- Email: [email protected]
Selective arterialization of the coronary venous system
Volume 77, Number J
January 1979
THORACIC AND CARDIOVASCULAR SURGERY The Journal of
Original Communications
Selective arterialization of the coronary venous system Encouraging long-term flow evaluation utilizing radioactive microspheres The long-term effectiveness of a retrograde coronary venous bypass graft (CVBG) to an ischemic left ventricle was evaluated in 18 dogs. A saphenous vein was interposed between the aorta and left anterior descending (LAD) vein. The LAD vein was ligated cephalad to the CVBG to prevent an arteriovenous fistula. The LAD artery was ligated at its origin to create anterior wall ischemia. Operative graft flow averaged 53 mi. per minute. The 14 surviving dogs were catheterized 3 to 5 months later. Ten of the 14 CVBC's were patent angiographically . The chests were opened and graft flow now averaged 50 mi. per minute. 141Ce microspheres were injected into the left atrium to measure myocardial flow to the anterior wall. In the 10 dogs with patent grafts, transmural flow was 39 ± I (5.E.M.) mi. per /00 Gm, of tissue per minute. The endocardiallepicardial flow ratio was 1.4/ I, indicating that retrograde venous perfusion effectively delivered blood to the subendocardium. After ligation of the CVBC, microsphere measured flow dropped to 15 ± I mi. per 100 Gm. per minute. In 15 control dogs, anterior wall flow was 100 ± 3 mi. per 100 Gm. per minute, decreasing to 13 ± 2 mi. 45 minutes after ligation of the LAD artery and vein. None of the eight control dogs with simple ligation of the LAD artery and vein survived more than 5 days. Histologic examination of the anterior wall of the left ventricle, the area served by the CVBC's for 3 to 5 months, disclosed no evidence of venous sclerosis or thrombosis and no evidence of interstitial edema or hemorrhage. Thus a CVBC permitted long-term survival in an otherwise nonviable anatomic preparation. Moreover, restoration offlow with a CVBG was effective because it perfused all layers of the myocardium, especially the subendocardium-the crucial layer of myocardial muscle.
Mark S. Hochberg, M.D. (by invitation), William C. Roberts, M.D. (by invitation), Andrew G. Morrow, M.D., and W. Gerald Austen, M.D., Bethesda, Md., and Boston, Mass.
From the Clinic of Surgery, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Md. 20014, and the Department of Surgery, Massachusetts General Hospital, Boston, Mass. 02114.
Read at the Fifty-eighth Annual Meeting of The American Association for Thoracic Surgery, New Orleans, La., May 8, 9, and 10, 1978. Address for reprints: Mark S. Hochberg, M.D., Department of Surgery, Massachusetts General Hospital, Boston, Mass. 02114.
The Journal 01
2 Hochberg et al .
Thoracic and Cardiovascular Surgery
Coronary artery bypass grafting (CABO) is now the procedure of choice for significant but discrete coronary arterial disease." However, of the 60,000 2 CABO patients operated upon yearly, a group of them have angiographically extensive atherosclerotic changes with poor runoff so that the CABO procedure is unlikely to succeed . This group is generally agreed to comprise 12 to 20 percent of revascularization candidates .":" In addition to these patients, many centers are beginning to see increasing numbers of patients with occluded bypass grafts and persistent angina. It is presumed that these individuals with failed CABO's have poor vascular runoff. Although we 7 have reported that coronary endarterectomy can be successful in certain circumstances, in general this procedure is not very useful for patients with (I) diffuse arterial disease (especially of the left system) or (2) previously failed CABO procedures.
CVBG (
Recently , weB have reported anterior left ventricular myocardial flow characteristics via selective retrograde coronary venous perfusion in a set of short-term experiments . In 15 animal experiments, the effectiveness of venous perfusion to the subendocardium, myocardium, and subepicardium was established with radioactive microspheres used to measure the flow. The present study is aimed at evaluating the longterm effectiveness of coronary venous bypass grafts (CVBO) functioning for 3 to 5 months. Methods Eighteen American foxhound dogs weighing 23 to 28 kilograms had saphenous vein-coronary venous bypass grafts interposed between the aorta and the left anterior descending (LAD) vein. The animals were anesthetized with thiamylal sodium and succinylcholine, intubated , and ventilated
! )
CORONARY VENOUS) . . . . BYPASS GRAFT
RCA \
\
<,
L. CIRCUMFLEX A.
LI GATION OF GREAT CARDIAC V.
LAD VEINS '-/
Fig. 1. Diagram of a completed coronary venous bypass graft (CVBG) . Note proximal ligation of LAD artery to render the anterior wall ischemic . The LAD vein was also ligated cephalad to the CVBG to prevent an arteriovenous fistula . RCA, Right coronary artery . LAD, Left anterior descend ing.
Fig. 2A. The photograph on the left shows normal coronary anatomy with the arteries red and the veins blue. The photograph on the right illustrates that during retrograde coronary venous perfusion , the arteries contain blue (desaturated) blood and the veins carry red (oxygenated) blood. (These photographs are from previously cited short-term experiments" with inflow into the LAD coronary veins accomplished via a shunt from the carotid artery) . Fig. 2B. Completed coronary venous bypass graft ju st prior to chest closure. Note that the diagonal coronary artery is now blue with desaturated blood , and its accompanying veins are now red with oxygenated blood (arrow).
RETROGRADE CORONARY VENOUS PERFUSION
NORMAL HEMODYNAMIC ANATOMY
/
I
I
I
/
,, ,, L.ANT. DESCENDING ARTERY
L.ANT. DESCENDING VEIN
/
/
I
L. ANT. DESCENDING ARTERY
L.ANT. DESCENDING VEIN
For legend see facing page.
Volume 77 Number 1
Arterialization of coronary venous system
5
January. 1979
Fig. 3. Dye injected into the aortic root of a dog revealing a patent coronary venous bypassgraft 4 months after operation. with a Harvard respirator. Halothane anesthesia was continued throughout each procedure. Measurements of arterial, left atrial, and central venous pressure were made with Statham P23Db transducers and continuously recorded on a Sanborn 350 recorder. The electrocardiogram was also monitored. The saphenous vein was removed from the lower part of the left hind leg. The chest was opened through the fourth left intercostal space. When the dissection was completed, the animal was heparinized and total cardiopulmonary bypass was instituted via the right atrium and right femoral artery . A left ventricular vent was placed through the left atrial appendage. The animal was cooled to 27° C. and the heart was electrically fibrillated. The aorta was cross-clamped and the vein graft was anastomosed with 6-0 Prolene to the great cardiac vein at the junction of the two LAD veins. Fig. I illustrates the exact anatomic location. The great cardiac vein was ligated cephalad to the anastomosis to prevent an arteriovenous fistula . The LAD artery was ligated at its origin (proximal to the first septal perforator) to precipitate acute ischemia to the anterior wall of the left ventricle (Fig. I). The aortic crossclamp was then removed and replaced with a partially occluding Satinsky clamp. The proximal CVBG anastomosis was performed during the warming phase . When the proximal suture line was secured, the heart was defibrillated. The LAD veins supplied by the CVBG turned red. As previously noted ," the LAD artery was now blue, because it contained desaturated blood . Fig . 2A illustrates this new physiological event. Cardiopulmonary bypass was discontinued and flow was measured through the newly constructed CVBG with Statham
Fig. 4, Selective injectionof a coronary venous bypass graft in an animal 5 months after operation. Dye illustrates perfusion of the anterior wall via the CVSG .
electromagnetic flow probes. Fig . 2B is an operative photograph of a CVBG just prior to the chest closure. Three to 5 months following operation, the animals were electively brought to the angiography suite and catheterized. Injections of dye into the aortic root and selective vein grafts were recorded on film. The dogs were then transported to an operating room and a left thoracotomy was performed. Graft flows were again measured via flow probes. Next, 900,000 141Ce microspheres of the 15 JL size were injected into the left atrium . Reference flow was collected for precisely I minute from a catheter in the left femoral artery. The 15 JL microsphere size was selected for its proved reliability. 9 . 10* After the first injection, the CVBG was ligated, and a second injection with 900,000 80Sr microspheres was made. The animals were put to death and the hearts were prepared for scintillation counting. A full-thickness piece of anterior left ventricular myocardium, approximately 4 by 2 cm ., lying directly in the area served by the LAD veins and artery, was excised . Epicardial fat and large vessels were stripped away . The fullthickness piece was dissected with a razor blade into its three anatomic components: the subendocardium, myocardium, and subepicardium. Each layer was placed in a separate counting vial and weighed. As an internal control on the method, a similar piece of posterior left ventricular wall was excised and dissected into the three layers. Because the posterior wall of the heart is served by the left circumflex coronary artery (which is dominant in the dog), there should have been no change in the flow it received during the *Utley JR: Personal communication .
