The clinical role of the cerebral collateral circulation in carotid occlusion

ORIGINAL ARTICLES The clinical role of the cerebral collateral circulation in carotid occlusion John W. Norris, MD, FRCP, Adam Krajewski, MD, and Natan M. Bornstein, MD, Toronto, Ontario, Canada The occurrence and severity ofischemic cerebral symptoms after carotid occlusion depends on the interdependency of cerebral collateral blood supply. Only those wifla the "fittest" collateral capacity survive this process of natural selection. Using the transcranial Doppler method in 55 patients with unilateral carotid occlusion, we tested the dependency of each cerebral hemisphere on the remaining patent carotid artery by digital carotid compression, and in 41 of these patients we also tested the carbon dioxide reactivity in each hemisphere. Both hemispheric dependency and carbon dioxide reactivity were compared to 15 healthy controls. Mean blood flow velocities in the middle cerebral artery were lower on the occluded side than on the patent side (p < 0.003). When the patent carotid artery was compressed middle cerebral artery blood flow velocities on the occluded side were mainly independent of the patent carotid artery, but on the patent side there was a high degree of dependency (p < 0.0001). Carbon dioxide reactivity did not differ between the hemispheres, but in hemispheres with total dependency, carbon dioxide reactivity was inversely proportional to the severity of stenosis (r = -0.63). Tests of cerebral collateral reserve in patients with unilateral carotid occlusion evaluated by carotid compression and cerebral carbon dioxide reactivity may discriminate between survivors and potential nonsurvivors before the patent carotid artery occludes. (J VAsc SUV.G 1990;12:113-8.) In 1859 Charles Darwin 1 postulated that the character and distribution of existing animal species were determined by the processes of natural selection--"survival of the fittest." It has also been known for many years that in patients with carotid occlusion, survival depends on the quality of residual collateral blood supply. 2 Our data resulting from this study in patients with unilateral carotid ocdusion suggest that the "fitness" of the remaining cerebral collaterals can be tested by measuring blood flow velocity (BFV) in each cerebral hemisphere and their response to carbon dioxide (CO2) inhalation and temporary carotid compression. Data on unselected patients with asymptomatic From the Stroke Research Unit, Sunnybrook Medical Centre~ University of Toronto. Dr. Krajewski is a Fellow of the Canadian Heart and Stroke Foundation. Dr. Bornstein's work is ftmded by the Medical Research Council of Canada., Dr. Bomstein is now at the Division of Neurology, Ichilov Hospital, Tel Aviv, Israel. Reprint requests: J. W. Norris, MD, Stroke Research Unit, Sunnyhrook Medical Centre, 2075 Bayview Ave,, Toronto, Ontario, Canada M4N 3M5. 24/1/21479 carotid stenosis indicate that surgical risks generally outweigh the spontaneous incidence of death or stroke. 3 A subgroup at high spontaneous risk may exist where this equation is reversed and surgery is justified. One such high risk group might be patients with unilateral carotid occlusion and a contralateral carotid stenosis where bilateral occlusion is imminent. In an attempt to identify this subgroup we determined the vascular dependency of each cerebral hemisphere in patients with unilateral carotid occlusion by compressing the contralateral, patent carotid artery, simultaneously measuring intracranial arterial velocities in each hemisphere by transcranial Doppler (TED).4 We also measured the vascular reactivity in each cerebral hemisphere to CO2 inhalation, since a compromised cerebral circulation has a decreased cerebrovascular reactivity, s All patients were referred to the carotid Doppler laboratory because of transient ischemic attacks or for the evaluation of asymptomatic neck bruits. In our laboratory, comparison of Doppler evaluation to angiography shows a sensitivity of 87%, specificity of 92%, and accuracy 90%. 6 113

