- Email: [email protected]
The internal mammary bypass graft: A superior second coronary artery
The internal mammary bypass graft: A superior second coronary artery In a study of the initial 1,004 consecutive patients who had direct coronary artery bypass at Ochsner Medical Institutions, computer methods of data processing were used to compare the clinical results between patients who had saphenous vein (SV) grafts and those who had internal mammary artery (IMA) grafts. The factors compared were the long-term mortality rates, nonfatal myocardial infarction rates, relief or persistence of angina, and the percentage of patients who acquired congestive heart failure. A simple comparison showed the patients with IMA grafts did better in all four categories; however, in a subsequent analysis in which maldistributed factors were removed, the rates of anginal relief and congestive heart failure were not significantly improved. The major benefit appears to be an increase in longevity among patients who had IMA bypasses.
James W. Jones, M.D.,* John L. Ochsner, M.D., Noel L. Mills, M.D., and Larry Hughes, Ph.D., New Orleans, La.
T
he prevailing attitude that the internal mammary artery (IMA) is a better conduit than the saphenous vein (5 V) for coronary bypass grafts is based on evidence of superior patency rates.!"" A less popular concept is that the IMA graft is inferior, except in certain circumstances, because of lessened flow rates.P- 6 Clinical results, and not angiographic patency or physiological measurements, remain the best means for determining superiority of an operative technique; therefore, the following study was done to provide a comparison of the long-term outcome between patients who had the IMA and those who had the 5 V used as coronary bypass grafts.
Patients and methods This analysis is based on computerized data obtained from the initial 1,004 consecutive patients who had direct coronary artery bypass at the Ochsner Medical Institutions. A previous report? dealt with other findings from this same study. The operative proceFrom the Department of Surgery, Ochsner Medical Institutions, New Orleans, La. Received for publication Oct. 4, 1977. Accepted for publication Nov. 7, 1977. Address for reprints: Dr. John L. Ochsner, Ochsner Clinic, 1514 Jefferson Highway, New Orleans, La. 70121. 'Present address: Department of Surgery, Tulane Medical Center, New Orleans, La. 70112.
0022-5223/78/0475-0625$00.70/0 © 1978 The C. V. Mosby Co.
dures were done or supervised by either of two surgeons (J. L. O. or N. L. M.) between 1968 and 1975. The follow-up averaged 48.6 months and was complete to the present in 98 percent of cases. Operative success was judged by relief of angina, over-all survival, prevention of congestive heart failure, and nonfatal myocardial infarction rates. Patients were counted as being improved unless they reported that the frequency and intensity of angina was unchanged from or worse than before operation. All deaths are included in the analysis regardless of cardiac relationship. Those patients requiring treatment for chronic heart failure are included as such. Patients classified as having myocardial infarction had to have definite electrocardiographic evidence supporting the diagnosis. Clinical success was evaluated by comparing all the patients who had the two types of grafts as to the presence of any variables indicating clinical result. A cross tabulation was made of the two types of grafts among 195 independent clinical variables; a partial listing of these variables is shown in Table I. Uneven distribution was found to exist in a number of independent variables, and these maldistributed variables were then cross-tabulated with increased mortality. The following variables were found to be associated with increased mortality: (1) female sex, (2) concomitant ventricular aneurysmectomy, (3) preoperative congestive heart 625
The Journal of
626
Jones et al.
Thoracic and Cardiovascular Surgery
Table I. Cross tabulation between IMA and SV grafts and variables* in total patient population Variable
No. with variable (%)
No. without variable (%)
Age
IMA SV
Below 50 yr. 141 (33%) 129 (33%)
50-59 yr. 196 (46%) 172 (44%)
Sex
IMA SV
Male 372 (88%) 310 (80%)
Female 51 (12%) 79 (20%)
Valve repair
IMA SV
7 (2%) 9 (2%)
Ventricular aneurysm
IMA SV
3 (7%) 9 (2%)
Aneurysm of ventriclet
IMA SV
State of ventriclet
p Value 60-69 yr. 80 (19%) 77 (20%)
70-79 yr. 6 (2%) 11 (3%)
5.379
0.25058
10.261
0.00136
416 (98%) 380 (98%)
0.455
0.50018
420 (99.3%) 380 (98%)
3.582
0.05840
9 (2%) 17 (4%)
414 (98%) 372 (96%)
3.287
0.06982
IMA SV
Normal 254 (60%) 202 (52%)
Seg. dysfunction III (26%) 114 (27%)
6.274
0.04341
Elevated triclycerides
IMA SV
183 (43%) 178 (46%)
240 (57%) 211 (54%)
0.510
0.4750
Diabetes
IMA SV
64(15%) 47 (12%)
359 (85%) 342 (87%)
1.594
0.2066
Hypertension
IMA SV
158 (37%) 153 (39%)
265 (73%) 236 (61%)
0.335
0.56255
Obesity
IMA SV
154 (36%) 154 (40%)
269 (64%) 235 (60%)
0.871
0.35063
Smoking
IMA SV
317(75%) 291 (75%)
106 (25%) 98 (25%)
0.002
0.96861
CHF
IMA SV
31 (7%) 57 (15%)
392 (93%) 332 (85%)
11.249
0.00080
Gen. dysfunction 58 (14%) 73 (21%)
Legend: IMA, Internal mammary artery. SV, Saphenous vein. CHF, Congestive heart failure.
'Partiallisting of 195 clinical variables statistically analyzed. t Aneurysm and state of ventricle as determined by preoperative ventriculogram.
Table II. Cross tabulation between IMA and SV grafts and variables* in unbiased patient population Variable
No. with variable (%)
No. without variable (%)
Below 50 yr. 117 (35%) 89 (35%)
50-59 yr. 159 (47%) 114 (45%)
60-69 yr. 57 (17%) 44 (18%)
70-79 yr. 2 (1%) 5 (2%)
Chi square
p Value
5.128
0.27443
Age
IMA SV
Valve repair
IMA SV
3 (1%) 2 (1%)
332 (99%) 250 (99%)
0.017
0.8951
Elevated cholesterol
IMA SV
30 (9%) 16(6%)
305 (91%) 236 (94%)
1.352
0.2444
Elevated triglycerides
IMA SV
151 (45%) 119 (47%)
184 (55%) 133 (53%)
0.266
0.6057
Hypertension
IMA SV
124 (37%) 82 (33%)
211 (63%) 170 (67%)
1.264
0.26089
Diabetes
IMA SV
47 (14%) 27 (11%)
288 (86%) 225 (89%)
1.435
0.2309
'Partial listing of variables analyzed.
