SG patients versus open-surgical patients was
85% ? 4% vs 82% ? 4%; (P ? 0.43).
Endoleaks. Endoleaks at any time interval occurred in
14 patients (13.3%). These included six (42.9%) type I
attachment site leaks, seven (50%) type II leaks, and one
(7.1%) delayed type III endoleak.
Among patients with type I leaks, three asymptomatic
patients were noted to have attachment site leaks with
progressive sac enlargement on follow-up CT scan. All
cases required reintervention. One patient initially thought
to have a type II endoleak via retrograde perfusion from the
0 12 24 36 48
Fig 1. Kaplan-Meier curve actuarial life-table analysis depicting intermediate (mid-term) survival between patients
treated with thoracic stent grafts and open-surgical repair. S.E., Standard error.
JOURNAL OF VASCULAR SURGERY
Volume 44, Number 6
Stone et al 1191
left subclavian artery underwent a left carotid subclavian
bypass. Postprocedure angiography revealed a type I en-
doleak, which was then repaired with an additional device.
One patient was noted to have a persistent distal attach-
ment site leak treated with two Palmaz stents (Cordis,
Miami Lakes, Fla). One patient was found to have proximal
device collapse on serial imaging with reperfusion of a
pseudoaneurysm and was repaired successfully with an ad-
ditional device deployed across the proximal seal zone.
Among the seven patients noted to have type II en-
doleaks on follow-up imaging, three have completely re-
solved. The remaining four patients continue to be moni-
tored and are asymptomatic. One patient presented 3 years
postprocedure with evidence of a type III endoleak. She
distal seal at the first procedure. Subsequently, the patient
underwent reintervention with additional device deploy-
ment and resolution of the leak.
Freedom from reintervention. Reintervention is de-
fined as all procedures performed on patients in either
group referable to or as a consequence of the initial proce-
dure performed. Reintervention for patients undergoing
stent grafting was performed in 11 (10.5%) of 105 patients.
These included a groin exploration for lymph leak, a carot-
id-subclavian bypass for a retrograde type II endoleak, and
explantation of the device secondary to fungal infection.
0 12 24 36 48
Fig 2. Kaplan-Meier curve actuarial life-table analysis depicting intermediate (mid-term) survival between patients
with degenerative aneurysms treated with thoracic stent grafts and open-surgical repair. S.E., Standard error.
JOURNAL OF VASCULAR SURGERY
1192 Stone et al
Additional reinterventions were performed for proximal
device collapse, stent fracture/separation, an expanding
aneurysm sac, and revision of a carotid-carotid bypass. Four
patients (3.8%) required deployment of additional endo-
vascular devices for endoleak.
By comparison, 9.7% (9/93) of patients who under-
went open surgical repair required reintervention. Two
patients required periprocedural ascending aorta and arch
replacement secondary to retrograde aortic dissection from
cross-clamp sites. Three patients warranted additional sur-
gery for aneurysmal disease in contiguous aortic segments.
One patient underwent abdominal exploration during the
same hospitalization for a perforated viscus. Two patients
required re-exploration for bleeding, and an additional
patient required tracheostomy for ventilator dependence.
Kaplan-Meier actuarial freedom from reintervention at
48 months’ follow-up was 80% ? 7% in the endograft
group vs 78% ? 7% in the open-surgery group (P ? 0.73;
Fig 4). Subgroup analysis of freedom from reintervention
in degenerative pathology cases only was 86% ? 5% in the
(P ? 0.9). Freedom from reintervention among patients,
excluding rupture cases, was 88% ? 5% in the SG group vs
83% ? 6% (P ? .89).
Conventional surgical repair of descending thoracic
aneurysms in centers of excellence is associated with mor-
tality rates ranging from 4% to 9%, with paraplegia rates
documented at roughly 4% to 14%.5,11,22-24 Data from the
0 12 24 36 48
Fig 3. Kaplan-Meier curve actuarial life-table analysis comparing mid-term survival between SG patients and open-
surgical controls, excluding cases of ruptured aneurysms. S.E., Standard error.
