- Email: [email protected]
Cod-liver oil in the prevention of intimal hyperplasia in autogenous vein grafts used for arterial bypass
J
THoRAc CARDIOVASC SURG
89:351-357, 1985
Cod-liver oil in the prevention of intimal hyperplasia in autogenous vein grafts used for arterial bypass Cod-liver oil, rich in eicosapentaenoic acid, an unsaturated fatty acid, was administered to 14 mongrel dogs to detennine if this acid would prevent platelet-mediated intimal hyperplasia. Twenty-eight 1 cm segments of undistended jugular vein were interposed between bilaterally divided femoral arteries. Seven control animals were fed a 2 % cholesterol diet 1 week before and for 6 weeks after the operation. A further seven animals received cod-liver oil capsules containing 1.8 gm of eicosapentaenoic acid daily 1 week before and for 6 weeks after autogenous vein implantation, in addition to the lipid-supplemented diet Baseline serum cholesterol was 4.6 ± 0.4 mmol/L The rise in serum cholesterol was similar in the two groups and increased to 7.4 ± 0.6 mmoljL (control group) and to 6.8 ± 0.2 mmol/L (eicosapentaenoic acid group) (p < 0.001). Prothrombin time, partial thromboplastin time, bleeding time, and platelet counts were unchanged in the two groups. Vein grafts, harvested at 6 weeks, were fixed in formaldehyde. Mean intimal thickness was measured from multiple vein graft cross sections with a Zeiss computerized interactive image analyzing system. A mean of 140 ± 11 measurements were computed from each graft Marked intimal hyperplasia occurred in the control group and increased from 4.3 ± 0.3 to 86.4 ± 14 ILm. In contrast, a high eicosapentaenoic acid diet inhibited intimal hyperplasia, with intimal thickness only increasing from 4.0 ± 0.4 to 24.8 ± 2.7 ILm (p < 0.001). These data indicate that eicosapentaenoic acid inhibits platelet-mediated intimal hyperplasia and suggest that cod-liver oil could be used to prevent intimal hyperplasia in vein grafts used for myocardial revascularization.
R. W. Landymore, M.D., F.R.C.S.(C), C. E. Kinley, M.D., J. H. Cooper, M.D., M. MacAulay, M.D., B. Sheridan, M.D., and C. Cameron, B.N.,
Halifax, Nova Scotia, Canada
Erous intimal hyperplasia is the most common cause of late graft failure after aorta-coronary bypass.!:' Considerable experimental evidence suggests that intimal thickening is platelet mediated.' This concept is supported by recent clinical trials that have shown improved vein graft patency with antiplatelet drug regimens after myocardial revascularization.v" Clinical studies have shown that increased dietary From Dalhousie University, Department of Surgery, Division of Thoracic and Cardiovascular Surgery, Halifax, Nova Scotia, Canada. Supported by a Canadian Heart Foundation Grant. Read at the Tenth Annual Meeting of the Western Thoracic Surgical Association, Maui, Hawaii, June 20-23, 1984. Address for reprints: R. W. Landymore, M.D., F.R.C.S.(C), Room 3065, R. C. Dickson Centre, Victoria General Hospital, Halifax, Nova Scotia, Canada B3H 2Y9.
intake of an unsaturated fatty acid, eicosapentaenoic acid, results in decreased platelet aggregation and total platelet counts." Eicosapentaenoic acid is found in high concentrations in salt water fish and synthetically refined cod-liver oil.'? 11 This study was initiated to determine the effects of cod-liver oil on plateletmediated intimal hyperplasia in autogenous vein grafts used for arterial bypass. Materials and methods Segments of autogenous external jugular vein were interposed between bilaterally divided femoral arteries in 14 adult mongrel dogs, weighing between 25 and 30 kg. The animals were divided into two equal groups, both receiving a 2% cholesterol diet* 1 week before and for 6 weeks after operation. Seven animals were used as "I.C.N. Nutritional Biochemicals, Cleveland, Ohio.
351
The Journal of Thoracic and Cardiovascular Surgery
3 5 2 Landymore et al.
Fig. 1. Vein graft cross sections (original magnification X250). A, Vein graft before implantation (hematoxylineosin stain). B, Vein graft from a control animal (Masson stain). C, Marked reduction in intimal thickness that occurs with eicosapentaenoic acid (Masson stain).
controls and seven other animals received 20 minim capsules of cod-liver oil containing 1.8 gm of eicosapentaenoic acid daily, 1 week before and for 6 weeks after implantation of the autogenous vein. The animals were given fentanyl (0.4 mg) and droperidol (20 mg) for preoperative sedation. Anesthesia was induced with sodium pentobarbitol (5 mg/kg), with maintenance doses of pentobarbitol (3 mg/kg) being administered as necessary. The animals were ventilated with the Bird Mark 7 pressure-regulated ventilator. A 3 ern segment of autogenous external jugular vein was harvested from the neck, and the common femoral arteries were exposed bilaterally by sterile technique. A segment of jugular vein was immediately fixed in formaldehyde for later examination under the light microscope. After heparin was given systemically in a dose of 2 mg/kg, 1 em segments of undistended external jugular vein were interposed between bilaterally divided common femoral arteries. The anastomoses were completed with running 6-0 Prolene suture, with the aid of optical magnification, Heparin was then reversed with protamine sulfate and the incisions were closed. Six weeks later the animals were given an overdose of sodium pentobarbitol, and the common femoral arteries, containing the interposed vein grafts, were harvested and fixed in formaldehyde. Cross sections of jugular vein were stained with hematoxylin-eosin and Masson stains, and intimal thickness was measured from multiple cross-sections of the vein grafts and a Zeiss computer-
ized interactive image analyzing system. * This system is composed of a Zeiss microscope, a computer, and a 12 inch television screen. Individual slides are projected on the television screen, and a mobile cursor, connected to the computer, is used to accurately measure intimal thickness in microns around the entire circumference of the graft. Hematocrit value, platelet count, prothrombin time, partial thromboplastin time, bleeding time, and serum cholesterol were determined prior to any dietary manipulation, after institution of the 2% cholesterol diet before the operation and 2 weeks and 4 weeks after implantation of autogenous vein grafts. Results have been reported as the arithmetic mean with standard error and the Student's t test was used for statistical comparison. Results An overall increase in total body weight of 3.5 ± 0.25 kg was observed in both groups while receiving the 2% cholesterol diet. Baseline serum cholesterol was 4.6 ± 0.4 mmoljL. Lipid supplementation resulted in a significant rise in serum cholesterol levels that was similar in the two groups. Serum cholesterol levels increased to 7.4 ± 0.6 mmol/L in the control animals and to 6.8 ± 0.2 mmoljL in those animals receiving eicosapentaenoic acid (p < 0.001). Hematocrit value, partial thromboplastin time, prothrombin time, bleeding time, and 'Karl Zeiss Canada Limited, Don Mills, Ontario, Canada.
