Safety of perioperative cerebral oxygen saturation during debranching in patients with incomplete circle of Willis

Transcription

1 Interactive CardioVascular and Thoracic Surgery 26 (2018) doi: /icvts/ivx443 Advance Access publication 19 January 2018 ORIGINAL ARTICLE Cite this article as: Takahashi Y, Murakami T, Sasaki Y, Bito Y, Fujii H, Nishimura S et al. Safety of perioperative cerebral oxygen saturation during debranching in patients with incomplete circle of Willis. Interact CardioVasc Thorac Surg 2018;26: a b Safety of perioperative cerebral oxygen saturation during debranching in patients with incomplete circle of Willis Yosuke Takahashi a,*, Takashi Murakami b, Yasuyuki Sasaki a, Yasuyuki Bito a, Hiromichi Fujii b, Shinsuke Nishimura a and Toshihiko Shibata b Department of Cardiovascular Surgery, Osaka City General Hospital, Osaka, Japan Department of Cardiovascular Surgery, Osaka City University Graduate School of Medicine, Osaka, Japan * Corresponding author. Department of Cardiovascular Surgery, Osaka City General Hospital, Hondori Miyakojima, Miyakojimaku, Osaka , Japan. Tel: ; fax: ; ysk@msic.med.osaka-cu.ac.jp (Y. Takahashi). Received 30 June 2017; received in revised form 6 December 2017; accepted 23 December 2017 Abstract OBJECTIVES: The consequences of common carotid artery (CCA) cross-clamping during debranching before thoracic endovascular aortic repair are unclear. We examined the safety of a simple CCA cross-clamping procedure under regional cerebral oxygen saturation monitoring (rso 2 ) in patients with a complete or incomplete circle of Willis (CoW) anatomy. METHODS: Twenty-eight patients with thoracic aneurysm underwent elective debranching thoracic endovascular aortic repair with bilateral frontal rso 2 monitoring at our institution between January 2012 and October Before CCA cross-clamping, we maintained a systemic mean arterial pressure of >100 mm Hg with a vasopressor. We recorded the bilateral frontal rso 2 before, during and after CCA cross-clamping. RESULTS: The CoW was incomplete in 11 (39.3%) patients. Of these, 6 patients had a complication of ischaemic potential. The left frontal rso 2 was <50% in 3 patients but did not fall below 40%. Compared with baseline values (mean ± SD 64.6 ± 6.9%), the left frontal rso 2 showed no significant change perioperatively in those with a complete CoW on the left CCA cross-clamping (during: 61.0 ± 7.9%, P = 0.17; after: 65.1 ± 5.9%, P = 0.09). In patients with an incomplete CoW with ischaemic potential, the left frontal rso 2 did not change significantly after cross-clamping (baseline: 59.8 ± 3.2%, during: 55.5 ± 5.0%; P = 0.10) but increased significantly on declamping (62.8 ± 4.5%, P = 0.023). The extent of the changes in the mean left frontal rso 2 on clamping and declamping decreased and increased by 7.3% and 11.7%, respectively, in patients with an incomplete CoW, when compared with 5.3% and 5.8% in those with a complete CoW (P = 0.65 and 0.31, respectively). No perioperative cerebrovascular events were observed. CONCLUSIONS: Simple CCA cross-clamping during debranching was safe when arterial pressure was supported and rso 2 was monitored, even with an incomplete CoW and ischaemic potential. Keywords: Cerebral infarction Thoracic endovascular repair Debranching INTRODUCTION Thoracic endovascular aortic repair (TEVAR) is a less invasive approach for treatment of aortic arch disease but may be complicated by perioperative stroke. In patients with an aneurysm involving the aortic arch, TEVAR with several debranching or fenestrated procedures have been reported [1, 2]. Our approach to TEVAR has been simply to cross-clamp the common carotid artery (CCA) during debranching while monitoring the regional cerebral oxygen saturation (rso 2 ). When the circle of Willis (CoW) is complete, simple cross-clamping of the CCA is associated with a lower risk of cerebral infarction because of the adequate cross-perfusion of the ipsilateral brain. On the other hand, simple cross-clamping of the CCA may increase the risk of cerebral ischaemia when the CoW is incomplete. The optimum approach for maintaining cerebral perfusion during debranching TEVAR simple cross-clamping of the CCA or shunting is a matter of debate, especially in patients in whom the CoW is incomplete. The purpose of this study was to examine the relationship between anatomical completeness of the CoW and neurological outcomes in patients undergoing simple CCA cross-clamping during debranching before TEVAR with rso 2 monitoring and arterial pressure support. PATIENTS AND METHODS Proximal landing zones were stratified by the extent of anatomical arch disease, using the classification proposed by Mitchell et al. [3]. Between January 2012 and October 2015, 131 patients VC The Author(s) Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.

