Decreased patency rates following endoscopic vein harvest in coronary artery bypass surgery

Abstract

Objectives. Endoscopic vein harvest has gained widespread use in coronary artery bypass surgery. However, potential negative mid- and long-term effects following endoscopic vein harvest have been described. We aimed to compare long-term clinical outcomes following endoscopic and open vein graft harvesting. Design. This study was a clinical follow-up with additional computed tomographic coronary angiography among 126 first-time bypass patients originally included in a randomized study comparing early leg wound complications and cosmetic results. Deceased patients were retrospectively followed up. Results. Follow-up was complete, but information on clinical endpoints was not available in all patients. A total of 111 patients were alive at follow-up. Median observation time was 6.3 (range: 0.2–9.1) years including three in-hospital deaths. Vein graft failure was significantly higher in the endoscopic vein harvest (EVH) group (13 of 31; 42%) compared with the open vein harvest (OVH) group (2 of 32, 6%) (P = 0.001). However, this difference was not reflected by differences in recurrence of angina (P = 0.44), myocardial infarction (P = 0.11), and all-cause mortality (P = 0.15). Conclusions. Using a median follow-up time of 6.3 years significantly more vein graft failures were identified following EVH compared with OVH without any differences in long-term clinical outcomes.

Trial registration: ClinicalTrials.gov identifier: NCT01480726.

Key words::

• coronary artery bypass grafting
• endoscopic vein harvest
• long-term follow-up
• vein graft patency

Introduction

Saphenous veins are commonly used as conduits in coronary artery bypass grafting (CABG), and endoscopic vein harvest (EVH) has gained widespread use as an alternative to conventional open vein harvest (OVH) due to its well-documented short-term benefits of EVH in reducing leg wound complications (1–6). However, potential negative mid- and long-term effects following EVH have been ventilated, that is, reduced vein graft patency, increased risk rates of repeat revascularization, myocardial infarction (MI), and death (7–9). In contrast, several reviews and meta-analyses have concluded that EVH can be performed safely without increasing the early and mid-term risks (2,4–6,10,11). This information is primarily based on results from non-randomized studies, and there is a lack of knowledge from randomized studies. Additionally, there has been a discussion regarding the number of EVH procedures needed for operators to overcome the EVH learning curve (12,13).

The aim of the present study was to perform a follow-up among CABG patients initially randomized to EVH or OVH between April 2004 and June 2007 focusing on long-term outcomes regarding MI, recurrence of angina, vein graft patency, and mortality.

Materials and methods

Study design

This study was a secondary clinical follow-up with additional computed tomographic coronary angiography (CT-CA). Deceased patients were followed up retrospectively using the hospitals’ administrative systems and patient records.

The study was approved by the Local Ethical Committee (date of issue: May 27, 2011; registration number: N-20110023). The study was in accordance with the Helsinki Declaration of 1975, as revised in 1983, registered at ClinicalTrials.gov (Identifier: NCT01480726) and approved by the Danish Data Protection Agency (record numbers: 2008-58-0028 and 2011-331-0587).

Operative procedures and postoperative routines

The exclusion criteria and operative procedures and were previously described (1). In brief, patients were randomized to either OVH or EVH using the Guidant VasoView 5 and 6 Endoscopic Harvesting Systems (Guidant Corporation, CA, USA). EVH was performed through a 2–3-cm skin incision made medially above or just below the knee area. Under direct visualization, the saphenous vein was dissected proximally. During OVH the saphenous vein was exposed through a longitudinal continuous incision starting just proximally to the medial malleolus but never extended beyond the knee. The veins were immediately flushed and gently manually pressure distended. All veins in the OVH group were harvested by seven different but experienced cardiothoracic surgeons, while all EVH procedures were performed by a single surgeon following a learning curve of 30 EVH procedures. No heparin was given prior to or during vein harvest. All grafts were determined patent intraoperative by transit time Doppler measurement at the end of surgery.

Acetyl salicylic acid 75 mg daily was initiated postoperatively. Clopidogrel 75 mg daily was added for 12 months if the patients suffered from acute MI surgery or unstable angina just prior to surgery. Statins were resumed before discharge.

Patients and clinical follow-up

Through the hospital administrative system all patients (n = 129) included in the original randomized study (1) were identified including their vital status. Pre-, peri- and immediate postoperative data regarding the CABG procedures were retrieved from the original dataset.

During the follow-up period from November 2011 to December 2013, all surviving patients were contacted by phone and mail before they were invited to the hospital. The patients were interviewed about symptoms of recurrent angina pectoris as well as referrals to other hospitals or private doctors due to cardiac symptoms. Information about actual medication and smoking status was collected. A diagnosis of angina was based on a history of chest discomfort including at least two of the following three clinical symptoms: Chest discomfort described as heaviness, pressure, or squeezing at the substernal site; chest discomfort related to physical exercise or emotion; and chest discomfort that is relieved by rest in less than 10 min or within 5 min following the use of nitroglycerin.

