AD

ULT

CA

RD

IAC

In Off-Pump Surgery, Skeletonized GastroepiploicArtery is Superior to Saphenous Vein in PatientsWith Bilateral Internal Thoracic Arterial GraftsTomoaki Suzuki, MD, PhD, Tohru Asai, MD, PhD, Keiji Matsubayashi, MD, PhD,Atsushi Kambara, MD, PhD, Takeshi Kinoshita, MD, Noriyuki Takashima, MD, andSo Hosoba, MD

Department of Cardiovascular Surgery, Shiga University of Medical Science, Shiga, Japan

Background. When the bilateral internal thoracic arter-ies are grafted to the left coronary arteries, it remainscontroversial whether the better conduit is provided bygrafting the saphenous vein graft (SVG) or the gastroepi-ploic artery (GEA) to the right coronary artery territory.From the beginning of the present study, we consistentlyused the GEA in a skeletonized fashion using ultrasoundscissors.

Methods. From January 2002 to December 2009, 320consecutive patients with triple-vessel disease under-went in situ bilateral internal thoracic artery grafting tothe left coronary arteries. Among the 320 patients, ofwhom 229 underwent GEA grafting to the right coronaryartery and 91 SVG grafting, 85 propensity score-matchedpairs were identified (C statistic, 0.68 [p < 0.001]).

Results. The mean follow-up duration was 5.1 � 2.2

years. Seven-year freedom from death from all causes

Surgery, Shiga University of Medical Science, Setatsukinowa, Otsu, Shiga520-2192, Japan; e-mail: suzukit@belle.shiga-med.ac.jp.

© 2011 by The Society of Thoracic SurgeonsPublished by Elsevier Inc

was 96.0% in the GEA group and 82.2% in the SVG group(p � 0.03); the rate of freedom from cardiac events(cardiac death, myocardial infarction, angina pectoris,repeat intervention, and heart failure) was 89.3% in theGEA group and 77.5% in the SVG group (p � 0.048).Multivariate Cox proportional hazard regression analysisshowed that SVG use (without GEA) (p � 0.04; hazardratio, 0.31; 95% confidence interval, 0.11 to 0.94) andsmoking history (p � 0.02; hazard ratio, 0.22; 95% confi-dence interval, 0.07 to 0.81) were independent predictorsof late cardiac event.

Conclusions. Skeletonized GEA grafting to the rightcoronary artery system is better than SVG grafting inpatients with left-sided bilateral internal thoracic arterialgrafts.

(Ann Thorac Surg 2011;91:1159–64)

© 2011 by The Society of Thoracic Surgeons

The use of the bilateral internal thoracic arteries (ITA)rather than a single ITA provides better long-term

survival and reduces the recurrence of cardiac events[1–4]. When the bilateral ITAs are grafted to the leftcoronary arteries, however, it remains controversial as towhich is the better conduit: the saphenous vein graft(SVG) or the gastroepiploic artery (GEA) to the rightcoronary artery (RCA) territory. The SVG is the mostpopular and easily handled graft to the RCA. The GEAwas first described in 1987 [5, 6], but it has not gainedworldwide acceptance among cardiovascular surgeons.The reasons for the less frequent use of GEA includeconcerns of insufficient flow capacity and vasospasm, theneed to open the abdomen, and competitive flow causinggraft failure. Some studies have reported the long-termpatency of GEA to be similar to that of SVG. However, inthese reports, the GEA was used as a pedicled graft. Wehave consistently used the GEA in a skeletonized fashionusing ultrasound scissors. The early functional patencyrate of the skeletonized GEA is reported to be better thanthat of the pedicled GEA. In the present report, we

Accepted for publication Dec 16, 2010.

Address correspondence to Dr Suzuki, Department of Cardiovascular

compare the long-term clinical outcome of patients un-dergoing left-sided bilateral ITAs, grafting between pro-pensity-matched groups, in which either the SVG or thein situ skeletonized GEA was used to graft to the distalright coronary artery.