The Journal of Thoracic and Cardiovascular Surgery
6 Hochberg et al.
Table I. Flow evaluation in 10 patent coronary venous bypass grafts (mI. 1100 Gm.lmin.) Exp. No. I
2 3 4 5 6 7 8 9 10
Subendocardial flow
Myocardial flow
Subepicardial flow
Endo/epi
Avg. transmural flow
36.8 59.3 23.5 55.5 44.8 37.8 34.6 40.5 27.6 35.0
34.5 36.6 47.8 46.9 50.6 37.5 39.4 43.3 36.1 30.4
41.4 35.2 41.3 20.1 13.1 51.6 25.5 39.0 60.1 52.5
0.89 1.69 0.57 2.76 3.42 0.73 1.36 1.04 0.46 0.67
36.8 41.5 39.2 42.8 40.2 41.0 33.1 41.3 40.5 37.9
Average
1.36
Table II. Summary of transmural myocardial flow in the ten patent and four occluded CVBG's (ml. 1100 Gm.Imin.) Ten patent
CVBG open
CVBG ligated
LV anterior wall
39.4 ::!: 0.9
15.2::!: 1.3
LV posterior wall (control)
88.5::!: 2.1
91.2 ::!: 1.9
CVBG's
CVBG occluded
Four occluded
LV anterior wall
14.3 ::!: 1.7*
LV posterior wall (control)
86.4 ::!: 3.9*
CVBG's
Legend: CVBG, Coronary venous bypass graft. LV, Left ventricular. -Average value of two microsphere injections.
two microsphere injections. If mean aortic pressure had differed during any two experimental injections, then the anterior wall flows would have been measured falsely. The knowledge that control posterior wall flows did not differ significantly between injections can permit this study to report that the measured anterior wall flows accurately portrayed the true hemodynamic picture. Histologic examination of the left ventricular anterior wall and saphenous vein bypass was also performed. Fifteen additional foxhounds were studied as a control of the CVBG experimental animals just described. The normal transmural flow to the nonoperated anterior wall was measured. The LAD artery and vein were then ligated and flow again was determined with microspheres at the time of sacrifice 45 minutes later. An additional series of eight dogs underwent ligation of the LAD artery and vein without the construction of a CVBG. It was planned that these dogs would serve as long-term controls for the 18 operated CVBG animals.
39.4::!: 0.9 (S.E.M.)
The tissue samples and the reference blood were counted for their gamma emission in a Beckman multichannel scintillation counter. * A computer program was employed to establish an equation matrix based upon the reference blood radioactivity and weight for each injection. Once the matrix was determined, the actual tissue counts per minute could be entered. The resultant flow in milliliters per 100 Gm. of tissue per minute was computed. This myocardial blood flow technique utilizing microspheres was developed by Rudolph and Heymann. II Buckberg 12 has recently reviewed the mathematical model and commented on the technical aspects of this microsphere method. Thus the hemodynamic effectiveness of long-term retrograde CVBG was evaluated by angiography, operative and postoperative graft flow measurements, histology, and the actual flow to the subendocardium, myocardium, and subepicardium with the use of radioactive tracer microspheres.
Results Fourteen of the 18 foxhounds that underwent CVBG survived the postoperative period to the time of sacrifice 3 to 5 months after operation, Angiography. Fig. 3 illustrates an aortic root injection of a CVBG which was patent 4 months after operation. Of the 14 surviving animals, 10 had patent grafts at the time of angiography and sacrifice. In most of these animals, the catheter could then be advanced into the graft itself so that the anatomy could be better delineated, The selective graft injection shown in Fig. 4 clearly demonstrates perfusion of the anterior wall of the left ventricle. This angiogram was performed in an *We are grateful to Mr. Richard Kagan for his advice and the use of his equiprnent.
Volume 77 Number 1 January, 1979
Arterialization of coronary venous system
7
.s: """ ~ ~ ~ .... ~
~
~~ ~ .....
~
~
~ ~ t-.:
100 90 80 70 60 50 40 30 20
to 0
LONGTERM NORMAL ANTERIOR WALL FLOW
PATENT CVBGs
LIGATION OF PATENT CVBGs
LIGATION LAD ARTERY and VEIN
Fig. 5. Bar graphs depicting the relative transmural anterior wall flows of all experiments in this report. Normal anterior wall flow is on the left. Flows were measured via radioactive microspheres.
animal 5 months after the venous bypass procedure. As reported above, four of the 18 animals in which the CVBO was performed did not survive the postoperative period. Necropsy was performed in these animals, and three of them had patent grafts. Because there was no evidence of congestive heart failure, it was thought that death was due to an arrhythmia or an acute myocardial infarction. Graft flow measurement. At the time of operation, operative graft flow was measured at 53.0 ± 3.1 * ml. per minute. At the time of sacrifice 3 to 5 months later, graft flow in the 10 patent grafts averaged 50.7 ± 4.9 ml. per minute. The operative CVBO flow in the four dogs found subsequently to have CVBO occlusion was 32.4 ± 11.4 ml. per minute. Myocardial flow measurements. Flow to fullthickness segments of the left ventricular anterior wall in the 10 animals with patent grafts averaged 39.4 ± 0.9 rnl. per 100 Om. per minute. Of importance, the flow to the subendocardium alone averaged 39.5 ± 3.5 rnl. per 100 Om. per minute with an endocardial/epicardial average flow ratio of 1.36 (Table I). Myocardial flow measurements were also performed in the four animals with occluded grafts. Table II summarizes the transmural flows in the 10 animals with patent CVBO's and the four animals with occluded CVBO's. This table also reports the control flows measured to the posterior left ventricular wall of all 14 dogs. This was done as an internal control on the microsphere method as noted previously. There was no significant difference (p = 0.05) be-
tween the average transmural flow for the four occluded grafts (14.3 ± 1.7 ml. per 100 Om. per minute) and the 10 acutely ligated grafts (15.2 ± 1.3 ml. per 100 Om. per minute).* However, there was a highly significant difference (p ::s: 0.00 I) when the transmural flow in the 10 patent grafts (39.4 ± 0.9 ml. per 100 Om. per minute) was compared to the flows of either the ligated CVBO's (10 animals) or the occluded CVBO's (four animals). These comparisons are graphically depicted in Fig. 5. Control series. The normal transmural flow to a nonmanipulated left ventricular anterior wall was 100 ± 2.5 ml. per 100 Om. per minute in the 15 control foxhounds. This flow fell to 13.5 ± 1.5 ml. per 100 Om. per minute 45 minutes after ligation of the LAD artery and vein (Fig. 5). Eight additional animals had control ligation of the LAD artery and vein. Their chests were closed in the hope of determining left ventricular transmural flow at the time of sacrifice of the experimental CVBO dogs (3 to 5 months after operation). Of great importance. none of these eight control dogs could withstand the acute effects of LAD artery ligation (without CVBO protection) and all died within 5 days of operation. Fig. 5 illustrates the normal anterior wall transmural flow compared to the perfusion provided by all the CVBO experiments and controls that have been described. Pathology. Histologic examination of typical areas perfused by the CVBO for the experimental period is shown in Fig. 6. There is no evidence of venous sclerosis or thrombosis and no evidence of interstitial
*All numbers following a ± sign indicate I standard error of the mean.
*As noted above. in all the experimental animals the LAD artery was ligated at operation to render the left ventricle ischemic.
8 Hochberg et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 6.A. This histologic section shows an occluded arteriole on the right with a patent adjacent venule on the left. (The venule contains some coagulated fibrin which occurred postmortem.) B. Two patent venules with an adjacent artery. There is no evidence of venous thrombosis or sclerosis. The myocardium shows no evidence of hemorrhagic infarction. C. A histologic section of a widely patent coronary venous bypass graft in an animal put to death 4 months after operation.
The experimental findings suggest that retrograde CVBG's can remain patent over a long period of time. This confirms the work of Magovem* and Chiu and Mulder. 13 The important finding of this experimental report is that, for the first time, the precise effectiveness of long-term selective venous perfusion to the three layers of the myocardium was determined with radioactive microspheres. Previous reports of successful
CVBG3-6. 13* have been based largely on coronary sinus oxygen data, electrocardiographic evidence of ischemic changes , microscopic evidence of infarction. radioisotope scanning, and animal death . These measures are at best indirect methods of evaluating retrograde coronary venous perfusion to the heart . The experimental findings of this report illustrate that a CVBG can provide almost three times more perfusion than the base-line flow to a left ventricular wall rendered acutely ischemic by simple LAD artery and vein ligation (or by ligation of a patent CVBG). More simply , a CVBG can sustain life, whereas the control setting resulted in the death of all eight dogs within 5
*Magovem GJ: Personal communication .
