Introduction

Despite widespread use of effective medical therapy aimed at atherosclerotic risk factor modification [1], coronary artery disease (CAD) remains the leading cause of morbidity and mortality in the USA. Coronary artery bypass grafting (CABG) is the preferred strategy for complete revascularization of multi-vessel coronary artery disease [2]. In patients with left main disease, advanced age, diabetes, and reduced left ventricular ejection fraction (EF), CABG has been shown to improve survival and lower the rates of major adverse cardiac events (MACE) [3] when compared to the medical therapy in stable CAD patients. At the time of surgical revascularization, both arterial and/or saphenous vein grafts (SVG) can be used to bypass native coronary artery stenosis or occlusion. However, due to improved outcomes and superior patency of internal thoracic artery (ITA) grafts compared with vein grafts, there has been a significant increase in the use of ITA as the bypass conduit of choice (from 88% in 2000 to 95% in 2009) [4,5,6,7]. Nevertheless, the use of arterial grafts has extended beyond ITA to include radial artery (RA), gastroepiploic artery (GEA), and inferior epigastric artery (IEA) grafting.

Although complete arterial revascularization seems theoretically advantageous, SVG continue to be the most widely used conduits in CABG with an average of 2 grafts per procedure [8]. SVG have been shown to have inferior patency as compared to arterial conduits with occlusion rates as high as 25% after the first year post-CABG and 75% over 10 years [8, 9]. From the clinical standpoint, vein graft failure is associated with increased risk of composite endpoint of death, myocardial infarction (MI), and repeat revascularization at 4 years [8]. Given the outlined issues with venous conduits, some surgeons elect to use arterial grafts (specifically RA) instead of SVG. In this review, we aim to discuss and inform clinicians of the current evidence supporting the use of RA conduits over venous grafts in patients undergoing CABG.

Radial Artery Grafts

Shortly after their introduction in the early 1970s, the use of RA grafts was abandoned. This was largely due to poor patency rates, which were thought to be secondary to arterial spasm, endothelial denudation, and graft hyperplasia [10]. Revival of RA graft use ensued two decades later after it was found that the long-term patency rates with RA grafts were significantly higher than previously reported [11]. In their 1992 publication, Acar et al. reported > 90% patency after a mean follow-up of 9 months, among 104 consecutive patients treated with RA grafts in addition to ITA [11]. One major contributor to the improvement in RA graft patency was the ability to reduce the incidence of RA spasm—thought to be the main cause of graft failure. Prevention of spasm was achieved by using the atraumatic “no-touch” technique during harvesting and the use of pharmacologic dilation as opposed to the more traumatic mechanical dilation, as well as optimal use of systemic vasodilator therapy during the postoperative course [12]. Furthermore, rates of graft occlusion can be lowered if RA grafting is used to bypass more severe coronary artery lesions. Use of RA conduits in less severe coronary artery stenosis is prone to spasm and subsequent graft failure because of competitive flow [13].

  1. a.

    Advantages: In patients without a suitable ITA, RA conduit represents a reasonable alternative. Given its relatively superficial course, RA is an excellent choice for harvesting at the time of bypass surgery. RA is typically harvested from the non-dominant arm, using either an open or endoscopic approach with an ultrasonic scalpel [14]. The usual length (> 20 cm) and diameter (2–3 mm) of the harvested RA conduit allows the surgeon to bypass virtually any coronary artery [12]. Furthermore, its thick muscular wall allows it to be easily anastomosed to both the aorta and the native coronary artery. Unlike SVG, RA graft is preconditioned to high arterial pressure which can explain the superior results associated with its use. Additional benefit of RA use over SVG is the avoidance of ambulatory dysfunction, which is a frequent complication associated with saphenous vein grafting [15]. Furthermore, graft disease is uncommon in RA grafts, a phenomenon that is found commonly in SVG within 5 years of surgery [16].

  2. b.

    Pitfalls: Although harvesting of the RA rarely results in motor dysfunction or compromised perfusion, sensation abnormality attributable to nerve damage has been described as the most common complication, affecting up to one-third of all patients [17]. Nevertheless, great care needs to be taken to ensure that adequate collateral circulation is present in the ipsilateral forearm prior to RA harvesting. Methods established to confirm the patency of collateral circulation via ulnar artery include Allen’s test, arterial Doppler testing, and oximetric plethysmography [18]. Aside from patency of the ulnar artery, other factors which can pose a challenge to using RA as a bypass conduit include visualization of RA plaque on ultrasound, presence of arterio-venous fistula for hemodialysis access, history of trauma, native RA disorders (i.e., vasculitis), and/or previous RA cannulation. Due to the histologic composition of RA as a single-layer elastic lamina with multiple fenestrations, there is a significantly higher prevalence of atherosclerosis in RA as compared to ITA [19]. Finally, the highly dense muscle cell composition of the RA renders it very sensitive to systemic vasopressor administration and to external trauma. [19, 20].

