Preparation of the internal mammary artery graft

J THORAC

CARDIOVASC SURG

1989;98:73-9

Preparation of the internal mammary artery graft Which is the best method? Early reports questioned the adequacy of flow of the internal mammary artery when used routinely as a bypass graft. "Adequate" mammary artery flow is now contested only in certain situations, that is, left ventricular hypertrophy, acute myocardial infarction, and reoperations. To compare the methods of mammary pedicle graft preparations with free mammary artery flow, we studied 31 patients who had the left internal mammary artery harvested for elective coronary artery bypass grafting. Group I comprised 14 patients whose mean body surface area was 1.91 m', Systolic, diastolic, and mean arterial blood pressures, left atrial pressure, and heart rate were recorded and stabilized during flow measurements. Free flow of the internal mammary artery was measured before any phannacologic manipulation and ranged from 5 to 44 m1jmin (mean 18 m1jmin). The grafts were sprayed and wrapped in sponges soaked in diluted papaverine solution (60 mg in 40 mI normal saline) for an average of 21 minutes. Free flow ranged from 10 to 108 m1jmin (mean 51 m1jmin). Intraluminal papaverine of the same dilution was then injected with hydrostatic dilatation. Immediate internal mammary artery flows rose from 150 to 333 m1jmin (mean 229 m1jmin). Group II comprised 17 patients who had internal mammary artery takedown under the exact conditions used in group I. Mean body surface area was 1.89 m', Mammary artery pedicles were injected with diluted papaverine throughout their lengths with size 25 needles. After an average of 19.5 minutes, free flow ranged from 28 to 132 m1jmin (mean 69 m1jmin). Intraluminal diluted papaverine was then administered as in group I, and flows increased from 144 to 280 m1jmin (mean 198 m1jmin). The distal internal mammary arteries in both groups were 1.75 to 2.5 mm in internal diameter at the site of arteriotomy for flow measurement. This study shows that all mammary arteries are in spasm immediately after harvest and that flow is inadequate before any pharmacologic intervention. Although extraluminal vasodilators will increase free mammary artery flow, intraluminal papaverine followed by hydrostatic dilatation raises free flow to maximal capacity. Subsequent graft spasm has not been observed. Intraluminal papaverine preparation of the internal mammary artery (1) is the method of choice to determine the true mammary flow capacity and to offer the best possible mammary flow to the myocardium, (2) improves mammary graft size and thereby makes technical anastomotic errors less likely, (3) is useful to distinguish mammary low flow caused by technical error from that caused by vasospasm, and (4) allows identification of unclipped or uncauterized internal mammary artery side branches that otherwise might necessitate reexploration for bleeding.

Noel L. Mills, MD, and Walter L. Bringaze III, MD, New Orleans. La.

EarlY reports questioned the ability of the internal mammary artery (IMA) to offer adequate flow as a coronary artery bypass. Over the past 5 years, however. From the Department of Surgery. The Cardiology Center. New Orleans. La. Read at the Fourteenth Annual Meeting of The Western Thoracic Surgical Association in Hawaii, June 22-25. 1988.

Address for reprints: Dr. Noel Mills. Surgical Director. Cardiology Center, 4500 Wichers Drive (Marrero). New Orleans. LA 70072.

there has been a marked increase in the use of the IMA, and reports have determined that in most situations flow is "adequate." 1. 2 Some investigators have avoided use of the IMA in patients with acute myocardial infarction, left ventricular hypertrophy, or reoperations in which a patient's saphenous vein graft is being replaced by an IMA graft. IMA grafts under such conditions have been considered not to have an adequate flow. We have used intraluminal papaverine for preparation of IMA grafts since 1972. In a prior report, free flow in 100 consecutive IMA grafts treated with intra73

74

The Journal of Thoracic and Cardiovascular Surgery

Mills and Bringaze

Table I. The pH of commercial papaverine alone and diluted in normal saline. electrolytes, and blood pH

Solution

Lab I

Lab /I

Papaverine net Electrolyte solution* Normal saline Papaverine (60 mg) + normal saline (20 ml) Papaverine (60 mg) + normal saline (40 ml) Papaverine (60 mg) + electrolyte solution* Papaverine (60 mg) + electrolyte solution* Blood Papaverine (60 mg) + blood (20 ml)

