Acta Neurochir (2009) 151:1099–1105 DOI 10.



How often does an incomplete circle of Willis predispose to cerebral ischemia during closure of carotid artery? Postmortem and clinical imaging studies
Hannu Manninen & Kimmo Mäkinen & Ritva Vanninen & Antti Ronkainen & Harri Tulla

Received: 2 February 2009 / Accepted: 4 July 2009 / Published online: 1 August 2009 # Springer-Verlag 2009

Abstract Purpose To evaluate the prevalence of anatomical variations in the circle of Willis predisposing to cerebral ischemia during intraoperative closure of a carotid artery. Materials Anatomy of the cerebral arteries of 92 deceased was assessed by angiography and permanent silicone casts. Cerebral ischemia during closure of a carotid artery with patent contralateral internal carotid artery (ICA) was considered possible in cases of simultaneous nonfunctioning anterior communicating artery (diameter <0.5 mm) and ipsilateral posterior communicating artery (PComA) (diameter <0.5 mm or fetal type posterior cerebral artery). In cases of contralateral ICA occlusion, cerebral ischemia was considered possible if ipsilateral PComA was nonfunctioning.

Results Cerebral ischemia during closure of the right or left carotid artery with patent contralateral ICA was estimated to be possible in 16 (17.4%) and 13 (14.1%) cases. In cases of occluded contralateral ICA, the corresponding numbers were 55 (59.8%) and 49 (53.3%). A review of magnetic resonance and catheter angiographies also identified other variants of the circle of Willis with increased risk. Conclusions Incomplete circle of Willis predisposes approximately one-sixth of individuals to cerebral ischemia during transient closure of carotid artery but the risk is more than threefold in case of contralateral ICA occlusion. Keywords Carotid surgery . Cerebral circulation Cerebral complications . Collateral blood flow

H. Manninen : R. Vanninen Diagnostic Medical Imaging Centre, Department of Radiology, Kuopio University Hospital and Kuopio University, Puijonlaaksontie 2, 70210 Kuopio, Finland K. Mäkinen : H. Tulla Department of Surgery, Kuopio University Hospital and Kuopio University, Puijonlaaksontie 2, 70210 Kuopio, Finland A. Ronkainen Department of Neurosurgery, Kuopio University Hospital and Kuopio University, Puijonlaaksontie 2, 70210 Kuopio, Finland H. Manninen (*) Department of Clinical Radiology, Kuopio University Hospital, Puijonlaaksontie 2, FI-70210 Kuopio, Finland e-mail:

Introduction The benefit and justification of carotid artery surgery is based on low procedure-related complication rate. Currently, surgery is thought to be beneficial if the combined stroke-death rate does not exceed 3% for treated asymptomatic lesions and 6% for symptomatic ones. Risks during surgery are related thromboembolic complications, hypoperfusion during carotid clamping, acute problems of the endarterectomy site and intracerebral hemorrhage. Although most of the perioperative strokes are the result of microemboli originating from the operation site [12], approximately 15–21% of ischemic strokes during surgery are thought to be due to hypoperfusion [18]. Several attempts have been made in order to improve safety and brain protection. These include routine and selective shunting during artery cross-clamping, carotid artery stump pressure measurement, intraoperative hypertension induction, local anesthesia, use of somatosensory

