Introduction
The first known use of the right GEA to supply blood to the heart was described by Bailey for implantation into the posterior territory of the heart in the 1960s . Two decades later, in 1987, Pym et al. and Suma et al. reported using the pedicled GEA as a bypass conduit with favorable long-term outcomes . Of note the typical proximal lumen diameter of pedicled GEA as reported by Pym was considerably smaller (in the range of 1.25–2.5 mm, and usually just under 1.5 mm) than one expects to achieve now through the use of skeletonization.
For various reasons the GEA has not been a widely popular choice of conduit for coronary revascularization. Most obviously, its use requires both the opening of a second body cavity, and the diversion of an artery through the diaphragm. Also, it is not easy to work with the GEA simultaneously with a thoracic conduit; in contrast, multiple operators can harvest a radial artery (RA) and/or a saphenous vein (SV) simultaneously with an internal thoracic artery (ITA). Further, the GEA being a muscular splanchnic artery, has a tendency to vasoconstrict. If the conduit is too narrow, the blood supply is inadequate, and that raises the risks of myocardial ischemia and of graft failure through competitive flow. Also, in the early years of its use, there was some concern related to the rate of late graft patency in pedicled GEA, which did not seem to be better than most SV grafts. This early phenomenon may have been due to primary anastomotic stenosis (with inadequate conduit dilatation during preparation), and to anastomosing the GEA to only mildly stenotic coronary arteries.
Concerns about vasoconstriction do not, however, apply when the GEA is skeletonized. With skeletonization the GEA is not only longer but also wide enough to supply adequate flow.
When properly managed, the skeletonized GEA offers a number of benefits. First, unlike the RA or SV, it can be used in situ, so that no proximal aortic anastomosis or complex Y-composite graft is required. Second, unlike a vein, the structure of the skeletonized GEA is well suited to arterial pressures, and it rarely develops significant atherosclerosis. Finally, intraoperative flow assessment of a GEA graft is technically simple.
Skeletonization of gastroepiploic artery
In 1998 Gagliardotto et al. first described the skeletonized preparation of the GEA . They found that skeletonization proved remarkably useful by preventing vasoconstriction, leaving the conduit significantly wider and longer. This creates two distinct advantages: dilation makes visual inspection and intraoperative graft assessments easier and allows easier sequential anastomoses. The first iteration of skeletonized harvesting techniques, using hemoclips, electrocautery, and scissors, was regrettably slow and difficult because of the fragile GEA arterial and venous branches. In 2001 we found an easier and safer way of harvesting and skeletonizing the GEA, using an ultrasonic scalpel (harmonic scalpel, Ethicon Endo-Surgery, Cincinnati, Ohio) . This chapter mainly describes the application of this improved harvesting method.
Anatomy of gastroepiploic artery
The right gastroepiploic artery (GEA) arises from the gastroduodenal artery. The gastroduodenal artery usually arises from the common hepatic artery of the celiac trunk ( Fig. 7.1 ). It occasionally arises from the right or left hepatic artery or from the accessory left hepatic artery. The gastroduodenal artery divides into the GEA and the superior pancreaticoduodenal artery behind the first portion of the duodenum, anterior to the head of the pancreas. In the rare case in which the gastroduodenal artery is absent, the GEA arises from the superior mesenteric artery. The GEA lies inferior to the pylorus and runs along the greater curvature of the stomach. It is the largest branch of the stomach arteries, giving off gastric branches ascending onto the surface of the stomach and epiploic branches descending into the greater omentum. In length the GEA has been measured as longer than one-half of the greater curvature of the stomach in 95% of people, and longer than two-thirds in a third of people . This length allows the GEA to reach the coronary arteries in almost all territories.
The GEA runs close to the inferior wall of the heart. When the GEA is used as an in situ coronary artery bypass conduit, we prefer to bring it into the pericardial space via an antegastric route, anterior to the pylorus and the liver, while others have reported to using a retrogastric route. Through the anterior vertical opening of the central tendon of the diaphragm, the in situ GEA is almost always long enough to reach not only the posterior descending artery (PDA) but also the left ventricular branches, and the circumflex branches with sufficient lumen size, especially when it is skeletonized.
A structural contrast between the GEA and the ITA exists in the balance between smooth muscle and elastic material in the media; the media of the GEA is biased more to the former, so it is classified as muscular, while the ITA is biased more toward the latter and is called elastic . An important functional correlate of the GEA’s muscularity is its tendency to vasoconstrict if inappropriately stimulated during harvesting. Other muscular arteries include the left anterior descending (LAD), the inferior epigastric artery, and the RA, though the RA has significantly thicker walls. The mean combined thicknesses of the intima and media in the GEA (291±109 μm), ITA (350±92 μm), and LAD (320±63 μm) are similar, and all are less than in the RA (529±52 μm) .