3.21 7.34 5.41 4.17

3.58 7.41 5.75 4.65

4.33

4.83

5.22

5.63

5.30

5.83

7.42 7.12

7.33 7.19

Laboratory I is NorthShore Regional Medical Center (Puckett Laboratory). Slidell. Louisiana. Laboratory" is West Jefferson General Hospital (Marrer~). New Orleans. Louisiana. "Plasma-Lyte. Travenol Laboratories. Inc.. Deerfield. Ill.

luminal papaverine averaged 195 nil/min.' In that study in a subgroup of patients with controlled conditions, it was found that the coronary peripheral vascular bed, and not characteristics of the IMA graft itself, was the major fact limiting flow. The method of preparation of the IMA graft appears to be the major determinant of IMA total flow capacity. Three techniques of preparation of IMA grafts were studied to identify the method for obtaining optimal IMA graft flow.

Patients and methods Group I. Fourteen consecutive patients (10 men and four women) underwent takedown of the left IMA in preparation for coronary bypass grafting by a previously reported technique" Body surface areas ranged from 1.53 to 2.21 m/, with an average area of 1.91 m'. Immediately after takedown of the IMA, total body heparinization was done in preparation for cardiopulmonary bypass. The IMAs were harvested 2 to 3 em distal to the point of bifurcation and were transected only after heparinization. The IMAs were opened with fine scissors proximal to the bifurcation and at a point near the estimated locus of anastomosis to the coronary artery. Thereafter, the arteriotomies were not extended more than of I em or past a significant branch for this study. The pericardium was opened, and a left atrial pressure line was put in place for continuous measurement throughout the operation and in the postoperative period. Mean arterial pressure and systolic and diastolic pressures were recorded. Cardiac outputs were measured by the thermal dilution technique in selected patients according to the personal preference of the anesthesiologist for monitoring with SwanGanz catheters (Baxter Healthcare Corporation, Bentley Laboratories, Inc., Irvine, Calif.),

Flow No.1: No papaverine. A free flow of the IMA was measured before any pharmacologic manipulation. Flow No.2: After papaverine-soaked sponge. Papaverine dilution (60 mg in 40 ml of normal saline) was sprayed on the IMA graft throughout its whole length with a small syringe and size 25 needle. A papaverine-soaked 4- by 8-inch sponge was then used to wrap the IMA graft in its entirety, and it was set aside during cannulation for cardiopulmonary bypass. The time interval that the graft was set aside until completion of cardiac cannulation was recorded. Free flows were again measured; if necessary, transfusion with the pump was done before measurement of flow so that the left atrial pressure and mean arterial pressure were within Y1c of those pressures recorded during the initial flow measurement. No vasopressors or vasodilators were used. Flow No.3: After IMA intraluminal dilatation. An olivetipped (I mm) metal needle (DLP, Inc., Grand Rapids, Mich., No. 31001) was then tied into the distal end of the IMA graft. The needle tip was withdrawn to the level of the tie to prevent abrasion of the IMA intima by the intraluminal portion of the needle. Papaverine, 3 to IS ml (mean, 10.5 ml), was infused. The graft was set aside for 4 to 30 minutes (mean 13 minutes) before any mechanical dilatation. Careful hydrostatic dilatation was then done sequentially with finger occlusion of the IMA graft to a point near its origin. If necessary, mean arterial pressure and mean left atrial pressure were again adjusted to the values recorded during previous IMA flow measurements. A final free flow of the IMAs after intraluminal papaverine was then measured. The dilated size of the IMA graft was measured with a calibrated probe. For the remainder of the operation, a standard technique for coronary artery bypass grafting was done" Group II. Seventeen patients (II men and six women) had takedown of IMAs as in group I. Body surface areas ranged from 1.55 to 2.20 me, with a mean of 1.89 me. Hemodynamic measurements were performed and controlled as in group I. Flow No.1: After IMA pedicle injection. The body of the IMA pedicles was injected using a size 25 needle throughout the whole length with diluted papaverine-saline solution at room temperature. No solution was injected into the wall or lumen of the grafts. After a period of 12 to 28 minutes (averaged 19.5 minutes), free flow was measured. Flow No.2: After IMA intraluminal dilatation. Intraluminal papaverine, 60 mg in 40 ml normal saline, was injected as in group I with amounts ranging from 6 to 12 ml (mean 7.2 ml). Hydrostatic dilatation was done. with the final dilatation being performed with finger compression of the IMA against the undersurface of the clavicle as in group I. The IMA distal lumen size was measured.

