Neuropsychological functioning 3—5 years after coronaryartery bypass grafting: does the pump make a difference?§

Nathalie Stroobant a,b,*, Guido van Nooten a, Dirk De Bacquer c,Yves Van Belleghem a, Guy Vingerhoets b

aCentre for Cardiac Surgery, University Hospital Ghent, Belgiumb Laboratory for Neuropsychology, Department of Internal Medicine - Section Neurology, Ghent University, Belgium

cDepartment of Public Health, Ghent University, Belgium

Received 10 September 2007; received in revised form 16 April 2008; accepted 5 May 2008; Available online 3 June 2008

Abstract

Objective: Neurocognitive sequelae after coronary artery bypass grafting (CABG) are frequently reported. The present study investigated thepossible difference between on- and off-pump CABG patients in a long-term perspective. Methods: We administered seven standardizedneuropsychological tests 1 day before, 6 days after, 6 months after and 3—5 years after CABG. We measured intra-operative high intensitytransient signals (HITS) as a reflection of embolic load by transcranial Doppler ultrasonography. Results:Of the 74 patients originally recruited, 54of the patients (73%) (mean age 59 years � 7.5 years) completed the neuropsychological battery. Cognitive decline was defined as 20% decline ontwo or more tests. Three to 5 years after the operation, 30% of the patients showed neurocognitive problems in domains of non-verbal immediatememory and attention, speed for visual search, visual attention and mental flexibility. Neither difference in frequency of cognitive dysfunctionnor a difference in evolution over the 5 year time period between on- and off-pump patients was observed. No significant predictors for cognitivedecline were found. Conclusions: Three to 5 years after the CABG surgery no difference in neurocognitive deficits is found between on- and off-pump CABG patients, indicating that the extracorporeal circulation (ECC) may not be the main cause of late neuropsychological decline.# 2008 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved.

Keywords: Coronary artery bypass grafting; Cognitive brain function; Neuropsychology; Cardiopulmonary bypass; Off-pump surgery

www.elsevier.com/locate/ejctsEuropean Journal of Cardio-thoracic Surgery 34 (2008) 396—401

1. Introduction

The neurological and neuropsychological outcomes ofcoronary artery bypass graft surgery (CABG) proceduresrange from the well-documented incidence of stroke to lessclearly defined and more subtle cognitive changes, whichinclude difficulties in memory, attention, concentration,vasoconstriction and speed of response [1].

The precise pathophysiological features of cognitiveimpairment are not certain but have been attributed to themicro-embolic, inflammatory and perfusion factors asso-ciated with extracorporeal circulation (ECC) [2,3]. Severalstudies using transcranial Doppler ultrasonography (TCD)have found an association between intra-operative cere-bral emboli, operationalized as high intensity transient

§ This research was supported by a grant from the Special Fund for ScientificResearch (University Ghent) and a postdoctoral research grant from the Fundfor Scientific Research Flanders, both awarded to Nathalie Stroobant.* Corresponding author. Address: Laboratory for Neuropsychology, Depart-

ment of Internal Medicine-Section Neurology, Ghent University Hospital, 4K3,De Pintelaan 185, B-9000 Ghent, Belgium. Tel.: +32 9 2404545;fax: +32 9 2404555.

E-mail address: Nathalie.Stroobant@ugent.be (N. Stroobant).

1010-7940/$ — see front matter # 2008 European Association for Cardio-Thoracic Sdoi:10.1016/j.ejcts.2008.05.001

signals (HITS), and neurocognitive problems [4,5]. CABGwithout ECC (off-pump surgery) on a beating heart is anincreasingly common procedure with a reduction incerebral embolic load [6] and an interesting alternativeto traditional revascularization avoiding some of themorbidity of on-pump CABG surgery. Nevertheless, recentstudies examining the cerebroprotective effect of off-pump procedures versus conventional CABG yieldedinconsistent results [7].

Although most studies of cognitive outcomes reportcognitive decline in on- and off-pump patients at theimmediate and early postoperative period [6,8], there isstill controversy regarding the degree and expected durationof these changes. Recently, the possibility of a delayed or latecognitive decline after CABG is given more attention. Someresearchers report a pattern of postoperative decline thatconsists of a relatively high frequency of early deteriorationwith short-term improvement within the first fewmonths andconsequent secondary decline after 5 years [9,10]. Incontrast, a more favorable long-term outcome was foundin the study of Mullges et al. [11] where none of the patientsshowed a neurocognitive decline after 4.5 years. Apart fromthe debate whether long-term cognitive decline over a

urgery. Published by Elsevier B.V. All rights reserved.