Volume75 Number 4 April,1978
IMA bypass graft
627
Table III. Severity of coronary artery disease cross tabulated with IMA or SV grafts >70% constricted
Coronary artery
<70% constricted
Chi square
p Value
Main left
IMA
3(1%) 6(3%)
39(11%) 25 (10%)
16 (5%) 8 (3%)
277 (83%) 213 (84%)
3.420
0.33123
LAD
IMA
63 (19%) 49(19%)
239(71%) 176 (70%)
14 (4%) 16 (7%)
19(6%) II (4%)
1.882
0.5972
4(1%) 4(2%)
83 (25%) 40 (16%)
13 (4%) 12 (5%)
235 (70%) 196(77%)
7.005
0.07173
37 (12%) 43 (17%)
98 (29%) 60 (24%)
21 (6%) 19(7%)
179 (53%) 130(52%)
5.840
0.11966
86 (26%) 75 (30%)
2.445
0.4852
274(82%) 210(83%)
1.721
0.63231
SV SV
Diagonal
IMA
SV Circumflex
IMA
SV Right main
IMA
112 (33%) 70 (28%)
101 (30%) 79 (31%)
36 (11%) 28 (11%)
Posterior descending
IMA
II (3%) II (5%)
35 (10%) 24 (9%)
15 (5%) 7 (3%)
SV
SV Legend: LAD, Left anterior descending.
Table IV. Comparison of clinical results between IMA and SV grafts in total patient population Variable
Graft type
No. without variable (%)
No. with variable (%)
Chi square
p Value
Angina
IMA
398 (94.1%) 345 (91.0%)
25 (5.9%) 44 (9.0%)
7.601
0.00583
Myocardial infarction
IMA
417 (98.5%) 364 (93.6%)
6(1.5%) 25 (6.7%)
13.842
0.00020
Death
IMA
404 (95.5%) 338 (86.9%)
19 (4.5%) 51 (13.1%)
22.242
0.00001
408 (96.4%) 359 (92.3%)
15 (3.6%) 30 (7.7%)
6.718
0.00954
SV
SV SV
Postoperative CHF
IMA
SV Legend: CHF, Congestive heart failure.
failure, (4) generalized ventricular dysfunction by ventriculogram, and (5) acute myocardial infarction. Bias was eliminated by removing these subgroups and repeating the distribution analysis. Table II indicates that the distribution of the two groups among the independent variables now appears uniform. The anatomic severity of the coronary artery disease is similar in the redefined subsets (Table III). Patients with IMA grafts averaged 2.06 major arteries with 70 percent or greater obstruction and those with SV grafts averaged 2.01 arteries similarly involved. The mean internal diameter of the coronary arteries bypassed, as measured with Parsonett probes, was 1.98 mm. for the IMA group and 2.0 mm. for the SV group. Also, the ventricular function as estimated by ventriculogram was almost identical after exclusion of those in the generalized dysfunction category. Therefore, after the maldistributed subsets were removed, the groups for comparison had similar severity of clinical disease, anatomic location of disease, and other variables.
Results The comparison of the clinical results of the entire group is shown in Table IV. For the four clinical variables examined, the patients who had internal mammary artery bypasses had significantly better clinical results. The determinants of clinical success in the unbiased groups also showed that the patients with IMA grafts had a better course, but anginal relief and postoperative congestive heart failure were not improved to a statistically significant extent (Table V). In the unbiased groups survival rates again were improved in the IMA group: 97.6 percent survival in the IMA group compared with 91.3 percent in the SV group. These results are highly significant (p = 0.000862) in favor of increased survival in the group with IMA conduits. Two circumstances that we thought deserved consideration regarding the comparison were (I) a longer follow-up of the patients with SV grafts because IMA grafts were not done between 1969 and 1971 and (2) a
The Journal of
628 Jones et al.
Thoracic and Cardiovascular Surgery
Table V. Comparison of clinical results between IMA and SV grafts in unbiased patient population Variable
Graft type
Angina
IMA
Myocardial infarction
IMA
Death
IMA
SV SV SV
Postoperative CHF
IMA
SV
No. without variable (%)
No. with variable (%)
316 (94.3%) 227 (90.1%)
Chi square
p Value
19 (5.7%) 25 (9.9%)
3.744
0.0530
331 (98.8%) 242 (96.0%)
4 (1.2%) 10 (4.0%)
4.754
0.02923
326 (97.3%) 231 (91.7%)
9 (2.7%) 21 (8.3%)
9.508
0.00862
331 (98.8%) 248 (98.4%)
4 (1.2%) 4 (1.6%)
0.165
0.68447
Legend: CHF, Congestive heart failure.
