Which Is the Graft of Choice for the Right Coronary and Posterior Descending Arteries?
Comparison of the Right Internal Mammary Artery and the Right Gastroepiploic Artery
Background The graft of choice for the left anterior descending coronary artery is the left internal mammary artery because of superior long-term patency. However, controversy exists regarding the graft of choice for the right coronary artery and for the posterior descending branch.
Methods and Results Two types of pedicled arterial grafts were used for the right coronary and the posterior descending arteries in patients undergoing coronary bypass surgery between January 1991 and September 1994. Group A comprised 114 patients with a right internal mammary artery (RIMA) graft, and group B consisted of 127 patients with an in situ right gastroepiploic artery (R-GEA) graft. Mean age was 56.9 years in group A and 63.3 years in group B; 7.9% (9 of 114) and 33.9% (43 of 127) were diabetics in groups A and B, respectively. Overall mortality was 2.6% (3 deaths) for group A and 3.9% (5 deaths) for group B (P=NS). However, the prevalence of perioperative myocardial infarction in the right coronary artery distribution was significantly higher for group A (5.3%, or 6 of 114) than for group B (0.8%, or 1 of 127; P<.05), and the reoperation rate for graft failure (from 0 to 12 months after surgery) was significantly higher for the RIMA (4.4%, or 5 of 114) than for the R-GEA (0%; P<.05). Also, the prevalence of deep sternal wound infection in diabetics was significantly higher in group A (22.2%, or 2 of 9) than in group B (4.6%, or 2 of 43; P<.05).
Conclusions Our preliminary results suggest that the failure rate of the RIMA graft is significantly higher, especially if used as a pedicled graft to the posterior descending artery. The risk of sternal wound complications is greater in diabetics if both internal mammary arteries are used for grafting. Therefore, the R-GEA graft is preferred in diabetics and whenever the posterior descending artery is the target vessel.
Both the left and right IMAs are excellent conduits for coronary revascularization, with no difference in patency rates when the vessel bypassed is the left anterior descending coronary artery.3 4 5 However, some studies suggest that patency of RIMA grafts used to bypass vessels other than the left anterior descending coronary artery is considerably lower,3 especially if the vessel bypassed is the right coronary artery.4 5
Another problem is the prevalence of sternal wound complications, which is significantly higher after bilateral IMA grafting, especially in the presence of diabetes, obesity, or chronic pulmonary disease.6 7 Thus, avoidance of bilateral mammary grafts is particularly important in such patients.
The R-GEA, when used as a pedicled in situ graft, has patency rates (up to 5 years) similar to the LIMA graft.8 9 Another advantage of the R-GEA graft is that it reaches the inferior wall of the heart without tension and has excellent flow characteristics.10 Its use does not devascularize the sternum, does not prolong hospital stay, and is associated with low morbidity.11
Comparative studies between the RIMA and the R-GEA used as pedicled grafts to the right coronary artery and to the posterior descending branch have not been reported yet. The present study compares retrospectively the early and midterm results obtained with these two types of arterial conduits.
Between January 1991 and September 1994, the RIMA was used as a pedicled graft in 114 patients undergoing isolated coronary artery bypass grafting (group A). The recipient vessel was the right coronary artery in 86 patients and the posterior descending branch in 28 patients.
During the same time period, the R-GEA was used as a pedicled graft in 127 patients (group B). The vessel grafted was the posterior descending artery in 107 patients, the posterior ventricular branch in 4, and the right coronary artery in 16.
All of these patients were operated on at Geisinger Medical Center in Danville, Pa, by five attending surgeons (group A) and by one of us (C.A.D.) in the case of the R-GEA grafts. All patients gave informed consent, and the procedures followed were in accordance with institutional guidelines. Patients in whom the RIMA or the R-GEA grafts were used as “free” grafts were not included in this study.
In general, the RIMA was preferred in younger and in nondiabetic patients, especially if the stenosis was in the proximal right coronary artery. Patients in group B were older and sicker, with a significantly higher prevalence of diabetes, obesity, and other risk factors (see Table 1⇓). The R-GEA graft was also preferred by one of us (C.A.D.) when the posterior descending artery was the target vessel.