2 966 Y. Takahashi et al. / Interactive CardioVascular and Thoracic Surgery underwent TEVAR in our institution. Of these patients, 37 with Zone 0, 1 or 2 lesions underwent debranching TEVAR for various aortic disorders. Twenty-eight of these patients underwent routine, planned surgery that included magnetic resonance angiography (MRA) as part of their preoperative assessment (Table 1). Patients with shaggy aorta or those requiring emergency surgery were excluded from the analysis. We examined frontal rso 2 during the debranching procedure and compared it between patients with a complete CoW (n = 17) and those with an incomplete CoW with ischaemic potential (n = 6). Ischaemic potential is defined with regard to a previous report [4]. The institutional review board of Osaka City General Hospital approved the data analysis for this retrospective study, and the board waived the need for patient consent. Magnetic resonance angiography We performed MRA using a 1.5-T unit (Signa EXCITE HT 1.5T, GE Healthcare Japan, Tokyo, Japan) to examine the CoW (Table 1). Arteries with diameters <50% of the expected diameter (when compared with the neighbouring portion or the contralateral side) were considered hypoplastic [5]. The absence of visualization was considered aplasia [5]. We defined incomplete CoW as hypoplasia or aplasia of the anterior and posterior communicating arteries, P1 segment and A1 segment [4]. Anatomical variations of the CoW are shown in Fig. 1. We classified incomplete CoW as Types I IV. According to a previous report [4], hypoplasia or aplasia of right A1 (Type IIa), left P1 (Type IVb), left posterior communicating artery (Type IIIb) and right vertebral artery potentially has a possibility of jeopardizing cross-perfusion of the left hemisphere during left CCA clamping. We defined these variations as an incomplete CoW with ischaemic potential. Table 1: Profiles of all patients, patients with a complete CoW and those with an incomplete CoW All (n = 28) Complete CoW (n = 17) Incomplete CoW(-) (n = 11) P-value Age (years), mean ± SD 73 ± ± ± Male gender Risk factor Diabetes mellitus Hypertension Current smoking CVA CAD PAD COPD CRF Zone Zone Zone TAA Chronic DAA Others CAD: coronary artery disease; COPD: chronic obstructive pulmonary disease; CoW: circle of Willis; CRF: chronic renal failure; CVA: cerebrovascular accident; DAA: dissecting artery aneurysm; PAD: peripheral artery disease; SD: standard deviation; TAA: thoracic artery aneurysm. Patients with proximal landing Zone 2 (1-debranching technique) Left CCA-to-left subclavian artery bypass was performed with a 6-mm Fusion Bioline heparin-coated vascular graft (Maquet Endovascular, Wayne, NJ, USA). After simple clamping of the left CCA, we anastomosed the graft from end-to-side as soon as possible while monitoring rso 2. The graft was then tunnelled under the left clavicle and anastomosed to the left subclavian artery from end-to-side (Fig. 2A). Surgical bypasses were performed before TEVAR. Patients with proximal landing Zone 1 (2-debranching technique) In cases that required right CCA-to-left CCA bypass, intraoperative transoesophageal echocardiography was used routinely in the retropharyngeal space. Right CCA-to-left CCA bypass was performed with 6-mm intra-ringed polytetrafluoroethylene grafts (Gore-Tex Intering vascular graft, W. L. Gore & Associates, Inc., Flagstaff, AZ, USA) from end-to-side and tunnelled retropharyngeally. The proximal end of the left CCA was ligated