higher degree of competence of the surgeons doing IMA grafts by virtue of having gained experience on the S V grafts. Examination revealed that these factors had little effect on the outcome. First, the results were I death/2,005 patient-months of observation in the IMA group compared with I death1716 patient-months of observation in the SV group. The mean length of follow-up (patient-months of observation) was not significantly different between the two groups (F = 2.86; df = 1,489; P < 0.05). Second, we eliminated the subjects who had operation before 1971. In the remaining subsets, 222 patients had SV grafts alone, and there were 13 deaths or 94 percent survival; by contrast, there were 298 patients in the IMA group, with six deaths or 98 percent survival. The increase in longevity remains valid (p = 0.0155). Therefore, our finding that longevity is significantly increased when patients have IMA grafts rather than S V grafts remains valid when examined from many viewpoints. In a previously reported study of these patients, we 7 matched the groups as to vessels bypassed, and improvement of survival was shown by the patients with IMA grafts; however, these results were statistically significant only in groups with high mortality rates. When the matched combinations were totaled, we found that of 260 patients with IMA grafts, 249 (96 percent) are living, and of 290 patients with SV grafts, 248 (86 percent) are living (p = 0.00005). Cumulative survival curves comparing the two groups emphasize the continued attrition rate in the groups who had SV bypasses and a remarkably stable survival rate of the patients with IMA grafts (Fig. I). When the two curves were compared with the expected survival curves of white men 53 years of age (the average age of the patients in this study) in Louisiana, the results indicated that the longevity of patients with IMA grafts equals or exceeds that of "normal" men by
about 3 years after operation and survival in the patients with saphenous vein grafts does not. Discussion Among all the factors determining outcome in the complex interaction of coronary artery disease and surgical treatment, the surgeon has the greatest control of the technical details of operation; therefore, methods which improve clinical results by altering the operative technique are of paramount importance. It is logical that a graft remaining patent longer should manifest an effect through improved long-term results. In this regard, longer life span in patients having IMA grafts allows a choice at the operating table to benefit the patient within the limitations imposed by the anatomic situation. Among the reported disadvantages of the IMA as a coronary bypass, none is as serious as that blood flow to the myocardium is reduced, perhaps to a critical level." A unique method developed by Flemma and associates" of comparing the blood flow of the two grafts to the same vascular bed in man clearly showed that the SV maintained higher flow rates. Greens criticized the study on the basis that the mammary arteries were probably in spasm at the time of measurement. The controversy is further demonstrated by some investigators who report decreased average flow rates in the IMA vessels at the time of operation"- 9 and in postoperative patients ,6, 10 and by others!' 10 who report no difference in flow rate in man at the time of operation. Also, in an experimental study" using dogs, no difference was found between the IMA and S V in base-line flow or in response to several vasoactive agents. The vein grafts in the latter study originated from the left subclavian artery, rather than the ascending aorta, and this fact might support the concept that increased flow in the S V grafts is secondary to a better
Volume 75
IMA bypass graft
Number 4
629
April, 1978
pulse wave contour resulting in a better flow during diastole. The real issue is not whether flow is less in IMA grafts, but whether the blood supply to the myocardium is adequate. An attempt to answer this question was made in a report which compared patients by graded exercise testing before and after single graft placement. 4 Single patent grafts of either IMA or S V were placed to the left anterior descending coronary artery, and improvement in graded exercise testing was no different. Our data support the concept that there is no difference in the function of the two grafts for relief of angina. An anatomic limitation of the IMA is lack of sufficient diameter for an adequate length and, as shown by Grondin and associates;" flow through vessels of inadequate size is markedly reduced. This problem is more frequently seen in women than men, but women are less frequently candidates for bypass. In the present study, the IMA was used in 57 percent of men and only 39 percent of women. The diameters of the IMA grafts in the present study were measured at operation, and only 34 patients had grafts placed that were smaller than 2.0 mm.; thus the problem of flow reduction may have been avoided in our patients. Unsuitable IMA grafts will produce a high percentage of technical failures. Suspected reasons for IMA failure include (I) inadequate conduit size (less than 1.5 mm. in diameter); (2) damage to the artery during operative preparation; (3) postoperative stretching of the graft by hyperinflated lungs; (4) anastomosis to excessively rigid, thickened coronary vessels without excising coronary wall; (5) exposure to undiluted papaverine solution; (6) insertion into marginally stenotic coronary vessels with competition of flow; (7) reduction of inflow by proximal stenoses of IMA or subclavian arteries; and (8) inclusion in suture line of surrounding tissues. Both preoperative and operative precautions are necessary to prevent these pitfalls and thereby insure the highest possible long-term patency. Selective IMA angiography is presently the preferred method of evaluation at the Ochsner Clinic. The injection of lidocaine hydrochloride before the dye will decrease the associated pain to tolerable levels. However, when patients are referred who already have had coronary angiograms without selective IMA studies, certain criteria help to determine whether the IMA can be used as the graft. Use of the IMA is not considered if subclavian murmurs are present, if a blood pressure gradient greater than 15 mm. Hg between the two arms is
10OppIiiiiO==---------------------,
95
.................~., »<.
~
Expected Survive]
.~ 90 .~
~ ~
85
...~
80
(' Survival curve for white moles in Louisiana at age 53 yrs)
0'---------':,------'-------'------'---------'------'-----------'-----'----'
2
3
4
5
>5
Years
Fig. 1. Comparison of survival curves for all patients in the SVG group and IMA group with the expected survival curve of normal 53-year-old men in Louisiana. *From U.S. Decennial Life Tables for 1969-1971. Vol 2 (19): DHEW publication No. HRA 75-1151.
present, or if hyperinflation of the lungs is seen on chest x-ray films. Operative criteria which must be met include (1) a prepared lumen 1.5 mm. in diameter; (2) a free flow of greater than 100 ml. per minute into a beaker; and (3) no pressure gradient between the end of the IMA and the radial artery. Also, any evidence of trauma to the vessel at the time of removal from the chest wall should preclude its use. The least valid criticism of the IMA, although it is true, is the greater technical difficulty in obtaining the graft and inserting it as a bypass. The difficulty can be overcome, and the improved results are worth the increased time of operation. Some explanation is required for our finding that, although IMA grafts increase longevity and decrease myocardial infarctions over SV grafts, there is no statistically significant difference in relief of angina, especially since longevity is increased in those in whom angina is resolved." Although improvement in angina is strongly associated with graft patency, some patients experience relief of pain with no patent bypass grafts. I I - 14 Achuff and associates's explained this paradox by the concept of an "angina-producing myocardial segment." They believe that this area could have become less painful because perfusion has been increased by a patent graft to the artery supplying the area or by a graft to a neighboring coronary vessel with collateral flow. In the absence of increased myocardial perfusion, a plausible explanation is that an asymptomatic myocardial infarction has destroyed the seg-
630
Jones et
The Journal of
at.