Register follow-up

For deceased patients, we obtained patient records in order to identify postoperative hospitalizations, diagnoses, treatments, and causes of death. Information on vein graft failure was obtained if repeat conventional coronary angiography (CAG) or an autopsy was performed since surgery. A diagnosis of MI was accepted if this diagnosis appeared in the patient record.

Laboratory follow-up

Patients were scheduled for a contrast-enhanced CT-CA in order to describe any stenosis > 50% or occlusion of the vein grafts. Both were accepted as vein graft failure. Further graduation of vein graft stenosis was not performed as any stenosis > 50% was considered clinical important without any measurement of the fractional flow reserve. Within seven days prior to the CT-CA a 12-lead electrocardiography (ECG) was obtained together with a blood sample for measurement of the hemoglobin concentration, p-creatinine level, and the estimated glomerular filtration rate (e-GFR). A diagnosis of MI since surgery was accepted if a new Q-wave was seen in the ECG.

Patients were excluded from CT-CA if they were not in sinus rhythm, had an e-GFR < 60 ml/min and/or p-creatinine level > 120 μmol/L, were allergic to the contrast used for CT-CA, or if they did not give their written informed consent for the CT-CA. CT scans were performed with a GE Lightspeed VCT 64-slice scanner in the first year of the follow-up period and thereafter with Siemens Definition Flash. All CT scans were performed with a dedicated protocol for graft visualization. The field of interest was set for full visualization of the vein grafts, and the actual scan was preceded by a small tracking bolus of contrast to ensure maximum contrast staining of the vein grafts during the scan. All scans were performed as helical scans. Before the scan, metoprolol was administered intravenously to reduce the heart rate to below 65 bpm, and nitroglycerin was administered sublingually to all patients. Analysis and description of the CT-CA was performed by cardiologists with experience in both CT-angiography and invasive cardiology. Due to practical reasons, the CT-CA investigators were not blinded for the type of vein harvest.

Statistical analysis

All patients were analyzed by the intention-to-treat principle and EVH patients converted to OVH were kept in the EVH group.

Occurrence of a clinical event was considered a binary outcome. The raw event counts were compared using chi-squared or Fischer's exact tests to test for any statistically significant difference between treatment groups. P values < 0.05 were considered statistically significant. The number of clinical events in this trial was low and the trial was not powered to show non-inferiority in any of the clinical events. In addition, a Cox proportional hazards model was used to compare all-cause mortality between the two treatment groups as the exact date of death or censoring was available. Due to the low number of events, treatment group was the only covariate in the Cox model, that is, no adjustment for other covariates was performed. The 95% confidence interval (CI) for the hazard ratio (HR) was calculated and CIs excluding 1.0 were considered statistically significant. Analyses were done using the statistical software package Stata version 13.1 (StataCorp LP, Texas, US).

Results

A flow chart for the follow-up of all patients is shown in Figure 1.

Figure 1. Flowchart showing details of postoperative follow-up.

Patient characteristics at follow-up

Three patients included in the originally randomized study were excluded from the follow-up study as they only received arterial grafts due to low quality of the vein graft harvested, leaving 126 patients for analyses. Pre- and perioperative patient characteristics of these patients are shown in Table I. A total of 111 patients were alive at follow-up, but one patient did not want to attend the clinical follow-up. Retrospective follow-up alone using patient records and the hospital register was thus performed in 16 patients.

Table I. Pre- and perioperative characteristics of patients included in the follow-up study.

Variable	Endoscopic vein harvesting (n = 64)	Open vein harvesting (n = 62)	P-value
Age (years), mean ± SD	64.7 ± 9.8	64.6 ± 8.2	0.94
Male, n (%)	57 (89)	52 (84)	0.39
Diabetes mellitus, n (%)	7 (11)	9 (15)	0.55
Smokers, n (%)	9 (14)	16 (26)	0.10
Hypercholesterolemia, n (%)	47 (73)	48 (77)	0.60
Obesity (BMI > 30 kg/m^2), n (%)	12 (19)	17 (27)	0.25
CABG + ECC, n (%)	52 (81)	42 (68)	0.08
No. of vein anastomoses, (mean ± SD)	1.5 ± 0.7	1.5 ± 0.7	0.80
LIMA, n (%)	64 (100)	61 (98)	0.49
RIMA, n (%)	39 (61)	32 (52)	0.29
Radial artery, n (%)	0 (0)	2 (3)	0.24
Use of tranexamic acid, n (%)	40 (63)	36 (58)	0.61
Use of aprotinin, n (%)	1 (2)	0 (0)	1.00

BMI, body mass index; CABG, coronary artery bypass grafting; Diabetes mellitus = treatment with either insulin or oral agents; ECC, extracorporeal circulation; LIMA, left internal mammary artery; RIMA, right internal mammary artery.