114 Norris, Krajewski, and Bornstein lournal of VASCULAR SURG~Y GCO~80 Right MCA 1 Left MCA... I,,i~, I +,k f~+'x f:.f~!+. ~..::+ Occluded LI ~ ~-~ 50% Stenosis Fig. 1. In an 80-year-old man (G. C.), with unilateral carotid occlusion, compression of the patent carotid artery produced complete obliteration of ipsilateral MCA flow but had no effect on contralateral flow in the occluded side; i.e., the patent hemisphere was completely dependent on the patent carotid artery. MATERIAL AND METHODS Hemispheric dependency Before the compression test was performed the carotid arteries in the neck were evaluated by duplex scanner to ensure safety of carotid compression. Compression was not performed when plaque was seen in the common carotid artery or if asystole occurred on electrocardiogram (ECG) monitor when the common carotid artery was gently pressed to determine carotid sinus hypersensitivity. Signed informed consent was obtained in all cases. The patent carotid artery was then digitally compressed low in the neck for 5 to 10 heart beats, while middle cerebral artery (MCA) velocities were evaluated in each hemisphere by use of a TCD (EME, TC 2 to 64). 3 This is a range-gated pulsed Doppler ultrasound system with frequency analysis and continuously computed BFVs with a 2 MHz probe. If a clear recording of the MCA velocities were not obtained because of thickness of the temporal bone, the patient was excluded from the study. We compared longer periods of compression (up to 30 seconds) but found no difference in MCA velocities than with shorter periods of time. During compression cardiac rhythm was monitored by continuous ECG, blood pressure was monitored by cuff, and the efficacy of carotid compression was assessed with use of a 10 khz probe over the superficial temporal artery to ensure complete obliteration of temporal artery flow. 7 No patients had symptoms during or after carotid compression. The fall in MCA-BFV in each hemisphere on carotid compression was expressed as a percentage of resting MCA velocities. For instance, if MCA velocity fell from 100 cm/sec to 40 cm/sec on carotid compression, hemispheric dependency was repo~kd as 60%. In some patients compression of the pateat carotid artery, causing in effect temporary bilateral carotid occlusion, had no effect on MCA velocities in either hemisphere, indicating adequate, bilateral compensatory flow from the vertebrobasilar circulation. In others there was complete obliteration of MCA flow, indicating 100% dependency (Fig. 1). Carbon dioxide reactivity By use of an anesthetic mask, 5% CO2 and 95% air were administered through a nonreturn valve with the patient in the same supine position as for TCD evaluation. End expiratory CO2 samples assayed with an infrared analyzer were assumed equivalent to PACO2 values. Middle cerebral artery velocities were evaluated before COz inhalation and after end expiratory CO2 values stabilized. The percentage change in MCA velocities per 1 mm change in CO2 in the expired air was expressed as the CO2 reactivity index. For instance, if PAC~

Volume i2 Number 2 Au~4st 1990 Cerebral collateral circulation in carotid occlusion 115 Total 25~(~ Partial 25/~._~ i~ F ~ Total,.%)-~ t24 Partial ~ 26 None CCLUDED SIDE Fig. 2. Schematic diagram of dependency (total, partial, none). Patent hemispheres are mainly dependent on the remaining carotid artery, whereas occluded hemispheres are mainly independent. Table I. Degree of hemispheric dependency related to degree of carotid stenosis of the patent artery in 55 patients with unilateral carotid occusion % Stenosis patent Patent side* Occluded side carotid artery Total (100%) Partial (1%-99%) None (0%) Total (100%) Partial (1%-99%) None (0%) 75-100 5 5 5 0 6 9 0-75 20 20 0 5 18 17 Total 25 25 5 5 24 26 Dependency on the patent carotid artery is expressed as total, partial, or none. ~Patent side repers to hemispheres ipsilateral to the patent carotid artery. changed from 32 to 38 mm Hg, and the MCA ve- ~,Jcities changed from 100 to 118, CO2 index was 3.0. In the I5 healthy subjects this index was 3.4 _+ 2.4 (SD). RESULTS Compression tests Fifty-five consecutive patients with unilateral carotid occlusion were referred to our Doppler laboratory, 43 men, 12 women, mean age 68 _+ 7 years. Twenty-two had symptoms (16 transient ischemic attacks and 6 strokes), and 33 had asymptomatic neck bruits. Mean resting systolic peak MCA flow velocity on the patent side was 96 + 4 cm/sec, compared to the occluded side, 79 -+ 5 cm/sec (p < 0.003). In the patent hemispheres (i.e., ipsilateral to the patent carotid artery), a high degree of dependency was found on