Results The pH of commercial papaverine hydrochloride alone and diluted in normal saline, electrolyte solution (Plasma-Lyte), and blood was measured by a hospital laboratory and a commercial laboratory. The results are shown in Table L Group I (Table II) Flow No.1. Measurement of the IMA immediately after takedown before any pharmacologic manipulation resulted in flows of 5 to 44 nil/min (mean 18 mIl min).

Volume 98 Number 1

Preparation of IMA graft

July 1989

75

Table II. Comparison offlows after topical and intraluminal papaverine preparation of left IMA grafts in 14patients

Age Patient

2 3 4

5 6 7 8

9 10 11 12 13

ivr]

54 50 63 5\ 63 70 69 59 60 69 72

14

68 60 69

Mean

62.6

Sex M M

F M M M

F M M M M

F F M

BSA (m:!

1.91 2.07 1.82 1.89 1299 1.96 1.53 1.89 1.94 2.05 2.21 1.71 1.81 1.92 1.91

Time: firstsecond flow (min)

Time: secondthird flow (min)

First flow

Second flow

Third flow

(mlfmin}

(mlfmin}

(mlfrnin)

Final dilated IMA size (mm)

36 24 20 16 12 12 8 12 20

84 40 108 40

162 260 204 276 228 208 150 304 168 228 334 318 150 216

2.5 2.5 2.0 2.0 2.0 2.5 2.0 2.5 2.0 2.0 2.5 2.5 2.25 2.5

19 21 21 8 19 20 12 35 18 23 18 27 37 10

4 11 11 13 10 15 25 30 6 14 19 14 7 5

12 12 15 12 5 12 15 6 5 5

229 (150-334 )

2.27

21

13

10.5

8

44 20 8 5 18 (5-44)

72

44 16 48 48 60 68 56 16 10 51 (10-108)

Total intraluminal papaverine (m/)

5 15 15 13

BS.·\. Body surface area.

Flow No.2. After topical application of papaverine and a mean waiting period of 21 minutes, free flows of the IMA grafts ranged from 10 to 108 ml/rnin (mean 51 ml/rnin). Flow No.3. After intraluminal injection of dilute papaverine with hydrostatic dilatation and a mean waiting time of 13 minutes, free IMA flows were measured from 150 to 333 ml/rnin (mean, 229 ml/ min).

Intraluminal diameters at the site of arteriotomy for the free flows measured 2 mm for six grafts, 2.25 mm for one graft, and 2.5 mm for seven grafts. Group II (Table III) Flow No.1. After IMA pedicle injection with papaverine solution and an average waiting period of 19.5 minutes, free flows ranged from 28 to 132 ml/rnin (mean 69 ml/rnin). Flow No.2. Intraluminal papaverine with hydrostatic dilatation raised the IMA flow rates to 144 to 280 ml/rnin (mean 188 ml/rnin). One of the IMAs measured 1.75 mm; six 2.0 mm; two 2.25 mm; and eight 2.5 mm in internal diameter at the site where the IMAs were opened for flow measurement and anastomosis. Discussion From the preceding data it is clear that there is wide variation in flow of the IMA graft before and after both external and internal application of diluted papaverine.

We have accepted 120 ml free flow as the minimum for the use of an IMA graft. If such flow is not achieved after intraluminal papaverine infusion, the sternal retractor is released and flow is again measured. If at that point the free flow is still below 120 ml, the IMA is divided at its origin and used as a free graft after patency is assumed with a warmed Parsonnet probe. The benefits of intraluminal administration of papaverine are more than that of increasing the free flow to over four times that of an IMA graft prepared with external papaverine. With a larger diameter vessel to work with, the surgeon is less apt to make a technical error (Fig. I). Potential bleeders from side branches are readily identified. A vascular paralysis occurs, and early spasm of the graft is avoided. In some centers, anecdotal cases of IMA spasm have been noted in patients with low cardiac output on vasopressor therapy.' Use of intraluminal papaverine with gentle mechanical dilatation when necessary, as described herein, has avoided spasm of the IMA. In contrast, spasm of the saphenous vein graft that necessitated manual redilatation of the graft before closing the sternum has been observed on rare occasions. This study does not address the actual use of the IMA during the setting of an acute myocardial infarction during the perioperative period. Takedown and preparation of the IMA is more time consuming than that of a saphenous vein graft, and it may be that in the setting of an acute infarction with hemodynamic changes it is

The Journal of

76

Thoracic and Cardiovascular Surgery

Mills and Bringaze

Fig. 1. Graft stenosis is not uncommon when the coronary anastomosis is performed with the IMA in spasm. With a larger vessel to work with after intraluminal papaverine, infusion, the surgeon is less apt to make a technical error.

