Cognition 6 years after surgical or medical therapy for coronary artery disease.код для вставкиСкачать
Cognition 6 Years after Surgical or Medical Therapy for Coronary Artery Disease Ola A. Selnes, PhD,1 Maura A. Grega, MSN,2 Maryanne M. Bailey, BA,3 Luu D. Pham, MS,4 Scott L. Zeger, PhD,4 William A. Baumgartner, MD,2 and Guy M. McKhann, MD1,3,5 Objective: The choice of coronary artery bypass grafting (CABG) as an intervention for coronary artery disease has been clouded by concerns about postoperative cognitive decline. Long-term cognitive decline after CABG has been reported, but without appropriate control subjects, it is not known whether this decline is specific to CABG or related to other factors such as cerebrovascular disease. Methods: This prospective, observational study of patients with diagnosed coronary artery disease included 152 CABG and 92 nonsurgical cardiac comparison patients from one institution. The main outcome measure was within-patient change in cognitive performance for eight cognitive domains from baseline to 12- and 72-month follow-up. Results: Mild late cognitive decline was observed for both study groups, but despite greater than 80% power to detect a 0.2 standard deviation difference, there were no statistically significant differences between the surgical and nonsurgical patients in the degree of change from 12 to 72 months for any cognitive domain. There was also no difference between groups in the degree of change from baseline to 72 months or in the number of patients with a Mini-Mental State Examination score in the clinically impaired range at 72 months. Interpretation: Late cognitive decline does occur in patients who have undergone CABG surgery, but the degree of this decline does not differ from that observed in patients of similar age with coronary artery disease who have not undergone CABG. Therefore, late cognitive decline after CABG is not specific to the use of cardiopulmonary bypass. Ann Neurol 2008;63:581–590 The choice among the types of intervention for coronary artery disease is complex. Recent reports suggest increased thrombotic complications with drug-eluting stents. On the other hand, neurological complications and cognitive decline have been associated with coronary artery bypass grafting (CABG) surgery. Adverse cerebral outcomes after CABG with the use of cardiopulmonary bypass (CPB) can range from clinically apparent events such as stroke to more subtle neurological injury such as cognitive decline.1,2 A review of contemporary studies concludes that 30 to 65% of patients have evidence of cognitive decline 1 month after surgery.3 The cause of this early postoperative cognitive decline has not yet been clearly established, but it has been commonly assumed that factors associated with the use of CPB, including hemodynamic fluctuations, microemboli, and systemic inflammatory response, may be responsible. Although the early cognitive decline does appear to improve over the first few months after surgery,4 longer term prospective follow-up studies have described subsequent late cognitive decline. In one study of cognitive outcomes after CABG, it was reported that 42% of the patients had worse cognitive performance at 5 years than at baseline.5 The occurrence of late cognitive decline 5 or more years after surgery has been reported by other investigators,6 although some studies have not found evidence of delayed decline several years after CABG.7 If the late cognitive decline is related to the use of CPB, one would predict that patients who have offpump surgery should not experience this late decline. The only randomized comparison of long-term outcomes after on- and off-pump CABG surgery did not find a difference in the incidence of late cognitive decline.8 This leaves open the possibility that the late cognitive decline might be caused by nonspecific effects of major surgery, anesthesia, progression of underlying cerebrovascular disease, or normal aging. Because most long-term studies have not included From the 1Departments of Neurology and Surgery; 2Division of Cardiac Surgery, Johns Hopkins University School of Medicine; 3 Zanvyl Krieger Mind/Brain Institute; 4Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health; and 5Department of Neuroscience, Johns Hopkins University School of Medicine, Baltimore, MD. This article includes supplementary materials available via the Internet at http://www.interscience.wiley.com/jpages/0364-5134/suppmat Received Nov 9, 2007, and in revised form Dec 18. Accepted for publication Feb 4, 2008. Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.21382 Address correspondence to Dr Selnes, Department of Neurology, Reed Hall East-2, 1620 McElderry Street, Baltimore, MD 212051910. E-mail: firstname.lastname@example.org © 2008 American Neurological Association Published by Wiley-Liss, Inc., through Wiley Subscription Services 581 patients of comparable age and health status, it is not known whether late cognitive decline may occur even in patients who do not undergo CABG surgery. In this study, we sought to determine whether patients undergoing CABG surgery have a greater degree of cognitive decline over 72 months than do patients with similar demographic and medical characteristics without surgical intervention. Subjects and Methods In this nonrandomized study, eligible patients completed a medical history review and a battery of standardized neuropsychological tests on the following longitudinal schedule: baseline, 3, 12, 36, and 72 months. The protocol was approved by the institutional review board, and subjects gave written informed consent. Description of Study Groups CORONARY ARTERY BYPASS GRAFTING GROUP. Patients scheduled to undergo elective or urgent CABG surgery with CPB at our institution between September 1997 and March 1999 were approached by the study coordinators for enrollment into the study. Eligibility criteria included being a native English speaker, not mechanically ventilated, able to sit upright, willing to return for follow-up, and able to provide informed consent. These procedures have been previously described.4,9 NONSURGICAL CARDIAC COMPARISON GROUP. Patients under medical management for coronary artery disease (diagnosed by cardiac catheterization) were recruited from cardiologists at our institution from September 1997 through December 2001. Eligibility criteria were the same as described earlier, except that patients who had previous cardiac surgery were excluded. Some of the nonsurgical comparison patients had a percutaneous coronary intervention either before or after entry into the study as part of their medical care. Neuropsychological Tests and Questionnaires The neuropsychological tests were selected to evaluate performance in specific areas of cognition and were structured into eight domains. A description of each test is included in Supplementary Appendix 1. To have a clinically meaningful measure of the overall cognitive status of the study patients, we also administered the Mini-Mental State Examination (MMSE) at all time points.10 The Center for Epidemiological Studies Depression scale (CES-D)11 to measure depression and the Functional Status Questionnaire (FSQ)12 to determine physical functioning were also administered at all study times. Statistical Analysis We used linear mixed models to estimate the difference between the CABG and comparison group in the mean withinsubject change in cognitive test scores from baseline (R: A Language and Environment for Statistical Computing, Vienna, Austria; http://www.R-project.org). This method allows each person to serve as his or her own control so that 582 Annals of Neurology Vol 63 No 5 May 2008 any differences between the two groups in baseline or other nonchanging characteristics do not bias the treatment comparison. We assumed that the average response for each group had its own mean change from baseline to 3 months, reflecting both learning and treatment effects, and then had an approximately linear change from 3 to 72 months. The average within-subject changes from baseline for each group, as estimated by this model, are shown in Figures 1 and 2. Because the performance on neuropsychological tests can be influenced by education, age, and sex, we controlled for these variables in a preliminary regression analysis. For education and age, we used regression splines with two degrees of freedom to allow for possible nonlinear relations. The linear mixed models account for missing data for which the probability of being missing depends on treatment group, time of follow-up, and other measured variables, but not on the unobserved missing data (missing at random). We did not distinguish deaths from refusals to participate or losses to follow-up. Hence, the inferences here pertain to the original cohort as if all subjects survived the follow-up period. This study was designed to have greater than 80% power to detect a difference in the average within-person change from baseline between the CABG and comparison group of 0.2 population standard deviation. Results Subject Characteristics We recruited 152 candidates for CABG and 92 comparison patients with coronary artery disease diagnosed by cardiac catheterization. Table 1 summarizes baseline characteristics for the surgical and nonsurgical patients. Risk factors for cerebrovascular disease were common in both study groups. The mean interval between baseline testing and the last follow-up was slightly longer for the CABG group (6.9 years) than for the nonsurgical patients (6.3 years) ( p ⫽ 0.006). A greater proportion of the