Patients and Methods

All patients previously had granted permission for use oftheir medical records for research purposes. The institu-tional review board approved the present study. FromJanuary 2002 to December 2009, 770 consecutive patientsunderwent isolated off-pump coronary artery bypasssurgery (OPCAB) at our institute by a single surgeon(T.A.) Of these, 320 patients with triple-vessel diseaseunderwent in situ bilateral ITA grafting to the left coro-nary arteries, with either the in situ GEA or the SVGanastomosed to the RCA system. Of the patients, 229underwent GEA grafting to the RCA and 91 SVG graft-ing. We calculated a propensity score for GEA grafting orSVG grafting to obtain one-to-one matched pairs withsimilar clinical characteristics. Logistic regression withbackward selection was performed to calculate the pro-pensity score based on the following patient characteris-

tics that might affect the surgeon’s decision on operative

0003-4975/$36.00doi:10.1016/j.athoracsur.2010.12.031

Tp

1160 SUZUKI ET AL Ann Thorac SurgSKELETONIZED GEA IS BETTER THAN SVG FOR RCA 2011;91:1159–64

AD

ULT

CA

RD

IAC

strategy: age, sex, body mass index, smoking history,diabetes mellitus, peripheral arterial disease, previouscerebrovascular disease, chronic kidney disease, NewYork Heart Association functional class, history of myo-cardial infarction, left ventricular ejection fraction, signif-icant left main trunk stenosis, intra-aortic balloon pumpuse, and urgency. By matching the propensity score, 85pairs were successfully matched in a one-to-one manner.We compared the clinical results between these propen-sity-matched groups.

We performed coronary artery bypass grafting (CABG)by using the off-pump technique in all CABG cases with noexclusion criteria. Patients with acute myocardial infarctionwere included, but patients who had undergone a salvageprocedure were excluded from the study. We adopted a“functional definition” of the completeness of revascular-ization in the study. “Functional” completeness was de-fined as the presence of at least one graft to all diseasedarterial territories (stenosis � 50%). Postoperative renalfailure was defined as the requirement for hemodialysis.Postoperative stroke was defined as a new neurologic eventpersisting for more than 24 hours after onset and wasconfirmed by computed tomography. Follow-up wasachieved by direct communication with the patient, thepatient’s family, or the attending physician. Significant leftmain coronary artery disease was defined as left maincoronary artery with stenosis greater than 50%, assessedvisually by the physician performing the coronaryangiography.

Anesthetic and Surgical TechniquesOur anesthetic and surgical techniques were performedas previously described [7]. All procedures were per-formed through a median sternotomy. The conduits(both internal thoracic arteries, the right gastroepiploicartery, and the saphenous vein) were harvested andskeletonized. We used the bilateral internal thoracicarteries routinely for three-vessel disease patients. Acommon combination for internal thoracic artery graftplacement was in situ grafting of the left internal thoracicartery to the circumflex area and of the right internalthoracic artery to the left anterior descending coronaryarea. We also preferentially used the skeletonized rightgastroepiploic artery to reconstruct the distal right coro-nary artery as an in situ graft.

At our institution, the GEA has been consistently usedsince 2002 in a skeletonized fashion using an ultrasoundscalpel (Harmonic Scalpel [Ethicon Endo-Surgery, Cin-cinnati, OH]) [8]. In the early stage of the study, the GEAwas used preferentially to reconstruct target coronaryarteries with proximal stenosis of less than 70%, but inthe later stage, we used the GEA only when the degree ofproximal stenosis of the target coronary artery was 70%or more because of concerns regarding competitive flow.During the operation, the size, pulsation, and calcifica-tion of the GEA were checked, and the final decision onits use was made. After harvesting the GEA, the distalend of the graft is divided, papaverine hydrochloridesolution is instilled into it, and a hemoclip is applied. The

skeletonized GEA is then wrapped in a sponge soaked in r

papaverine hydrochloride. With the papaverine hydro-chloride preparation, the GEA becomes a maximallydilated, spasm-free arterial conduit.