Physiology and vascular biology of gastroepiploic artery
The GEA has a similar vascular wall thickness to ITAs and LAD, but the GEA is a muscular artery and responds differently to pharmacologic stimuli. Its vascular tonus is variably affected by surgical manipulation, pharmacological stimuli, and diet. Many surgeons have found that the GEA has a higher tendency to vasospasm than does the ITA , a major concern in using the GEA, so correct preparation and management of the GEA is important for optimal graft performance in coronary artery bypass grafting (CABG).
He et al. proposed a three-part functional classification for arterial grafts: Type I, somatic arteries such as the ITA, IEA, and intercostal arteries; Type II, splanchnic arteries, including the GEA and splenic artery; and Type III, limb arteries, including the RA and lateral femoral circumflex artery. Anatomically, somatic arteries (Type I) such as the ITA are located in and supply blood to the body wall. The GEA, Type II artery is prone to spasm because of the higher contractility of splanchnic arteries, responsible for the changes in splanchnic blood flow to accommodate the function of the alimentary tract. The flow increases after meals and decreases in critical situations.
The ITA and the GEA have different pharmacological responses. Histamine causes contraction of the ITA but dilatation of the GEA . By the use of an implantable ultrasonic Doppler miniprobe, Takayama et al. found that blood flow through the GEA increases after a meal, consistent with the known response to histamine, just as one would expect of a stomach artery. This response is observed in the GEA graft connected to the coronary artery early after operation.
Toda et al. have reported that dopamine produces dose-dependent contraction of the GEA at the proximal portion, but at the distal site, it dilates at low concentrations and contracts at high concentrations, suggesting that the GEA smooth muscle contains α-adrenoceptors and has dopaminergic DA1 receptors in the distal region.
Our group investigated segmental differences in vasoreactivity of the GEA . In 2018 Kinoshita et al. divided full length GEA taken from total gastrectomy cases and divided them into three sections. The vasoreactivity of each segment was studied using vasoconstrictors and vasodilators. We found that segments in the distal region showed significantly more contraction than those at the middle and proximal sections, regardless of the type of vasoconstrictor, but no significant difference was found in vasodilator-induced relaxation. We propose that the distal portion of the GEA should be trimmed off and not be used as an anastomotic site wherever possible.
Gagliardotto et al. described using a skeletonized GEA, aiming to overcome some of the technical difficulties encountered with the conventional pedicled graft . Skeletonization has been reported to reduce vasoconstriction and to be as effective as intraluminal papaverine injection for a pedicled ITA graft . Gagliardotto et al. emphasized that the main advantage of skeletonizing GEA is the width of the conduit, but they recognized also that the skeletonization technique required extra time and care. After developing a simple and safe technique for harvesting skeletonized GEA , we further investigated the properties of the skeletonized GEA.
Two members of our group (Phung and Kinoshita) examined the histological and morphometric properties of the skeletonized GEA . The median lumen diameter at the most distal anastomosis was 3.8 mm (range 2.4−6.4 mm), markedly larger than the previously reported diameter of the pedicled GEA with surrounding tissue . Even when sequentially grafting the skeletonized GEA to a branch of the circumflex artery, the median distal lumen diameter was 3.9 mm (range 3.2−5.2 mm). The skeletonized GEA was found to have a sufficient luminal diameter at the distal anastomosis, with excellent graft flow, and graft patency. We noted that the first visual inspection and manual palpation of the GEA did not predict the potential full size of the artery, and care should be taken not to underestimate the available size of the GEA at the beginning of the surgery.
There have been some speculations about why the skeletonized GEA becomes so enlarged. One reason may be that in the conventional pedicled GEA, the surrounding omental tissues constrain and prevent the GEA trunk from dilating and elongating fully. Another possible reason is denervation of the periarterial sympathetic nerves. Yokoyama et al. demonstrated that trimming of perivascular connective tissue including the periarterial nerve plexus decreased smooth muscle contraction in the GEA, while preserving endothelial and contractile function of the smooth muscle . The denervation may play a role in inhibiting vasoconstriction in the skeletonized state.
Target selection for gastroepiploic artery grafting
The ideal target for GEA grafting is any totally occluded or severely stenotic coronary artery on the inferior surface of the heart such as the PDA or the posterior left ventricular (PLV) branches of the right coronary artery. Any branches of the circumflex artery with severe stenosis are also good targets if the skeletonized GEA can reach them with sufficient lumen size. The region of the inferior to infero-lateral wall of the heart is usually too far for in situ ITAs to reach with sufficient lumen size, so good long-term ITA patency rates are not expected there. In contrast, this is the target region nearest to the GEA.