Table m. Comparison offlows after IMA pedicle injection and intraluminal administration of papaverine in 17 patients ToTal Age Patient

(yr)

Sex

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

62

M M

Mean

71

78 48 68 66 54 65 70 74 63 58 67 56 73 76 59 65.2

F

M F

M F

M M M F

M M F

M F

M

BSA (me)

2.11 1.88 1.59 2.20 2.01 1.82 1.80 1.84 1.95 1.69 1.89 2.19 1.83 1.79 1.85 1.77 1.90 1.89

Timefirstsecond flow (min)

intraluminal

22 14 15 10 22 18 9 14 36 18 26

6 12 9 9 12 12 6 12 9 9 5 15 12 12

(low

Second flow

Dilated IMA size

tmlfmin!

(mlfmin}

tmm]

28 96 60 36 42 112 96 114 42 90 132 36 30 84 56

228 168 144 198 216 186 133 252 192 216 168 172 280 204 222 162 210

2.0 2.5 1.75 2.0 2.5 2.5 2.25 2.5 2.5 2.5 2.0 2.5 2.5 2.0 2.25 2.0 2.0

10 17 24 34 16

69 (28-132)

198 (144-280)

2.25

19.5

Firsl

48 72

unwise to use the IMA, even though the flow after proper preparation is more than adequate to supply the appropriate area of the heart. With administration of papaverine there rarely may

27

papaverine tmli

l'i

9 6 10

be a transient blood pressure drop. However, if the anesthesiologist is alerted, this can be easily treated by drugs, or fluid volume can be given by way of the arterial cannula by the perfusionist.

Volume 98 Number 1

Preparation of IMA graft

July 1989

Commercial papaverine is an acidic solution with a pH of 3.0 to 4.5. The pH of normal saline is 5.2, and dilution of 2 ml of papaverine (2 ml = 60 mg = I ampule) does not appreciably alter the pH of normal saline. Therefore, we have not been concerned about papaverine injury to the IMA graft. Review of postoperative angiograms since the inception of this technique in 1974 has not revealed any instances in which papaverine could possibly be implicated as causing IMA injury. Papaverine added to heparinized crystalloid solutions will result in a white precipitate with loss of the drug effect. Therefore, heparinized solution without blood buffers is not used. Heparinized blood with papaverine may be used with confidence (Table I). Inadequate size of an IMA is not a realistic problem when the graft is properly prepared. In carefully analyzing 2053 consecutive IMAs, we found the arteries to be too small on four occasions (0.19%). The one concern noted with our use of intraluminal papaverine* over the years is that of incurring a dissection with manual dilatation. This is usually due to IMA injury at a side branch incurred during dissection. If the dissection is detected immediately, the IMA may be salvaged as a free graft. Apart from this study, our routine is to insert the olive-tipped needle into the larger of the distal branches of the IMA after its division and to tie it in place. Papaverine is then slowly injected while the artery is carefully observed with loupe magnification. Dissection is avoided because no hydrostatic dilatation is performed at this time. After 3 to 6 ml of diluted papaverine solution is instilled into the graft, it is set aside, the pericardium is opened, and cannulations are done. Often the IMA becomes fully dilated in the interim. If not, careful manual hydrostatic dilatation up to the clavicle is done. Rarely we have found spasm near the origin of the IMA. This might be due to bleed-off through a sizable pericardiophrenic artery, which would have left a proximal segment of IMA that had not had contact with the dilating solution (Fig. 2). Although we prefer to avoid any mechanical dilatation, this condition may be treated by passage of a 2 or 2.5 mm Parsonnet probe (placed momentarily in warm saline with the stylet removed) to the level of the subclavian artery. A graft that must be discarded because of "inadequate size" is extremely rare. Likewise, a legitimate concern arises when papaverine solution is used to inject the IMA pedicle. Inadvertant instillation of papaverine into the arterial wall with graft damage and/or obstruction has been observed. Undiluted papaverine with its acidic pH injected adjacent to

'Eli Lilly & Company. Indianapolis. Ind.