the ipsilateral (patent) carotid arteries (Fig. 2). In the occluded hemispheres most MCA-BFVs were largely independent, with perfu-.;on presumably from the vertebrobasilar system (p < 0.0001, analysis of variance [ANOVA]) (Table I). In 15 heal@ volunteers (mean age 27-+ i2 years), compression tests were performed on each carotid artery separately (Table II). Generally, each hemisphere was dependent on its ipsilateral carotid artery, quite unlike the degree of independency observed in the contralateral hemispheres in patients with unilateral occlusion. In two of the heal@ controls a paradoxical increase in MCA velocities occurred in the same cerebral hemisphere as the side of the compressed carotid artery, whereas in eight others, arterial velocities increased on the other side (Fig. 3). This was never seen in the patient group and presumably represents the effect of age on the collateral reserve. Carbon dioxide reactivity This procedure was performed in 41 of the 55 patients undergoing the compression test. In I4 patients it was not possible to perform the procedure

116 Norris, Krajewski, and Bornstein Journal of VASCULAR SURGEry Ipsilateral Contralateral 1 140 1 100 Table II. Dependency of MCA flow in 15 healthy subjects having carotid compression independently on each side Dependency Compressed side Contralteral side Total 9 0 Partial 19 0 None 0 22 Paradoxical* 2 8 *MCA flow increased when artery is compressed. 50 Table IlL CO2 reactivity (mean +- SD) in 41 patients with unilateral carotid occlusion and contralateral carotid stenosis Contralateral stenosis % No. Occluded hemisphere Patent hemsipher~ <50 23 4.5 + 3.9 4.3 + 1.7 51-75 8 2.5 + 3.7 2.6 + 1.3 >75 10 1.7 + 1.2 3.4 + 2.0 Total 41 3.2 _ 3.3 3.7 -+ 1.9 Controls 15 3.4 + 2.4 (30 hemispheres) v 0 Fig. 3. Effects of carotid compression on hemispheric dependency in 15 healthy controls. Ipsilateral, effect on dependency on the same hemisphere as the compressed artery. Contralateral, effect of dependency on opposite hemispheres as the compressed artery. Calibration is percent dependency. Note that most controls have total hemispheric dependency on the ipsilateral carotid artery (mark 100 on left-hand calibration), but that the compression has no effect on dependency of the contralateral hemisphere (0 mark on right-hand calibration). either because they refused consent or because it was contraindicated in view of intercurrent illness. Carbon dioxide reactivities were lower in occluded hemispheres when severe contralateral carotid stenosis was present, but this difference did not reach statistical significance (Table III). In hemispheres without dependency, no correlation was found between CO2 reactivity and severity ofstenosis in the patent carotid artery. However, in hemispheres with total dependency, CO2 reactivity fell with increased stenosis of the artery (r = -0.63) (Fig. 4). DISCUSSION Our data indicate that the effect of unilateral carotid occlusion on cerebral perfusion differs in each hemisphere. On the side of the occluded artery, hemi- spheric blood flow velocities are largely independent of the patent, contralateral carotid artery, having already established collateral flow from noncarotid (vertebrobasilar) sources. However, on the patent artery side, cerebral blood flow velocities are heavily dependent on the remaining ipsilateral carotid artery, much like the hemispheric streaming of ipsilateral carotid flow that we found in our normal controls. Patients with hemispheres largely dependent on the patent contralateral carotid artery presumably d~ or suffer strokes in many instances and so do nut enter our study data. The intact survivors, with the "fittest" collateral cerebral circulation account for the reversed appearance of carotid dependency observed between occluded and patent hemispheres (Fig. 2). The critical influence of vertebrobasilar flow is demonstrated by those patients with unilateral carotid occlusion and contralateral severe stenosis who have normal MCA-BFVs yet no significant carotid flow. Clearly, reversed flow through the orbit from external carotid arteries is largely insignificant, since compression of the occluded carotid arteries had no effect on MCA-BFV in 95% of patients and very little effect in the remaining 5%. Presumably external carotid arterial stenosis produces stroke, by embolic and not hemodynamic means. In patients with normal cerebral circulation (our control group), hemispheric flow is predominantly from the ipsilateral carotid artery (Table II, Fig. 3), and vertebrobasilar flow remains a potential collater~

