Comparison of alternate preparative techniques on wall thickness in coronary artery bypass grafts: The HArVeST randomized controlled trial

The success of coronary artery bypass grafting surgery (CABG) is dependent on long‐term graft patency, which is negatively related to early wall thickening. Avoiding high‐pressure distension testing for leaks and preserving the surrounding pedicle of fat and adventitia during vein harvesting may reduce wall thickening.


| INTRODUCTION
Saphenous vein remains the most commonly used conduit for coronary artery bypass because of its predictable handling qualities and ready availability. 1 However, as many as 40% of vein grafts become occluded at 12 years postsurgery. 2,3 Early thrombosis occurs in up to 10% of vein grafts due to spasm or technical error. 4,5 Late vein graft failure occurs as a consequence of early neointimal hyperplasia with later superimposed atherosclerosis. 5 Therefore, the clinical implications of any intervention to improve long-term graft patency are significant.
Surgical preparation of the harvested vein is often overlooked but may be an important contributor to vein graft failure. Typically, the harvested vein is distended at uncontrolled high pressure with blood or fluid using a syringe to test for leaks and relieve spasm before grafting; this has been shown to result in endothelial loss and medial disruption. 6,7 Our group proposed an alternative protocol that tests for leaks by connecting the harvested vein to a side branch of the aortic return cannula on the cardiopulmonary bypass circuit. 8 In this way, the grafts are then distended at the patient's own systemic pressures. We hypothesized that this should attenuate medial disruption, platelet adherence, the release of mitogenic growth factors and subsequent neointima formation, as supported by our observations that avoiding high-pressure distension reduces medial damage and preserves release of prostacyclin and nitric oxide (NO), both of which have potential antiproliferative actions. [6][7][8] Our technique has also been shown to reduce neointima formation in cultured human saphenous vein grafts in vitro compared with conventional harvest. 9 However, while avoiding pressure distension increases graft patency in vivo in a porcine saphenous vein to carotid artery interposition model, this appears to be due to a reduction in early thrombosis rather than an effect on neointima formation. 10 Furthermore, it is not known whether avoiding high-pressure distension by our method improves long-term graft patency.
Souza and colleagues described a nontraumatic vein harvesting technique that combines avoidance of pressure distension with preservation of the surrounding adventitia and fat or "pedicle," which is stripped during conventional vein harvesting. 11 The pedicle has been shown to preserve wall architecture and endothelial function in vitro and experimental studies have shown that the preserved fat tissue is a rich source of NO. 12,13 It has been hypothesized that preservation of the adventitial microcirculation may reduce subsequent atherogenesis by improving wall oxygenation and reducing oxidative stress. 14 Until recently, clinical trials of therapies designed to prevent vein graft disease were limited by the large sample sizes required to measure changes in the lumen (i.e., new lesions) or patency rates by postoperative angiography. Intravascular ultrasound (IVUS) accurately and reproducibly measures vein graft wall dimensions and quantifies levels of plaque and wall fibrosis. [15][16][17] Using this technology, we set out to determine the effects of the vein graft harvesting technique and pressure distension on the development of vein graft wall thickness in a factorial randomized controlled trial (RCT), which allowed both techniques to be evaluated in the same trial.

| Trial design
A single-centre, parallel group, factorial RCT with blinding of participants and clinical and research staff not involved in the operation.
Participants were randomly allocated in a 1:1:1:1 ratio to one of four treatment groups; conventional harvest and high-pressure test, conventional harvest, and low-pressure test, pedicled harvest and high-pressure test, or pedicled harvest and low-pressure test.
The study was approved by a National Health Service Research Ethics Committee (Wiltshire ref. 09/H0104/28). The trial was registered as ISRCTN10567790. Patients and the public did not contribute to the design or conduct of the trial.

| Participants
Adults aged 18 and over undergoing first time nonemergency coronary artery bypass grafting surgery (CABG; either on or off-pump) at the Bristol Heart Institute, with at least one saphenous vein graft and not requiring valve replacement/repair or an aortic procedure were eligible to take part. 18,19 Patients with congestive heart failure, ejection fraction < 30%, preoperative serum creatinine >104 μmol/L, peripheral vascular disease, allergy to iodinated contrast media, participating in another interventional study, or unwilling to participate in follow-up were excluded. All participants provided written informed consent.

