Carotid-subclavian bypass and the carotid steal phenomenon

UTTERWORTH E~NEMANN

Cardiovascular

Surgery, Vol. 3, No. 6, pp. 637-642, 1995

Copyright 0 1995 Iflvevier ScienceLtd Printed in Great Britain. i\ll rights reserved W67-210’?/95 $10.07 + 0.00

-----

---l___-~--

Carotid-subclavianbypassand the carotid steA phenomenon - .-.._. __..-_-.-----S. Hadjipetrou. R. S. A. Lord, J. A. Crozier, K. Sommerville and A. C. Meek SurgiccalProfesssorial Unit, St ¢ k Hospit& University of New South Wales,Sydney. Austraha

Twelve patients with patent carotid-subclavian bypass grafts were investigated by colour flow duplex imaging to look for a carotid steal. Centre-stream peak systolic flow velocities in the common carotid arteries distal to the take-off of the bypass graft were equivalent to those in the contralateral common carotid artery. On the side of the bypass, common carotid flow proximal to the graft was almost double that on the distal side. Steal phenomena were not identified in the carotid artery either with the arm at rest or during hyperaemia. However, in two instances arm hyperaemia caused reversal of flow in the ipsilateral vertebral artery, suggesting the induction of a subclavian steal. The method of centre-stream peak systolic velocity as an index of flow confirmed ttiat carotid-subclavian bypass can be safely carried out without depriving the cerebral circulation of the normal carotid inflow. Keywords: carotid steal, carotid-subclavian bypass, reactive hyperaemia, ultrasonography

Several techniques are available to treat occlusive atherosclerotic disease involving branches of the aortic arch but there is no firm agreement regarding the best option for subclavian lesions. Bypass procedures, percutaneous angioplasty, endarterectomy and transposition all have particular advantages and specific potential complications I-4 . The authors have favoured carotid-subclavian bypass when the donor carotid artery is free of significant proximal stenosis in the treatment of subclavian artery occlusion with or without evidence of subclavian steal. There is a potential reduction in blood flow in the donor carotid (carotid steal) by diverting blood down the carotid-subclavian bypass and so far this has not been empirically discounted in human?. Previous canine studies have shown that a steal develops from the carotid vascular bed only when a critical stenosis is present proximal to or at the site of carotid graft anastomosis ‘. These studies also indicated that the degree of stenosis in the donor carotid artery necessary to induce a carotid steal was inversely proportional to the graft flow rate. Nevertheless there have been no systematic clinical studies confirming that carotid-subclavian bypass does not compromise hemispheric circulation by reducing inflow to the ipsilateral carotid territory.

Correspondence to: Professor R. S. A. Lord

CARDIOVASCULAR SURGERY DECEMBER1995 VOL3 NO 6

The present study uses non-invasive carotid Doppler colour flow imaging to determine if a carotid steal phenomenon occurred in patients following carotid-subclavian bypass. In 12 patients with established carotid-subclavian bypass grafts, blood flaw was measured in the carotid artery proximal to and distal to the inflow anastomosis to investigate the possibility of carotid steal at rest or following hyperaemia.

Patients and rzmthods During a 20-yea.r period ending December 1992, 41 carotid-subclavian bypass procedures were carried out in the Surgical Professorial Unit, St Vincent’s Hospital, Sydney, Australia. Fourteen patients agreed to participate in the present study but two were excluded when their carotid-subclavian grafts were found to be occluded. The 12 patients with patent carotid-subclavian grafts form the basis of the present report and comprise four men and eight women, mean(s.d.) age 64.8(9.8) years. At the time of follow-up all the patients were symptom-free although before operation all suffered from vertebrobasilar ischaemia and some also experienced fatiguability and pain in the ipsilateral arm. All patients were assessed by inrra-arterial digital subtraction angiography which demonstrated that the proximal common carotid artery was free of haemodynamically significant stenosis and also confirmed that the graft was widsely patent. 637

Carotid-subclavian bypass and carotid steal: 5 Hadjipetrou et al.

The technique of carotid-subclavian bypass grafting has been described in detail elsewhere’. Proximally, the grafts were anastomosed end-to-side to the midcommon carotid artery. The distal anastomosis was located end-to-side to the second part of the subclavian artery beyond the vertebral orifice (Figure 1). Reversed saphenous vein taken from the groin was used for six grafts, five were 6 mm diameter polytetrafluoroethylene and one was 6mm Dacron. An Acuson scanner (Model 128X, 1986, Acuson Corporation, Mountain View, California, USA) was used for Doppler colour flow imaging of the large extracranial vessels and for assessing graft patency on

follow-up. The scanner combined real-time B-mode imaging with a S-MHz Doppler transducer. Bilateral Doppler measurements of centre-stream peak systolic velocity and vessel diameter were made in the common, vertebral and subclavian arteries. The graft and its proximal and distal anastomoses were also examined for stenosis or dilation. Retrograde vertebral flow was identified by a reversal of the velocity spectrum in the insonated lumen of the vertebral artery. Duplex scans were recorded on an industrial video tape recorder (VHS Panasonic Hi-Fi model AG-7350) and reviewed later for data collection. To assess the influence of the carotid-subclavian bypass on distal carotid flow, comparisons were made of the flow characteristics between: (1) the donor common carotid artery proximal to the graft (PCCA) and the common carotid artery distal to the, graft anastomosis (DCCA); (2) the PCCA and the contralateral common carotid artery (CCCA); and (3) the DCCA and the CCCA. The sites where measurements were taken are illustrated in Figure 2. The initial studies were carried out once on each patient by the same ultrasound technician. Blood velocity was recorded in m/s and vessel diameter in mm. A second sonographer confirmed the initial results and also carried out reactive hyperaemia tests on seven of the 12 patients to test the capacity of the donor carotid artery to provide adequate flow to the distal circulation. In this circumstance, flow through the graft would be expected to increase as a result of dilatation of arm vessels, thereby enhancing the potential to induce a carotid steal. Reactive hyperaemia was tested in the arm

BASILAR

Figure 1 Angiograms of a patent left common carotid to left subclavian vein bypass and b patent left common carotid artery to left subclavian polytetrafluoroethylene bypass. The arrows indicate the bypass graft

638

ARTERY

Figure 2 Anatomy of the extracranial arteries shown in the left anterior oblique projection and the sites where measurements were obtained. The basilar artery has been drawn out of scale for completeness. PCCA: proximal common carotid artery: DCCA: distal common carotid artery; CCCA: contralateral common carotid artery: LSCA: left subclavian artery

CARDIOVASCULAR SURGERY DECEMBER 1995 VOL 3 NO 6

Carotid-sub&vim

on the side of the patent graft. The response following hyperaemia was then compared with the results obtained in six normal unoperated volunteers. Reactive hyperaemia was induced by occluding the arterial inflow to the arm with a sphygmomanometer pressure cuff inflated to 20 mmHg above the systolic blood pressure. Arm exercise commenced when the cuff was inflated and ceased when the cuff was released 1 min later. The arm exercise involved alternating forced extension and flexion of the fingers at a rate of one per second. In each patient separate measurements of the PCCA, DCCA, vertebral and subclavian flow profiles were obtained following reactive hyperaemia with a rest period of 5 min between hyperaemic episodes. This meant that the maximum peak systolic velocity was recorded in each of the insonation sites with the arm at rest, while the cuff was inflated during the period of exercise, and after the cuff was released. Similar measurements were taken in the six normal control patients in each of whom duplex examination confirmed that there was no demonstrative atherosclerotic disease of the extracranial arteries. Statistical analysis was carried out using Student’s t test for non-paired and paired data as appropriate, for velocity and diameter among the sites from which readings were acquired. Values are reported as mean(s.e.m.).

