Click here to read this article as a CME activity
Ann Thorac Surg 2008;85:1571-1578. doi:10.1016/j.athoracsur.2008.01.090
© 2008 The Society of Thoracic Surgeons
Original Articles: Adult Cardiac
Absence of Cognitive Decline One Year After Coronary Bypass Surgery: Comparison to Nonsurgical and Healthy Controls
Jerry J. Sweet, PhDa,
Eileen Finnin, BS,
Penny L. Wolfe, PhDb,
Jennifer L. Beaumont, MSa,
Elizabeth Hahn, MAa,
Jesse Marymont, MDa,
Timothy Sanborn, MDa,
Todd K. Rosengart, MDa,*
a Evanston Northwestern Healthcare, Evanston, Illinois; Northwestern University Feinberg School of Medicine, Chicago, Illinois
b National Rehabilitation Hospital, Washington, DC
Accepted for publication January 28, 2008.
* Address correspondence to Dr Rosengart, Stony Brook University Medical Center, Stony Brook, NY 11794 (Email: todd.rosengart{at}stonybrook.edu).
| Cardiothoracic anesthesiology:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
|
 |
Abstract
|
|---|
Background: Cognitive decline after open-heart surgery has been the subject of a number of conflicting reports in recent years. Determination of possible cognitive impairment due to surgery or use of cardiopulmonary bypass is complicated by numerous factors, including use of appropriate comparison groups and consideration of practice effects in cognitive testing.
Methods: Neuropsychological data were gathered from 46 healthy controls, 42 cardiac patients referred for percutaneous coronary intervention (PCI), and 43 cardiac patients referred for coronary artery bypass grafting (CABG). Fourteen cognitive function tests were utilized at baseline and at three time points after surgery (3 weeks, 4 months, 1 year). Measures showing acceptable test-retest reliability based on intraclass correlations were compared using regression-based reliable change indices.
Results: No clear pattern of group differences or change at follow-up emerged. A greater percentage of CABG patients than controls worsened in seven tests (three at 1 year), but a greater percentage of PCI patients than controls also worsened in seven tests (three at 1 year). Generalized estimating equations showed only two tests (Wechsler Adult Intelligence Scale, Third Edition, Digit Symbol, and Hopkins Verbal Learning Test, Revised, Total Recall) to be significantly different between groups from baseline to 1 year. Interestingly, compared with healthy controls, more PCI patients than CABG patients worsened in the former of those two tests, whereas more PCI and CABG patients improved on the latter.
Conclusions: Using healthy controls and a relevant nonsurgical comparison group to contend with important methodological considerations, current CABG procedure does not appear to create cognitive decline.
Numerous studies have been published concerning the likelihood and extent of cognitive decline after coronary artery bypass graft (CABG) surgery. Results of these studies have changed across the decades in which they were carried out, as a consequence of (1) improvements in CABG procedure, (2) use of the procedure in individuals at greater risk for cognitive dysfunction, such as increasingly more elderly patients, (3) changes in selection of specific cognitive tests, (4) inclusion of better methods to evaluate change in scores from repeated testings, and (5) changes in use and selection of appropriate comparison groups. To be certain, the weight of evidence across decades of clinical research has suggested that, at least using past practices, CABG could be demonstrated to produce some measurable cognitive dysfunction. Methodologic differences, variability in surgical procedure (for example, rewarming rate, blood pressure during cardiopulmonary bypass, attention to aortic atherosclerosis), and even choice of patients (for example, older, preexistent cerebrovascular disease) across the literature have, however, created a literature that is difficult to interpret [1].
Refinements in study design and testing methodologies have improved the reliability of reported results in the recent literature. Unfortunately, even recent studies have continued the controversy regarding negative effects of CABG on cognition. Unsubstantiated conclusions continue to be drawn, for example, based on the use of inadequate neuroprotective techniques during surgery, insufficient measurements and statistical treatment of neurocognitive function after CABG, and a lack of appropriate "normal" and "atherosclerotic" age-matched comparison groups [2–7]. Such methodology leads to a skewed interpretation of "statistical" outcomes as being clinically significant, and has the effect of continuing the impression that CABG is detrimental to cognition, when the data more likely suggest that any such changes are mild, generally reversible, and most relevant to a high-risk population with preoperative risk factors, such as the presence of preexistent cerebrovascular disease.
