Improved Preservation of Saphenous Vein Grafts by the Use of Glyceryl Trinitrate–Verapamil Solution During Harvesting
Background High-pressure distension during harvesting damages the saphenous vein (SV) and may contribute to subsequent coronary artery bypass graft (CABG) occlusion. Application of vasodilator agents to the SV during harvesting may reduce the need for high-pressure distension and improve graft quality. We tested the effects of a vasodilator solution containing glyceryl trinitrate and verapamil (GV) or the conventional agent papaverine (Pap) on the pressure necessary to overcome SV spasm and on the structure and biochemistry of the SV graft.
Methods and Results Thirty-six patients undergoing CABG were randomly allocated to receive an application of either topical and intraluminal GV solution, topical Pap, or topical and intraluminal Ringer’s solution (untreated) to the SV during harvesting. The peak and mean pressures required to distend the vein were recorded. Samples of SV were taken for microscopy and biochemical analysis just before we performed the anastomosis. The percentage of endothelial coverage was calculated by area measurements of stained en face preparations of the vein intima. The results for peak pressures (mm Hg) were: untreated, 479.2±27.5; Pap, 384.8±29.0; and GV, 309.5±28.3 (P<.001, GV plus Pap versus untreated); and the results for mean pressures (mm Hg) were untreated, 136.2±9.6; Pap, 102.2±10.8; and GV, 98.0±8.3 (P<.01, GV plus Pap versus untreated). The results for endothelial cover (%) were: untreated, 43.7±7.0; Pap, 44.1±9.2; and GV, 68.7±7.0 (P<.05, GV versus Pap); and the results for ATP (nmol/g wet wt) were: untreated, 67.3±12.7; Pap, 112.0±19.4; and GV, 132.5±22.7 (P<.05, GV plus Pap versus untreated).
Conclusions First, pharmacological treatment of SV during harvesting, especially with GV solution, allows the use of a lower distension pressure and reduces the breakdown of high-energy phosphates in the vein wall. Second, topical and intraluminal use of GV solution during vein harvesting improves endothelial coverage compared with the topical use of Pap or no pharmacological treatment.
Despite the widespread use and superior patency of the internal mammary artery and other arterial conduits, the saphenous vein continues to be the most commonly used conduit for coronary artery bypass graft surgery. However, the long-term patency of saphenous vein grafts is low. The occlusion rate in the first year is 15% to 26%.1 2 By 10 to 12 years, approximately 50% of saphenous vein grafts are occluded,3 4 and of those still patent, 50% show marked atherosclerotic changes.5
Morphological examination of venous grafts from reoperation or autopsy suggests three different modes of occlusion.1 In the first month after operation, occlusion is usually due to thrombosis related to faulty surgical technique. Between 1 month and 1 year, graft occlusion is due to intimal proliferation and medial fibrosis. Beyond 1 year, atherosclerosis is the most common cause of graft occlusion.
There is abundant experimental evidence that the use of high-pressure distension to overcome venospasm during harvesting causes endothelial cell loss and medial damage.6 7 Endothelial cell loss denudes the surface of the intima, causing deposition of platelets and fibrin, which predisposes to enhanced lipid uptake by the vein wall,8 increased intimal hyperplasia,9 10 11 and reduced patency, at least in the short term.12 Pharmacological relaxation of the saphenous vein has been recommended to reduce spasm and obviate the need for high-pressure distension.13 14 Papaverine is the most commonly used vasodilator. However, there is evidence that papaverine used intraluminally may damage the endothelium due to its acidic pH15 and may reduce prostacyclin production in the vein wall. Furthermore, if papaverine enters the systemic circulation, it may cause hypotension.
We developed, on the basis of organ bath testing, a new vasodilator solution that is suitable for intraluminal use; it is composed of glyceryl trinitrate (GTN) and verapamil (GV) in optimal concentrations for venodilation made up in neutral Ringer’s solution.16 We also developed a simple technique of applying the solution topically and intraluminally to the saphenous vein during harvesting.17 In our experience of more than 5 years of clinical use, this technique appeared to provide a relaxed venous conduit for grafting. However, we had not quantified the degree of relaxation or shown that the quality of the graft conduits was improved. We therefore set out to study the effects of GV solution and papaverine on the distension pressure required to prepare the saphenous vein for grafting, endothelial morphology and coverage, and the energy status of the vein wall.