A midline sternotomy incision was used. The side branches of the IMAs were divided by use of scissors between metallic clips to prevent thermal injury to the IMA, which may be caused by electrocautery.12 On several occasions, the RIMA was dissected beyond the distal bifurcation to obtain a longer graft. The internal mammary pedicle was wrapped with a gauze soaked with a solution containing papaverine hydrochloride. After systemic heparinization, the distal end of the RIMA was ligated and divided; a solution containing papaverine was introduced into the distal end of the RIMA in selected cases, which included all the RIMA grafts taken distal to the bifurcation. Free flows were measured while the mean arterial pressure was maintained between 75 and 90 mm Hg. The mean free flow for the RIMA graft was 118 mL/min (range, 80 to 180 mL/min), which is similar to that of the LIMA graft (mean, 126 mL/min).
The R-GEA graft was approached by extending inferiorly the median sternotomy incision halfway to the umbilicus. The gastrocolic ligament was freed by division of the omental branches between ligatures of 2-0 silk. The branches to the greater curvature of the stomach were divided between 3-0 silk ligatures with scissors instead of electrocautery to avoid thermal injury to the intima of the R-GEA. Automatic staplers and hemostatic clips were avoided because clips may become dislodged during manipulation of the graft, causing serious bleeding or a hematoma in the pedicle, which may be difficult to manage. In addition, silk ligatures facilitate proper orientation of the R-GEA graft, minimizing the risk of twisting the pedicle.
Dissection of the R-GEA pedicle was continued proximally to the level of the pylorus and distally to the mid–greater curvature of the stomach. A gauze soaked with a solution containing papaverine was wrapped around the pedicle, and the R-GEA was ligated distally and divided after systemic heparinization. A solution of papaverine hydrochloride was introduced into the distal end of the R-GEA in all patients, and the distal 3 to 5 cm of the graft was discarded. The artery was then opened lengthwise until a lumen at least 2 mm in diameter was identified. An attempt was made to keep the R-GEA graft as short as possible to obtain a larger lumen and better flows. The length of the R-GEA graft that was used ranged from 11 to 15 cm; the mean internal diameter was 2.3 mm (range, 1.5 to 3 mm). The mean free flow for the R-GEA graft was 128 mL/min (range, 60 to 270 mL/min).
The R-GEA graft was passed through the fibrous portion of the diaphragm, after a 2-cm segment was excised adjacent to the posterior descending artery (Fig 1⇓). The retrogastric route was preferred for several reasons: first, when the R-GEA is positioned behind the stomach, the length of the graft is shorter (Fig 2⇓), which correlates with better flows and a larger lumen for the anastomosis; second, the R-GEA graft is less likely to be stretched by gastric dilation when placed behind the stomach (Fig 2⇓); third, the R-GEA may kink at the anastomosis if the graft is positioned anterior to the stomach and the liver; and fourth, the R-GEA is less likely to be injured during a future laparotomy if placed behind the stomach and the left lobe of the liver.13
Myocardial protection was accomplished with antegrade cold blood cardioplegia. In 15 patients, however, no clamps were applied to the ascending aorta because it was heavily calcified. In these patients, the distal anastomoses were performed with intermittently induced ventricular fibrillation or with a beating heart.14
The distal anastomosis was performed with continuous 7-0 polypropylene suture, and the “toe” of the anastomosis was oriented “upstream” (Fig 3⇓) to reduce the possibility of twisting or kinking the R-GEA. The triangular ligament of the liver was usually reattached to the diaphragm anterior to the R-GEA to prevent compression of the graft.
Intravenous diltiazem was given for 48 hours at 0.1 mg · kg−1 · h−1, followed by oral calcium channel blockers for 3 to 6 months to reduce the possibility of arterial spasm.
Mean age for the two groups and mean number of distal anastomoses per patient were tested using unpaired Student’s t test. Comparison of all other risk factors (Table 1⇑) was carried out using the χ2 test, and comparison of results for groups A and B (Table 2⇓) was analyzed using multiple logistic regression analysis. A value of P<.05 was considered significant.
There were two cardiac-related deaths in each group (1.8% and 1.6% for groups A and B, respectively). One patient in group A died in low cardiac output after a reoperation for failure of the RIMA graft; another death was caused by an anterior myocardial infarction. In group B, both deaths were caused by an anterior myocardial infarction (one of these patients underwent concomitant endarterectomy of the anterior descending coronary artery).