A relative contraindication for GEA grafting is moderate or mild stenosis in the right coronary artery. Competitive flow from a native coronary artery might suppress flow through a GEA graft and induce narrowing or nonfunctionality . The same limitation applies to other arterial conduits, including ITA and radial arteries. Long-term graft patency results need to be followed in such patients. Di Mauro et al. cautioned that arterial grafts, particularly the GEA, demonstrated inferior clinical results when used on target vessels with less than 80% stenosis. Glineur et al. compared RCA revascularization with SV, GEA, and right ITA used in a Y-composite fashion. They found higher GEA and RITA graft patency rates with smaller target artery lumen size, especially when minimum lumen diameter of the proximal stenosis by quantitative angiography was below 1.1 mm, though SV patency was less sensitive to lumen diameter.
These findings show an important limitation for GEA usage. However, there have been insufficient comparative data on the details of skeletonized GEA interaction with flow competition.
Harvesting of the skeletonizing gastroepiploic artery using the harmonic scalpel
We believe that there are three important points to successfully perform GEA grafting in CABG surgery, based on the anatomical and physiological characteristics :
- 1.
The GEA should always be mobilized and harvested up to the proximal portion close to the pylorus of the stomach. This portion not only has a wider lumen but is also less prone to vasospasm.
- 2.
The GEA should always be skeletonized during harvesting, as this allows it to become a significantly larger and longer conduit than if nonskeletonized.
- 3.
The best graft target sites are severely stenotic or totally occluded coronary arteries on the inferior or the infero-lateral wall of the heart.
These crucial points provide the best effective early outcomes as well as durable long-term results. A simple surgical technique is described.
A midline chest skin incision is extended 5 cm from the xiphoid process toward the umbilicus; this is usually adequate to find and harvest the GEA and no large abdominal incision is needed. Following median sternotomy, we enter the abdominal cavity and incise the central tendon of the diaphragm deeper up to the attachment of the liver. This makes an opening from the abdominal cavity with a minimal upper abdominal skin incision. The wide diaphragmatic opening is later closed with a running nonabsorbable suture (such as 0 silk), leaving only a small loophole for the GEA into the pericardial cavity.
As a first step before approaching the GEA, we always check its appearance and pulsation before starting ITA harvesting. A sternal retractor is applied in reverse fashion (with the handle up toward the cranial direction) and opened widely. The retractor yields excellent exposure for the GEA harvesting. We always inspect the GEA and palpate its pulsation at the major curvature of the stomach. It is particularly important also to palpate the proximal portion of the GEA at the level of the pylorus to ensure that the GEA is of usable quality; if not, a limb vessel can be harvested while work proceeds on the ITAs.
After we harvest one or both ITAs, we reapply the sternal retractor to expose the GEA. We mostly use the harmonic scalpel with the coagulating shears tip to harvest and skeletonize the GEA. The exposure is normally facilitated with just the sternal retractor with the handle up toward the cranial direction. With traction applied to the border of the laparotomy with a large spatula, the most proximal portion of the GEA is readily accessed. We always palpate the pylorus as a marker of the proximal limit of the GEA harvesting. Decompression of the stomach cavity is important and is achieved by sucking air from the nasogastric tube; this significantly facilitates visualization of the whole stomach and the course of the GEA and makes harvesting easier. The stomach is gently pulled to show the greater curvature.
Next we pass thin blue vessel loops under only the GEA, omitting the vein, at 5-cm intervals ( Fig. 7.2 ). We make a small opening on the anterior layer of the greater omentum just above the trunk of the GEA using electrocautery. The soft tissue between the GEA and its satellite vein is carefully undermined using “mosquito” forceps so that only the artery is encircled with the thin blue vessel loops. We carefully avoid any injury to the gastric branches and epiploic arteries. We carry out this encircling process throughout the length of the GEA from the most proximal portion near the pylorus to the distal part.
We then unroof the anterior omental layer and tissue surrounding the GEA ( Fig. 7.3 ). The anterior layer of the greater omentum is carefully divided with the harmonic scalpel coagulating shears between the vessel loops. The nonactivated “tissue pad” jaw of the shears is inserted through the soft tissue ( Fig. 7.4 ) so that the GEA trunk can be protected from heat injury. With this procedure, all of the GEA is exposed. Should we encounter fatty omentum, we clear away the fat around the GEA by stroking it gently with the activated tip of the harmonic scalpel. A number of thin-walled gastric branches and epiploic arteries branch off the GEA; if those branches are large, we may seal and cut them together during this unroofing step.