These studies also give reliable data if there are multiple lesions and if lesion morphology or site can predict technical difficulties in intraluminal shunt placement. nonfunctioning circle. Preoperative imaging methods like conventional angiography. The criteria for a nonfunctioning anterior communicating artery (AComA) were either a diameter of <0. that it carries a potential risk of intimal damage and embolization [7. The aim of the present study was to evaluate the prevalence of anatomical variations of the circle of Willis predisposing to cerebral ischemia during carotid artery closure. Analysis of the postmortem casts and angiographies The existence and visibility of the arteries in the anatomical casts and cast angiography were assessed by two experienced radiologists. A detailed description of the demographic data of the individuals is presented elsewhere [15]. in spite of these facts. Briefly. in cases with an incomplete. Stereotopic images (at 5° projection difference) were taken at frontal direction on high resolution film. Materials and methods The autopsy material of 92 individuals (mean age 59 years. liquefied silicone rubber is radiopaque.. thus making an assessment of collateral flow possible. even high quality randomized studies have been carried out in which the preoperative imaging and determination of stenosis were made only by carotid ultrasound examination [6]. magnetic resonance (MR) angiography and computed tomographic (CT) angiography can give reliable data about intracranial circulation. After the solidification of the silicone rubber anterior and lateral X-rays were taken. The study was approved by the ethical committee of our hospital but the ethics committee waivered the need for . During interpretation. ultrasound examination neglects the assessment of comprehensive cerebrovascular anatomy. Arterial diameters were measured with a high precision digimetric calibre device. Although collateral flow through the circle of Willis is usually sufficient to prevent cerebral ischemia during carotid clamping.e. In the case of contralateral internal carotid artery (ICA) occlusion. 19]. However. three illustrative cases of different types of anatomical variants in clinical catheter angiographies. the risk is increased [10. This made the cervical and cerebral arteries visible. The deceased’s aortic arch and the arteries branching from it (brachiocephalic trunk. Angiography Because lead oxide is used as the contrast medium. The anatomy of the cerebral arteries was assessed postmortem by angiography and permanent casts. MR and CT angiographies are presented to demonstrate other variants of the circle of Willis with increased risk. evoked potentials. The study is based on the data received from postmortem permanent casts and angiography of the cerebral arteries. autopsies dictated by law.5 mm or a persistent fetal type of posterior cerebral artery (PCA) supply associated with an absent or hypoplastic P1 segment. 20]. the radiographies were placed on the X-ray board next to each other and examined with a specially designed binoculars that created a threedimensional view of the cerebral arteries. the risk for cerebral ischemia was considered to be possible if ipsilateral PComA was nonfunctioning. Then right and left subclavian arteries were exposed and clamped and the mixture was led into the cerebral arteries using a portable perfusion device with a physiological pressure of 110–140 mmHg. death during surgical intervention. Consequently. Intraluminal shunt placement can also be technically demanding or even impossible in cases of high lesion or very narrow vessel. Some authors preferred routine use of a shunt [4. A plastic Y-shaped tube was connected so that one arm of the tube was fastened to the brachiocephalic trunk and the other to left carotid artery. range 14–90 years) was collected as a part of normal forensic medicine autopsies (i. which were made by experienced autopsy technicians. controversy still persists as to the optimal method for doing carotid surgery safely and how to avoid hypoperfusion. it is also known.5 mm or a hypoplastic/aplastic A1 segment of anterior cerebral artery (ACA) by the same diameter criterion. left common carotid artery and left subclavian artery) were exposed for a range of 3 cm. 5]. the permanent casts of cerebral arteries were performed before the autopsy and 210 g of lead oxide (Pb3O4) was used per 1 kg of liquid silicone rubber. consent from the patient’s next of kin because autopsy was dictated by law. The criteria for a nonfunctioning posterior communicating artery (PComA) were either a diameter of <0. Furthermore. Hannu et al. especially focusing on the anatomy of the circle of Willis.1100 M. electroencephalography. transcranial Doppler. and brain oxygen saturation measurement. such as cases of accidental death. Permanent casts The exact description of manufacturing and radiographic imaging of the casts is given elsewhere [16. In addition to the fact that estimation of the degree of stenosis is not without controversy. However. Cerebral ischemia during perioperative closure of carotid artery with patent contralateral ICA was considered possible in case of simultaneous nonfunctioning AComA and ipsilateral PComA. or suspicion of crime). The brain was removed and placed on a plate with the base upwards. 17].

Clinical MR and CT angiographies Clinical MR angiography. While a complete circle of Willis was registered in only one-fifth the individuals. cerebral ischemia during closure of the right or left carotid artery with patent contralateral ICA was estimated to be possible in 16 (17.missing .fetal type PCA Nonfunctioning left PComA .8.5 mm) .missing .5 mm) Nonfunctioning right PComA . a totally undeveloped circle of Willis was registered in about one-eighth of angiographies and almost every fourth had a nonfunctioning AComA and more than half had nonfunctioning right or left PComA. 2 a case with missing AComA and PComAs.missing .1%) cases.3% Fig. 3.9% 59. 4. Evaluation of these cases revealed several additional types of vascular variations or anomalies predisposing the patient to cerebral ischemia during carotid artery closure. regardless of whether the contralateral ICA is patent or occluded due to aplasia of the AComA and PComAs .fetal type PCA Number 19 11 22 21 1 55 40 5 10 49 40 4 5 Percentage 20.4%) and 13 (14.5 mm) . 1 Cerebral cast angiography demonstrating a complete circle of Willis Table 1 Frequency of anatomical variants of the circle of Willis in cast angiographies (n=92) Anatomy Complete Totally undeveloped Nonfunctioning AComA . the corresponding numbers for the right carotid artery was 55 (59.How often does an incomplete circle of Willis predispose to cerebral ischemia during closure of carotid artery? 1101 Results Anatomical study Table 1 summarizes the frequency of anatomical variants of the circle of Willis in cast angiography.3%). Anatomical cast angiography in Fig. respectively.hypoplastic (<0. In case of occluded contralateral ICA.0% 23.7% 12. Three illustrative cases are shown in Figs. By the above-mentioned criteria.8% 53. CT angiography and digital subtraction angiography studies demonstrating different types of anatomical variations and anomalies of the cervicocranial arteries have been prospectively collected into a separate teaching file in our hospital.hypoplastic (<0. These examples show the wide variability of vascular anatomy of the cerebral circulation Fig.hypoplastic (<0. 2 Cerebral cast angiography.7%) and for the left carotid artery 49 (53. and 5. Possible risk for cerebral ischemia during carotid closure on both right and left sides. 1 illustrates a case with a complete circle of Willis and Fig.

