A been described after it has been used as a coronary

Transcription

1 Competitive Flow From a Fully Patent Coronary Artery Does Not Limit Acute Mammary Graft Flow Paul A. Spence, MD, Robert M. Lust, PhD, Richard S. Zeri, MD, Stanley R. Jolly, PhD, Prabodh M. Mehta, MD, Masaki Otaki, MD, You Su Sun, MD, and W. Randolph Chitwood, Jr, MD Divisions of Cardiothoracic Surgery and Cardiology, ast Carolina University School of Medicine, Greenville, North Carolina The shriveled, stenotic mammary graft sometimes observed after internal mammary artery (IMA) to coronary artery bypass grafting has been attributed to competitive flow from the insufficiently stenosed native coronary vessel. To study further the effects of native coronary artery competing flow on IMA graft flow, 10 dogs (mean weight, 23.5 f 3.69 kg) underwent coronary artery bypass grafting using the pedicled left IMA anastomosed to a normal, fully patent proximal circumflex (CFX) coronary artery. The procedure was performed through a left thoracotomy, off pump, using a brief local occlusion to perform the anastomosis. Native in situ IMA flow, CFX flow distal to the anastomosis, and IMA graft flow were measured using calibrated electromagnetic flow probes. When the CFX proximal to the anastomosis was occluded transiently, IMA flow increased to supply 100% of the previously measured distal CFX flow (60.2 f 7.9 ml/ min). When both the IMA graft and CFX proximal to the anastomosis were patent, total distal perfusion was maintained (58.9 f 7.8 ml/min) and relative IMA graft flow (26.5 f 3.3 ml/min) was proportional to the relative diameter of the IMA graft to the native coronary artery (T = 0.96). The mean flow in the IMA in situ on the chest wall before its division was ml/min. These results suggest that, at least acutely in a canine model, IMA graft flow is maintained above in situ levels even when grafted to a completely patent coronary artery and that acute competitive flow probably does not cause mammary artery shriveling. ( 1992;54:21-6) shriveled and stenotic mammary artery graft has A been described after it has been used as a coronary artery bypass conduit. Some have suggested that the appearance of a string sign or the distal thread phenomenon, which is observed arteriographically when mammary artery grafts to coronary arteries are studied, is due to competitive flow [l-31. That is, high flow rates from the native coronary vessel result in reduced mammary artery flow and ultimately in mammary artery shrinkage. In 1974, Barner introduced the term disuse atrophy to describe this phenomenon, although he later noted that the mammary narrowing may be reversed [l, 21. A number of factors led us to wonder whether this initial concept were true. With the widespread use of the internal mammary artery as the conduit of choice for coronary artery bypass grafting, there has been surprisingly little clinical documentation of the development of mammary artery shrinkage despite the pressure to place a mammary graft to many minimally obstructed left anterior descending coronary arteries. A report by Urschel and associates [4] suggested that coronary graft patency was not adversely affected by successful intraoperative balloon dilation of lesions proximal to the sites of graft insertion. These observations led us to wonder whether the effect of competitive flow from a patent coronary artery vessel was, in fact, responsible for the early reports of mammary shrinkage. This study was conducted to evaluate the effect of competitive flow from a fully patent and normal coronary artery on mammary artery graft flow in an acute experimental canine model. Material and Methods Ten mongrel dogs, weighing 14 to 27 kg (mean weight, 23.5 * 3.69 kg), were anesthetized with sodium pentobarbital (26 mg/kg intravenously), intubated, and ventilated with a volume ventilator. A cutdown was performed on the left femoral artery, and arterial blood pressures were monitored with high-fidelity catheter micromanometers (Millar Instruments, Houston, TX). Blood gases were monitored and bicarbonate was added to maintain ph between 7.35 and The electrocardiogram was monitored continuously. A left fifth interspace thoracotomy was performed and the left internal mammary artery was dissected from the chest wall with a wide pedicle including the accompanying veins and adjacent muscle. The dissection was continued proximally to the vessel s origin at the left subcla- Presented at the Twenty-eighth Annual Meeting of The Society of Thoracic vian and to the level Of the seventh Or Surgeons, Orlando, FL, Feb S5, eighth costal interspace. With the mammary pedicle taken Address reprint requests to Dr Spence, Division Cardiothoracic Surgery, down, but with the distal end of the vessel Still intact in ast Carolina University School of Medicine, Greenville, NC the circulation, flow in the mammary artery was mea by The Society of Thoracic Surgeons /92/$5.00