JOURNAL OF VASCULAR SURGERY
Volume 44, Number 6
Stone et al 1193
TAG investigators’ pivotal trial comparing SG repair with
open operation for descending thoracic aneurysms, and
conducted in 16 major academic medical centers, indicate
significantly diminished mortality (2.1% vs 11.7%) and SCI
complications (3% vs 14%) for SG vs open repair.11,14 Stent
graft repair of the thoracic aorta for the treatment of the
spectrum of aortic pathologies appears to offer a feasible
and safe alternative compared with conventional re-
pair.10,11,17-19 Most reports of thoracic stent grafting offer
promising results; however, these are generally retrospec-
tive single-center experiences.1,25,26 The recent European
Collaborators on Stent-Graft Techniques for AAA and
Thoracic Aortic Aneurysm and Dissection Repair (EURO-
STAR) and United Kingdom Thoracic Endograft regis-
tries, which comprise among the largest published experi-
ence with thoracic devices, suggest that SG repair is a safe
and viable treatment strategy for patients with thoracic
aortic pathology.13 Most patients undergoing thoracic aor-
tic stent grafting have been treated outside the confines of
prospective clinical trials. To our knowledge, this report is
among the first to offer both perspective and direct com-
parison with conventional surgical repair as well as mid-
term follow-up for SG treatment. Furthermore, unlike
pivotal trial data, this study included a significant percent-
candidates and perhaps better reflects patients typically
encountered in contemporary practice.
The periprocedural mortality rate of 7.6% in this report
among patients treated with SG is consistent with other
studies, which often include patients treated in desperate
Freedom From Re-intervention
0 12 24 36 48
Fig 4. Kaplan-Meier curve actuarial analysis depicting freedom from reintervention in patients treated with stent-graft
repair and open surgery. S.E., Standard error.
JOURNAL OF VASCULAR SURGERY
1194 Stone et al
clinical circumstances and those unfit for open surgery, or
both.13,27,28 Despite borderline statistical significance, op-
erative mortality in our study was halved with SG vs open
repair; these data are further impressive because nearly 30%
of SG patients were not open surgery candidates. The TAG
investigators’ pivotal trial data of 2.1% operative mortality
clearly reflect the stringent inclusion criteria and elective
nature of such cases. With respect to open operation, the
15.1% (10% excluding ruptures) reported herein is quite
similar to the W. L. Gore pivotal trial data of 11.7% mor-
tality in the open surgery control group with intact aneu-
rysms.11,14 Recently reported data from high-volume cen-
ters are similar (8% mortality) and likely reflect a “best case
scenario” for open operation.24 Similar to the experience
with endovascular stent grafting for AAA,30 the accumulat-
ing evidence indicates a clear advantage for SG repair with
respect to operative mortality.
Actuarial survival at 48 months in our endograft cohort
was 54% ? 7%, and the corresponding figure for open
surgery patients was 64% ? 6% (P ? 0.91; Fig 1). These
data are similar to the TAG investigator’s actuarial 2-year
survival of 78% in patients treated with thoracic SGs vs 76%
in patients treated with open surgery (P ? 0.48). Likewise,
EUROSTAR and United Kingdom Thoracic Endograft
registries report a 1-year actuarial survival of 80% among
patients treated with thoracic SGs for degenerative aneu-
rysms.13 In another report documenting mid-term fol-
low-up survival, Demers et al12 report a 5-year survival of
49% ? 5% among patients treated with thoracic SGs vs 78%
? 6% for open surgery patients, yet the authors appropri-
ately emphasize that these data reflect the inclusion of
patients who were deemed reasonable surgical candidates
Our experience with a 95.2% technical success at im-
plantation rate likely reflects improved second-generation
devices.11,13,27 Currently, anatomic restrictions such as
severe thoracic aortic tortuosity, short landing and sealing
zones, and extensive mural thrombus have been limiting
factors, although a seemingly infinite variety of debranch-
ing and bypass procedures can be applied to extend either
the proximal or distal sealing zones.30,31 An important
consideration in the deployment of thoracic SGs remains
arterial access. Current SG device parameters and accom-
panying sheath diameters frequently warrant alternative
arterial access using either iliac or aortic conduits; for
example 23.8% of the SG cohort in this series required
alternative arterial access. Gore phase II data document a
15% rate of alternative access required for device deploy-
ment.11 White et al32 noted a 27% access site complication
rate. The rate of access site complications for SG patients in
this report was about 15.2% (n ? 16), and this led to a
fatality in one instance. The prevalence of such procedure-
associated complications related to thoracic aortic stent
grafting highlights the importance of performing such pro-
cedures in an operating room environment. Specific strat-
egies to both avoid and deal with access-related issues are
In any discussion of the relative merits of stent grafting
and open surgery of the thoracic aorta, consideration of
secondary interventions or reinterventions figures promi-
nently. Specifically, Demers et al12 document a rate for
actuarial freedom from reintervention of 77% ? 5% at 5
years, similar to our findings of 80% ? 7% at 48 months.