*Magovem GJ : Personal communication.
edema or hemorrhage (venous infarction) . A cross section of a CVBG shows a wide-open channel (Fig. 6, C) .
Discussion
Volume 77 Number 1 January, 1979
days of LAD artery and vein ligation. These latter dogs, of course, did not have the protection of venous retroperfusion. Beck ,':' Zajtchuk," and their colleagues found that ligation of the dog's LAD artery was fatal in 70 percent and 77 percent of animals, respectively, observations which confirm our control series findngs. * Not only can the CVBG's sustain life in an otherwise nonviable control setting, but the perfusion is also effective, because the microsphere technique demonstrated flow to all layers of the myocardium. Initial fears that coronary venous perfusion would be ineffective because of venovenous epicardial anastomoses are dispelled by the demonstration that the subendocardium receives, via a CVBG, a level of flow not dissimilar to that of the myocardium and subepicardium. This is proved by an average endocardial/epicardial ratio of 1.4/1. Furthermore, successful perfusion of the subendocardium cannot be explained by the development of arterial collateral channels. The patent CVBG's were occluded just prior to sacrifice, and flow dropped from 39.4 ml. per 100 Gm. per minute to 15.2 (Fig. 5), suggesting true perfusion via the CVBG's. Although the data showed significant reversal of ischemia with retrograde coronary venous perfusion, this study can make no statement as to whether this level of perfusion can prevent an infarct or even relieve angina. Indeed, we expected some measure of infarction, since LAD artery ligation was chosen as the experimental model of ischemia for evaluation of selective retrograde coronary venous perfusion. Since Beck.tv 16 Bakst;" and their associates first attempted to globally retroperfuse the heart by arterializing the coronary sinus, the effectiveness of reversing flow in the coronary venous system has been debated. More recently, many authors":" 8. 13. 15. 18- 2 7 t have studied the possibility of venous retroperfusion to a selective area of the heart. The work has been well summarized by Bates" and Chiu." Most of these models used the left internal mammary artery (LIMA) as the inflow to the coronary venous system. Two important technical points emerge from the literature. The first finding is that inflow must be at least 50 mi. *When ligating the LAD artery. it is important to ensure that occlusion occurs proximal to the first septal perforator which arises very close to the bifurcation of the left main coronary artery. This may account for the varying mortality rates following LAD arterial ligation reported in the literature. t Magovern Gl: Personal communication.
Arterialization of coronary venous system
9
per minute through the graft. Three reports 20. 29. 30 suggest that an LIMA-LAD vein anastomosis usually cannot supply this necessary level of flow. Consequently, in this report a reversed saphenous vein was utilized to establish inflow into the coronary venous system. Second, Hammond and Austen" nicely demonstrated in 1967 that greater than 50 percent of LAD artery blood flow was returned to the heart by routes other than the coronary sinus (anterior cardiac veins as well as direct communications to the heart chambersthe thebesian and arteriosinusoidal channels). Consequently, reports of retrograde perfusion based on coronary sinus sampling must be carefully evaluated. Selective retrograde coronary venous perfusion has stirred a significant amount of controversy in the literature. Encouraging CVBG experiments with various means of evaluation have been cited earlier. 3-6. 8. 13. 24* However Zajtchuk ," Razzuk," Marco;" and their co-workers, using seemingly similar methods of experimental evaluation, found venous perfusion to be of little benefit. The explanation for these less sanguine reports is not immediately obvious. However, it can be speculated that their experiments may have failed for four reasons. First, LIMA inflow may have been inadequate as discussed": 29. 30 None of these papers reported CVBG flow at the time of operation and sacrifice. The second potential technical problem may have been the point of LAD artery ligation. As noted, fatal arterial infarction can be achieved only if the artery is occluded proximal to the first septal perforator. If ligation is performed distal to this branch, "competitive" arterial inflow via this perforator and collaterals may render the LIMA inflow far less significant at the capillary level. This conceivably could explain the finding of LIMA patency with obstruction at the LIMALAD vein anastomosis in one of these reports. 27 Conversely, as illustrated in Fig. 6, we found no histologic evidence of intimal hyperplasia of the coronary venous system with graft flow averaging greater than 50 ml. per minute. The third problem may have been metabolic data sampling from the coronary sinus. As suggested earlier, the greatest portion of blood draining the left ventricular anterior wall returns by noncoronary sinus routes. The fourth question raised by these three pessimistic reports is their operative approach. Two of these three papers reported that the anastomosis was achieved without the aid of cardiopulmonary bypass (the third report does not state the operative technique *Magovern Gl: Personal communication.
I0
Hochberg et
at.
employed). We found a more precise anastomosis could be achieved with a quiet heart. Precise quantification of the hemodynamic effects of retrograde CVBG's to the subendocardium, myocardium, and subepicardium has now been accomplished with the use of radioactive microspheres. In both short-term" and the present long-term experiments, effective tissue perfusion has been achieved. There have been scattered reports of the clinical use of CVBG's with some success.t?: 24. 32. 33 Because the numbers are small, no significant conclusions can yet be drawn. On the basis of the encouraging data presented in this report, it may now be reasonable to offer a saphenous vein-CVBG to patients with (I) diffuse atherosclerosis (not amenable to endarterectomy) or to patients (2) who have failed to benefit from the conventional coronary artery bypass procedure.
Conclusion I. In a series of long-term animal experiments, a saphenous vein-coronary venous bypass graft (CVBG) permitted survival in the presence of LAD artery ligation. By contrast, no animals survived LAD artery ligation without CVBG protection. 2. Flow via a CVBG is effective because it perfuses all layers of the myocardium. 3. The encouraging long-term patency data in animals suggest that judicious use of a venous bypass may be of value in selected patients with diffuse atherosclerosis or a previously failed coronary artery bypass. We wish to acknowledge the excellent technical assistance of Miss Connie Martino and Mrs. Joan B. Fuller.
2
3
4
5
6
REFERENCES Mundth ED. Austen WG: Surgical measures for coronary heart disease (three parts). N Engl J Med 293: 13. 75. 124. 1975 Miller DW. Hessel EA II. Winterscheid LC, Merendino KA. Dillard DH: Current practice of coronary artery bypass surgery. Results of a national survey. J THORAC CARDIOVASC SURG 73:75. 1977 Bhayana IN. Olsen DB. Byrne JP. KolffWJ: Reversal of myocardial ischemia by arterialization of the coronary vein. J THORAC CARDIOVASC SURG 67:125. 1974 Gardner RS. Magovern GJ. Park SB. Dixon CM: Arterialization of coronary veins in the treatment of myocardial ischemia. J THORAC CARDIOVASC SURG 68:273, 1974 Moll JW. Dziathowiak A. Rybinski K. Edelman M, Ratajczak-Pakalska E: Arterialisierung des Sinus Coronarius. lndikationen , Technik , Ergebnisse. Thoraxchirurgie 21 :295. 1973 Arealis, EG. VoIder JG. Kolff WJ: Arterialization of the
The Journal of Thoracic and Cardiovascular Surgery
coronary vein coming from an ischemic area. Chest 63:462. 1973 7 Hochberg MS. Merrill WHo Michaelis LL. Mcintosh CL: Results of combined coronary endarterectomy and coronary bypass for diffuse coronary artery disease. J THORAC CARDIOVASC SURG 75:38. 1978 8 Hochberg MS: Hemodynamic evaluation of selective arterialization of the coronary venous system: An experimental study of myocardial perfusion using radioactive microspheres. J THORAC CARDIOVASC SURG 74:774. 1977 9 Marshall WG, Boatman GB. Dickerson G. Perlin A. Todd EP. Utley JR: Shunting. release. and distribution of nine and fifteen micron spheres in myocardium. Surgery 79:631. 1976 10 Utley JR. Carlson EL. Hoffman HE. Martinez HM. Buckberg GD: Total and regional myocardial blood flow measurements with 25 JL. 15 JL. 9 JL. and filtered I to 10 JL diameter microspheres and antipyrine in dogs and sheep. Circ Res 34:391. 1974 II Rudolph AM. Heymann MA: The circulation of the fetus in utero. Methods for studying distribution of blood flow. cardiac output and organ blood flow. Circ Res 21: 163. 1967 12 Buckberg GD: Studies of regional coronary flow using radioactive microspheres. Ann Thorac Surg 20:46. 1975 13 Chiu CJ. Mulder DS: Selective arterialization of coronary veins for diffuse coronary occlusion. An experimental evaluation. J THORAC CARDIOVASC SURG 70:177. 1975 14 Beck CS. Stanton E. Batiuchok W. Leiter E: Revascularization of the heart by a graft of systemic artery into the coronary sinus. JAMA 137:436. 1948 15 Zajtchuk R. Heydorn WHo Miller JG. Strevey TE. Treasure RL: Revascularization of the heart through the coronary veins. Ann Thorac Surg 21:318. 1976 16 Beck CS. Leighninger DS: Scientific basis for the surgical treatment of coronary artery disease. JAMA 159: 1264. 1955 17 Bakst AA. Costas-Durieux 1. Goldberg H. Bailey CP: Protection of the heart by arterialization of the coronary sinus. II. Coronary flow in dogs with aorticocoronary sinus anastomosis. J THoRAc SURG 27:442. 1954 18 Andreadis P. Natsikas N. Arealis E. Lazarides DP: The aortocoronary venous anastomosis in experimental acute myocardial ischemia. Vase Surg 8:45. 1974 19 Kay EB. Suzuki A: Coronary venous retroperfusion for myocardial revascularization. Ann Thorac Surg 19:327. 1975 20 Benedict JS. Buhl TG. Henney RP: Cardiac vein myocardial revascularization. An experimental study and report of three clinical cases. Ann Thorac Surg 20:550. 1975 21 Williams GD. Burnett HF, Derrick BL. Miller CH: Retrograde venous cardiac perfusion for myocardial revascularization. An experimental evaluation. Ann Thorac Surg 22:322. 1976