Volume 89 Number 3
Intimal hyperplasia in vein grafts
March, 1985
353
Arachidonic Acid Eicosapentaenoic Acid
I
Cyclo-oxygenase
Cyclic Endoperoxides PGG z PGH z
Prostacyclin PGI 3
/\
Prostacyclin PGl z (Arterial Wall)
Thromboxane Aa TXA 3
Thromboxane Az TXA z (Platelet Membrane)
Fig. 2. Pathways for prostaglandin biosynthesis. Arachidonic acid is normally metabolized to proaggregatory thromboxane A 2 by platelet membranes, whereas blood vessel walls metabolize arachidonic acid to antiaggregatory prostacyclin 12, Eicosapentaenoic acid competes with arachidonic acid for cyclooxygenase, resulting in decreased production of thromboxane A 2• In addition, eicosapentaenoic acid may be metabolized to the weakly aggregatory prostanoid thromboxane A] by platelet membranes and to antiaggregatory prostacyclin I, by blood vessels.
platelet counts did not change significantly in either group during the study. Light microscopic examination of representative vein graft cross sections prior to implantation indicated that the intima was intact and had not been damaged during harvesting. Intimal thickness was measured from multiple vein graft cross sections with the Zeiss computerized interactive image analyzing system. A mean of 140 ± 11 measurements were computed from each vein graft cross section. Marked intimal hyperplasia occurred in the control animals receiving lipid supplementation and increased from 4.3 ± 0.3 to 86.4 ± 1.4 ~m. In contrast, eicosapentaenoic acid markedly reduced intimal hyperplasia, with intimal thickness only increasing from 4.0 ± 0.4 to 24.8 ± 2.7 ~m (p < 0.001). Photomicrographs of representative vein cross sections are shown in Fig. 1. Fig. 1, A is a vein cross section prior to implantation that has been stained with hematoxylineosin. The intima in this cross section is only one cell thick. Fig. 1, B shows marked intimal hyperplasia in a representative section from a control animal. In contrast to D, C illustrates the marked reduction in intimal hyperplasia that occurs in animals receiving eicosapentaenoic acid. Discussion Myocardial revascularization usually results in the immediate and dramatic relief of angina, with optimum results being directly related to vein graft patency." Although 85% of vein grafts are open at 1 year,I3 15 the 0
long-term patency rate is disappointing, with over 50% of vein grafts occluding within 10 years after myocardial revascularization." Early graft failure is usually related primarily to thrombosis, whereas late graft failure is secondary to progressive fibrous intimal hyperplasia-'' Experimental investigations have shown that intimal hyperplasia is platelet mediated and that platelets release a mitogenic serum factor that stimulates intimal proliferation.t" Metke and colleagues IS have added further experimental evidence to support this concept and have demonstrated that antiplatelet drug therapy with a combination of aspirin-dipyridamole reduces vein graft intimal hyperplasia in dogs undergoing aorta-left anterior descending coronary bypass. Subsequent clinical trials have demonstrated that this medical regimen decreases fibrous intimal hyperplasia and improves long-term graft patency after myocardial revascularization.' Although this early work is encouraging, optimum antiplatelet therapy has yet to be described. Epidemiologic studies indicate that myocardial infarction is uncommon in the Greenland Eskimo population."?' The low incidence of atherosclerotic heart disease appears to be related to a high dietary intake of fish oil, which is rich in the unsaturated fatty acid eicosapentaenoic acid." Clinical trials have demonstrated that consumption of cod-liver oil, rich in eicosapentaenoic acid, not only reduces total platelet counts but markedly interferes with platelet aggregation.>" Arachidonic acid is the precursor for prostaglandin biosynthesis in platelet membranes and arterial blood
The Journal of
354
Landymore et al.
vessels." Platelet membranes enzymatically convert arachidonic acid to proaggregatory thromboxane A 2 (Fig. 2), whereas blood vessels convert arachidonic acid to antiaggregatory prostacyclin (PGI 2) . 26 Eicosapentaenoic acid competes with arachidonic acid for the enzyme cyclooxygenase (Fig. 2), which results in decreased production of proaggregatory thromboxane A 2 by platelet membranes. Recent reports also indicate that eicosapentaenoic acid is metabolized to thromboxane A J , a weakly aggregatory prostanoid, whereas some of the intermediary metabolites of eicosapentaenoic acid may be metabolized to PGI J by arterial blood vesselS. 8- 11, 24, 27 Regardless of the actual mechanism of action of eicosapentaenoic acid, substitution of this unsaturated fatty acid for arachidonic acid in platelet membranes results in decreased thromboxane A b which causes a marked reduction in platelet aggregation. The known effects of eicosapentaenoic acid on platelet metabolism prompted the present study to determine whether eicosapentaenoic acid would inhibit platelet-mediated intimal hyperplasia in autogenous vein grafts for arterial bypass. Previously we" developed and reported upon a canine model that may be used to produce accelerated shortterm intimal hyperplasia in vein grafts used for arterial bypass. This model was used in this study to determine the effects of eicosapentaenoic acid on vein graft intimal hyperplasia. Interestingly, our data clearly demonstrate that eicosapentaenoic acid .inhibits platelet-mediated intimal hyperplasia. Intimal thickening was reduced by 350% in animals receiving cod-liver oil. The use of cod-liver oil to prevent intimal thickening after autogenous vein implantation is a particularly attractive concept because eicosapentaenoic acid has been ingested by the Eskimo population for over 100 years without untoward side effects. Although this study is only a preliminary investigation, our data suggest that cod-liver oil might be of use in the prevention of intimal hyperplasia in vein grafts used for myocardial revascularization. REFERENCES Chesebro J, Fuster V: Drug trials in prevention of occlusion of aorta-coronary artery vein grafts. J THoRAc CARDIOVASC SURG 83:90, 1982 2 Unni K, Kottke B, Titus J, Frye R, Wallace R, Brown A: Pathologic changes in aortocoronary saphenous vein graft. Am J Cardiol 34:526, 1974 3 Vlodaver Z, Edwards J: Pathologic changes in aorticcoronary arterial saphenous vein grafts, Circulation 64:719, 1971 4 Friedman R, Burns E: Role of platelets in the proliferative response of the injured artery. Prog Hemost Thromb 4:249, 1978
Thoracic and Cardiovascular Surgery