Four additional noncardiac deaths occurred: 1 (0.9%) in a group A patient, who had a postoperative cerebrovascular accident; and 3 (2.4%) in group B, 2 of which were secondary to cerebrovascular accidents and 1 caused by bilateral pneumonia with gram negative septicemia. Thus, the overall hospital mortality was 2.6% (3 patients) and 3.9% (5 patients) for groups A and B, respectively (P=NS). In both groups, mortality was higher in patients aged 70 or older but not significantly different between the two groups (Table 2⇑).
Other postoperative complications are listed in Table 2⇑. There was a significantly higher prevalence of a new Q-wave inferior wall myocardial infarction in group A (5.3%) compared with group B (0.8%) (P<.05). In group A, only 1 patient out of 6 who sustained a perioperative inferior wall myocardial infarction underwent concomitant endarterectomy of the right coronary artery; in group B, the only patient with an inferior wall infarction after surgery also had an extensive endarterectomy that might have predisposed this event.
Another significant difference observed was the percentage of graft failures requiring reoperation from 0 to 12 months after surgery, which was significantly higher for group A (4.4%) than for group B (0%) (P<.05). Three patients required revision of the RIMA graft during the initial 24 hours after surgery because of acute ischemic changes in the inferior leads of the ECG. At reoperation, severe spasm was noted in the distal segment of the RIMA graft. In another patient, who required a reoperation 6 months after surgery for recurrent angina, the RIMA graft, which had been anastomosed to the posterior descending artery, was under tension and diffusely narrowed. A fifth patient, with recurrent angina, underwent reoperation 12 months after the initial surgery because of an anastomotic stricture, presumably caused by anastomosing the RIMA to a thick-walled right coronary artery proximal to the bifurcation. In addition, one patient required angioplasty plus atherectomy of the right coronary artery because of failure of the RIMA graft, also caused by anastomosing the RIMA to a right coronary artery with a thick wall.
In group B, however, there were no reoperations for graft failure. The difference is statistically significant (P<.05) (Table 2⇑).
With respect to DSWI, there was no significant difference in the overall rate of infections when both groups were compared. In group A, 2 of 105 (1.9%) nondiabetic patients had DSWI; in group B, 2 of 84 (2.4%) nondiabetic patients had DSWI (P=NS). However, the prevalence of DSWI in diabetic patients was significantly higher in group A (22.2%) compared with group B (4.6%) (P<.05) (see Table 2⇑).
In our present series, three patients with an R-GEA graft had abdominal complications related to its use. In one patient, who was obese and diabetic, the abdominal incision dehisced. Another patient developed an incisional hernia 1 year after surgery. Abdominal wound complications such as these have been avoided since we started using retention sutures. Another patient developed a herniation of the gastric fundus through the diaphragmatic opening. This complication occurred because the opening in the diaphragm was more posterior and larger than usual and can be prevented by creating the window for the R-GEA graft in the fibrous portion of the diaphragm.13
During a mean follow-up of 16.0 months (range, 1 to 36 months), there were two deaths in group A: one sudden death, presumably caused by ventricular arrhythmia, and one death caused by congestive heart failure. No late deaths occurred in group B.
Six patients (5.3%) in group A, five of whom underwent repeat coronary angiography, had recurrent angina. Two of these patients required reoperation, and another patient had angioplasty and atherectomy as already mentioned. The other three patients were managed with medications.
In group B, two patients (1.6%) had recurrent angina. Repeat catheterization showed no significant abnormalities in one patient; another patient had a stenotic vein graft to an obtuse marginal artery and a widely patent R-GEA graft (Fig 4⇓). Both patients were managed conservatively.
A stress thallium scan was performed 1 year after operation in 41 patients in group A and in 69 patients in group B, all of whom were free from angina. Stress-induced defects were observed in the right coronary distribution in 3 patients (7.3%) in group A and in 4 patients (5.8%) in group B (P=NS).