We used a suction-type mechanical stabilizer (Octopus4.3 [Medtronic, Minneapolis, MN]) to immobilize thetarget coronary artery, but we did not use heart position-ing. An intracoronary shunt tube and CO2 blower wereroutinely used. The distal anastomosis was constructedwith 7-0 polypropylene using a standard technique. Ared blood cell-saving device was used in all cases.

Statistical AnalysisData are presented as the mean � standard deviation.Categorical variables were analyzed using the �2 test orFisher’s exact test. Continuous variables were examinedusing the Student’s t test, or the Mann–Whitney U test.

We calculated a propensity score for GEA grafting orSVG grafting to obtain one-to-one matched pairs withsimilar clinical characteristics. Logistic regression withbackward selection was performed to calculate the pro-pensity score. The C statistic was calculated to assess thediscriminatory ability of the model. Each patient withSVG grafting was then matched to 1 patient with GEAgrafting using propensity scores identical to within 3%.

Univariate and multivariate Cox proportional hazardregression analyses were performed for the analysis oflate mortality and cardiac events. The multivariate anal-yses were performed with a stepwise forward regressionmodel in which each variable with a probability value ofless than 0.20 in the univariate analysis, which wasentered into the model. Actuarial survival and event-freesurvival curves were estimated using the Kaplan-Meiermethod and differences between groups were comparedusing the log-rank test. Calculated p values of less than0.05 were considered significant. Data were analyzedusing SPSS 11.5.1 (SPSS Inc, Chicago, IL) for Windows(Microsoft Corp, Redmond, WA).

Results

The preoperative characteristics of the patients are sum-marized in Table 1. Preoperative patient comorbiditiesand cardiac characteristics were equally distributed inthe two matched groups. The discriminatory ability of thelogistic model as measured by the C statistic was 0.68(p � 0.001), whereas the Hosmer-Lemeshow goodness-of-fit test was not statistically significant (p � 0.75),indicating good discriminative power and acceptablecalibration of the model, respectively.

Short-Term ResultsThe distribution of grafts and outcome after surgery areshown in Tables 2 and 3. There were no significantdifferences in the number of grafts per patient (4.09 � 2.1in GEA group vs 4.16 � 2.4 in SVG group; p � 0.26) or therate of achievement of complete revascularization (100%in the GEA group vs 98.8% in the SVG group; p � 0.5).

here was no significant difference in other morbidities:erioperative myocardial infarction (0% vs 1.2%), bleeding

eoperation (2.4% vs 1.2%), cerebrovascular accident (0% vs

3is1

HHDICPC

HCTPLPUI

d

d

1161Ann Thorac Surg SUZUKI ET AL2011;91:1159–64 SKELETONIZED GEA IS BETTER THAN SVG FOR RCA

AD

ULT

CA

RD

IAC

0%), acute renal failure (0% vs 2.4%), prolonged ventilatorsupport for respiratory insufficiency (� 24 hours; 2.4% vs.5%), and deep sternal infection (0% vs 2.4%). One patientn the GEA group (1/85; 1.2%) died within 30 days afterurgery of pneumonia and 1 patient in the SVG group (1/85;.2%) died of multisystem organ failure.

Long-Term ResultsThe mean follow-up duration was 5.1 � 2.2 years. Seven-year freedom from death from all causes was 96.0% in theGEA group and 82.2% in the SVG group (p � 0.03); therate of freedom from the combined end point of cardiacdeath, myocardial infarction, angina pectoris, repeat cor-onary intervention, and heart failure requiring treatmentwere 89.3% in the GEA group and 77.5% in the SVGgroup (p � 0.048) (Figs 1, 2).