In two prior studies, we addressed the implications of baseline cognitive function and initial effects of CABG and percutaneous coronary intervention (PCI) on cognitive function [6, 7]. The present study was undertaken to determine if cognitive dysfunction was present in these groups after 1 year. Specifically, we sought to control for effects of repeated administration of cognitive measures in healthy controls and provide a statistical determination of change that controlled for normal variation in the measurement process, while controlling for the effects of cardiac disease across time using patients not undergoing open-heart surgery as a control [8, 9].
|
Patients and Methods
|
|---|
Enrollment
Institutional Review Board approval was obtained in February 2002, and fully informed written consent was obtained from patients selected for participation in this study. Prospective participants in the two cardiac groups (CABG and PCI) were screened by project personnel subsequent to physician referral based on standard clinical indications for either CABG or PCI. Exclusion criteria for cardiac patients primarily reflected excessive risk of neurologic events, or potential inability to successfully complete valid neurocognitive assessments before PCI or isolated, primary CABG procedure [6, 7]. Exclusion criteria for cardiac patients were as follows: a history of stroke or symptomatic carotid artery disease, dementia, substance abuse, renal dysfunction (blood urea nitrogen greater than 50 mg/dL, creatinine greater than 2.5 mg/dL), or hepatic dysfunction (serum glutamic-oxaloacetic transaminase [SGOT]/aspartate aminotransferase [AST] or serum glutamic-pyruvic transaminase [SGPT]/alanine aminotransferase [ALT] more than three times upper limits of normal), language or physical deficiency not allowing test completion, participation in another clinical trial of an investigational device or drug, or other factors suggesting the potential inability to successfully complete neurocognitive assessments in patients undergoing PCI or isolated, primary CABG procedure [6, 7]. Patients with renal or hepatic dysfunction were excluded in an attempt to exclude individuals who might be suffering from a preexisting independent dementing disorder, as were patients whose mini-mental status evaluation (MMSE) scores were below 24.
A cohort of ("healthy") control subjects without specific evidence by history of cardiac, cerebrovascular or neurologic disease and similar to the cardiac patients with regard to age and education was recruited from the community. The presence of potential risk factors for cardiac or neurologic disease, such as diabetes mellitus or hypertension, were not criteria for exclusion from this control group.
Demographics of the controls and cardiac patients are presented in Table 1
and are further described in our previous reports [6, 7]. Education was not significantly different across groups (p = 0.572), with the median completed educational level in all groups being that of a college graduate (range, CABG: eighth grade graduate to advanced degree; control and PCI: high school graduate to advanced degree). A greater number of men than women were present in both cardiac groups compared with the control group (p = 0.010), reflective of the expected distribution of sex in these populations. Race was also significantly different across groups (p = 0.023), with greater proportion of non-Hispanic whites in the cardiac patient groups.
Neurocognitive Testing Battery
All consenting CABG, PCI, and control patients meeting inclusion criteria underwent neurocognitive testing at baseline and 3 weeks, 4 months, and 12 months after surgery. Each study subject was tested on all occasions by the same experienced psychometrician in a comfortable, well-lit environment generally free from visual and auditory distractions over approximately 1 hour.
The test battery included only examinations with age-specific normative standards that would remain valid on retesting, in the following domains: attention: Digit Span from the Wechsler Adult Intelligence Scale, Third Edition (WAIS-III); fine motor dexterity: Grooved Pegboard (dominant and nondominant hands); processing speed: Digit Symbol (WAIS-III), Trail-Making A, Stroop Color-Word Test (word and color pages); language: Controlled Oral Word Association (COWA) and Visual Naming of the Multilingual Aphasia Examination (MAE); executive function: Stroop Color-Word Test (color-word page), Trail-Making B; verbal learning and memory: Hopkins Verbal Learning Test, Revised (HVLT-R; alternate form at second testing).
Operative Technique
Coronary artery bypass graft surgery was performed utilizing standard cardiopulmonary bypass technique with moderate hypothermia (32°C.). Fentanyl and inhalation agents were administered to allow "early" extubation (mean, 4 hours postoperative). Arterial cannulation was performed utilizing a 22F right-angle or soft-flow cannula in an area seen to be devoid of plaque as determined by the epiaortic analysis. Mean arterial perfusion pressure on bypass was maintained at a pressure greater than 60 mm Hg, and hematocrit was maintained at a level greater than 18%.