Patients revascularized with the LIMA have improved 10-year survival rates and reduced risk for late cardiac events compared with patients who underwent myocardial revascularization with vein grafts only.15 However, patients with patent LIMA grafts are still at risk of recurrent angina, late myocardial infarction, or death.16 Thus, bilateral IMA grafting has been advocated by several authors.16 17 However, bilateral IMA grafting should be avoided in insulin-dependent diabetic patients because of the increased risk of sternal wound complications.6 7 18
To avoid excessive devascularization of the sternum, the R-GEA graft may be employed as an alternative arterial conduit.8 9 10 11 When used as a pedicled graft, the patency rate of the R-GEA (up to 5 years follow-up) is similar to the LIMA graft.8 9 However, patency of arterial grafts is lower when used as “free” grafts. Loop and associates19 observed that only 77% of free IMA grafts were patent 18 months after surgery, compared with the 85% to 95% patency rate of the pedicled LIMA graft after 10 years.15 Likewise, the early patency rate (up to 1 year) of the free R-GEA graft is only 75%, compared with 95% patency of the in situ R-GEA graft.8
Thus, the IMA and the R-GEA, when used as pedicled grafts, are the two grafts of choice because of their superior patency rates.8 9 Even at advanced age, both have a low susceptibility to the development of atherosclerosis, which may be a major determinant in the higher long-term patency of these conduits.20
All arterial conduits are susceptible to spasm.21 Spasm of the R-GEA may be triggered by surgical manipulation or by catheter stimulation at angiography. However, in a study reported by Suma and associates,8 there was no evidence of myocardial ischemia or angina pectoris caused by spontaneous spasm of the R-GEA after the operation.
Early postoperative spasm of the LIMA may also occur, usually with catastrophic consequences.22 23 He24 observed that the contractility of the IMA increases toward the distal end. The implication is that the distal IMA, or its branches, should not be used for the anastomosis.24 25 Morin and associates26 also observed a high prevalence of recurrent angina and a 40% occlusion rate when the IMA branches were used.
This may be an explanation for the increased rate of RIMA graft failure observed in our present series, because three patients in group A who required early reoperation were noted to have severe spasm of the distal end of the RIMA graft. Other authors4 27 have also observed that patency of the RIMA was significantly lower than the LIMA graft when the vessel bypassed was not the left anterior descending artery. An explanation offered by Huddleston and associates27 is that a more distal segment of the RIMA was used for the anastomosis relative to the length of the LIMA graft because of the distance involved in reaching the target vessel.
Although it may be argued that antispasm treatment was not the same in our two groups of patients, because papaverine hydrochloride was not routinely introduced in all RIMA grafts and postoperative calcium channel blockers were used selectively in group A, we believe that this did not contribute to an increased failure rate of the RIMA graft because papaverine hydrochloride was introduced in all RIMA grafts taken distal to its bifurcation. Postoperative diltiazem was also used in these patients to minimize the risk of spasm.
Another cause of RIMA graft failure in our experience was anastomosing the RIMA to the posterior descending artery under tension. To avoid this complication, some authors prefer to use the RIMA as a free graft in some cases.5 17 However, we currently prefer to use the in situ R-GEA graft to bypass the posterior descending artery because it reaches the inferior wall of the heart without any tension, especially if the retrogastric route is used (see Fig 2⇑).
In the majority of our patients, the RIMA was anastomosed to the right coronary artery, especially in younger patients and when the obstruction was located in the proximal segment of this vessel. However, the wall of the right coronary artery is frequently diseased down to the bifurcation, and the anastomosis between the RIMA and a thick-walled right coronary artery may result in anastomotic stricture, as occurred in two patients in group A who required reintervention.
Although it has been suggested that bilateral IMA grafting does not increase surgical risk,28 a recent article indicates that use of the RIMA is an incremental risk factor for elderly patients undergoing coronary revascularization.29 In the present study, the overall mortality was slightly higher (but not significantly; P=.729) for group B, which can be explained because of the significantly higher prevalence of associated risk factors present in group B. Nevertheless, the mortality for septuagenarians was less in group B than in group A (P=NS) (Table 2⇑). In general, we prefer to use the R-GEA in elderly patients in whom the saphenous veins are not suitable, because the posterior descending artery is less likely to be diseased in these patients.