Multivariate Cox proportional hazard regression anal-ysis showed that SVG use (without GEA) (p � 0.04;hazard ratio, 0.31; 95% confidence interval, 0.11 to 0.94)and smoking history (p � 0.02; hazard ratio, 0.22; 95%confidence interval, 0.07 to 0.81) were independent pre-dictors of late cardiac event (Table 4).

Comment

Bilateral internal thoracic artery grafting to the left ante-rior descending coronary artery and circumflex coronary

Table 1. Preoperative Patient Characteristics in Propensity-Matched Groups

CharacteristicGEA Group SVG Group p

Value(n � 85) (n � 85)

Age (mean � SD) 66.9 � 9.6 66.9 � 10.1 0.60BMI 24.0 � 5.6 23.6 � 5.4 0.37Female gender 12 (14%) 11 (13%) 0.82Smoking history 46 (54%) 41 (48%) 0.44

ypertension 63 (74%) 59 (69%) 0.31yperlipidemia 45 (53%) 42 (49%) 0.64iabetes mellitus 32 (38%) 32 (38%) 1.0

nsulin dependent 15 (18%) 16 (19%) 0.84OPD 16(19%) 21 (17%) 0.36revious stroke 10 (12%) 8 (9%) 0.62hronic renal failure(Cre � 1.5)

22 (26%) 22 (26%) 1.0

emodialysis 7 (8%) 11 (13%) 0.32ongestive heart failure 12 (14%) 8 (9%) 0.34hree-vessel disease 85 (100%) 85 (100%) 1.0revious myocardial infarction 31 (36%) 32 (38%) 0.87VEF � 40% 7 (8%) 10 (12%) 0.44revious PCI 27 (31%) 30 (35%) 0.34rgency 10 (12%) 12 (14%) 0.65

ABP use 5 (6%) 4 (5%) 0.50

BMI � body mass index; COPD � chronic obstructive pulmonaryisease; Cre � serum creatine; GEA � gastroepiploic artery;

IABP � intra-aortic balloon pumping; LVEF � left ventricular ejectionfraction; PCI � percutaneous coronary intervention; SD � stan-

ard deviation; SVG � saphenous vein graft.

artery offers the best long-term survival and lowest rates

of reintervention in patients undergoing coronary arterybypass grafting surgery [1–4]. In the decade since Buxtonand coworkers [9], and Lytle and coworkers [4] revealedthe long-term efficacy of bilateral ITA grafting, it hasbeen gaining acceptance among surgeons, and there isno doubt that it affords the best long-term outcome.Furthermore, CABG with grafting of the bilateral ITAs tothe left coronary system and the GEA to the distal RCAhas been reported to provide good long-term outcome [1,10, 11].

When bilateral ITAs are grafted to the left coronaryarteries, it remains controversial whether or not thebetter conduit is provided by grafting the saphenous veinor the GEA to the RCA territory. The saphenous vein isthe most popular and easily handled graft to the RCAand the relevant long-term, clinical outcome, flow capac-ity, patency, and long-term complications are wellknown. The efficacy of grafting with the GEA in patientsundergoing bilateral ITA grafting has not yet been clearlyelucidated. The GEA was first described in 1987 [5, 6], buthas not gained worldwide acceptance among cardiovas-cular surgeons. Reasons for its less frequent use includeconcerns of insufficient flow capacity and vasospasm, andcompetitive flow causing graft failure.

Lev-Ran and colleagues [12] reported that GEA graft-ing to the RCA system in left-sided bilateral ITA graftingshowed no particular benefit after a 4-year follow-up,and Pevni and colleagues [13] reported the same conclu-sion after a 6-year follow-up. Glineur and coworkers [14]presented a prospective, randomized clinical and angio-graphic midterm evaluation that compared the use of theSVG and the GEA for grafting to the RCA at 6 monthsand 3 years, but found no advantage of the GEA incomparison with the SVG. Many reports have thus failedto demonstrate the benefits of the GEA in comparisonwith the SVG. However, in these studies the GEA graftwas used in a pedicled fashion.