more recent studies imply that the lowest limit for a functional PComA is as low as 0. Frontal projection of the left carotid angiography demonstrates a missing A1 segment of the ACA (a). and advocate these noninvasive imaging modalities with a capability to demonstrate even complicated arterial communications and collateral pathways in this anatomical area. 3]. 3 Clinical catheter angiography of a 40-year-old woman. We found no cases with absent A1 segment of ACA in our postmortem material. although this variation is fairly common and was easily found in our clinical patient files (Fig. [11] reported an entirely complete circle of Willis in 42% and complete posterior circulation was seen in 52% of cases. The corresponding numbers. A complete configuration of its posterior part was present in 48. we found a complete. Discussion Macchi et al. According to another study 1. We selected to use a 0. The relatively small sample size makes our estimate of anomalies subject to large variation. Carotid angiographies in a lateral projection show missing PComAs on both the left (c) and right sides (d).5%. functional circle of Willis in 21% of cases. 3). The anatomy exposes the patient to possible risk for cerebral ischemia during carotid closure on both sides both in case of patent and occluded contralateral ICA (MCA middle cerebral artery) M. Although a diameter of 1 mm has been used in anatomical studies as a criterion for hypoplasia for the collaterals in the circle of Willis [1. There is some evidence also that patients who .5-mm cut-off value. based on transcranial color-coded duplex ultrasonography and carotid compression test were 33% and 57% among patients who had suffered ischemic stroke in the anterior circulation [18]. the PComA may be hypoplastic or absent on one or both sides of the brain in about 25–30% of patients referred to carotid endarterectomy.1102 Fig. [14] demonstrated by MR angiography that the circle of Willis has a complete configuration in 47% of subjects. however. In healthy volunteers studied by MR angiography. both pericallosal arteries (PericA) are filling from the right (b). By this criterion.6 mm: a transcranial Doppler US revealed that the cross-flow through posterior circulation was insufficient in 45% of individuals in an atherosclerotic population [9].4–0. while a nonfunctioning AComA was registered in 24% and PComA in 53–60% of individuals. Krabbe-Hartkamp et al. It is to be noted also that in an elderly population undergoing carotid artery surgery the percentage of risk patients is probably higher due to adjunctive atherosclerotic obstructions. Hannu et al.

Closure of the right carotid artery would endanger ipsilateral anterior cerebrovascular territory for ischemia and closure of the left carotid artery also the posterior brain territory for ischemia 1103 suffer ischemic stroke in the anterior circulation have even higher incidence of collateral deficient circle of Willis than those with atheroscrotic disease without ischemic cerebrovascular disease [8]. in patients that had neither an AComA nor a PComA 83% showed signs of cerebral ischemia. However. However. such as MR and CT angiography to predict cerebral ischemia during carotid cross-clamping and selecting indications for shunting. TOF MR angiography has some limitations.g. Lee et al. [13] found transient ischemia during clamping in 11% (11/96) and 57% (12/21) in the corresponding groups of patients. d) since the distal basilar artery (BA) is hypoplastic (d). Interesting observations have been made by the use noninvasive imaging techniques. circle of Willis morphology could not predict the development of intraoperative ischemia. in patients without contralateral ICA occlusion. b). especially with respect to spatial resolution. an incomplete circle of Willis would have been predisposed to cerebral ischemia in 14. but modern contrast-enhanced MR angiography and CT angiography with multiple-row detector equipment are probably better suited for evaluation of the circle of Willis. However. e. necessitating shunting. On the left side. supplying the main portion of the posterior cerebral territory (c. Interestingly.8% if the contralateral ICA is occluded. In the present study population. demonstrating normal anatomy of the right carotid artery but missing AComA and PComA (a.3–59. Kim et al. This is well in line with clinical studies. Bagan et al. [10] reported that shunting was needed only in 7% (4/55) of cases when an AComA or PComA was patent according to four-vessel angiogram. one may speculate that the collaterals are not fully dilated since there is no ongoing ischemia and the . one may argue that even in case of chronically occluded contralateral ICA and nonfunctioning ipsilateral PComA.1– 17. [2] concluded that absence of visible collaterals of circle of Willis in three-dimensional phase-contrast MRI is significantly predictive of early ischemia and an indication for intraluminal shunting. On the other hand. In another study [13]. a primitive trigeminal artery arises from the ICA. the patent contralateral PComA may in some cases provide sufficient blood flow to both anterior hemispheres if AComA is patent. 4 Clinical catheter angiography of a 42-year-old man..4% of individuals if the contralateral ICA is patent and in 53.How often does an incomplete circle of Willis predispose to cerebral ischemia during closure of carotid artery? Fig. the incompleteness of the posterior part of the circle of Willis in three-dimensional time-of-flight (TOF) MR angiography was found to associate with a significant risk of ischemia during vascular clamping of ICA.