2 22 SPNC T AL COMPTITIV CORONARY GRAFT FLOW 1992;51:214 of the anastomosis, the snares occluding the circumflex artery were released and the mammary occluder was released. All hearts were allowed to recover for approximately 30 minutes, during which time cardiac function normalized. No evidence of ischemia was apparent on electrocardiogram. The electromagnetic flow prolbes were placed around the left internal mammary artery and the circumflex coronary artery proximal and distal to the site of the anastomosis (Fig 2). The probes were then calibrated electronically. The left internal mammary artery was occluded and the flow in the distal circumflex artery was measured. After 10 minutes of reequilibration, the circumflex coronary artery was snared proximal to the site of the mammary anastomosis and mammary artery flow was measured. After a second interval of 10 minutes, flow in both the mammary artery and the distal circumflex coronary artery was measured with both the mammary and proximal circumflex arteries patent. The animals were then sacrificed by fibrillation. All anastomoses were probed to ensure their patency. The mammary artery and the circumflex artery were measured and their sizes recorded. In conducting the experiment we adhered strictly to the guiding principles for the Care and Use of Laboratory Fig 1. Technique of anastomosis of left internal mamma y artery to circumflex artey with local occlusion off pump. sured using a calibrated flow probe (Howell Instruments, Camarillo, CA). The pericardium was then incised and its edges were sutured to the margins of the thoracotomy incision to create a cradle for the heart. The left atrial appendage was ligated at its tip and gently retracted with this suture to expose the atrioventricular groove. The coronary anastomosis was performed using a modification of a technique described by McCarthy and Schaff [5]. The left circumflex artery was dissected proximal to the takeoff of its first marginal branch. Small overlying veins were ligated as necessary. When approximately 1.5 cm of the circumflex artery was prepared, snares of number 2 silk were placed around the vessel at each end of the dissected segment. Lidocaine 1% (Abbott Laboratories, North Chicago, IL) was then administered intravenously at a dose of 1.5 mg/kg. The snares were then tightened around the vessel and lifted to reduce the vessel movement with cardiac activity (Fig 1). A coronary arteriotomy approximately 5 mm long was made, and the spatulated distal end of the mammary vessel was anastomosed to the circumflex artery with a continuous suture of 7-0 polypropylene suture (thicon, Somerville, NJ). The time required for this anastomosis was between 5 and 10 minutes. During this period of occlusion, no episode of fibrillation occurred, but early paradoxic lateral wall motion could be observed. Immediately after the completion Fig 2. Position of the flow probes on the internal mammy artey, proximal circumflex artery, and distal circumflex artery.

3 ~~~ ~ 1992;54:21-6 SPNC T AL 23 COMPTITIV CORONARY GRAFT FLOW _ c IMA and Proximal CFX Open 0 IMA Occluded, Proximal CFX Open IMA open, Proximal CFX Occluded.- C - Y g 20 0 In Situ 0 As Graft In Competition I I 20 1c 10 0 Fig 3. Total flow in the distal circumflex artery (CFX) remained constant when both the internal mammary artery (IMA) and the proximal CFX were open, the IMA was occluded with the proximal CFX open, or the proximal CFX was occluded with the IMA open. Animals of the American Physiological Society. All data are expressed as the mean and standard deviation of the mean. Linear regression was employed to express the relationship between ratios of vessel sizes and vessel flows. Results Flow in the distal circumflex artery was the same whether the circumflex was patent with the mammary occluded, the mammary was patent with the circumflex occluded, or when both vessels were patent (Fig 3). The relationship of flow in the mammary and the proximal circumflex arteries as well as the total circumflex flow when both inflows were patent (competing flow condition) is summarized in Fig 4. Note that both the mammary and circumflex vessels contributed substantially to the total flow in the distal vessel when both were patent. The flow in the left internal mammary artery before its division distally (in situ flow) is compared with the flow in the mammary artery after it was anastomosed to the circumflex vessel (flow in competition) in Figure 5. ven when the mammary was competing with the native C Fig 5. Internal mammary artery flow was higher after it was anastomosed as a graft than while in situ on the chest wall. coronary vessel, its flow was higher than it was in its in situ condition. The diameter of the internal mammary artery is compared with the diameter of the circumflex vessel in Figure 6. When both the mammary