Original Article

Endovascular Aortic Aneurysm Repair (EVAR) Has

Significantly Lower Perioperative Mortality in
Comparison to Open Repair: A Systematic Review

Muhammad S. Sajid, Mittal Desai, Zishan Haider, Daryll M. Baker and George Hamilton,
Department of Vascular Surgery, Royal Free Hospital, Hampstead, London, UK.

OBJECTIVE: The objective of this meta-analysis was to evaluate the effectiveness of endovascular
abdominal aortic aneurysm repair (EVAR) in reducing inhospital mortality against open graft replace-
ment for aortic aneurysm.
METHODS: Generic terms including EVAR, endovascular aneurysm repair and aortic endografting were
used to search a variety of electronic databases. Based on selection criteria, decisions regarding inclusion
and exclusion of primary studies were made.
RESULTS: A total of three randomized controlled trials on 1,468 patients were included. In the EVAR
group, 12 of 759 (1.5%) patients died, compared to 33 of 709 (4.6%) patients who died in the open surgery
group. In both the fixed and random effect models, EVAR was associated with statistically significantly
lower perioperative mortality when compared to open surgical repair of aortic aneurysm. The risk ratio
of 0.33 indicates that mortality is 3.3 times more likely in the open surgery group compared to the EVAR
group.
CONCLUSION: EVAR carries a threefold lower risk of perioperative death in comparison to open repair
of abdominal aortic aneurysm. This early advantage must be offset against the increased need for later
re-intervention and probable equivalence of long-term outcome. In older and high operative risk patients,
EVAR should be the treatment of choice. [Asian J Surg 2008;31(3):119–23]

Key Words: aortic aneurysm, endograft, endovascular aneurysm repair, EVAR

Introduction rupture and its fatal consequences. The incidence of symp-

In the last six decades, open aortic surgery has remained
the treatment of choice for elective infrarenal abdominal
aortic aneurysm (AAA). Endovascular AAA repair (EVAR)
was first reported in 1991 by Parodi et al.1 It involves
inserting an endograft through the common femoral
arteries or iliac arteries and positioning it within the
aneurysm, thus excluding the aneurysm sac from circula-
tion, reducing systemic pressure on the sac and avoiding
tomatic AAA is known to increase with age, occurring
in 25 per 100,000 men aged 50 years, rising to 78 per
100,000 in those over 70 years of age.2 Minimally invasive
endovascular repair is associated with less physiological
stress on the body, more than three times less mortality,3
less morbidity and less intensive therapy unit/anaesthetic
requirements. However, apart from causing less routine
systemic complications, EVAR can cause unique technical
problems like endoleak (perigraft blood flow, sometimes

Address correspondence and reprint requests to Mr Muhammad S. Sajid, Washington Suite, North Wing, Worthing
Hospital, West Sussex BN11 2DH, United Kingdom.
E-mail: surgeon1wrh@hotmail.com ● Date of acceptance: 3 January 2008

© 2008 Elsevier. All rights reserved.

ASIAN JOURNAL OF SURGERY VOL 31 • NO 3 • JULY 2008 119

 ■ SAJID et al ■
responsible for aneurysm rupture), endograft migration
and endograft fracture. These technical problems need
further intervention after endovascular repair and are
responsible for the uncertain long-term success of this
procedure.
This study evaluated the effectiveness of EVAR in
reducing perioperative mortality by a systematic review of
available evidence in the literature.
Methods
Studies published between January 1996 and December
2007 were identified through the MEDLINE, EMBASE,
CINAHL, and COCHRANE LIBRARY databases with a
search in all languages, through review of Current Contents.
Generic terms including EVAR, endovascular aneurysm repair,
and aortic endografting were used to search the electronic
databases. The search was also performed with a combina-
tion of medical subject headings such as “mortality after
EVAR”, “complications of EVAR”, and “endovascular versus
open abdominal aortic aneurysm surgery”. Other electronic
databases covering biomedical issues, health-related arti-
cles, science and social science were also searched. Hand
searching and looking through references quoted in use-
ful articles were also carried out and yielded a few more
articles. The detailed search strategy is given in Figure 1.
Database searched Articles identified
MEDLINE 213
EMBASE 289
PubMed 517
CENTRAL 16
CINAHL 6
From bibliography 2 Abstracts reviewed = 732
Excluded = 701
Trials comparing nonsurgical techniques
Trials comparing other procedures
Trials comparing laparoscopic techniques
Trials comparing open techniques
Trials comparing endovascular devices
Reviews
Meta-analyses
Case series
Figure 1. Flow chart of the literature search.
120
Two authors to assess the eligibility for inclusion
(Table 1) critically reviewed each article. The study method
of all potentially eligible publications was reviewed.
Studies on EVAR were eligible for inclusion in the analysis
if they satisfied our predefined methodological criteria,
i.e. randomized.
Statistical analyses were performed by a senior statis-
tician using Microsoft Excel 2003 for Windows XP. Binary
data (mortality) were summarized as risk ratios (RR)
using the Mantel-Haenszel method under the fixed
effects model and the DerSimonian and Laird method
under the random effects model.4 In each case, a hetero-
geneity test was carried out to determine whether the
fixed effects model was appropriate. Forest plots were
used for the graphical display of results from this review.
Subgroup and sensitivity analyses were not feasible due
to the limited number of studies. Publication bias was
also difficult to assess due to the small number of studies
in this review.
Table 1. Inclusion criteria
• Prospective randomized controlled trial on endovascular
versus open abdominal aortic aneurysm repair
• Trial on patients of any age and sex
Duplicates = 312
Total citations =
1,044
Full text analyses = 31
Excluded = 28
Nonrandomized trials
Comparative trials
Duplicate publications
of same trial
Included trials = 3 (on 1,468 patients)
ASIAN JOURNAL OF SURGERY VOL 31 • NO 3 • JULY 2008
 ■ PERIOPERATIVE MORTALITY AFTER EVAR ■