and closed, and the distal end was anastomosed with the graft from end-toside. The graft was then tunnelled under the left clavicle and anastomosed to the left subclavian artery from end-to-side (Fig. 2B). Surgical bypasses were performed before TEVAR. Patients with proximal landing Zone 0 (3-debranching technique) The chest was opened via median sternotomy. With a side-biting clamp placed as proximal as possible on the ascending aorta, the first anastomosis with a 3-branched prosthetic graft (HEMA- SHIELD GOLD Woven 3 Branch Graft; Maquet Cardiovascular) was performed end-to-side. Next, the brachiocephalic trunk (end-to-end), left carotid artery (end-to-end) and left subclavian artery (end-to-end) were anastomosed to the smaller branches of the prosthesis. The proximal ends of the neck arteries were ligated and closed. Thoracic endovascular aortic repair All procedures were performed under general anaesthesia and with systemic heparinization. After debranching the neck vessels, the left subclavian artery was occluded with a balloon proximal to the vertebral artery to prevent thromboembolism during TEVAR. Stent grafts (Gore TAG, W. L. Gore & Associates, Inc.; Zenith TX2, Cook Medical, Bloomington, IN, USA; or Relay Plus, ABS Bolton Medical, Barcelona, Spain) were deployed through an open common femoral access. Then, coiling of the left subclavian artery was performed using Tornado platinum coils (Cook Medical) while occluding with the balloon. Maintenance and monitoring of cerebral oxygenation All patients underwent perioperative bilateral frontal rso 2 monitoring with a device approved by the US Food and Drug Administration (INVOS 5100C Cerebral Oximeter; Covidien,

3 Y. Takahashi et al. / Interactive CardioVascular and Thoracic Surgery 967 Boulder, CO, USA). We measured rso 2 before, during clamping and after CCA declamping. We proposed 2 levels of intervention as follows: if rso 2 decreased by 10 20% below the baseline values on CCA cross-clamping, we performed a physiological intervention [6]. If the rso 2 decreased to <_40% or by >20% persistently below the baseline values during debranching, we prepared to perform femoral arterial shunting. In the physiological intervention group, an intraoperative management protocol was used to maintain rso 2 values. With a decrease in rso 2, the patient s head position was checked to ensure that it had not been Figure 1: Anatomical variations of the circle of Willis: (A) complete circle of Willis; (B) absent AComA (Type I); (C) absent proximal segment of the right anterior cerebral artery (A1) (Type IIa) and left A1 (Type IIb); (D) absent right PComA (Type IIIa) and left PComA (Type IIIb); (E) absent proximal segment of the right PCA (P1) (Type IVa) and left P1 (Type IVb). Square in red means ischaemic potential. AComA: anterior communicating artery; BA: basilar artery; PCA: posterior cerebral artery; PComA: posterior communicating artery; VA: vertebral artery. Figure 2: Representative cases of thoracic endovascular aortic repair (TEVAR): (A) 3D computed tomography image of 1-debranching TEVAR with left carotid arteryto-left subclavian artery bypass (arrowhead) and coiling of the proximal left subclavian artery; (B) 3D computed tomography image of 2-debranching TEVAR with right common carotid artery-to-left carotid artery bypass, left carotid artery-to-left subclavian artery bypass (arrow), closure of the proximal left common carotid artery and coiling of the proximal left subclavian artery.

