c© 2012 Wiley Periodicals, Inc. 211

Effective Cerebral Protection UsingNear-Infrared Spectroscopy Monitoringwith Antegrade Cerebral PerfusionDuring Aortic SurgeryEshan Senanayake, M.B.Ch.B.,∗ Mohamed Komber, M.D.,∗ Ahmed Nassef, M.D.,†Nicholas Massey, M.D.,‡ and Graham Cooper, M.D.∗

∗Department of Cardiothoracic Surgery; †Department of Vascular Surgery; and ‡Departmentof Cardiothoracic Anaesthesia, Northern General Hospital, Sheffield, United Kingdom

ABSTRACT Objective: Antegrade cerebral perfusion (ACP) under moderate hypothermia may improve cere-

bral protection. Intraoperative measurement of cerebral regional oxygen saturations (rSO2) using near-

infrared spectroscopy (NIRS) can provide accurate monitoring of cerebral perfusion during ACP. We eval-

uated the role, outcomes, and advantages of using NIRS in providing effective cerebral protection with

ACP. Methods: Between May 2006 and March 2009, 27 patients (mean age 60%, 93% elective) under-

went ascending aorta replacement with ACP monitored by NIRS. ACP was established through the right

axillary artery (n = 26). All patients had continuous intraoperative measurement of both anterior cere-

bral rSO2 using NIRS (INVOS; Somanetics Corporation, Troy, MI, USA). Posterior cerebral perfusion was

measured using left radial artery pressures. Quality of life (QoL) was assessed using a Short Form 36

questionnaire. Results: Mean cardiopulmonary bypass and aortic cross clamp time were 169 ± 27 and

95 ± 22 minutes, respectively. Mean ACP rate of 1.27 ± 0.35 L/min provided a mean left radial artery

pressure of at least 60 mmHg. All patients’ cerebral rSO2 were maintained above their baseline using

NIRS. Mean ACP time was 14.3 ± 2.6 minutes at a mean core temperature of 23.4 ◦C ± 2.0 ◦C. Tempo-

rary neurological deficit was observed in two patients (7.4%). No permanent neurological dysfunction was

observed. Thirty-day mortality was 3.7%. At median follow-up of 18.3 (interquartile range 10.8 to 23.3)

months survival was 92% and mean norm-based QoL score was above average at 52.5 ± 6.5. Conclu-

sion: Cerebral rSO2 and left radial artery pressure monitoring with ACP during aortic surgery provides

accurate measurement of cerebral perfusion resulting in minimal neurological and perioperative complica-

tions and good midterm QoL outcomes. doi: 10.1111/j.1540-8191.2012.01420.x (J Card Surg 2012;27:211-216)

Aortic surgery extending into the transverse aorticarch increases the potential for neurological injury.1

Despite the existence of a number of cerebral pro-tection strategies there still remains controversy overthe optimum method to be used.2-4 Selective ante-grade cerebral perfusion (SACP) has been developed todecrease the potential for cerebral damage during hy-pothermic circulatory arrest (HCA).1 SACP maintains

Conflict of interest: None.

Funding source: No additional sources of funding have been used forthis work.

This paper was presented at the AATS Aortic Symposium, New York;May 2010.

Address for correspondence: Graham J Cooper, M.D., Departmentof Cardiothoracic Surgery, Northern General Hospital, Herries Road,Sheffield S5 7AU, United Kingdom. Fax: 0114 2610350; e-mail:graham.cooper@sth.nhs.uk

cerebral perfusion during open distal anastomosis.However, existence of continuous and adequate cere-bral perfusion through both cerebral hemispheres iscontroversial as there is yet no standardized techniquein monitoring cerebral perfusion. Adequate cerebralperfusion through both cerebral hemispheres is morepertinent during unilateral antegrade cerebral perfu-sion. Despite the advantages of this technique, it re-lies on a patent vertebral artery and Circle of Willissystem to allow perfusion of the contralateral cerebralhemisphere. In these instances, continuous cerebralperfusion monitoring is imperative.