| Interventions
Harvest technique: Vein grafts were harvested either with (pedicle harvest) or without (conventional harvest) the pedicle of surrounding fat and adventitia, as described by Souza. 11 All grafts were harvested using a no-touch technique and were left in situ until systemic heparinisation.
Pressure test for leaks: vein grafts, excised following systemic heparinisation, were either flushed with heparinised blood from a 10 ml syringe (conventional high-pressure test) or attached to a side arm of the aortic canulae in patients undergoing on-pump surgery, or anastomosed first to the ascending aorta in off-pump surgery, and flushed with blood at systemic pressure (low pressure test). 8

| Outcomes
The protocol-defined primary outcomes were vein graft disease as measured by (i) wall thickness and (ii) lumen diameter assessed using IVUS at 12 months postsurgery. Multiple IVUS measurements were taken per graft and the patient-level mean for each measurement was used. Wall thickness at baseline was measured by histological analysis of a short segment of the harvested vein from the end of each graft retrieved before completion of the proximal anastomoses.
Secondary outcomes were lumen diameter and graft patency assessed by quantitative angiography at 12 months; serious adverse events (SAEs); wound infection using the ASEPSIS scoring system; duration of postoperative stay; leg wound pain or dysaesthesia at 3 and 12 months assessed using the neuropathic pain symptom inventory (NPSI) scoring system; and readmissions to hospital within 12 months. 20,21 SAEs not listed in the study protocol were coded using the Medical Dictionary for Regulatory Activities (version 19.0; McLean).

| Sample size
The sample size was set at 96 patients (24 per group), which is sufficient to detect an effect size of 0.5 standard deviations (SD) with 80% power and 5% statistical significance, assuming no interaction between the method of harvesting the graft and the method of testing for leaks, a correlation of 0.7 between the one pre-and one postrandomization measurement and allowing for 25% loss to follow-up.
An effect size of 0.5 SD equates to differences of ≈1.2 mm in the mean graft wall thickness and ≈2.4 mm 2 in lumen area between pedicled and conventional harvest groups, or between high-and lowpressure test groups, assuming estimated within-group SDs of 2.33 and 4.69, respectively. 16 The sample size calculation assumes only one graft per patient whereas, in practice, some patients received two or more vein grafts. 18 Therefore, the study was powered to detect differences somewhat smaller than 0.5 SD.

| Randomization
Allocations were generated by computer using block randomization with varying block sizes in advance of the study. A passwordcontrolled secure database concealed the allocation until data had been entered to confirm identity and eligibility. Randomization took place as close to the start of surgery as possible. The team member responsible for randomization was not involved in data collection for the study.

| Blinding
Participants were blinded to treatment allocation. The surgical team involved in the operation were unblinded. Research nurses collecting postoperative data, and assessors measuring wall thickness and lumen diameter with IVUS and quantitative angiography were blinded to treatment allocation. Laboratory staff conducting histological analyses of short segments of prepared vein could not be blinded to the harvest technique used.

| Follow-up
Postoperative management was in accordance with hospital protocols. Participants were followed up at 3 and 12 months by questionnaire to assess wound infection, leg wound pain and dysesthesia, and readmissions to hospital; and attended hospital for IVUS and angiographic imaging at 12 months.

| Statistical methods
Analyses were based on a prespecified statistical analysis plan, were performed on an intention-to-treat basis, and are reported in line with the CONSORT reporting guidelines. Patient level means of the multiple IVUS measurements per graft were calculated for wall thickness, lumen diameter, lumen area, and wall area. Binary outcomes were compared using logistic regression, counts using Poisson regression, time to event outcomes using Cox proportional hazards models, and continuous longitudinal outcomes using hierarchical mixed-effects models, to account for participants with multiple grafts and IVUS frames of each graft. Analyses were adjusted for IVUS frame number as a continuous fixed covariate and patient and graft number as random effects. Multiple imputation by predictive mean matching using chained equations was used to impute missing data for the primary outcomes and angiographic assessment of lumen diameter. Data missing due to graft occlusion were not imputed. See Supporting Information Appendix for further details. All analyses used the conventional harvest and high pressure as the reference groups and included harvest technique, pressure test and the interaction between two (if significant at the 10% level) and were adjusted for baseline measurements where available. If a significant interaction is indicated, pressure testing is compared and reported for each harvest technique separately (and harvest technique is   (Table 1, Supporting Information   Appendix Table A1). In total, 17 patients randomized to pedicled harvest received conventional harvest, 6 patients randomized to low-pressure received high-pressure. Unblinding of treatment allocation occurred for one patient randomized to the pedicled harvest and high-pressure group.
A total of 20 of the 96 patients did not attend for the 12-month IVUS and angiography assessment, 12 of whom returned at least one questionnaire at 12 months. Seventeen patients formally withdrew from the trial, 15 from 12-month IVUS and angiography assessment only and 2 from all follow-up ( Figure 1, Supporting Information Appendix Table A2).