5 Chesebro J, Fuster V, Elveback L, Clements I, Smith H, Holmes 0, Bardsley W, Pluth J, Wallace R, Puga F, Orszulak T, Piehler J, Danielson G, Schaff H, Frye R: Effect of dipyridamole and aspirin on late vein graft patency after coronary bypass operations, N Engl J Med 310:209, 1984 6 Mayer JE, Maj MC, Lindsay WG, Castaneda A, Nicoloff OM: Influence of aspirin and dipyridamole on patency of coronary artery bypass grafts, Ann Thorac Surg 31:204, 1981 7 Siess W, Scherer B, Bohlig B, Roth P, Kurzmann I, Weber PC: Platelet-membrane fatty acids, platelet aggregation, and thromboxane formation during a mackerel diet. Lancet 1:441, 1980 8 Dyerberg, J" Bang, H.O" Stoffersen, E, Eicosapentaenoic acid and prevention of thrombosis and atherosclerosis? Lancet 2:117, 1978 9 Hay C, Durber A, Saynor R: Effect of fish oil on platelet kinetics in patients with ischemic heart disease, Lancet 1:1269,1982 10 Sanders T, Vickers M, Haines A: Effect on blood lipids and haemostasis of a supplement of cod-liver oil, rich in eicosapentaenoic and docosahexaenoic acids, in healthy young men. Clin Sci 61:317,1981 II Hirai A, Terano T, Hamazaki T, Sajiki J, Kondo S, Ozawa A, Fujita T, Miyamoto T, Tamura Y, Kumagai A: The effects of the oral administration of fish oil concentrate on the release and the metabolism of arachidonic acid and eicosapentaenoic acid by human platelets, Thromb Res 28:285, 1982 12 Cukingnan R, Carey J, Wittig J, Brown G: Influence of complete coronary revascularization on relief of angina, J THoRAc CARDIOYASC SURG 79: 188, 1980 13 Geha A, Baue A: Early and late results of coronary revascularization with saphenous vein and internal mammary artery grafts. Am J Surg 137:456, 1979 14 Little B, Loop F, Thurer R, Groves L, Taylor P, Cosgrove 0: Isolated left anterior descending coronary atherosclerosis, Long-term comparison of internal mammary artery and venous autografts, Circulation 61:869, 1980 15 Tyras 0, Barner H, Kaiser G, Codd J, Pennington 0, Willman V: Bypass grafts to the left anterior descending coronary artery, J THoRAc CARDIOVASC SURG 80:327, 1980 16 Campeau L, Enjalbert M, Lesperance J, Bourassa MG: Arteriosclerosis and late closure of aortocoronary saphenous vein grafts, Sequential angiographic studies I year, 5-7 years, and 1O-12 years after surgery (abstr). Circulation 66:Suppl 2:11-94, 1982 17 Ross R, Glomset J, Kariya B, Harker L: A plateletdependent serum factor that stimulates the proliferation of arterial smooth muscle cells in vitro, Proc Nat! Acad Sci USA 71:1207,1974 18 Metke M, Lie J, Fuster V, Jova M, Kaye M: Reduction of intimal thickening in canine coronary bypass vein grafts with dipyridamole and aspirin, Am J Cardiol 43:1144, 1979
Volume 89 Number 3 March, 1985
19 Harris WS, Connor WE, Goodnight SH Jr: Dietary fish oils, plasma lipids and platelets in man. Prog Lipid Res 20:75,1981 20 Dyerberg J, Bang HO: Atherogenesis and haemostasis in Eskimos. The role of the prostaglandin-S family. Haernostasis 8:227, 1979 21 Bang HO, Dyerberg J, Hjorne N: The composition of food consumed by Greenland Eskimos. Acta Med Scand 200:69, 1976 22 Dyerberg J, Bang HO, Stoffersen E, Moncada S, Vane JR: Eicosapentaenoic acid and prevention of thrombosis and atherosclerosis. Lancet 2: 117, 1978 23 Hay C, Durber A, Saynor R: Effect of fish oil on platelet kinetics in patients with ischemic heart disease. Lancet 1:1269,1982 . 24 Brox J, Killie J, Osterud B, Holme S, Nordoy A: Effects of cod-liver oil on platelets and coagulation in familial hypercholesterolemia (Type IIA). Acta Med Scand 213:137,1983 25 McKean M, Smith J, Silver M, Ahern D: Metabolism of radioactive 5, 8, II, 14, 17-eicosapentaenoic acid by human platelets. Prog Lipid Res 20:435, 1981 26 Willis A, Smith J: Some perspectives on platelets and prostaglandins. Prog Lipid Res 20:387, 1981 27 Fischer S, Weber P: Thromboxane A] (TXA 3) is formed in human platelets after dietary eicosapentaenoic acid. Biochem Biophys Res Comm 116:1091, 1983 28 Landymore R, Kinley C, Cameron C: Intimal hyperplasia in autogenous vein grafts used for arterial bypass. A canine model. Cardiovasc Res (in press) 29 Brox J, Killie J, Gunnes S, Nordoy A: The effect of cod-liver oil and corn oil on platelets and vessel wall in man. Thromb Haemost 46:604, 1981
Discussion DR. D. CRAIG MILLER Stanford, Calif
Having been working on various models of experimental atherogenesis in the laboratory for 7 or 8 years and being aware of the differential biological functions of the dienoic and trienoic prostanoids, I was intrigued to see this paper on the program. I felt that it would be unfair to comment on whether cod-liver oil is good or bad without having tried it myself, but I was spared this fate when I learned that my colleague Dr. Scott Mitchell had previously attempted to augment his intake of eicosapentaenoic acid. Dr. Mitchell, however, could not tolerate this regimen for more than a couple of months before he developed an intense aversion to cod-liver oil. Thus, clinical applicability of such secondary risk factor modification still has a long way to go. The topic under discussion is real, widespread, and potentially quite amenable to therapy. The specimen in this slide may appear to be an atherosclerotic abdominal aortic aneurysm, but this is a gross photograph of a 7-year-old coronary artery bypass graft harvested from a patient with type II hyperlipoproteinemia. Gross atherosclerosis with secondary
Intimal hyperplasia in vein grafts
355
ulceration and hemorrhage is readily apparent. This problem is not limited solely to bypass grafts. This slide demonstrates extensive subendothelial myointimal cell proliferation in a native coronary artery many years following cardiac transplantation. This process of accelerated atherogenesis is probably mediated by platelet release and degranulation; liberation of platelet alpha granule constituents (including plateletderived growth factor) is instrumental in this characteristic "phenotypic modulation" of smooth muscle cells from their normal contractile state to a synthetic state. Microscopically, this is represented in this slide by an extensive layer of myointimal cells and some fibrosis between the internal elastic lamella and the endothelial surface. This same pathological process may also be implicated in other phenomena that, unfortunately, we who deal with congenital heart disease see all too frequently. This slide represents exuberant neointimal proliferation inside a woven Dacron valved conduit, which had become severely stenotic 5 years after implantation in a child with dextro-transposition, ventricular septal defect, and left ventricular outflow tract obstruction. I have several questions for Dr. Landymore. First, I would like to inject a word of caution, because he has implied that increased early (I month and I year) coronary bypass graft patency documented in the Mayo Clinic aspirin/dipyridamole randomized trial is due to reduced myointimal cell proliferation. Although none of us can argue with the beautiful design of the Chesebro/Fuster study, this study did not elucidate the actual mechanism of action of aspirin and dipyridamole. In point of fact, such early improvement in graft patency rates might actually be an unexpected side effect of therapy. This is because, as Dr. Landymore has stated, early graft thromboses are more related to technical flaws, low graft flow, and thrombosis than to prolific subendothelial proliferation with resultant stenosis and graft occlusion. Thus, it is plausible that the main potential therapeutic benefits of anti platelet therapy my not really be appreciated until 5 to 10 years have elasped, when graft atherosclerosis becomes manifest in large proportions of patients. Clearly, serial I, 5, and 10 year angiograms in large randomized cohorts (such as the current Veterans Administration Cooperative Study) will be necessary before we have a definitive answer. Why does this terrible tasting cod-liver oil work? I thought it would be informative to refresh your memories regarding the differences between the monoenoic, dienoic, and trienoic prostanoids. In the monoenoic pathway dihydro-garnmalineoleic acid is the substrate, whereas in most of our diets (i.e., sunflower seed oil) lineoleic acid is the principal substrate. With the cold fish diet of Greenland Eskimos and northern Japanese fishermen, the trienoic cascade is prominent, and eicosapentaenoic acid is the primary substrate. For most of us, desaturation and chain elongation of dietary lineoleic acid produces arachidonic acid, which is the most common fatty acid present in cellular phospholipids. Arachidonic acid is converted by the enzyme cyclooxygenase into the cyclic endoperoxides PGG, and PGH,. These intermediates can then be enzymatically metabolized into either prostacyclin (PGI,) or thromboxane (TXA 2 ) - a very potent platelet aggregator
3 5 6 Landymore et at.