N. Stroobant et al. / European Journal of Cardio-thoracic Surgery 34 (2008) 396—401 397

certain time period is existing or not, none of these studiesinvestigated if on- and off-pump CABG patients differedseveral years after surgery.

In the present study we studied the longitudinalneurocognitive problems after CABG and aimed to find outwhether there is a relation between pre-, peri- andpostoperative variables and these late sequelae. We alsowanted to investigate the potential long-term difference inneuropsychological outcome between on- and off-pumpCABG.

2. Materials and methods

2.1. Patient selection, anesthesia, surgical managementand intra-operative HITS counting

Exclusion criteria were carotid artery stenosis (asdetermined by Duplex B mode Doppler scan), emergencyoperation, cerebrovascular, neurological, or psychiatricdisease, drug and/or alcohol addiction. The incidence ofdementia and other neurological disturbances with asso-ciated cognitive impairment increases dramatically withadvancing age. Therefore we excluded patients >70 years.Each patient who qualified for the study signed aninformed consent document that had been approved byour institutional ethical committee (University HospitalGhent). A Dutch version of the National Adult Reading Test[12] was used to obtain an estimated intelligence quotient(IQ).

All patients were potential candidates for off-pumpsurgery except in case of unstable angina, severe left mainstenosis (>95%) and/or hemodynamic instability, or anato-mical inaccessibility. In multiple vessel disease, completerevascularization, especially in the region of the distalcircumflex or right coronary artery, using the off-pumptechnique necessitates important prolonged tilting of theheart sometimes leading to hazardous hemodynamic instabil-ity. As a consequence, more patients with multiple distalanastomoses in the region of the circumflex and rightcoronary were selected to be included in the on-pump group.However, since the beginning of the study, the used types ofstabilizers have improved considerably, thus broadening theindications for off-pump surgery.

For details of the intra-operative HITS measurement andinsonation technique we refer to the methods described inearlier research [6,13].

2.2. Neuropsychological testing

One examiner (NS) administered a battery of sevenstandardized neuropsychological tests 1 day before (T1), 6days after (T2), 6 months after (T3), and 3—5 years after (T4)surgery. The battery included a broad range of cognitiveabilities (attention and concentration, verbal and non-verbalmemory, language, visuo-spatial functions, executive func-tions, motor and psychomotor speed). The tests used in thestudy are listed below.

(1) T

he Rey auditory verbal learning test (AVLT) assessesverbal memory. The measure retained is the total

number of words immediately recalled over the first fivetrials.

(2) T

he trail making test (TMT Part B) assesses speed forvisual search, attention and mental flexibility. The timeto completion is the measure taken.

(3) T

he grooved pegboard test (GPT) measures finger andhand dexterity. The sum of the times to completion ofthe left and right placement of all pegs is the measuretaken.

(4) T

he block taps test (TAPS) assesses non-verbal immediatememory and attention. The number of errors is themeasure retained.

(5) T

he line bisection test (LBT) assesses unilateral visualinattention. Total deviation from the true center is themeasure taken.

(6) T

he controlled oral word association test (COWAT,orthographic categories) assesses word fluency. Thetotal number of words (with four given letters) is themeasure retained.

(7) J

udgment of line orientation (JLO) examines the abilityto estimate angular relationships between line seg-ments. The measure retained is the correct number ofanswers.

2.3. Definition of cognitive decline

We used the 20%-definition (a decrease in two or moretests by at least 20% from the preoperative baseline) becauseit is believed to be the most sensitive method for identifyingpatients with signs of cognitive decline [14].

2.4. Statistical analysis

All continuous variables are expressed as mean (SD).Categorical variables are expressed as percent of thoseexhibiting the trait out of all patients for whom data wereavailable. The statistical analysis (SPSS/PC+; Chicago, IL)was based on two independent groups of patients (on-pump vs off-pump). We compared baseline characteristics,outcome measures, and complications between the twogroups with the Mann—Whitney test. We analyzed nominaldata with chi-square statistics if all expected cellfrequencies were greater or equal to five. Otherwise,Fisher’s exact test was used. Postoperative changes inabsolute neuropsychological test scores and in percentageswere compared to the preoperative baseline level withWilcoxon signed rank tests. We calculated associationsbetween number of HITS, age, IQ and neuropsychologicaldata by Spearman rho correlation coefficients. A mixedeffects model was used to study the evolution over timeand to observe differences between on- and off-pumppatients.