ment. 15 Several other investigators'< 16 have reported myocardial infarctions in those with anginal relief without patent grafts following aorta-coronary bypass. Finally, the high degree of efficacy of anginal relief by both of the types of conduit predisposes to statistical nonrecognition of further improvements except in comparisons of extremely large groups of patients. Thus angina may be relieved by a variety of different mechanisms which could mask the rather narrow margin of differences between the two types of conduits. The mechanism whereby late closure of vein or artery bypasses occurs is fundamental to understanding the problem. If late closure is not related to technical problems of the anastomosis or continuation of atherosclerosis in the coronary circulation, one is left to consider disease of the conduit itself. Neither the IMA 6 nor the SY17-20 is immune to arteriosclerosis. Experimentally, the S Y exhibits a greater propensity to the development of arteriosclerosis than either the IMA or coronary vessels." This finding may account for a minor portion of late graft closures. A more important pathological change which many of the S Y grafts undergo is subendothelial proliferation;" In one study these changes were present in varying degrees in all patients who had vein grafts patent for over a month.F The hemodynamic effect depends on the degree of constriction of the lumen, whether by extension of intimal thickening'? or thrombosis." In a study of the probable causes of late graft closure, approximately 10 percent of vein grafts closed because of stenosis of the aortic anastomosis." Thus the fact that the proximal anastomosis of the IMA graft is by nature rather than by surgical manipulation would eliminate at least one possible site of late closure. Mortality correlates with graft occlusion, which in the SY continues at a diminished rate after the subject is discharged from the hospital. 24 The IMA pedicle grafts, on the other hand, show a diminished or nonexistent rate of occlusion. Another important correlate of mortality is the importance of the vessel that occludes. Multiple bypasses with collateral flow may have overcome angina, but if the bypass upon which the heart has become dependent occludes, the result would be a pain-free individual until an early death. Evidence of increased longevity from aorta-coronary bypass surgery has long been awaited. With the prevalence of heart disease at epidemic proportions, such evidence should open entirely new indications for application of the procedure, especially since it appears that an irreversible stage is reached in chronically ischemic hearts. The life expectancy for 53-year-old persons (the average age of patients bypassed in the pres-
Thoracic and Cardi ovascular Surgery
ent series) is 20 additional years. Therefore, patency of grafts should be considered not only for the immediate operative and postoperative periods, but also for periods of 20 or more years. It is hoped that the final answer for arteriosclerosis will be found before the present generation of these grafts has reached this age. However, until then, a strong plea is made to use the IMA whenever possible to secure the best chance of survival for the patient. REFERENCES
2
3
4
5
6
7
8 9
10
II
12
13
14
Geha AS, Krone RJ, McConnick JR, et al: Selection of coronary bypass. Anatomic, physiological, and angiographic consideration of vein and mammary artery grafts. J THORAC CARDIOVASC SURG 70:414-431, 1975 Kay EB, Naraghipour H, Beg RA, et al: Internal mammary artery bypass graft. Long-term patency rate and folIow-up. Ann Thorac Surg 18:269-279, 1974 McConnick JR, Kaneko M, Baue AE, et al: Blood flow and vasoactive drug effects in internal mammary and venous bypass grafts. Circulation 51, 52:Suppl 1:72-80, 1975 Siegel W, Loop FD, Proudfit WL, et al: Comparison of internal mammary artery and saphenous vein grafts for myocardial revascularization by exercise testing (abstr). Circulation 51, 52:Suppl 2:142, 1975 Flemma RJ, Singh HM, Tector AJ, et al: Comparative hemodynamic properties of vein and mammary artery in coronary bypass operations. Ann Thorac Surg 20:619627,1975 Hamby RI, Aintablian A, Wisoff BG, et al: Comparative study of the postoperative flow in the saphenous vein and internal mammary artery bypass grafts. Am Heart J 93:306-315, 1977 Jones JW, Ochsner JL, Mills NL, et al: Impact of multiple variables on operative and extended mortality rates. Surgery 83:20-26, 1978. Green GE: Discussion of Flemma RJ, et al 5 Singh H. Flemma RJ, Tector AJ, et al: Direct myocardial revascularization. Determinants in the choice of vein graft or internal mammary artery. Arch Surg 107:699-703, 1973 Grondin CM, Lesperance J, Bourassa MG, Campeau L: Coronary artery grafting with the saphenous vein or internal mammary artery. Ann Thorac Surg 20:605-618, 1975 Benchimol A, dos Santos A, Desser KB: Relief of angina pectoris in patients with occluded coronary bypass grafts. Am J Med 60:339-343, 1976 Benchimol A, dos Santos A, Desser KB: Relief of angina pectoris in patients with occluded coronary bypass grafts. Circulation 51, 52:Suppl 2: 158, 1975 DiLuzio V, Roy PR, Sow ton E: Angina in patients with occluded aorto-coronary vein grafts. Br Heart J 36:139147, 1974 Frick MH: An appraisal of symptom relief after coronary bypass grafting. Postgrad Med J 52:765-769, 1976
Volume 75
IMA bypass graft
Number 4 April,1978
15 Achuff SC, Griffith LS, Conti R, et al: The "anginaproducing" myocardial segment. An approach to the interpretation of results of coronary bypass surgery. Am J Cardiol 36:723-733, 1975 16 Sheldon WC, Rincon G, Pichard A: Results of graft surgery for coronary artery disease: A progress report. Cardiovasc Clin 7:93-107, 1975 17 Barboriak JJ, Batayia GE, Pintar K, et al: Pathological changes in surgically removed aortocoronary vein grafts. Ann Thorac Surg 21:524-527, 1976 18 Barboriak JJ, Pintar K, Korns ME: Atherosclerosis in aortocoronary vein grafts. Lancet 2:621-624, 1974 19 Farry J, Hammond GL, Cohen LS, et al: Development of atherosclerosis in a saphenous vein coronary bypass graft. Ann Intern Med 85:478-479, 1976 20 Suzuki A, Hardy JD: Comparison of coronary bypass
21
22
23
24
63 I
procedure. Saphenous vein graft versus internal mammary artery graft. Bull Soc Int Chir 34: 197-205, 1975 Rossiter SJ, Brody WR, Kosek JC, et al: Internal mammary artery versus autogenous vein for coronary artery bypass graft. Circulation 50:1236-1243, 1974 Unni KK, Kottke BA, Titus JL, et al: Pathologic changes in aortocoronary saphenous vein grafts. Am J Cardiol 34:526-532, 1974 Grondin CM, Lesperance J, Bourassa MG, et al: Serial angiographic evaluation in 60 consecutive patients with aorto-coronary artery vein grafts 2 weeks, I year, and 3 years after operation. J THoRAc CARDIOVASC SURG 67: 1-6, 1974 Cheanvechai C, Irarrazaval MJ, Loop FD, et al: Aortacoronary bypass grafting with the internal mammary artery. J THoRAc CARDIOVASC SURG 70:278-281,1975
American Board of Thoracic Surgery Examination The 1979 annual certifying examination of the American Board of Thoracic Surgery (written and oral) will be held on March 22,23, and 24, 1979, in Chicago, Illinois. Final date for filing application is August I, 1978. Please address all communications to the American Board of Thoracic Surgery, 14640 East Seven Mile Road, Detroit, Michigan 48205.