Our data challenge the tacit assumption that reinterven-
in SG patients, although this depends on how such inter-
ventions are defined. For example, we consider it legitimate
to include tracheostomy as a reintervention, even though
not specific to the surgical graft. Freedom from reinterven-
tion in our SG group approached 80% ? 7% vs 78% ? 7% in
In most series, similar to our findings, endoleak accounts
for most reinterventions in the patients treated with a
SG.11,12 By comparison, reintervention for patients under-
going open surgery often reflects persistent or recurrent
disease in contiguous aortic segments. Reintervention rates
in Gore phase II clinical trial reports document seven
reinterventions (5%) in the stent-graft cohort after 2 years
of follow-up, whereas no patients in the open surgical
cohort required reintervention at 2 years.11 Reintervention
rates will also reflect clinical decision making in treating
patients with TAAs in regard to the extent of resection. We
previously reported that 20% of our TAA cohort had resid-
ual aneurysm disease after resection and 25% of our TAA
patients had prior aortic resections. As many as 50% of
patients undergoing TAA repair will have synchronous or
metachronous aneurysm disease.5,34 The increased avail-
ability of SG treatment options further permits the oppor-
tunity to stage resection of diffuse disease, with reliance on
SG reintervention to treat more proximal aortic segments.
As this staged hybrid open and endovascular treatment
paradigm becomes more widespread, it may cloud reported
reintervention data associated with specific treatments.
There remains a paucity of late follow-up reintervention
data among SG series. It is likely that the number of
reinterventions among thoracic SG patients will increase as
longer follow-up data become available.
Accumulating evidence suggests that the risk of SCI
complications is significantly decreased in SG groups com-
pared with patients undergoing open repair.11,33 Indeed, in
the first available comparative trial of SG vs open repair for
descending thoracic aneurysms, there was a significant re-
duction in repair with SG (2.9% vs 13.8% incidence; P ?
0.003).11,13 The data reported in this study are quite
different, with an essentially equivalent SCI rate (6.7% vs.
8.6%; P ? 0.44). Previously, we reported an overall SCI
rate of 11% in a series of 337 patients undergoing thoraco-
abdominal aneurysm repair, recognizing that these patients
are at higher risk than patients with isolated aneurysm
disease of the descending thoracic aorta.5,23 The relatively
high paraplegia rate in the pivotal trial open-surgical con-
trol cohort may reflect a broader cross section of surgeons
with different backgrounds and the inconsistent or absent
application of spinal cord–protective adjuncts (although
78% were repaired with distal aortic perfusion techniques).
JOURNAL OF VASCULAR SURGERY
Volume 44, Number 6
Stone et al 1195
The modest rate of paraplegia in our cohort likely reflects a
consistent (since 1993) application of spinal cord–protec-
tive adjuncts that have been previously documented to
decrease the incidence of SCI.5,20
It has been suggested that the incidence of spinal cord
complications may correlate with long aortic segment ex-
clusion with SG devices and previous AAA repair.21,36
Despite the concern for extensive intercostal arterial exclu-
sion during SG repair, the reported incidence of neurologic
complications remains low, from 2.9% to 12%.13,21,35 Piv-
otal trial data cite a 4.7% incidence of paraplegia in patients
treated previously for AAA, and a 2% incidence among
patients without prior aortic replacement.11,14 The data
reported in this study indicate no correlation with prior
AAA repair, however the statistical power of these findings
is unreliable owing to small patient numbers. It is our
posture to incorporate spinal cord–protective adjuncts in
patients in whom long segment aortic coverage is antici-
pated or patients with previous abdominal aortic resection.
We manage these patients in an intensive care unit setting
with arterial catheters and cerebrospinal fluid drains for at
least 24 to 48 hours postprocedure, specifically to avoid
perioperative hypotension. Our open surgical patients are
routinely treated with spinal cord–protective adjuncts.
Stroke risk during SG repair remains an important
consideration, particularly in cases where arch fixation is
series was 9.5% (10/105). Among these patients, eight
(80%) had proximal device fixation in the transverse arch.