and the proximal circumflex were patent (competitive flow), the ratio of internal mammary artery diameter to total diameter (the sum of internal mammary artery and circumflex artery diameter) was linearly related (y = (0.902)~ ) and significantly correlated (Y = 0.96; p < 0.01) to the ratio of internal mammary artery flow to distal circumflex artery flow (Fig 7). In other words, in this model, the relative internal mammary artery diameter determined the internal mammary artery flow contribution to distal perfusion. Comment The results of these studies suggest that, at least acutely, mammary artery graft flow and flow from a normal nonstenotic coronary artery both contribute to distal coronary perfusion. Furthermore, it appears that the relative amount of flow each provides is related to each vessel's diameter, with the larger vessel providing more flow. It was interesting that at least in this study, the flow in the 50._ 1 40 v 70 1 I 0 Proximal CFX i 20 I 4 0 IMA CFX 10 0 Fig 4. When all vessels were open, the total circumflex artery (CFX) flow was the sum of proximal CFX flow and internal mammary artery (IMA) flow. The relative contribution of each vessel is demonstrated. Fig 6. The mean circumflex artery (CFX) diameter was higher than the internal mammary artery (IMA) diameter.

4 24 SPNC T AL COMPTITIV CORONARY GRAFT FLOW 1992;54: b m m r IMA/Distal CFX Flow Fig 7. Linear regression demonstrates that the ratio between pow in the internal mammay artey (IMA) and total flow in the distal circumflex artey (CFX) was highly correlated to the ratio of IMA diameter to total diameter (sum of ZMA and proximal CFX diameters), r = mammary artery when used as a graft on a normal coronary artery was greater than its flow when in situ on the chest wall. Thus the acute measurement of flow in the mammary graft would suggest that mammary graft flow is not overwhelmed by the normal coronary vessel and that, in fact, its flow is greater than when attached in situ. If the same conditions occur in humans, these data would suggest that factors other than acute competitive flow result in the mammary string sign. It is, however, possible that this model provided an exceptional situation for study because the circumflex coronary artery was large with a high baseline flow and the mammary artery was just slightly smaller with a relatively low in situ flow. Longer time periods must be studied with a greater variety of mammary and coronary vessel sizes to further clarify these observations. Despite the widespread use of the mammary artery for coronary surgery, there has been little recent clinical reporting of mammary arteries shriveling. Many observations of the mammary string sign occurred early after its use as a coronary bypass conduit. It is conceivable that some of these episodes may represent technical problems related to inexperience with mammary artery harvesting and grafting. In their early experience, Dincer and Barner [2] noted an 11% incidence of diffuse mammary artery narrowing that decreased to only 2% shortly thereafter. If the criteria for mammary grafting were not changed, technical factors may have played an important role in some of the earlier observations about mammary narrowing. Barner himself wondered if trauma may have been a factor. He questioned whether medial necrosis and fibrosis secondary to the ischemic insult of mobilization, as has been described for vein grafts may have been responsible [l]. Other early experimental work on the mammary artery suggests that trauma may play an important role in mammary artery narrowing. Mullerworth and associates [6] found that in a canine model, mammary arteries used as intramyocardial implants were much more prone to development of subendothelial fibrocellular proliferation similar to that seen in vein grafts than were mammary arteries used as direct coronary artery bypass grafts. The intramyocardial segment of the mammary implant was particularly prone to development of this complication. There is other evidence that competitive coronary flow is less harmful to graft patency. Urschel and colleagues [4] recently described a series of patients in whom intraoperative balloon angioplasty was performed in addition to coronary artery bypass grafting,. Despite good angioplasty results reducing proximal obsiruction to coronary artery flow, the patency of bypass grafts did not seem to be adversely affected. The available data leave considerable doubt about the etiology of mammary artery narrowing. There are examples in the circulation in which a number of arterial inflows result in end-organ perfusion, which is particularly obvious in the extremities. Both the hand and the foot receive inflow from a number of major sources (ulnar and radial in the hand and peroneal, posterior tibial, and anterior tibial in the foot),,and a stable relationship between competing sources of flow