Near-infrared spectroscopy (NIRS) monitoring al-lows continuous noninvasive measurements of bothanterior cerebral regional oxygen saturations (rSO2).It is suggested that cerebral oxygenation can beassessed by NIRS and NIRS monitoring may optimizeneurological outcome. Clinical and experimental use ofNIRS monitoring during SACP has been used5,6 and

212 SENANAYAKE, ET AL.CEREBRAL PROTECTION DURING ANTEGRADE CEREBRAL PROTECTION

J CARD SURG2012;27:211-216

TABLE 1

Patient Characteristics

Characteristics Number (%)

Number of patients 27Age (IQR) 59 (51 to 62)Sex (M/F) 16/11Renal failure 0Pulmonary disease 3 (11.1)Myocardial infarction 0Diabetes 0Hypertension 16 (59)Peripheral vascular disease 0Aortic valve lesion

Aneurysmal 25 (92.6)dissection 2 (3.4)

Logistic euroscore 10.2 ± 7.25Timing of operation

Elective 25 (92.6)Urgent 1 (3.7)Emergency 1 (3.7)

Previous cardiac operation 0NYHA 2.1 ± 1.14CCS 0.6 ± 1.1

IQR = inter quartile range; M/F = male/female; NYHA =New York Heart Association; CCS = CanadianCardiovascular Status.

results have been promising,5-8 but its reliability hasbeen equivocal.9,10

In our institution, unilateral antegrade cerebral perfu-sion is used as opposed to antegrade selective cerebralperfusion. In this study, we aim to evaluate the value ofNIRS during unilateral antegrade cerebral perfusion andevaluate neurological outcome. Secondary goals wereto evaluate midterm outcomes and quality of life.

MATERIALS AND METHODS

Between May 2006 and March 2009, 27 patientshad an ascending aorta replacement with unilateralantegrade cerebral perfusion, performed by a singlesurgeon at our institution. Indications for aortic replace-ment were acute Type A dissection in two (7.4%) pa-tients and degenerative aneurysm in 25 (92.6%) pa-tients. All case notes were retrospectively analyzed forclinical variables including pre-operative status, intra-operative data, and early postoperative complications.Patient demographics are outlines in Table 1. This studywas authorized by the Sheffield Teaching HospitalsNHS Foundation Trust.

All patients had yearly computed tomography ormagnetic resonance imaging follow-up. Mortality wasdefined as all-cause in-hospital mortality. Permanentneurological deficit (PND) was defined as any newpostoperative neurological deficit that included newfocal stroke or global coma which did not resolve bydischarge, and confirmed by a new cerebral infrac-tion on computer tomography. Temporary neurolog-ical deficit (TND) was defined as any postoperativeneurological deficit that included motor deficit, con-

fusion, agitation, or transient delirium that resolvedspontaneously before discharge with no new cere-bral infarction on computer tomography. Neurologicaldeficit was initially identified and a clinical diagnosismade by the surgical team, followed by a radiologicalconfirmation of this diagnosis.

Near infrared spectroscopy

The principles of NIRS have been described in de-tail previously.10 NIRS measures intracerebral oxygensaturation by spectroscopy of reflected near-infraredlight. Near-infrared red light between 650 and 1100 nmwavelengths is used to measure absorption spectra ofoxygenated and deoxygenated forms of hemoglobin.Regional cerebral oxygen saturations can then be cal-culated according to the Lambert-Beer law, wherelight absorbance is proportional to the concentrationof these chromophores.

The INVOS 4100 cerebral oximeter (SomaneticsCorp, Troy, MI, USA) was used for continuous frontalcerebral rSO2 monitoring throughout the operation. Af-ter induction of anesthesia, bilateral sensors (single useadhesive patches) were placed on the forehead justto the right and left of the midline. Near-infrared lightwas emitted form a light emitting diode source and de-tected at 3 and 4 cm from this source. This allows forsubtraction of the contribution to the signal by superfi-cial tissues.