Volume 77 Number 1 January, 1979
22 Meerbaum S, Lang TW, Osher JV, Hashimoto K, Lewis GW, Feldstein C, Corday E: Diastolic retroperfusion of acutely ischemic myocardium. Am J Cardiol 37:588, 1976 23 Demos S, Brooks H, Holland R, Balderman S, Anagnostopoulos C: Retrograde coronary venous perfusion to reverse and prevent acute myocardial ischemia. Circulation 49,50:Suppl 3:168, 1974 24 Park SB, Magovern GJ, Liebler GA, Dixon CM, Begg FR, Fischer DL, Dosios TJ, Gardner RS: Direct selective myocardial revascularization by internal mammary artery-coronary vein anastomosis. J THORAC CARDIOVASC SURG 69:63, 1975 25 Bates RJ. Toscano M, Balderman Sc, Anagnostopoulos CE: The cardiac veins and retrograde coronary venous perfusion. Ann Thorac Surg 23:83, 1977 26 Razzuk MA: Discussion of Benedict et al'" 27 Marco JD, Hahn JW, Barner HB, Jellinek M, Blair OM, Standeven JW, Kaiser GC: Coronary venous arterialization. Acute hemodynamic, metabolic and chronic anatomical observations. Ann Thorac Surg 23:449, 1977 28 Chiu CJ: Myocardial revascularization in diffuse coronary atherosclerosis. Recent experimental progress, Coronary Artery Medicine and Surgery: Concepts and Controversies. JC Norman, ed .. New York, 1975, AppletonCentury-Crofts 29 Dart CH Jr. Kato Y, Scott SM, Fish RG, Nelson WM, Takaro T: Internal thoracic (mammary) arteriography. A questionable index of myocardial revascularization. J THORAC CARDIOVASC SURG 59: I 17, 1970 30 Wanibuchi Y, Mundth ED, Wright JEC, Austen WG: The effect of indirect myocardial revascularization on left ventricular function. Ann Thorac Surg 12:93, 1972 31 Hammond GL, Austen WG: Drainage patterns of coronary arterial flow as determined from the isolated heart. Am J Physiol 212: 1435, 1967 32 Moll JW. Dziatkowiak AJ. Edelman M, Iljin W, Ratajczyk-Pakalska E, Stengert K: Arterialization of the coronary veins in diffuse coronary arteriosclerosis. J Cardiovasc Surg 16:520, 1975 33 Lawrie GM, Morris GC, Winters WL: Aortocoronary saphenous vein autograft accidentally attached to a coronary vein. Follow-up angiography and surgical correction of the resultant arteriovenous fistula. Ann Thorac Surg 22:87, 1976
Discussion DR. GEORGE A. LIEBLER Pittsburgh. Po.
We previously presented our initial experimental and clinical data to this organization in 1974. At this time. we have performed reverse myocardial perfusion in 10 patients. Eight patients have had internal mammary-to-anterior descending vein grafts and two have had saphenous vein grafts to the anterior descending vein. All patients were retested within 2 weeks and all grafts were patent. Six of the first seven patients
Arterialization of coronary venous system
I I
were retested I year postoperatively and the grafts were patent. The sixth patient was asymptomatic and refused to be retested. All of these patients had internal mammary grafts. The two saphenous vein grafts were retested and both were found to be occluded. Follow-up in the tenth patient is only 2 months. This slide shows the long-term follow-up of Ilh to 4'/2 years in the first nine patients having myocardial reverse perfusion. One patient died suddenly of unknown causes. An autopsy was not performed. Seven of the eight remaining patients are clinically doing well. One patient with an occluded saphenous vein graft is having angina on mild exertion. Both the experimental animals and patients have undergone myocardial scanning. We have noted up to a 20 percent shunt with no deleterious effects. We have sacrificed dogs up to 3 years postoperatively and have found no evidence of myocardial infarction or fibrosis. The coronary vein showed no evidence of fibrosis and the internal mammary artery grafts were patent except in one dog. 3 years postoperatively that had mild fibrosis of the anterior descending coronary veins although the graft was patent. Over 100 dogs have been subjected to internal mammary artery-coronary vein anastomoses, with Ameroid constrictors used to slowly occlude the native anterior descending coronary artery. Myocardial scanning did show perfusion of the anterior myocardium and catheterizations showed occlusion of the native anterior descending coronary artery and patency of the internal mammary-coronary vein graft. In conclusion, we feel that this procedure can be effective in improving myocardial perfusion in selected patients. In our hands, the saphenous vein grafts have not been successful, for they subsequently became occluded in both patients. There has been no operative death. DR. OTTO GAGO Ann Arbor. Mich.
We also agree that this is a good alternative procedure in patients with diffuse coronary artery disease in whom direct coronary artery bypass cannot be accomplished. We have done this operation on six different occasions since 1973. Our initial observations reveal that the bypasses are open but, unfortunately, restudy one year after the procedure has shown occlusion of the vein-to-coronary vein bypass despite initially large flows of 350 to 400 c.c. per minute. We were discouraged further by the presence of a large hemorrhagic infarction in one of our patients who had a bypass graft to the posterior descending coronary vein. We have limited the indication to smaller branches of the coronary venous system, like the obtuse marginal vein or the diagonal vein of the left anterior descending system. In our experimental data, the left anterior descending vein graft with an internal mammary artery graft in the dog did not stay open. I agree with the group presenting this paper that a larger flow is not needed initially to maintain the grafts open.
The Journal of
I 2 Hochberg et al.
DR. H 0 CH BERG (Closing) I am delighted to learn of the venous bypass experiences presented by Dr. Liebler and Dr. Gago. The paper given by Dr. Liebler and Dr. George J. Magovem at this meeting in 1974 stands as one of the very few clinical reports of coronary venous bypass. Dr. Magovem has been kind enough to share some of his recent data with me and they are described in the text of this paper. To answer Dr. Gage's questions: One of the key points is that graft flow in the coronary venous bypass is very critical. Contrary to what might be expected, a large graft flow is probably deleterious in the immediate postoperative stage and may well lead to venous engorgement and hemorrhagic infarction. That is why I was careful to point out that our flows
Thoracic and Cardiovascular Surgery
were approximately 50 ml. per minute in the bypasses both at the time of operation and at the time of sacrifice 3 to 5 months later. Four of our 18 dogs died, and all of them had flows greater than 100 ml. per minute. This might account for Dr. Gage's problems with bypass flows of 350 ml. per minute. The concept of selective venous retroperfusion is an attractive one for the selected patient because the venous system is never ravaged by atherosclerosis. which occurs only on the arterial side. It is my hope that animal and clinical investigation will continue to appear on the intriguing concept of coronary venous perfusion. I would like to thank the Association for the privilege of presenting this work and especially President Scannell who has been a valued teacher.
Information for authors Most of the provisions of the Copyright Act of 1976 became effective on January I, 1978. Therefore, all manuscripts must be accompanied by the following written statement, signed by one author: "The undersigned author transfers all copyright ownership of the manuscript (title of article) to The C. V. Mosby Company in the event the work is published. The undersigned author warrants that the article is original, is not under consideration by another journal, and has not been previously published. I sign for and accept responsibility for releasing this material on behalf of any and all co-authors." Authors will be consulted, when possible, regarding republication of their material.