3. Results

3.1. Patient characteristics

Of the 74 patients originally recruited into the study, 54 ofthe patients (73%) completed the neuropsychological assess-ment at T1, T2, and T4 at the University Hospital Ghent

N. Stroobant et al. / European Journal of Cardio-thoracic Surgery 34 (2008) 396—401398

(Belgium). The group of 54 participants consisted of 48 menand 6 women, age range 42—70,mean age 59 years � 7.5, on-pump (n = 33; 61%), off-pump (n = 21, 39%). Reasons for notattending the long-term follow-up were as follows: 12patients refused (too busy/not interested), and 8 patientsdied. Causes of death were CVA (on-pump: n = 1; off-pump:n = 1), sudden death (on-pump: n = 2), traffic accident(on-pump: n = 1), cancer (on-pump: n = 1), perioperativecomplication (dental surgery) (on-pump: n = 1), suicide (on-pump: n = 1). The dropout group (n = 20) showed a sig-nificantly worse score for preoperative ( p = 0.013) andimmediate postoperative verbal memory (AVLT) ( p = 0.002)compared to the follow-up group. Six months after surgery(T3) 43 patients returned.

Demographic, pre-, peri-, and postoperative variables ofthe patient group are described in Table 1. Although therewas a trend towards a lower IQ in the off-pump group, meanIQs were comparable to the (lower) mean IQ level of thegeneral population. On-pump patients were comparable tooff-pump patients with the exception of a number ofvariables, indicating a surgical selection bias (see patient

Table 1Patient characteristics (SD)

Characteristics On-pump Off-pump p value

DemographicsMean age (years) 58.0 � 7.4 61.1 � 7.5 0.19Sex (% male) 87.9 90.5 0.6IQ estimate (NLV) 95.2 � 10.8 88.5 � 12.2 0.07State-anxiety 41.4 43.1 0.36

Preoperative risk factorsHypertension (%) 36.4 71.4 0.01Diabetes (%) 24.2 33.3 0.47Tobacco use (%) 69.7 90.5 0.07Recent MI (%) 24.2 23.8 0.97Ejection fraction (%) 68.7 � 12.3 75.9 � 11.1 0.04

Number of diseased vessels 0.021 14.32 21.2 38.13 78.8 47.6

BMI 27.3 � 4.9 28.6 � 4.1 0.20

Perioperative factorsNumber of grafts 0.001 (%) 4.82 (%) 19.03 (%) 33.3 52.44 (%) 45.5 23.85 (%) 18.26 (%) 3.0

Operation time (min) 191.5 � 36.3 165.4 � 38.2 0.01Anesthesia time 245.4 � 47.0 220.3 � 44.4 0.05Total number of emboli (n = 24) (n = 13)

478.5 � 499.8 97.0 � 106.5 0.001

Postoperative factors (3—5 years)Cardiac/vascular problemsCardiac (%) 6.1Peripheral (%) 12.1 9.5Arrhythmias (%) 12.1 9.5Angina (%) 6.1 9.5

CVA 4.8Current smoking 3.0 33.3 0.01

selection). Except from one CVA, none of our patients in thefollow-up group showed postoperative complications such asTIAs or strokes.

3.2. Pre- and postoperative neuropsychological testresults (group level)

Except for a higher preoperative word fluency score in theon-pump group, no significant differences between on- andoff-pump group were found (Mann—Whitney tests). We usedWilcoxon signed rank tests for the significance of each timepoint versus preoperative baseline level. Table 2 shows us thata significant deterioration was obtained 6 days after surgeryfor GPT (p < 0.001), TAPS (p < 0.001), LBT (p = 0.01) and JLO(p = 0.03). Six months after surgery a significant improvementwas obtained for AVLT (p < 0.001), GPT (p = 0.02) and COWAT(p = 0.002). Three to 5 years after surgery no significantdeterioration or improvements were observed.

We also calculated differences in neuropsychologicaloutcome compared to preoperative baseline level withchanges in percentages and obtained the same results(Wilcoxon signed rank tests). No differences were observedbetween the on- and off-pump groups.