77

Fig. 2. Spasm of the proximal IMA between its origin and the takeoff of the pericardiophrenic artery.

the IMA wall may also result in an injury with late IMA stenosis (string sign). Other agents are available for use to prepare such arterial grafts as the IMA or right gastroepiploic arteries. To our knowledge there are no formal studies testing such vasodilators as nitroglycerin, sodium nitroprusside, or calcium-channel blockers. REFERENCES 1. Okies JE, Page US, Bigelow JC, et al. The left internal mammary artery: the graft of choice. Circulation 1984; 70(Pt 2):1213.

2. Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internal mammary artery graft on IO-year survival and other cardiac events. N Engl J Med 1986;314: 1. 3. Mills NL. Physiologic and technical aspects of internal mammary artery--coronary artery bypass graft. In: Cohn LH, ed. Modern technics in surgery. Cardiac-thoracic surgery. Mount Kisco, New York: Futura, 1982:1-19. 4. Ochsner JL, MillsNL. Coronary artery surgery. Philadelphia. Lea & Febiger, 1978. 5. Sarabu MR, McClung JA, Fass A, Reed GE. Early postoperative spasm in left internal mammary artery bypass grafts. Ann Thorac Surg 1987;44: 199.

The Journal of

78

Mills and Bringaze

Discussion Dr. Quentin R. Stiles (Los Angeles, Calif). I like this paper. It may even make me modify the way I do coronary surgery. Fifteen years ago, when I was one of the few surgeons in our area using the IMA routinely, and not absolutely sure that this was the correct thing to do. there were criticisms that the flow from the IMA would be inadequate to meet the demands of the left anterior descending (LAD) vascular bed. To satisfy myself, I routinely measured the flows in all the coronary grafts just before closing the chest. We calibrated the electromagnetic flowmeter frequently and believed it was reasonably accurate. Bert Meyer never used IMAs and I routinely did. I compared his flows in 100 consecutive LAD vein grafts to 100 consecutive IMA grafts to the LAD that I did. The average vein graft flow was 86 mljmin compared to 84 mljmin for the IMA grafts. That indicated to me that the limiting factor to flow, at least for the anesthetized patient at rest, was the vascular resistance of the LAD coronary bed and not the conduit. This study was never published. but it gave me great confidence in the use of the IMA. I have never done anything to try to distend the vessel. Frequently I have noticed that the arterial spasms develop from the dissection while the artery is being taken down. I always use the IMA as the last graft that I construct, and by that time the spasm seems to be gone. I always have avoided dilating or distending the artery for fear of damaging it. Only rarely would I discard an IMA because of poor flow when I opened it just before doing the anastomosis. For years I saw no reason to change this policy because I was not recognizing any problems. In recent years I have been using both IMAs and frequently using them in sequential fashion. Once again, some doubts of the flow adequacy of the IMA are creeping into my thoughts. Dr. Mills has given me some reasons to consider altering my technique. I know him well and have great respect for his thinking. I must admit that the larger the IMA and the greater the flow from it, the greater is the surgeon's satisfaction. I really have no good way to know for sure whether in some cases residual spasm of the IMA may be limiting the flow to the grafted coronary bed. Also, my naive confidence in the IMA in the past has very likely led me to occasionally use an intact IMA with flow limited by ostial stenotic disease in the subclavian artery. Dr. Mills' method gives us a guideline whereby these patients would be better served by dividing the IMA proximally and using it as a free graft. Dr. Mills' experience tends to assure me that the risk of injury to the delicate IMA is minimal from intraluminal distention. He has done enough of these operations and has restudied sufficient numbers of patients that he would know by now. I think I will begin to try his method. I have a few questions. Dr. Mills. you imply. but do not state, the conditions under which you measured the flows. Will you confirm them for me? As I understand the method. you made an arteriotomy in the IMA at approximately the site that was to be used for the anastomosis, and it was from this arteriotomy that the free flow was measured. Furthermore, all of the flow measurements were made before bypass was begun, when the heart was supporting the circulation. Are these assumptions correct? Dr. Mills. That is correct. All of this flow is measured before we begin using the pump. That has been our routine over the years except in the rare instance when someone becomes hypotensive and unresponsive to drugs, and with

