Endoscopic radial artery harvesting for coronary artery bypass grafting

This article summarizes the current research on endoscopic technique of radial artery harvesting as a graft for coronary artery bypass grafting (CABG) surgery. Based on the available data, we reviewed various grafts available for CABG. Radial artery as a graft in CABG surgery has recently gained popularity. We sought to investigate the impact of radial artery harvesting techniques on clinical outcomes. Endoscopic harvest approach was found to be feasible in all patients when performed by skillful surgeon while local arm complications were found to be infrequent. However, when compared to open approach for harvest, it takes longer but provides higher patient satisfaction and cosmetic result.


INTRODUCTION
Coronary artery bypass grafting (CABG) is considered one of the mainstays of myocardial revascularization particularly in patients with left main coronary artery disease, multi-vessel coronary artery disease, and diabetes [1,2] .Conduits available in CABG can be broadly classified into arterial and venous grafts.Saphenous vein (SV), Internal Thoracic Artery (ITA), Radial Artery (RA), Ulnar Artery (UA), and right gastroepiploic artery can be used as grafts [3,4] .Venous grafts have been found to be more prone to thrombosis, intimal hyperplasia, and atherosclerosis in comparison with arterial grafts [5] .Arterial grafts are thus more widely preferred over venous grafts [6] .
The RA has been suggested as a suitable conduit for CABG in the 70's [7] in virtue of advantages like uniform size and easy availability [8] .Its use was initially abandoned due to concerns of vasospasm, but progressively reconsidered in light of the positive long-term results described in the last decade.At least 6 randomized clinical trials and several observational studies have positively compared patency rates and outcomes of RA vs. RITA and saphenous vein graft (SVG), and very recent evidence from a large patient-level meta-analysis has declared the actual "renaissance" of the RA [9] .Bilateral internal thoracic artery (BITA) grafting is not widely used due to a higher incidence of sternal complications when compared to single internal thoracic artery grafting (SITA) (0.6% vs. 1.9%, 95% CI: 1.5-6.8)which necessitates additional measures to improve prognosis.Current recommendations are: Use of ITA in left anterior descending (LAD) bypass in patients where benefits greatly outweigh risks, i.e., when the procedure has to be performed.
An alternative to LITA should be used in patients where benefits outweigh risks, i.e., when additional studies with broad objectives are required.
BITA should be considered in cases where risk of sternal complications is minimal and benefits outweigh risks, i.e., when additional studies with broad objectives are required.
Risk of infection with BITA, should be reduced by skeletonizing the grafts, encouraging smoking cessation, having tight glycemic control, and ensuring adequate sternal stabilization [10] .
Both endoscopic and open techniques are available options for RA harvesting and carry different advantages and complications.This review will focus on RA and will examine the main features of this conduit along with the impact of the available harvesting techniques on clinical outcomes.

PREOPERATIVE CONSIDERATIONS
While there are no major absolute contraindications to the use of RA, some disadvantages include its increased tendency to spasm owing to a thick tunica media.It is also associated with higher risk of atherosclerosis and intimal hyperplasia compared to ITA [3] .Despite initial reports of lower patency rate of RA grafts in diabetic patients when compared to SVG at 1 year (89.28% were patent in RA vs. 97.05%patent in SVG) [11] , the more recent literature and the results of a large patient-level meta-analysis confirm the suitability of RA in diabetics and its theoretical advantage in terms of sternal wound infections when compared to BITA [9,12,13] .
Pre-operative suitability of the radial artery is most commonly assessed by the modified Allen's test which is used to determine the patency of the vessels supplying the hand.If the patient's hand flushes within 5-15 s after the examiner releases the occlusive pressure applied on both the RA and the UA, it is considered a positive modified Allen's test.If the hand takes longer than 15 s the test is considered negative.A positive modified Allen's test translates into good blood circulation in the forearm [14] .However, the reliability of this test has been repeatedly questioned.Jarvis et al. [15] recommended a Doppler ultrasound to be the gold standard.They determined the sensitivity and specificity of the modified Allen's test to be 54.5% and 91.7% respectively with a diagnostic accuracy of 81.7% at the conventional cut off value of 6 second.Doppler ultrasound, on the other hand had sensitivity and specificity of 100% and 27% respectively with a diagnostic accuracy of 52% [15,16] .Starnes et al. [17] raised concerns that the high false negative rate of the modified Allen's test could lead to unnecessary exclusion of some patients by placing some patients at an incorrectly high risk of digit ischemia.They recommended direct digit pressure measurement, which is simpler and more accurate to determine the adequacy of collateral circulation in the hand prior to CABG.
In 1998, Buxton et al. [18] were the first to describe the successful use of the ulnar artery as a conduit in a series of 8 patients.The idea of collateral circulation to the hand allowing the use of either the radial or ulnar artery for harvest was reiterated in a larger, more recent study of 25 patients described by Newcomb et al. [19] in 2006.Although, both studies conclude that routine use of the ulnar artery is not recommended due to its close proximity to the ulnar nerve with significant potential for resultant motor and sensory deficits in the hand along with the fact that the ulnar artery tends to be the dominant artery to the hand [18,19] .
Another factor to be considered before choosing the RA for graft is the previous use of the conduit for invasive diagnostic procedures, which exposes the artery to post trans-radial access (TRA) occlusion.Interventional cardiologists commonly use the radial artery as access during procedures like percutaneous cardiovascular intervention (PCI) and angiography [20] .Our group previously reported significant endothelial damage in the vessel post-TRA leading to reduced vasodilatory function of the vessel with no clear evidence of return of baseline function with time.Along with the biological dysfunction, the risk of RA thrombotic occlusion has been estimated at 7.7% at 1 day and 5.5% at > 7 days after the procedure in a recent large meta-analysis with specific clinical factors (i.e., age, diabetes, reduced artery size, female gender, peripheral vascular disease, smoking, low body weight and lack of statin use) and procedural factors (i.e., use of non-hydrophilic catheters, prolonged postprocedural high pressure compression, etc.) being described as mainly responsible for occurrence of RA occlusion [21] .Additionally, an association between TRA and graft occlusion has been described, leading to the recent recommendations on the use of non-punctured RA as graft for CABG and on the preservation of the RA during angiographic diagnostic procedures in surgical candidates [20,22] .In situations with limited graft options, it is recommended that the surgeon should perform Doppler ultrasound preoperatively to positively determine the patency and the diameter of the vessel.The use of the distal end of the artery should also be avoided [22] .It was found that the incidence of the RA graft not being suitable for use after findings of non-satisfactory collateral circulation pre-operatively was around 5% [23][24][25] .This does not take away from the versatility of the RA graft which can be used as a single free graft, as a Y graft, or as a sequential graft [21] .
Apart from the previously mentioned reasons for graft failure amongst venous and arterial grafts, another very important aspect to address is the degree of native artery stenosis when considering the RA for use as conduit.Tatoulis et al. [26] explained that arterial graft patency rate is proportional to native artery stenosis severity so the more severe the stenosis the higher the patency rate of the arterial conduit.Radial arteries are known to be more sensitive and show best results when the native artery has at least 80% stenosis [26] .This phenomenon is termed competitive flow and results in graft failure if the flow through the graft matches flow through the native artery after bypass.Vein grafts do not experience this phenomenon since they have much less resistance with a larger diameter [27] .