Gore phase II trial reports document 5 patients among 140
who experienced periprocedural cerebrovascular acci-
dent,11 4 of whom had a planned bypass procedure second-
ary to proximal aneurysmal disease that required device
fixation in the arch. We strongly consider the quality of the
arch fixation sites as important in clinical decision-making.
Debranching operative strategies are likely safer than fixa-
tion in diseased arch segments.
Our SG cohort had half the periprocedural mortality of
open surgery patients, despite overall higher risk. Mid-term
survival appears to be equivalent between the two groups.
Arterial access remains an important consideration, as does
scrupulous follow-up. Although long-term durability re-
patients with degenerative conditions and finite survival.
Because SG repair offers lower periprocedural risk, it is our
current practice to perform SG repair when anatomically
Conception and design: DHS, RPC
Analysis and interpretation: DHS, TKC, RPC
Data collection: DHS, MFC, TKC
Writing the article: DHS, DCB, RPC
Critical revision of the article: DHS, DCB, CJK, RPC
Final approval of the article: DHS, DCB, CJK, GML,
MFC, TKC, RPC
Statistical analysis: DHS, TKC, RPC
Overall responsibility: DHS
1. Ehrlich M, Grabenwoeger M, Cartes-Zumelzu F, et al. Endovascular
stent graft repair for aneurysms on the descending thoracic aorta. Ann
Thorac Surg 1998;66:19-24.
LJ III. Improved prognosis of thoracic aortic aneurysms: a population-
based study. JAMA 1998;280:1926-9.
3. Cambria RA, Gloviczki P, Stanson AW, et al. Outcome and expansion
Am J Surg 1995;170:213-7.
4. Juvonen T, Ergin MA, Galla JD, et al. Prospective study of the natural
history of thoracic aortic aneurysms. Ann Thorac Surg 1997;63:1533-
5. Cambria RP, Clouse WD, Davison JK, Dunn PF, Corey M, Dorer D.
Thoracoabdominal aneurysm repair: results with 337 operations per-
formed over a 15-year interval. Ann Surg 2002;236:471-9.
6. Galloway AC, Schwartz DS, Culliford AT, et al. Selective approach to
descending thoracic aortic aneurysm repair: a ten-year experience. Ann
Thorac Surg 1996;62:1152-7.
7. Hamerlijnck RP, Rutsaert RR, De Geest R, Brutel de la Riviere A,
Defauw JJ, Vermeulen FE. Surgical correction of descending thoracic
aortic aneurysms under simple aortic cross-clamping. J Vasc Surg 1989;
8. Kouchoukos NT, Dougenis D. Surgery of the thoracic aorta. N Engl
J Med 1997;336:1876-88.
9. Dake MD, Miller DC, Semba CP, Mitchell RS, Walker PJ, Liddell RP.
Transluminal placement of endovascular stent-grafts for the treatment
of descending thoracic aortic aneurysms. N Engl J Med 1994;331:
10. Cambria RP, Brewster DC, Lauterbach SR, et al. Evolving experience
with thoracic aortic stent graft repair. J Vasc Surg 2002;35:1129-36.
11. Makaroun MS, Dillavou ED, Kee ST, et al. Endovascular treatment of
thoracic aortic aneurysms: results of the phase II multicenter trial of the
GORE TAG thoracic endoprosthesis. J Vasc Surg 2005;41:1-9.
12. Demers P, Miller DC, Mitchell RS, et al. Midterm results of endovas-
cular repair of descending thoracic aortic aneurysms with first-genera-
tion stent grafts. J Thorac Cardiovasc Surg 2004;127:664-73.
13. Leurs LJ, Bell R, Degrieck Y, Thomas S, Hobo R, Lundbom J.
Endovascular treatment of thoracic aortic diseases: combined experi-
ence from the EUROSTAR and United Kingdom Thoracic Endograft
registries. J Vasc Surg 2004;40:670-9.
14. Cho JS, Haider SE, Makaroun MS. US multicenter trials of endopros-
theses for the endovascular treatment of descending thoracic aneu-
rysms. J Vasc Surg 2006;43(Suppl A):12A-19A.
15. Black JH III, Cambria RP. Contemporary results of open surgical repair
of descending thoracic aortic aneurysms. Semin Vasc Surg 2006;19:
16. Fairman RM FM, White RA. Preliminary results of the Medtronics
vascular thoracic stent graft system for patients with thoracic aortic
disease, the Valor trial: high risk/non-surgical arm. J Vasc Surg. In
17. Amabile P, Collart F, Gariboldi V, Rollet G, Bartoli JM, Piquet P.
Surgical versus endovascular treatment of traumatic thoracic aortic
rupture. J Vasc Surg 2004;40:873-9.
