Author + information
- Received July 24, 2001
- Revision received November 13, 2001
- Accepted November 30, 2001
- Published online February 20, 2002.
- Mahmoud H. Al-Bustami, MD*,*,*,
- M. Amrani, FETCS*,† (, )
- Adrian H. Chester, PhD†,
- Charles J. Ilsley, FRCP* and
- Magdi H. Yacoub, FRS†
- ↵*Reprint requests and correspondence:
Dr. M. Al-Bustami, Department of Cardiology, Harefield Hospital, Harefield, Middlesex, UB9 6JH, United Kingdom.
Objectives We sought to evaluate the in vivo endothelial function of the radial artery (RA) used as a coronary graft.
Background The RA is becoming a recognized alternative coronary bypass conduit. In vivo endothelial function is a possible predictor of long-term performance.
Methods Sixty consecutive patients underwent coronary artery bypass graft surgery (CABG); all received RA and left internal mammary artery (LIMA) grafts. Three weeks after CABG, 36 patients underwent angiography under basal conditions, during pacing and after intragraft injection of glyceryl trinitrate (GTN). Angiography was repeated at six months in 20 patients.
Results The estimated mean difference of 66 segments of the radial graft on the first QCA study was 0.170 mm (95% confidence interval [CI] 0.101 to 0.258, p < 0.001) between baseline and pacing, and 0.310 mm (CI 0.225 to 0.401, p < 0.001) between baseline and GTN. At six months, the differences between baseline and pacing and baseline and GTN were 0.112 mm (CI 0.062 to 0.162, p < 0.001) and 0.274 (CI 0.192 to 0.353, p < 0.001), respectively. The difference between baseline values at three weeks and six months was 0.416 mm (CI 0.236 to 0.603, p < 0.001). In the LIMA segments, the difference between baseline and pacing and baseline and GTN were 0.206 mm (CI 0.136 to 0.278, p < 0.001) and 0.304 mm (CI 0.213 to 0.396, p < 0.001), respectively. At six months, the differences between baseline and pacing and baseline and GTN were 0.098 mm (CI 0.014 to 0.173, p < 0.001) and 0.218 mm (CI 0.130 to 0.298, p < 0.001). The difference between baseline values at three weeks and six months was 0.061 mm (CI 0.064 to 0.176, p > 0.05).
Conclusions In vivo flow-mediated dilation of the RA is comparable to that of pedicled LIMA. The increased dilation both at baseline and after pacing at six months represents a time-related improvement in the vasomotor function of the RA, which could have implications for its performance as a coronary conduit.
Superiority of the internal mammary artery (IMA) used as a bypass graft, compared with the saphenous vein graft (SVG), is well established. Postoperative follow-up at 15 years showed significantly better survival and freedom from cardiac events in patients who received one or two IMA grafts (1). In patients who received a bilateral IMA, there were decreased risks of death, re-operation and angioplasty, compared with those who received a single IMA (2,3). In free IMA grafts, 100% patency at two weeks to six months was initially reported (4), and a four-year patency rate of 92% was reported by Loop et al. (5). Other arteries—namely, the right gastroepiploic and the inferior epigastric arteries—proved disappointing, with an 80% patency rate at one year (6,7). The inferior epigastric artery was hailed as an arterial conduit, with characteristics nearly identical to those of the IMA (8). However, more recently, Perrault et al. (9)reported their experience with 18 inferior epigastric artery grafts; a poor 57% patency rate was established before hospital discharge.
Use of the radial artery (RA) as a free graft is increasing. The initial work in 1973 yielded encouraging results, with 90% patency at one to 10 months by Carpentier et al. (10). Despite initial, promising results in subsequent studies, other investigators found that RA grafts became occluded at a higher rate than did simultaneously placed IMA and SVGs (11,12). Fisk et al. (12)advocated that the artery should not be used, and indeed it was abandoned.
Revival of the RA graft came about when grafts were shown to be widely patent 15 years after coronary artery bypass graft surgery (CABG). The philosophy regarding the role of the RA is changing rapidly, from a graft of last resort to the graft of second choice after the established left internal mammary artery (LIMA) pedicled graft. This increase in the use of the RA graft is occurring despite limited published data on the in vivo biologic function of the graft. The capacity of a conduit to dilate in response to an increased demand could be important in enhancing patency and performance. The aim of this study was to characterize the extent of flow-mediated dilation of the RA, both early and at mid-term.
Sixty consecutive patients who underwent CABG between January and July 1998 and received aorto-coronary RA (five of whom had the RA anastomosed proximally to a vein graft, creating a Y graft) and pedicled LIMA grafts were approached. Thirty-six patients (35 men; mean [±SD] age 60 ± 6.5 years) gave written, informed consent to undergo angiography. All patients were well and free of angina at the time of the investigation. Patients who declined to participate had similar characteristics.