4 968 Y. Takahashi et al. / Interactive CardioVascular and Thoracic Surgery inadvertently rotated, and the face was observed to detect plethora. If the PaCO 2 or end-tidal carbon dioxide (CO 2 ) was 35 mmhg during positive pressure ventilation, ventilation was reduced to achieve a PaCO 2 of 40 mmhg. Before CCA clamping, the systemic mean arterial pressure (MAP) was maintained at >100 mmhg with phenylephrine and/or ephedrine to maintain the rso 2. In patients with persistent rso 2 below treatment, fraction of inspired oxygen increased. If haematocrit count was below 30%, red blood cell transfusion was administered [6]. The extent of the changes in the rso 2 on CCA clamping (1) and declamping (2) were calculated as follows: rso 2 before clamping - rso 2 during clampingþ 100; (1) rso 2 before clamping ðrso 2 after declamping - rso 2 during clampingþ 100: (2) rso 2 during clamping Data collection and statistical analysis Descriptive statistics for categorical variables were reported as frequency and percentage, and continuous variables were given as mean and standard deviation (SD). Categorical data were compared between the 2 groups (complete CoW group versus incomplete CoW with ischaemic potential group) using the Fisher s exact test. Continuous data between the 2 groups were compared using the Student s t-test. Statistical analysis was performed with StatView (SAS Institute, Cary, NC, USA). Statistically significant differences were assumed when P-values were <0.05. RESULTS Preoperative magnetic resonance angiography We performed MRA before surgery in 28 patients. Of these patients, 11 had an incomplete CoW; the findings are summarized in Table 2. Of the 11 patients with an incomplete CoW, 7, 3 and 1 underwent 1-, 2- and 3-debranching TEVAR, respectively. Three (27.2%) cases were complicated with hypoplastic vertebral artery, and another 3 cases were complicated with stenosis of the internal carotid artery and/or middle cerebral artery. Three (27.2%) patients had a complication of more than 2 hypoplastic or aplastic lesions of CoW. Six (54.5%) patients had a complication of ischaemic potential. Seventeen (60.7%) patients had a complete CoW anatomy, of whom 11 patients underwent 1-debranching TEVAR and 6 patients underwent 2-debranching TEVAR. Of the patients with a complete CoW, 3 patients had hypoplasia of the right vertebral artery and 1 patient had bilateral hypoplasia of the vertebral artery. Perioperative cerebral oxygenation Interventions to increase regional cerebral oxygen saturation monitoring. All patients underwent simple CCA clamping without CCA perfusion. Physiological interventions to increase arterial CO 2 tension and to adjust anaesthetic depth were performed in 4 (23.5%) patients with a complete CoW and in 5 (45.5%) patients with an incomplete CoW (P = 0.70). The left frontal rso 2 decreased to <50% in 3 patients, of whom 2 had a complete CoW anatomy with left frontal rso 2 of 66.0%, 48.0% and 65.0% in 1 patient and 63.0%, 48.0% and 68.0% in the other patient before left CCA clamping, after cross-clamping and after declamping, respectively. The CoW was incomplete in the other patient, with the left frontal rso 2 of 56.0% before left CCA clamping, 47% after cross-clamping and 54% after declamping. Physiological interventions were performed in all 3 cases. Femoral arterial shunting to the CCA was not necessary in all patients. Changes in the left frontal rso 2 before, during and after left CCA clamping in all 6 patients with ischaemic potential complication are presented in Table 3. Physiological interventions to increase arterial CO 2 tension and to adjust anaesthetic depth were performed in 3 (50.0%) patients. Changes in the frontal regional cerebral oxygen saturation monitoring at any time points. The changes in the bilateral frontal rso 2 before, during and after left CCA Table 2: Variations of an incomplete CoW and lesions of the vertebral and carotid arteries CoW variations VA Carotid artery Ischaemic potential Type Patient 1 Aplasia of bilateral PComA Hypoplasia of right VA (+) IIIa + IIIb Patient 2 Aplasia of AComA Right ICA stenosis (+) I Patient 3 Aplasia of right P1 IVa Patient 4 Aplasia of right P1 IVa Patient 5 Aplasia of right PComA, Hypoplasia of left VA (+) IIa + IIIa + IIIb hypoplasia of left PComA, hypoplasia of right A1 Patient 6 Aplasia of right P1 IVa Patient 7 Hypoplasia of right PComA IIIa Patient 8 Aplasia of left P1 Left MCA stenosis (+) IVb Patient 9 Hypoplasia of left A1, hypoplasia of left PComA Left ICA stenosis (+) IIb + IIIb Patient 10 Hypoplasia of left A1 Hypoplasia of left VA IIb Patient 11 Hypoplasia of AComA, bilateral aplasia of P1 (+) I + IVa + IVb AComA: anterior communicating artery; CoW: circle of Willis; ICA: internal carotid artery; MCA: middle cerebral artery; PComA: posterior communicating artery; VA: vertebral artery.