and vasoconstrictor. If, as in the case of the Eskimos and fishermen, eicosapentaenoic acid (cod-liver oil) is substituted for the substrate, the resultant end products are PGI 3 and TXA 3• TXA 3 is very benign compared to TXA, in terms of vasoconstrictor and platelet aggregation properties. Thus, the balance is theoretically tipped in favor of the antiaggregatory and vasodilatory effects of the prostacyclins. A word is necessary about the end points in any canine or subhuman primate model of accelerated atherosclerosis. By not concentrating solely on graft patency, Dr. Landymore has not fallen into a common trap. Results of any experimental protocol can be completely discordant depending upon which quantitative end point an investigator uses. All too many papers in the surgical and pathology literature have focused purely on patency rate, which can be quite misleading because it is manifestation of many factors other than plateletmediated myointimal cell proliferation. Meaningful assessment of treatment effects must rely on more sophisticated methodology; as an example, one of the methods we use is similar to that of Dr. Landymore, e.g., digitizing the thickness of the proliferative lesion at a large number of radii in the graft wall. This does, however, exaggerate the magnitude of the lesion if it is quite eccentric. To compensate for this artifact, one can digitize the circumference of the internal and external elastic lamellae and the luminal surface, extrapolate these measurements to a perfect circle, and then derive a computerized percent area of vessel wall thickening or of myointimal cell proliferation. This is probably the most sensitive and specific end point, as shown in this slide which summarizes the results of various antiplatelet regimens in over 200 venoarterial canine allografts. You can readily appreciate the discordance results depending upon which end point is used. By virtue of the fact that Dr. Landymore and his colleagues performed an average of 140 measurements of lesion thickness in each ring section of each graft (a very exhaustive and time-consuming process), their results probably avoid some of the pitfalls inherent in simple one-dimensional measurements of subintimal proliferation. However, this does raise a question concerning the very short (1 ern) grafts used in this study. Perhaps Dr. Landymore could elaborate on this aspect. Second, though we know that the incidence of fatal complications of atherosclerotic coronary and peripheral vascular disease is low in Greenland Eskimos and northern Japanese fishermen, it has been stated that some of these persons die prematurely. Although there was no difference in the coagulation profile between the control and treatment groups in Dr. Landymore's canine study, is it true that some of these premature deaths are due to hemorrhagic problems? Perhaps alcohol abuse and cirrhosis are also a compounding problem here. A third question pertains to what minimum dose of cod-liver oil would be necessary to optimize production of TXA 3 in man? Would not a more practical approach involve substitution of dihydrogamma-linoleic acid for lineoleic acid plus a specific thromboxane synthetase inhibitor? Fourth, I was struck by the marked fibrosis in the subendothelial and medial layers of the grafts in the eicosapentaenoic acid treatment group. Could Dr. Landymore comment on this? Finally, since at least two subpopula-
The Journal of Thoracic and Cardiovascular Surgery
tions of mongrel dogs are known to exist in terms of sensitivity to ex vivo (arachidonic acid) platelet aggregation, could this possibly influence the results of your study considering the small number of dogs in each group? In closing, the dienoic prostanoids (TXA, and PGI,) may not be all bad. If we overwhelm the metabolic systems with eicosapentaenoic acid, the relatively weaker biological activity of PGI 3 might actually outweigh theoretical benefit of TXA J • Alexis Carrel demonstrated over 70 years ago that venoarterial grafts thickened over time, a morphologic adaptive process he termed "arterialization." In recent work in my laboratory, Dr. Vernon Henderson found it most interesting that not only venoarterial autografts, but also venoarterial allografts, recover their capability to produce PGI 2 over time. At 10 to 12 weeks after implantation in the canine model, the luminal surface of both the autografts and the allografts produce normal arterial levels of PG12, which far exceed the levels produced by veins. Thus, both functional (biochemical) and morphologic adaptation occur in vein grafts placed into the arterial system; substituting PGI 3 for PGI, in this context may possibly be deleterious. DR. LANDY MORE (Closing) The animal model that we used in this study was developed in my laboratory primarily to look at intimal hyperplasia in vein grafts used for arterial bypass. Previous studies, investigating the effects of antiplatelet drug therapy on intimal hyperplasia, have not separated vein graft patency from the effects of intimal hyperplasia, and I believe this is a mistake. If you consider the patency of the vein bypass grafts, the patency rate reflects not only the effects of intimal hyperplasia but also a multitude of technical factors, including graft angulation and size of the vein graft in relationship to the host artery. Pathological studies indicate that long-term occlusion is related to intimal hyperplasia and that the intima becomes thickened, injured, and may take up cholesterol; these processes ultimately result in vein graft occlusion. We developed this animal model using interposed segments of vein graft in the femoral position to avoid all the technical problems associated with end-to-side anastomoses, Our model results in a 100%graft patency and was designed to investigate intimal hyperplasia and to avoid stimulation of intimal proliferation by other factors, such as angulation of the vein grafts. The model depends upon cholesterol induction of platelet membranes. It has been shown over the past 5 years that one may induce platelets with cholesterol. A high cholesterol diet results in the increased production of thromboxane A" which significantly increases platelet aggregation. Cholesterol induction enables the investigator to look at the effects of antiplatelet drug regimens on intimal hyperplasia in vein grafts used for bypass over the short term and avoids the necessity for keeping animals over a period of 4 to 6 months. We have been using this model to study the effects of anti platelet drug regimens on intimal hyperplasia in vein grafts for over I V2 years. I have analyzed all our previous vein graft cross sections over the past couple of months, not just the 14 animals that we have reported here but over some 50
Volume 89 Number 3 March, 1985
experiments. This represents a considerable number of vein graft cross sections, considering that each animal has two vein grafts. I increased the number of intimal measurements to 350 to 400 measurements per side. The fact that there was no statistical difference between the original measurements and the repeat measurements demonstrates the reproducibility of our results. Dr. Miller has commented that the short segment of vein graft used in this study may have affected the results. We have looked statistically at the site of sampling and have shown that cross sections can be taken anywhere throughout the length of the vein graft and that anyone cross section is representative; thus, I do not believe that the length of the vein graft will have adversely affected our results. Interestingly, the measurements of intimal hyperplasia are similar when one compares the right to the left vein graft, indicating that technical factors have not influenced our results. Dr. Miller wondered whether cod-liver oil would have any adverse side effects. I would like to emphasize that this is only a preliminary investigational study and that we are not recommending the use ~f cod-liver oil for the prevention of intimal hyperplasia in vein grafts used for myocardial revascu-
Intimal hyperplasia in vein grafts
357
larization. Much more work needs to be done in this area before we can conclusivelyconfirm the efficacy of cod-liver oil. Clinical trials in man have been carried out in Europe, and in the majority of these studies the subjects were ingesting a dose of 3.5 gm of eicosapentaenoic acid daily without untoward side effects. We have done some preliminary human studies in volunteers at our center using 1.5 gm per day, and none of these individuals has had any adverse side effects. I am not sure that platelet aggregation, as measured by response to adenosine diphosphate or collagen, must be inhibited to prevent intimal hyperplasia. As you have seen from our studies, we did not find any change in anticoagulation parameters. We have been using one of the other unsaturated fatty acids, dihomo-gamma-linoleic acid, and have found that small doses that have no effect on platelet aggregation will inhibit intimal hyperplasia. I would like to reemphasize that this is a preliminary investigation. However, our data are exciting and indicate that prostaglandin inhibition with unsaturated fatty acids may be of use in the future for preventing intimal hyperplasia in vein grafts used for arterial bypass.