James W. Jones, M.D.,* John L. Ochsner, M.D., Noel L. Mills, M.D., and Larry Hughes, Ph.D., New Orleans, La.
T
he prevailing attitude that the internal mammary artery (IMA) is a better conduit than the saphenous vein (5 V) for coronary bypass grafts is based on evidence of superior patency rates.!"" A less popular concept is that the IMA graft is inferior, except in certain circumstances, because of lessened flow rates.P- 6 Clinical results, and not angiographic patency or physiological measurements, remain the best means for determining superiority of an operative technique; therefore, the following study was done to provide a comparison of the long-term outcome between patients who had the IMA and those who had the 5 V used as coronary bypass grafts.
Patients and methods This analysis is based on computerized data obtained from the initial 1,004 consecutive patients who had direct coronary artery bypass at the Ochsner Medical Institutions. A previous report? dealt with other findings from this same study. The operative proceFrom the Department of Surgery, Ochsner Medical Institutions, New Orleans, La. Received for publication Oct. 4, 1977. Accepted for publication Nov. 7, 1977. Address for reprints: Dr. John L. Ochsner, Ochsner Clinic, 1514 Jefferson Highway, New Orleans, La. 70121. 'Present address: Department of Surgery, Tulane Medical Center, New Orleans, La. 70112.
0022-5223/78/0475-0625$00.70/0 © 1978 The C. V. Mosby Co.
dures were done or supervised by either of two surgeons (J. L. O. or N. L. M.) between 1968 and 1975. The follow-up averaged 48.6 months and was complete to the present in 98 percent of cases. Operative success was judged by relief of angina, over-all survival, prevention of congestive heart failure, and nonfatal myocardial infarction rates. Patients were counted as being improved unless they reported that the frequency and intensity of angina was unchanged from or worse than before operation. All deaths are included in the analysis regardless of cardiac relationship. Those patients requiring treatment for chronic heart failure are included as such. Patients classified as having myocardial infarction had to have definite electrocardiographic evidence supporting the diagnosis. Clinical success was evaluated by comparing all the patients who had the two types of grafts as to the presence of any variables indicating clinical result. A cross tabulation was made of the two types of grafts among 195 independent clinical variables; a partial listing of these variables is shown in Table I. Uneven distribution was found to exist in a number of independent variables, and these maldistributed variables were then cross-tabulated with increased mortality. The following variables were found to be associated with increased mortality: (1) female sex, (2) concomitant ventricular aneurysmectomy, (3) preoperative congestive heart 625
The Journal of
626
Jones et al.
Thoracic and Cardiovascular Surgery
Table I. Cross tabulation between IMA and SV grafts and variables* in total patient population Variable
No. with variable (%)
No. without variable (%)
Age
IMA SV
Below 50 yr. 141 (33%) 129 (33%)
50-59 yr. 196 (46%) 172 (44%)
Sex
IMA SV
Male 372 (88%) 310 (80%)
Female 51 (12%) 79 (20%)
Valve repair
IMA SV
7 (2%) 9 (2%)
Ventricular aneurysm
IMA SV
3 (7%) 9 (2%)
Aneurysm of ventriclet
IMA SV
State of ventriclet
p Value 60-69 yr. 80 (19%) 77 (20%)
70-79 yr. 6 (2%) 11 (3%)
5.379
0.25058
10.261
0.00136
416 (98%) 380 (98%)
0.455
0.50018
420 (99.3%) 380 (98%)
3.582
0.05840
9 (2%) 17 (4%)
414 (98%) 372 (96%)
3.287
0.06982
IMA SV
Normal 254 (60%) 202 (52%)
Seg. dysfunction III (26%) 114 (27%)
6.274
0.04341
Elevated triclycerides
IMA SV
183 (43%) 178 (46%)
240 (57%) 211 (54%)
0.510
0.4750
Diabetes
IMA SV
64(15%) 47 (12%)
359 (85%) 342 (87%)
1.594
0.2066
Hypertension
IMA SV
158 (37%) 153 (39%)
265 (73%) 236 (61%)
0.335
0.56255
Obesity
IMA SV
154 (36%) 154 (40%)
269 (64%) 235 (60%)
0.871
0.35063
Smoking
IMA SV
317(75%) 291 (75%)
106 (25%) 98 (25%)
0.002
0.96861
CHF
IMA SV
31 (7%) 57 (15%)
392 (93%) 332 (85%)
11.249
0.00080
Gen. dysfunction 58 (14%) 73 (21%)
Legend: IMA, Internal mammary artery. SV, Saphenous vein. CHF, Congestive heart failure.
'Partiallisting of 195 clinical variables statistically analyzed. t Aneurysm and state of ventricle as determined by preoperative ventriculogram.
Table II. Cross tabulation between IMA and SV grafts and variables* in unbiased patient population Variable
No. with variable (%)
No. without variable (%)
Below 50 yr. 117 (35%) 89 (35%)
50-59 yr. 159 (47%) 114 (45%)
60-69 yr. 57 (17%) 44 (18%)
70-79 yr. 2 (1%) 5 (2%)
Chi square
p Value
5.128
0.27443
Age
IMA SV
Valve repair
IMA SV
3 (1%) 2 (1%)
332 (99%) 250 (99%)
0.017
0.8951
Elevated cholesterol
IMA SV
30 (9%) 16(6%)
305 (91%) 236 (94%)
1.352
0.2444
Elevated triglycerides
IMA SV
151 (45%) 119 (47%)
184 (55%) 133 (53%)
0.266
0.6057
Hypertension
IMA SV
124 (37%) 82 (33%)
211 (63%) 170 (67%)
1.264
0.26089
Diabetes
IMA SV
47 (14%) 27 (11%)
288 (86%) 225 (89%)
1.435
0.2309
'Partial listing of variables analyzed.