The early functional patency rate of the skeletonizedGEA is reported to be better than that of the nonskel-etonized GEA [15]. Kim and coworkers [16] evaluated theearly and 1-year postoperative results of grafting theskeletonized GEA to the RCA and found an excellent

Table 2. Distribution of Grafts

GEA Group (n � 85) SVG Group (n � 85)

RITA LITA GEA RITA LITA SVG

LAD 83 2 0 79 6 0Diag 3 22 0 2 39 9OM 2 60 0 6 53 4PL 0 46 8 0 30 13PDA 0 0 122 0 0 113Sequential

grafting3 41 38 2 43 47

Diag � diagonal branch; GEA � gastroepiploic artery; LAD � leftanterior descending coronary artery; LITA � left internal thoracicartery; OM � obtuse marginal branch; PDA � posterior descend-

ing coronary artery; PL � posterolateral branch; RITA � rightinternal thoracic artery; SVG � saphenous vein graft.

spltntG

y;

1162 SUZUKI ET AL Ann Thorac SurgSKELETONIZED GEA IS BETTER THAN SVG FOR RCA 2011;91:1159–64

AD

ULT

CA

RD

IAC

early patency rate of 98.3% and 1-year patency rate of92.0%. To avoid flow competition, it has been recom-mended that the GEA not be used as an in situ graft if theinner diameter at the anastomotic site is less than 1.5 mm[17]. Ochi and associates [18] also recommended that theGEA should have a large luminal diameter (2 to 3 mm) atits anastomotic point to generate adequate perfusionpressure and thereby avoid flow competition betweenthe GEA and the native coronary artery. Having used theGEA in a skeletonized fashion since early times, Gagli-ardotto and coworkers [19] claimed many advantages,uch as avoidance of early spasm, easy identification ofotential bleeding, quality of the vessel, functionally

engthened and larger graft, ease in performing sequen-ial anastomosis, and preservation of lymphatic and ve-ous drainage to the stomach. They further emphasize

hat the observed advantage of using the skeletonizedEA is primarily graft size. From the beginning of the

Table 3. Operative and Postoperative Data

GEA Group

Operation time (min) 239 � 5No. of distal anastomoses 4.09 � 2Complete revascularization 85 (100BITA use 85 (100GEA use 85 (100All arterial reconstruction 85 (100Prolonged ventilation (� 24 h) 2 (2%)ICU stay (h) 17.0 � 1Reoperation for bleeding 2 (2%)Deep sternal infection 2 (0%)Permanent stroke 0 (0%)Perioperative myocardial infarction 0 (0%)Atrial fibrillation 19 (22%Renal failure requiring dialysis 0 (0%)Mortality (30 days) 1 (1%)

BITA � bilateral internal thoracic arteries; GEA � gastroepiploic artergraft.

Fig 1. Seven-year freedom from death from all causes was 96.0% inthe gastroepiploic artery (GEA) group and 82.2% in the saphenous

vein graft (SVG) group.

present study, we also consistently used the GEA in askeletonized fashion using ultrasound scissors. Our tech-nique for harvesting the skeletonized GEA using anultrasonic scalpel was previously described and hasproved to be very simple and safe [8]. In our experience,skeletonization of the whole GEA takes only 7 to 13minutes and does not increase operating time. Removalof the surrounding tissue makes the GEA longer andwider, so that it can be anastomosed at a more proximalposition than the pedicled GEA. The muscular regulatorsthat cause vasospasm are located at the two ends of theartery [20]. As the smaller distal portion is more suscep-tible to spasm, proximal anastomosis has the advantageof leaving the GEA spasm-free. Moreover, if the GEA isnot maximally dilated when performing anastomosis, the

85) SVG Group (n � 85) p Value

252 � 63 0.284.16 � 2.4 0.2684 (99%) 0.585 (100%) 1.00 (0%) . . .0 (0%) . . .3 (4%) 0.5