Baseline frontal cerebral rSO2 was measured afteranesthetic induction and commencement of cardiopul-monary bypass (CPB). Once ACP was commenced,rSO2 was maintained above base line by adjusting theACP flow rate. If there was a decrease in rSO2, the ACPflow rate was increased to achieve a reading higherthan baseline. Cerebral rSO2 was recorded every10 minutes and during key events, that is, initiation ofCPB, commencement of aortic cross clamp and ACP,termination of ACP, release of aortic cross clamp, ter-mination of CPB and every five minutes during ACP.From the rSO2 data the number of patients and timein which rSO2 was below 45%, 50%, 55%, and 60%were determined. The cerebral hemisphere that had alonger decrease was used as the representative valuefor each patient.

Operative technique

All patients received a standardized anesthetic andcardiopulmonary bypass management. Arterial cannu-lation and ACP was established via the right subclavianor right axillary artery. The patient’s head was packed inice and 10 mg/kg of thiopentone was given one minutebefore ACP was commenced. At 24 ◦C, the patient wasplaced head down, systemic circulation was arrested,and supra-aortic vessels were clamped. ACP wascommenced and confirmed by adequate continuousmeasurements of anterior cerebral rSO2, which wasmaintained above baseline using NIRS. Posterior cere-bral perfusion was monitored using continuous leftradial artery pressure monitoring. ACP flow rate wasmaintained to achieve a left radial artery pressure of

J CARD SURG2012;27:211-216

SENANAYAKE, ET AL.CEREBRAL PROTECTION DURING ANTEGRADE CEREBRAL PROTECTION

213

TABLE 2

Operative Data

Variables Number (%)

Cardiopulmonary bypass time (min) 169 ± 27Cross clamp time (min) 95 ± 22ACP rate 1.3 ± 0.35ACP time (min) 14.3 ± 2.5Core temp during ACP 23.4 ± 2.0Aneurysm diamter (cm) 5.7 ± 0.8Arterial cannulae site

Right subclavien artery 26 (96.3)Right axillary artery 1 (3.7)

Cell saver loss (IQR) 1301 (1095 to 1652)Valve graft type

Mechanical 22Biological 2

ACP = antegrade cerebral perfusion; IQR = interquartilerange.

≥60 mmHg and a rSO2 greater than baseline. Opendistal anastomosis was carried out after removing theaortic cross clamp. Acid-base balance was managedwith the alpha-stat strategy. After open distal anas-tomosis was complete, whole body cardiopulmonarybypass was recommenced through the side arm of theascending aortic graft. Operative details are shown inTable 2.

Quality of life assessment

Quality of life was assessed using a Short-Form 36(SF-36) health survey, conducted by telephone inter-view at a median follow-up period of 18.3 (IQR 10.8to 23.3) months. This assessed eight health domainswhich included physical functioning, role-physical, bod-ily pain, general health, vitality, social functioning, role-emotional, and mental health. These scores were usedto calculate a physical component summary (PCS) andmental component summary (MCS) and were com-pared to age-and sex-matched individuals.

Statistical analysis

All continuous data are presented as a mean ± stan-dard deviation. All categorical variables are presentedas percentages. Actuarial survival curves were esti-mated with the standard nonparametric Kaplan-Meiermethod using a 95% confidence interval.

RESULTS

Operative procedures and concomitant surgery areoutlined in Table 3. In all patients, ACP was maintainedabove their baseline cerebral rSO2 and left radial arterypressures were maintained >60 mmHg. A typical cere-bral rSO2 tracing is shown in Figure 1, where measure-ments were recorded every 10 minutes before ACPand every five minutes during ACP. All patients hadNIRS monitoring, and data for analysis were availablefor 18 patients. NIRS data were not stored for the re-maining nine patients.

TABLE 3

Operative Procedure and Concomitant Surgery

Procedure Number

Aortic root and ascending aorta replacement 12Aortic valve and ascending aorta replacement 11Ascending aorta replacement 3Aortic root and proximal arch replacement 1Concomitant CABG 3Concomitant mitral valve repair 1

CABG = coronary artery bypass graft.

Thirty days in hospital, all-cause mortality was 3.7%(n = 1). This patient died from low cardiac output syn-drome secondary to a long aortic cross clamp time.TND was seen in two patients (7.4%). One patientdeveloped amaurousis fugax and diplopia, which spon-taneously resolve by discharge. The second developedprolonged confusion, which resolved by discharge.There were no PNDs in this cohort.