All patients received diltiazem orally, 90 mg twice daily, which was discontinued after six months. Vasoactive medications, including nitrates, beta-blockers, calcium antagonists and angiotensin-converting enzyme inhibitors, were discontinued 48 h before angiography. The femoral artery and vein were percutaneously cannulated. A 6F bipolar pacing lead was positioned in the right atrium, and the capture threshold and Wenckebach rate were established. A nonionic contrast medium (Omnipaque) was used. Two diagnostic pictures of the left coronary artery and, similarly, two of the right coronary artery were taken. Two sequences of selective injection into the radial graft were then done; the first sequence was under the basal condition and the second was during the last 10 s of 2-min pacing at a sub-Wenckebach rate. Both sequences were done at the same settings of the X-ray tube, table and intensifier, as previously described (13). The second graft was handled the same way, and the third graft was studied in three sequences. Two minutes later, these two grafts had an intragraft injection of 200 μg of glyceryl trinitrate (GTN), and the final sequence of pictures was taken. Re-cannulation of the RA and the third graft (SVG or right internal mammary graft) was then done, followed by intragraft injection of GTN at 200 μg, followed 1 min later by the final picture sequence. The heart rate and blood pressure were recorded before and after each contrast injection.
Three RA grafts and two vein grafts were occluded. All LIMA grafts were patent. Confirmation of occlusion was made by using two oblique imaging views.
Quantitative coronary angiography (QCA)
Quantitative coronary angiography was done using Medical Imaging Systems (MEDIS; Nuenen, The Netherlands). A well-opacified, end-diastolic frame was captured. Each graft was divided visually into three segments, and an attempt was made at measuring all segments. When this was impossible, two segments were measured, at baseline, during pacing and after nitrate administration, using validated automated edge-detection algorithms (14). Mean segmental lumen diameters were obtained. Calibration was done with reference to the known diameter of the angiographic catheter 2 cm from its tip. Intraobserver and interobserver variabilities were measured and found to be insignificant.
The mean differences between baseline and post-interventions and between baseline values early and at mid-term are expressed in millimeters, with 95% confidence intervals (CIs) and p values. Because the values were not normally distributed, and because some patients had two segments measured and some had three, depending on the angiographic suitability for measurement, the Bootstrap method (15)was used to determine the mean difference, CI and p value between baseline, pacing and GTN both at three weeks and six months and between baseline values at three weeks and six months. Two thousand bootstrap samples were created to provide an estimate of the mean difference, to obtain CIs for the mean difference and to find a p value (for testing whether the mean difference was zero).
The blood pressure was similar at baseline, during pacing and after GTN injection. The heart rate was also similar at baseline and after GTN. In the second QCA study, this similarity held true, and there was no significant difference between the heart rate and blood pressure between the first and second QCA studies (Table 1).
Angiography at three weeks after CABG showed that the baseline diameter of the RA was smaller than that of the IMA (Tables 2 and 3). ⇓⇓Both arterial grafts became dilated in response to atrial pacing. The estimated mean differences (mm, with 95% CI and p value) of 66 segments of the radial graft on the first examination were 0.170 mm between baseline and pacing (CI 0.101 to 0.258, p < 0.001) and 0.310 mm between baseline and GTN (CI 0.225 to 0.401, p < 0.001). At six months, the mean differences of 60 segments between baseline and pacing and between baseline and GTN were 0.112 mm (CI 0.062 to 0.162, p < 0.001) and 0.274 mm (CI 0.192 to 0.353, p < 0.001), respectively. The estimated mean difference between baseline values at three weeks and six months was 0.416 mm (CI 0.236 to 0.603, p < 0.001). The estimated mean differences of 69 LIMA segments (at three weeks) between baseline and pacing and between baseline and GTN were 0.206 mm (CI 0.136 to 0.278, p < 0.001) and 0.304 mm (CI 0.213 to 0.396, p < 0.001), respectively. At six months, the mean differences of 51 segments between baseline and pacing and between baseline and GTN were 0.098 mm (CI 0.014 to 0.173, p < 0.001) and 0.218 mm (CI 0.130 to 0298, p < 0.001). The estimated mean difference between baseline values at three weeks and six months was 0.061 mm (CI 0.064 to 0.176, p > 0.05) (Tables 4 and 5). ⇓⇓
Three radial grafts were occluded (92% patency rate) and two vein grafts were also occluded. All LIMA grafts were patent.
This study has shown that the RA, used as a coronary bypass conduit, retains the ability to dilate in vivo in response to an increased heart rate induced by atrial pacing. This dilation was comparable to that of the pedicled LIMA and was evident both early (at 3 weeks) and at mid-term (6 to 8 months) after CABG. In addition, we have shown that the baseline size of the RA increases in the mid-term, although no such change was observed in the IMA.
The dilatory response of the conduit arteries to an augmentation of blood flow (flow-dependant dilation) was observed over 60 years ago (16). Atrial pacing increases the myocardial metabolic requirement, causing an increase in coronary blood flow (17,18). In normal epicardial coronary arteries, this increase in blood flow is brought about by an increase in blood flow velocity and by vasodilation (19). Local endothelial-dependant factors modulate this response, which requires intact endothelial functional integrity (20); the process is largely unrelated to innervation (21,22). Our group previously studied endothelial function in mammary artery grafts by using substance P and found no significant difference in the endothelium-dependent and endothelium-independent vasodilatory responses between free and pedicled mammary grafts (23). The central role of the endothelium in controlling vascular tone has only been appreciated since the discovery of the potent vasodilators prostacyclin and nitric oxide (NO) (24,25).
