The significant findings of this study are as follows. (1) In the early stage after the operation, skeletonized RA grafts maintained larger diameter in compared to non-skeletonized grafts. Unlike non-skeletonized grafts, skeletonized grafts did not show remodeling of the graft diameter. (2) In the mid-term follow-up CT angiogram, the skeletonized RA graft changed its diameter. There was no difference in this remodeling capacity between the skeletonized and the non-skeletonized graft.

Our first finding is consistent with several past reports. Yamaguchi et al. [12] reported that on angiography at an early postoperative stage, the mean diameter of the RA was wider and spasm or stenosis was seen less often in skeletonized grafts. Although the underlining mechanism for larger diameter and less vasoreactivity of skeletonized RA remains unknown, there are several possible explanations. First explanation relates to the use of an ultrasonic scalpel as a harvesting tool. Erkut et al. [13] and Ronan et al. [14] reported that ultrasonic dissection of RA has a positive effect on endothelial preservation and is associated with increased free blood flow through the graft. Maruo et al. [15] reported that sonication on canine ITA with an ultrasonic scalpel induces vasorelaxation almost completely by endothelial nitric oxide (NO) and prostacyclin release. In our study, we harvested the RA either as a skeletonized or a non-skeletonized graft, but used an ultrasonic scalpel in both groups. However, the sonication may have more directly and effectively stimulated endothelium in the skeletonized artery as it lacks surrounding tissue outside the adventitia. Second, RA is known to consist of relatively thick media packed with smooth muscle cells [5]. RA smooth muscle cells have higher receptor-mediated contractility in compared to ITA. RA is known to be an α-adrenergic receptor dominant artery with weak β-adrenergic receptor function [16]. RA exhibits greater contraction to several chemical mediators such as serotonin, thromboxane A2, and norepinephrine than ITA [16]. Skeletonization may influence the sensitivity of the receptor of the RA or release of mediator from the endothelium. Further physiological and pharmacological examination is required to elucidate the effect of skeletonization of RA graft.

It has been well known that arterial grafts change its diameter depending on flow requirement [17]. The present study demonstrated first time that this remodeling capacity differs depending on the harvesting technique. Skeletonization impaired the remodeling capacity in the early phase but that is regained in the later phase. The remodeling response largely depends on endothelial function that tries to maintain shear stress in the vessel constant. NO released from endothelium plays a crucial role in the remodeling of vascular diameter depending on shear stress [18]. Fukui and colleagues [19] reported that a non-skeletonized RA graft adjusts its caliber to suit the flow requirement, which eventually normalizes shear stress at the endothelial interface and allows the conduit to deliver an appropriate flow. Low shear stress induces the attachment of endothelial cells and leukocytes, resulting in thrombus, intimal hyperplasia, and subsequent graft failure [20]. If the postoperative diameter of the RA did not reduce to match the flow requirement, shear stress would be reduced as a result of the decreased flow velocity, which would eventually result in graft failure. In our study, the reduced vasoreactive capacity to the blood flow of the skeletonized radial artery graft seems to be recovered 1 year after the operation. The recovered vasoreactivity may contribute to avoid graft failure caused by the lack of the caliber change in response to the blood flow, although it is hard to justify the speculation from our study.

There are several points that could have influenced our results. First, RA straight composite graft with ITA can behave differently from that used as aortocoronary bypass. Composite grafting allows an aorta no touch technique when combined with off pump CABG, which has major advantage of decreasing perioperative stroke. On the contrary, decreased drive pressure in the composite graft compared to direct aorta anastomosis could compromise the flow through the graft [21, 22]. The opposite argument relates to the fact that ITA has increased capacity to release NO compared to other grafts [23], and thus may have rather beneficial impact on RA graft, which was anastomosed at its downstream. Currently, advantage of the straight composite graft is still unclear.

Second, multiple sequential grafting may also affect the fate of arterial grafts. There are conflicting reports on graft patency between sequential and single RA grafting. Nakajima et al. [24] reported that graft patency and competitive flow of sequential RA grafts is influenced by the severity of coronary artery stenosis in the most distal target artery. We preferred to use multiple sequential grafting to achieve complete arterial revascularization. This grafting strategy may have changed reaction of RA grafts.

Third, it has been well known that arterial grafts change its diameter depending on severity of target vessel stenosis [25]. In our study, severity of the target vessel stenosis was comparable between each group. However, visual assessment of the percentage of stenosis at angiography does not necessarily reflect accurate flow disturbance. Current studies demonstrated that fractional flow reserve (FFR) is more predictive of graft patency [26]. Our study does not include the assessment of FFR, thus flow requirement or flow competition of target vessel could have been different in each group.

This study has several limitations. First, the number of the patients in each cohort was small and selection of each harvesting methods was not randomized. Although the patient backgrounds were not significantly different between the groups, there was possible selection bias. Second, we could not assess several patients 1 year after the operation. Therefore, we emphasize the need for a large, randomized study.

In conclusion, at an early stage after the operation, the RA graft skeletonized with an ultrasonic scalpel maintains larger diameter irrespective of graft flow. It may contribute to avoid graft failure or flow competition in early postoperative period, when the grafts are especially susceptible to spastic stimuli. However, 1 year after the operation, luminal diameter of the skeletonized radial artery graft reduces possibly as physiological adaptation to graft flow. Further studies will be needed to confirm whether those characteristics of skeletonized radial artery graft result in any beneficial short- and long-term graft function.