Left and right hemispheric rSO2 curves remainedparallel in all patients but differences between hemi-spheres were present in individual patients (meandifference 1.9% ± 6.6%). In 50% of patients rSO2

were maintained above 60% throughout surgery. rSO2

dropped below 45%, 50%, and 55% in two, two,and five patients, respectively, on at least one oc-casion. These drops in rSO2 were not sustained forover five minutes. Of those that dropped rSO2 below45% and 50%, one in each group sustained a TND.In one patient with TND, rSO2 dropped below 50%on four separate occasions and once below 45% dur-ing ACP. In the other with TND, rSO2 dropped be-low 50% on three separate occasions. In both pa-tients who sustained TND, rSO2 ran below 60% for upto 144 minutes. However, in these patients, baselinereadings remained below 60% even before the initia-tion of CPB and after aortic cross clamping. As thesereadings were close to baseline, no further manipu-lations in ACP were performed. In one other patient,where rSO2 dropped below baseline, ACP rate wasincreased. This, however, showed no improvementin rSO2. Therefore, this patient was cooled further to16 ◦C, which improved rSO2. Early postoperative andlate complications are outlined in Table 4. Median du-ration of intensive care stay was three days (ICR 1.25to 4) and median length of hospital stay was nine days(ICR 3 to 20).

Of the 26 patients in follow-up, quality of life mea-sures were available for 17 patients. The remainingnine patients were not contactable by phone or hadmoved without a forwarding address. The mean qual-ity of life score was 80.3 ± 16.9 at median follow-up of 18.3 months. Mean norm based quality of lifescore was above average at 52.5 ± 6.5. Of this, PCSand MCS were both above average at 51.1 ± 8.7 and52.8 ± 10.8, respectively. After a median follow-uptime of 18.3 months (ICR 10.8 to 23.3) survival was92% ± 5%.

214 SENANAYAKE, ET AL.CEREBRAL PROTECTION DURING ANTEGRADE CEREBRAL PROTECTION

J CARD SURG2012;27:211-216

Figure 1. Left and right hemispheric rSO2 of one patient during an AVR and ascending aortic replacement. Marked events are asfollows: (A) Initiation of CPB; (B) aortic cross clamping; (C) initiation of ACP; (D) end of ACP; (E) end of aortic cross clamping; (F)end of CPB. AVR = aortic valve replacement; CPB = cardiopulmonary bypass; ACP = antegrade cerebral perfusion.

TABLE 4

Early and Late Complications

Complication Number (%)

Atrial fibrillation 8 (20)Renal failure requiring CVVH 1 (3)Drainage of pericardial effusion 3 (11)Chest sepsis causing prolonged ventilation 1 (3)Permanent pacemaker 3 (11)Sternal osteomyelitis 1 (3)

CVVH = continuous veno-venous hemofiltration.

DISCUSSION

In this study, there were no PND and TND was7.3%. Quality of life as assessed by SF-36 question-naire was above average and survival was 92% ± 5%at a median follow-up of 18.3 months. This study hasshown that near-infrared spectroscopy can be a use-ful tool in assessing adequate cerebral perfusion dur-ing unilateral antegrade cerebral perfusion, resultingin low neurological dysfunction. Patency of the pos-terior cerebral circulation was confirmed and main-tained by keeping left radial artery pressure above60 mmHg.

In this study, patency of the right vertebral arteryand Circle of Willis was assumed. There were no prior

investigations to confirm this before surgery. There-fore, NIRS monitoring before and during ACP providedan assurance that unilateral ACP is safe consideringthe initial assumption of vessel patency. On the ba-sis of NIRS monitoring, if there is evidence that thereis inadequate cerebral perfusion to the cerebral hemi-spheres, surgical technique can be adjusted to improverSO2. Failure to achieve adequate rSO2 above base-line should alert the surgical team to initiate changes.rSO2 can be achieved by increasing the flow rate orachieving further hypothermia. In addition, arterial can-nula position and cerebral oximetery equipment shouldbe tested. Further options would be to establish bilat-eral or SACP or perform the procedure under deepHCA.