Radial artery graft harvest could be performed open or endoscopically [open radial artery harvest (ORAH)
or endoscopic radial artery harvest (ERAH)], preferably from the non-dominant arm [28] .The vessel might be harvested in skeletonized or pedicled fashion.Despite some advocating an advantage in conduit length and diameter when harvested in skeletonized fashion (P < 0.01) [29] , others point out the disadvantage of longer harvesting time and the risk for endothelial damage especially when the Harmonic scalpel is used [30] .The Harmonic scalpel uses ultrasonic energy for tissue dissection which is well known to cause increased release of nitric oxide leading to vasodilation and endothelial damage [31] .Considering the lack of clear evidence pointing towards a significant improvement in patency rate using the skeletonizing technique, this approach should be discouraged [16] .
The open technique requires the forearm to be incised in its entirety.The artery is accessed via a curvilinear incision along the edge of the brachioradialis muscle.The incision is initiated 1 cm distal to the elbow and extends up to 1 cm proximal to the wrist.After retraction of the superficial veins, the fascia is incised to expose the radial artery in the mid arm under the belly of brachioradialis muscle [Figure 1] [32] .Endoscopic approach of RA harvesting involves a longitudinal incision about 2-3 cm long proximal to the wrist crease.Dissection is done using bipolar scissors or bisector under direct endoscopic vision.Dissection is done as a pedicle with accompanying venae comitantes included.Carbon dioxide (CO 2 ) is insufflated at 10-12 mmHg to provide a working tunnel.Addition of CO 2 insufflation prevents spasms in the artery.This is followed by anterior dissection along the radial artery bilaterally up to the level of the antecubital fossa.Posterior dissection is then done up to the level of the radial plexus.Dissection is carried out laterally to achieve adequate branch length for sealing later.Fasciotomy enhances visualization and prevents the development of compartment syndrome.This is followed by branch division using a C-ring to stabilize the pedicle.The graft is then finally retrieved [33] .Navia et al. [34] found that ERAH requires more equipment in comparison with ORAH, hence it is more expensive than ORAH.Shapira et al. [35] showed that the cost of an endoscopic kit, including the disposable Harmonic shears, is $550.ERAH was also found to take a longer duration as compared to ORAH with an associated steep learning curve in inexperienced hands [36] .The reported learning curve for endoscopic harvest ranges from 5 to 30 cases [37] .Although, Kiaii et al. [38] found that the endoscopic approach requires a significantly lower harvest time (36.5 ± 9.4 min) compared to the open approach (57.7 ± 9.4 min) when performed by a surgeon adequately experienced in endoscopic harvest.