Another major concern is the increased risk of DSWI in diabetic patients, which may be 12% when both IMAs are used.6 18 In our experience, although the overall rate of DSWI was similar for both groups, the prevalence of DSWI in diabetic patients was significantly higher in group A than in group B (P<.05). However, there was no significant difference in nondiabetic patients (see Table 2⇑).
In summary, the assumption that the RIMA is the graft of choice for the right coronary artery system is not necessarily correct. We believe that the R-GEA is more versatile because, unlike the RIMA, the R-GEA may be used in most patients. It is also a more dependable graft, because in our experience the prevalence of ischemic events and of graft failure requiring reoperation was significantly higher for the RIMA graft (P<.05).
Unlike the RIMA, the in situ R-GEA graft reaches the posterior descending artery without tension; this artery is usually a more suitable vessel for the anastomosis than the right coronary artery itself. In addition, bilateral IMA grafting may result in an increased rate of sternal wound complications, especially in insulin-dependent diabetic patients. Our current approach is to use the RIMA as a pedicled graft in younger, nondiabetic patients if the right coronary artery can be grafted in an area free from disease and the anastomosis is performed avoiding the distal segment of the RIMA and its branches.
At present, the R-GEA is our arterial conduit of choice for the posterior descending artery, whether it arises from a dominant right or left coronary system. The R-GEA is also our preferred graft in patients in whom the risk of sternal wound complications is increased, including diabetic and obese patients and those who are taking steroids or who have a history of radiotherapy to the chest. A previous abdominal operation is not a contraindication for using the R-GEA, except for a previous gastric resection. A laparotomy in the near or distant future also does not preclude its use.13
Another important issue when using the R-GEA is to avoid long grafts, because the distal lumen is usually smaller and more prone to spasm. The most effective method to obtain good flows from the R-GEA is to use a shorter graft placed behind the stomach. This usually correlates with a much wider lumen (at least 2 mm), which is less likely to undergo spasm. Under these conditions, the use of topical and systemic vasodilators becomes an issue of secondary importance.
Since submission of the original manuscript, 58 additional patients underwent coronary bypass surgery using the R-GEA, with no deaths. The overall mortality is now 2.7% (5/185) for all patients in whom the R-GEA graft was used.
Selected Abbreviations and Acronyms
|DSWI||=||deep sternal wound infection|
|IMA||=||internal mammary artery|
|LIMA||=||left internal mammary artery|
|R-GEA||=||right gastroepiploic artery|
|RIMA||=||right internal mammary artery|
The authors are very grateful to Lisa Peñalver, BA (medical illustrator, Fairbanks, Alaska) for artistic illustrations; to Carmen Acuña, PhD (Department of Mathematics, Bucknell University, Lewisburg, Pa), for statistical analysis; and to Ellen K. Smith, MD, and John A. Baxter, MD, for providing postoperative angiograms.
Reprint requests to Charles A. Dietl, MD, Department of Cardiovascular Surgery, Geisinger Medical Center, Danville, PA 17822-1343.
Presented in part at the 67th Scientific Sessions of the American Heart Association, Dallas, Tex, November 14-17, 1994, and published in abstract form (Circulation. 1994;90[pt 2]:I-251).
- Copyright © 1995 by American Heart Association
Tector AJ, Schmahl TM, Canino VR. The internal mammary artery graft: the best choice for bypass of the diseased left anterior descending coronary artery. Circulation. 1983;68(suppl II):II-214-II-217.
Okies JE, Page US, Bigelow JC, Krause AH, Salomon NW. The left internal mammary artery: the graft of choice. Circulation. 1984;70(suppl I):I-213-I-221.
Chow MST, Sim E, Orszulak TA, Schaff HV. Patency of internal thoracic artery grafts: comparison of right versus left and importance of vessel grafted. Circulation. 1994;90(suppl II):II-129-II-132.
Dion R, Etienne PY, Verhelst R, Baele P, Wyns W, Chalant CH. The right mammary artery (RIMA) as coronary bypass graft: techniques and results in 200 consecutive patients. J Cardiovasc Surg. 1991;32:71. Abstract.
Dietl CA, Deitrick JE, West JC, Pagana TJ. Laparotomy after using the gastroepiploic artery graft: retrogastric versus antegastric route. Ann Thorac Surg.1995;60:382-386.