but the risk is more than threefold in case of contralateral ICA occlusion. for example. Hannu et al. prediction of intraoperative cerebral ischemia during carotid endarterectomy. such as in operations on a malignant tumor surrounding the carotid artery. patency of the collateral pathways is usually not so critical. if a stenting procedure is performed during carotid flow reversal or an occlusion balloon is used for embolic protection. The delay in venous phase during temporary contralateral carotid occlusion also predicts poor clinical toleration after carotid sacrifice but our postmortem angiographic study did not facilitate performance this dynamic test. However. . However. preferably during induction of hypotension and possibly with simultaneous brain perfusion imaging with SPECT. However.1104 Fig. the ICA is usually occluded only for a short period of time during dilatation with a balloon and. During carotid artery stenting. We underline that the criteria and possible problems for insufficient collateral flow pathways in the circle of Willis were selected for temporary carotid occlusion during carotid surgery. includes reasonable uncertainty. 5 A volume-rendered MR angiographic image of a 50year-old man indicates that the right A1 segment is aplastic (arrow in a). In these patients. a hypoplastic A1 segment may enlarge due to hemodynamic shear stress in the case of contralateral carotid occlusion. In our postmortem study. It must be emphasized that the flow demand may be much higher if there is a risk that the carotid artery needs to be sacrificed. therefore. When using general anesthesia. since chronic hypoperfusion can also induce secondary collateral pathways via ophthalmic artery and leptomeningeal vessels that may compensate for sudden decrease of blood supply to the brain during vascular clamping. careful evaluation of the original MR angiography slices reveals that the A1 segment on the right is patent but hypoplastic (b). emphasizing the difficulties in the interpretation of very small vascular structures and the differences between different types on reconstruction of MR angiography data. based on morphological anatomy of the circle of Willis. demand of blood flow through these collaterals before ICA clamping and thus the flow is not visualized even in catheter angiography without. the criteria applied in the present study for insufficient collateral pathways in the circle of Willis may also be utilized to select those patients who can tolerate these type of interventions. collateral pathways should be evaluated preoperatively by ICA test occlusion. We conclude that an incomplete circle of Willis predisposes about one-sixth of individuals to cerebral ischemia during intraoperative closure of the carotid artery. Patency of A1 is confirmed in subsequent digital subtraction angiography (c) M. carotid compression provocation. Further. the mixture of lead oxide and liquid silicone rubber was perfused with physiological pressure simultaneously through both carotid arteries and filled even tiny collaterals very effectively. it is useful to evaluate the morphology of the circle of Willis by preoperative imaging and to identify unfavorable anatomical variations indicating intraoperative shunting or use of alternative endovascular intervention to avoid prolonged carotid closure with possible cerebral hypoperfusion.