Saphenous Venous Graft

Despite the inferior patency rates of saphenous vein grafts, SVG conduits are utilized in most bypass surgeries, especially in those patients who undergo repeat CABG or have poor arterial conduits. Saphenous vein is a large superficial vein which courses through the medial aspect of the lower extremity and drains into the common femoral vein at the level of the femoral head. Compared to arterial conduits, SVG are thin walled, large in caliber, and have an ill-defined laminar layer. Sparse amount of smooth muscle cells, presence of valves, and abundance of vasa vasorum make these conduits structurally and functionally different from arterial conduits [21].

  1. a.

    Advantages: In patients who have limited amounts of available graft conduits (i.e., patients with prior bypass surgery, presence of peripheral artery disease, or an arterio-venous fistula for hemodialysis access), single SVG can be used to achieve multi-vessel revascularization. These sequential grafts, which have a single proximal aortic anastomosis and multiple distal coronary anastomoses (m-SVG), are preferred over (s-SVG) grafts in which one SVG has only a single distal coronary anastomosis. Furthermore, the use of sequential vein grafting techniques allows for shorter surgical procedure times and more significant improvement in hemodynamics as well as a more complete revascularization. Interestingly, however, recent data suggest that patients undergoing bypass with s-SVG have superior 1-year graft patency rates and improved clinical outcomes as compared to patients receiving m-SVG [22].

  2. b.

    Pitfalls: Large caliber, thin walls, and ill-defined elastic lamina make SVG more prone to ischemic injury, lipid deposition, and atherosclerosis as compared to RA grafts [21]. Given their high compliance and the natural adaptation to low pressure, exposure of SVG conduits to arterial pressures can lead some conduits to undergo early graft failure.

Comparison of RA Grafts with SVG in Observational Studies

Survival benefit associated with CABG in high-risk patients is predominantly determined by graft patency. Several observational studies utilizing angiography have shown that RA conduits have superior patency over SVG at a short- and long-term follow-up. RA grafts, either used separately or as composite grafts, have also been shown to have equivalent patency to other arterial grafts [23]. In one of the earliest retrospective studies, Acar et al. [11] showed RA conduits to have 100% patency at 2 weeks following revascularization and 93.5% patency at 9 months of follow-up, thus establishing RA as a reasonable substitute for IMA. In a more recent report, the long-term patency (10 years of follow-up) and perfect patency (widely patent vessel without irregularity, stringed, or occlusion) of the RA grafts were 91.6% and 88% versus 67.1% and 53.4% for SVG, respectively [24]. The most important determinant of patency in this study was the severity of target vessel stenosis, whereas there was no association between patency and either target vessel location or use of calcium channel blockers [24]. In a prospective review of 175 consecutive patients who underwent bypass grafting by a single surgeon, there were no RA harvest site hematoma or infection. Transient dysesthesia occurred in 2.6% patients with no death reported. In the follow-up of 3 months, patency rate exceeding 95% was observed via elective coronary angiography [25]. Zacharias and colleagues evaluated the comparative outcomes between RA or SV grafts as an additional conduit for patients undergoing CABG with LIMA to LAD grafting. In this study, which had similar distribution of demographics and operative characteristics between the two groups, RA grafting was associated with superior estimated survival (92% vs. 87%, p < 0.03) at 6 years of follow-up as compared with SVG [26]. In a short- and mid-term follow-up in a Japanese Institute [27], no adverse effects were noted after CABG with RA grafting. The cumulative graft patency rate was 93% after 1.5 years of angiographical follow-up with graft calculated patency rates at 1, 2, and 3 years of 98.2%, 91%, and 86.2%, respectively [27]. “String sign,” a marker of diffuse severe occlusion of RA graft was found in 13 patients in a mean follow-up of 13 years [27].

