| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Thoracic and Cardiovascular Surgery, West German Heart Center, Department of Neurology, Institute of Diagnostic and Interventional Radiology and Neuroradiology, Institute of Medical Informatics, Biometry and Epidemiology, University Clinic of Essen, Essen, Germany
Accepted for publication October 24, 2007.
* Address correspondence to Dr Knipp, Department of Thoracic and Cardiovascular Surgery, West German Heart Center, University Clinic of Essen, Hufelandstrasse 55, Essen, 45122, Germany (Email: stephan.knipp{at}uk-essen.de).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Methods: Thirty-nine patients undergoing on-pump CABG completed preoperative neuropsychologic examination and were followed up prospectively at discharge, 3 months, and 3 years after surgery. Cognitive performance was assessed with a battery of 11 standardized psychometric tests assessing 7 cognitive domains. Cognitive outcome was analyzed by determining (1) mean changes in within-patient scores over time (identifying cognitive functions with decline), and (2) the incidence of cognitive deficit for each individual (identifying patients with decline). Objective evidence of acute cerebral ischemia was obtained by diffusion-weighted magnetic resonance imaging. Prospectively collected data were used to identify predictors of cognitive deficits.
Results: From baseline to discharge, cognitive test scores significantly declined in 7 measures. Most tests improved by 3 months. Between 3 months and 3 years, late decline was observed in 2 measures with persistent deterioration in 1 measure (verbal memory) relative to baseline. Postoperative cognitive deficits (drop of 
1 SD in scores on 3 tests) were observed in 56% of patients at discharge, 23% at 3 months and 31% at 3 years. On postoperative diffusion-weighted magnetic resonance imaging, there were new ischemic cerebral lesions in 51% of patients. The presence of cognitive deficit at discharge was a significant univariate predictor of late cognitive decline (p = 0.025). A relation between the presence of new diffusion-weighted magnetic resonance imaging detected lesions and cognitive decline, however, was not found.
Conclusions: Longitudinal cognitive performance of patients with CABG showed a two-stage course with early improvement followed by later decline. Long-term cognitive deficit was predicted by early cognitive decline, but not by ischemic brain lesions on magnetic resonance imaging.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Until recently, the majority of studies on cognitive changes after coronary bypass surgery have focused on the early postoperative period, and impairment in cognition was generally assumed to be transient and reversible within a few months [4]. In 2001, Newman and colleagues [5] reported on a prospective longitudinal study on patients who had undergone CABG in the late 1990s. The authors found that a surprisingly high rate of 42% of their patients performed below their baseline performance when reevaluated 5 years after surgery. Predictors of long-term cognitive decline included older age, fewer years of education, greater baseline scores, and cognitive decline at the time of discharge. In a similar 5-year study, a greater than expected late decline was found in certain cognitive domains [6]. However, late cognitive decline after CABG may not be inevitable. In a study from Germany on elective CABG patients who were followed for as long as 5 years, none of the patients had clinically significant cognitive decline [7]. Thus, there is substantial evidence that late decline in cognitive functions does occur in patients after CABG, but the degree and duration of these changes are still controversial.
The application of magnetic resonance imaging (MRI) to the study of neurologic and neuropsychologic complications of cardiac surgery is of particular interest as cerebral ischemia is the crucial pathophysiologic factor for postoperative stroke and possibly may be related to persistent cognitive deficits [2]. Some studies have demonstrated that new focal brain lesions, detected on postoperative MRI scans, can develop after CABG [8, 9]. Advanced MRI techniques, including diffusion-weighted imaging (DWI), offer important diagnostic advantages over conventional MRI sequences and computed tomography for the detection of early signs of ischemia [10]. It has been suggested that the presence of new brain lesions on MRI is associated with a decline in cognitive test performance, but reports with no correlation between cognitive decline and MRI abnormalities also exist [11].