is clearly evident. Whether such a situation can exist when coronary grafts are piggybacked on top of native vessels without intervening vascular arches is uncertain. However, if this were a stable situation with satisfactory mammary graft flows and long-term patency, it may be necessary to rethink the current strategy for the management of coronary disease. At the time of coronary bypass grafting, perhaps all vessels, including minimally diseased ones, should be grafted with arterial grafts. For. cost effectiveness perhaps the approach to single-vessel disease should not be angioplasty but arterial revascularization of all vessels. As stated previously, it will be necessary to greatly expand this line of research and to undertake chronic studies of graft patency to determine if, in fact, long-term patency of arterial grafts occurs when they are anastomosed to normal vessels. If patency of arterial grafts to normal vessels does occur, it is highly possible that particular attention to the mechanism of patency will be necessary. It is clear that total flow is not enough to explain mammary graft patency; the normal-sized mammary artery flowing into a tiny residual coronary artery lumen years after operation has been seen frequently. It is possible that the pressure dynamics of a pulsating tube keep this artery expanded. In other situations where there is good outflow, it may be necessary to size-match mammary vessels to their recipient undiseased vessels. Application of a graft to a much larger undiseased vessel may result in an unacceptably low rate of graft flow and, without the benefit of high flow or expanding pressure, the graft may shrivel. It may also be necessary to select an arterial conduit that has a lower flow in its native circulation than it cariries in a competitive flow situation as a graft. It is clear that a number of studies will be necessary to clarify this issue. In summary, mammary artery graft flow is not eliminated by competitive flow from a normal coronary artery

5 1992;54:214 SPNC T AL 25 COMPTITIV CORONARY GRAFT FLOW in this acute canine model. It appears that the amount of flow in a mammary graft depends on the flow required in the run-off bed and the relative sizes of the native coronary artery and the mammary graft. Further studies are necessary to determine if these early observations persist over the long term. References 1. Barner HB. Double internal mammary-coronary artery bypass. Arch Surg 1974;109: Dincer B, Barner HB. The occluded internal mammary artery graft: restoration of patency after apparent occlusion associated with progression of coronary disease. J Thorac Cardiovasc Surg 1983;85: Geha AS, Bane A. arly and late results of coronary revascularization with saphenous vein and internal mammary artery grafts. Am J Surg 1979;13745& Urschel HC Jr, Razzuk MA, Miller, Chung SY. Operative transluminal balloon angioplasty. J Thorac Cardiovasc Surg 1990;99: McCarthy PM, Schaff HV. A cost effective technique for experimental coronary artery bypass. J Thorac Cardiovasc Surg 1988;96: Mullerworth MH, Daniel FJ, Lie JT. The fate of internal mammary arterial implants and bypass conduits for myocardial revascularization. J Thorac Cardiovasc Surg 1975; DISCUSSION DR STVN R. GUNDRY (Loma Linda, CA): I congratulate you on an elegant study and also offer a word of caution. Anyone who has done a midline sternotomy on a dog knows that the mammary artery size relative to the body size in a dog is much different than it is in a human. I am not sure that you can extrapolate these data directly from the dog, which has a huge mammary artery relative to its body size, to the human, which has a relatively smaller mammary artery. Also, the second word of caution is that you have anastomosed your mammary arteries into a run-off bed, which by your demonstration is better than its native bed, and I am not sure that the diseased coronary artery (which is more and more often diseased in my practice) is analogous to an ideal run-off bed. DR SPNC: Thank you very much Dr Gundry. We agree that the mammary artery in a dog is a very large vessel. The vessels were harvested distally where they tapered considerably. However, as you saw, the average size of these mammary arteries was about 2.5 mm in dogs with an average weight of about 23.5 kg. Your comment about extrapolating these data to humans is well taken. We too were concerned about the size of the run-off bed. Interestingly, our concern related to using a smaller mammary artery in a larger run-off circulation. DR DWARD D. VRRIR (Seattle, WA): Could you tell us what the timing was of your data sampling? Did the vessel you were studying still have the ability to autoregulate its blood flow, or was it vasodilated in a reactive hyperemic phase? The size of the vessel is not necessarily a meaningful measurement. Flow has to do with the metabolic needs of the bed