THoRAc CARDIOVASC SURG
89:351-357, 1985
Cod-liver oil in the prevention of intimal hyperplasia in autogenous vein grafts used for arterial bypass Cod-liver oil, rich in eicosapentaenoic acid, an unsaturated fatty acid, was administered to 14 mongrel dogs to detennine if this acid would prevent platelet-mediated intimal hyperplasia. Twenty-eight 1 cm segments of undistended jugular vein were interposed between bilaterally divided femoral arteries. Seven control animals were fed a 2 % cholesterol diet 1 week before and for 6 weeks after the operation. A further seven animals received cod-liver oil capsules containing 1.8 gm of eicosapentaenoic acid daily 1 week before and for 6 weeks after autogenous vein implantation, in addition to the lipid-supplemented diet Baseline serum cholesterol was 4.6 ± 0.4 mmol/L The rise in serum cholesterol was similar in the two groups and increased to 7.4 ± 0.6 mmoljL (control group) and to 6.8 ± 0.2 mmol/L (eicosapentaenoic acid group) (p < 0.001). Prothrombin time, partial thromboplastin time, bleeding time, and platelet counts were unchanged in the two groups. Vein grafts, harvested at 6 weeks, were fixed in formaldehyde. Mean intimal thickness was measured from multiple vein graft cross sections with a Zeiss computerized interactive image analyzing system. A mean of 140 ± 11 measurements were computed from each graft Marked intimal hyperplasia occurred in the control group and increased from 4.3 ± 0.3 to 86.4 ± 14 ILm. In contrast, a high eicosapentaenoic acid diet inhibited intimal hyperplasia, with intimal thickness only increasing from 4.0 ± 0.4 to 24.8 ± 2.7 ILm (p < 0.001). These data indicate that eicosapentaenoic acid inhibits platelet-mediated intimal hyperplasia and suggest that cod-liver oil could be used to prevent intimal hyperplasia in vein grafts used for myocardial revascularization.
R. W. Landymore, M.D., F.R.C.S.(C), C. E. Kinley, M.D., J. H. Cooper, M.D., M. MacAulay, M.D., B. Sheridan, M.D., and C. Cameron, B.N.,
Halifax, Nova Scotia, Canada
Erous intimal hyperplasia is the most common cause of late graft failure after aorta-coronary bypass.!:' Considerable experimental evidence suggests that intimal thickening is platelet mediated.' This concept is supported by recent clinical trials that have shown improved vein graft patency with antiplatelet drug regimens after myocardial revascularization.v" Clinical studies have shown that increased dietary From Dalhousie University, Department of Surgery, Division of Thoracic and Cardiovascular Surgery, Halifax, Nova Scotia, Canada. Supported by a Canadian Heart Foundation Grant. Read at the Tenth Annual Meeting of the Western Thoracic Surgical Association, Maui, Hawaii, June 20-23, 1984. Address for reprints: R. W. Landymore, M.D., F.R.C.S.(C), Room 3065, R. C. Dickson Centre, Victoria General Hospital, Halifax, Nova Scotia, Canada B3H 2Y9.
intake of an unsaturated fatty acid, eicosapentaenoic acid, results in decreased platelet aggregation and total platelet counts." Eicosapentaenoic acid is found in high concentrations in salt water fish and synthetically refined cod-liver oil.'? 11 This study was initiated to determine the effects of cod-liver oil on plateletmediated intimal hyperplasia in autogenous vein grafts used for arterial bypass. Materials and methods Segments of autogenous external jugular vein were interposed between bilaterally divided femoral arteries in 14 adult mongrel dogs, weighing between 25 and 30 kg. The animals were divided into two equal groups, both receiving a 2% cholesterol diet* 1 week before and for 6 weeks after operation. Seven animals were used as "I.C.N. Nutritional Biochemicals, Cleveland, Ohio.
351
The Journal of Thoracic and Cardiovascular Surgery
3 5 2 Landymore et al.
Fig. 1. Vein graft cross sections (original magnification X250). A, Vein graft before implantation (hematoxylineosin stain). B, Vein graft from a control animal (Masson stain). C, Marked reduction in intimal thickness that occurs with eicosapentaenoic acid (Masson stain).
controls and seven other animals received 20 minim capsules of cod-liver oil containing 1.8 gm of eicosapentaenoic acid daily, 1 week before and for 6 weeks after implantation of the autogenous vein. The animals were given fentanyl (0.4 mg) and droperidol (20 mg) for preoperative sedation. Anesthesia was induced with sodium pentobarbitol (5 mg/kg), with maintenance doses of pentobarbitol (3 mg/kg) being administered as necessary. The animals were ventilated with the Bird Mark 7 pressure-regulated ventilator. A 3 ern segment of autogenous external jugular vein was harvested from the neck, and the common femoral arteries were exposed bilaterally by sterile technique. A segment of jugular vein was immediately fixed in formaldehyde for later examination under the light microscope. After heparin was given systemically in a dose of 2 mg/kg, 1 em segments of undistended external jugular vein were interposed between bilaterally divided common femoral arteries. The anastomoses were completed with running 6-0 Prolene suture, with the aid of optical magnification, Heparin was then reversed with protamine sulfate and the incisions were closed. Six weeks later the animals were given an overdose of sodium pentobarbitol, and the common femoral arteries, containing the interposed vein grafts, were harvested and fixed in formaldehyde. Cross sections of jugular vein were stained with hematoxylin-eosin and Masson stains, and intimal thickness was measured from multiple cross-sections of the vein grafts and a Zeiss computer-
ized interactive image analyzing system. * This system is composed of a Zeiss microscope, a computer, and a 12 inch television screen. Individual slides are projected on the television screen, and a mobile cursor, connected to the computer, is used to accurately measure intimal thickness in microns around the entire circumference of the graft. Hematocrit value, platelet count, prothrombin time, partial thromboplastin time, bleeding time, and serum cholesterol were determined prior to any dietary manipulation, after institution of the 2% cholesterol diet before the operation and 2 weeks and 4 weeks after implantation of autogenous vein grafts. Results have been reported as the arithmetic mean with standard error and the Student's t test was used for statistical comparison. Results An overall increase in total body weight of 3.5 ± 0.25 kg was observed in both groups while receiving the 2% cholesterol diet. Baseline serum cholesterol was 4.6 ± 0.4 mmoljL. Lipid supplementation resulted in a significant rise in serum cholesterol levels that was similar in the two groups. Serum cholesterol levels increased to 7.4 ± 0.6 mmol/L in the control animals and to 6.8 ± 0.2 mmoljL in those animals receiving eicosapentaenoic acid (p < 0.001). Hematocrit value, partial thromboplastin time, prothrombin time, bleeding time, and 'Karl Zeiss Canada Limited, Don Mills, Ontario, Canada.