January 1979
THORACIC AND CARDIOVASCULAR SURGERY The Journal of
Original Communications
Selective arterialization of the coronary venous system Encouraging long-term flow evaluation utilizing radioactive microspheres The long-term effectiveness of a retrograde coronary venous bypass graft (CVBG) to an ischemic left ventricle was evaluated in 18 dogs. A saphenous vein was interposed between the aorta and left anterior descending (LAD) vein. The LAD vein was ligated cephalad to the CVBG to prevent an arteriovenous fistula. The LAD artery was ligated at its origin to create anterior wall ischemia. Operative graft flow averaged 53 mi. per minute. The 14 surviving dogs were catheterized 3 to 5 months later. Ten of the 14 CVBC's were patent angiographically . The chests were opened and graft flow now averaged 50 mi. per minute. 141Ce microspheres were injected into the left atrium to measure myocardial flow to the anterior wall. In the 10 dogs with patent grafts, transmural flow was 39 ± I (5.E.M.) mi. per /00 Gm, of tissue per minute. The endocardiallepicardial flow ratio was 1.4/ I, indicating that retrograde venous perfusion effectively delivered blood to the subendocardium. After ligation of the CVBC, microsphere measured flow dropped to 15 ± I mi. per 100 Gm. per minute. In 15 control dogs, anterior wall flow was 100 ± 3 mi. per 100 Gm. per minute, decreasing to 13 ± 2 mi. 45 minutes after ligation of the LAD artery and vein. None of the eight control dogs with simple ligation of the LAD artery and vein survived more than 5 days. Histologic examination of the anterior wall of the left ventricle, the area served by the CVBC's for 3 to 5 months, disclosed no evidence of venous sclerosis or thrombosis and no evidence of interstitial edema or hemorrhage. Thus a CVBC permitted long-term survival in an otherwise nonviable anatomic preparation. Moreover, restoration offlow with a CVBG was effective because it perfused all layers of the myocardium, especially the subendocardium-the crucial layer of myocardial muscle.
Mark S. Hochberg, M.D. (by invitation), William C. Roberts, M.D. (by invitation), Andrew G. Morrow, M.D., and W. Gerald Austen, M.D., Bethesda, Md., and Boston, Mass.
From the Clinic of Surgery, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Md. 20014, and the Department of Surgery, Massachusetts General Hospital, Boston, Mass. 02114.
Read at the Fifty-eighth Annual Meeting of The American Association for Thoracic Surgery, New Orleans, La., May 8, 9, and 10, 1978. Address for reprints: Mark S. Hochberg, M.D., Department of Surgery, Massachusetts General Hospital, Boston, Mass. 02114.
The Journal 01
2 Hochberg et al .
Thoracic and Cardiovascular Surgery
Coronary artery bypass grafting (CABO) is now the procedure of choice for significant but discrete coronary arterial disease." However, of the 60,000 2 CABO patients operated upon yearly, a group of them have angiographically extensive atherosclerotic changes with poor runoff so that the CABO procedure is unlikely to succeed . This group is generally agreed to comprise 12 to 20 percent of revascularization candidates .":" In addition to these patients, many centers are beginning to see increasing numbers of patients with occluded bypass grafts and persistent angina. It is presumed that these individuals with failed CABO's have poor vascular runoff. Although we 7 have reported that coronary endarterectomy can be successful in certain circumstances, in general this procedure is not very useful for patients with (I) diffuse arterial disease (especially of the left system) or (2) previously failed CABO procedures.
CVBG (
Recently , weB have reported anterior left ventricular myocardial flow characteristics via selective retrograde coronary venous perfusion in a set of short-term experiments . In 15 animal experiments, the effectiveness of venous perfusion to the subendocardium, myocardium, and subepicardium was established with radioactive microspheres used to measure the flow. The present study is aimed at evaluating the longterm effectiveness of coronary venous bypass grafts (CVBO) functioning for 3 to 5 months. Methods Eighteen American foxhound dogs weighing 23 to 28 kilograms had saphenous vein-coronary venous bypass grafts interposed between the aorta and the left anterior descending (LAD) vein. The animals were anesthetized with thiamylal sodium and succinylcholine, intubated , and ventilated
! )
CORONARY VENOUS) . . . . BYPASS GRAFT
RCA \
\
<,
L. CIRCUMFLEX A.
LI GATION OF GREAT CARDIAC V.
LAD VEINS '-/
Fig. 1. Diagram of a completed coronary venous bypass graft (CVBG) . Note proximal ligation of LAD artery to render the anterior wall ischemic . The LAD vein was also ligated cephalad to the CVBG to prevent an arteriovenous fistula . RCA, Right coronary artery . LAD, Left anterior descend ing.
Fig. 2A. The photograph on the left shows normal coronary anatomy with the arteries red and the veins blue. The photograph on the right illustrates that during retrograde coronary venous perfusion , the arteries contain blue (desaturated) blood and the veins carry red (oxygenated) blood. (These photographs are from previously cited short-term experiments" with inflow into the LAD coronary veins accomplished via a shunt from the carotid artery) . Fig. 2B. Completed coronary venous bypass graft ju st prior to chest closure. Note that the diagonal coronary artery is now blue with desaturated blood , and its accompanying veins are now red with oxygenated blood (arrow).
RETROGRADE CORONARY VENOUS PERFUSION
NORMAL HEMODYNAMIC ANATOMY
/
I
I
I
/
,, ,, L.ANT. DESCENDING ARTERY
L.ANT. DESCENDING VEIN
/
/
I
L. ANT. DESCENDING ARTERY
L.ANT. DESCENDING VEIN
For legend see facing page.
Volume 77 Number 1
Arterialization of coronary venous system
5
January. 1979
Fig. 3. Dye injected into the aortic root of a dog revealing a patent coronary venous bypassgraft 4 months after operation. with a Harvard respirator. Halothane anesthesia was continued throughout each procedure. Measurements of arterial, left atrial, and central venous pressure were made with Statham P23Db transducers and continuously recorded on a Sanborn 350 recorder. The electrocardiogram was also monitored. The saphenous vein was removed from the lower part of the left hind leg. The chest was opened through the fourth left intercostal space. When the dissection was completed, the animal was heparinized and total cardiopulmonary bypass was instituted via the right atrium and right femoral artery . A left ventricular vent was placed through the left atrial appendage. The animal was cooled to 27° C. and the heart was electrically fibrillated. The aorta was cross-clamped and the vein graft was anastomosed with 6-0 Prolene to the great cardiac vein at the junction of the two LAD veins. Fig. I illustrates the exact anatomic location. The great cardiac vein was ligated cephalad to the anastomosis to prevent an arteriovenous fistula . The LAD artery was ligated at its origin (proximal to the first septal perforator) to precipitate acute ischemia to the anterior wall of the left ventricle (Fig. I). The aortic crossclamp was then removed and replaced with a partially occluding Satinsky clamp. The proximal CVBG anastomosis was performed during the warming phase . When the proximal suture line was secured, the heart was defibrillated. The LAD veins supplied by the CVBG turned red. As previously noted ," the LAD artery was now blue, because it contained desaturated blood . Fig . 2A illustrates this new physiological event. Cardiopulmonary bypass was discontinued and flow was measured through the newly constructed CVBG with Statham
Fig. 4, Selective injectionof a coronary venous bypass graft in an animal 5 months after operation. Dye illustrates perfusion of the anterior wall via the CVSG .
electromagnetic flow probes. Fig . 2B is an operative photograph of a CVBG just prior to the chest closure. Three to 5 months following operation, the animals were electively brought to the angiography suite and catheterized. Injections of dye into the aortic root and selective vein grafts were recorded on film. The dogs were then transported to an operating room and a left thoracotomy was performed. Graft flows were again measured via flow probes. Next, 900,000 141Ce microspheres of the 15 JL size were injected into the left atrium . Reference flow was collected for precisely I minute from a catheter in the left femoral artery. The 15 JL microsphere size was selected for its proved reliability. 9 . 10* After the first injection, the CVBG was ligated, and a second injection with 900,000 80Sr microspheres was made. The animals were put to death and the hearts were prepared for scintillation counting. A full-thickness piece of anterior left ventricular myocardium, approximately 4 by 2 cm ., lying directly in the area served by the LAD veins and artery, was excised . Epicardial fat and large vessels were stripped away . The fullthickness piece was dissected with a razor blade into its three anatomic components: the subendocardium, myocardium, and subepicardium. Each layer was placed in a separate counting vial and weighed. As an internal control on the method, a similar piece of posterior left ventricular wall was excised and dissected into the three layers. Because the posterior wall of the heart is served by the left circumflex coronary artery (which is dominant in the dog), there should have been no change in the flow it received during the *Utley JR: Personal communication .