Thirty-six consenting patients who were undergoing coronary bypass graft surgery with or without valve replacement were randomized to receive GV solution, papaverine solution, or Ringer’s solution (no pharmacological treatment) during harvesting of the long saphenous vein. The protocol was approved by the Alfred Hospital Human Ethics Committee under the guidelines of the National Health and Medical Research Council of Australia.
The saphenous vein was removed from the leg by atraumatic surgical technique. The pressure used during distension and testing of the vein was measured with a pressure transducer connected to a monitor and chart recorder. The distending pressure used was the minimum judged necessary by the surgeon in each instance to overcome spasm and to distend the vein adequately for grafting. During harvesting and preparation of the vein, the surgeon was unaware of the distension pressure measurements and the nature of the test solution. Pressure was recorded continuously while the vein was harvested, flushed, and distended. The peak pressure was read from the pressure tracing, and the mean pressure was calculated from the area under the pressure curve. After harvesting, all veins were stored in autologous heparinized blood. In patients receiving GV treatment, the blood also contained 1 mg GTN/100 mL (0.1 mmol/L).
All solutions were prepared in the operating room using room temperature Ringer’s solution containing 1.6 U/mL heparin. GV solution was made up by adding verapamil hydrochloride, GTN, sodium bicarbonate, and heparin to 300 mL of Ringer’s solution (Table 1⇓) to provide concentrations shown in the organ bath to be optimal for relaxation of human saphenous vein rings.16 Papaverine solution was made by adding 120 mg papaverine hydrochloride to 100 mL Ringer’s solution (1.2 mg/mL).15
The three solutions were used in a blinded fashion so that both the principal surgeons and the four harvesting surgeons were unaware of the identity of the solution in use.
The saphenous vein was exposed via a short incision at the ankle, and a metal cannula was inserted for injection of the dilator solution. A 50-mL syringe containing GV solution was attached to the cannula with the use of a short extension tube.16 17 From 1 to 2 mL of solution was then injected into the vein and also applied topically as a spray onto its surface with the use of a separate syringe. As the vein was progressively exposed, GV solution was again injected into the lumen and sprayed onto the surface of the vein. Thus, the vein was exposed both internally and externally to the vasodilators in optimal concentrations. After removal from the leg, the vein was distended with GV solution and then stored at room temperature in heparinized blood containing 0.1 mmol/L GTN until grafting.
The cannulation and injection technique was the same as that used for GV except that papaverine hydrochloride solution (1.2 mg/mL) was sprayed over the exterior of the vein (topical application) and unmodified Ringer’s solution was used intraluminally during harvesting and distension. The prepared vein was stored in heparinized blood.
No Pharmacological Treatment
The cannulation and injection technique was the same as that used for the vasodilator techniques. During harvesting, the surface of the vein was sprayed with heparinized Ringer’s solution. Ringer’s solution was also injected intraluminally and was used for distension. The prepared vein was stored in heparinized blood.
To preserve the blinded design, for each technique the surgeon used solutions from two bowls—one for intraluminal use and the other for external use (the surgeon had no knowledge of the identity of either of the solutions).
Two samples of vein were obtained for morphological and biochemical analyses. The first sample (control sample) was obtained immediately after the vein was exposed. The second sample (prepared sample) was obtained from the proximal end of the vein when it was trimmed to length before construction of the aortic anastomosis.
All samples of vein collected during the operation were stored in phosphate buffer. Immediately after the second sample was obtained, the samples were taken to the laboratory, where they were cut longitudinally and pinned onto a nylon tile as an en face preparation of the intimal surface. These preparations were fixed by immersion in 1.2% glutaraldehyde in phosphate buffer with an osmolarity of 370 mosmol/L, which is optimal for preservation of vessel architecture.18 After fixation for at least 4 hours, the preparations were washed and irrigated with toluidine blue. This stain colors endothelial cell nuclei pale blue and denuded areas purple.19 The preparations were then viewed under a stereomicroscope. The microscopist was blinded to the treatments applied to the veins. Endothelial cell coverage was quantified with the use of an eyepiece graticule to measure the area covered by endothelium. Endothelial coverage was expressed as a percentage of the total area of the specimen.
To confirm the light microscopy findings, the same samples that had shown typical appearances under the stereomicroscope were also prepared for examination under the scanning electron microscope. These samples were postfixed in 2% osmium tetroxide, rapidly dehydrated in a series of ethanols, critical-point–dried, sputter-coated, and then viewed with a Cambridge S250 scanning electron microscope.