CABG patients (32%) than nonsurgical comparison patients (20%) had surgery with general anesthesia during the follow-up interval ( p value is not significant). Additional interval medical history is described in Table 2. Completeness of Follow-up Twenty-six of the 152 CABG patients (17%) died before the 72-month visit. Contact was made with 119 patients of whom 96 (76%) completed cognitive testing at 72 months, and 23 patients refused. Seven patients were lost to follow-up. Nineteen of the 92 nonsurgical patients (21%) died before the 72-month visit. Contact was made with 69 patients, of whom 61 (84%) completed cognitive testing and 8 refused. No contact was made with four patients, who were therefore lost to follow-up. There were eight patients in the comparison group who during the course of the study follow-up had subsequently undergone CABG surgery (two before the Fig 1. Longitudinal trends by individual cognitive domains: mean Z-scores for the eight cognitive domains and the global cognitive domain showing trends in cognitive performance for both groups from baseline to 3-, 12-, 36-, and 72-month follow-up. The Z-scores are standardized relative to the baseline performance of the nonsurgical cardiac patients, whose mean baseline value is therefore zero. The Z-scores have been adjusted for age, sex, and education. Error bars indicate the limits of the 95% confidence interval for the mean of each measure. Solid line indicates nonsurgical cardiac comparison (NSCC) group; dashed line indicates coronary artery bypass graft (CABG) group. 12-month visit, three before the 36-month visit, and three before the 72-month visit). These patients completed follow-up according to the study protocol and were included in the 72-month data as nonsurgical patients under the intent-to-treat methodology. We did, however, complete our primary statistical analysis outlined later, both with and without these eight patients, and concluded that there were no material differences in the study results. Cognitive Outcomes The mean unadjusted scores for each neuropsychological variable by study group at baseline, 12, and 72 months are presented in Supplementary Appendix 2. Selnes et al: Cognitive Decline with CAD 583 72 months, the only domain where the two groups differed statistically significantly was Motor Speed ( p ⫽ 0.04), with the nonsurgical patients showing greater decline. Fig 2. Changes from baseline in overall cognitive performance by group: mean changes from baseline (with 95% confidence intervals) for the global average of all eight cognitive domains for the coronary artery bypass graft (CABG) group (dashed line) and nonsurgical patients (solid line). Although the CABG group had lower mean scores for several of the neuropsychological tests at baseline, there were no statistically significant differences between the CABG group and nonsurgical patients in these unadjusted raw scores. The trends in the levels of longitudinal test performance by cognitive domain for the two study groups are shown in Figure 1. Both groups showed improvement from baseline to 12 months, and this was followed by a trend of subsequent decline in performance for all cognitive domains. The mean of the eight cognitive domain scores is summarized in the global domain score (see Fig 2). In Figure 2, the baseline for both groups is set at zero, to compare differences between the groups in the degree of change over time. Comparison of Change in Cognitive Performance between Study Groups The question of principal interest to this study was whether the degree of late decline was greater in the CABG group than in the nonsurgical group. Using the linear mixed models described in Subjects and Methods, we compared the within-patient change in test performance from baseline to each follow-up point for each cognitive domain for the two study groups (Table 3). In this model, there were no statistically significant differences in the degree of decline/change between the two study groups from baseline to 72 months (see Fig 2). Comparing the degree of “late” decline from 12 to 584 Annals of Neurology Vol 63 No 5 May 2008 Degree of Cognitive Change from Baseline to 72 Months for Each Group To evaluate the amount of decline in cognitive test performance over time for the two groups, we examined change from baseline to 72 months for each cognitive domain. As seen in Figure 1, both study groups declined in all domains except attention. For the CABG patients, however, the only statistically significant declines were in visuoconstruction ( p ⬍ 0.01) and executive function ( p ⫽ 0.02). For the nonsurgical comparison group, declines were statistically significant in visuoconstruction ( p ⬍ 0.01) and motor speed ( p ⫽ 0.01). The degree of change from baseline to 72 months was relatively