Table 2. Characteristics of included trials

Number of patients
Trial Year Type of trial
EVAR Open surgery

Cuypers et al5 2001 Randomized 57 19

DREAM trial6 2004 Randomized 171 174
EVAR I trial3 2005 Randomized 531 516

Results Table 3. Outcome variables

Trial Perioperative mortality

Three trials3,5,6 on EVAR encompassing 1,468 patients
were retrieved from the electronic databases. The charac- Cuypers et al5
teristics of each trial are given in Table 2. There were 759 Endovascular repair 1/57
patients in the EVAR group and 709 patients in the open Open repair 1/19

surgery group. If standard deviations were not reported DREAM trial6

in the studies, then where necessary, they were estimated Endovascular repair 2/171
either from the ranges or p values. The outcome variables Open repair 8/174
extracted from these trials are shown in Table 3. In the EVAR I trial3
EVAR group, 12 of 759 (1.5%) patients died in hospital, Endovascular repair 9/531
compared to 33 of 709 (4.6%) patients who died in the Open repair 24/516
open surgery group.

Methodological quality of included studies
The methodological quality of included trials is explained
comprehensively in Table 4. The Mantel-Haenszel fixed
effects model was used to compute robustness and sus-
ceptibility to any outliers among these trials. The alloca-
tion concealment and blinding of investigator or assessor
was clearly reported; consequently, the methodological
quality of all three trials was considered adequate.
Perioperative mortality
Both in the fixed (RR, 0.33; 95% confidence interval [CI],
0.17−0.65; z = −3.23; p < 0.001) and random (RR, 0.23; 95%
CI, 0.12−0.43; z = −4.6; p = 0.0001) effects models (Figure 2),
EVAR was associated with statistically significantly lower
perioperative mortality as compared to open surgical
repair of aortic aneurysm. The RR of 0.33 indicates that
mortality was 3.3 times more likely in the open surgery
group than in the EVAR group. There was no heterogeneity
(Q = 0.16; p = 0.91) among trials.
Discussion
More than 6,000 patients die from leaking AAA every year
in the UK, which corresponds to 2% of deaths in men over
the age of 65 years. This review considered the use of
EVAR for asymptomatic AAA. Despite a general reduction
in mortality and morbidity due to cardiovascular ailments,
there is evidence of an increasing incidence of aneurysm
disease in the Western world.7 Conventional treatment of
AAA involves a major laparotomy, opening of the aneurysm
sac and replacing it with either tube or bifurcated graft.
Aneurysm repair is carried out to reduce the subsequent
risk of complications, particularly rupture, which is associ-
ated with significant mortality and morbidity. Conven-
tional treatment has been considered the treatment of
choice for more than six decades. Yet, according to the
Vascular Society’s National Vascular Database, the aver-
age postoperative mortality rate for open aneurysm repair
is 6.8% in the UK. The efficacy and safety of a new proce-
dure needs to be considered in the context of the natural
history of the condition and outcome after conventional
treatment.8 The last decade has seen a dramatic shift towards
EVAR after a giant leap in technology. With more sophis-
ticated devices, about 90% of AAA can be excluded from
systemic circulation, and the risk of subsequent rupture
has become markedly low, particularly in the early years of
follow-up. There are significant reductions in periopera-
tive mortality, postoperative pain, critical care admissions

ASIAN JOURNAL OF SURGERY VOL 31 • NO 3 • JULY 2008 121

 ■ SAJID et al ■

Table 4. Methodological quality of included trials

Quality variables Cuypers et al5 DREAM trial6 EVAR I trial3

Inclusion criteria Listed Listed Listed

Exclusion criteria Listed Listed Listed
Randomization technique Computer-based Computer-based Computer-generated
Sample size calculation Stated Stated Stated
Baseline comparable Listed Listed Listed
Masked/blinded Yes Yes Yes
Crossover Not stated Not stated No
Lost to follow-up No Stated Stated
Allocation concealment Yes Yes Yes
Analysis by intention-to-treat Yes Yes Yes