Patient characteristics at the time of the clinical follow-up are shown in Table II. Median observation time at follow-up for all patients was 6.3 (range: 0.2–9.1) years including three in-hospital deaths. There was no difference in the median follow-up time between the EVH and OVH group (6.25 years and 6.34 years, respectively).

Table II. Patient characteristics stratified by type of vein harvesting technique at the time of follow-up.

Variable	Endoscopic vein harvesting (n = 58)	Open vein harvesting (n = 52)	P-value
Active smokers, n (%)	5 (9)	13 (25)	0.02
p-creatinine (μmol/l), mean ± SD	96.3 ± 36.8^a	92.3 ± 37.2^a	0.65
Estimated glomerular filtration rate < 60 ml/min, n (%)	8 (22) ^a	6 (17) ^a	0.55
Ongoing atrial fibrillation, n (%)	8 (19)^b	10 (21)^c	0.75
Medication, n (%) Platelet inhibitors β-blockers ACE inhibitors Calcium antagonists Statins Anticoagulation^g	54 (93) 47 (83)^d 25 (43) 4 (7) 55 (95) 3 (5)^f	48 (92) 38 (73) 27 (53)^e 4 (8)^e 44 (85) 7 (14)	1.00 0.24 0.31 1.00 0.08 0.19

^an = 36; ^bn = 43; ^cn = 47; ^dn = 57; ^en = 51; ^fn = 56 ^gEither Warfarin or Dabigatran etexilate mesylate.

Recurrent angina

Results from the follow-up are shown in Table III stratified by type of vein harvesting technique. Information on symptoms of recurrent angina was available for 110 patients (52 in the EVH group and 58 in the OVH group). A total of 18 patients had experienced or were suffering from recurrent angina (eight patients in the EVH group and 10 in the OVH group). Of these, 12 patients had been referred to hospital for further evaluation and treatment due to chest discomfort. CT-CA showed complete revascularization without any new coronary stenosis in five of these patients. A new stenosis in the genuine coronary arteries was diagnosed in one patient and vein or arterial graft failure was demonstrated in six of the patients.

Table III. Long-term postoperative outcomes stratified by type of vein harvesting technique.

Outcome	Endoscopic vein harvesting	Open vein harvesting	P-value
Patency follow-up	(n = 31)	(n = 32)
Vein graft failure, n (%)	13 (42)	2 (6)	0.001
Clinical follow-up	(n = 58)	(n = 52)
Recurrence of angina, n (%)	8 (14)	10 (19)	0.44
Register follow-up	(n = 64)	(n = 62)
Myocardial infarction, n (%)	0 (0)	3 (5) ^b	0.11
All-cause death, n (%)	5 (8)	10 (16)	0.15
Cardiac mortality, n (%)	0 (0)	2 (3)	0.24

^an = 61; ^bn = 60.

Myocardial infarction and mortality

Three patients had suffered a MI since surgery (Table III). One patient in the OVH group had suffered a non-fatal MI three years following CABG and underwent repeat CABG. Another two patients in the OVH group died from ventricular fibrillation outside hospital one and six years postoperatively, respectively, and were classified as having fatal MIs.

Fifteen (12%) patients died before clinical follow-up. There were two cardiac deaths. Both occurred in the OVH group. All-cause mortality was lower in the EVH (5/64) than in the OVH group (10/62) with a non-significant HR of 0.48 (95% CI: 0.16–1.39).

Vein graft failure

Information on vein graft patency was obtained in 63 of 126 patients (50%) either by CT-CA, CAG, or autopsy with 31 patients in the EVH group (Table III). CT-CA was performed in 52 of 111 (47%) surviving patients. The main reasons why a CT-CA was not performed in all surviving patients were ongoing atrial fibrillation, decreased renal function, and patient who preferred only to participate in the clinical follow-up without having a CT-CA performed as shown in Figure 1. Information regarding vein graft patency was obtained from CAGs performed in 10 patients prior to follow-up and information on vein graft patency was obtained from the autopsy report in one patient.

Overall vein graft patency at follow-up was 76% among patients where information on vein graft patency was obtained. The incidence of vein graft failure in patients was significantly higher in the EVH group (13 of 31) compared with that in the OVH group (2 of 32) (P = 0.001). Only four of 15 patients with vein graft failures experienced angina. No specific information on potential vein graft failures was obtained in patients who died from MI.

Discussion

This secondary follow-up study among CABG patients initially randomized to either EVH or OVH in order to compare short-term outcomes reports clinical and angiographic results and has to date the longest follow-up time among patients included in a randomized study comparing the two vein harvesting methods. The primary finding was that there was significantly more vein graft failures following EVH compared with OVH without any differences in clinical endpoints.