The Journal of Thoracic and Cardiovascular Surgery
6 Hochberg et al.
Table I. Flow evaluation in 10 patent coronary venous bypass grafts (mI. 1100 Gm.lmin.) Exp. No. I
2 3 4 5 6 7 8 9 10
Subendocardial flow
Myocardial flow
Subepicardial flow
Endo/epi
Avg. transmural flow
36.8 59.3 23.5 55.5 44.8 37.8 34.6 40.5 27.6 35.0
34.5 36.6 47.8 46.9 50.6 37.5 39.4 43.3 36.1 30.4
41.4 35.2 41.3 20.1 13.1 51.6 25.5 39.0 60.1 52.5
0.89 1.69 0.57 2.76 3.42 0.73 1.36 1.04 0.46 0.67
36.8 41.5 39.2 42.8 40.2 41.0 33.1 41.3 40.5 37.9
Average
1.36
Table II. Summary of transmural myocardial flow in the ten patent and four occluded CVBG's (ml. 1100 Gm.Imin.) Ten patent
CVBG open
CVBG ligated
LV anterior wall
39.4 ::!: 0.9
15.2::!: 1.3
LV posterior wall (control)
88.5::!: 2.1
91.2 ::!: 1.9
CVBG's
CVBG occluded
Four occluded
LV anterior wall
14.3 ::!: 1.7*
LV posterior wall (control)
86.4 ::!: 3.9*
CVBG's
Legend: CVBG, Coronary venous bypass graft. LV, Left ventricular. -Average value of two microsphere injections.
two microsphere injections. If mean aortic pressure had differed during any two experimental injections, then the anterior wall flows would have been measured falsely. The knowledge that control posterior wall flows did not differ significantly between injections can permit this study to report that the measured anterior wall flows accurately portrayed the true hemodynamic picture. Histologic examination of the left ventricular anterior wall and saphenous vein bypass was also performed. Fifteen additional foxhounds were studied as a control of the CVBG experimental animals just described. The normal transmural flow to the nonoperated anterior wall was measured. The LAD artery and vein were then ligated and flow again was determined with microspheres at the time of sacrifice 45 minutes later. An additional series of eight dogs underwent ligation of the LAD artery and vein without the construction of a CVBG. It was planned that these dogs would serve as long-term controls for the 18 operated CVBG animals.
39.4::!: 0.9 (S.E.M.)
The tissue samples and the reference blood were counted for their gamma emission in a Beckman multichannel scintillation counter. * A computer program was employed to establish an equation matrix based upon the reference blood radioactivity and weight for each injection. Once the matrix was determined, the actual tissue counts per minute could be entered. The resultant flow in milliliters per 100 Gm. of tissue per minute was computed. This myocardial blood flow technique utilizing microspheres was developed by Rudolph and Heymann. II Buckberg 12 has recently reviewed the mathematical model and commented on the technical aspects of this microsphere method. Thus the hemodynamic effectiveness of long-term retrograde CVBG was evaluated by angiography, operative and postoperative graft flow measurements, histology, and the actual flow to the subendocardium, myocardium, and subepicardium with the use of radioactive tracer microspheres.
Results Fourteen of the 18 foxhounds that underwent CVBG survived the postoperative period to the time of sacrifice 3 to 5 months after operation, Angiography. Fig. 3 illustrates an aortic root injection of a CVBG which was patent 4 months after operation. Of the 14 surviving animals, 10 had patent grafts at the time of angiography and sacrifice. In most of these animals, the catheter could then be advanced into the graft itself so that the anatomy could be better delineated, The selective graft injection shown in Fig. 4 clearly demonstrates perfusion of the anterior wall of the left ventricle. This angiogram was performed in an *We are grateful to Mr. Richard Kagan for his advice and the use of his equiprnent.
Volume 77 Number 1 January, 1979
Arterialization of coronary venous system
7
.s: """ ~ ~ ~ .... ~
~
~~ ~ .....
~
~
~ ~ t-.:
100 90 80 70 60 50 40 30 20
to 0
LONGTERM NORMAL ANTERIOR WALL FLOW
PATENT CVBGs
LIGATION OF PATENT CVBGs
LIGATION LAD ARTERY and VEIN
Fig. 5. Bar graphs depicting the relative transmural anterior wall flows of all experiments in this report. Normal anterior wall flow is on the left. Flows were measured via radioactive microspheres.
animal 5 months after the venous bypass procedure. As reported above, four of the 18 animals in which the CVBO was performed did not survive the postoperative period. Necropsy was performed in these animals, and three of them had patent grafts. Because there was no evidence of congestive heart failure, it was thought that death was due to an arrhythmia or an acute myocardial infarction. Graft flow measurement. At the time of operation, operative graft flow was measured at 53.0 ± 3.1 * ml. per minute. At the time of sacrifice 3 to 5 months later, graft flow in the 10 patent grafts averaged 50.7 ± 4.9 ml. per minute. The operative CVBO flow in the four dogs found subsequently to have CVBO occlusion was 32.4 ± 11.4 ml. per minute. Myocardial flow measurements. Flow to fullthickness segments of the left ventricular anterior wall in the 10 animals with patent grafts averaged 39.4 ± 0.9 rnl. per 100 Om. per minute. Of importance, the flow to the subendocardium alone averaged 39.5 ± 3.5 rnl. per 100 Om. per minute with an endocardial/epicardial average flow ratio of 1.36 (Table I). Myocardial flow measurements were also performed in the four animals with occluded grafts. Table II summarizes the transmural flows in the 10 animals with patent CVBO's and the four animals with occluded CVBO's. This table also reports the control flows measured to the posterior left ventricular wall of all 14 dogs. This was done as an internal control on the microsphere method as noted previously. There was no significant difference (p = 0.05) be-
tween the average transmural flow for the four occluded grafts (14.3 ± 1.7 ml. per 100 Om. per minute) and the 10 acutely ligated grafts (15.2 ± 1.3 ml. per 100 Om. per minute).* However, there was a highly significant difference (p ::s: 0.00 I) when the transmural flow in the 10 patent grafts (39.4 ± 0.9 ml. per 100 Om. per minute) was compared to the flows of either the ligated CVBO's (10 animals) or the occluded CVBO's (four animals). These comparisons are graphically depicted in Fig. 5. Control series. The normal transmural flow to a nonmanipulated left ventricular anterior wall was 100 ± 2.5 ml. per 100 Om. per minute in the 15 control foxhounds. This flow fell to 13.5 ± 1.5 ml. per 100 Om. per minute 45 minutes after ligation of the LAD artery and vein (Fig. 5). Eight additional animals had control ligation of the LAD artery and vein. Their chests were closed in the hope of determining left ventricular transmural flow at the time of sacrifice of the experimental CVBO dogs (3 to 5 months after operation). Of great importance. none of these eight control dogs could withstand the acute effects of LAD artery ligation (without CVBO protection) and all died within 5 days of operation. Fig. 5 illustrates the normal anterior wall transmural flow compared to the perfusion provided by all the CVBO experiments and controls that have been described. Pathology. Histologic examination of typical areas perfused by the CVBO for the experimental period is shown in Fig. 6. There is no evidence of venous sclerosis or thrombosis and no evidence of interstitial
*All numbers following a ± sign indicate I standard error of the mean.
*As noted above. in all the experimental animals the LAD artery was ligated at operation to render the left ventricle ischemic.
8 Hochberg et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 6.A. This histologic section shows an occluded arteriole on the right with a patent adjacent venule on the left. (The venule contains some coagulated fibrin which occurred postmortem.) B. Two patent venules with an adjacent artery. There is no evidence of venous thrombosis or sclerosis. The myocardium shows no evidence of hemorrhagic infarction. C. A histologic section of a widely patent coronary venous bypass graft in an animal put to death 4 months after operation.
The experimental findings suggest that retrograde CVBG's can remain patent over a long period of time. This confirms the work of Magovem* and Chiu and Mulder. 13 The important finding of this experimental report is that, for the first time, the precise effectiveness of long-term selective venous perfusion to the three layers of the myocardium was determined with radioactive microspheres. Previous reports of successful
CVBG3-6. 13* have been based largely on coronary sinus oxygen data, electrocardiographic evidence of ischemic changes , microscopic evidence of infarction. radioisotope scanning, and animal death . These measures are at best indirect methods of evaluating retrograde coronary venous perfusion to the heart . The experimental findings of this report illustrate that a CVBG can provide almost three times more perfusion than the base-line flow to a left ventricular wall rendered acutely ischemic by simple LAD artery and vein ligation (or by ligation of a patent CVBG). More simply , a CVBG can sustain life, whereas the control setting resulted in the death of all eight dogs within 5
*Magovem GJ: Personal communication .