The aim of this study was to characterize long-term cognitive performance in a cohort of patients enrolled in a prospective study of cognitive outcomes after on-pump coronary bypass surgery, and to elucidate the significance of ischemic brain injury as detected by objective MRI for the development of early and late cognitive changes. Additionally, we sought to identify predictors of late cognitive decline.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
2 mg/dL), active liver disease, unstable angina, previous cardiac surgery or contraindications for MRI (eg, severe claustrophobia, pacemaker) were excluded.
Neuropsychologic Testing
Study participants were administered a battery of standardized neuropsychologic tests at baseline, early before discharge, 3 months, and 3 years after coronary artery bypass surgery. All examinations were performed individually by the same experienced neuropsychologist. Discharge examination was administered when the patient was able to sit in a chair, free of sedative and narcotic medication, and when chest drains were removed. Special care was taken to ensure that investigation was performed at the same time during day in silent rooms outside the hospital wards. At 3-year follow-up examination, all patients were tested as outpatients except 1 patient who was in reduced general condition and therefore being tested at home. To minimize learning effects, parallel test forms were used and randomly assigned to the examination. The following tests were administered to evaluate performance in 7 major areas of cognitive functioning known to be vulnerable to organic injury: (1) executive function: Trail Making Test B, a timed task requiring the patient to connect numbers and letters alternately in a sequential order; (2) attention: Zimmermann Joint Attention Test, measures the time to react on auditory and visual signals; (3) psychomotor speed: Trail Making Test A, a timed task requiring to connect numbers sequentially; (4) verbal memory: Verbal Learning Test, delivering measures of Immediate Recall and Delayed Recognition; Wechsler Memory Scale-Revised Digit Span Test, measures short-term and working memory for numbers (subtests Forward and Backward, respectively); (5) visual memory: Corsi Block Tapping Test: requires the patient to reproduce a sequence of blocks the investigator had tapped on (subtests Forward and Backward); (6) logical thinking: Horn Performance Test 55+ subtest 3, a task to identify irregularities in a set of 8 geometrical designs; and (7) visuoconstruction: Horn Performance Test 55+ subtest 9, measures visuospatial abilities.
Magnetic Resonance Imaging of the Brain
Magnetic resonance imaging was performed on a 1.5-T Sonata whole body imaging system (Siemens AG, Erlangen, Germany). All patients were imaged before surgery and at discharge, 3 months, and 3 years after surgery. The protocol included the following five sequences: (1) transaxial T1 (repetition time [TR] 500 ms, echo time [TE] 14 ms, averages 2); (2) transaxial T2 turbo spin-echo (TR 5120 ms, TE 115 ms, average 1); (3) coronal T2 (TR 4810 ms, TE 113 s, averages 2); (4) transaxial fluid-attenuated inversion recovery (TR 9000 ms, TE 115 ms, average 1); and (5) transaxial diffusion-weighted (DW) performed as a single shot, spin echo planar sequence of the whole brain (TR 4600 ms, TE 137 ms, averages 2, gradients of b-volume: 0, 500, 1000 s/mm2). Matrix of images was 256 x 256 for all images except DWI where it was 128 x 128. Field of view was 230 mm and slice thickness 6 mm for all sequences. Brain and ventricular abnormalities (eg, subcortical atherosclerotic encephalopathy, infarction, atrophy, punctuate lesion) were documented. Scans were read by two experienced neuroradiologists blinded to the clinical and neuropsychologic data of the patient. For volumetric quantification of new DWI abnormalities on postoperative scans, the area of lesion was delineated manually in each slice, and volume was calculated using standard scanner software (Synego; Siemens AG).
Anesthesia and Surgery
Standard anesthetic techniques were administered using narcotics (flunitrazepam, ethmidate), analgesics (fentanyl), inhalation agents (isofluran), and paralytics (rocuronium). Median sternotomy and cardiopulmonary bypass (CPB) in mild hypothermia (28°C to 32°C) were performed in all patients. The CPB technique included the use of a roller pump (SIII; Stöckert GmbH, Munich, Germany), membrane oxygenator (Dideco Avant D903; Stöckert GmbH), 40-µm arterial line filter and 
-stat blood gas management. Nonpulsatile pump flow rates were established to achieve a mean arterial pressure between 50 and 70 mm Hg. Myocardial protection was administered with a cardioplegia delivering system antegrade cold Bretschneider solution and ice cold water for topical hypothermia. Central anastomoses were fashioned using a partial side biting aortic clamp.