SACP avoids deep hypothermic circulatory arrest(DHCA) during aortic arch surgery. Olsson and Thelin8

report that bilateral cerebral perfusion using SACP re-sulted in higher rSO2 values in both hemispheres com-pared with patients managed with unilateral ACP, withhigher right hemispheric readings. This however didnot correlate with perioperative neurological dysfunc-tion. We demonstrate good rSO2 correlation betweenright and left hemispheres and have shown unilateralperfusion to be associated with low neurological com-plications.

In this study, TND was 7.4% with no PND. Olssonand Thelin report a similar TND of 4.3% during SACPwith the aid of NIRS, but their techniques of cerebral

J CARD SURG2012;27:211-216

SENANAYAKE, ET AL.CEREBRAL PROTECTION DURING ANTEGRADE CEREBRAL PROTECTION

215

perfusion varied between unilateral and bilateralSACP.8 However their PND was 13%, similar to thatreported by Orihashi et al.6 Similarly Di Eusanio et al.,comparing DHCA with ASCP report a PND and TNDof 7.6% and 8.7%, respectively.11 However, Kazuiet al. report a lower TND and PND of 4.2% and 2.4%,respectively, with the aid of NIRS.3 Despite the evi-dence of low neurological dysfunction in these stud-ies with SACP and the aid of NIRS, there is also ev-idence to suggest high neurological dysfunction withSACP. Immer et al. report a higher major perioperativecerebrovascular injury with SACP of 9.8% comparedto 1.1% using continuous cerebral perfusion using theright subclavian artery during DHCA,12 supporting theuse of unilateral cerebral perfusion. Further studies us-ing unilateral flow have shown good neurological out-comes.13,14 Therefore, this study re-enforces the pos-itive neurological outcomes associated with unilateralACP. Mortality in this study was 3.7%. Other studiesreport a higher mortality of 11.4% to 25% in a higherrisk category of patients that include a high proportionof dissections and redo surgery.8,11,15

Cerebral rSO2 was maintained above baseline val-ues for each patient. Orihashi et al. reported thatneurological events increased when a drop in rSO2 be-low 55% is sustained for longer than five minutes6

and this is consistent with similar reports by Hirofumiet al.16 Other studies have demonstrated a cutoff point;a decrease of 20% from baseline, which had a highsensitivity and specificity for stroke,8 and a desatura-tion of 25% to 30% from baseline corresponded toan episode of severe cerebral ischemia.17 This studydid not assess a cutoff to predict perioperative strokebut provides assurance that maintaining rSO2 abovebaseline and >60% is associated with no neurologicaldysfunctions. All patients had their rSO2 maintainedabove baseline by either increasing ACP rate or achiev-ing a deeper hypothermia. The left radial artery pres-sures were used to confirm posterior cerebral circula-tion. Although the aim was to maintain left radial arterypressure >60 mmHg, ACP rate was influenced moreby the rSO2 readings to maintain reading above base-line. In both patients with a TND, baseline readings re-mained below 60% before and after commencementof CPB compared to those that had no neurologicaldeficient. In addition, these patients had ≥ 3 separateepisodes of rSO2 below 50%. These episodes werenot for more than five minutes on each occasion. There-fore, although it is pertinent to maintain rSO2 abovebaseline, a lower baseline rSO2 of <60% may be as-sociated with TND. Furthermore, a drop in rSO2 below50% on ≥3 separate occasions may also be associatedwith TND. In such patients a change in technique, asdescribed above, to improve cerebral perfusion may bebeneficial. In this study, brief periods of rSO2 <55%,50%, and 45% do occur but are not associated withTND if they are not sustained for more than five min-utes or occur on less than three separate occasions.