A linear mixed effects model affirmed these results.Moreover, no interaction effect was found between type ofsurgery and time, thus indicating that the evolution in timewas not different between both surgical groups.

Table 2Neuropsychological test scores: absolute mean scores (SD), median and Wil-coxon signed rank tests between preoperative baseline level and each post-operative score

Test Time-point N Mean (SD) Median Significancevs baseline( p value)

AVLT T1 54 45.1 � 9.6 45.5T2 54 43.6 � 9.6 42.5 0.22T3 43 49.6 � 9.7 50 0.000T4 54 44.4 � 9.3 45 0.31

TMT T1 50 123.7 � 77.5 101.5T2 49 127.4 � 58.0 114 0.64T3 40 108.0 � 55.5 102 0.40T4 51 114.7 � 53.7 106 0.92

GPT T1 54 182.5 � 36.8 175.5T2 54 203.1 � 48.9 188 0.000T3 43 177.4 � 34.9 180 0.02T4 54 183.7 � 46.5 182 0.53

TAPS T1 54 36.0 � 18.3 34T2 54 45.0 � 20.0 41.5 0.000T3 43 34.6 � 20.3 31 0.72T4 54 38.0 � 19.4 34 0.33

LBT T1 53 61.9 � 25.7 58T2 53 68.4 � 25.2 63 0.01T3 42 60.2 � 23.7 54 0.51T4 54 56.9 � 21.8 55 0.12

COWAT T1 51 29.1 � 9.9 30T2 52 29.6 � 10.8 30.5 0.60T3 42 32.1 � 9.8 32 0.002T4 54 28.5 � 9.7 29 0.63

JLO T1 54 24.1 � 3.3 24.5T2 54 23.2 � 3.7 23 0.03T3 43 24.4 � 4.1 25 0.37T4 54 24.2 � 3.4 24 0.93

N. Stroobant et al. / European Journal of Cardio-thoracic Surgery 34 (2008) 396—401 399

3.3. Relations between neuropsychological test resultsand number of HITS, age and IQ (group level)

We correlated absolute neurocognitive test scores withage, IQ, and HITS. Age was correlated with worse scores onAVLT (r range between �0.35 and 0.56: p between 0.009 and<0.001), LBT (rT3 = �0.42) ( p = 0.006) and TMT (rT1 = 0.3,p = 0.036; rT4 = 0.32, p = 0.023), GPT (rT3 = 0.3, p = 0.04),and TAPS (rT1 = 0.27, p = 0.046) ( p = 0.042). A lower IQ wascorrelated with most tests at most time points. HITS andneurocognitive test scores were not correlated.

We correlated changes in percentages in neurocognitivetest scores with age, IQ, and HITS. Both age (r = �0.31,p = 0.025) and IQ (r = 0.34, p = 0.025) were correlated withJLO ( p = 0.042). HITS and neurocognitive test scores werenot correlated.

3.4. Pre- and postoperative neuropsychological testresults (individual level)

According to the 20% definition, 16/54 (30%) showedcognitive decline 3—5 years after the operation. 9/33 (27.3%)belonged to the on-pump group and 7/21 (33.3%) belonged tothe off-pump group. Neurocognitive problems were found inthe domains of non-verbal immediate memory and attention,speed for visual search, visual attention and mentalflexibility. No significant differences between the on- andthe off-pump group were found.

To investigate whether 3—5 year postoperative neurop-sychological dysfunction in the patients undergoing cardiacsurgery was associated with certain pre-, peri- and post-operative variables, we allocated patients to the group of nodeficit (n = 38, postoperative decline in �1 test) or deficitgroup (n = 16, postoperative decline in�2 tests). Both groupswere compared on a wide range of demographic, medical andsurgical data (age, IQ, years of education, depression scoreon different time assessments, anxiety score on differenttime assessments, number of emboli, BMI, left ventricleejection fraction, number of vessels, number of grafts,operation time, anesthesia time, ventilation time, stay onintensive care unit, smoking, diabetes, hyperuricemia,hyperlipidemia, hypertension, COPD, familial history, recentMI, atrial fibrillation, postoperative cardiac and non-cardiacmedical complications 3—5 years after surgery, CABG on- oroff-pump. None of these variables showed a significantdifference between both groups.