and therefore how much ability that vessel has of autoregulating its flow over a tremendously wide range of pressures. DR SPNC: Your question about reactive hyperemia is a good one. The occlusion time to perform these anastomoses was short. We allowed about half a hour after the completion of each anastomosis before any other measurements were taken. In several animals, we sequentially measured blood flow to ensure that a baseline level had been reached. In a number of these animals, we also injected adenosine to examine the response. The response to adenosine did suggest that they were not in a maximally dilated state. DR VRRIR Both vessels had the ability to augment flow to adenosine? DR SPNC: Correct. It was very interesting that the mammary arteries appeared to have a flow capacity at least equal to and perhaps slightly better than the native coronary vessels. However, we only injected adenosine in several animals and so we did not report these data. DR VRRIR Do you have any preliminary data on the longterm implications, because the short term is only a curiosity. In a patient with coronary disease, it is the long term that is most essential. DR SPNC: We intended to have a whole series completed by this time, but angiographic equipment problems occurred, and to this date, only 2 have been studied at 4 months and 3 months, respectively. The dog at 4 months had a beautifully patent mammary artery flowing in competition. When he was restudied, his flow behavior was identical to the acute measurements we reported in this presentation. The second animal, studied at 3 months, had developed a mammary artery string sign. However, what was most interesting was the fact that although this artery did not fill at arteriography, it was able to provide about 50% of native circumflex flow when the proximal circumflex artery was acutely occluded. In other words, although it did not fill on the arteriogram, the vessel was able to function as a collateral. This, in fact, may be the most interesting observation we have seen. It is remarkable to see failed angioplasty patients with bail-out catheters in place who are in stable condition for prolonged periods with minimal flow. It is possible that these shrunken mammary arteries may act as bail-out catheters and provide adequate collateral circulation after an acute occlusion, preventing infarction and arrhythmias. DR VRRIR But that has tremendous implications in terms of your conclusions. DR SPNC: You are certainly correct. However, these chronic studies are only in 2 dogs, so I would not want to make any general or long-term conclusions based on these few observations. DR CHARLS B. BCKMAN (New Haven, CT): I am wondering if we could have a show of hands of all those people in this room who have seen a vessel develop a string sign and then open up later on. (Show of approximately 40 hands.) I think this phenomenon is still happening.

6 26 SPNC T AL COMPTITIV CORONARY GRAFT FLOW 1992;54:21-6 DR SPNC: I did not mean to imply that the mammary artery string sign does not occur. Stringing of the mammary artery may not be a serious problem, if in fact the mammary artery can respond when an acute occlusion of the native coronary artery occurs. There are two clinical reports discussing this problem; one comes from Dr Barner and one was just recently published from Japan by Dr Kitamura. These reports suggest that these mammary grafts change in size and respond to their environment. Perhaps this is the most important message from these mammary artery studies. We have also wondered whether these acute mammary artery flows could be maintained pharmacologically, perhaps with the use of calcium channel blockers. We have no information on this question. DR J. LANCLOT LSTR (West Palm Beach, FL): I believe that we have a chronic model to evaluate competitive flow in patients with a string sign of an internal mammary artery graft. These patients are seen for cardiac catheterization and angioplasty. This would allow temporary occlusion of the grafted coronary artery. The morphology of the mammary graft would allow an assessment of competitive flow if dilatation of the string occurred. These events would provide evidence of protection of the region by the internal mammary artery and thus justify the clinical experiment. Have you had experience with this? DR SPNC: I agree with you. In fact, this technique has been reported in The Annals of Thoracic Surgery by Dr Kitamura and his associates [l] from Japan. They temporarily occluded the recipient coronary artery with a percutaneous angioplasty balloon, and were able to demonstrate anatomic patency of a shrunken mammary artery in 2 patients. Reference 1. Kitamura S, Kawachi K, Seki T, Sawabata N, Morita R, Kawata T. Angiographic demonstration of no-flow anatomical patency of internal thoracic-coronary artery bypass grafts. Ann Thorac Surg 1992;53:15&9.