Volume 89 Number 3
Intimal hyperplasia in vein grafts
March, 1985
353
Arachidonic Acid Eicosapentaenoic Acid
I
Cyclo-oxygenase
Cyclic Endoperoxides PGG z PGH z
Prostacyclin PGI 3
/\
Prostacyclin PGl z (Arterial Wall)
Thromboxane Aa TXA 3
Thromboxane Az TXA z (Platelet Membrane)
Fig. 2. Pathways for prostaglandin biosynthesis. Arachidonic acid is normally metabolized to proaggregatory thromboxane A 2 by platelet membranes, whereas blood vessel walls metabolize arachidonic acid to antiaggregatory prostacyclin 12, Eicosapentaenoic acid competes with arachidonic acid for cyclooxygenase, resulting in decreased production of thromboxane A 2• In addition, eicosapentaenoic acid may be metabolized to the weakly aggregatory prostanoid thromboxane A] by platelet membranes and to antiaggregatory prostacyclin I, by blood vessels.
platelet counts did not change significantly in either group during the study. Light microscopic examination of representative vein graft cross sections prior to implantation indicated that the intima was intact and had not been damaged during harvesting. Intimal thickness was measured from multiple vein graft cross sections with the Zeiss computerized interactive image analyzing system. A mean of 140 ± 11 measurements were computed from each vein graft cross section. Marked intimal hyperplasia occurred in the control animals receiving lipid supplementation and increased from 4.3 ± 0.3 to 86.4 ± 1.4 ~m. In contrast, eicosapentaenoic acid markedly reduced intimal hyperplasia, with intimal thickness only increasing from 4.0 ± 0.4 to 24.8 ± 2.7 ~m (p < 0.001). Photomicrographs of representative vein cross sections are shown in Fig. 1. Fig. 1, A is a vein cross section prior to implantation that has been stained with hematoxylineosin. The intima in this cross section is only one cell thick. Fig. 1, B shows marked intimal hyperplasia in a representative section from a control animal. In contrast to D, C illustrates the marked reduction in intimal hyperplasia that occurs in animals receiving eicosapentaenoic acid. Discussion Myocardial revascularization usually results in the immediate and dramatic relief of angina, with optimum results being directly related to vein graft patency." Although 85% of vein grafts are open at 1 year,I3 15 the 0
long-term patency rate is disappointing, with over 50% of vein grafts occluding within 10 years after myocardial revascularization." Early graft failure is usually related primarily to thrombosis, whereas late graft failure is secondary to progressive fibrous intimal hyperplasia-'' Experimental investigations have shown that intimal hyperplasia is platelet mediated and that platelets release a mitogenic serum factor that stimulates intimal proliferation.t" Metke and colleagues IS have added further experimental evidence to support this concept and have demonstrated that antiplatelet drug therapy with a combination of aspirin-dipyridamole reduces vein graft intimal hyperplasia in dogs undergoing aorta-left anterior descending coronary bypass. Subsequent clinical trials have demonstrated that this medical regimen decreases fibrous intimal hyperplasia and improves long-term graft patency after myocardial revascularization.' Although this early work is encouraging, optimum antiplatelet therapy has yet to be described. Epidemiologic studies indicate that myocardial infarction is uncommon in the Greenland Eskimo population."?' The low incidence of atherosclerotic heart disease appears to be related to a high dietary intake of fish oil, which is rich in the unsaturated fatty acid eicosapentaenoic acid." Clinical trials have demonstrated that consumption of cod-liver oil, rich in eicosapentaenoic acid, not only reduces total platelet counts but markedly interferes with platelet aggregation.>" Arachidonic acid is the precursor for prostaglandin biosynthesis in platelet membranes and arterial blood
The Journal of
354
Landymore et al.
vessels." Platelet membranes enzymatically convert arachidonic acid to proaggregatory thromboxane A 2 (Fig. 2), whereas blood vessels convert arachidonic acid to antiaggregatory prostacyclin (PGI 2) . 26 Eicosapentaenoic acid competes with arachidonic acid for the enzyme cyclooxygenase (Fig. 2), which results in decreased production of proaggregatory thromboxane A 2 by platelet membranes. Recent reports also indicate that eicosapentaenoic acid is metabolized to thromboxane A J , a weakly aggregatory prostanoid, whereas some of the intermediary metabolites of eicosapentaenoic acid may be metabolized to PGI J by arterial blood vesselS. 8- 11, 24, 27 Regardless of the actual mechanism of action of eicosapentaenoic acid, substitution of this unsaturated fatty acid for arachidonic acid in platelet membranes results in decreased thromboxane A b which causes a marked reduction in platelet aggregation. The known effects of eicosapentaenoic acid on platelet metabolism prompted the present study to determine whether eicosapentaenoic acid would inhibit platelet-mediated intimal hyperplasia in autogenous vein grafts for arterial bypass. Previously we" developed and reported upon a canine model that may be used to produce accelerated shortterm intimal hyperplasia in vein grafts used for arterial bypass. This model was used in this study to determine the effects of eicosapentaenoic acid on vein graft intimal hyperplasia. Interestingly, our data clearly demonstrate that eicosapentaenoic acid .inhibits platelet-mediated intimal hyperplasia. Intimal thickening was reduced by 350% in animals receiving cod-liver oil. The use of cod-liver oil to prevent intimal thickening after autogenous vein implantation is a particularly attractive concept because eicosapentaenoic acid has been ingested by the Eskimo population for over 100 years without untoward side effects. Although this study is only a preliminary investigation, our data suggest that cod-liver oil might be of use in the prevention of intimal hyperplasia in vein grafts used for myocardial revascularization. REFERENCES Chesebro J, Fuster V: Drug trials in prevention of occlusion of aorta-coronary artery vein grafts. J THoRAc CARDIOVASC SURG 83:90, 1982 2 Unni K, Kottke B, Titus J, Frye R, Wallace R, Brown A: Pathologic changes in aortocoronary saphenous vein graft. Am J Cardiol 34:526, 1974 3 Vlodaver Z, Edwards J: Pathologic changes in aorticcoronary arterial saphenous vein grafts, Circulation 64:719, 1971 4 Friedman R, Burns E: Role of platelets in the proliferative response of the injured artery. Prog Hemost Thromb 4:249, 1978
Thoracic and Cardiovascular Surgery