3.5. Relation between neuropsychological test resultsand number of HITS (individual level)

We tested whether patients showing significant cognitivedecline (as defined by the 20% definition) received significantmore HITS during operation than patients showing nosignificant decline. A Mann—Whitney test with total numberof HITS as dependent variable and absence/presence ofcognitive decline after 3—5 years as independent variableshowed no significant differences between both groups.

To examine which variables were predictive of thecognitive decline after 5 years following CABG, we performedan explorative multivariate regression analysis. The numberof independent variables was selected on the basis of

evidence in previous studies and limited to preserve anadequate ratio of participants per variable (HITS, age,cognitive decline after 6 days, IQ). No significant predictorswere found.

4. Discussion

The present study investigated neurocognitive sequelaeafter CABG in a long-term perspective that is of even greaterconcern than transient acute postoperative changes [15].Still, it is complicated to determine whether the long-termsymptoms are a consequence of the surgery, the result ofnormal aging in a population with cardiovascular risk factorsor a combination of these and other factors.

Our attrition rate of 27% is similar (or even better) thandropout rates of other long-term follow-up studies thatmention dropout rates of 33—46% [9,10,15]. Baseline data ofthe dropout patients were comparable to those of our studypatients, which argue against a major bias through dropouts.The only significant difference we found was the worseperformance of the dropout group on one memory task(AVLT).

Compared to 42% of the patients showing cognitivedecline after 5 years [9], our rate of 30% is more favorablebut disagrees with the results of Mullges et al. [11] where noECC-patient showed a long-term significant cognitivedecline. Strict control of vascular risk factors such ashypertension and hypercholesterolemia was suggested.The hypothesis of Mullges et al. [11] is interesting becauseit implies prevention of late cognitive decline throughstricter postoperative control of risk factors. Nevertheless,although the majority of our patient population also tooklipid-lowering medication cognitive decline was found. Thisissue remains to be further investigated.

In contrast to a more global screening like factor analysis[9] or conceptually defined constructs of domains [16], wepreferred documenting individual domains of cognitivefunction. Note that, because of the spontaneous recoveryof central nervous system functions, the frequency ofmeasurable deficits in any cognitive domain could decreaseover time, which is especially interesting for a long-termstudy. Measures of visuomotor and psychomotor speed arethemost frequently reported vulnerable domains [10,16] andthis was confirmed by the present study. In addition, wefound deficits to occur in domains of non-verbal immediatememory and attention, visual attention and mental flex-ibility. The potential impact of these clinically significantneuropsychological sequelae in a variety of the above-mentioned domains on the social, professional and recrea-tional daily life of 30% of the patients may not beunderestimated.

Many authors believe that the use of ECC has a specificdamaging effect on cerebral function [4,17]. In contrast tothe expectations, several studies found no difference in earlyneurological and neuropsychological outcomes between on-and off-pump patients [18—20]. More recent randomizedprospective studies [8,21] found more cognitive decline inthe on-pump group than in the off-pump group after 10 weeks[21] and 3 months [8] but the effects were limited andbecame negligible after 12 months [8]. In one of our own

N. Stroobant et al. / European Journal of Cardio-thoracic Surgery 34 (2008) 396—401400

previous studies [22] more favorable results were revealedfor the off-pump group after 6 months. This could suggestthat at least in the short term (up to 6 months after thesurgery) ECC is not favorable for the brain.

Our study was limited by the fact that, although weplanned a randomized study, the decision regarding thesurgical strategy (on-pump vs off-pump) was left to theindividual surgeon due to technical reasons. As an implica-tion, patients with more severe pathologies were included inthe on-pump group. However, when including the patient,the neuropsychologist was not aware of the surgicaltechnique that was going to be used.

Another potential bias was introduced by exclusioncriteria (e.g. neurological dysfunction and emergencysurgery). So, the true incidence of cognitive disability inthe entire population having bypass surgery thus might beunderestimated [15].

To our knowledge, this is one of the first reports aboutcognitive sequelae that compare the two surgical groups onthe long-term (more than 1 year after surgery). The mostimportant finding of this study is, that we found no significantdifference between on- and off-pump patients for thepercentage patients with cognitive decline, for the percen-tage cognitive impaired tests, for the evolution over time andfor the domain of impaired neurocognitive ability 3—5 yearsafter the surgery. However, our ability to detect a significantdifference in cognitive function between patients with andwithout off-pump might be limited by the relatively smallsample size of our study. Due to a limited power, falsenegatives could have occurred. Nevertheless, a recentlypublished large randomized trial was unable too todemonstrate any benefit from avoiding cardiopulmonarybypass on long-term cognitive outcomes [23].