Thoracic and Cardiovascular Surgery

good anesthesia that problem does not occur much anymore. Dr. Stiles. When I start distending the IMA, I want to make sure that I am doing it in a safe way. Exactly how is the intraluminal infusion performed? Is it done with a hand-held syringe? If so, is there any means to measure or limit the pressure applied to the inside of the artery? We know that applying excess pressure to a saphenous vein damages the delicate intima and decreases the patency rate of the vein when it is used as a coronary graft. Is there any assurance that we won't introduce a similar risk for IMA grafts by your method unless we carefully control the conditions of the distending maneuver? Dr. Mills. I will start with the last part of your question. As you know, occasionally there will be an anesthesia mistake and too much anesthetic will be given: A blood pressure of 280 mm Hg will develop because hypotension was treated with a vasopressor drug. We then have to wait for 5 to 10 minutes until the blood pressure stabilizes. That has led me to believe that a transient increase in pressure in the IMA probably is not detrimental. In any event, carefulness is the answer. We use a I mm DLP olive-tipped needle, tie it into the IMA, and with the first injection we avoid any hydrostatic pressure at all. We put in about 5 to 9 ml of dilute papaverine and let it sit. It is amazing that that, in itself, will cause the IMA to dilate. If the vessel is not fully dilated, then we pinch 3 or 4 em from the needle and carefully begin to give it hydrostatic dilatation sequentially all of the way up to the undersurface of thr clavicle. Dr. Stiles. But when you first do it, you don't pinch the IMA? Dr. Mills. When we first do it, we put the papaverine in, set it aside, and let nature do the dilating, so to speak. The key thing is to be sure there are no little side branches that have been tugged on too hard to incur a dissection. Then pressure is carefully applied to the 3 ml syringe. Dr. Arun K. Mittal (Torrance. Calif). I have one suggestion, one observation, and one question. We used the IMA in 50 or 60 patients in 1972 and abandoned its use because we did not know the long-term result. After seeing those patients restudied, we have started using the IMAs again in the past 4 or 5 years. Lately, we have been using a sequential graft with three to five grafts from one or two IMAs. My suggestion is that the IMA be dissected to the terminal branch, a small size 23 or 25 arterial polyethylene line needle be inserted, and papaverine solution be injected gently. If there is any resistance the injection should be stopped, because it is very easy to dissect the IMA. After injecting the solution, we apply a clamp and proceed with the plastic nontraumatic cannulas and the saphenous vein bypass graft. Last, we implant the IMAs. In some patients, we noticed atheromatous material when we opened the IMA for 1.5 or 2 ern. Dr. Mills just clarified that this material actually is crystallized papaverine, but this papaverine was diluted in saline solution and was given before the heparin was given. After transecting the IMA for another 1.5 or 2 ern, we found that the diameter varied from 2 to 4 mm and the flow was always good. At first we worried that these patients had atheromatous changes, although the material looked like crystallized material. My question is this: What would be the long-term effect of