| Baseline data
The mean age of participants was 66.0 years (SD 8.9), and 87/96 (91%) were male. By chance, patients randomized to pedicled harvest were slightly older than those randomized to the conventional harvest group (mean 67.3 vs. 64.5 years), and the low-pressure group had proportionally more males than the high-pressure group (96% vs. 85%). From histology of veins after surgical preparation, the median wall thickness was greater in the low-pressure (0.39 vs. 0.32 mm) and pedicled harvest (0.39 vs. 0.31 mm) groups and the median lumen diameter was smaller in these groups (pressure comparison 2.87 vs. F I G U R E 1 Flow of participants. Angio/IVUS, angiogram/intravascular ultrasound; NPSI, neuropathic pain symptom inventory. 1 Some patients may be ineligible for more than one reason 3.30 mm; harvest comparison 2.87 vs. 3.35 mm) ( Table 2, Supporting   Information Appendix Table A3).

| Operative details
The median duration of surgery was 3 h and was similar across groups. Overall, 28/96 (29%) procedures were performed on-pump.

| In-hospital outcomes
The median postoperative hospital stay was 6 days irrespective of harvest technique groups. However, in the low-pressure group, the

| Wound infection and pain
Two patients had an ASEPSIS score > 20 within their index hospital admission, increasing to five patients at any time from surgery to 12 months postoperatively (Supporting Information Appendix Table A8).
Fewer patients randomized to the low-pressure group reported pain (NPSI pain score > 0) at 3 months compared with the high-pressure   thickness increases two to threefold. 23 Vein wall thickening occurs rapidly within the first 2 months, stabilizing at 6-9 months. 5,24 Hence, by 12 months early fluctuations have abated, establishing the choice of primary endpoint, although atherosclerosis within 12 months is rare. 2,3,5 Owing to its effect on bulk fluid transfer and lipoprotein retention, total wall thickness is thought to be a key predisposing factor for subsequent graft atherosclerosis and late occlusion. Wall thickness may be considered a surrogate for longterm patency, although this remains to be investigated directly.

| Clinical implications
The success of CABG is dependent on long-term graft patency because vein graft occlusion is associated with a return of cardiac symptoms, an increase in myocardial infarction and death. 5 0.53 ± 0.08 mm, p < .01). 16 Similar excellent long-term results were reported by this group in a comparison of pedicled vein grafts with radial artery grafts. 22 Although the improved graft patency with the pedicled technique is encouraging the results derive from a single surgeon study and have not been confirmed by others.
Compared with low pressure distention of conventionally harvested veins, pedicle grafting requires more extensive dissection (which led to a number of protocol deviations) and increases the frequency of leg wound morbidity. 11 Our data demonstrate that it is possible to achieve similar benefits in terms of reduced wall thickening to pedicled grafts by using low pressure distension of conventionally harvested veins.

| Strengths and limitations
Strengths include the inclusive eligibility criteria, with few patients being ineligible and minimization of bias through concealed allocation. Personnel conducting the analyses, participants and staff not involved in the surgical procedure were blinded to the group allocation. The factorial study design allows two interventions to be evaluated in one study.
Limitations include that participants were recruited from a single centre, limiting the generalizability of the findings. There were protocol deviations, mostly in the pedicle harvest group, which reduces the power of the study to detect differences between the interventions. Some patients withdrew and did not attend the follow-up at 12 months, although the impact of these missing data were minimized by estimating the missing data using multiple imputation.
Finally, the trial was not powered to detect differences in clinical outcomes, and while the analyses were prespecified, it is possible that the differences in clinical outcome observed are due to a type I error.
In conclusion, conventional saphenous vein graft preparation with low pressure distension and harvesting the vein with a surrounding pedicle yielded similar graft wall thickness after 12 months, but lowpressure distention was associated with fewer short-and long-term adverse events.

ACKNOWLEDGEMENT
The British Heart Foundation and the NIHR Biomedical Research

Centre at University Hospitals Bristol and Weston NHS Foundation
Trust and the University of Bristol supported this study. This trial was designed and delivered in collaboration with the Clinical Trials and Evaluation Unit, Bristol Trials Centre, a UKCRC registered clinical trials unit, which is in receipt of NIHR CTU support funding.
The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care. The funders played no role in the design of the study, in the collection, analysis, and interpretation of data, or in the decision to submit the manuscript for publication.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.