5 Y. Takahashi et al. / Interactive CardioVascular and Thoracic Surgery 969 clamping in patients with a complete CoW (n = 17) and those with an incomplete CoW with ischaemic potential (n = 6) are shown in Fig. 3A and B. The mean ± SD right frontal rso 2 in patients with a complete CoW was 60.6 ± 7.2% before left CCA clamping, 61.7 ± 8.1% after cross-clamping (P = 0.70) and 63.0 ± 7.1% after declamping (P = 0.63). The mean ± SD left frontal rso 2 in patients with a complete CoW was 64.6 ± 6.9% before left CCA clamping and 61.0 ± 7.9% after cross-clamping (P = 0.17) and 65.1 ± 5.9% after declamping (P = 0.09). No significant differences were found between the right and the left frontal rso 2 in the complete CoW anatomy group at any of the time points. The mean ± SD right frontal rso 2 in patients with an incomplete CoW with ischaemic potential was 60.2 ± 5.6% before left CCA clamping, 61.8 ± 6.9% after cross-clamping (P = 0.65) and 63.3 ± 9.2% after declamping (P = 0.76). The mean ± SD left frontal rso 2 in patients with an incomplete CoW was 59.8 ± 3.2% before left CCA clamping and 55.5 ± 5.0% after cross-clamping (P =0.10) but increased significantly to 62.8 ± 4.5% after declamping (P = 0.023). During left CCA clamping, the left frontal rso 2 was lower than the right frontal rso 2 (P = 0.098). The comparison of the left frontal rso 2 at any time point between the complete and incomplete CoW groups is shown in Fig. 3C. No significant difference in perioperative rso 2 was found between those with a complete and those with an incomplete CoW anatomy with ischaemic potential at any of the time points. Table 3: rso 2 before, during and after the left CCA clamping in patients with an incomplete CoW with the risk of hypoperfusion Patient rso 2 before ischaemia (%) rso 2 during ischaemia (%) rso 2 after ischaemia (%) Patient Patient Patient Patient Patient Patient Mean ± SD 60 ± ± ± 4.5 CCA: common carotid artery; CoW: circle of Willis; rso 2 : regional cerebral oxygen saturation monitoring; SD: standard deviation. Extent of change in the regional cerebral oxygen saturation monitoring on common carotid artery clamping or declamping. The extent of the changes in the rso 2 on CCA clamping or declamping is shown in Fig. 4A. The extents of the changes in the mean left frontal rso 2 on clamping and declamping decreased and increased by 7.3% and 11.7%, respectively, in patients with an incomplete CoW, when compared with 5.3% and 5.8% in those with a complete CoW (P = 0.65 and 0.31, respectively). Effect of the vertebral arteries on all patients. The changes in the left frontal rso 2 before, during and after left CCA clamping in patients with a complete CoW with normal vertebral artery (n = 13), a complete CoW with hypoplasia or aplasia of the vertebral artery (n = 4), an incomplete CoW with normal vertebral artery (n = 8) and an incomplete CoW with hypoplasia or aplasia of vertebral artery (n = 3) are shown in Fig. 4B. No significant difference in the left frontal rso 2 at any of the time points was found among the 4 groups (Fig. 4B). Perioperative data The mean left CCA clamping time was 14.5 ± 4.1 min in patients with a complete CoW and 15.2 ± 2.5 min in an incomplete CoW with ischaemic potential (P = 0.70). Evidence of clinically apparent stroke was not detected on routine clinical examination. One of the patients was diagnosed as having transient ischaemic attack 6 months after surgery, and 1 patient was diagnosed as having the regional dissection of the left CCA due to cross-clamping. The recovery of these 2 patients was complicated by haemorrhage from the CCA or left subclavian artery suture site, because of consumption disseminated intravascular coagulation. No postoperative infections were observed. Two patients developed Type Ia endoleak after surgery, and both required additional endovascular treatment at the proximal landing sites during the same episode of hospitalization. Four patients developed Type II endoleak, 2 of whom required reintervention by plug insertion in the left subclavian artery after conservative management failed. DISCUSSION We found no statistically significant changes in the left frontal rso 2 from baseline to during left CCA clamping or after left CCA Figure 3: (A) Changes in bilateral frontal rso 2 before, during and after left CCA clamping in patients with a complete CoW anatomy. (B) Changes in the bilateral rso 2 before, during and after left CCA clamping in patients with an incomplete CoW anatomy with ischaemic potential. (C) The changes in the left frontal rso 2 before, during and after left CCA clamping in patients with a complete CoW anatomy or an incomplete CoW anatomy with ischaemic potential. Data are expressed as mean- ± standard deviation. *P < CCA: common carotid artery; CoW: circle of Willis; Lt: left; rso 2 : regional cerebral oxygen saturation monitoring.