Reference Favours EVAR Favours open surgery Estimate L95% CI U95% CI

Cuypers et al
0.33 0.31 4.47
(2001)5

DREAM trial
0.25 0.20 0.93
(2004)6

EVAR I trial
0.36 0.19 0.41
(2004)3

Overall fixed
0.34 0.16 0.31
effects

Overall
0.24 0.11 0.20
random effects

0.01 0.10 1.0 10.0 100.0

Risk ratio (95% CI)

Figure 2. Perioperative mortality.

and total length of hospital stay.3 However, due to EVAR-
related unique complications such as endograft migra-
tion or occlusion, renal infarction, endoleak and higher
re-intervention rate, this procedure is still controversial
among vascular surgeons. The early advantages of EVAR
must be offset against the increased need for later re-
intervention and probable equivalence of long-term out-
come because early and late results of the UK EVAR trials
have confirmed markedly low 30-day mortality after EVAR.
This review, from the point of view of service provi-
sion, has demonstrated that EVAR carried threefold less
risk of perioperative mortality when compared with con-
ventional repair. When planning an open procedure in
a high-risk patient, it can be difficult to organize the
required critical care services and, consequently, EVAR,
with its fewer requirements, will have many advantages.
However, there is also a clear need for a highly skilled vas-
cular team, vascular theatre nurse and interventional
radiologist with appropriate facilities of technology in
order to carry out a successful EVAR. This review provides
food for thought and suggests continuing with EVAR due
to very significant early advantages. Based on the avail-
able evidence, patients, media and industry will no doubt
drive EVAR forward due to its minimal invasiveness and
a 3% improvement in early survival. However, long-term
assessment of secondary outcome measures such as
re-intervention rate, cost-effectiveness and psychological
impact on patients due to frequent scanning need further
exploration before EVAR can be accepted as a standard
procedure for the management of AAA.

122 ASIAN JOURNAL OF SURGERY VOL 31 • NO 3 • JULY 2008

 ■ PERIOPERATIVE MORTALITY AFTER EVAR ■
The cost of the procedure itself and cost of continuous
long-term surveillance following EVAR are major con-
cerns for health sector managers. Although the conclusions
regarding the cost-effectiveness of aneurysm treatment
options are time-dependent and vary by institutional per-
spective, EVAR is still considered an expensive procedure.9
Despite the cost implications of EVAR and its failure to
improve mid-term all-cause mortality10 over open AAA
repair, it is likely that the bias of both patients and sur-
geons toward this minimally invasive procedure means
that it will continue to have a significant role. Experience
and endograft developments have the potential to reduce
postoperative complications; surveillance strategies could
then be amended to reduce cost implications.
References
1. Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal
graft implantation for abdominal aortic aneurysm. Ann Vasc Surg
1991;5:491–9.
2. Yii LG. Epidemiology of abdominal aortic aneurysm in an Asian
population. ANZ J Surg 2003;73:393–5.
ASIAN JOURNAL OF SURGERY VOL 31 • NO 3 • JULY 2008
3. Greenhalgh RM, Brown LC, Kwong GP, et al, for EVAR trial
participants. Comparison of endovascular aneurysm repair with
open repair in patients with abdominal aortic aneurysm (EVAR
I trial), 30-day mortality results: randomised controlled trial.
Lancet 2004;364:843–8.
4. Egger M, Smith GD, Altman DG. Systematic Reviews in Healthcare.
London: BMJ Publishing, 2006.
5. Cuypers PWM, Gardien M, Buth J, et al. Randomised study com-
paring cardiac response in endovascular and open abdominal
aortic aneurysm repair. Br J Surg 2001;88:1059–65.
6. Prinssen M, Verhoven EL, Buth J, et al. A randomised trial
(DREAM) comparing conventional and endovascular repair of
abdominal aortic aneurysm. N Engl J Med 2004;351:1607–18.
7. Filipovic M, Goldacre MJ, Roberts SE, et al. Trends in mortality
and hospital admission for abdominal aortic aneurysm. Br J Surg
2005;92:968–75.
8. The Vascular Society of Great Britain and Ireland. Fourth National
Vascular Database Report 2004. Oxfordshire: Dendrite Clinical
Systems, 2005. [ISBN 1-903968-12-7]
9. Jonk YC, Kane RL, Lederle FA, et al. Cost-effectiveness of abdomi-
nal aortic aneurysm repair: a systematic review. Int J Technol Assess
Health Care 2007;232:205–15.
10. Brooks MJ, Brown LC, Greenhalgh RM. Defining the role of
endovascular therapy in the treatment of abdominal aortic
aneurysm: results of a prospective randomized trial. Adv Surg
2006;40:191–204.
123