*Magovem GJ : Personal communication.
edema or hemorrhage (venous infarction) . A cross section of a CVBG shows a wide-open channel (Fig. 6, C) .
Discussion
Volume 77 Number 1 January, 1979
days of LAD artery and vein ligation. These latter dogs, of course, did not have the protection of venous retroperfusion. Beck ,':' Zajtchuk," and their colleagues found that ligation of the dog's LAD artery was fatal in 70 percent and 77 percent of animals, respectively, observations which confirm our control series findngs. * Not only can the CVBG's sustain life in an otherwise nonviable control setting, but the perfusion is also effective, because the microsphere technique demonstrated flow to all layers of the myocardium. Initial fears that coronary venous perfusion would be ineffective because of venovenous epicardial anastomoses are dispelled by the demonstration that the subendocardium receives, via a CVBG, a level of flow not dissimilar to that of the myocardium and subepicardium. This is proved by an average endocardial/epicardial ratio of 1.4/1. Furthermore, successful perfusion of the subendocardium cannot be explained by the development of arterial collateral channels. The patent CVBG's were occluded just prior to sacrifice, and flow dropped from 39.4 ml. per 100 Gm. per minute to 15.2 (Fig. 5), suggesting true perfusion via the CVBG's. Although the data showed significant reversal of ischemia with retrograde coronary venous perfusion, this study can make no statement as to whether this level of perfusion can prevent an infarct or even relieve angina. Indeed, we expected some measure of infarction, since LAD artery ligation was chosen as the experimental model of ischemia for evaluation of selective retrograde coronary venous perfusion. Since Beck.tv 16 Bakst;" and their associates first attempted to globally retroperfuse the heart by arterializing the coronary sinus, the effectiveness of reversing flow in the coronary venous system has been debated. More recently, many authors":" 8. 13. 15. 18- 2 7 t have studied the possibility of venous retroperfusion to a selective area of the heart. The work has been well summarized by Bates" and Chiu." Most of these models used the left internal mammary artery (LIMA) as the inflow to the coronary venous system. Two important technical points emerge from the literature. The first finding is that inflow must be at least 50 mi. *When ligating the LAD artery. it is important to ensure that occlusion occurs proximal to the first septal perforator which arises very close to the bifurcation of the left main coronary artery. This may account for the varying mortality rates following LAD arterial ligation reported in the literature. t Magovern Gl: Personal communication.
Arterialization of coronary venous system
9
per minute through the graft. Three reports 20. 29. 30 suggest that an LIMA-LAD vein anastomosis usually cannot supply this necessary level of flow. Consequently, in this report a reversed saphenous vein was utilized to establish inflow into the coronary venous system. Second, Hammond and Austen" nicely demonstrated in 1967 that greater than 50 percent of LAD artery blood flow was returned to the heart by routes other than the coronary sinus (anterior cardiac veins as well as direct communications to the heart chambersthe thebesian and arteriosinusoidal channels). Consequently, reports of retrograde perfusion based on coronary sinus sampling must be carefully evaluated. Selective retrograde coronary venous perfusion has stirred a significant amount of controversy in the literature. Encouraging CVBG experiments with various means of evaluation have been cited earlier. 3-6. 8. 13. 24* However Zajtchuk ," Razzuk," Marco;" and their co-workers, using seemingly similar methods of experimental evaluation, found venous perfusion to be of little benefit. The explanation for these less sanguine reports is not immediately obvious. However, it can be speculated that their experiments may have failed for four reasons. First, LIMA inflow may have been inadequate as discussed": 29. 30 None of these papers reported CVBG flow at the time of operation and sacrifice. The second potential technical problem may have been the point of LAD artery ligation. As noted, fatal arterial infarction can be achieved only if the artery is occluded proximal to the first septal perforator. If ligation is performed distal to this branch, "competitive" arterial inflow via this perforator and collaterals may render the LIMA inflow far less significant at the capillary level. This conceivably could explain the finding of LIMA patency with obstruction at the LIMALAD vein anastomosis in one of these reports. 27 Conversely, as illustrated in Fig. 6, we found no histologic evidence of intimal hyperplasia of the coronary venous system with graft flow averaging greater than 50 ml. per minute. The third problem may have been metabolic data sampling from the coronary sinus. As suggested earlier, the greatest portion of blood draining the left ventricular anterior wall returns by noncoronary sinus routes. The fourth question raised by these three pessimistic reports is their operative approach. Two of these three papers reported that the anastomosis was achieved without the aid of cardiopulmonary bypass (the third report does not state the operative technique *Magovern Gl: Personal communication.
I0
Hochberg et
at.
employed). We found a more precise anastomosis could be achieved with a quiet heart. Precise quantification of the hemodynamic effects of retrograde CVBG's to the subendocardium, myocardium, and subepicardium has now been accomplished with the use of radioactive microspheres. In both short-term" and the present long-term experiments, effective tissue perfusion has been achieved. There have been scattered reports of the clinical use of CVBG's with some success.t?: 24. 32. 33 Because the numbers are small, no significant conclusions can yet be drawn. On the basis of the encouraging data presented in this report, it may now be reasonable to offer a saphenous vein-CVBG to patients with (I) diffuse atherosclerosis (not amenable to endarterectomy) or to patients (2) who have failed to benefit from the conventional coronary artery bypass procedure.
Conclusion I. In a series of long-term animal experiments, a saphenous vein-coronary venous bypass graft (CVBG) permitted survival in the presence of LAD artery ligation. By contrast, no animals survived LAD artery ligation without CVBG protection. 2. Flow via a CVBG is effective because it perfuses all layers of the myocardium. 3. The encouraging long-term patency data in animals suggest that judicious use of a venous bypass may be of value in selected patients with diffuse atherosclerosis or a previously failed coronary artery bypass. We wish to acknowledge the excellent technical assistance of Miss Connie Martino and Mrs. Joan B. Fuller.
2
3
4
5
6
REFERENCES Mundth ED. Austen WG: Surgical measures for coronary heart disease (three parts). N Engl J Med 293: 13. 75. 124. 1975 Miller DW. Hessel EA II. Winterscheid LC, Merendino KA. Dillard DH: Current practice of coronary artery bypass surgery. Results of a national survey. J THORAC CARDIOVASC SURG 73:75. 1977 Bhayana IN. Olsen DB. Byrne JP. KolffWJ: Reversal of myocardial ischemia by arterialization of the coronary vein. J THORAC CARDIOVASC SURG 67:125. 1974 Gardner RS. Magovern GJ. Park SB. Dixon CM: Arterialization of coronary veins in the treatment of myocardial ischemia. J THORAC CARDIOVASC SURG 68:273, 1974 Moll JW. Dziathowiak A. Rybinski K. Edelman M, Ratajczak-Pakalska E: Arterialisierung des Sinus Coronarius. lndikationen , Technik , Ergebnisse. Thoraxchirurgie 21 :295. 1973 Arealis, EG. VoIder JG. Kolff WJ: Arterialization of the
The Journal of Thoracic and Cardiovascular Surgery
coronary vein coming from an ischemic area. Chest 63:462. 1973 7 Hochberg MS. Merrill WHo Michaelis LL. Mcintosh CL: Results of combined coronary endarterectomy and coronary bypass for diffuse coronary artery disease. J THORAC CARDIOVASC SURG 75:38. 1978 8 Hochberg MS: Hemodynamic evaluation of selective arterialization of the coronary venous system: An experimental study of myocardial perfusion using radioactive microspheres. J THORAC CARDIOVASC SURG 74:774. 1977 9 Marshall WG, Boatman GB. Dickerson G. Perlin A. Todd EP. Utley JR: Shunting. release. and distribution of nine and fifteen micron spheres in myocardium. Surgery 79:631. 1976 10 Utley JR. Carlson EL. Hoffman HE. Martinez HM. Buckberg GD: Total and regional myocardial blood flow measurements with 25 JL. 15 JL. 9 JL. and filtered I to 10 JL diameter microspheres and antipyrine in dogs and sheep. Circ Res 34:391. 1974 II Rudolph AM. Heymann MA: The circulation of the fetus in utero. Methods for studying distribution of blood flow. cardiac output and organ blood flow. Circ Res 21: 163. 1967 12 Buckberg GD: Studies of regional coronary flow using radioactive microspheres. Ann Thorac Surg 20:46. 1975 13 Chiu CJ. Mulder DS: Selective arterialization of coronary veins for diffuse coronary occlusion. An experimental evaluation. J THORAC CARDIOVASC SURG 70:177. 1975 14 Beck CS. Stanton E. Batiuchok W. Leiter E: Revascularization of the heart by a graft of systemic artery into the coronary sinus. JAMA 137:436. 1948 15 Zajtchuk R. Heydorn WHo Miller JG. Strevey TE. Treasure RL: Revascularization of the heart through the coronary veins. Ann Thorac Surg 21:318. 1976 16 Beck CS. Leighninger DS: Scientific basis for the surgical treatment of coronary artery disease. JAMA 159: 1264. 1955 17 Bakst AA. Costas-Durieux 1. Goldberg H. Bailey CP: Protection of the heart by arterialization of the coronary sinus. II. Coronary flow in dogs with aorticocoronary sinus anastomosis. J THoRAc SURG 27:442. 1954 18 Andreadis P. Natsikas N. Arealis E. Lazarides DP: The aortocoronary venous anastomosis in experimental acute myocardial ischemia. Vase Surg 8:45. 1974 19 Kay EB. Suzuki A: Coronary venous retroperfusion for myocardial revascularization. Ann Thorac Surg 19:327. 1975 20 Benedict JS. Buhl TG. Henney RP: Cardiac vein myocardial revascularization. An experimental study and report of three clinical cases. Ann Thorac Surg 20:550. 1975 21 Williams GD. Burnett HF, Derrick BL. Miller CH: Retrograde venous cardiac perfusion for myocardial revascularization. An experimental evaluation. Ann Thorac Surg 22:322. 1976