It is suggested that the number of emboli in the brain hasbeen a surrogate marker for cerebrovascular disease [11].Generally less emboli are found in off-pump surgery [24].Moreover, the number of emboli has been associated withcognitive sequelae as well [4,5]. Other studies, like thepresent one, have failed to find any relationship [6]. Thecomposition and size of emboli as well as the ultimatelocation in the brain might be the best predictors instead ofthe absolute number [6,10]. Interesting for future perspec-tives, solid and gaseous cerebral HITS can now bedifferentiated with multi-frequency TCD [25].

The lack of an association between emboli and cognitiveimpairment together with the lack of difference in cognitivedecline between the on- and off-pump group could suggestthat chronic vascular disease, other than ECC, may underlielonger-term cognitive loss in some of the patients. Furtherresearch on this topic is needed.

Besides emboli during surgery [10] predictors of long-termcognitive decline in other studies, specifically after CABG,were older age [23], fewer years of education [23], higherbaseline score [9] cognitive decline at time of discharge[9,10], and degree of recovery between discharge andfollow-up testing at 6—8 weeks after surgery [10]. Themajority of these variables reflect more general processesrather than the ones specific to CABG [10]. In our study nospecific pre-, peri- and postoperative characteristics for theimpaired group were found. Neither did we find predictors forcognitive decline.

The increased vulnerability for cognitive decline ofolder patients has been clear in many other short-termstudies. Long-term research yielded more inconsistentresults [9,10,16]. We found age to be correlated with themajority of neuropsychological tests. In common with twoother reported studies that have followed CABG patientsup to 5 years, an important limitation in the present studyis that no control group was included to assess the impactof aging in relation to changes in cognitive performance[9,16]. Controlled studies are needed to demonstratewhether the cognitive decline is caused by the surgeryitself or by age-related changes or underlying cerebrovas-cular disease [15].

To conclude, cognitive decline was found in one third ofthe CABG patients 3—5 years after surgery in variousneurocognitive domains. No differences were found betweenthe on- and off-pump patients, indicating that ECC may notbe the only cause of late cognitive deficits. Further researchwith control groups is recommended.

References

[1] Newman SP, Stygall J. Neuropsychological outcome following cardiacsurgery. In: Newman SP, Harrison MJG, editors. The brain and cardiacsurgery: causes of neurological complications and their prevention.Amsterdam: Harwood; 2000. p. 21—49.

[2] Blauth CI. Macroemboli and microemboli during cardiopulmonary bypass.Ann Thorac Surg 1995;59:1300—3.

[3] Harrison MJ. Neurologic complications of coronary artery bypass grafting:diffuse or focal ischemia? Ann Thorac Surg 1995;59:1356—8.

[4] Pugsley W, Klinger L, Paschalis C, Treasure T, Harrison M, Newman S. Theimpact of microemboli during cardiopulmonary bypass on neuropsycho-logical functioning. Stroke 1994;25:1393—9.

[5] Clark RE, Brillman J, Davis DA, Lovell MR, Price TR, Magovern GJ.Microemboli during coronary artery bypass grafting. Genesis and effecton outcome. J Thorac Cardiovasc Surg 1995;109:249—57.

[6] Stroobant N, Van Nooten G, Van Belleghem Y, Vingerhoets G. Relationbetween neurocognitive impairment, embolic load, and cerebrovascularreactivity following on- and off-pump coronary artery bypass grafting.Chest 2005;127:1967—76.

[7] Stump DA, Rorie KD, Jones TJ. Does off-pump coronary artery bypasssurgery reduce the risk of brain injury? Heart Surg Forum 2001;4(Suppl.1):S14—8.

[8] Van Dijk D, Jansen EWL, Hijman R, Nierich AP, Diephuis J, Moons KGM,Lahpor JR, Borst C, Keizer AMA, Nathoe HM, Grobbe DE, De JaegerePPT, Kalkman CJ. Cognitive outcome after off-pump and on-pumpcoronary artery bypass graft surgery—a randomized trial. JAMA2002;287:1405—12.