6 970 Y. Takahashi et al. / Interactive CardioVascular and Thoracic Surgery Figure 4: (A) Extent of the changes in the rso 2 monitoring on the left CCA clamping or declamping in patients with a complete CoW anatomy or an incomplete CoW anatomy with ischaemic potential. (B) Changes in the left frontal rso 2 monitoring before, during and after left CCA clamping in patients with a complete or an incomplete CoW anatomy with normal or abnormal vertebral arteries. CCA: common carotid artery; CoW: circle of Willis; rso 2 : regional cerebral oxygen saturation. declamping in patients with a complete CoW anatomy. Although rso 2 decreased slightly on left CCA clamping in patients with an incomplete CoW with ischaemic potential, it increased significantly after declamping. No statistically significant difference in the rso 2 was found between those with a complete CoW and those with an incomplete CoW anatomy at any of the perioperative time points; however, the left rso 2 in patients with an incomplete CoW with ischaemic potential was slightly lower than those in patients with a complete CoW during left CCA clamping. Simple cross-clamping of the CCA under rso 2 monitoring with vasopressor support and appropriate physiological intervention was not associated with perioperative or postoperative stroke, even in patients with an incomplete CoW with ischaemic potential. Cerebral oximetry has been steadily adopted as the routine monitoring for aortic operations by major high-volume centres in North America and Western Europe [7], and some evidence demonstrates its benefits during cardiovascular surgery [8, 9]. Slater et al. [9] reported that rso 2 predicted cognitive dysfunction after on-pump coronary artery bypass grafting, demonstrating that prolonged cerebral desaturation of <50% was associated with postoperative cognitive decline [9]. Yao et al. [8] reported a significant relationship between intraoperative cerebral desaturation of <40% and postoperative cognitive decline after cardiac surgery. Ours is the first report of rso 2 monitoring during debranching procedures and cerebral outcomes after debranching. In our study, we found a cerebral desaturation of <50% in 3 (11.1%) patients, but no incidences of clinically apparent stroke were detected on routine clinical examination, which could be explained by the relatively shorter duration of cerebral desaturation than that in previously reported cases [10, 11]. The reason the rso 2 in 1 patient with a complete CoW decreased to <50% might be the complication of stenosis in the M1 portion. The reason the rso 2 of 1 patient with an incomplete CoW decreased to <50% might be because of the ischaemic potential (Type I). Augmentation of the systemic MAP during debranching might also allow adequate cross-perfusion of the ipsilateral brain. Furthermore, interventions such as elevation of arterial CO 2 tension, adjustment of the anaesthetic depth and blood transfusion may have contributed to positive postoperative outcomes. On the basis of clinical observation, congenital incompleteness of the CoW is present in >50% of the population [12, 13]; in our study, the CoW was incomplete in 11 (39.3%) patients. Sussman et al. [14] reported that poor collateral circulation within the CoW predicts cognitive decline after carotid endarterectomy, particularly when the posterior communicating arteries were absent. Pennekamp et al. [15] reported that incompleteness of the CoW is associated with electroencephalography-based shunting during carotid endarterectomy. Assessment of the cerebral circulation and configuration of the CoW using MRA can identify patients with a low and high likelihood of the need for shunt use during surgery. Additional perfusion of the CCA during CCA clamping is believed to reduce the incidence of perioperative stroke in high-risk patients [16 18]. In contrast, Rerkasem and Rothwell [19] found no evidence to