The period of ACP in this study was relatively lowat 14.3 ± 2.5 minutes, in comparison to other stud-ies in the literature where ASCP times range between19 and 96 minutes.6,8,15 This is in relation to the ex-

tent of surgery, with longer ASCP times associatedwith more complex and more extensive arch replace-ments. Therefore, it can be hypothesized that the lowTND is as a result of short ACP time and moder-ate hypothermia rather than the aid of NIRS. How-ever, this study has shown an association betweenlow rSO2 and TND. Therefore, although a short ACPtime is beneficial, NIRS monitoring may aid better cere-bral protection. A mean perfusion rate of 1.3 L/minwas observed in this study to maintain a left radialartery pressure of >60 mmHg and rSO2 above base-line. This flow rate has been shown to be adequateto maintain adequate cerebral perfusion and is simi-lar to those reported in other series where perfusionrate was maintain around 1 L/min or 10 mL/kg/minto maintain right radial artery pressure between 40and 70 mmHg.3,4,6,8,11

Quality of life was measured using a SF-36 ques-tionnaire, through telephone interview at a medianfollow-up of 18 months. The mean quality of life scorewas 80.3 ± 16, similar to quality of life reported by Im-mer et al.,12 reporting a good quality of life with arterialaccess through the right subclavian artery in surgeryof the aorta. The norm based quality of life score wasabove average at 52.5 ± 6.5. The individual PCS andMCS were both above average at 51.1 ± 8.7 and 52.8± 10.8, respectively. Although this represents only65% of the study population, it provides good patientperception of recovery after aortic surgery, taking intoconsideration all measures of physical, mental, andemotional well being. This is consistent with earlierstudies reporting good quality of life with antegradecerebral perfusion compared to DHCA.18 Survival atmedial follow-up was 92% and is similar to otherstudies.3

NIRS provides an additional tool for the assessmentof cerebral protection and in this study has shownto be a reliable method of assessing cerebral per-fusion. Furthermore, Daubeney et al. have demon-strated good correlation between NIRS and jugular bulbsaturations,5,19 and other studies have shown goodcorrelations between rSO2 and magnetic resonancefindings,20 thus providing an alternative to invasivemeasurements of jugular bulb saturations. Therefore,NIRS has demonstrated to be a desirable intraopera-tive tool, providing noninvasive and continuous data.However, a more significant disadvantage as shownby Orihashi et al.6 is that cerebral infarction caused byarterial embolism was not demonstrated by a drop inrSO2. Similarly, rSO2 can remain unchanged if cerebralischemia is at a site other than the NIRS sampling area.Although a drop in rSO2 is likely to arise due to malper-fusion, which can be corrected intraoperatively, NIRScannot accurately differentiate this with other causesof ischemia such as atheroma or air emboli.

This study carries the risks inherent to the use of ret-rospective data and includes a relatively low risk smallpopulation. Furthermore, there is no control group forcomparison of NIRS data. The findings of this studyneed to be confirmed in larger studies randomized todifferent ACP strategies with continuous intraoperativeNIRS data. Although there is an association between

216 SENANAYAKE, ET AL.CEREBRAL PROTECTION DURING ANTEGRADE CEREBRAL PROTECTION

J CARD SURG2012;27:211-216

low rSO2 and TND, a statistical model has not beenmade to show statistical significance.

In conclusion, NIRS aids the management of cere-bral protection and provides a guide for adequate cere-bral perfusion. A unilateral approach to ACP with theaid of NIRS is associated with low neurological dys-functions and confirms adequate perfusion. The im-portance of maintaining circulation pressures to main-tain rSO2 above baseline is paramount for effectivecerebral protection. ACP with the aid of NIRS is associ-ated with good midterm quality of life and NIRS can beconfidently used as an adjunct for established cerebralprotection strategies.

REFERENCES

1. Ergin MA, Galla JD, Lansman L, et al: Hypothermiccirculatory arrest in operations on the thoracic aorta.Determinants of operative mortality and neurologic out-come. J Thorac Cardiovasc Surg 1994;107:788-797; dis-cussion 97-99.

2. Ueda Y, Miki S, Kusuhara K, et al: Surgical treatment ofaneurysm or dissection involving the ascending aorta andaortic arch, utilizing circulatory arrest and retrograde cere-bral perfusion. J Cardiovasc Surg (Torino) 1990;31:553-558.

3. Kazui T, Yamashita K, Washiyama N, et al: Usefulnessof antegrade selective cerebral perfusion during aorticarch operations. Ann Thorac Surg 2002;74:S1806-S1809;discussion S25-S32.