5 Chesebro J, Fuster V, Elveback L, Clements I, Smith H, Holmes 0, Bardsley W, Pluth J, Wallace R, Puga F, Orszulak T, Piehler J, Danielson G, Schaff H, Frye R: Effect of dipyridamole and aspirin on late vein graft patency after coronary bypass operations, N Engl J Med 310:209, 1984 6 Mayer JE, Maj MC, Lindsay WG, Castaneda A, Nicoloff OM: Influence of aspirin and dipyridamole on patency of coronary artery bypass grafts, Ann Thorac Surg 31:204, 1981 7 Siess W, Scherer B, Bohlig B, Roth P, Kurzmann I, Weber PC: Platelet-membrane fatty acids, platelet aggregation, and thromboxane formation during a mackerel diet. Lancet 1:441, 1980 8 Dyerberg, J" Bang, H.O" Stoffersen, E, Eicosapentaenoic acid and prevention of thrombosis and atherosclerosis? Lancet 2:117, 1978 9 Hay C, Durber A, Saynor R: Effect of fish oil on platelet kinetics in patients with ischemic heart disease, Lancet 1:1269,1982 10 Sanders T, Vickers M, Haines A: Effect on blood lipids and haemostasis of a supplement of cod-liver oil, rich in eicosapentaenoic and docosahexaenoic acids, in healthy young men. Clin Sci 61:317,1981 II Hirai A, Terano T, Hamazaki T, Sajiki J, Kondo S, Ozawa A, Fujita T, Miyamoto T, Tamura Y, Kumagai A: The effects of the oral administration of fish oil concentrate on the release and the metabolism of arachidonic acid and eicosapentaenoic acid by human platelets, Thromb Res 28:285, 1982 12 Cukingnan R, Carey J, Wittig J, Brown G: Influence of complete coronary revascularization on relief of angina, J THoRAc CARDIOYASC SURG 79: 188, 1980 13 Geha A, Baue A: Early and late results of coronary revascularization with saphenous vein and internal mammary artery grafts. Am J Surg 137:456, 1979 14 Little B, Loop F, Thurer R, Groves L, Taylor P, Cosgrove 0: Isolated left anterior descending coronary atherosclerosis, Long-term comparison of internal mammary artery and venous autografts, Circulation 61:869, 1980 15 Tyras 0, Barner H, Kaiser G, Codd J, Pennington 0, Willman V: Bypass grafts to the left anterior descending coronary artery, J THoRAc CARDIOVASC SURG 80:327, 1980 16 Campeau L, Enjalbert M, Lesperance J, Bourassa MG: Arteriosclerosis and late closure of aortocoronary saphenous vein grafts, Sequential angiographic studies I year, 5-7 years, and 1O-12 years after surgery (abstr). Circulation 66:Suppl 2:11-94, 1982 17 Ross R, Glomset J, Kariya B, Harker L: A plateletdependent serum factor that stimulates the proliferation of arterial smooth muscle cells in vitro, Proc Nat! Acad Sci USA 71:1207,1974 18 Metke M, Lie J, Fuster V, Jova M, Kaye M: Reduction of intimal thickening in canine coronary bypass vein grafts with dipyridamole and aspirin, Am J Cardiol 43:1144, 1979
Volume 89 Number 3 March, 1985
19 Harris WS, Connor WE, Goodnight SH Jr: Dietary fish oils, plasma lipids and platelets in man. Prog Lipid Res 20:75,1981 20 Dyerberg J, Bang HO: Atherogenesis and haemostasis in Eskimos. The role of the prostaglandin-S family. Haernostasis 8:227, 1979 21 Bang HO, Dyerberg J, Hjorne N: The composition of food consumed by Greenland Eskimos. Acta Med Scand 200:69, 1976 22 Dyerberg J, Bang HO, Stoffersen E, Moncada S, Vane JR: Eicosapentaenoic acid and prevention of thrombosis and atherosclerosis. Lancet 2: 117, 1978 23 Hay C, Durber A, Saynor R: Effect of fish oil on platelet kinetics in patients with ischemic heart disease. Lancet 1:1269,1982 . 24 Brox J, Killie J, Osterud B, Holme S, Nordoy A: Effects of cod-liver oil on platelets and coagulation in familial hypercholesterolemia (Type IIA). Acta Med Scand 213:137,1983 25 McKean M, Smith J, Silver M, Ahern D: Metabolism of radioactive 5, 8, II, 14, 17-eicosapentaenoic acid by human platelets. Prog Lipid Res 20:435, 1981 26 Willis A, Smith J: Some perspectives on platelets and prostaglandins. Prog Lipid Res 20:387, 1981 27 Fischer S, Weber P: Thromboxane A] (TXA 3) is formed in human platelets after dietary eicosapentaenoic acid. Biochem Biophys Res Comm 116:1091, 1983 28 Landymore R, Kinley C, Cameron C: Intimal hyperplasia in autogenous vein grafts used for arterial bypass. A canine model. Cardiovasc Res (in press) 29 Brox J, Killie J, Gunnes S, Nordoy A: The effect of cod-liver oil and corn oil on platelets and vessel wall in man. Thromb Haemost 46:604, 1981
Discussion DR. D. CRAIG MILLER Stanford, Calif
Having been working on various models of experimental atherogenesis in the laboratory for 7 or 8 years and being aware of the differential biological functions of the dienoic and trienoic prostanoids, I was intrigued to see this paper on the program. I felt that it would be unfair to comment on whether cod-liver oil is good or bad without having tried it myself, but I was spared this fate when I learned that my colleague Dr. Scott Mitchell had previously attempted to augment his intake of eicosapentaenoic acid. Dr. Mitchell, however, could not tolerate this regimen for more than a couple of months before he developed an intense aversion to cod-liver oil. Thus, clinical applicability of such secondary risk factor modification still has a long way to go. The topic under discussion is real, widespread, and potentially quite amenable to therapy. The specimen in this slide may appear to be an atherosclerotic abdominal aortic aneurysm, but this is a gross photograph of a 7-year-old coronary artery bypass graft harvested from a patient with type II hyperlipoproteinemia. Gross atherosclerosis with secondary
Intimal hyperplasia in vein grafts
355
ulceration and hemorrhage is readily apparent. This problem is not limited solely to bypass grafts. This slide demonstrates extensive subendothelial myointimal cell proliferation in a native coronary artery many years following cardiac transplantation. This process of accelerated atherogenesis is probably mediated by platelet release and degranulation; liberation of platelet alpha granule constituents (including plateletderived growth factor) is instrumental in this characteristic "phenotypic modulation" of smooth muscle cells from their normal contractile state to a synthetic state. Microscopically, this is represented in this slide by an extensive layer of myointimal cells and some fibrosis between the internal elastic lamella and the endothelial surface. This same pathological process may also be implicated in other phenomena that, unfortunately, we who deal with congenital heart disease see all too frequently. This slide represents exuberant neointimal proliferation inside a woven Dacron valved conduit, which had become severely stenotic 5 years after implantation in a child with dextro-transposition, ventricular septal defect, and left ventricular outflow tract obstruction. I have several questions for Dr. Landymore. First, I would like to inject a word of caution, because he has implied that increased early (I month and I year) coronary bypass graft patency documented in the Mayo Clinic aspirin/dipyridamole randomized trial is due to reduced myointimal cell proliferation. Although none of us can argue with the beautiful design of the Chesebro/Fuster study, this study did not elucidate the actual mechanism of action of aspirin and dipyridamole. In point of fact, such early improvement in graft patency rates might actually be an unexpected side effect of therapy. This is because, as Dr. Landymore has stated, early graft thromboses are more related to technical flaws, low graft flow, and thrombosis than to prolific subendothelial proliferation with resultant stenosis and graft occlusion. Thus, it is plausible that the main potential therapeutic benefits of anti platelet therapy my not really be appreciated until 5 to 10 years have elasped, when graft atherosclerosis becomes manifest in large proportions of patients. Clearly, serial I, 5, and 10 year angiograms in large randomized cohorts (such as the current Veterans Administration Cooperative Study) will be necessary before we have a definitive answer. Why does this terrible tasting cod-liver oil work? I thought it would be informative to refresh your memories regarding the differences between the monoenoic, dienoic, and trienoic prostanoids. In the monoenoic pathway dihydro-garnmalineoleic acid is the substrate, whereas in most of our diets (i.e., sunflower seed oil) lineoleic acid is the principal substrate. With the cold fish diet of Greenland Eskimos and northern Japanese fishermen, the trienoic cascade is prominent, and eicosapentaenoic acid is the primary substrate. For most of us, desaturation and chain elongation of dietary lineoleic acid produces arachidonic acid, which is the most common fatty acid present in cellular phospholipids. Arachidonic acid is converted by the enzyme cyclooxygenase into the cyclic endoperoxides PGG, and PGH,. These intermediates can then be enzymatically metabolized into either prostacyclin (PGI,) or thromboxane (TXA 2 ) - a very potent platelet aggregator
3 5 6 Landymore et at.