15.3 � 16.6 0.391 (1%) 0.52 (2%) 0.240 (0%) 1.01 (1%) 0.5

13 (15%) 0.242 (2%) 0.241 (1%) 1.0

ICU � intensive care unit; No. � number; SVG � saphenous vein

Fig 2. Seven-year actuarial freedom from any cardiac event (eg, car-diac death, myocardial infarction, angina pectoris, repeat coronaryreintervention, and heart failure) were 89.3% in the gastroepiploicartery (GEA) group and 77.5% in the saphenous vein graft (SVG)

(n �

9.1%)%)%)%)

9.1

)

group.

p

CCiass

1163Ann Thorac Surg SUZUKI ET AL2011;91:1159–64 SKELETONIZED GEA IS BETTER THAN SVG FOR RCA

AD

ULT

CA

RD

IAC

anastomotic lesion will be narrow when the GEA dilatesafter surgery, which may cause graft trouble, such asocclusion and reduced flow. Therefore, we believe thatintraoperative maximal dilatation and maximal lengthmay make an outstanding contribution to better long-term clinical outcome.

There are a number of clinical studies comparing theuse of the GEA with the SVG for grafting to the RCA inpatients with bilateral ITA grafting to the left coronaryartery system. Lev-Ran and coworkers [12] and Esaki andcoworkers [21] compared the use of the GEA and theSVG for grafting to the RCA in patients undergoingbilateral ITA grafting to the left coronary arteries with4-year and 7-year follow-ups, respectively. They similarlyconcluded that GEA grafting to the RCA did not signifi-

Table 4. Univariate and Multivariate Cox ProportionalHazard Regression Analyses of Late Cardiac Eventsa

VariableHazardRatio

95% ConfidenceInterval

pValue

Univariate modelAge 0.7 0.4–1.3 0.26Female gender 1.1 1.1–2.0 0.07Smoking history 3.5 1.1–11.1 0.02Hypertension 0.9 0.3–3.0 0.67Hyperlipidemia 1.1 0.4–3.0 0.87Diabetes mellitus 0.5 0.1–1.5 0.16Insulin-dependent 1.4 0.4–4.7 0.37COPD 1.1 0.3–3.7 0.52Previous stroke 0.5 0.1–4.1 0.43Chronic renal failure

(Cre � 1.5)1.2 0.4–3.7 0.46

LMCA 0.7 0.4–1.3 0.27Congestive heart failure 1.7 0.4–6.6 0.32Previous myocardial

infarction1.2 0.4–3.3 0.71

LVEF � 40% 3.3 0.9–11.6 0.07Previous PCI 1.2 0.7–2.2 0.51Emergency 0.3 0.04–2.5 0.22IABP use 1.1 1.1–1.2 0.37Atrial fibrillation 1.9 0.6–6.0 0.19SVG use (without GEA) 2.6 0.9–7.8 0.06

(Multivariate model)SVG use (without GEA) 0.3 0.1–0.9 0.04Smoking history 0.2 0.07–0.9 0.02Female gender NSAtrial fibrillation NSDiabetes mellitus NSLVEF � 40% NS

a Cardiac events, such as cardiac death, myocardial infarction, anginaectoris, repeat coronary intervention, and heart failure.

OPD � chronic obstructive pulmonary disease; Cre � serum creatine;RF � chronic renal failure; GEA � gastroepiploic artery; IABP �

ntra-aortic balloon pump; LMCA � left main coronaryrtery; LVEF � left ventricular ejection fraction; NS � notignificant; PCI � percutaneous coronary intervention; SVG �aphenous vein graft.