Whereas some reports found RA grafts to be superior to SVG, others found either similar or superior patency and outcomes with the use of SVG. In a 5-year follow-up of 50 asymptomatic patients who underwent bypass surgery with at least one RA graft, coronary angiography was performed to determine graft patency and presence of narrowing, graft patency rates were 100% for LIMA, 94% for RIMA, 89% for RA, and 92% for SV grafts [28]. Within the cohort of 62 patients with RA conduits, “String sign” (diffusely narrowed and diseased RA graft) was found in 7 patients (11%). In an angiographic review of 310 patients who underwent CABG using IMA, SVG, and RA conduits between 1996 and 2011 at Cleveland Clinic and presenting with signs and symptoms of ischemia, RA graft patency was found to be 51% as compared to 64% in SVG (p = 0.0016) and 91% in LIMA grafts (p = 0.0001) [29]. Interestingly, within the subsets of patients with RA grafts, women had lower patency as compared to their male counterparts 38% versus 56%, respectively. Lower RA graft patency rates in this study were attributed to increased vasoreactivity, intimal hyperplasia, atherosclerosis, and graft dependence on high level of blood flow. Major limitation of this study was the fact that only those patients with active ischemia were studied, thereby introducing significant selection bias. Summary of these results illustrates that there is substantial variability within the non-randomized literature, and therefore equipoise regarding superiority of RA over SVG conduits. Although RA grafts appear to have superior patency rates and outcomes in some reports, these same endpoints are better in the SVG arm of other studies.

Randomized Controlled Trials Comparing RA and SV Conduits

To date, six RCTs have been conducted to address the inconsistent results of the non-randomized studies regarding the superiority of RA versus SV grafts (Table 1). Radial Artery Patency and Clinical Outcomes (RAPCO) trial was the first single-center RCT published in 2003 which compared patency rates and event free survival between RA, RIMA, and SV grafts after CABG over a 10-year follow-up [35]. In group 1, RA grafts were compared to RIMA grafts in patients under 70 years of age, whereas in group II RA grafts were compared with SVG in patients 70 years and over. Graft patency estimates after 5 years were 95% (RA) vs. 100% (RIMA), p = 0.40 in group 1, and 87% (RA) vs. 94% (SVG), p = 0.50 in group 2. The event-free survival at 5 years was 91% (RA) vs. 82% (RIMA), p = 0.70 in group 1, and 85% (RA) vs. 89% (SVG), p > 0.90 in group 2. Therefore, the 5-year interim results did not support the superiority of RA graft over either SV or RIMA grafts in terms of graft patency or clinical outcomes [35]. In another single-center prospective study, 142 patients were randomized to receive either RA or SVG conduits for aortocoronary bypass grafting of the native left circumflex artery [33•]. After 5 years post-CABG, 98.3% of RA and 86.4% of SV (p = 0.04) grafts were found to be patent on follow-up angiography. RA was superior to SVG in terms of perfect patency and absolute patency. Limitations of this study included a small patient sample size, imbalance of clinical characteristics between the two groups, male predominance, and lack of angiographic follow-up in a significant number of patients. In a subsequent RCT published by Nasso and colleagues in 2009 [34], 815 patients with coronary artery disease and planned to undergo surgical revascularization were randomized into 4 groups, e.g., LIMA-LAD and RIMA for secondary vessel grafting, RIMA-LAD and LIMA for secondary grafting, LIMA-LAD and RA for secondary grafting, and LIMA-LAD and SVG for secondary grafting. Among the 4 surgical strategies, all the 3 approaches that utilized arterial conduits for secondary vessel grafting had comparable cardiac and overall survival at 40 months. However, adverse cardiac event-free survival was significantly lower in the single arterial conduit (with SVG used for secondary vessel grafting) group as compared to the double arterial bypass groups. This was irrespective of which second arterial graft (LIMA, RIMA, or RA) was used. Lack of consistent angiographic follow-up was the major limitation of this study.