and vasoconstrictor. If, as in the case of the Eskimos and fishermen, eicosapentaenoic acid (cod-liver oil) is substituted for the substrate, the resultant end products are PGI 3 and TXA 3• TXA 3 is very benign compared to TXA, in terms of vasoconstrictor and platelet aggregation properties. Thus, the balance is theoretically tipped in favor of the antiaggregatory and vasodilatory effects of the prostacyclins. A word is necessary about the end points in any canine or subhuman primate model of accelerated atherosclerosis. By not concentrating solely on graft patency, Dr. Landymore has not fallen into a common trap. Results of any experimental protocol can be completely discordant depending upon which quantitative end point an investigator uses. All too many papers in the surgical and pathology literature have focused purely on patency rate, which can be quite misleading because it is manifestation of many factors other than plateletmediated myointimal cell proliferation. Meaningful assessment of treatment effects must rely on more sophisticated methodology; as an example, one of the methods we use is similar to that of Dr. Landymore, e.g., digitizing the thickness of the proliferative lesion at a large number of radii in the graft wall. This does, however, exaggerate the magnitude of the lesion if it is quite eccentric. To compensate for this artifact, one can digitize the circumference of the internal and external elastic lamellae and the luminal surface, extrapolate these measurements to a perfect circle, and then derive a computerized percent area of vessel wall thickening or of myointimal cell proliferation. This is probably the most sensitive and specific end point, as shown in this slide which summarizes the results of various antiplatelet regimens in over 200 venoarterial canine allografts. You can readily appreciate the discordance results depending upon which end point is used. By virtue of the fact that Dr. Landymore and his colleagues performed an average of 140 measurements of lesion thickness in each ring section of each graft (a very exhaustive and time-consuming process), their results probably avoid some of the pitfalls inherent in simple one-dimensional measurements of subintimal proliferation. However, this does raise a question concerning the very short (1 ern) grafts used in this study. Perhaps Dr. Landymore could elaborate on this aspect. Second, though we know that the incidence of fatal complications of atherosclerotic coronary and peripheral vascular disease is low in Greenland Eskimos and northern Japanese fishermen, it has been stated that some of these persons die prematurely. Although there was no difference in the coagulation profile between the control and treatment groups in Dr. Landymore's canine study, is it true that some of these premature deaths are due to hemorrhagic problems? Perhaps alcohol abuse and cirrhosis are also a compounding problem here. A third question pertains to what minimum dose of cod-liver oil would be necessary to optimize production of TXA 3 in man? Would not a more practical approach involve substitution of dihydrogamma-linoleic acid for lineoleic acid plus a specific thromboxane synthetase inhibitor? Fourth, I was struck by the marked fibrosis in the subendothelial and medial layers of the grafts in the eicosapentaenoic acid treatment group. Could Dr. Landymore comment on this? Finally, since at least two subpopula-
The Journal of Thoracic and Cardiovascular Surgery
tions of mongrel dogs are known to exist in terms of sensitivity to ex vivo (arachidonic acid) platelet aggregation, could this possibly influence the results of your study considering the small number of dogs in each group? In closing, the dienoic prostanoids (TXA, and PGI,) may not be all bad. If we overwhelm the metabolic systems with eicosapentaenoic acid, the relatively weaker biological activity of PGI 3 might actually outweigh theoretical benefit of TXA J • Alexis Carrel demonstrated over 70 years ago that venoarterial grafts thickened over time, a morphologic adaptive process he termed "arterialization." In recent work in my laboratory, Dr. Vernon Henderson found it most interesting that not only venoarterial autografts, but also venoarterial allografts, recover their capability to produce PGI 2 over time. At 10 to 12 weeks after implantation in the canine model, the luminal surface of both the autografts and the allografts produce normal arterial levels of PG12, which far exceed the levels produced by veins. Thus, both functional (biochemical) and morphologic adaptation occur in vein grafts placed into the arterial system; substituting PGI 3 for PGI, in this context may possibly be deleterious. DR. LANDY MORE (Closing) The animal model that we used in this study was developed in my laboratory primarily to look at intimal hyperplasia in vein grafts used for arterial bypass. Previous studies, investigating the effects of antiplatelet drug therapy on intimal hyperplasia, have not separated vein graft patency from the effects of intimal hyperplasia, and I believe this is a mistake. If you consider the patency of the vein bypass grafts, the patency rate reflects not only the effects of intimal hyperplasia but also a multitude of technical factors, including graft angulation and size of the vein graft in relationship to the host artery. Pathological studies indicate that long-term occlusion is related to intimal hyperplasia and that the intima becomes thickened, injured, and may take up cholesterol; these processes ultimately result in vein graft occlusion. We developed this animal model using interposed segments of vein graft in the femoral position to avoid all the technical problems associated with end-to-side anastomoses, Our model results in a 100%graft patency and was designed to investigate intimal hyperplasia and to avoid stimulation of intimal proliferation by other factors, such as angulation of the vein grafts. The model depends upon cholesterol induction of platelet membranes. It has been shown over the past 5 years that one may induce platelets with cholesterol. A high cholesterol diet results in the increased production of thromboxane A" which significantly increases platelet aggregation. Cholesterol induction enables the investigator to look at the effects of antiplatelet drug regimens on intimal hyperplasia in vein grafts used for bypass over the short term and avoids the necessity for keeping animals over a period of 4 to 6 months. We have been using this model to study the effects of anti platelet drug regimens on intimal hyperplasia in vein grafts for over I V2 years. I have analyzed all our previous vein graft cross sections over the past couple of months, not just the 14 animals that we have reported here but over some 50
Volume 89 Number 3 March, 1985
experiments. This represents a considerable number of vein graft cross sections, considering that each animal has two vein grafts. I increased the number of intimal measurements to 350 to 400 measurements per side. The fact that there was no statistical difference between the original measurements and the repeat measurements demonstrates the reproducibility of our results. Dr. Miller has commented that the short segment of vein graft used in this study may have affected the results. We have looked statistically at the site of sampling and have shown that cross sections can be taken anywhere throughout the length of the vein graft and that anyone cross section is representative; thus, I do not believe that the length of the vein graft will have adversely affected our results. Interestingly, the measurements of intimal hyperplasia are similar when one compares the right to the left vein graft, indicating that technical factors have not influenced our results. Dr. Miller wondered whether cod-liver oil would have any adverse side effects. I would like to emphasize that this is only a preliminary investigational study and that we are not recommending the use ~f cod-liver oil for the prevention of intimal hyperplasia in vein grafts used for myocardial revascu-
Intimal hyperplasia in vein grafts
357
larization. Much more work needs to be done in this area before we can conclusivelyconfirm the efficacy of cod-liver oil. Clinical trials in man have been carried out in Europe, and in the majority of these studies the subjects were ingesting a dose of 3.5 gm of eicosapentaenoic acid daily without untoward side effects. We have done some preliminary human studies in volunteers at our center using 1.5 gm per day, and none of these individuals has had any adverse side effects. I am not sure that platelet aggregation, as measured by response to adenosine diphosphate or collagen, must be inhibited to prevent intimal hyperplasia. As you have seen from our studies, we did not find any change in anticoagulation parameters. We have been using one of the other unsaturated fatty acids, dihomo-gamma-linoleic acid, and have found that small doses that have no effect on platelet aggregation will inhibit intimal hyperplasia. I would like to reemphasize that this is a preliminary investigation. However, our data are exciting and indicate that prostaglandin inhibition with unsaturated fatty acids may be of use in the future for preventing intimal hyperplasia in vein grafts used for arterial bypass.