support the use of a carotid shunt during endarterectomy, possibly because of the risk of arterial wall injury. Although rso 2 did not decrease to <40% in our cohort, rso 2 monitoring can provide information to guide the decision to initiate additional physiological intervention or CCA perfusion, especially if the CoW is incomplete. We found that although rso 2 did not decrease significantly on CCA clamping in patients with an incomplete CoW with ischaemic potential, a significant recovery occurred after declamping. However, no significant differences were found at any time point in comparison with patients with a complete CoW anatomy. The significant rebound in rso 2 on declamping may indicate the effect of physiological intervention during debranching or may, nonetheless, indicate the presence of substantial brain ischaemia when the CoW is incomplete. The lower left rso 2 compared with the right rso 2 during left CCA clamping may also indicate substantial brain ischaemia when the CoW is incomplete. Montisci et al. [20] reported that 2.9% of patients with a complete CoW and 4.8% of those with single-vessel CoW hypoplasia or aplasia were intolerant of CCA cross-clamping, increasing to 60.0% if 2 or more CoW vessels were hypoplastic or absent. In our study, 4 patients had 2 or more hypoplastic vessels in the CoW. We judge that none of these patients experienced perioperative cerebrovascular events, as we strictly maintained the MAP and monitored the rso 2 during debranching. Limitations Our study had several limitations. First, it was retrospective in nature and involved a small number of patients. There is a possibility that a statistically significant difference in the rso 2 might be found between those with a complete CoW and those with specific incomplete CoW anatomy during ischaemia. These issues should be addressed in future prospective studies by including

7 Y. Takahashi et al. / Interactive CardioVascular and Thoracic Surgery 971 larger numbers of patients. Second, we only used MRA to assess the CoW. Although both computed tomography angiography and MRA are considered reliable tools for the assessment of the CoW, Pennekamp et al. [15] reported that anomalies identified on MRA were more likely to be associated with cerebral ischaemia during carotid endarterectomy rather than with anomalies observed on computed tomography angiography. Third, we supported MAP with phenylephrine and/or ephedrine to maintain the rso 2 before CCA clamping. Nissen et al. [21] demonstrated that phenylephrine but not ephedrine reduced the rso 2 after anaesthesia-induced hypotension. Further studies are necessary to compare the effect of phenylephrine and that of ephedrine and to examine their effects on the rso 2 during debranching. Finally, we monitored rso 2 on both sides of the forehead; consequently, we examined the territory of the middle cerebral arterial area (the watershed area) and, therefore, may have missed ischaemia in the territories of the posterior cerebral or basilar circulations. Further studies with global cerebral oxygenation monitoring will also be needed. CONCLUSION In conclusion, simple clamping of the left CCA might be safe, even in patients with an incomplete CoW with ischaemic potential. Augmentation of the systemic MAP increased the collateral cerebral blood flow and may have reduced the risk of perioperative stroke. Monitoring rso 2 allows debranching to be performed safely. Conflict of interest: none declared. REFERENCES [1] Pecoraro F, Lachat M, Hofmann M, Cayne NS, Chaykovska L, Rancic Z et al. Mid-term results of zone 0 thoracic endovascular aneurysm repair after ascending aorta wrapping and supra-aortic debranching in highrisk patients. Interact CardioVasc Thorac Surg 2017;24: [2] Hori D, Okamura H, Yamamoto T, Nishi S, Yuri K, Kimura N et al. Early and mid-term outcomes of endovascular and open surgical repair of non-dissected