Table 1 Characteristics of all randomized controlled trials comparing radial artery grafts with saphenous vein grafts
Full size table

In a randomized study of 60 elderly patients (age > 70 years old) scheduled to undergo off-pump CABG, clinical outcomes and graft patency (by computed tomography–CT angiography) were compared between patients receiving either RA or SV grafts [31]. After a 1-year follow-up, there was no significant difference in terms of graft patency or overall clinical outcomes (in-hospital and postoperative mortality) between the two groups. Assessment of graft patency using CT angiography was a novel approach as compared to traditional invasive coronary angiography. The study was limited by its short-term follow-up (mean of 8 months). In an RCT conducted by Petrovic and colleagues, 200 relatively young patients (mean age 57 years) with significant coronary artery disease were randomized to surgical revascularization with either LIMA and RA or LIMA and SVG conduits [30••]. After 8 years of follow-up, there was no significant difference in either the absolute survival or clinical outcomes (i.e., MI, need for percutaneous coronary intervention, or redo surgery). Within the cohort of patients that underwent coronary angiography for ischemia, RA patency rate was 92% as compared with 86% for SVG (p = 0.65). Despite the limitation of being a single-center study with a relatively small patient population, the trial was important in demonstrating equivalent clinical outcomes for both the arterial and venous grafts after a long-term follow-up (8 years).

Finally, the Radial Artery Patency Study (RAPS), which was the first multi-center trial to compare arterial and venous surgical conduits, randomized 510 patients from 9 centers in Canada to receive either RA or SV grafting at the time of CABG [32••]. The primary endpoint was angiographic evidence of functional graft occlusion (Lack of Thrombolysis in Myocardial Infarction flow grade 3); secondary endpoint was the rate of complete graft occlusion (Thrombolysis in Myocardial Infarction flow grade 0). After a mean follow-up of 7.7 years, the frequency of functional graft occlusion (12% vs. 19.7%, p = 0.03) and complete graft occlusion (8.9% vs. 18.6%, p = 0.002) was significantly lower in patients undergoing RA as compared to SV grafting, respectively. Recent meta-analysis of all RCTs to date (total of 1036 patients) showed RA to be the preferred conduit as compared to SVG. RA grafting was associated with lower rates of major adverse cardiac events (HR 0.67, p = 0.01) and higher patency rates at 5 years of follow-up (HR 0.44, p = < 0.001). Although no difference in mortality was found between RA and SV conduits, RA use was associated with lower incidence of MI (HR 0.72, p = 0.04) and rates of repeat revascularization (HR 0.50, p = < 0.001) [36].

CABG Conduit Guidelines and Contemporary Practice

Despite the US and European guidelines endorsing complete arterial revascularization in patients requiring CABG, SVG are still commonly used as second conduit in the USA as well as in Europe. As per 2011 American College of Cardiology (ACC)/American Heart Association (AHA) guidelines [37], complete arterial revascularization should be considered in patients < 60 years of age (class IIb). RA grafts are preferred when grafting left-sided arteries (left circumflex and LAD) with severe stenosis (> 70%) and right-sided artery (RCA) when there is critical stenosis (> 90%) (class IIb). According to the joint 2014 focused guideline update from Society of Thoracic Surgeons (STS)/American Association of Thoracic Surgeons/AHA/ACC, bilateral ITA should be considered in patients who are at low risk of sternal complications (class IIa, level of evidence–LOE B) [38]. As an adjunct to LIMA, RA graft should be considered as second conduit when target vessel has severe stenosis, especially in patients with higher risk of sternal complications (i.e., patient who are obese and/or have diabetes mellitus) (class IIa, LOE B) [38]. Use of pharmacological agents is also recommended in the pre-operative and peri-operative period to reduce the incidence of RA spasm (class IIb, LOE B). In 2014 guideline, European Society of Cardiology (ESC) and the European Association for Cardiothoracic Surgery (EACTS) recommended bilateral ITA grafting in patients younger than 70 years of age (class IIa, LOE B) [36]. RA grafting is a reasonable alternative for patients with significant coronary artery stenosis who have a contraindication to bilateral ITA harvesting (i.e., obese, older females, and diabetics) (class I, LOE B) [39]. Total arterial revascularization was recommended in patients with reasonable life expectancy (class IIa, LOE B) and in patients with poor vein quality independent of age (class I, LOE C).

Conclusion

In the current era, CABG remains to be the treatment modality of choice for patients with complex multi-vessel coronary artery disease, left main disease, and in patients with diabetes. Both arterial and venous conduits are part of a surgeon’s armamentarium at the time of surgery. Although saphenous vein grafting is the most commonly used conduit in addition to ITA, current data suggests that total arterial bypass (using RA conduit in addition to ITA) may be the better strategy. Both the American and European consensus guidelines advocate for the use of arterial over SV grafting for most patients. Nevertheless, in some patients (such as those who are older, diabetic, obese), SV grafting may be preferable. Therefore, the ultimate revascularization strategy for a given patient should be derived following a multidisciplinary heart team discussion as well as consideration of the individual’s risk factors and preferences.