In addition to early postoperative vasospasm, the determinants of vasomotor dysfunction of coronary conduits can have an impact on the long-term success of the bypass grafts, as well. The IMAs have been shown to have superior long-term patency (26). The IMA produces NO, which induces vasodilation and inhibits platelet adhesion and the development of atherosclerosis (27). It is postulated that the greater release of NO by the endothelium in IMA grafts, compared with SVGs, contributes to the superior results with this conduit in bypass grafting (28,29).
Endothelial function of the RA versus LIMA
We have demonstrated that the in vivo function of the endothelium is comparable between the RA and IMA. It is known that the RA has a greater degree of contractility, compared with other conduits, in response to vasopressor agents such norepinephrine, serotonin, endothelin-1 and angiotensin II (30–32). It appears that the greater propensity of the RA toward vasospasm is therefore due to the muscular nature of the media, rather than specific differences in the protection offered by the endothelium (33,34).
This study demonstrates that the vasomotor tone of the RA, used as a coronary bypass conduit, is modulated by changes in blood flow demand caused by an increased heart rate.
The baseline diameter of the RA at three weeks was smaller than that of the IMA; this could possibly be due to differences in the recipient vessels, as the LIMA is grafted to the left anterior descending coronary artery (LAD), which carries a greater volume of blood than that in the circumflex coronary artery. At six months, the baseline size of the RA increased, whereas the size of the LIMA did not change. These findings indicate that the thick wall of the RA has no detrimental effect on the baseline size of the RA, and that the artery is capable of responding to a long-term persistent increase in demand by increasing its baseline size. Both arteries responded to increased blood flow, by significant vasodilation. This early response to pacing contrasts with the findings of Gurne et al. (35), who reported no vasodilation in response to atrial pacing in either the mammary or vein grafts, as well as coronary blood flow that was enhanced by increased velocity only. Their findings could be partially explained by the timing of angiography (mean 9 days), compared with that in our protocol (3 weeks).
Time-related changes in graft size and endothelial function.
Non–endothelial-dependent vasodilation was also demonstrated, both early and at mid-term. however, in contrast to the ra, which showed a significant increase in its mean baseline diameter at six to eight months, repeat angiography of the lima demonstrated a mean diameter similar to that at baseline. in the lima, this may be due to an early achievement of maximally relaxed vasomotor tone, which, apparently, took the ra longer to attain. the response to increased blood flow on the repeat angiogram was comparable between the ra and lima. these changes differed from the findings of gurne et al. (36), in their study of the inferior epigastric artery, where pacing induced vasodilation late after cabg, but not early postoperatively. similar vasomotor changes in the lima (pedicled and free) have been previously reported by hanet et al. (13), in their study showing the 6- to 82-month follow-up results after cabg. the latter study did not include early assessment, which could have shown a finding similar to ours. the limitation of the study is the fact that the ra was anastomosed to the circumflex or right coronary artery, and the lima to the lad. this, too, could have been a factor in the early occlusions. two of the ra occlusions were anastomosed to poor vessels (one was anastomosed to a vessel with mild stenosis), and these two coronary features now preclude use of the ra graft. the fact that there were no further occlusions at six months lends support to the mechanism of failure of the ra graft as being due to poor runoff or competitive native flow. a randomized study comparing the ra and lima, both anastomosed to the lad, is probably difficult to justify. the long-term patency will require angiographic analysis similar to that performed in landmark studies of ima grafts and svgs (1,37).
Conclusions and clinical implications
This study has shown that the in vivo endothelial function of the RA is retained both early and in the mid-term in a fashion similar to that of the IMA. In addition, the baseline dimension of the RA appears to increase with time. These findings encourage further evaluation and use of the RA graft.
- coronary artery bypass graft surgery
- confidence interval
- glyceryl trinitrate
- internal mammary artery
- left anterior descending coronary artery
- left internal mammary artery
- nitric oxide
- quantitative coronary angiography
- radial artery
- saphenous vein graft
- Received July 24, 2001.
- Revision received November 13, 2001.
- Accepted November 30, 2001.
- American College of Cardiology
- Cameron A.,
- Kemp H.J.,
- Green G.E.
- Reiber J.H.C.,
- Serruys P.W.,
- Koojiman C.J.,
- et al.
- Efron B.,
- Tibshirani R.J.
- Yoshida S.,
- Ganz W.,
- Donoso R.,
- Marcus H.S.,
- Swan H.J.
- Hanet C.,
- Pouleur H.,
- Harlow B.J.,
- Rousseau M.F.
- Nabel E.G.,
- Selwijn A.P.,
- Ganz P.
- Joannides R.,
- Haefeli W.E.,
- Linder L.,
- et al.
- Chardigny C.,
- Jebara V.A.,
- Acar C.,
- et al.
- Gurne O.,
- Chenu P.,
- Buche M.,
- et al.
- Gurne O.,
- Chenu P.,
- Buche M.,
- et al.