The overall vein graft patency at follow-up was 76% in the present study and this is in the range of five-year patency rates reported in other studies (14,15). Reduced patency rate of EVH compared with OVH have also been shown in a few previous studies (7–9) evaluating early and mid-term patency rates of EVH compared with OVH; however, these studies were not randomized according to the vein graft harvest method. Lopez et al. (8) showed an association of EVH with higher rates of vein graft failure (46.7% vs. 38.0%, P < 0.001), reanalyzing a subset of patients who underwent angiographic assessment of the vein grafts in the Project of Ex-vivo Vein Graft Engineering via Transfection IV trial (PREVENT IV). In a recent study it was later concluded that this difference was not related to a specific EVH device (10). An increased vein graft failure rate following EVH was shown in a subset of patients undergoing angiographic follow-up at one year from the Randomized On/Off Bypass (ROOBY) trial which evaluated differences in clinical outcomes between patients undergoing on- and off-pump CABG (9).

In contrast to these results and the results from the present study several recent reviews and meta-analyses of pooled data have concluded that EVH can be performed without compromising vein graft patency rates (4–7,11,16). These discrepancies have been explained, for example, by the fact that previous studies showing decreased patency rates following EVH were not randomized according to the vein graft harvest technique. However, long-term results from properly powered randomized studies are still needed but until then results from the present secondary analysis add to the pool of knowledge.

Patency of vein grafts depend on several factors, for example, patient characteristics, technical skills of the surgeon, conditions of the target vessels at the time of surgery, the runoff area, postoperative medication, the postoperative arteriosclerotic process in the coronary arteries, etc. Thus, the harvest technique may only play a minor role.

Similar to the PREVENT IV study (10) in which there were no protocol requirements for pre-harvest use of heparin, all grafts in the present study were harvested by the EVH technique without heparinization. Lack of pre-procedural heparinization may, however, result in residual clot strands in the vein grafts and have the potential to decrease vein graft patency rates. Administration of heparin prior to EVH has been shown to decrease the saphenous vein clot burden (17,18) and attention should be directed toward the use of pre-procedural heparin administration.

Operator experience with EVH plays an important role for the quality of veins harvested by the endoscopic technique (17) although there is no consensus regarding the number of EVH procedures needed to overcome the learning curve. It has been demonstrated that operators with less than 100 cases of experience had nearly 50% more discrete vein injuries compared with more EVH harvesters having experience from more than 900 cases (12), but other investigators reported that about 20–35 cases were required to overcome the learning curve (13). The inferior patency rate following EVH in the present study may direct attention to the number of EVH cases needed to overcome the EVH learning curve. All veins in the EVH procedures were performed by a single surgeon following a learning curve of 30 EVH procedures (1). This number of EVH procedures may not be enough even if the surgeon was experienced in OVH.

In accordance with reviews and meta-analyses (2,4–6,10,11) we did not find any increased risk of recurrent angina, MI, or death following EVH. Recurrence of angina was only found in three of 15 patients in whom vein graft failure was diagnosed. Explanations for this may be development of collateral blood flow, medical treatment, and a possible reduced physical activity.

Study limitations

There are some limitations of this study that should be acknowledged. First, the study was not originally powered in order to study clinical long-term results following EVH and OVH, and the number of patients was relatively small. Thus, the study may only be hypothesis generating. However, no other studies have reported clinical and angiographic results following EVH beyond a median of 6 years, and we therefore believe that these results should add to the pool of knowledge for future meta-analyses. Second, the invasive cardiologists who performed and interpreted the CT-CAs were not blinded to the harvest technique for practical reasons thus imposing a risk of bias. Third, invasive CAG which is still accepted as the gold standard was not used for graft patency assessment, and although coronary bypass graft patency can be assessed using CT-CA (19,20), we might still have missed or misinterpreted some graft failures. Fourth, we were only able to obtain information on vein graft patency in 50% of the patients included in the original study. Some patients did not want to have a CT-CA performed inducing the risk of selection bias. Although we have no reason to believe that exclusion from CT-CA was related to the vein harvest method there is however still a risk of selection bias. Fifth, we have no information on compliance to antithrombotic medical treatment postoperatively and no precise information regarding the postoperative smoking history in details except for the smoking status at follow-up. Sixth, this study was a single-center study with only one surgeon performing EVH and results may not be externally valid for all centers; however, results from the present study should bring attention toward the number of EVH procedures needed for operators to overcome the EVH learning curve.

Conclusions

Using a median follow-up time of 6.3 years significantly more vein graft failures were identified following EVH compared with OVH without any differences in long-term clinical outcomes.

Declaration of interest: The authors report no declarations of interest. The authors alone are responsible for the content and writing of the paper.