Volume 12 Number 2 Au.mist 1990 Cerebral collateral circulation in carotid occlusion 117 10 8 m OO CO 2 Reactivity Index 6 4 "0 ~ 0 O :. 2. ~ r=-0.63 -- 0 I I I I I 20 40 60 80 1 O0 Stenosis (%) Fig. 4. CO2 reactivity in 16 patients with totally dependent hemispheres on the patent side. Percent stenosis refers to the patent artery. supply only pressed into use when the normal carotid channel is obstructed. Only one control subject had hemispheric flow unaltered by compressing either carotid artery, so both hemispheres were predominantly perfused by the vertebrobasilar circulation (Fig. 3). In this case there would be no cerebral hemodynamic effect of bilateral carotid occlusion. In two patients contralateral MCA velocities actually increased on ipsilateral compression, presumably re- ~-:cting overcompensation from the posterior cerebral artery. This paradoxical effect was not observed in the study group of older patients with atherosclerosis. Carbon dioxide reactivity did not vary with the severity of stenosis of the patent carotid artery in the hemispheres on the patent side. On the occluded side it fell progressively with the severity of stenosis, but not significantly. However, CO2 reactivity did fall significantly with progressively severe carotid stenosis in totally dependent hemispheres where the only source of cerebral perfusion was the patent carotid artery (Fig. 4). If collateral supply is adequate to maintain hemispheric flow, CO2 reactivity will remain unaffected by progressing stenosis or even occlusion of the remaining carotid artery, that is, reactivity to CO2 relates to the degree of carotid dependency, not whether the side is occluded or patent. This may explain the difference between our results ad others where carotid dependency was not taken into account, s,9 Ringelstein et al., 8 using TCD techniques similar to ours, found vasomotor reactivity to CO2 significantly lower in both the occluded and nonoccluded sides. They also found a poorer reactivity in symptomatic as compared to asymptomatic patients, inferring that atherosclerosis is more severe in the symptomatic group. In 73% of their patients vasomotor reactivity was either normal or only moderately reduced, which they- attributed to adequate collaterals. The reduced CO2 reactivity in cerebral hemispheres ipsilateral to carotid occlusion described by Bishop et al.9 implies that in general, collateral blood supply is inadequate. However, although reactivity was significantly reduced on the occluded side, the data divide clearly into two groups; one with normal reactivity and the other with very low reactivities. Presumably the low reactivity group would show high or total dependency. Our data are similar to those of Norrving et al) measuring cerebral blood flow with xenon 133 in patients with unilateral carotid occlusion. Carbon dioxide reactivity was normal where angiography revealed good collateral flow, but it was progressively impaired with contralateral carotid stenosis greater than 50%. The correlation between blood velocities and blood flow is only linear when there is little or no change in vessel caliber. Variations in MCA blood flow were investigated by Lindegaard et aly and

118 Norris, Krajewski, and Bornstein Journal of VASCULAR SURG,E~Y Benetos et al) 2 by comparing TCD velocities to ipsilateral carotid artery flow volumes. In patients without significant extracranial or intracranial arterial stenosis, brain volume perfused by the MCA remains constant, and velocities generally reflect changes in brain perfusion, since flow velocity and volume are linearly related. Carbon dioxide reactivity, like autoregulation, depends on changes in arteriolar perfusion and very little on the larger arteries that are evaluated by TCD. Therefore flow changes in these arteries should reflect those in the more distal vascular beds. The clinical effects of cerebral blood flow distal to a unilateral carotid artery occlusion therefore depend entirely on cerebral collateral capacity. Some patients suffer devastating stroke with unilateral occlusion, whereas others remain asymptomatic even with bilateral