bypassandcarotidsteav3

lfa~#IetrouetaL

a A.--

2,

PCkA

D&A

CdCA

b

Results In 12 patients with carotid-subclavian bypass comparison of flow velocity at rest showed significantly higher flow proximal to the graft anastomosis in the common carotid artery than distal to the anastomosis (PCCA = 1.62(0.08) m/s; DCCA = 0.94(0.06) m/s) (P < 0.01) (F@re 3). The proximal carotid velocity was almost double that recorded in the contralateral common carotid artery (CCCA = 0.86(0.07) m/s) (P < 0.01). The distal common carotid velocity did not significantly differ from that in the contralateral common carotid artery (P = 0.38). As flow is a function of velocity and vessel size, comparison of the arterial diameters showed no significant difference between proximal and distal measurements (PCCA = 6.7(0.4) mm; DCCA = 6.ljO.3) mm) (P = 0.27) and the contralateral common carotid arteries were found to have similar dimensions to the operated sides (CCCA = 6.5(0.24)mm) (P = 0.58) (Figure 3). Duplex colour flow examination of the common carotid and the subclavian arteries demonstrated normal non-turbulent flow. However, in each of the 12 patent grafts, the velocity profiles recorded by Doppler colour flow imaging were turbulent, which made velocity readings in the graft unreliable as an index of flow. Nonetheless, when the subclavian artery itself was insonated, a triphasic velocity waveform was observed similar to the pattern observed in the normal subclavian arterv (Figz4w 4). CARDIOVASCULAR SURGERY DECEMBER 199s VOL 3 NO 6

0

PCCA

DCCA

CCCA

Figure 3 a Peak systolic blood flow velocity. as measured by pulsed Doppler ultrasonography proximal to the graft, increases so that normal perfusion of the distal vascular bed is maintained. Blood flow vetocity (m/s: mean(s.d.)). *Significantly different to PCCAvetocity (P < O.Ol}. b Common carotid artery diameters (mm; mean(s.d.)) at the correspondinq sites where blood flow velocity was sampled

Figure 4 A triphasic waveform was observed in a the normal subclavian arteries and b the subclavian artery ipsilateral to the qraR

639

Carotid-subclavian bypass and carotid steal: S. Hadjipetrou et al.

Reactive hyperaemia The reactive hyperaemia tests insonating each of the PCCA, DCCA, CCCA and subclavian arteries were generally completed within 25 min, with 5 min between recordings. Peak response to hyperaemia was observed in the subclavian artery with maximal velocity seen approximately five heart beats following cuff release returning to resting levels within 1 min. No neurological symptoms or deficits were experienced during the tests apart from paraesthesia in the arm during the period of cuff inflation. In the six normal volunteers, the peak systolic velocities measured at rest were similar in the subclavian, common carotid and vertebral arteries when right

a

b

C

Q 8

.9 .8i

Figure B a Subclavian flow velocity in a group of six healthy subjects following induction of reactive hyperaemia. b Subctavian flow velocity in a subgroup of seven patients. c Mean vertebral flow velocity in the subgroup of seven patients. Contralateral artery: arm resting. 0 lpsilateral artery: arm resting; Wpsilateral artery arm cuff inflated: I ipsilateral artery: arm cuff released/reactive hyperaemia

n

640

and left sides were compared. The authors were unable to demonstrate a reduction in proximal subclavian velocity during brachial artery compression with cuff inflation (Figure 5). However, although no reduction in subclavian flow tias demonstrated with cuff inflation, there was a significant increase in subclavian flow velocity with hyperaemia from 1.07(0.08) m/s at rest to 1.44(0.11) m/s following cuff release (P < 0.01). No significant changes were observed in common carotid flow velocity during cuff inflation and when reactive hyperaemia was induced the flow velocities remained unchanged from the resting and cuff inflated values. In the group of seven patients with carotid-subclavian bypass challenged with reactive hyperaemia the following results were found. In keeping with the original study, the PCCA velocities ipsilateral to the graft were higher at rest (1.32(0.03) m/s) than those in the CCCA (0.81(0.1)m/s) and those in the DCCA (0.86(0.06) m/s) (Figure 5). Flow velocity in the DCCA was not significantly different from that in the CCCA. ‘There were no differences in the velocities recorded from the left and right vertebral or subclavian arteries with carotid-subclavian bypass. In contrast to the group of normal healthy subjects, on cuff inflation the subclavian artery flow velocity tended to decrease from 0.94(0.26) to 0.71(0.12) m/s, as did the vertebral artery velocity from 0.46(0.08) to 0.40(0.09) m/s although neither of these changes reached statistical significance. Velocities in the PCCA and DCCA remained unaltered with cuff inflation (PCCA = 1.39(0.14) and DCCA = 0.98(0.09) m/s). Following cuff release there was no detectable change in PCCA velocity measured at 1.41(0.22)m/s or DCCA at l.O(O.13) m/s, which suggested that there was no carotid steal with arm hyperaemia. In keeping with the normal group, a hyperaemic response was recorded in the subclavian artery where velocity rose to 1.02(0.16) m/s but failed to reach statistical significance when compared to the resting or cuff inflated values. Measurement of carotid artery blood velocity distal to the carotid/subclavian take-off was not affected by cuff inflation or hyperaemia and therefore the authors were unable to demonstrate a carotid steal with this technique. In all seven patients there was a paradoxical decrease in vertebral artery flow ipsilateral to the carotid-subclavian bypass following cuff inflation. In five there was a further reduction in vertebral flow in response to hyperaemia although it remained anterograde. In the two remaining patients, vertebral flow reversed and became retrograde. This reversal of vertebral flow persisted for over 30s after cuff release and then progressively resumed a normal anterograde pattern by about 1 min. In one of the two patients in whom vertebral flow reversed during arm hyperaemia, the resting waveform in the vertebral artery was also abnormal, being biphasic with a longer anterograde component alternating with a short phase of retrograde flow (Figure 6).