Volume75 Number 4 April,1978
IMA bypass graft
627
Table III. Severity of coronary artery disease cross tabulated with IMA or SV grafts >70% constricted
Coronary artery
<70% constricted
Chi square
p Value
Main left
IMA
3(1%) 6(3%)
39(11%) 25 (10%)
16 (5%) 8 (3%)
277 (83%) 213 (84%)
3.420
0.33123
LAD
IMA
63 (19%) 49(19%)
239(71%) 176 (70%)
14 (4%) 16 (7%)
19(6%) II (4%)
1.882
0.5972
4(1%) 4(2%)
83 (25%) 40 (16%)
13 (4%) 12 (5%)
235 (70%) 196(77%)
7.005
0.07173
37 (12%) 43 (17%)
98 (29%) 60 (24%)
21 (6%) 19(7%)
179 (53%) 130(52%)
5.840
0.11966
86 (26%) 75 (30%)
2.445
0.4852
274(82%) 210(83%)
1.721
0.63231
SV SV
Diagonal
IMA
SV Circumflex
IMA
SV Right main
IMA
112 (33%) 70 (28%)
101 (30%) 79 (31%)
36 (11%) 28 (11%)
Posterior descending
IMA
II (3%) II (5%)
35 (10%) 24 (9%)
15 (5%) 7 (3%)
SV
SV Legend: LAD, Left anterior descending.
Table IV. Comparison of clinical results between IMA and SV grafts in total patient population Variable
Graft type
No. without variable (%)
No. with variable (%)
Chi square
p Value
Angina
IMA
398 (94.1%) 345 (91.0%)
25 (5.9%) 44 (9.0%)
7.601
0.00583
Myocardial infarction
IMA
417 (98.5%) 364 (93.6%)
6(1.5%) 25 (6.7%)
13.842
0.00020
Death
IMA
404 (95.5%) 338 (86.9%)
19 (4.5%) 51 (13.1%)
22.242
0.00001
408 (96.4%) 359 (92.3%)
15 (3.6%) 30 (7.7%)
6.718
0.00954
SV
SV SV
Postoperative CHF
IMA
SV Legend: CHF, Congestive heart failure.
failure, (4) generalized ventricular dysfunction by ventriculogram, and (5) acute myocardial infarction. Bias was eliminated by removing these subgroups and repeating the distribution analysis. Table II indicates that the distribution of the two groups among the independent variables now appears uniform. The anatomic severity of the coronary artery disease is similar in the redefined subsets (Table III). Patients with IMA grafts averaged 2.06 major arteries with 70 percent or greater obstruction and those with SV grafts averaged 2.01 arteries similarly involved. The mean internal diameter of the coronary arteries bypassed, as measured with Parsonett probes, was 1.98 mm. for the IMA group and 2.0 mm. for the SV group. Also, the ventricular function as estimated by ventriculogram was almost identical after exclusion of those in the generalized dysfunction category. Therefore, after the maldistributed subsets were removed, the groups for comparison had similar severity of clinical disease, anatomic location of disease, and other variables.
Results The comparison of the clinical results of the entire group is shown in Table IV. For the four clinical variables examined, the patients who had internal mammary artery bypasses had significantly better clinical results. The determinants of clinical success in the unbiased groups also showed that the patients with IMA grafts had a better course, but anginal relief and postoperative congestive heart failure were not improved to a statistically significant extent (Table V). In the unbiased groups survival rates again were improved in the IMA group: 97.6 percent survival in the IMA group compared with 91.3 percent in the SV group. These results are highly significant (p = 0.000862) in favor of increased survival in the group with IMA conduits. Two circumstances that we thought deserved consideration regarding the comparison were (I) a longer follow-up of the patients with SV grafts because IMA grafts were not done between 1969 and 1971 and (2) a
The Journal of
628 Jones et al.
Thoracic and Cardiovascular Surgery
Table V. Comparison of clinical results between IMA and SV grafts in unbiased patient population Variable
Graft type
Angina
IMA
Myocardial infarction
IMA
Death
IMA
SV SV SV
Postoperative CHF
IMA
SV
No. without variable (%)
No. with variable (%)
316 (94.3%) 227 (90.1%)
Chi square
p Value
19 (5.7%) 25 (9.9%)
3.744
0.0530
331 (98.8%) 242 (96.0%)
4 (1.2%) 10 (4.0%)
4.754
0.02923
326 (97.3%) 231 (91.7%)
9 (2.7%) 21 (8.3%)
9.508
0.00862
331 (98.8%) 248 (98.4%)
4 (1.2%) 4 (1.6%)
0.165
0.68447
Legend: CHF, Congestive heart failure.