carotid occlusion. Usually, cerebral collateral flow must compensate since carotid occlusion is seldom accompanied by stroke. Most studies indicate a stroke risk after carotid occlusion of only 1% to 2% per year. r'~3'14 In the longest documented prospective follow-up, Furlan et al. 14 found an annual risk of only 3% in 138 patients followed for an average of 5 years. This relatively benign outcome indicates that by the time the artery has stenosed to the point of occlusion, patients have either experienced stroke or have adequate collaterals to remain asymptomatic. Although the present balance of opinion favors an embolic rather than hemodynamic pathogenesis for most ischemic strokes, recent pathologic studies of surgically removed carotid plaques emphasize the effect Of progressive stenosis on flow in producing symptoms? ~,16 In any case whether focal cerebral ischemia results from direct (embolic) occlusion of distal cerebral vessels or represents a secondary (hemodynamic) effect from more remote internal carotid artery occlusion, neuronal survival still depends on collateral flow. Darwin stated "I do not believe that variability is an inherent and necessary contingency... something must be attributed to the conditions of life. "1 Whether the capacity of cerebral collaterals to dictate survival is inherent or acquired remains unanswered. Dependency may slowly decrease as collaterals respond to progressive ischemia in the cerebral hemisphere subsequent to progressive atherosclerosis. Serial testing of hemispheric dependency as the patent carotid artery progresses to occlusion, in conjunction with observation of outcome in this potentially high-risk subgroup, may yield the answer. In those developing carotid independence, no further intervention is needed, whereas in the less fit group who remain largely dependent on the remaining carotid artery, surgical intervention could be justified., REFERENCES 1. Darwin C. The origin of the species. London: John Murray, 1859. 2. Fields WS, Lemak MA. Collateral circulation. A history of stroke--its recognition and treatment. New York: Oxford University Press, 1989:29-56. 3. Chambers BR, Norris JW. Outcome of patients with asymptomatic neck b~its. N Engl J Med 1986;315:860-5. 4. Aaslid R, Markwalder T-M, Nornes H. Noninvasive traf~ cranial Doppler ultrasound recording of flow velocity in bas~l cerebral arteries. J Neurosurg 1982;57:769-74. 5. Dyken ML, Campbell RL, Frayser R. Cerebral blood flow, oxygen utilisation, and vascular reactivity. Internal carotid artery complete occlusion versus incomplete occlusion with infarction. Neurology 1970;20:1127-32. 6. D'Alton JG, Norris JW. Carotid Doppler evaluation in cerebrovascular disease. Can Med Assoc J 1983;129:1184-4. 7. Bornstein NM, Norris JW. Benign outcome of carotid occlusion. Neurology 1989;39:6-8. 8. Ringelstein EB, Sievers C, Ecker S, Scheider PA, Otis SM. Noninvasive assessment of CO2 induced cerebral vasomotor response in normal individuals and patients with internal carotid artery occlusions. Stroke 1988;19:963-9. 9. Bishop CC, Powell S, Insall M, Rutt D, Brose NL. Effect of internal carotid artery occlusion on middle cerebral artery blood flow at rest and in response to hypercapnia. Lancet 1986;710-2. 10. Norrving B, Nilsson B, Risbert J. rcbf in patients with carotid occlusion. Resting and hypercapnic flow related to collateral pattern. Stroke 1982;13:155-62. 11. Lindegaard KF, Lundar T, Wiberg J, Sjoberg D, Aaslid k;' Nornes H. Variations in middle cerebral artery blood flow investigated with noninvasive transcranial blood velocity measurements. Stroke 1987;18:1025-30. 12. Benetos A, Simon A, Levenson J, Lagneau P, Bouthier J, Safar M. Pulsed Doppler: an evaluation of diameter, blood velocity and blood flow of the common carotid artery in patients with isolated unilateral stenosis of the internal carotid artery. Stroke 1985;16:969-72. 13. Cote R, Caron J-L. Management of carotid artery occlusion. Stroke 1989;20:123-6. 14. Furlan A1, Whisnant JP, Bates AL. Long-term prognosis after carotid artery occlusion. Neurology 1980;39:986-8. 15. Fisher CM, Ojemann RG. A clinico-pathologic study of carotid endarterectomy plaques. Rev Neurol (Paris) 1986; 142: 6-7, 573-589. 16. Fisher CM, Ojemann RG. A clinico-pathologic study of carotid endarterectomy plaques. Rev Neurol (Paris) 1986;142: 573-89: