Elderly patients have an increased risk of neurologic

Craniocervical and Aortic Atherosclerosis as Neurologic Risk Factors in Coronary Surgery Tomoko Goto, MD, Tomoko Baba, MD, Atsushi Yoshitake, MD, Yoshihiro Shibata, MD, Masashi Ura, MD, and Ryuzo Sakata, MD Deartments of Anesthesiology and Cardiovascular Surgery, Kumamoto Chuo Hosital, Kumamoto, Jaan Background. Advanced age is associated with increased systemic atherosclerosis and is a consistent neurologic risk factor after coronary artery byass grafting (CABG). Methods. We studied rosectively whether varying degrees of a total atherosclerotic score derived from the brain, carotid arteries, and ascending aorta redicted ostoerative neurosychologic (NP) dysfunction and stroke in 177 elderly atients (> 60 years) undergoing CABG. Results. Grou L (low total atherosclerotic score) had rates of NP dysfunction of 25% and 4%, grou I (intermediate) had rates of 33% and 22%, and grou H (high) had rates of 79% and 43% on ostoerative days 1 and 7, resectively ( < 0.001). The incidence of stroke was higher in grou H (14.3%) than in grous I and L (7.8% and 0.9%; 0.013). Stewise logistic regression analysis demonstrated the significant redictors of NP dysfunction on ostoerative day 7 to be total atherosclerotic score, eriheral vascular disease, and diabetes mellitus, and those of stroke to be total atherosclerotic score, eriheral vascular disease, and hyerliidemia. Conclusions. Perioerative evaluation of craniocervical and aortic atherosclerosis is useful to identify a high-risk atient at ostoerative NP dysfunction and stroke after CABG. (Ann Thorac Surg 2000;69:834 40) 2000 by The Society of Thoracic Surgeons Elderly atients have an increased risk of neurologic dysfunction after coronary artery byass grafting (CABG) [1]. Given their higher revalence of systemic atherosclerosis, cerebrovascular disease, and underlying cognitive dysfunction, elderly atients are at increased risk of cerebral injury. Our revious study reorted that small cerebral infarctions were common in elderly atients undergoing CABG [2]. Cerebral ischemic disorders and carotid stenosis tend to imair cerebral autoregulation mechanisms [3, 4], which may cause hyoerfusion during cardioulmonary byass (CPB). Advanced atheromatous disease of the ascending aorta is a major risk factor for erioerative stroke [5 7]. However, there are no data to relate severity of craniocervical and aortic atherosclerosis to ostoerative neurosychologic (NP) dysfunction and stroke after CPB in combination in a redictive model. We focused on those three major risk factors for erioerative cerebral ischemia in elective CABG, which is erformed in atients older than 60 years. Thus we studied rosectively whether varying degrees of the total atherosclerotic score of the brain, carotid arteries, and ascending aorta by using magnetic resonance imaging (MRI), magnetic resonance angiograhy (MRA), and echocardiograhy redicted ostoerative NP dysfunction and stroke in atients undergoing CABG. Acceted for ublication Se 13, 1999. Address rerint requests to Dr Goto, Deartment of Anesthesiology, Kumamoto Chuo Hosital, 96 Tainoshima Tamukaemachi, Kumamoto 862-0965, Jaan; e-mail: togoto@bronze.ocn.ne.j. Material and Methods After institutional review board aroval, we obtained written informed consent from 177 Jaanese atients older than 60 years who underwent elective CABG between January 1995 and Aril 1997 at Kumamoto Chuo Hosital. There were 148 asymtomatic atients and 29 with reexisting cerebrovascular disease (CVD) who had had strokes or transient ischemic attack (TIA). Patients who underwent cardiac valve relacement or ventricular aneurysmectomy were excluded to eliminate the effects of intracardiac air embolism and left ventricular thrombus. Four atients who had had ostoerative rolonged intubation ( 24 hours) induced by resiratory comlications were removed from the study. Demograhic and historic data were defined as follows: age, sex, history of hyertension requiring treatment with antihyertensive medication, history of diabetes mellitus (insulin or oral hyoglycemic theray or a carbohydrate-restricted diet), hyerliidemia (total cholesterol 240 mg/dl or triglyceride 150 mg/dl or antihyerliidemic theray), history of eriheral vascular disease (PVD), and history of CVD with strokes or TIA. Neurologic and Cerebral Evaluation Preoerative cerebrovascular evaluation was made by cerebral MRI and cervical MRA (0.5 T Suerconductive MR unit, Philis, Amsterdam, The Netherlands) as described reviously [2]. The scatter reresenting ischemic changes or infarctions in the brain on MRI was classified as 0 (almost normal); 1 (leukoaraiosis); 2 (some small infarctions); and 3 (multile small infarctions or broad 2000 by The Society of Thoracic Surgeons 0003-4975/00/$20.00 Published by Elsevier Science Inc PII S0003-4975(99)01421-6

Ann Thorac Surg GOTO ET AL 2000;69:834 40 ATHEROSCLEROSIS AND NEUROLOGIC RISK 835 infarctions). The degree of stenosis in the carotid arteries was graded by MRA as 0 (almost normal); 1 (mild narrowing of 50%); 2 (moderate narrowing of 50% to 75%); and 3 (severe narrowing of 75% or obstruction). The lesions on MRI and MRA were evaluated indeendently by two radiologists who were blinded from reoerative risk factors and symtoms. Cognitive status was measured using the Hasegawa dementia score (HDS; score 0 to 30, with 30 best), a modification of the Mini-Mental Status (MMS) Examination, an NP test widely utilized in Jaan. Scores less than 24 on the HDS are indicative of cognitive imairment (equivalent to 24 on the MMS). The HDS consists of nine questions that test three orientations, three recall memory functions, attention, common knowledge and calculations. All atients received neurologic and NP examinations before oeration by one of three investigators (T.G., T.B., Y.S.) and on the 1st and 7th ostoerative days by the same investigator (T.B.). NP examiners were blinded to atients reoerative neurologic findings and atherosclerotic score. NP dysfunction for each ostoerative assessment was defined as a decrease in erformance from baseline greater than or equal to 4 (equal to 2 standard deviations in baseline). Postoerative MRI or comuted tomograhy (CT) was erformed only on atients with neurologic deficits lasting more than 24 hours or the decrement of HDS from baseline greater than or equal to 8 on ostoerative day 7. Patients with new ostoerative neurologic symtoms and ositive findings on ostoerative MRI or CT of the brain were examined by a staff neurologist to confirm intraoerative and ostoerative stroke. Total stroke was the sum of intraoerative stroke and ostoerative stroke. Patient Management Anesthesia and atient management during CPB were conducted as described reviously [2]. Distal coronary anastomoses and roximal anastomoses were erformed during a single aortic cross-clam. To limit release of atherosclerotic emboli from the ascending aorta, we changed the standard site for clam and cannulation if there was greater than or equal to 3 mm intimal thickening of the aorta. All oerations were erformed by the same oerator (R.S.). No attemt was made to standardize the sedative or analgesic agents in the ostoerative eriod. Evaluation of Atherosclerotic Lesions in the Ascending Aorta We evaluated atherosclerotic lesions in the ascending aorta by echocardiograhy (Sonolayer SSA-260A, Toshiba, Tokyo, Jaan) using an eiaortic robe (linear, 7.5 MHz, IOE 702V). Grading of the degree of atherosclerosis in the ascending aorta was done according to the modified Wareing and associates [5] method: 0 (almost normal); 1 (mild, 3 mm intimal thickening); 2 (moderate, 3 mm intimal thickening involving one segment of the ascending aorta); 3 (severe or diffuse, 3 mm intimal thickening involving two or all three segments). Statistical Analyses To relate combined severity of the three atherosclerotic scores derived from the brain, carotid arteries, and ascending aorta to ostoerative NP dysfunction and stroke, the 177 atients were divided into three grous according to comlications with grade 3 scores. Grou L was low grade score (no grade 3 atherosclerotic scores, n 112). Grous I and H consisted of atients with intermediate and high grade scores (one grade 3 atherosclerotic score, n 51; two or three grade 3 atherosclerotic scores, n 14, resectively). Statistical comarison between the three grous was erformed using the 2 test and one-way analysis of variance (ANOVA). Pearson s correlation coefficient was calculated to determine associations between atherosclerotic diseases in the ascending aorta, brain, carotid arteries, and eriheral arteries. To assess the risk of NP dysfunction on ostoerative day 7 and total stroke, we used univariate analysis ( 2 test or unaired t test, significance level, 0.05) and stewise logistic regression analysis by keeing redictors with values no greater than 0.15. We included hyertension, diabetes mellitus, hyerliidemia, PVD, CVD (absence 0, resence 1), age, aortic clam time, CPB time, the three atherosclerotic scores (MRI, MRA, and aorta), the reoerative value of HDS as ordinal value, and total atherosclerotic score (none of grade 3 1, one of grade 3 2, two or three of grade 3 3) into the regression model to select the best set. Odds ratios were calculated for each factor in the resence of the others in the final model. All statistical analyses were comleted using the SAS Institute Inc (Cary, NC) statistical ackage, version 6.12. Results Patient demograhics, oerative characteristics, and degrees of atherosclerosis of the three grous are summarized in Table 1. The incidence of reexisting CVD was higher in grou H than the other two grous, whereas the incidence of PVD was higher in grou I than the other two grous. Aortic clam and CPB time were lower in grou H than in the other two grous, as 7 atients in grou H (n 14) underwent CABG without aortic clam. In grou H (n 14), there were 13 (93%) atients with multile small infarctions or broad infarction; 6 (43%) atients with carotid stenosis or obstruction; and 11 (79%) atients exhibited severe or diffuse atherosclerosis of the ascending aorta. As shown in Table 1, 22 (43%) atients in grou I and 11 (79%) atients in grou H had severe or diffuse atherosclerosis in the ascending aorta. Artery grafts were used in all 33 atients, and vein grafts were also used in 10 of these atients. Of these 33 atients, 19 underwent CABG at an alternate site of aortic cannulation and claming that was identified by ultrasonograhy. In 14 atients with severe or diffuse atherosclerosis, consisting of 7 atients each in both grou I and grou H; 6 atients underwent CABG on beating, 8 atients underwent

836 GOTO ET AL Ann Thorac Surg ATHEROSCLEROSIS AND NEUROLOGIC RISK 2000;69:834 40 Table 1. Preoerative and Oerative Characteristics of the Three Grous a Total Atherosclerotic Score Characteristic Grou L (n 112) Grou I (n 51) Grou H (n 14) b Age (y) c 70.2 5.0 70.3 5.2 70.6 4.1 0.966 Gender (male/female) 73/39 33/18 11/3 0.589 Hyertension 72 (64) 41 (80) 10 (71) 0.116 Diabetes mellitus 45 (40) 15 (29) 4 (29) 0.343 Hyerliidemia 47 (42) 22 (43) 2 (14) 0.120 PVD 3 (3) 8 (16) 1 (7) 0.009 Preexisting CVD 13 (12) 8 (16) 8 (57) 0.001 Aortic clam time (min) c 77.5 16.0 60.3 27.8 24.7 27.3 0.001 CPB time (min) c 108.8 21.5 103.3 20.3 90.2 20.4 0.006 MRI (brain) Normal or leukoaraiosis 67 (60) 9 (18) 0 0.001 Some small infarctions 45 (40) 15 (29) 1 (7) Multile small infarctions or broad infarctions 0 27 (53) 13 (93) MRA (carotid arteries) Normal or mild 108 (96) 48 (94) 6 (43) 0.001 Moderate 4 (4) 1 (2) 2 (14) Severe or obstruction 0 2 (4) 6 (43) Ascending aorta Normal or mild 96 (86) 24 (47) 1 (7) 0.001 Moderate 16 (14) 5 (10) 2 (14) Severe or diffuse 0 22 (43) 11 (79) a s in arentheses reresent ercentages within grou. b s refer to comarisons between the three grous. c Exressed as mean standard deviation. CPB cardioulmonary byass; CVD cerebrovascular disease; H high; I intermediate; L low; MRA magnetic resonance angiograhy; MRI magnetic resonance imaging; PVD eriheral vascular disease. CABG on hyothermic fibrillation or hyothermic circulatory arrest without aortic claming, and 5 atients with femoral artery cannulation underwent CABG. Of the 177 atients (age, 70.3 5.0 years), NP dysfunction occurred in 56 (32%) and 21 (12%) atients, resectively, on ostoerative days 1 and 7. The HDS was lower in grou H than in the other two grous reoeratively and on ostoerative days 1 or 7 ( 0.001; Table 2). Grou L had low rates of ostoerative NP dysfunction of 25% and 4%, grou I had rates of 33% and 22%, and grou H had rates of 79% and 43% on ostoerative days 1 and 7, resectively ( 0.001). Intraoerative stroke occurred in 4 atients of grous I and H who had severe or diffuse atherosclerosis of the ascending aorta: 2 in grou I had small infarctions and PVD with aortic clam and weakness of the uer extremities ostoeratively. Table 2. Postoerative Neurologic Comlications in the Three Grous a Total Atherosclerotic Score Grou I (n 51) Comlication Grou L (n 112) Grou H (n 14) Overall (n 177) b HDS c Preoerative 27.4 2.0 26.5 2.6 24.7 4.2 27.0 2.5 0.001 Postoerative day 1 25.5 3.3 23.7 4.5 18.0 8.8 24.4 4.8 0.001 Postoerative day 7 27.4 2.5 24.7 5.3 19.0 10.3 26.0 5.0 0.001 NP dysfunction Postoerative day 1 28 (25) 17 (33) 11 (79) 56 (32) 0.001 Postoerative day 7 4 (4) 11 (22) 6 (43) 21 (12) 0.001 Intraoerative stroke 0 2 (3.9) 2 (14.3) 4 (2.3) 0.002 Postoerative stroke 1 (0.9) 2 (3.9) 0 3 (1.7) 0.345 Total stroke 1 (0.9) 4 (7.8) 2 (14.3) 7 (4.0) 0.013 a s in arentheses reresent ercentages within grou. b s refer to comarisons between the three grous. c Exressed as mean standard deviation. H high; HDS Hasegawa dementia scale; I intermediate; L low; NP neurosychologic.

Ann Thorac Surg GOTO ET AL 2000;69:834 40 ATHEROSCLEROSIS AND NEUROLOGIC RISK 837 Table 3. Univariate Analysis Between Three Study Factors and Neurosychologic Dysfunction and Total Stroke Neurosychologic Dysfunction Total Stroke Variable Yes (n 21) No (n 156) Yes (n 7) No (n 170) Age (y) 71.6 4.4 70.1 5.1 0.202 69.7 5.0 70.3 5.0 0.769 Aortic clam time (min) 52.5 29.8 70.5 24.7 0.003 47.1 33.6 69.3 25.3 0.027 Cardioulmonary byass time (min) 100.4 20.6 106.5 21.7 0.230 96.4 12.3 106.1 21.8 0.245 The remaining 2 atients in grou H had multile small infarctions without aortic clam and induced unconsciousness ostoeratively. Postoerative CT in atients with intraoerative stroke showed multile infarctions due to emboli from the ascending aorta. In grous L and I, 3 atients develoed focal deficits on the 2nd or 3rd ostoerative days: 2 had hemiaresis and 1 had homonymous hemianosia. The incidence of total stroke was higher in grou H (14.3%) than in grous I and L (7.8% and 0.9%; 0.013). The outcome of stroke resulted in severe disability in 3 atients and mild disability in 2 atients; the remaining 2 atients had no serious functional disability. Six atients with a decrement of HDS from baseline greater than or equal to 8 on ostoerative day 7 had no new findings on ostoerative CT. In the univariate analysis, there were no significant differences in age and CPB time, whereas aortic clam time was shorter in atients with NP dysfunction on ostoerative day 7 and total stroke than in those without (Table 3). NP dysfunction on ostoerative day 7 and total stoke were not associated with reoerative risk factors, such as age, sex, hyertension, diabetes mellitus, hyerliidemia, or revious CVD. PVD was associated only with total stoke ( 0.002; Table 4). The correlates of NP dysfunction on ostoerative day 7 and total stoke were carotid stenosis, severe aortic atherosclerosis, and total atherosclerotic score, whereas only multile small infarctions was correlated with NP dysfunction on ostoerative day 7. The severity of atherosclerosis in the ascending aorta and the brain was significantly correlated with the grade of the carotid stenosis (r 0.171, 0.023, and r 0.314, 0.001, resectively). The severity of atherosclerosis in the ascending aorta and carotid stenosis were significantly correlated with PVD (r 0.199, 0.008, and r 0.201, 0.007, resectively). In the multivariate analysis, some covariates significant in single variate models droed out, whereas diabetes mellitus and hyerliidemia, which were not significant in single variate models, became significantly associated with NP dysfunction on ostoerative day 7 and total stroke (Table 5). Stewise logistic regression analysis revealed the simultaneously determined redictive variables of NP dysfunction on ostoerative day 7 and total stroke. The variables significantly associated with NP dysfunction and total stroke in the final models were total atherosclerotic score (OR 5.063, 0.001), PVD (OR 3.230, 0.111), and diabetes mellitus (OR 2.593, 0.078) in NP dysfunction on ostoerative day 7; and total atherosclerotic score (OR 6.498 0.010), PVD (OR 10.718, 0.010), and hyerliidemia (OR 8.324, 0.036) in total stroke. Comment The incidence of NP dysfunction and stroke after CPB was demonstrated to increase in atients who had a higher total atherosclerotic score. Multivariate analysis confirmed the indeendent risk of NP dysfunction to be total atherosclerotic score, PVD, and diabetes mellitus, and stroke was redicted by total atherosclerotic score, PVD, and hyerliidemia. The overall incidence of ostoerative NP dysfunction in our study was lower than the 69% to 79% rate reorted in revious studies [7 9]. Variations in the incidence of ostoerative NP dysfunction after CPB among the different studies may have been due to differences in the time of examination, methods, and definition of NP dysfunction. We determined NP dysfunction as a decrease in erformance from baseline by 2 standard deviations, whereas revious studies have used the oulation standard deviations of baseline reoerative tests [8]. Therefore, unlike Shaw and coworkers [8], we found a lower incidence of NP dysfunction in this study. Murkin and colleagues [9] reorted that 75% of CABG atients had cognitive dysfunction at 1 week and the rate was reduced to 33% at 2 months. The HDS and MMS do not require written resonses and have been shown to be reliable, valid, and concise. However, NP tests have been recommended as two or more cognitive tests including attention, concentration, erceived motor seed, and memory [10]. Further studies are needed to assess the correlation between systemic atherosclerosis and ostoerative NP dysfunction by long-term follow-u studies and the use of several NP tests. Small cerebral infarctions are common in elderly atients undergoing CABG [2]. Most atients with multile small infarctions are asymtomatic and manifest no clinical signs. However, reoerative HDS was lower in atients with a higher total score for atherosclerotic disease. Yoshitake and coworkers [11] showed that asymtomatic small infarctions were an imortant factor in the develoment of vascular dementia in a 7-year follow-u study of normal elderly residents of Hisayama town. In our study, the HDS score of grou H aeared to redict the develoment of asymtomatic multile small infarctions. Because NP tests such as the HDS are considered to reflect an early stage of cognitive imair-

838 GOTO ET AL Ann Thorac Surg ATHEROSCLEROSIS AND NEUROLOGIC RISK 2000;69:834 40 Table 4. Univariate Analysis Between Ten Study Factors and Neurosychologic Dysfunction and Total Stroke Neurosychologic Dysfunction (ostoerative day) Total Stroke Variable No. of Patients No. % No. % Gender Female 60 6 10.0 0.761 0 0 0.127 Male 117 15 12.8 7 6.0 Hyertension No 54 4 7.4 0.336 1 1.9 0.594 Yes 123 17 13.8 6 4.9 Diabetes mellitus No 113 11 9.7 0.356 6 5.3 0.408 Yes 64 10 15.6 1 1.6 Hyerliidemia No 106 14 13.2 0.661 2 1.9 0.183 Yes 71 7 9.9 5 7.0 PVD No 165 17 10.3 0.055 4. 2.4 0.002 Yes 12 4 33.3 3 25.0 Preexisting CVD No 148 16 10.8 0.506 4 2.7 0.159 Yes 29 5 17.2 3 10.3 Ascending aorta Normal or mild 121 9 7.4 0.001 3 2.5 0.024 Moderate 23 2 8.7 0 0 Severe or diffuse 33 10 30.3 4 12.1 MRI (brain) Normal or leukoaraiosis 76 3 3.9 0.004 0 0 0.063 Some small infarctions 61 8 13.1 4 6.6 Multile small infarctions 40 10 25.0 3 7.5 or broad infarctions MRA (carotid arteries) Normal or mild 162 15 9.3 0.001 5 3.1 0.007 Moderate 7 1 14.3 0 0 Severe 8 5 62.5 2 25.0 Total atherosclerotic score Low 112 4 3.6 0.001 1 0.9 0.013 Intermediate 51 11 21.6 4 7.8 High 14 6 42.9 2 14.3 CVD cardiovascular disease; MRA magnetic resonance angiograhy; MRI magnetic resonance imaging; PVD eriheral vascular disease. ment they may be useful for identifying the onset of cognitive imairment in the elderly. In addition, an NP test may serve as an indicator of neurologic and NP dysfunction after CPB. Carotid stenosis was another statistically significant risk factor for NP dysfunction. A recent study suggested that carotid disease is a redictor of erioerative stroke [12]. Although atients with severe stenosis of the carotid arteries have not been at increased risk for stroke during CABG [13], distal hemodynamic deficiencies deend, to a large degree, on the resence of collateral flow in these atients. Besides a reduced cerebral blood flow (CBF), cerebrovascular reactivity is frequently imaired distal to areas of severe stenosis or occlusion of the cervical carotid artery [4]. Several studies also reorted that in atients with asymtomatic small infarctions CBF reactivity to carbon dioxide was significantly lower than in normal controls [3]. However, there are no data to define otimal mean arterial ressure (MAP) during CPB in these atients. Further studies should examine whether higher MAP may reduce the incidence of ostoerative NP dysfunction and stroke after CPB in atients with multile small infarctions and carotid stenosis. Atherosclerotic disease of the ascending aorta was a significant factor for NP dysfunction and stroke after CPB [5 7]. Microemboli have been imlicated in the cause of NP dysfunction after CPB. Many studies reorted that NP deficits after routine CPB were related to the number of microemboli delivered during the surgical rocedure [14, 15]. In an imortant athologic study, Moody and

Ann Thorac Surg GOTO ET AL 2000;69:834 40 ATHEROSCLEROSIS AND NEUROLOGIC RISK 839 Table 5. Multivariate Analysis of Neurosychologic Dysfunction on Postoerative Day 7 and Total Stroke Variable Regression Coefficient Standard Error Odds Ratio 95% Confidence Interval NP dysfunction Intercet 6.234 1.226 0.001 Total score 1.622 0.376 0.001 5.063 2.423 10.578 PVD 1.173 0.736 0.111 3.230 0.763 13.673 Diabetes mellitus 0.953 0.541 0.078 2.593 0.898 7.487 Total stroke Intercet 10.374 2.774 0.001 Total score 1.871 0.728 0.010 6.498 1.559 27.083 PVD 2.372 0.926 0.010 10.718 1.745 65.816 Hyerliidemia 2.119 1.011 0.036 8.324 1.147 60.408 NP neurosychologic; PVD eriheral vascular disease. colleagues [16] demonstrated a high ercentage of small caillary and arteriolar dilatations in atients undergoing CPB. Many investigators reorted that microemboli occurred during CPB, esecially aortic cannulation, claming, and declaming [14, 17]. These data suggested that microdebris from the ascending aorta induced NP dysfunction after CPB. On the other hand, those with the greatest number of microemboli generated by the CPB system aeared to tolerate them well, but had greater decrements in cognitive function than those with fewer microemboli counts [15]. Our data suggested that microemboli from the ascending aorta may have lodged in greater numbers in the arteriolar vasculature with cerebral atherosclerosis. In our multivariate analysis, diabetes mellitus became a significant factor associated with NP dysfunction on the 7th ostoerative day, whereas hyerliidemia was a significant factor with total stroke. Although the association between hyerliidemia and stroke is unclear, hyerliidemia has been associated with carotid artery atherosclerosis in several studies that used ultrasonograhy [18]. This study used a small number of atients to identify the redictors for stroke. Thus, further study is needed to determine the relation between hyerliidemia and stroke. On the other hand, diabetes mellitus atients are unable to comensate, when necessary, with an increased CBF during and after CPB [19]. In addition, it is conceivable that diffuse damage to the cerebral microvasculature is the mechanism of increased NP dysfunction in diabetes mellitus after CPB. Our univariate and multivariate analyses showed that total atherosclerotic score was a significant redictive factor for NP dysfunction and total stroke after CPB. In our series of 177 atients older than 60 years who underwent CABG, 14 (grou H, 7.9%) had severe systemic atherosclerosis. PVD was a significant factor associated with NP dysfunction and total stroke. We found that PVD was correlated with atherosclerotic disease of the ascending aorta and carotid stenosis. Criqui and associates [20] reorted that the resence of PVD may indicate an increased robability of atherosclerosis in the coronary or cerebrovascular arteries. Patients with history of CVD and PVD, or atients with signs of vascular dementia should be referred for assessment of craniocervical and aortic atherosclerosis. We elected to discriminate atients based on the total atherosclerotic score and reoerative risk factors alone, without exlicitly considering oerative interventions. Further studies should validate rosectively whether the strategies of erioerative management reduce NP dysfunction and stroke in the higher risk atients. The total atherosclerotic score, a measure of illness, with a higher score reresenting a greater burden of comorbidity, redicted ostoerative NP dysfunction and stroke after CPB. Routine screening for ostoerative NP dysfunction should be erformed with romt worku of incident cases by screening test, such as MMT or HDS, to detect underlying cognitive dysfunction and focal deficit. Given the multifactorial nature of NP dysfunction and stroke after CPB, high-risk atients may be best managed by modifications in surgical techniques and by roviding stable erioerative hemodynamics. Perioerative assessment using cerebral MRI, cervical MRA, and eiaortic ultrasonograhy could reduce the incidence of ostoerative NP dysfunction and stroke in elderly atients undergoing CABG. The authors thank Dr Yasuhiko Matsukado for his critical comments and suggestions, and Dr Jon Moon for his editorial assistance. This work was suorted in art by the Okinaka Memorial Institute for Medical Research of Jaan. References 1. Tuman KJ, McCarthy RJ, Najafi H, Ivankovich AD. Differential effects of advanced age on neurologic and cardiac risks of coronary artery oerations. J Thorac Cardiovasc Surg 1992; 104:1510 7. 2. Goto T, Yoshitake A, Baba T, Shibata Y, Sakata R, Uozumi H. Cerebral ischemic disorders and cerebral oxygen balance during cardioulmonary byass surgery: reoerative evaluation using magnetic resonance imaging and angiograhy. Anesth Analg 1997;84:5 11. 3. Maeda H, Matsumoto M, Handa N, et al. Reactivity of cerebral blood flow to carbon dioxide in various tyes of ischemic cerebrovascular disease: evaluation by the transcranial Doler method. Stroke 1993;24:670 5. 4. Thiel A, Zickmann B, Stertmann WA, Wyderka T, Hemel-

840 GOTO ET AL Ann Thorac Surg ATHEROSCLEROSIS AND NEUROLOGIC RISK 2000;69:834 40 mann G. Cerebrovascular carbon dioxide reactivity in carotid artery disease. Relation to intraoerative cerebral monitoring results in 100 carotid endarterectomies. Anesthesiology 1995;82:655 61. 5. Wareing TH, Davila-Roman VG, Barzilai B, Murhy SF, Kouchoukos NT. Management of the severely atherosclerotic ascending aorta during cardiac oerations. J Thorac Cardiovasc Surg 1992;103:453 62. 6. Davila-Roman VG, Barzilai B, Wareing TH, Murhy SF, Schechtman KB, Kouchoukos NT. Atherosclerosis of the ascending aorta. Prevalence and role as an indeendent redictor of cerebrovascular events in cardiac atients. Stroke 1994;25:2010 6. 7. Hammon JW Jr, Stum DA, Kon ND, et al. Risk factors and solutions for the develoment of neurobehavioral changes after coronary artery byass grafting. Ann Thorac Surg 1997;63: 1613 8. 8. Shaw PJ, Bates D, Cartlidge NEF, et al. Neurologic and neurosychological morbidity following major surgery: comarison of coronary artery byass and eriheral vascular surgery. Stroke 1987;18:700 7. 9. Murkin JM, Martzke JS, Buchan AM, Bentley C, Wong CJ. A randomized study of the influence of erfusion technique and H management strategy in 316 atients undergoing coronary artery byass surgery. II. Neurologic and cognitive outcomes. J Thorac Cardiovasc Surg 1995;110:349 62. 10. Murkin JK, Newman SP, Stum DA, Blumenthal JA. Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery. Ann Thorac Surg 1995;59:1289 95. 11. Yoshitake T, Kiyohara Y, Kato I, et al. Incidence and risk factors of vascular dementia and Alzheimer s disease in a defined elderly Jaanese oulation: the Hisayama study. Neurology 1995;45:1161 8. 12. D Agostino RS, Svensson LG, Newmann DJ, Balkhy HH, Williamson WA, Shahian DM. Screening carotid ultrasonograhy and risk factors for stroke in coronary artery surgery atients. Ann Thorac Surg 1996;62:1714 23. 13. Johnsson P, Algotsson L, Ryding E, Stahl E, Messeter K. Cardioulmonary erfusion and cerebral blood flow in bilateral carotid artery disease. Ann Thorac Surg 1991;51: 579 84. 14. Pugsley W, Klinger L, Paschalis C, Treasure T, Harrison M, Newman S. The imact of microemboli during cardioulmonary byass on neurosychological functioning. Stroke 1994;25:1393 9. 15. Clark RE, Brillman J, Davis DA, Lovell MR, Price TRP, Magovern GJ. Microemboli during coronary artery byass grafting: genesis and effect on outcome. J Thorac Cardiovasc Surg 1995;109:249 58. 16. Moody DM, Bell MA, Challa VR, Johnston WE, Prough S. Brain microemboli during cardiac surgery or aortograhy. Ann Neurol 1990;28:477 86. 17. Barbut D, Hinton RB, Szatrowski TP, et al. Cerebral emboli detected during byass surgery are associated with clam removal. Stroke 1994;25:2398 402. 18. O Leary DH, Anderson KM, Wolf PA, Evans JC, Poehlman HW. Cholesterol and carotid atherosclerosis in older ersons: the Framingham study. Ann Eidemiol 1992;2: 147 53. 19. Croughwell N, Lyth M, Quill TJ, et al. Diabetic atients have abnormal cerebral autoregulation during cardioulmonary byass. Circulation 1990;82(Sul IV):407 12. 20. Criqui MH, Langer RD, Froeke A, et al. Mortality over a eriod of 10 years in atients with eriheral arterial disease. N Engl J Med 1992;326:381 6.