higher degree of competence of the surgeons doing IMA grafts by virtue of having gained experience on the S V grafts. Examination revealed that these factors had little effect on the outcome. First, the results were I death/2,005 patient-months of observation in the IMA group compared with I death1716 patient-months of observation in the SV group. The mean length of follow-up (patient-months of observation) was not significantly different between the two groups (F = 2.86; df = 1,489; P < 0.05). Second, we eliminated the subjects who had operation before 1971. In the remaining subsets, 222 patients had SV grafts alone, and there were 13 deaths or 94 percent survival; by contrast, there were 298 patients in the IMA group, with six deaths or 98 percent survival. The increase in longevity remains valid (p = 0.0155). Therefore, our finding that longevity is significantly increased when patients have IMA grafts rather than S V grafts remains valid when examined from many viewpoints. In a previously reported study of these patients, we 7 matched the groups as to vessels bypassed, and improvement of survival was shown by the patients with IMA grafts; however, these results were statistically significant only in groups with high mortality rates. When the matched combinations were totaled, we found that of 260 patients with IMA grafts, 249 (96 percent) are living, and of 290 patients with SV grafts, 248 (86 percent) are living (p = 0.00005). Cumulative survival curves comparing the two groups emphasize the continued attrition rate in the groups who had SV bypasses and a remarkably stable survival rate of the patients with IMA grafts (Fig. I). When the two curves were compared with the expected survival curves of white men 53 years of age (the average age of the patients in this study) in Louisiana, the results indicated that the longevity of patients with IMA grafts equals or exceeds that of "normal" men by
about 3 years after operation and survival in the patients with saphenous vein grafts does not. Discussion Among all the factors determining outcome in the complex interaction of coronary artery disease and surgical treatment, the surgeon has the greatest control of the technical details of operation; therefore, methods which improve clinical results by altering the operative technique are of paramount importance. It is logical that a graft remaining patent longer should manifest an effect through improved long-term results. In this regard, longer life span in patients having IMA grafts allows a choice at the operating table to benefit the patient within the limitations imposed by the anatomic situation. Among the reported disadvantages of the IMA as a coronary bypass, none is as serious as that blood flow to the myocardium is reduced, perhaps to a critical level." A unique method developed by Flemma and associates" of comparing the blood flow of the two grafts to the same vascular bed in man clearly showed that the SV maintained higher flow rates. Greens criticized the study on the basis that the mammary arteries were probably in spasm at the time of measurement. The controversy is further demonstrated by some investigators who report decreased average flow rates in the IMA vessels at the time of operation"- 9 and in postoperative patients ,6, 10 and by others!' 10 who report no difference in flow rate in man at the time of operation. Also, in an experimental study" using dogs, no difference was found between the IMA and S V in base-line flow or in response to several vasoactive agents. The vein grafts in the latter study originated from the left subclavian artery, rather than the ascending aorta, and this fact might support the concept that increased flow in the S V grafts is secondary to a better
Volume 75
IMA bypass graft
Number 4
629
April, 1978
pulse wave contour resulting in a better flow during diastole. The real issue is not whether flow is less in IMA grafts, but whether the blood supply to the myocardium is adequate. An attempt to answer this question was made in a report which compared patients by graded exercise testing before and after single graft placement. 4 Single patent grafts of either IMA or S V were placed to the left anterior descending coronary artery, and improvement in graded exercise testing was no different. Our data support the concept that there is no difference in the function of the two grafts for relief of angina. An anatomic limitation of the IMA is lack of sufficient diameter for an adequate length and, as shown by Grondin and associates;" flow through vessels of inadequate size is markedly reduced. This problem is more frequently seen in women than men, but women are less frequently candidates for bypass. In the present study, the IMA was used in 57 percent of men and only 39 percent of women. The diameters of the IMA grafts in the present study were measured at operation, and only 34 patients had grafts placed that were smaller than 2.0 mm.; thus the problem of flow reduction may have been avoided in our patients. Unsuitable IMA grafts will produce a high percentage of technical failures. Suspected reasons for IMA failure include (I) inadequate conduit size (less than 1.5 mm. in diameter); (2) damage to the artery during operative preparation; (3) postoperative stretching of the graft by hyperinflated lungs; (4) anastomosis to excessively rigid, thickened coronary vessels without excising coronary wall; (5) exposure to undiluted papaverine solution; (6) insertion into marginally stenotic coronary vessels with competition of flow; (7) reduction of inflow by proximal stenoses of IMA or subclavian arteries; and (8) inclusion in suture line of surrounding tissues. Both preoperative and operative precautions are necessary to prevent these pitfalls and thereby insure the highest possible long-term patency. Selective IMA angiography is presently the preferred method of evaluation at the Ochsner Clinic. The injection of lidocaine hydrochloride before the dye will decrease the associated pain to tolerable levels. However, when patients are referred who already have had coronary angiograms without selective IMA studies, certain criteria help to determine whether the IMA can be used as the graft. Use of the IMA is not considered if subclavian murmurs are present, if a blood pressure gradient greater than 15 mm. Hg between the two arms is
10OppIiiiiO==---------------------,
95
.................~., »<.
~
Expected Survive]
.~ 90 .~
~ ~
85
...~
80
(' Survival curve for white moles in Louisiana at age 53 yrs)
0'---------':,------'-------'------'---------'------'-----------'-----'----'
2
3
4
5
>5
Years
Fig. 1. Comparison of survival curves for all patients in the SVG group and IMA group with the expected survival curve of normal 53-year-old men in Louisiana. *From U.S. Decennial Life Tables for 1969-1971. Vol 2 (19): DHEW publication No. HRA 75-1151.
present, or if hyperinflation of the lungs is seen on chest x-ray films. Operative criteria which must be met include (1) a prepared lumen 1.5 mm. in diameter; (2) a free flow of greater than 100 ml. per minute into a beaker; and (3) no pressure gradient between the end of the IMA and the radial artery. Also, any evidence of trauma to the vessel at the time of removal from the chest wall should preclude its use. The least valid criticism of the IMA, although it is true, is the greater technical difficulty in obtaining the graft and inserting it as a bypass. The difficulty can be overcome, and the improved results are worth the increased time of operation. Some explanation is required for our finding that, although IMA grafts increase longevity and decrease myocardial infarctions over SV grafts, there is no statistically significant difference in relief of angina, especially since longevity is increased in those in whom angina is resolved." Although improvement in angina is strongly associated with graft patency, some patients experience relief of pain with no patent bypass grafts. I I - 14 Achuff and associates's explained this paradox by the concept of an "angina-producing myocardial segment." They believe that this area could have become less painful because perfusion has been increased by a patent graft to the artery supplying the area or by a graft to a neighboring coronary vessel with collateral flow. In the absence of increased myocardial perfusion, a plausible explanation is that an asymptomatic myocardial infarction has destroyed the seg-
630
Jones et
The Journal of
at.