small for both groups, ranging from an improvement of 0.1 to a decline of 0.3 standard deviation units. To address the question of the magnitude of the late decline, we compared degree of change in cognitive performance from 12 to 72 months. With the single exception that the nonsurgical group did not have significant decline in the cognitive domain of attention, both groups had statistically significant decline in all other cognitive domains ( p values ranging from ⬍0.01– 0.02). The degree of change from 12 to 72 months was of greater magnitude than that observed from baseline to 72 months, ranging from a decline of 0.1 to a decline of 0.58 standard deviation units. There were no group differences in the mean change in the global score regardless of whether we also controlled for baseline medical variables that were different between the groups. Mini-Mental State Examination Performance We also examined changes in scores on the MiniMental State Examination from baseline to follow-up points for the two groups. Mean MMSE scores at baseline and at 72 months did not differ between the two study groups (see Tables 1 and 2). At baseline, 5% of the CABG and 6% of the nonsurgical patients had MMSE scores below normal (⬍24) (not significant). From baseline to 72 months, the average withinpatient change on the MMSE was a decline of 0.22 for the CABG group and 0.16 point for the nonsurgical patients (not significant). The majority of CABG (57%) and nonsurgical patients (67%) had stable or improved MMSE scores at 72 months. Six of the CABG patients (6.3%) had a clinically meaningful decline (4 or more points)13 on the MMSE at 72 months compared with 7 patients (11.5%) in the comparison group. The number of patients with MMSE scores less Table 1. Baseline Participant Characteristics Characteristics CABG (n ⴝ 152) NSCC (n ⴝ 92) p Mean age ⫾ SD, yr 63.6 ⫾ 9.4 65.9 ⫾ 9.2 0.06 Mean education ⫾ SD, yr 14.0 ⫾ 4.0 14.4 ⫾ 3.4 0.47 Male sex, n (%) 115 (76) 71 (77) 0.46 White, n (%) 137 (90) 86 (93) 0.26 8 (5) 4 (4) 0.50 Carotid bruit, n (%) 26 (17) 6/33 (8) 0.04 Hypertension, n (%) 98 (65) 46 (50) 0.01 Diabetes mellitus, n (%) 45 (30) 21 (23) 0.16 Peripheral vascular disease, n (%) 25 (16) 6 (7) 0.01 Transient ischemic attack, n (%) 8 (5) 2 (2) 0.20 Myocardial infarction, n (%) 71 (47) 45 (49) 0.42 History of atrial fibrillation, n (%) 16 (10) 13 (15) 0.23 Family history of Alzheimer’s disease, n (%) 20 (13) 9 (10) 0.27 Past smoker, n (%) 106 (70) 51 (55) 0.01 Current smoker, n (%) 17 (11) 3 (3) 0.02 2.73 ⫾ 0.5 1.95 ⫾ 0.7 0.0001 30 (20) 49 (53) 0.0001 Mean number of percutaneous coronary interventions ⫾ SD 1.6 ⫾ 0.9 1.4 ⫾ 0.6 0.24 Past chest pain/angina, n (%) 121 (80) 56 (61) 0.001 56.0 ⫾ 11.1 55.8 ⫾ 9.5 0.87 27.6 ⫾ 2.4 27.9 ⫾ 2.0 0.16 Past stroke, n (%) Mean number of diseased coronary vessels ⫾ SD Percutaneous coronary interventions, n (%) Mean cognitive testing time ⫾ SD, min Functional Measures Mean Mini-Mental State Examination score ⫾ SDa Mini-Mental State Examination score ⬍ 24, n (%) 7 (5) 6 (6) 0.55 Mean CES-D score ⫾ SD 13.2 ⫾ 9.6 8.8 ⫾ 7.6 0.0001 CES-D score ⬎ 15, n (%) 50 (33) 14 (15) 0.002 30.6 ⫾ 5.1 33.0 ⫾ 3.9 0.0001 Mean Functional Status Questionnaire, Physical function score ⫾ SDb Physiological Data Any ApoE-ε4 genotype, n (%) 38 (25) 29 (32) 0.14 Mean systolic blood pressure ⫾ SD, mm Hg 131 ⫾ 22 141 ⫾ 26 0.001 Mean diastolic blood pressure ⫾ SD, mm Hg 73 ⫾ 11 79 ⫾ 13 0.0001 Mean MAP ⫾ SD, mm Hg 92 ⫾ 13 100 ⫾ 16 0.0001 Mean heart rate ⫾ SD, beats/min 66 ⫾ 14 66 ⫾ 11 0.93 Mean pulse pressure ⫾ SD, mm Hgc 58 ⫾ 19 62 ⫾ 21 0.15 Pulse pressure ⬎ 65 mm Hg, n (%) 41 (27) 29 (33) 0.21 Selnes et al: Cognitive Decline with CAD 585 Table 1. (Continued) Operative Data (CABG only) Mean cardiopulmonary bypass time ⫾ SD, min 99 ⫾ 32.3 NA — Mean aortic cross clamp time ⫾ SD, min 68 ⫾ 24.4 NA — Mean number of coronary grafts placed ⫾ SD 3.1 ⫾ 0.9 NA — Left anterior descending graft placed, n (%) 149 (98) NA — Right coronary artery graft placed, n (%) 119 (78) NA — Circumflex graft placed, n (%) 110 (72) NA — 38 (25) NA — Mean general anesthesia time ⫾ SD, min 291 ⫾ 68 NA — Mean lowest esophageal temperature ⫾ SD, °C 29.7 ⫾ 2.3 NA — 69 ⫾ 6.2 NA — Red blood cell transfusion in first 24 hours, n (%) 54 (35) NA — Inotropic support in the ICU, n (%) 80 (53) NA — Mean awake time from general anesthesia ⫾ SD, hr 3.6 ⫾ 2.6 NA — 10.3 ⫾ 10.2 NA — 5.7 ⫾ 2.8 NA — Aortic disease by palpation, n (%) Mean MAP ⫾ SD, mm Hg ICU and Postoperative Data Mean time to extubation ⫾ SD, hr Mean postoperative hospital length of stay ⫾ SD, days Mini-Mental State Examination (30 ⫽ best score), score ⬍ 24 indicative of dementia. Beth Israel Functional Status Questionnaire, Physical functioning score (36 ⫽ best score). c Pulse pressure is calculated as systolic blood pressure minus diastolic blood pressure. CABG ⫽ coronary artery bypass grafting group; NSCC ⫽ nonsurgical cardiac comparison group; SD ⫽ standard deviation; CES-D ⫽ Center for Epidemiological Study-Depression, score ⬎ 15 indicative of depressive symptoms; ApoE ⫽ apolipoprotein E; MAP ⫽ mean arterial pressure; NA ⫽ not available; ICU ⫽ intensive care unit. a b than 24 at 72 months was similar for the two groups (see Table 2). Discussion This observational study compared longitudinal changes in cognition in patients 72 months after coronary artery bypass surgery with those occurring in a group of demographically and medically similar nonsurgical patients with coronary artery disease. Despite evidence of mild cognitive decline in both study groups, we found no statistically significant or clinically meaningful differences in the degree of decline in cognitive test performance over time between the CABG patients and the nonsurgical comparison patients. Comparing the degree of change from 12 to 72 months, which in previous studies has been considered as late decline, there was again no evidence of statistically significant greater late decline in the CABG group than in the nonsurgical patients. Relative to their baseline performance, both groups had modest decline in performance in all cognitive domains when reevaluated at 72 months. The relatively small change from baseline to 72 months may partially reflect improvements from practice effects counteract- 586 Annals of Neurology Vol 63 No 5 May 2008 ing any decline in cognitive test performance that may be expected over a 6-year follow-up period.14 In our CABG population, there are comorbidities that might be associated with worse cognitive performance. These include greater incidence of cardiovascular risk factors, greater severity of coronary artery disease, and more surgery requiring general anesthesia during the follow-up period. Nonetheless, despite these factors, which might have been expected to produce greater cognitive decline in the CABG population, no differences between the two groups were found. On the other hand, the CABG group had higher scores than the nonsurgical patients on a screening measure of depression at baseline. This difference was not statistically significant at 72-month follow-up. There is evidence from several studies, however, that improvement in depression scores does not predict improved performance on cognitive tests.14 –16 Therefore, the lack of greater late decline among the CABG patients cannot be attributed to an improvement in their mood status. Similar to the findings from this study, Newman and colleagues5 report evidence of late cognitive decline in their prospective 5-year study. They hypothesize that the late cognitive decline might be related to Table 2. Characteristics of the Study Population at 72 Months Variables CABG NSCC p Died since enrollment, n (%) 26/152 (17) 19/92 (21) 0.35 Completed cognitive testing, n (%) 96/126 (76) 61/73 (84) 0.14 Refused cognitive testing, n (%) 23/126 (18) 8/73 (11) 0.12 7/126 (6) 4/73 (5) 0.62 Mean number of days since enrollment ⫾ SD 2,515 ⫾ 434 2,313 ⫾ 380 0.006 Mean number of years since enrollment ⫾ SD 6.9 ⫾ 1.2 6.3 ⫾ 1.0 0.006 Tested patient at home, n (%) 47/96 (49) 19/61 (31) 0.02 62 ⫾ 14 61 ⫾ 13 0.71 Testing Status No contact/lost to follow-up, n (%) Mean cognitive testing time ⫾ SD, min n ⴝ 96 n ⴝ 61 70.99 ⫾ 9.6 71.91 ⫾ 9.7 0.56 Actively employed for wages, n (%) 39/94 (41) 15/59 (25) 0.03 Hospitalized since last visit, n (%) 53 (55) 30 (49) 0.28 Surgery with general anesthesia, n (%) 31 (32) 12 (20) 0.06 Angina/chest pain, self-report, n (%) 28 (29) 21 (34) 0.30 18/28 (64) 12/21 (57) 0.41 Seen by doctor since last visit, n (%) 94(98) 61 (100) 0.37 Electrocardiogram done, n (%) 86 (90) 57 (93) 0.30 Stress testing done, n (%) 66 (69) 43 (73) 0.35 4 (4) 1 (2) 0.35 19 (20) 13 (21) 0.48 8 (8) 5 (8) 0.61 0 8 (13) 0.0001 17 (18) 8 (13) 0.29 Cerebrovascular accident, n (%) 4 (4) 1 (2) 0.35 Currently smoking, n (%) 5 (5) 5 (8) 0.33 Cancer/cancer therapy, n (%) 14 (15) 9 (15) 0.57 Vision changes for the worse, n (%) 36 (37) 19 (31) 0.26 New diagnosis of hypertension, n (%) Interim Medical History Mean current age ⫾ SD, yr If yes to chest pain, NTG used, n (%) Myocardial infarction, n (%) Repeat coronary angiogram, n (%) Percutaneous coronary intervention, n (%) New CABG surgery, n (%) Atrial fibrillation, n (%) 13 (13) 7 (12) 0.46 New diagnosis of diabetes mellitus, n (%) 9 (9) 3 (5) 0.24 Thyroid problems, n (%) 9 (9) 8 (14) 0.27 Takes medication as prescribed, n (%) 88 (93) 57 (98) 0.12 Hormone replacement therapy, n (%) 5/24 (21) 3/16 (19) 0.60 69 (72) 42 (69) 0.40 7 (7) 10 (16) 0.06 81 (84) 47 (76) 0.13 3 (3) 4 (7) 0.26 Taking a ␤-blocker, n (%) Taking an antidepressant, n (%) Taking a statin, n (%) Taking cognitive-enhancing drugs, n (%) Selnes et al: Cognitive Decline with CAD 587 Table 2. (Continued) Variables CABG NSCC p 27.4 ⫾ 2.5 27.9 ⫾ 2.6 0.18 Functional Measures Mean Mini-Mental Examination score ⫾ SDa Mini-Mental State Examination score ⬍ 24, n (%) Mean CES-D score ⫾ SD CES-D score ⬎ 15, n (%) b Mean Functional Status Questionnaire, Physical functioning score ⫾ SD 7 (7) 5 (8) 0.53 9.5 ⫾ 9.2 9.0 ⫾ 8.6 0.73 13 (13) 9 (15) 0.50 29.9 ⫾ 6.4 32.1 ⫾ 5.5 0.03 Physiologic Data Mean systolic blood pressure ⫾ SD, mm Hg 132 ⫾ 20 136 ⫾ 20 0.29 Mean diastolic blood pressure ⫾ SD, mm Hg 71 ⫾ 10 73 ⫾ 10 0.21 Mean MAP ⫾ SD, mm Hg 91 ⫾ 11 94 ⫾ 11 0.17 Mean heart rate ⫾ SD, beats/min 65 ⫾ 10 64 ⫾ 10 0.69 Mean pulse pressure ⫾ SD, mm Hg 61 ⫾ 19 63 ⫾ 18 0.66 Pulse pressure ⬎ 65mm Hg, n (%) 42 (45) 22 (39) 0.28 c Mini-Mental State Examination (30 ⫽ best score), score ⬍ 24 indicative of dementia. Beth Israel Functional Status Questionnaire, Physical functioning score (36 ⫽ best score). c Pulse pressure is calculated as systolic blood pressure minus diastolic blood pressure. CABG ⫽ coronary artery bypass grafting group; NSCC ⫽ nonsurgical cardiac comparison group; SD ⫽ standard deviation; NTG ⫽ sublingual nitroglycerin; CES-D ⫽ Center for Epidemiological Study-Depression, score ⬎ 15 indicative of depressive symptoms; MAP ⫽ mean arterial pressure. a b delayed effects of the CABG 5 years earlier, but because their study, as well as other 5-year studies,6 did not include a comparison group, they could not rule out that late cognitive decline might also occur in a population of older patients with cardiovascular disease. The lack of any difference in the long-term cognitive trajectories between the surgery patients and such a comparison group in our study suggests that late cognitive decline 5 or more years after CABG is not specific to the use of CPB. Additional support for this interpretation comes from studies that have compared long-term cognitive outcomes after conventional CABG and off-pump surgery. In a recent study, van Dijk and colleagues8 com- Table 3. Differences in Change from Baseline to 72 Months and from 12 to 72 Months between Coronary Artery Bypass Grafting Group and Nonsurgical Cardiac Comparison Group Cognitive Domains Verbal Memory Visual memory Visuoconstruction Language Motor speed Psychomotor speed Attention Executive function Global Difference in Change from Baseline to 72 Months Difference in Change from 12-72 Months Differencea (95% CI) p Differencea (95% CI) p 0.01 (⫺0.23, 0.26) 0.10 (⫺0.25, 0.44) 0.02 (⫺0.24, 0.28) ⫺0.07 (⫺0.26, 0.13) 0.20 (⫺0.03, 0.43) ⫺0.10 (⫺0.32, 0.13) ⫺0.02 (⫺0.30, 0.26) ⫺0.13 (⫺0.48, 0.22) 0.02 (⫺0.15, 0.18) 0.94 0.58 0.88 0.48 0.10 0.36 0.89 0.47 0.80 ⫺0.15 (⫺0.41, 0.11) 0.05 (⫺0.24, 0.35) ⫺0.08 (⫺0.35, 0.19) ⫺0.06 (⫺0.25, 0.13) 0.21 (0.00, 0.41) ⫺0.09 (⫺0.29, 0.12) 0.04 (⫺0.21, 0.29) ⫺0.07 (⫺0.42, 0.28) ⫺0.03 (⫺0.17, 0.11) 0.25 0.74 0.57 0.55 0.04b 0.37 0.76 0.70 0.67 a Difference numbers that are positive indicate that the coronary artery bypass grafting (CABG) group has declined less than the nonsurgical cardiac comparison group. b Statistically significant values. CI ⫽ confidence interval. 588 Annals of Neurology Vol 63 No 5 May 2008 pared cognitive outcomes 5 years after surgery in patients who had been randomized to undergo either conventional CABG or off-pump surgery. They found that a significant proportion of both study groups had declined at 5 years, but there was no difference in the incidence of decline between the two groups. They concluded, therefore, that avoiding the use of CPB does not materially affect cognitive outcomes 5 years after the surgery. The cause of the late cognitive changes after CABG and in nonsurgical patients with cardiovascular disease remains uncertain. There is considerable evidence in support of the hypothesis that patients with risk factors for vascular disease and markers of atherosclerosis have greater rates of cognitive decline over time than do patients without such risk factors.17–21 Because both the surgical and nonsurgical patients had one or more risk factors for cerebrovascular disease, it is likely, therefore, that the cognitive decline over time is related either to normal aging in the context of cerebrovascular disease or to progression of the underlying vascular disease over the 6-year follow-up. This hypothesis is also supported by data from recent studies suggesting that degree of coronary atherosclerosis is predictive of greater degree of cerebrovascular disease on magnetic resonance imaging.21–23 We did not have MRI studies in our patients, but previous studies have shown that not only are lacunar infarcts common before surgery in candidates for CABG,24 but new silent infarcts occurring between 3 and 12 months after surgery are also frequent.25 If rate of cognitive decline in elderly patients is related to vascular risk factors, one might expect that better control of such risk factors might reduce the degree of late cognitive decline. Mullges and colleagues7 studied 52 patients more than 4 years after CABG and reported that none of the patients in this group had cognitive decline at follow-up. They hypothesized that strict postoperative control of modifiable risk factors, including hypertension, hyperlipidemia, and diabetes, may explain the lack of late cognitive decline in their group of patients. Although these findings are intriguing, prospective randomized trials with larger cohorts would be required to test the hypothesis that better postoperative control of risk factors for cerebrovascular disease may attenuate the degree of late cognitive decline in patients with coronary artery disease. It has also been suggested that there may be an increased risk for Alzheimer’s disease several years after surgery.26 In a recent retrospective cohort study of Veterans Affairs patients with either CABG or percutaneous coronary intervention 5 years earlier, the authors conclude that the risk for development of Alzheimer’s disease was significantly greater among the CABG patients.27 In contrast with these findings, Knopman and colleagues28 did not find any evidence of an association of CABG and dementia in a case–control study from the Mayo Clinic. In our study, we found no differences between CABG and nonsurgical patients in the proportion of patients with Mini-Mental State Examination scores in the clinically impaired range at 72 months. Similarly, there were no differences between the two groups in the number of patients with a meaningful decline (four points or more) from baseline to 72 months. Furthermore, as with previous studies, we did not find a disproportionate decline in memory that is typically associated with Alzheimer’s disease.28,29 The principal limitation of this study is that it was not a randomized comparison of CABG patients and nonsurgical control subjects. Such a study would not have been ethically possible in our environment at the time when the study was initiated. However, we compared the mean change from a person’s own baseline level so that constant (eg, baseline) characteristics are less likely to bias the comparison. As was the case with previous studies reporting late cognitive decline, this was a single-site study of relatively educated patients, and it may not be representative of a more diverse population. Finally, in cohort studies of this duration, there is an inevitable loss to follow-up of study participants. The statistical methods we used adjusted for differential dropout rates that are predictable from observed variables, but not for those that cannot be observed. The loss to follow-up in this study is comparable or better than what has been reported in previous long-term studies of cognition after CABG.5 As an observational study, there were differences in the medical and demographic characteristics of the patients in the two study groups. Because we focused the analysis on change from each person’s own baseline level, we controlled for the influence of nonchanging variables. As in every nonrandomized study, there may remain unmeasured factors that bias the results reported. Strengths of this study are the duration of the follow-up (6 years), and to our knowledge, it is also the first study to compare cognitive outcomes after CABG with those of nonsurgical patients with coronary artery disease. In addition to addressing the impact of aging and cardiovascular disease on cognitive test performance over time, such comparison patients also allow for practice effects and random errors associated with repeated neuropsychological testing to be taken into account. As with previous long-term follow-up studies, we did observe decline in cognitive test performance over the follow-up period. However, in contrast with past work, this study with nonsurgical patients did not find evidence of greater late cognitive decline compared with patients undergoing CABG 6 years earlier. Late cognitive decline after CABG is therefore not specific to the use of CPB. Selnes et al: Cognitive Decline with CAD 589 This study was supported by the NIH (NINCDS, 35610, [G.M.M.]), the Charles A. Dana Foundation (90030633, [G.M.M.]), and the Johns Hopkins Medical Institution (GCRC 00052, [G.M.M.]). We thank Drs P. Talalay, R. Gottesman, and C. Hogue for their help during the preparation of this manuscript. L.M. Borowicz, Jr. helped with the data collection. We also thank the cardiologists, cardiac surgeons, and anesthesiologists at our institution and at Johns Hopkins Bayview Medical Center who helped with this study. Special thanks are extended to our study participants who volunteered their time and energy to make this study possible. References 1. Gottesman RF, Wityk RJ. Brain injury from cardiac bypass procedures. Semin Neurol 2006;26:432– 439. 