Volume 77 Number 1 January, 1979
22 Meerbaum S, Lang TW, Osher JV, Hashimoto K, Lewis GW, Feldstein C, Corday E: Diastolic retroperfusion of acutely ischemic myocardium. Am J Cardiol 37:588, 1976 23 Demos S, Brooks H, Holland R, Balderman S, Anagnostopoulos C: Retrograde coronary venous perfusion to reverse and prevent acute myocardial ischemia. Circulation 49,50:Suppl 3:168, 1974 24 Park SB, Magovern GJ, Liebler GA, Dixon CM, Begg FR, Fischer DL, Dosios TJ, Gardner RS: Direct selective myocardial revascularization by internal mammary artery-coronary vein anastomosis. J THORAC CARDIOVASC SURG 69:63, 1975 25 Bates RJ. Toscano M, Balderman Sc, Anagnostopoulos CE: The cardiac veins and retrograde coronary venous perfusion. Ann Thorac Surg 23:83, 1977 26 Razzuk MA: Discussion of Benedict et al'" 27 Marco JD, Hahn JW, Barner HB, Jellinek M, Blair OM, Standeven JW, Kaiser GC: Coronary venous arterialization. Acute hemodynamic, metabolic and chronic anatomical observations. Ann Thorac Surg 23:449, 1977 28 Chiu CJ: Myocardial revascularization in diffuse coronary atherosclerosis. Recent experimental progress, Coronary Artery Medicine and Surgery: Concepts and Controversies. JC Norman, ed .. New York, 1975, AppletonCentury-Crofts 29 Dart CH Jr. Kato Y, Scott SM, Fish RG, Nelson WM, Takaro T: Internal thoracic (mammary) arteriography. A questionable index of myocardial revascularization. J THORAC CARDIOVASC SURG 59: I 17, 1970 30 Wanibuchi Y, Mundth ED, Wright JEC, Austen WG: The effect of indirect myocardial revascularization on left ventricular function. Ann Thorac Surg 12:93, 1972 31 Hammond GL, Austen WG: Drainage patterns of coronary arterial flow as determined from the isolated heart. Am J Physiol 212: 1435, 1967 32 Moll JW. Dziatkowiak AJ. Edelman M, Iljin W, Ratajczyk-Pakalska E, Stengert K: Arterialization of the coronary veins in diffuse coronary arteriosclerosis. J Cardiovasc Surg 16:520, 1975 33 Lawrie GM, Morris GC, Winters WL: Aortocoronary saphenous vein autograft accidentally attached to a coronary vein. Follow-up angiography and surgical correction of the resultant arteriovenous fistula. Ann Thorac Surg 22:87, 1976
Discussion DR. GEORGE A. LIEBLER Pittsburgh. Po.
We previously presented our initial experimental and clinical data to this organization in 1974. At this time. we have performed reverse myocardial perfusion in 10 patients. Eight patients have had internal mammary-to-anterior descending vein grafts and two have had saphenous vein grafts to the anterior descending vein. All patients were retested within 2 weeks and all grafts were patent. Six of the first seven patients
Arterialization of coronary venous system
I I
were retested I year postoperatively and the grafts were patent. The sixth patient was asymptomatic and refused to be retested. All of these patients had internal mammary grafts. The two saphenous vein grafts were retested and both were found to be occluded. Follow-up in the tenth patient is only 2 months. This slide shows the long-term follow-up of Ilh to 4'/2 years in the first nine patients having myocardial reverse perfusion. One patient died suddenly of unknown causes. An autopsy was not performed. Seven of the eight remaining patients are clinically doing well. One patient with an occluded saphenous vein graft is having angina on mild exertion. Both the experimental animals and patients have undergone myocardial scanning. We have noted up to a 20 percent shunt with no deleterious effects. We have sacrificed dogs up to 3 years postoperatively and have found no evidence of myocardial infarction or fibrosis. The coronary vein showed no evidence of fibrosis and the internal mammary artery grafts were patent except in one dog. 3 years postoperatively that had mild fibrosis of the anterior descending coronary veins although the graft was patent. Over 100 dogs have been subjected to internal mammary artery-coronary vein anastomoses, with Ameroid constrictors used to slowly occlude the native anterior descending coronary artery. Myocardial scanning did show perfusion of the anterior myocardium and catheterizations showed occlusion of the native anterior descending coronary artery and patency of the internal mammary-coronary vein graft. In conclusion, we feel that this procedure can be effective in improving myocardial perfusion in selected patients. In our hands, the saphenous vein grafts have not been successful, for they subsequently became occluded in both patients. There has been no operative death. DR. OTTO GAGO Ann Arbor. Mich.
We also agree that this is a good alternative procedure in patients with diffuse coronary artery disease in whom direct coronary artery bypass cannot be accomplished. We have done this operation on six different occasions since 1973. Our initial observations reveal that the bypasses are open but, unfortunately, restudy one year after the procedure has shown occlusion of the vein-to-coronary vein bypass despite initially large flows of 350 to 400 c.c. per minute. We were discouraged further by the presence of a large hemorrhagic infarction in one of our patients who had a bypass graft to the posterior descending coronary vein. We have limited the indication to smaller branches of the coronary venous system, like the obtuse marginal vein or the diagonal vein of the left anterior descending system. In our experimental data, the left anterior descending vein graft with an internal mammary artery graft in the dog did not stay open. I agree with the group presenting this paper that a larger flow is not needed initially to maintain the grafts open.
The Journal of
I 2 Hochberg et al.
DR. H 0 CH BERG (Closing) I am delighted to learn of the venous bypass experiences presented by Dr. Liebler and Dr. Gago. The paper given by Dr. Liebler and Dr. George J. Magovem at this meeting in 1974 stands as one of the very few clinical reports of coronary venous bypass. Dr. Magovem has been kind enough to share some of his recent data with me and they are described in the text of this paper. To answer Dr. Gage's questions: One of the key points is that graft flow in the coronary venous bypass is very critical. Contrary to what might be expected, a large graft flow is probably deleterious in the immediate postoperative stage and may well lead to venous engorgement and hemorrhagic infarction. That is why I was careful to point out that our flows
Thoracic and Cardiovascular Surgery
were approximately 50 ml. per minute in the bypasses both at the time of operation and at the time of sacrifice 3 to 5 months later. Four of our 18 dogs died, and all of them had flows greater than 100 ml. per minute. This might account for Dr. Gage's problems with bypass flows of 350 ml. per minute. The concept of selective venous retroperfusion is an attractive one for the selected patient because the venous system is never ravaged by atherosclerosis. which occurs only on the arterial side. It is my hope that animal and clinical investigation will continue to appear on the intriguing concept of coronary venous perfusion. I would like to thank the Association for the privilege of presenting this work and especially President Scannell who has been a valued teacher.
Information for authors Most of the provisions of the Copyright Act of 1976 became effective on January I, 1978. Therefore, all manuscripts must be accompanied by the following written statement, signed by one author: "The undersigned author transfers all copyright ownership of the manuscript (title of article) to The C. V. Mosby Company in the event the work is published. The undersigned author warrants that the article is original, is not under consideration by another journal, and has not been previously published. I sign for and accept responsibility for releasing this material on behalf of any and all co-authors." Authors will be consulted, when possible, regarding republication of their material.