Statistical Analysis
In a primary data analysis, we examined within-patient changes in neuropsychologic test scores between different testing points (eg, from baseline to discharge, from discharge to 3 months, and so forth). This approach allowed identifying changes in cognitive test performance over time. To do so, raw test scores were transformed into z scores based on the mean and standard deviation (SD) of the preoperative cognitive scores of the study group. Normality of distribution was tested with the Kolmogorov-Smirnov test. For timed tests (Trail Making Tests A and B, Zimmermann Joint Attention Test), the sign of the z score was reversed so that improved performance resulted in a higher score in all tests. Mean changes of within-patient z scores (later score minus earlier score) between two testing points were analyzed using one-sample Student t test and Wilcoxon test, respectively. In a second approach, the incidence of cognitive decline was defined for each individual. A clinically significant cognitive decline was defined as a drop of 1 SD or more in z scores in at least three tests. To identify factors that may be related to cognitive decline at long-term follow-up, univariate logistic (cognitive decline as a dichotomous variable) and linear (within-patient z score change) regression analyses were performed, respectively. The univariate relationship between two categorical variables was analyzed using 
2 or Fisher exact test as appropriate. Potentially important covariates included demographic and medical history variables, factors related to the operation, and postoperative factors. All analyses were conducted using SPSS software for windows (SPSS 13.0; SPSS, Chicago, Illinois). Continuous data are presented as mean ± SD and categorical data are presented as counts (%). All p values less than 0.05 were considered to be statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
|
|
|
|
Between 3 months and 3 years, scores declined in 7 measures, with significant declines in 2 measures (Trail Making Test B, Horn Test No. 9). A marked decrease in cognitive function was also seen in Delayed Recognition, but the difference did not reach the level of significance (p = 0.08).
Comparing cognitive performance at 3-year follow-up with baseline, scores did not differ for most tests. However, the cognitive test score in Delayed Recognition (verbal memory) was statistically significantly worse than at baseline (p = 0.011). Additionally, there was marked, although nonsignificant decline in Trail Making Test B (p = 0.064).
In addition to the determination of mean changes in within-subject test scores between different testing points, we defined the incidence of cognitive deficit in each patient. Postoperative cognitive deficits were observed in 56% (22 of 39) of the patients at the time of hospital discharge, in 23% (9 of 39) at 3 months, and in 31% (10 of 32) at 3 years.
Finally, we sought to identify factors that were associated with cognitive decline. The occurrence of cognitive deficit at discharge (p = 0.025, 2 test) and decline in verbal learning test at discharge (subtests Immediate Recall, p = 0.039; Delayed Recognition, p = 0.043, logistic regression) were found to be univariate predictors of late cognitive deficit. An association of early or late cognitive changes to the number or volume of new ischemic brain lesions demonstrated on postoperative DWI was not found in this cohort. Cognitive changes were also unrelated to MRI detected preexisting brain abnormalities. There was also no correlation between demographic, medical history, or perioperative surgical factors and postoperative cognitive test performance. The results of the logistic regression analysis for factors possibly related to cognitive outcome are given in Table 6.
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
For obvious reasons, long-term decline in cognition after cardiac surgery is of greater concern than transient cognitive changes that generally resolve within 3 months or sooner [4]. From 3 months to 3 years after surgery, there was a decline in some cognitive measures, and the rate of cognitive deficits slightly increased from 23% to 31%. Like others [5, 16], we found that early cognitive decline was a predictor of long-term cognitive dysfunction in both logistic and linear regression analyses. In 2001, Newman and associates [5] reported on a series on 261 CABG patients who were examined preoperatively and serially after surgery. At 5 years, a surprising 42% of patients had late cognitive decline on a global measure of cognition. In two studies from the group of Selnes and colleagues [6, 17], significant decline was observed in certain cognitive domains at 5 years. The only previous prospective long-term study that failed to demonstrate late cognitive decline after CABG was reported on a smaller group of 52 patients [7]. Thus, several studies including our own show that some degree of late decline does occur between baseline and 3- to 5-year follow-up testing [5, 6, 18, 22].