cantly improve late outcomes in patients with bilateral

ITA grafting to the left coronary artery system. However,these studies did not use risk-adjusted models and thepatient characteristics differed between groups. Ourstudy, which compared 85 propensity score-matchedpairs, is the first to use such a technique to compare theuse of the GEA with the use of the SVG in grafting to theRCA with reference to long-term clinical outcome inpatients with bilateral ITA grafting to the left coronaryartery system. Although our study population was notlarge, the statistical power and reliability are muchgreater than in previous studies. The present study foundthe 7-year survival rate for all causes of death to be 96.0%in the GEA group and 82.2% in the SVG group, repre-senting a significant difference (p � 0.03). Freedom fromany cardiac event (eg, cardiac death, myocardial infarc-tion, angina pectoris, percutaneous coronary interven-tion, cardiac failure) was 89.3% in the GEA group and77.5% in the SVG group (p � 0.04). Multivariate Coxproportional hazard regression analysis showed thatSVG use (without GEA) (p � 0.04; hazard ratio, 0.31; 95%confidence interval, 0.11 to 0.94) and smoking history (p �0.02; hazard ratio, 0.22; 95% confidence interval, 0.07 to0.81) were independent predictors of late cardiac event.

The difference in clinical result between the two groupsis small, the reason for which is presumably that bilateralITA grafting to the left coronary arteries is a much moredecisive factor than the choice between GEA and SVG forgrafting to the RCA. However, although the impact maybe small, we were able to demonstrate the superiority ofskeletonized GEA grafting in comparison with SVGgrafting to the RCA using the propensity-matchedmodel.

A limitation of the present study is that it is a nonran-domized, retrospective study with a small study popula-tion. On the other hand, all of the procedures wereconsecutive and conducted within a single institution bya single surgeon. In addition, comparison of preoperativedemographic and risk factors demonstrated that the twocohorts were well matched.

We conclude that skeletonized gastroepiploic arterygrafting to the right coronary artery system is superior tosaphenous vein grafting in patients with left-sided bilat-eral internal thoracic artery grafting.

References

1. Endo M, Nishida H, Tomizawa Y, Kasanuki H. Benefit ofbilateral over single internal mammary artery grafts formultiple coronary artery bypass grafting. Circulation 2001;104:2164–70.

2. Rizzoli G, Schiavon L, Bellini P. Does the use of bilateralinternal mammary artery (IMA) grafts provide incrementalbenefit relative to the use of a single IMA graft? A meta-analysis approach. Eur J Cardiathorac Surg 2002;22:781–6.

3. Taggart DP, D’Amico R, Altman DG. Effect of arterial revas-cularization on survival: a systematic review of studiescomparing bilateral and single internal mammary arteries.Lancet 2001;358:870–5.

4. Lytle BW, Blackstone EH, Loop FD, et al. Two internalthoracic artery grafts are better than one. J Thorac Cardio-

vasc Surg 1999;117:855–72.

1164 SUZUKI ET AL Ann Thorac SurgSKELETONIZED GEA IS BETTER THAN SVG FOR RCA 2011;91:1159–64

AD

ULT

CA

RD

IAC

5. Pym J, Brown PM, Charrete EJ, Parker JO, West RO. Gastro-epiploic-coronary anastomosis. A viable alternative bypassgraft. J Thorac Cardiovasc Surg 1987;94:256–9.

6. Suma H, Fukumoto H, Takeuchi A. Coronary artery bypassgrafting by utilizing in situ right gastroepiploic artery: basicstudy and clinical application. Ann Thorac Surg 1987;44:394–7.

7. Suzuki T, Asai T. Early and midterm results of off-pumpcoronary artery bypass grafting without patient selection.Heart Surgery Forum 2008;11:196–200.

8. Asai T. Technique and results for skeletonized GEA usingthe Harmonic Scalpel in combination with other arterialgrafts in off-pump coronary artery bypass surgery. In: HeGW(ed). Arterial grafting for coronary artery bypass sur-gery. 2nd ed. Berlin: Springer-Verlag; 2006:196–200.

9. Buxton BF, Komeda M, Fuller JA, Gordon I. Bilateral internalthoracic artery grafting may improve outcomes of coronaryartery surgery, risk-adjusted survival. Circulation 1998;98:II-1–6.