Volume 98 Number 1 July 1989

papaverine on endothelial lining with an IMA graft? Lately, we have not used papaverine infusion because of this observation, and we did not know whether it might cause damage or affect the long-term patency of the IMA. On four occasions last year, we did not notice atheromatous changes in the IMA. We have had one case of subclavian artery steal in a patient with bilateral IMA implants. Dr. Mills. I will address the question primarily by noting that I think the "atheromatous changes" are precipitate of heparin and papaverine combination. The drug company clearly says that the two drugs are incompatible. If there is heparin in blood with the blood buffers, there is not a problem. Insofar as injuring the IMA itself, I don't think it is a pH problem, because saline has a low pH. The test of time with repeat angiography over the course of years has led us not to worry about giving diluted papaverine in the IMA, and we believe the benefits far outweigh any possible ill effect. The benefits are that one has an immediate adequate flow in the graft and the chance of a technical error is much less. Dr. Thomas J. Fogarty (Palo Alto. Calif). Noel Mills visitedour group about 8 years ago, and he is really responsible for anything we know about handling the IMA. I think we have been good students. He left us with several premises-to optimize flow, optimize size, and minimize mechanical trauma. For the first 3 years we used essentially the technique he has outlined, and then we began to change somewhat in response to our practice. Currently, 30% of our volume represents repeat coronary artery operations, and in a significant number of patients we have found that the intraarterial injection of papaverine and hydrostatic dilatation did not always optimize flow. We altered the technique in two ways. the first of which involved incising the investing fascia that overlies the IMA. This gives several advantages. The first, and in this we agree with Dr. Mills. is that most of the damage that will occur in the IMA actually occurs during the dissection, or takedown, including heat trauma and mechanical trauma. On incising that investing fascia, one will encounter and recognize any intimal damage that has occurred, because the incision exposes the undersurface of the IMA. Second, it also exposes any fibrous bands that may serve to constrict flow. We have found that this occurs in approximately 3% of the repeat operations. Thus, if the fascia is directly incised, the fibrous bands are exposed and can be served. Incising the fascia also increases the linear length of the IMA by approximately 10% to 15%, which allows the advantage of doing increased numbers of sequential grafts. The dilatation process, as Dr. Mills has outlined in his manuscript, must be undertaken with considerable caution. Overdistention will cause damage. Hydrostatic dilatation is an appropriate way to dilate the IMA in most situations. However, with proximal spasm the IMA becomes extremely difficult. We are so concerned about the damage that we took measurements in the laboratory and found that, no matter what dilatation method is used, if the dilatation pressure does not exceed a sheer force of 50 grn, there is no histologic difference in the artery at 1 month, whether by electron or light microscopy. As a result, we take dilatation sequence quite seriously and now use an air-filled balloon. To limit the sheer force, we use a gauge that actually limits the sheer force to less

Preparation of IMA graft

79

than 50 gm. This is easily accomplished by inserting the balloon in the uninflated position, inflating it with 0.12 cc of air when it reaches a size of approximately 2 to 2.5 mm, and then pulling up on the sheer force gauge. This method has been very effective. We have done 900 procedures now. The complication rate since we began using a sheer force gauge has been zero, and the mean flow rates approach 40 ml, effecting an increase in the flow rate of threefold to eighteenfold. The mechanism by which one dilates the artery is critically important. Dr. Mills outlined in his study the use of a Parsonnet probe, which I think is very appropriate. There were two reasons for using this probe-to identify any damage that may occur with the proximal portion to increase flow. I have a question to ask that may help us with another study with mechanical dilatation of the proximal segment. Have you in any of these patients measured the difference in flow after hydrostatic dilatation and after mechanical dilatation? Dr. Mills. Yes, we have. In situations in which the flow does not seem adequate, we use a warmed Parsonnet probe that is very carefully inserted up to the subclavian artery. It is that last inch or so that is in spasm, and the spasm is probably due to the bleed-off from the pericardiophrenic artery. Then we again measure flow, and we have found the flow adequate in the majority of patients. If not, there is something wrong, such as a subclavian stenosis that has been missed or some other problem obstructing free flow. Dr. Fogarty. The second question I would like to ask because you have been so helpful in other areas. I know this isn't the subject of this paper, but have you uncovered any tricks in increasing the linear length of the IMA short of transection and using it as a free graft? Dr. Mills. In one of our previous publications, we advised opening up the fascia over the last 5 to 6 ern routinely, not only for a reoperation, but in the routine case. Occasionally, the electrocautery heat that is generated causes little fibrous bands over the distal part of the IMA, and this problem can be prevented. You have made an important point, which is to open up the fascia. We do have three tricks to get the right IMA to the posterior descending in the majority of patients. The only patient who would cause difficulty would be the extremely obese, very short patient with a transverse heart. The three tricks for the right IMA are these: First, divide the pericardium down to the superior vena cava, being careful not to injure the phrenic nerve. Second, on the right side the IMA and internal mammary vein come off at different levels. The vein comes off first, and the IMA continues on for another 3 or 4 ern. On that side only, we skeletonize the IMA almost to its origin. That provides an inch or more of length. Because the vein is connected posteriorly on this pedicle, it will easily drop distally. The third trick for increasing the right IMA length is to dissect the IMA down across the undersurface of the costal cartilages into its two, or sometimes three, branches. Choose the larger branch for the anastomosis, again measuring prebypass flow and using the 120 mljmin volume as the minimum acceptable volume. I have rarely resorted to Dr. Cosgrove's technique of sequential transverse incisions in the IMA to gain length.