CARDIOVASCULAR SURGERY DECEMBER 199s VOL 3 NO 6

Carotid-subclavian bypassand carotidsteak S h’a@ipetrouet al.

F@re 6 Reversal of vertebral flow velocity was observed with carotid-subclavian bypass after reactive hyperaemia. a A normal vertebral flow waveform with healthy subject at rest b Resting vertebral flow waveform in a patient with carotid-subclavian bypass. Note small retrograde component in diastole. c Retrograde vertebral flow following’ reactive hyperaemia

Discussion Current angiographic methods and magnetic resonance imaging techniques can provide qualitative information on flow directions but cannot measure flow augmentation or reduction after revascularization procedures. To determine whether a steal develops after extra-anatomic bypass, other technology needs to be applied. Direct arterial flowmetry was previously used to investigate the carotid steal in canines but it is not appropriate for chronic human studies. Others, like positron emission tomography and single photon emission computed tomography can detect hemispheric hypoperfusion but cannot identify flow reduction in the carotid artery in the neck7. Doppler colour flow imaging was chosen to investigate the postulated carotid steal despite the limitation that this method does not actually quantitate flow but rather permits an examination of indirect indices of flow. Of several Doppler flow parameters which might be derived, the peak systolic velocity was selectedfor the present study. In each case the peak systolic flow velocity readings were taken from a centre-stream position where velocity oscillations more accurately reflect volumetric flow. Consistent measurements of blood velocity in the bypass grafts were not obtained as turbulent flow was encountered, probably as a consequence of the angle of take-off of the proximal anastomoCARDIOVASCULAR SURGERY DECEMBER 1BBS VOL 3 NO 6