In the operating room, samples for biochemical analysis were collected, blotted to remove blood, wrapped in foil, and immediately frozen in liquid nitrogen. Subsequently, the frozen samples were ground to a powder and extracted with perchloric acid. The supernatant was collected and analyzed with high-performance liquid chromatography20 for levels of ATP, ADP, AMP, hypoxanthine, inosine, and nicotinamide adenine dinucleotide. The total adenine nucleotide level, ATP-to-ADP ratio, and energy charge [(ATP+0.5 ADP)/(ATP+ADP+AMP)] were calculated.
One-way ANOVA with orthogonal comparisons21 was used for statistical analysis of the distension pressures, endothelial coverage, and energy metabolite levels. The maximum number of valid individual comparisons for this method is one less than the number of groups being compared. As there were three treatment groups, two individual comparisons were permitted. The comparisons were kept uniform throughout the three sets of results. The first comparison tested for a difference between the combined pharmacological treatment groups (GV and papaverine) and the untreated group. The second comparison tested for a difference between the GV and papaverine treatment groups. For the biochemical results, the preharvesting control values were also included in the analysis. Here, a third comparison was made, between the control and the average value of all the postharvesting (prepared sample) levels.
To test for nonrandom distribution of age and sex among the three treatment groups and the four harvesting surgeons, χ2 analysis was performed.
All values are given as mean±SEM. A probability value of <.05 was considered statistically significant.
Patient characteristics are given in Table 2⇓. Age and sex distributions in the three groups were similar (P=.201, χ2 analysis), as was the number of grafts per patient (P=.5). There were no significant differences in the distribution of the three treatments among the four harvesting surgeons (χ2=6.21, P=.4).
Satisfactory measurements of distension pressure were obtained in 32 patients. The average peak pressure reached during surgical preparation in the group receiving no pharmacological treatment (untreated, n=11), was 479.2±27.5 mm Hg (range, 318 to 560 mm Hg) (Fig 1⇓). In the papaverine group (n=11), the peak pressure was 384.8±29.0 mm Hg (range, 285 to 530 mm Hg). In the GV group (n=10), the peak pressure was 309.5±28.3 mm Hg (range, 220 to 450 mm Hg). The combined pharmacological treatment groups had a significantly lower average distension pressure than the untreated group (P<.001). The GV group had the lowest distension pressure, but the difference between this group and the papaverine group failed to reach statistical significance (P=.066).
The average mean pressure in the untreated group was 136.2±9.6 mm Hg (range, 76 to 190 mm Hg). In the papaverine group, the average mean pressure was 102.2±10.8 mm Hg (range, 45 to 173 mm Hg). In the GV group, the average mean pressure was 98.0±8.3 mm Hg (range, 65 to 158 mm Hg). The combined pharmacological treatment groups had a significantly lower average distension pressure than the untreated group (P<.01), but there was no difference between the GV and papaverine groups (P=.77).
Satisfactory samples for toluidine blue staining were obtained from 27 patients before preparation (control samples) and 29 patients after preparation (prepared samples). Fig 2⇓ shows typical appearances of well-preserved and damaged areas. Scanning electron microscopy of en face preparations confirmed the appearances of satisfactory preservation and endothelial loss as seen under the stereomicroscope. Typical scanning electron micrographs of well-preserved and damaged intima are shown (Fig 3⇓).
The mean value for endothelial coverage in the control samples (n=27) was 93.1±1.2% (Fig 4⇓). The untreated group (n=10) had a mean endothelial coverage of 43.7±7.0%, the papaverine treatment group (n=9) had 44.1±9.2%, and the GV treatment group (n=10) had 68.7±7.0%. There was no significant difference between the combined pharmacological treatment groups and the untreated group (P=.17). The GV group had significantly greater endothelial coverage than the papaverine group (P<.05). The control sample data were not included in the statistical analysis.
Table 3⇓ shows levels of metabolites and indexes of energy status in the saphenous vein wall. In general, measurements made in the freshly isolated (control) samples were greater than those for all the prepared samples except for ADP and energy charge.
The ATP levels in the pharmacologically treated groups (combined GV and papaverine) were significantly greater than those of the untreated group (P<.05) (Fig 5⇓).The ATP-to-ADP ratio and the energy charge in the pharmacologically treated groups were markedly greater than in the untreated group (P<.01). The levels of breakdown products of high-energy phosphates (AMP, inosine, and hypoxanthine) were significantly lower in the pharmacologically treated groups than in the untreated group (P<.05 in all cases). However, there were no statistically significant differences in energy compound levels between the GV and the papaverine groups.
