Key Words: Arterial graft; Coronary artery bypass grafting; Gastroepiploic artery;
Ischemic heart disease; Myocardial infarction

The right gastroepiploic artery (GEA) has been used for coronary artery bypass grafting (CABG) in the past 2 decades since its successful clinical application by Pym et al[1] and Suma et al[2] in 1987. Although the GEA graft is primarily harvested by detaching it from the greater curvature of the stomach as a pedicle with surrounding tissues, skeletonized GEA harvesting was introduced by Gagliardotto et al[3] in 1998. This differs from classic pedicle harvesting in the removal of the satellite veins and surrounding tissue together with the main trunk of the GEA. The method prolongs the usable length of GEA and makes it easier to handle the GEA for the anastomosis to the coronary artery. It may also optimize the response to intraluminal vasodilator injection and decrease the incidence of graft spasm. We have been using the skeletonization technique as a routine procedure for GEA harvest since 2000 and report our midterm results for skeletonized GEA grafts.

Methods
Patients
In a 6-year period from 2000 to 2006, the skeletonized GEA has been used in 223 patients at the Cardiovascular Institute Hospital, Tokyo, and Hayama Heart Center, Kanagawa, Japan (208 males, 15 females; mean age 64 years [range 38-76 years]). The extent of coronary artery disease was 1-vessel disease in 1 patients, 2-vessel disease in 28 patients, 3-vessel disease in 122 patients and left main trunk disease in 72 patients; 118 patients (53%) had a previous myocardial infarction with a mean ejection fraction of 53.1% (Table 1)

Surgical Procedure
The skeletonized GEA is harvested using the Harmonic Scalpel with a coagulating shears tip. The anterior layer of the greater omentum is divided and the GEA is exposed along its entire length. The small omental and gastric branches of the GEA are divided. The distal end of the GEA graft is then divided and clipped, and papaverine solution is infused intraluminally to relieve spasm of the GEA. The skeletonized GEA is then wrapped in a papaverine-soaked sponge. The papaverine preparation ensures that the GEA becomes a maximally dilated, relaxed arterial conduit. The skeletonized GEA is brought anteriorly to the pylorus and introduced into the pericardial cavity through the diaphragm (Fig 1).
Off-pump CABG was used in 134 patients (60%). GEA was used as a single in-situ graft in 154 patients. Free, composite and sequential grafts were used in 33, 36 and 56 patients, respectively. The sites of GEA grafting were 1 anterior descending, 10 diagonal, 97 circumflex and 185 right coronary arteries (Table 2). Concomitantly used grafts were internal thoracic artery (ITA) in 217 patients (bilateral ITA in 32 patients), radial artery in 73 patients and saphenous vein in 53 patients. The mean number of distal anastomosis was 3.5 (range 1-6) and 3.1 coronary arteries were bypassed with arterial grafts (Table2). All patients were routinely followed up for at least 6 months after operation. Further follow-up data were obtained by telephone or mail questionnaire or from the relevant cardiologists. Graft patency was assessed by conventional catheter coronary angiography (CAG) or multidetector computer tomography (MDCT).

Results
There were 1 (0.4%) early and 3 (1.3%) late deaths (stroke, renal failure and sudden death in 1 patient each). New Q waves were noted in 2 patients (0.9%). Intra-aortic balloon pump was used in 9 patients (4%), mostly those with combined ventriculoplasty for ischemic cardiomyopathy. Postoperative angiography was performed in the early postoperative period (within 1 month) in 189 patients, and the patency rate of the GEA graft was 97.6% (242/248 distal anastomoses). The early patency rates of the GEA graft in each anastomotic sites were 100% (1/1) to the anterior descending, 100% (10/10) to the diagonal, 97.5% (79/81) to the circumflex, and 97.4% (152/1 56) to the right coronary artery. A repeat midterm study was conducted in 32 patients between 1 and 5 years (mean 27 months) and the patency rate of the GEA graft was 91.5% (43/47 distal anastomoses)(Table3). In that group of 32 patients, the GEA graft occlusion sites were 1 circumflex and 3 right coronary arteries, among 47 distal anastomoses. A cumulative patency rate of the GEA graft at 4 years after operation was 86.4% (Fig2). Among the patent grafts, significant stenosis (>50%) at the anastomosis site or in the GEA trunk was found in 4 grafts. A string sign was found in 4 grafts. The other patent GEA grafts were patent with good flow (Fig 3). Graft assessment was made by MDCT in 18 patients: combined with catheter CAG in 15 patients and MDCT alone in 3 patients. There was no mismatch in evaluation of the GEA graft patency between catheter CAG and MDCT in the 15 patients.

Discussion
With recognition of the superiority of the ITA graft on long-term outcome,[4,5] other reliable arterial conduits have been needed to bypass affected coronary arteries. The GEA has been utilized in CABG following successful clinical reports by Pym et al[1] and Suma et al[2] in 1987. Our 20-year experience with 1,352 patients has shown that operative mortality is 1.26% and the respective actuarial 5-, 10- and 15-year survival rates are 91.7%, 81.4% and 71.3%, and cardiac death-free survival rates are 95.8%, 91.7% and 88.6%.[6]
The recent technique of skeletonization of the GEA, introduced by Gagliardotto et al[3] seems to achieve superior patency in the early postoperative period, even with off-pump CABG. Kamiya et al have reported that their early (2 postoperative weeks) patency rate of the skeletonized GEA graft used in off-pump CABG was 98.3% (118/120 distal anastomosis).[7] Ryu et al demonstrated satisfactory clinical and early patency results of the skeletonized GEA composite graft, with a patency rate of 97.3% within 2 weeks and 92.5% at 1 year postoperatively.[8] Kim et al have recently reported that their early patency rate of the skeletonized GEA graft to the right coronary artery territory was 98.3% (170/173),[10] which was not significantly different from that for the ITA anastomosed to the left coronary artery.
The skeletonized technique for harvesting the GEA achieves a longer graft for easier use in sequential anastomoses. In addition, the Y- or I-composite graft technique further increases the length of the GEA. Other advantages of using the skeletonized GEA graft technique are being abel to avoid or detect spasm at the time of surgery, obtain a longer and larger graft, and preservation of lymphatic and venous drainage to the stomach.
In the present study, our early patency rate for the skeletonized GEA graft was 97.6% and a cumulative 4-year patency rate was 86.4%, which was better than for the classic pedicle GEA graft in our previous report (80.5% at 4 years).[9] The patency status of the skeletonized GEA graft, such as string sign or flow physiology against competitive flow, is a matter of interest in comparison with the conventional pedicle GEA graft. Our study is ongoing following an increasing number of repeat studies with more precise analysis. MDCT is very useful in evaluating CABG grafts because it is a less invasive procedure than catheter CAG and, particularly for the GEA graft, clear visualization of the graft regardless of the difficulty of catheterization of abdominal arteries. On some occasions, the GEA originates from the superior mesenteric artery rather than the gastroduodenal artery. If the GEA graft is not visualized through the celiac or gastroduodenal angiography alone on CAG, it could be judged to be occluded when it is patent and instead arose from the superior mesenteric artery. MDCT can avoid such a misjudgment.
In conclusion, the skeletonized GEA graft enables a longer and larger conduit to be obtained and facilitates the use of composite and sequential grafts. The procedure is safe and effective, and better late patency rates can be expected.