aortic arch aneurysm. Interact CardioVasc Thorac Surg 2017;24: [3] Mitchell RS, Ishimaru S, Criado FJ, Ehrlich MP, Ivancev K, Lachat M et al. Third International Summit on Thoracic Aortic Endografting: lessons from long-term results of thoracic stent-graft repairs. J Endovasc Ther 2005;12: [4] Papantchev V, Stoinova V, Aleksandrov A, Todorova-Papantcheva D, Hristov S, Petkov D et al. The role of Willis circle variations during unilateral selective cerebral perfusion: a study of 500 circles. Eur J Cardiothorac Surg 2013;44: [5] Ohkura K, Shiiya N, Washiyama N, Yamashita K, Takahashi D, Tsuda K et al. Vertebral artery variations in thoracic aortic patients. Eur J Cardiothorac Surg 2014;46: [6] Murkin JM, Adams SJ, Novick RJ, Quantz M, Bainbridge D, Iglesias I et al. Monitoring brain oxygen saturation during coronary bypass surgery: a randomized, prospective study. Anesth Analg 2007;104:51 8. [7] Olsson C, Thelin S. Regional cerebral saturation monitoring with nearinfrared spectroscopy during selective antegrade cerebral perfusion: diagnostic performance and relationship to postoperative stroke. J Thorac Cardiovasc Surg 2006;131: [8] Yao FS, Tseng CC, Ho CY, Levin SK, Illner P. Cerebral oxygen desaturation is associated with early postoperative neuropsychological dysfunction in patients undergoing cardiac surgery. J Cardiothorac Vasc Anesth 2004;18: [9] Slater JP, Guarino T, Stack J, Vinod K, Bustami RT, Brown JM III et al. Cerebral oxygen desaturation predicts cognitive decline and longer hospital stay after cardiac surgery. Ann Thorac Surg 2009;87: [10] Vets P, ten Broecke P, Adriaensen H, Van Schil P, De Hert S. Cerebral oximetry in patients undergoing carotid endarterectomy: preliminary results. Acta Anaesthesiol Belg 2004;55: [11] Kragsterman B, Pärsson H, Bergqvist D. Local haemodynamic changes during carotid endarterectomy: the influence on cerebral oxygenation. Eur J Vasc Endovasc Surg 2004;27: [12] Eftekhar B, Dadmehr M, Ansari S, Ghodsi M, Nazparvar B, Ketabchi E. Are the distributions of variations of circle of Willis different in different populations? Results of an anatomical study and review of literature. BMC Neurol 2006;6:22. [13] Vrselja Z, Brkic H, Mrdenovic S, Radic R, Curic G. Function of circle of Willis. J Cereb Blood Flow Metab 2014;34: [14] Sussman ES, Kellner CP, Mergeche JL, Bruce SS, McDowell MM, Heyer EJ et al. Radiographic absence of the posterior communicating arteries and the prediction of cognitive dysfunction after carotid endarterectomy. J Neurosurg 2014;121: [15] Pennekamp CW, van Laar PJ, Hendrikse J, den Ruijter HM, Bots ML, van der Worp HB et al. Incompleteness of the circle of Willis is related to EEG-based shunting during carotid endarterectomy. Eur J Vasc Endovasc Surg 2013;46: [16] Aburahma AF, Mousa AY, Stone PA. Shunting during carotid endarterectomy. J Vasc Surg 2011;54: [17] Tambakis CL, Papadopoulos G, Sergentanis TN, Lagos N, Arnaoutoglou E, Labropoulos N et al. Cerebral oximetry and stump pressure as indicators for shunting during carotid endarterectomy: comparative evaluation. Vascular 2011;19: [18] Okita Y, Matsumori M, Kano H. Direct reperfusion of the right common carotid artery prior to cardiopulmonary bypass in patients with brain malperfusion complicated with acute aortic dissection. Eur J Cardiothorac Surg 2016;49: [19] Rerkasem K, Rothwell PM. Routine or selective carotid artery shunting for carotid endarterectomy (and different methods of monitoring in selective shunting). Cochrane Database Syst Rev 2009;4:CD [20] Montisci R, Sanfilippo R, Bura R, Branca C, Piga M, Saba L. Status of the circle of Willis and intolerance to carotid cross-clamping during carotid endarterectomy. Eur J Vasc Endovasc Surg 2013;45: [21] Nissen P, Brassard P, Jørgensen TB, Secher NH. Phenylephrine but not ephedrine reduces frontal lobe oxygenation during anesthesia-induced hypotension. Neurocrit Care 2010;12:17 23.