The present study showed that pharmacological treatment with GV or papaverine significantly reduced the pressure required to overcome spasm in vein grafts, and that of the two treatments, GV tended to have the greater effect. The GV-treated vein grafts showed significantly greater endothelial coverage than did papaverine-treated or untreated grafts. Both pharmacological treatments improved preservation of high-energy phosphate compounds in the vein wall.
If the endothelium is lost during harvesting, the denuded areas become covered by regenerated endothelium after 1 to 2 weeks.22 However, this neoendothelium is laid over a carpet of platelets and fibrin that has been deposited on the thrombogenic basement membrane. Later, this thrombus is replaced by proliferating smooth muscle cells. Platelets are known to release various factors that are implicated in the development of intimal hyperplasia, which, if progressive, may lead to graft occlusion. A variety of growth factors are involved in endothelial segmentation and subsequent intimal hyperplasia.23 Thus, preservation of the intima is vital to graft integrity.
The present study has confirmed the findings of others7 8 14 that high-pressure distension during preparation damages vein grafts. In addition, we have shown that by using a powerful vasodilator solution, it is possible to reduce the level of distension pressure required to prepare a vein graft and thereby produce a graft with superior structural and metabolic integrity. However, it remains to be proved that the better-quality graft will have longer patency. A study by Angelini et al12 in pigs showed that compared with undistended grafts, distension of vein grafts to 600 mm Hg reduced the patency in the carotid circulation studied up to 5 weeks. The fact that vein grafts of primates distended at high pressure have an increased uptake of lipid from the plasma 7 months later suggests that the late patency would also be reduced.8 However, until the results of a long-term study of the patency of distended and nondistended vein grafts are available, this question will remain open. In the meantime, we believe that there is sufficient circumstantial evidence that overdistended vein grafts are damaged grafts and that given the poor long-term results of saphenous vein grafts, it is preferable to implant an undamaged vein graft than a damaged one. The GV technique is simple to use, reduces the damage to vein grafts assessed by structural and biochemical measures, and therefore should be a useful adjunct to coronary artery bypass graft surgery.
High distension pressure has been shown to cause loss of endothelium. The excessive tension in the vein wall causes separation of endothelial cells, which then contract and are easily washed away when the graft is perfused. Extremes of temperature, pH, and osmolarity in the storage solution can also have similar effects on the endothelium.24 High distension pressures can also cause loss of high-energy compounds from the vein wall.12 25
In the present study, the peaks of distension pressure generated by the syringe were maintained for only a few seconds at a time, but this would be sufficient to disrupt the medial smooth muscle and to stretch the intima so that the endothelial cells separate, contract, and are able to be removed after perfusion of the graft. The damaged vein graft samples were not exposed to the flow of blood, but we observed under the microscope that gentle irrigation of a damaged vein was sufficient to dislodge the separated endothelial cells. The mean pressure measured was derived from all episodes of distension and quantified the entire damaging insult. A higher mean pressure is the result of higher peaks being reached during each episode of distension and thus indicates a greater potential for damage to the vein wall.
The endothelium was best preserved with GV treatment. A likely reason for this would be the lower peak distension pressure in this group, which reduced shedding of endothelial cells. Although this was not statistically significant (P=.066), this could be due to the small group numbers. Another possible explanation is the calcium-blocking effect of GV solution on the endothelium itself, which could inhibit contraction of endothelial cells. Another factor could be the provision of a physiological pH in GV solution (pH 7.4), which is lacking in unbuffered Ringer’s solution (approximate pH 5; range, 4.5 to 7.5), which was used in the vein lumen in the papaverine-treated and untreated groups. Although the endothelial coverage was greatest in the GV group, there was still considerable cell loss compared with the preharvesting control samples.
Many other factors are involved in damage to the endothelium other than physical trauma during harvesting and preparation. Most important are the storage conditions for the vein between harvesting and grafting, which may subject the vein to ischemic damage.26 Our results suggest that storage conditions also need to be improved to enhance the endothelial coverage and energy preservation in vein grafts.
The vast majority of ATP in the saphenous vein wall resides in the smooth muscle cells of the media.6 This is a consequence of the fact that the most common cell type in the vein wall is the smooth muscle cell. Angelini et al6 have shown that high distension pressure causes a reduction in the ATP content in the vein wall, which is therefore in large part a reflection of medial damage rather than loss of endothelium. Loss of ATP is presumably caused by contraction of the smooth muscle against high distending pressure. We found that in terms of energy status, pharmacologically treated veins were better preserved than the untreated veins. This was particularly evident in the parameters of cellular energy such as the levels of ATP, the ATP-to-ADP ratio, and the energy charge. Significantly lower values for AMP, inosine, and hypoxanthine compared with the untreated group are consistent with reduced breakdown of high-energy phosphates.