Diffusion-weighted MRI detected new focal brain abnormalities on postoperative scans in 51% of patients. We assume that the observed abnormalities correspond to small areas of brain infarctions caused by emboli arising from the heart or the ascending aorta during surgical intervention. This assumption is based on the acuity of new lesions, their radiographic appearance (small, rounded), and their distribution and multiplicity (half of the patients had multiple lesions, often in different cerebral artery territories), although the finding of discrete areas of signal hyperintensity is nonspecific. One substantial limitation of previous studies employing conventional MRI is that many CABG patients have old white matter hyperintensities on baseline scans that may hide new superimposed lesions [19]. Other limitations refer to the use of low-field strength scanners (0.15 to 1.0 T), broad slice thickness (sometimes 8 mm), and limited sequences (often T-1 and T-2 images) [11, 20]. Therefore, it is supposed that some of the previous studies actually reported false negative MRI examinations or underestimated the extent of injury. In the few studies that have assessed cognition, the correlation between the presence of new lesions on MRI and cognitive dysfunction is variable. In a study on 13 patients undergoing CABG, those with postoperative new DWI defects (31%) had a larger cognitive decline than their counterparts with stable MRI [21]. In a larger series on 35 CABG patients, DWI detected new lesions in 26% of patients [22]. Very recently, a study on 50 cardiac surgical patients reported an incidence of acute perioperative cerebral ischemia in 32% of patients [23]. In these studies including our own, the impairment of cognitive function was not related to evidence of structural cerebral ischemia as detected by DW-MRI.
There are some possible explanations for the lack of correlation between postoperative cognitive dysfunction and presence of cerebral ischemic lesions. First, one could criticize the DWI technology as insufficiently sensitive to detect very small cerebral ischemia. While this is theoretically possible, DW-MRI is the most sophisticated neuroimaging modality available today. The sensitivity and specificity of DWI for subcortical infarction has been reported at 95% and 94%, respectively [24]. The resolution is in the range of 2 to 3 mm. Imaging was also conducted at a time when acute ischemic lesions would be evident. Second, it is conceivable that the neuropsychologic measures were not appropriate or not sensitive enough to detect subtle changes in cognitive functioning. Although this may be true for some studies, we do not believe that it pertains to this investigation because we detected, by use of a broad battery of standardized tests to several domains, significant cognitive decline at various postoperative testing points. Third, as far as early cognitive function is concerned, transient metabolic disorder due to cerebral and systemic inflammatory activation may be responsible for decline [22]. Identification of significant cognitive decline after noncardiac major surgery [25], similar longer term outcomes in interventional cardiology and CABG patients [26], and the absence of a meaningful effect of avoiding cardiopulmonary bypass [27] prompt the proposal that late cognitive decline after CABG may not be specific to the use of cardiopulomonary bypass, but rather represents a separate process related to underlying microvascular disease of the brain or other age-related changes.
Interpretation of the findings from our study is limited by several factors. Loss to follow-up is almost inevitable in a study that follows up patients for 3 years. The follow-up rate was 74% (39 of 53) at 3 months and 60% (32 of 53) at 3 years and is thereby comparable to previous long-term studies of cognition after CABG (follow-up rates: 57% to 66%) [4, 5, 7]. It may be that we enrolled a population of patients undergoing CABG that was biased toward those with a better outcome. Any study of postoperative outcomes that includes a comprehensive neuropsychologic test battery almost inevitably implies a certain selection bias. Because we analyzed cognitive test scores intraindividually across follow-up time, such a bias would probably be of minor importance in our study.