18. Greenberg R, Khwaja J, Haulon S, Fulton G. Aortic dissections: new
perspectives and treatment paradigms. Eur J Vasc Endovasc Surg 2003;
19. Wellons ED, Milner R, Solis M, Levitt A, Rosenthal D. Stent-graft
repair of traumatic thoracic aortic disruptions. J Vasc Surg 2004;40:
20. Cambria RP, Davison JK, Carter C, et al. Epidural cooling for spinal
cord protection during thoracoabdominal aneurysm repair: a five-year
experience. J Vasc Surg 2000;31:1093-102.
21. Gravereaux EC, Faries PL, Burks JA, et al. Risk of spinal cord ischemia
after endograft repair of thoracic aortic aneurysms. J Vasc Surg 2001;
JOURNAL OF VASCULAR SURGERY
1196 Stone et al
22. Coselli JS, LeMaire SA, Conklin LD, Adams GJ. Left heart bypass
during descending thoracic aortic aneurysm repair does not reduce the
incidence of paraplegia. Ann Thorac Surg 2004;77:1298-303.
23. Estrera AL, Miller CC III, Chen EP, et al. Descending thoracic aortic
aneurysm repair: 12-year experience using distal aortic perfusion and
cerebrospinal fluid drainage. Ann Thorac Surg 2005;80:1290-6.
24. Safi HJ, Estrera AL, Miller CC, et al. Evolution of risk for neurologic
deficit after descending and thoracoabdominal aortic repair. Ann Tho-
rac Surg 2005;80:2173-9.
25. Lepore V, Lonn L, Delle M, Mellander S, Radberg G, Risberg B.
Treatment of descending thoracic aneurysms by endovascular stent
grafting. J Card Surg 2003;18:436-43.
26. Ramaiah V, Rodriguez-Lopez J, Diethrich EB. Endografting of the
thoracic aorta. J Card Surg 2003;18:444-54.
27. Cambria RP. Invited commentary: endovascular treatment of thoracic
aortic diseases—combined experience from the EUROSTAR and
United Kingdom Thoracic Endograft registries. J Vasc Surg 2004;40:
28. Bell RE, Taylor PR, Aukett M, Sabharwal T, Reidy JF. Mid-term results
for second-generation thoracic stent grafts. Br J Surg 2003;90:811-7.
29. Orend KH, Scharrer-Pamler R, Kapfer X, Kotsis T, Gorich J, Sunder-
Plassmann L. Endovascular treatment in diseases of the descending
thoracic aorta: 6-year results of a single center. J Vasc Surg 2003;
30. Hua HT, Cambria RP, Chuang SK, et al. Early outcomes of endovas-
cular versus open abdominal aortic aneurysm repair in the National
Surgical Quality Improvement Program-Private Sector (NSQIP-PS). J
Vasc Surg 2005;41:382-9.
31. Diethrich EB, Ghazoul M, Wheatley GH III, Alpern JB, Rodriguez-
Lopez JA, Ramaiah VG. Great vessel transposition for antegrade deliv-
ery of the TAG endoprosthesis in the proximal aortic arch. J Endovasc
32. Flye MW, Choi ET, Sanchez LA, et al. Retrograde visceral vessel
revascularization followed by endovascular aneurysm exclusion as an
alternative to open surgical repair of thoracoabdominal aortic aneu-
rysm. J Vasc Surg 2004;39:454-8.
33. White RA, Donayre CE, Walot I, et al. Endovascular exclusion of
descending thoracic aortic aneurysms and chronic dissections: initial
clinical results with the AneuRx device. J Vasc Surg 2001;33:927-34.
how to avoid them. Semin Vasc Surg 2006;19:3-10.
35. Clouse WD, Marone LK, Davison JK, et al. Late aortic and graft-related
events after thoracoabdominal aneurysm repair. J Vasc Surg 2003;37:
ing thoracic aortic aneurysms: an early experience with intermediate-
term follow-up. J Vasc Surg 2000;31(1 Pt 1):147-56.
Submitted May 11, 2006; accepted Aug 1, 2006.
JOURNAL OF VASCULAR SURGERY
Volume 44, Number 6
Stone et al 1197