sis of the carotid-subclavian grafts. Flow in the subclavian artery distal to the graft was nevertheless normal, consistent with the concept that the velocity profile within an artery is largely determined by the resistance of the distal vascular bed. The haemoclynamic results of this investigation support the observations of previous animal research6 that carotid-subclavian bypass in itself does not lead to a steal from the carotid artery. In the present study, distal carotid flow was maintained on the side of the carotid-subclavian bypass and found to be equivalent to that of the contralateral common carotid artery. This was confirmed at rest and after a period of reactive hyperaemia, which increased subclavian flow to the arm such that any potential steal might be enhanced. The authors were unable therefore to demonstrate a carotid steal beyond a functioning carotid-subclavian graft. Of significance in this investigation was the increase in velocity which was consistently recorded in the proximal common carotid artery, suggesting that the diameter of the vessel was sufficiently wide to maintain adequate cerebral and upper limb perfusion, even after hyperaemia. There may be a critical diameter below which this dual perfusion cannot be sustained so that if the common carotid artery is of small diameter or affected by stenotic disease,a cerebral steal may yet he identified. In other anatomical regions where extra-anatomic bypass may feasibly produce steal phenomena, it is recognized that a femorofemoral cross-over graft is unlikely to induce steal from the donor side if the inflow from the iliac vessel is widely patent anti normal. The finding that in two patients vertebral flow reversed during arm hyperaemia, despite :.Iwidely patent carotid-subclavian graft, has not been adequately explained. An increase in anterograde vertebral flow was expected with cuff inflation but this was nor shown in the patients examined. Similarly, a patent carotid-subclavian bypass would be expected to correct any tendency to reverse flow in the vertebral artery, even with induced hyperaemia. Reversed vertebral flow in itself is typically benign unless hind brain inflow is compromised by inadequate collaterals, especiallv an incomplete circle of Willis’. In a reporr ot 45 patients with severe subclavian stenosis, 64% demonstrated reversal of vertebral flow provoked by inducing reactive hyperaemia, although none of these patients was symptomatic’. This evidence, including that presented in the current study, confirms that in some patients vertebral flow becomes retrograde only when arm hyperaemia is induced and in these circumstances the subclavian steal causesno cerebral ischaemic symptoms. In the present six normal volunteers, a l-mm period of hand exercise during cuff occlusion induced a prompt and consistent in’crease in subclavian flow with hyperaemia induced by cuff deflation. While a similar change in velocity signals indicated augmented subclavian flow with with hyperaemia, in some of the patients carotid-subclavian grafts this was nor- SCE’IIuniformly. 641

Carotid-subclavian bypass and carotid steal: 5. Hadjipetrou et al.

It is not clear why some patients demonstrate detectable velocity responses to hyperaemia and others do not. While the inflow may be adequate under resting conditions, the proximal common carotid flow may be insufficient to maintain both cerebral and upper limb circulation during hyperaemia. In this case it is likely that cerebral perfusion is protected at the expense of flow through the bypass; however, the mechanism of this remains obscure. Alternatively, it may be explained by sub-optimal haemodynamic characteristics of the carotid-subclavian bypass, which results in flow resistance if perfusion is reduced and routing the greater portion of blood flow to the carotid circulation. The present study supports the following conclusions. First, carotid-subclavian bypass does not siphon blood away from the distal common carotid artery. Second, in patients with a patent graft, subclavian flow usually but not invariably increases when vasodilatation is induced in the upper limb by cuff occlusion and arm exercise. Third, that despite a functioning carotid-subclavian graft arm hyperaemia may lead to reduction or reversal of vertebral flow; however, the patients developed no cerebral symptoms.

642

References 1. 2. 3.

4.

5. 6. 7.

8. 9.

Lord RSA. Surgery of Occlusive Cerebrovascular Disease. St Louis: CV Mosby, 1986. Perler BA, Williams GM. Carotid-subclavian bypass - A decade of experience. ] Vast Surg 1990; 12: 716-23. Sterpetti AV, Schultz RD, Farina C, Feldhaus C, Feldhaus RJ. Subclavian artery revascularization: a comparison between carotid-subclavian artery bypass and subclavian-carotid transposition. Surgery 1989; iO61-624-32. Wilms G. Baert A. Dewale D. Vermenlen _I, I. Neuelsteen A.I SuvI R. Percutaneous transluminal angioplasty of the subclavian artery: early and late results. Cardiovasc Intervent Radio1 1987; 10: 123-8. Otis S, Rush M, Thomas M, Dilley R. Carotid steal syndrome following carotid subclavian bypass. ] Vast Surg 1984; 1: 649-52. Lord RSA, Ehrenfeld WK. Carotid-subclavian bypass: a haemodvnamic studv. Suraerv 1969: 66: 521-6. Lord RSA. Haemodynamir carotid insufficiency. In: Bernstein EF, Callow AD, Nicolaides AN, Shifrin EG, eds. Cerebral Revascularisation. London: Med-Orion, 1993: 39-50. Lord RSA, Adar R, Stein RL. Contribution of the circle of Willis to the subclavian steal syndrome. Circulation 1969; 40: 871-8. Bornstein NM, Norris JW. Subclavian steal: a harmless haemodynamic phenomenon? Lancet 1986; 2: 303-5.

Paper accepted 4 October 1994

CARDIOVASCULAR SURGERY DECEMBER 1995 VOL 3 NO 6