ment. 15 Several other investigators'< 16 have reported myocardial infarctions in those with anginal relief without patent grafts following aorta-coronary bypass. Finally, the high degree of efficacy of anginal relief by both of the types of conduit predisposes to statistical nonrecognition of further improvements except in comparisons of extremely large groups of patients. Thus angina may be relieved by a variety of different mechanisms which could mask the rather narrow margin of differences between the two types of conduits. The mechanism whereby late closure of vein or artery bypasses occurs is fundamental to understanding the problem. If late closure is not related to technical problems of the anastomosis or continuation of atherosclerosis in the coronary circulation, one is left to consider disease of the conduit itself. Neither the IMA 6 nor the SY17-20 is immune to arteriosclerosis. Experimentally, the S Y exhibits a greater propensity to the development of arteriosclerosis than either the IMA or coronary vessels." This finding may account for a minor portion of late graft closures. A more important pathological change which many of the S Y grafts undergo is subendothelial proliferation;" In one study these changes were present in varying degrees in all patients who had vein grafts patent for over a month.F The hemodynamic effect depends on the degree of constriction of the lumen, whether by extension of intimal thickening'? or thrombosis." In a study of the probable causes of late graft closure, approximately 10 percent of vein grafts closed because of stenosis of the aortic anastomosis." Thus the fact that the proximal anastomosis of the IMA graft is by nature rather than by surgical manipulation would eliminate at least one possible site of late closure. Mortality correlates with graft occlusion, which in the SY continues at a diminished rate after the subject is discharged from the hospital. 24 The IMA pedicle grafts, on the other hand, show a diminished or nonexistent rate of occlusion. Another important correlate of mortality is the importance of the vessel that occludes. Multiple bypasses with collateral flow may have overcome angina, but if the bypass upon which the heart has become dependent occludes, the result would be a pain-free individual until an early death. Evidence of increased longevity from aorta-coronary bypass surgery has long been awaited. With the prevalence of heart disease at epidemic proportions, such evidence should open entirely new indications for application of the procedure, especially since it appears that an irreversible stage is reached in chronically ischemic hearts. The life expectancy for 53-year-old persons (the average age of patients bypassed in the pres-
Thoracic and Cardi ovascular Surgery
ent series) is 20 additional years. Therefore, patency of grafts should be considered not only for the immediate operative and postoperative periods, but also for periods of 20 or more years. It is hoped that the final answer for arteriosclerosis will be found before the present generation of these grafts has reached this age. However, until then, a strong plea is made to use the IMA whenever possible to secure the best chance of survival for the patient. REFERENCES
2
3
4
5
6
7
8 9
10
II
12
13
14
Geha AS, Krone RJ, McConnick JR, et al: Selection of coronary bypass. Anatomic, physiological, and angiographic consideration of vein and mammary artery grafts. J THORAC CARDIOVASC SURG 70:414-431, 1975 Kay EB, Naraghipour H, Beg RA, et al: Internal mammary artery bypass graft. Long-term patency rate and folIow-up. Ann Thorac Surg 18:269-279, 1974 McConnick JR, Kaneko M, Baue AE, et al: Blood flow and vasoactive drug effects in internal mammary and venous bypass grafts. Circulation 51, 52:Suppl 1:72-80, 1975 Siegel W, Loop FD, Proudfit WL, et al: Comparison of internal mammary artery and saphenous vein grafts for myocardial revascularization by exercise testing (abstr). Circulation 51, 52:Suppl 2:142, 1975 Flemma RJ, Singh HM, Tector AJ, et al: Comparative hemodynamic properties of vein and mammary artery in coronary bypass operations. Ann Thorac Surg 20:619627,1975 Hamby RI, Aintablian A, Wisoff BG, et al: Comparative study of the postoperative flow in the saphenous vein and internal mammary artery bypass grafts. Am Heart J 93:306-315, 1977 Jones JW, Ochsner JL, Mills NL, et al: Impact of multiple variables on operative and extended mortality rates. Surgery 83:20-26, 1978. Green GE: Discussion of Flemma RJ, et al 5 Singh H. Flemma RJ, Tector AJ, et al: Direct myocardial revascularization. Determinants in the choice of vein graft or internal mammary artery. Arch Surg 107:699-703, 1973 Grondin CM, Lesperance J, Bourassa MG, Campeau L: Coronary artery grafting with the saphenous vein or internal mammary artery. Ann Thorac Surg 20:605-618, 1975 Benchimol A, dos Santos A, Desser KB: Relief of angina pectoris in patients with occluded coronary bypass grafts. Am J Med 60:339-343, 1976 Benchimol A, dos Santos A, Desser KB: Relief of angina pectoris in patients with occluded coronary bypass grafts. Circulation 51, 52:Suppl 2: 158, 1975 DiLuzio V, Roy PR, Sow ton E: Angina in patients with occluded aorto-coronary vein grafts. Br Heart J 36:139147, 1974 Frick MH: An appraisal of symptom relief after coronary bypass grafting. Postgrad Med J 52:765-769, 1976
Volume 75
IMA bypass graft
Number 4 April,1978
15 Achuff SC, Griffith LS, Conti R, et al: The "anginaproducing" myocardial segment. An approach to the interpretation of results of coronary bypass surgery. Am J Cardiol 36:723-733, 1975 16 Sheldon WC, Rincon G, Pichard A: Results of graft surgery for coronary artery disease: A progress report. Cardiovasc Clin 7:93-107, 1975 17 Barboriak JJ, Batayia GE, Pintar K, et al: Pathological changes in surgically removed aortocoronary vein grafts. Ann Thorac Surg 21:524-527, 1976 18 Barboriak JJ, Pintar K, Korns ME: Atherosclerosis in aortocoronary vein grafts. Lancet 2:621-624, 1974 19 Farry J, Hammond GL, Cohen LS, et al: Development of atherosclerosis in a saphenous vein coronary bypass graft. Ann Intern Med 85:478-479, 1976 20 Suzuki A, Hardy JD: Comparison of coronary bypass
21
22
23
24
63 I
procedure. Saphenous vein graft versus internal mammary artery graft. Bull Soc Int Chir 34: 197-205, 1975 Rossiter SJ, Brody WR, Kosek JC, et al: Internal mammary artery versus autogenous vein for coronary artery bypass graft. Circulation 50:1236-1243, 1974 Unni KK, Kottke BA, Titus JL, et al: Pathologic changes in aortocoronary saphenous vein grafts. Am J Cardiol 34:526-532, 1974 Grondin CM, Lesperance J, Bourassa MG, et al: Serial angiographic evaluation in 60 consecutive patients with aorto-coronary artery vein grafts 2 weeks, I year, and 3 years after operation. J THoRAc CARDIOVASC SURG 67: 1-6, 1974 Cheanvechai C, Irarrazaval MJ, Loop FD, et al: Aortacoronary bypass grafting with the internal mammary artery. J THoRAc CARDIOVASC SURG 70:278-281,1975
American Board of Thoracic Surgery Examination The 1979 annual certifying examination of the American Board of Thoracic Surgery (written and oral) will be held on March 22,23, and 24, 1979, in Chicago, Illinois. Final date for filing application is August I, 1978. Please address all communications to the American Board of Thoracic Surgery, 14640 East Seven Mile Road, Detroit, Michigan 48205.