Improved ATP levels and ATP-to-ADP ratio in response to improved preservation techniques have also been demonstrated by Angelini et al.25 27 The most likely reason for improved energy status is the reduction of energy expenditure by the vein due to reduced contraction of medial smooth muscle against the lower distension pressure in a relaxed vein.
Papaverine is the vasodilator most described in the literature. Previous studies have suggested that the use of solutions containing papaverine for distension and storage13 15 or for perivenous infiltration or topical application26 28 during vein harvesting have resulted in better histological appearance of the grafts. However, there is the possibility that papaverine used intraluminally may damage the endothelium due to its acidic pH, and there is evidence that papaverine may reduce prostacyclin production in the vein wall.15 Therefore, in the present study to avoid endothelial damage and the risk of systemic hypotension, in the papaverine-treated group, papaverine was applied only to the outside of the vein, and heparinized Ringer’s solution was used intraluminally.
The benefit of GV solution over papaverine shown in the present study was the superior endothelial coverage. There also was a trend toward a lower distension pressure. In our clinical experience, no episodes of clinically significant systemic hypotension attributable to the use of GV solution were observed when the solution was used as described. Thus, GV solution has the additional advantage of being able to be injected relatively freely when the vein is still in continuity with the circulation.
We found the en face preparation of the intima of the saphenous vein stained with toluidine blue to be a reliable technique for visualizing and quantifying the endothelial coverage of the saphenous vein. This method has been described by Merrilees et al.19 Our confidence in this method was strengthened by the concordance of the morphological appearances by light microscopy with those seen by scanning electron microscopy of the same vein samples. The blinded design of this study during both data collection and analysis eliminated any investigator or observer bias.
We reached three conclusions. First, all types of surgical preparation and storage cause some loss of the endothelium and a decrease in the energy status of the vein wall. Second, treatment of saphenous veins with vasodilators allows the use of lower distension pressure during harvesting, reduces the loss of endothelium, and retards breakdown of high-energy phosphates in the vein wall. Third, topical and intraluminal use of GV solution improves endothelial coverage compared with the topical use of papaverine.
This work was supported by a Baker Institute block grant from the National Health and Medical Research Council of Australia. We acknowledge the assistance of Prof David H. Barkla and Christopher Mayberry of the Department of Anatomy at Monash University; Dr Rodney Dilley, Marion Attwater, and Neil Potter of the Baker Institute; and the operating theater nurses, surgeons, and residents of the Cardiac Surgery Department, Alfred Hospital.
- Copyright © 1995 by American Heart Association
Bourassa MG, Campeau L, Lesperance J, Grondin CM. Changes in grafts and coronary arteries after saphenous vein aortocoronary bypass surgery: results at repeat angiography. Circulation. 1982;65(suppl II):II-90-II-97.
Campeau L, Enjalbert M, Lesperance J, Vaislic C, Grondin CM, Bourassa MG. Atherosclerosis and late closure of aortocoronary saphenous vein grafts: sequential angiographic studies at 2 weeks, 1 year, 5 to 7 years, and 10 to 12 years after surgery. Circulation. 1983;68(suppl II):II-1-II-7.
Grondin CM, Campeau L, Lesperance J, Enjalbert M, Bourassa MG. Comparison of late changes in internal mammary artery and saphenous vein grafts in two consecutive series of patients 10 years after operation. Circulation. 1984;70(suppl I):I-208-I-212.
Boerboom LE, Bonchek LI, Kissebah AH, Werner PH, Pepper JR, Olinger GN, Korns ME, Garancis JD. Effect of surgical trauma on tissue lipids in primate vein grafts: relation to plasma lipids. Circulation. 1980;62(suppl I):I-144-I-147.
Haudenschild C, Gould KE, Quist WC, LoGerfo FW. Protection of endothelium in vessel segments excised for grafting. Circulation. 1981;64(suppl II):II-101-II-107.
Hayat MA. Fixation for Electron Microscopy. London, UK: Academic Press; 1981.
Snedecor GW, Cochran WG. Statistical Methods, 7th ed. Ames, Iowa: Iowa State University; 1980:226-227.
Brindle NPJ. Growth factors in endothelial regeneration. Cardiovasc Res. 1993;27:1162-1172.