10. Chavanon O, Durand M, Hacini R, et al. Coronary arterybypass grafting with left internal mammary artery and rightgastroepiploic artery, with and without bypass. Ann ThoracSurg 2002;73:499–504.

11. Tavilla G, Kappetein AP, Braum J, Gopie J, Tjien ATJ, DionRAE. Long-term follow-up of coronary artery bypass graft-ing in three-vessel disease using exclusively pedicled bilat-eral internal thoracic and right gastroepiploic arteries. AnnThorac Surg 2004;77:794–9.

12. Lev-Ran O, Mohr R, Uretzky G, et al. Graft of choice to rightcoronary system in left-sided bilateral internal thoracic ar-tery grafting. Ann Thorac Surg 2003;75:88–92.

13. Pevni D, Uretzky G, Yosef P, et al. Revascularization of right

Fifty-Eighth Annual Meeting—

via a link on the STSA website at www.stsa.org. Abstracts

© 2011 by The Society of Thoracic SurgeonsPublished by Elsevier Inc

14. Glineur D, Hanet C, Poncelet A, et al. Comparison ofsaphenous vein graft versus right gastroepiploic artery torevascularize the right coronary artery: A prospective ran-domized clinical, functional, and angiographic midtermevaluation. J Thorac Cardiovasc Surg. 2008;136:482–8.

15. Kamiya H, Watanabe G, Takemura H, Tomita S, NagamineH, Kanamori T. Skeletonization of gastroepiploic artery graftin off-pump coronary artery bypass grafting: early andangiographic assessment. Ann Thorac Surg 2004;77:2046–50.

16. Kim KB, Cho KR, Choi JS, Lee HJ. Right gastroepiploic arteryfor revascularization of right coronary territory in off-pumptotal arterial revascularization: strategies to improve pa-tency. Ann Thorac Surg 2006;81:2135–41.

17. Shimizu T, Suesada H, Cho M, Ito S, Ikeda K, Ishimaru S.Flow capacity of gastroepiploic artery versus vein grafts forintermediate coronary artery stenosis. Ann Thorac Surg2005;80:124–30.

18. Ochi M, Hatori N, Fujii M, Saji Y, Tanaka S, Honma H.Limited flow capacity of the right gastroepiploic artery graft:postoperative echocardiographic and angiographic evalua-tion. Ann Thorac Surg 2001;71:1210–4.

19. Gagliardotto P, Coste P, Lazerg M, Dor V. Skeletonized rightgastroepiploic artery used for coronary artery bypass graft-ing. Ann Thorac Surg 1998;66:240–2.

20. He GW. Clinical physiology and related biological char-acteristics. In: He GW (ed). Arterial grafting for coronaryartery bypass surgery. 2nd ed. Berlin: Springer-Verlag;2006:24 –30.

21. Esaki J, Koshiji T, Okamoto M, Tsukashita M, Ikuno T,Sakata R. Gastroepiploic artery grafting does not improve

coronary artery in bilateral internal thoracic artery grafting.Ann Thorac Surg 2005;79:564–9.

the late outcome in patients with bilateral internal thoracicartery grafting. Ann Thorac Surg. 2007;83:1024–9.

Southern Thoracic Surgical Association:

Call for Abstracts

You are invited to submit abstracts and surgical motionpictures for the Southern Thoracic Surgical Association(STSA) Fifty-Eighth Annual Meeting to be held at the JWMarriott San Antonio Hill Country Resort in San Antonio,TX, November 9–12, 2011.

To submit an abstract, access the online submission site

must be submitted by Monday, April 11, 2011 at 11:59 PM,Eastern Time. Accepted abstracts will be presented at theSTSA Fifty-Eighth Annual Meeting as oral presentations orsurgical videos.

Please direct any questions regarding abstract submis-sion to Katie Bochenek, STSA Affiliate Manager, at

kbochenek@stsa.org or call 312-202-5835.

Ann Thorac Surg 2011;91:1164 • 0003-4975/$36.00