Statistical Analysis Methods
Evaluation of change was based on methods for multiple regression–based change scores [9]. Multiple linear regression was applied to control group data, including age, sex, education level, baseline score, square and cube of baseline score, and length of time between tests as predictors of follow-up scores. Only predictors that were significant at the 0.10 level were retained in the models, with the exception of the baseline score, which was retained regardless of significance. Intercept and regression coefficients from these models were used to estimate the predicted follow-up scores for all patients. Standardized z-scores were then calculated as the difference between observed retest score and predicted retest score divided by the residual standard deviation from the multiple regression model. Patients were categorized as worsened, unchanged, or improved based on standardized z-scores ([observed minus predicted] divided by residual standard deviation) of ± 1.645. Generalized estimating equation methodology [8] was used to analyze repeated measures. Generalized estimating equation models allow the use of all available data but require missing data to be missing completely at random. Dependent variables in these models were categorical change status from baseline to follow-up on neurocognitive measures. Independent variables were group (control, CABG, PCI), time, and the interaction of group and time. The primary study goal of determining whether or not the pattern of change over time differed between groups was evaluated by the group-by-time interaction term.
|
Results
|
|---|
Forty control subjects (87% of those enrolled in this group), 37 PCI patients (88%), and 29 CABG patients (67%) participated in all four study assessments (Fig 1). Most incomplete assessments were due to scheduling difficulties, lack of interest, or failure to commit to further participation. No one in the control group and only 1 PCI patient withdrew for reasons known to be related to ill health (gastric cancer). In the CABG group, however, 5 patients withdrew for reasons related to ill health or death (n = 2 died, n = 3 too ill; Fig 1). There were, however, no interval neurologic events or interventions or operations in any patient in the study. Baseline characteristics of those who completed all four assessments did not differ from those who did not, with the exception of the Trail Making Test Parts A and B (p = 0.048 and p < 0.001, respectively).
View larger version (16K):
[in this window]
[in a new window]
|
Fig 1. Study participation and drop-out. Numbers in boxes indicate patients who participated at that time point; study dropouts are indicated between the boxes with reason for dropout, where available. Patients with "missed visits" did not drop out of the study but returned to complete later assessments. These missed visits were generally due to scheduling difficulties. (CABG = coronary artery bypass graft surgery; PCI = percutaneous coronary intervention.)
|
|
Test-retest reliability in the control group was fair (> 0.60) for most measures (Table 2); however, the Recognition Discrimination Index exhibited extremely poor reliability (intraclass correlation coefficient: 0.134 to 0.319). On this basis, subsequent analyses did not include this variable.
View this table:
[in this window]
[in a new window]
|
Table 2 Intraclass Correlation Coefficients for Test-Retest Reliability in the Control Group Compared With Baseline
|
|
Group outcomes measured by multiple regression-based reliable change methodology were categorized, as compared with baseline scores, as "worsened," "unchanged," or "improved" using cutoff values for the z-scores of ± 1.645 (Table 3). These methods were designed so that approximately 5% of normal, stable controls would be classified as worsened and 5% as improved.
No clear pattern of change in cognitive function or differences between groups was demonstrated utilizing the above criteria. There were no differences between groups on Grooved Pegboard Dominant Hand, WAIS-III Digit Span, Stroop Color-Word, and MAE Visual Naming. A greater percentage of CABG patients than controls worsened in seven tests (three at 1 year): Grooved Pegboard Nondominant Hand (4 months), Trail Making Test Part A (1 year), Trail Making Test Part B (all times), Stroop Word (1 year), Stroop Color (3 weeks), MAE COWA (3 weeks and 4 months), and HVLT-R Delayed Recall (3 weeks). However, a greater percentage of PCI patients than controls also worsened in seven tests (three at 1 year): Grooved Pegboard Nondominant Hand (4 months), Trail Making Test Part B (3 weeks and 1 year), WAIS-III Digit Symbol basic score (1 year), Stroop Word (1 year), Stroop Color (3 weeks), MAE COWA (4 months), and HVLT-R Delayed Recall (3 weeks). Interestingly, a substantial number of CABG and PCI patients also demonstrated improvement in various scores at each time point.
Generalized estimating equations were used for the longitudinal analysis of the change categories above. The primary question of whether the time trend differs between groups was evaluated by the group-by-time interaction term (Table 4). Utilizing this statistical tool, the only measures with a significant (p < 0.05) interaction were the WAIS III Digit Symbol basic score and the HVLT-R Total Recall: at 1 year, more PCI patients than CABG patients worsened in Digit Symbol, whereas more of both the PCI and CABG groups than healthy controls improved on HVLT-R Total Recall (Figs 2, 3).