Patency rates
Radial artery as a graft for CABG was first introduced by Carpentier et al. [7] , in 1973, however, its use was abandoned because of concerns of spasm and intimal hyperplasia.In 1975, Curtis et al. [8] found that in the On the other hand, later in the early 90's, Acar et al. [39] demonstrated a 93.5% patency rate at 9 month followup concluding that the RA was still a reasonable alternative to complement IMA.Ikeda et al. [40] found the patency rates of RA graft to be 91% (24 out 26 grafts), which was comparable to ITA graft patency rate of 97% (139 out 143 grafts) in the mid post-operative period (27 ± 10 months).Subsequently, Desai et al. [41] also found that RA grafts were less likely to develop occlusion at one year when compared to saphenous venous grafts.RA grafts were found to have a higher patency rate at 5 year follow up when compared to SVG [42] .Angiographic studies have demonstrated a patency rate of 80%-90% at 7-10 year follow-up [43] .A more recent study reported an 84.4% patency rate at 20 years with a probability of graft failure at a similar time to LITA (19.0% ± 0.2% for LITA vs. 25.0%± 0.2% for the RA) [44] .
Moreover, a large patient-level meta-analysis, involving 1,036 patients from 6 trials, comparing RA with saphenous vein grafts showed a significant reduction in the incidence of adverse cardiac events like myocardial infarction, repeat revascularization, and death from cardiac causes with the use of the RA [9] .
Vasospasm was one of the major reasons which lead to a significant degree of reluctance in the use of the RA as a graft conduit in CABG procedures soon after its introduction.These concerns were related to the pronounced muscular profile of the RA wall as opposed to the more elastic wall of the ITA.Perioperative arterial spasm is reported at 0.43% in all CABG surgeries, and He et al. [45] have suggested this to be an underestimate owing to the chance of mild spasms going unreported.For this reason, a number of pharmacology prophylaxis protocols aimed at avoiding vasospasm have been developed [46] .Treatment includes the use of calcium channel blockers with or without long-acting nitrates in the postoperative management of these patients [46] or during surgery as per Reyes et al. [47] .Chanda et al. [48] recommend a combination of a calcium channel blocker like nifedipine, diltiazem, or verapamil with nitroglycerine for prevention of spasm.Organic nitrates like nitroglycerine have a short half-life but have a faster onset of action as compared to calcium channel blockers like nifedipine, verapamil, etc.These drugs are usually started immediately post operatively and continued for the first postoperative year.However, biological studies have demonstrated a progressive remodeling towards a more elastomuscular phenotype after implantation of the RA as a graft [49] , and clinical reports deny a benefit from antispastic pharmacological therapy [50] .

COMPARISON OF OUTCOMES BETWEEN ERAH AND ORAH
Harvesting technique may have an impact on postoperative outcomes and surgical complications.In early experiences in the late 90's, Royse et al. [51] reported that post-operative numbness and paresthesia of the hand were a commonly observed complication although there was a 98.9% resolution within 3-6 months after surgery.Neurological complications such as sensory loss occurred in 1.6%-18.1%,while motor complications such as diminished thumb strength were observed in 5.5% of patients in other large series [51][52][53] .
After the introduction of endoscopic harvesting, the incidence of these complications significantly decreased.In an early prospective study from Patel et al. [54] comparing ERAH with ORAH, major neurological complication restricting motor function post-operatively (8% vs. 1% patients at 1 month, P < 0.05), wound erythema, ecchymosis, mild numbness, or tingling were found to be significantly increased when using open approaches [Table 1].
More recently, the results of a propensity-matched study showed significantly lower hand ischemia (open 7.3% vs. endoscopic 0%, P = 0.007), by performing a modified Allen test prior to graft selection to confirm good collateral blood flow to the hand, and wound infection rate in the ERAH group, as well as better minor neurological outcomes (open 19.5% vs. endoscopic 3.6%, P < 0.001).Interestingly, freedom from cardiacrelated mortality (open 96.3% ± 2.1% vs. endoscopic 98.1% ± 1.8%, P = 0.448) as well as survival free from major cardiac and cerebrovascular adverse events (open 93.9% ± 2.6% vs. endoscopic 93% ± 3.4%, P = 0.996) at 5 year follow-up were similar among the groups.This suggests that ERAH could provide additional shortterm benefits in terms of improved cosmesis and reduced wound and neurologic complications without compromising the long-term clinical outcomes [55] .These results were confirmed by Burns et al. [56] by demonstrating non-inferiority of ERAH with regards to patency rates at 5 years when compared to ORAH (91.2% ERAH vs. 87.5% in ORAH, P = 0.705).
Lastly, ERAH has a steep learning curve considering the need to master manipulation of the conduit along with the endoscope which requires an advanced hand-eye coordination.Initial experiences do indeed describe harvest times longer than one hour [58][59][60] [Table 1], but other reports from the neurosurgical arena, in which the RA is also widely used as a conduit, show that the learning curve associated with the endoscope can be overcome by practice on cadavers [61] .

CONCLUSION
It is suggested that standard treatment for patients with multivessel disease is use of single or bilateral ITA along with additional arterial conduit [62] .When appropriate, the use of RA is recommended over SV graft since it is associated with better 5 year patency rates and improved patient longevity.The RA is preferred in patients at risk for sternal wound complications, such as diabetics who cannot tolerate BITA grafting [18,59] .