[9] Newman MF, Kirchner JL, Phillips-Bute B, Gaver V, Grocott H, Jones RH,Mark DB, Reves JG, Blumenthal JA. Longitudinal assessment of neuro-cognitive function after coronary-artery bypass surgery. N Engl J Med2001;344:395—402.

[10] Stygall J, Newman SP, Fitzgerald G, Steed L, Mulligan K, Arrowsmith JE,Pugsley W, Humphries S, Harrison MJ. Cognitive change 5 years aftercoronary artery bypass surgery. Health Psychol 2003;22:579—86.

[11] Mullges W, Babin-Ebell J, Reents W, Toyka KV. Cognitive performanceafter coronary artery bypass grafting: a follow-up study. Neurology2002;59:741—3.

[12] Schmand B, Lindeboom J, van Harskamp F. Nederlandse Leestest voorVolwassenen (Dutch Adult Reading Test). Lisse, the Netherlands: Swetsand Zeitlinger BV; 1992.

[13] Ringelstein EB, Kahlscheuer B, Niggemeyer E, Otis SM. TranscranialDoppler sonography: anatomical landmarks and normal velocity values.Ultrasound Med Biol 1990;16:745—61.

[14] Mahanna EP, Blumenthal JA, White WD, Croughwell ND, Clancy CP, SmithLR, Newman MF. Defining neuropsychological dysfunction after coronaryartery bypass grafting. Ann Thorac Surg 1996;61:1342—7.

[15] Selnes OA, Grega MA, Borowicz LM, Royall RM, McKhann GM, BaumgartnerWA. Cognitive changes with coronary artery disease: a prospective study

N. Stroobant et al. / European Journal of Cardio-thoracic Surgery 34 (2008) 396—401 401

of coronary artery bypass graft patients and nonsurgical controls. AnnThorac Surg 2003;75:1377—86.

[16] Selnes OA, Royall RM, Grega MA, Borowicz L, Quaskey S, McKhann GM.Cognitive changes 5 years after coronary artery bypass grafting: is thereevidence of late decline? Arch Neurol 2001;58:598—604.

[17] Murkin JM. Neurologic dysfunction after CABG or valvular surgery: is themedium the miscreant? Anesth Analg 1993;76:213—4.

[18] Taggart DP, Browne SM, Halligan PW, Wade DT. Is cardiopulmonary bypassstill the cause of cognitive dysfunction after cardiac operations? J ThoracCardiovasc Surg 1999;118:414—20.

[19] Lloyd CT, Ascione R, Underwood MJ, Gardner F, Black A, Angelini GD.Serum S-100 protein release and neuropsychologic outcome during cor-onary revascularization on the beating heart: a prospective randomizedstudy. J Thorac Cardiovasc Surg 2000;119:148—54.

[20] Jensen BO, Hughes P, Rasmussen LS, Pedersen PU, Steinbruchel DA.Cognitive outcomes in elderly high-risk patients after off-pump versusconventional coronary artery bypass grafting: a randomized trial. Circu-lation 2006;113:2790—5.

[21] Zamvar V, Williams D, Hall J, Payne N, Cann C, Young K, Karthikeyan S.Assessment of neurocognitive impairment after off-pump and on-pumptechniques for coronary artery bypass graft surgery: prospective rando-mized controlled trial. BMJ 2002;325:1268—72.

[22] Stroobant N, Van Nooten G, Van Belleghem Y, Vingerhoets G. Short-termand long-term neurocognitive outcome in on-pump versus off-pumpCABG. Eur J Cardiothorac Surg 2002;22:559—64.

[23] Van Dijk D, Spoor M, Hijman R, Nathoe HM, Borst C, Jansen EWL, GrobbeeDE, de Jaegere PPT, Kalkman CJ. Cognitive and cardiac outcomes 5 yearsafter off-pump vs on-pump coronary artery bypass graft surgery. JAMA2007;297:701—8.

[24] Bowles BJ, Lee JD, Dan CR, Taoka SN, Johnson EW, Lau EM, Nekomoto K.Coronary artery bypass performed without the use of cardiopulmonarybypass is associated with reduced cerebral microemboli and improvedclinical results. Chest 2001;119:25—30.

[25] Russell D, Brucher R. Online automatic discrimination between solid andgaseous cerebral microemboli with the first multifrequency transcranialDoppler. Stroke 2002;33:1975—80.