View larger version (13K):
[in this window]
[in a new window]
|
Fig 2. Wechsler Adult Intelligence Scale, Third Edition, Digit Symbol basic score change from baseline to 1 year. (CABG = coronary artery bypass graft surgery; PCI = percutaneous coronary intervention; black area = improved; open area = unchanged; gray area = worsened.)
|
|
View larger version (12K):
[in this window]
[in a new window]
|
Fig 3. Hopkins Verbal Learning Test, Revised, Total Recall change from baseline to 1 year. (CABG = coronary artery bypass graft surgery; PCI = percutaneous coronary intervention; black area = improved; open area = unchanged; gray area = worsened.)
|
|
|
Comment
|
|---|
Improvements in CABG surgical technique that have been introduced over the past decade may have resulted in better preservation of cognitive function after open heart surgery compared with earlier analyses [10–12]. Further, numerous studies have demonstrated that the accurate assessment of cognitive function after open heart surgery is critically dependent upon careful selection of comparison groups, use of memory tests that have alternate forms, and use of statistical methods that control for effects of repeated cognitive testing [4–9, 13]. To account for these factors, the present study has endeavored to examine "neurocognitive best practice" CABG procedures utilizing consensus-recommended cognitive testing [6, 7, 13]. In this context, no clear pattern of group differences or change in cognitive function was found in the present study comparing CABG and PCI groups to healthy control groups at 1 year of follow-up.
More specifically, longitudinal models in the present study showed only the WAIS-III Digit Symbol (processing speed) and HVLT-R Total Recall (verbal learning and memory) to be significantly different across groups from baseline to 1 year. However, a greater proportion of PCI patients, rather than CABG patients, were actually "worse" on Digit Symbol, with the proportion of worsened CABG patients comparable with that of healthy controls. Furthermore, the proportions of PCI and CABG groups "worse" on HVLT-R Total Recall at 1 year were comparable and the proportions of both groups that were "improved" were substantially higher than controls. Thus, by the statistical analyses employed in the current study, no evidence of cognitive impairment was uniquely associated with CABG.
These findings are consistent with several recent studies. For example, Rankin and coworkers [14] found no difference between on-pump and off-pump CABG and no change in either group from presurgical to postsurgical level of impairment. Selnes and colleagues [15] similarly found no differences between CABG patients and nonsurgical CAD controls.
This study's limitations include its observational design, relatively small sample size, and loss to follow-up. While there are unavoidable limitations inherent in observational study designs, our use of a healthy control group and a relevant comparison group with cardiac disease provides some advantage over other observational studies that do not provide such control data. With regard to the sample size, the observed differences in the current study were small enough in magnitude that, even with sufficient power to find them statistically significant, they are not clinically meaningful. Finally, the assumption that missing data are completely random for the longitudinal GEE models may have been violated by the loss to follow-up, particularly in the CABG group, although the lack of clear difference between study completers and noncompleters is reassuring.
On the basis of the current data, it seems reasonable to conclude that CABG procedures per se undertaken using appropriate neurocognitive "protection" techniques may not cause decline in cognitive function, at least within the limitations imposed by the currently recommended testing battery and in the relatively low risk patients encompassed by the present study. Some presurgical factors may, however, place subgroups of CABG candidates at increased risk for postsurgical cognitive impairment [6, 14–20]. Such factors as age, education/occupation, hypertension, and low ejection fraction have been found to predict presurgical cognitive impairment, as has the presence of cerebrovascular disease that is often associated with coronary atherosclerosis. In this regard, we and others have previously demonstrated cognitive dysfunction in patients scheduled for CABG or PCI at baseline, which may then predispose them to poor postprocedure cognitive testing [6, 14, 18, 19]. Furthermore, the performance of surgery and utilization of anesthetics themselves can lead to cognitive dysfunction, especially in the elderly [20]. Conversely, we and others have demonstrated that race and sex do not affect neurocognitive testing independent of the above-noted variables [6, 21]. Discrepancies in these demographics in the present study thus should not bear relevance to the reported outcomes.