Most studies using MRI have intrinsic methodologic limitations, particularly small sample size. The largest study that applied conventional MRI after CABG comprised 38 patients [20]. Using DW-MRI, larger series included 35 [19] and 50 patients [23], respectively. However, in the latter, only 27 of the patients had CABG. Thus, to date the present cohort of 39 CABG patients represents the largest that was administered DW-MRI and cognitive assessment before and serially after surgery for as long as 3 years. A limitation of this study is also the lack of a control group. Therefore, it is not possible to determine whether early and late cognitive dysfunction are two phenomena of the same process resulting from the use of CPB or the operation itself or if late decline represents a separate process caused by aging or other age-related factors in a population of patients with cardiovascular risk factors. In a large randomized trial that tested the hypothesis that elimination of cardiopulmonary bypass would improve cognitive outcome, no significant difference was found in the frequency of cognitive dysfunction at 3 months and 12 months between patients undergoing on-pump versus off-pump surgery [27]. Similar results were shown for elderly high-risk patients [28]. When comparing patients undergoing on-pump CABG with nonsurgical controls, a possible link between late cognitive decline and previous surgery has been suggested. By means of objective P300 auditory evoked potentials as a global measure of cognition, it was shown that cognitive measures in surgical patients were significantly impaired as compared with nonsurgical controls and cognitive deficit at four months was predictive of long-term deficit at 3 years [16].
In summary, the present prospective investigation showed a two-stage course of cognition after CABG characterized by early decline in most domains during the initial postoperative days that improved by 3 months. A second later decline was observed in certain measures at 3 years. Acute perioperative ischemic cerebral lesions on DW-MRI were not associated with early or late cognitive decline. Early postoperative decline was predictive of late cognitive outcome.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  P. Wang, M. A. Acker, M. Bilello, E. R. Melhem, E. Stambrook, S. J. Ratcliffe, T. F. Floyd, and DENOVO (Determining Neurologic Outcomes from Valve Sex, Aging, and Preexisting Cerebral Ischemic Disease in Patients With Aortic Stenosis Ann. Thorac. Surg., October 1, 2010; 90(4): 1230 - 1235. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. W. J. Kruis, F. A. E. Vlasveld, and D. Van Dijk The (Un)Importance of Cerebral Microemboli Seminars in Cardiothoracic and Vascular Anesthesia, June 1, 2010; 14(2): 111 - 118. [Abstract] [PDF]
|
|
|
|
 |
|
 C. N. Hallas, J. L. Burke, D. G. White, and D. T. Connelly A Prospective 1-Year Study of Changes in Neuropsychological Functioning After Implantable Cardioverter-Defibrillator Surgery Circ Arrhythm Electrophysiol, April 1, 2010; 3(2): 170 - 177. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. J. Tully, R. A. Baker, J. L. Knight, D. A. Turnbull, and H. R. Winefield Neuropsychological Function 5 Years after Cardiac Surgery and the Effect of Psychological Distress Arch Clin Neuropsychol, December 1, 2009; 24(8): 741 - 751. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Grigore, C. F. Murray, H. Ramakrishna, and G. Djaiani A Core Review of Temperature Regimens and Neuroprotection During Cardiopulmonary Bypass: Does Rewarming Rate Matter? Anesth. Analg., December 1, 2009; 109(6): 1741 - 1751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. A. Barber, L. J. Tippett, A. Merry, S. Hach, C. Frampton, and P. Milsom Postoperative Ischemia and Cognitive Impairment in Cardiac Surgery Patients Ann. Thorac. Surg., February 1, 2009; 87(2): 672 - 673. [Full Text] [PDF]
|
|
|
|
|
|
S. C. Knipp, H. Jakob, C. Losch, N. Matatko, H. Wilhelm, H. C. Diener, and M. Schlamann Reply. Ann. Thorac. Surg., February 1, 2009; 87(2): 673 - 674. [Full Text] [PDF]
|
|
|
|
|
|
H. P. Grocott Invited Commentary Ann. Thorac. Surg., March 1, 2008; 85(3): 879 - 879. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