2. Newman MF, Mathew JP, Grocott HP, et al. Central nervous system injury associated with cardiac surgery. Lancet 2006;368: 694 –703. 3. Hogue CW Jr, Palin CA, Arrowsmith JE. Cardiopulmonary bypass management and neurologic outcomes: an evidence-based appraisal of current practices. Anesth Analg 2006;103:21–37. 4. Selnes OA, Grega MA, Borowicz LM Jr, et al. Cognitive changes with coronary artery disease: a prospective study of coronary artery bypass graft patients and nonsurgical controls. Ann Thorac Surg 2003;75:1377–1384. 5. Newman MF, Kirchner JL, Phillips-Bute B, et al. Longitudinal assessment of neurocognitive function after coronary-artery bypass surgery. N Engl J Med 2001;344:395– 402. 6. Stygall J, Newman SP, Fitzgerald G, et al. Cognitive change 5 years after coronary artery bypass surgery. Health Psychol 2003; 22:579 –586. 7. Mullges W, Babin-Ebell J, Reents W, et al. Cognitive performance after coronary artery bypass grafting: a follow-up study. Neurology 2002;59:741–743. 8. van Dijk D, Spoor M, Hijman R et al. Cognitive and cardiac outcomes 5 years after off-pump vs on-pump coronary artery bypass graft surgery. JAMA 2007;297:701–708. 9. McKhann GM, Goldsborough MA, Borowicz LM Jr, et al. Cognitive outcome after coronary artery bypass: a one-year prospective study. Ann Thorac Surg 1997;63:510 –515. 10. Folstein MF, Folstein SE, McHugh PR. “Mini-Mental State”: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189 –198. 11. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Measurement 1977;1:385– 401. 12. Jette AM, Davies AR, Cleary PD, et al. The Functional Status Questionnaire: reliability and validity when used in primary care. J Gen Intern Med 1986;1:143–149. 13. Tombaugh TN. Test-retest reliable coefficients and 5-year change scores for the MMSE and 3MS. Arch Clin Neuropsychol 2005;20:485–503. 590 Annals of Neurology Vol 63 No 5 May 2008 14. Wesnes K, Pincock C. Practice effects on cognitive tasks: a major problem? Lancet Neurol 2002;1:473. 15. Andrew MJ, Baker RA, Kneebone AC, et al. Mood state as a predictor of neuropsychological deficits following cardiac surgery. J Psychosom Res 2000;48:537–546. 16. Airaksinen E, Wahlin A, Larsson M, et al. Cognitive and social functioning in recovery from depression: results from a population-based three-year follow-up. J Affect Disord 2006; 96:107–110. 17. Xiong GL, Plassman BL, Helms MJ, et al. Vascular risk factors and cognitive decline among elderly male twins. Neurology 2006;67:1586 –1591. 18. Piguet O, Grayson DA, Creasey H, et al. Vascular risk factors, cognition and dementia incidence over 6 years in the Sydney Older Persons Study. Neuroepidemiology 2003;22:165–171. 19. Saxton J, Ratcliff G, Newman A, et al. Cognitive test performance and presence of subclinical cardiovascular disease in the cardiovascular health study. Neuroepidemiology 2000;19: 312–319. 20. Elkins JS, O’Meara ES, Longstreth WT Jr, et al. Stroke risk factors and loss of high cognitive function. Neurology 2004;63: 793–799. 21. Knopman D, Boland LL, Mosley T, et al. Cardiovascular risk factors and cognitive decline in middle-aged adults. Neurology 2001;56:42– 48. 22. Rosano C, Naydeck B, Kuller LH, et al. Coronary artery calcium: associations with brain magnetic resonance imaging abnormalities and cognitive status. J Am Geriatr Soc 2005;53: 609 – 615. 23. de Leeuw FE, de Groot JC, Oudkerk M, et al. Aortic atherosclerosis at middle age predicts cerebral white matter lesions in the elderly. Stroke 2000;31:425– 429. 24. Goto T, Baba T, Honma K, et al. Magnetic resonance imaging findings and postoperative neurologic dysfunction in elderly patients undergoing coronary artery bypass grafting. Ann Thorac Surg 2001;72:137–142. 25. Kohn A. Magnetic resonance imaging registration and quantitation of the brain before and after coronary artery bypass graft surgery. Ann Thorac Surg 2002;73:S363–S365. 26. Brown WR, Moody DM, Tytell M, et al. Microembolic brain injuries from cardiac surgery: are they seeds of future Alzheimer’s disease? Ann NY Acad Sci 1997;826:386 –389. 27. Lee TA, Wolozin B, Weiss KB, et al. Assessment of the emergence of Alzheimer’s disease following coronary artery bypass graft surgery or percutaneous transluminal coronary angioplasty. J Alzheimers Dis 2005;7:319 –324. 28. Knopman DS, Petersen RC, Cha RH, et al. Coronary artery bypass grafting is not a risk factor for dementia or Alzheimer disease. Neurology 2005;65:986 –990. 29. Schmidtke K, Hull M. Neuropsychological differentiation of small vessel disease, Alzheimer’s disease and mixed dementia. J Neurol Sci 2002;203-204:17–22.