4. Bakhtiary F, Dogan S, Zierer A, et al: Antegrade cerebralperfusion for acute type A aortic dissection in 120 con-secutive patients. Ann Thorac Surg 2008;85:465-469.

5. Daubeney PE, Pilkington SN, Janke E, et al: Cerebraloxygenation measured by near-infrared spectroscopy:Comparison with jugular bulb oximetry. Ann Thorac Surg1996;61:930-934.

6. Orihashi K, Sueda T, Okada K, et al: Near-infraredspectroscopy for monitoring cerebral ischemia duringselective cerebral perfusion. Eur J Cardiothorac Surg2004;26:907-911.

7. Sakamoto T, Hatsuoka S, Stock UA, et al: Predictionof safe duration of hypothermic circulatory arrest bynear-infrared spectroscopy. J Thorac Cardiovasc Surg2001;122:339-350.

8. Olsson C, Thelin S: Regional cerebral saturation monitor-ing with near-infrared spectroscopy during selective an-tegrade cerebral perfusion: Diagnostic performance andrelationship to postoperative stroke. J Thorac CardiovascSurg 2006;131:371-379.

9. Reents W, Muellges W, Franke D, et al: Cerebral oxy-gen saturation assessed by near-infrared spectroscopyduring coronary artery bypass grafting and early post-operative cognitive function. Ann Thorac Surg 2002;74:109-114.

10. Wahr JA, Tremper KK, Samra S, et al: Near-infrared spec-troscopy: Theory and applications. J Cardiothorac VascAnesth 1996;10:406-418.

11. Di Eusanio M, Wesselink RM, Morshuis WJ, et al: Deephypothermic circulatory arrest and antegrade selectivecerebral perfusion during ascending aorta-hemiarch re-placement: A retrospective comparative study. J ThoracCardiovasc Surg 2003;125:849-854.

12. Immer FF, Moser B, Krahenbuhl ES, et al: Arterial ac-cess through the right subclavian artery in surgery ofthe aortic arch improves neurologic outcome and mid-term quality of life. Ann Thorac Surg 2008;85:1614-1618;discussion 8.

13. Ozatik MA, Kucuker SA, Tuluce H, et al: Neurocognitivefunctions after aortic arch repair with right brachial arteryperfusion. Ann Thorac Surg 2004;78:591-595.

14. Kucuker SA, Ozatik MA, Saritas A, et al: Arch repair withunilateral antegrade cerebral perfusion. Eur J Cardiotho-rac Surg 2005;27:638-643.

15. Rubio A, Hakami L, Munch F, et al: Noninvasive con-trol of adequate cerebral oxygenation during low-flowantegrade selective cerebral perfusion on adults and in-fants in the aortic arch surgery. J Card Surg 2008;23:474-479.

16. Hirofumi O, Otone E, Hiroshi I, et al: The effectivenessof regional cerebral oxygen saturation monitoring usingnear-infrared spectroscopy in carotid endarterectomy. JClin Neurosci 2003;10:79-83.

17. Kirkpatrick PJ, Lam J, Al-Rawi P, et al: Defining thresholdsfor critical ischemia by using near-infrared spectroscopyin the adult brain. J Neurosurg 1998;89:389-394.

18. Immer FF, Lippeck C, Barmettler H, et al: Improvementof quality of life after surgery on the thoracic aorta:Effect of antegrade cerebral perfusion and short dura-tion of deep hypothermic circulatory arrest. Circulation2004;110:II250-II255.

19. Kim MB, Ward DS, Cartwright CR, et al: Estimation ofjugular venous O2 saturation from cerebral oximetry or ar-terial O2 saturation during isocapnic hypoxia. J Clin MonitComput 2000;16:191-199.

20. Mehagnoul-Schipper DJ, van der Kallen BF, Colier WN,et al: Simultaneous measurements of cerebral oxygena-tion changes during brain activation by near-infrared spec-troscopy and functional magnetic resonance imaging inhealthy young and elderly subjects. Hum Brain Mapp2002;16:14-23.