Ronkainen A (2008) Endangered cerebral blood supply after closure of left subclavian artery: postmortem and clinical imaging studies. Federico C. Grubhofer G. Brain 90:747–758. Bagan P. Hutchinson EC. Bernhard VM.4. Mali WP (1998) Circle of 13.1007/s00508-007-0837-9 Riles TS. Dworschak M (2007) Contralateral stroke during carotid endarterectomy due to abnormalities in the circle of Willis. Lancet 363:1491–1502. Ann Vasc Surg 15:19–24 Salvian AJ.747 4. Breteler MM. Alpers BJ. Gumerlock MK. Karhunen PJ (1999) Post-mortem cast angiography in the diagnostics of graft complications in patients with fatal outcome following coronary artery bypass grafting (CABG). van Huffelen AC (2002) High embolic rate early after carotid endarterectomy is associated with early cerebrovascular complications. 15. Adelman MA. Ronkainen A. Csiba L.1067/mva. Paddison RM (1959) Anatomical studies of the circle of Willis in normal brain. doi:10. AMA Arch Neurol Psychiatry 81:409–418 2. 18. Krabbe-Hartkamp MJ. Hippelainen M (2006) Incomplete circle of Willis and right axillary artery perfusion. Steinlechner B. Jacobowitz GR. doi:10. Koivisto T. Marro J. Legemate DA. Hunter GC (1994) Critical carotid artery stenosis: diagnosis. Destable MD. Battacharji SK. Siro P. doi:10. Nasel C. J Vasc Surg 36:278–284. doi:10. Lee HK.1093/brain/90. Corcos L. van der Grond J. Ital J Anat Embryol 101:115–123 Manninen H. Soppi V. Cardiovasc Surg 5:481–485. Vidal R.1007/s004140050211 . Berman SS. 14. doi:10. 19. Azorin JF (2006) Cerebral ischemia during carotid artery cross-clamping: predictive value of phase-contrast magnetic resonance imaging. Holzenbein T. doi:10. Stroke 23:1583–1587 8. especially in women. Hillen B.2006. Dumas JL.1016/j. Erly WK. doi:10. Gulisano M. Oksala A. 20. Fulesdi B (2003) Absent collateral function of the circle of Willis as risk factor for ischemic stroke. Salminen US. Hildebrand HD. Ann Thorac Surg 82:74–79. Int J Legal Med 112:107–114. Halliday A. Giangola G.1016/j. 16. doi:10. Choi CG. timing of surgery. Pacini P.034 Moritz A. de Groot JC.02. Algra A. Peto C. Landis R (1994) Causes of perioperative stroke after carotid endarterectomy: special considerations in symptomatic patients. Potter J (2004) Prevention of disabling and fatal strokes by successful carotid endarterectomy in patients without recent neurological symptoms: randomised controlled trial. Neurol Res 24:237– 240. Martinod E.1007/ S10016-006-9126-8 3. Vanninen R. Taylor DC. MacDonald DB (1997) Selective shunting with EEG monitoring is safer than routine shunting for carotid endarterectomy.08. Koci G. Suh DC (2004) Relationship between circle of Willis morphology on 3D time-offlight MR angiograms and transient ischemia during vascular clamping of the internal carotid artery during carotid endarterectomy. Lim SM (2002) The anatomy of the circle of Willis as a predictive factor for intra-operative cerebral ischemia (shunt need) during carotid endarterectomy.1016/ S0140-6736(04)16146-1 7. Kim DK. Catini C. Hoksbergen AW.1016/ S0967-2109(97)00044-6 Weman SM. van Duijn H. Csati G. The International Transcranial Doppler Collaborators. 17. Jacobs MJ (2000) Collateral variations in circle of willis in atherosclerotic population assessed by means of transcranial color-coded duplex ultrasonography. Radiology 207:103–111 Laman DM. Erdoes LS. Kim GE.1159/000071115 9.035 Merkkola P. Willis: morphologic variation on three-dimensional time-of-flight MR angiograms. Hsiang YN. Halsey JH Jr (1992) Risks and benefits of shunting in carotid endarterectomy. Tulla H. McIntyre KE. Berry RG. Ann Vasc Surg 20:747–752. Neuwelt EA (1988) Carotid endarterectomy: to shunt or not to shunt. AJNR Am J Neuroradiol 25:558–564 Macchi C. Wieneke GH. Tulla H. Brizzi E (1996) Magnetic resonance angiographic evaluation of circulus arteriosus cerebri (circle of Willis): a morphologic study in 100 human healthy subjects. Legemate DA.2007. McCall AJ (1967) The Circle of Willis—the incidence of developmental abnormalities in normal and infarcted brains.How often does an incomplete circle of Willis predispose to cerebral ischemia during closure of carotid artery? 1105 References 12. and outcome. Kim GE. Lamparello PJ. Humer MF. Teal PA. J Vasc Surg 20:499–508 discussion 508–510 5. Hoksbergen AW. Ann Thorac Surg 85:120–125. Cerebrovasc Dis 16:191–198. Penttila A. doi:10.1179/016164102101199846 11.athoracsur. 1. doi:10.2002. Mansfield A. Peto R. Mannikko A. de Leeuw FE. Ubbink DT. Stroke 31:1656–1660 10. Cecchi F. Wien Klin Wochenschr 119:669–673. Tremblay B.125796 Lee JH. Imparato AM. Litherland HK. doi:10. Cho YP.athoracsur. Stroke 19:1485–1490 6.

Sign up to vote on this title
UsefulNot useful