Given these considerations and prior consensus [13], we recommend that future CABG research addressing the possibility of cognitive dysfunction obtain baseline measures and use healthy controls and a relevant non-CABG comparison group to contend with important statistical and methodological considerations. Importantly, large study samples that allow for investigation of subgroup risk factors may lead to better identification of CABG candidates who could suffer cognitive impairment, despite the improvements in CABG techniques that have served to reduce postsurgical cognitive dysfunction.
|
Acknowledgments
|
|---|
This work was supported by a grant from the American Heart Association. We wish to thank Milicia Vukovic, Anne Galioto, Jenna Duffecy, and Lina Nayak for their assistance on this project.
|
References
|
|---|
- Selnes O, McKhann G. Neurocognitve complications after coronary artery bypass surgery Ann Neurol 2005;57:615-621.[Medline]
- Dijk D. Cognitive and cardiac outcomes 5 years after off-pump vs on-pump coronary artery bypass graft surgery JAMA 2007;297:701-708.[Abstract/Free Full Text]
- Zimpfer D, Czerny M, Vogt F, et al. Neurocognitive deficit following coronary artery bypass grafting: a prospective study of surgical patients and nonsurgical controls Ann Thorac Surg 2004;78:513-518.[Abstract/Free Full Text]
- Vingerhoets G, Van Nooten G, Jannes C. Neuropsychological impairment in candidates for cardiac surgery J Int Neuropsychol Soc 1997;3:480-484.[Medline]
- Keith JR, Puente AE, Malcolmson KL, Tartt S, Coleman AE, Marks Jr HF. Assessing postoperative cognitive change after cardiopulmonary bypass surgery Neuropsychology 2002;16:411-421.[Medline]
- Rosengart T, Sweet JJ, Finnin E, et al. Neurocognitive functioning in patients undergoing coronary artery bypass graft or percutaneous coronary intervention: evidence of impairment prior to intervention compared with normal controls Ann Thorac Surg 2005;80:1327-1335.[Abstract/Free Full Text]
- Rosengart T, Sweet JJ, Finnin E, et al. Stable cognition after coronary bypass: comparisons to percutaneous intervention and normal controls Ann Thorac Surg 2006;82:597-607.[Abstract/Free Full Text]
- Liang K, Zeger S. Longitudinal data analysis using generalized linear models Biometrika 1986;73:13-22.[Abstract/Free Full Text]
- Temkin N, Heaton R, Grant I, Dikmen S. Detecting significant change in neuropsychological test performance: a comparison of four models J Int Neuropsych Soc 1999;5:357-369.[Medline]
- 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.[Abstract/Free Full Text]
- Roach GW, Kanchuger M, Mangano CM, et al. Adverse cerebral outcomes after coronary bypass surgery N Engl J Med 1996;335:1857-1863.[Abstract/Free Full Text]
- Moody DM, Brown WR, Challa VR, Stump DA, Reboussin DM, Legault C. Brain microemboli associated with cardiopulmonary bypass: a histologic and magnetic resonance imaging study Ann Thorac Surg 1995;59:1304-1307.[Abstract/Free Full Text]
- Murkin J, Newman S, Stump D, Blumenthal J. Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery Ann Thorac Surg 1995;59:1289-1295.[Free Full Text]
- Rankin K, Kochamba G, Boone K, Petitti D, Buckwalter JG. Presurgical cognitive deficits in patients receiving coronary artery bypass graft surgery J Int Neuropsych Soc 2003;9:913-924.[Medline]
- Selnes OA, Grega MA, Borowicz Jr LM, Royall RM, McKhann GM, Baumgartner WA. 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.[Abstract/Free Full Text]
- Ernest C, Elliott P, Murphy B, et al. Predictors of cognitive function in candidates for coronary artery bypass graft surgery J Int Neuropsych Soc 2007;13:257-266.[Medline]
- Ho PM, Arciniegas DB, Grigsby J, et al. Predictors of cognitive decline following coronary artery bypass graft surgery Ann Thorac Surg 2004;77:597-603.[Abstract/Free Full Text]
- Baird DL, Murkin JM, Lee DL. Neurologic findings in coronary artery bypass patients: pre-operative or pre-existing? J Cardiothorac Vasc Anesth 1997;1196:694-698.
- Millar K, Asbury AJ, Murray GD. Pre-existing cognitive impairment as a factor influencing outcome after cardiac surgery Br J Anesthes 2001;86:63-67.
- Treasure T, Smith PL, Newman S, et al. Impairment of cerebral function following cardiac and other major surgery Eur J Cardiothorac Surg 1989;3:216-221.[Abstract]
- Hogue CW, Lillie R, Hershey T, et al. Gender influence on cognitive function after cardiac operation Ann Thorac Surg 2003;76:1119-1125.[Abstract/Free Full Text]
Related Article
-
Invited Commentary
- Paul Kurlansky
Ann. Thorac. Surg. 2008 85: 1578.
[Extract]
[Full Text]
[PDF]
This article has been cited by other articles:
|
|
|
|
 
P. Kurlansky
Invited Commentary
Ann. Thorac. Surg.,
May 1, 2008;
85(5):
1578 - 1578.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
