| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;75:1377-1386
© 2003 The Society of Thoracic Surgeons
a Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
b Division of Cardiac Surgery, Department of Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
c Department of Neuroscience, Johns Hopkins University School of Medicine, , USA
d Department of Biostatistics, The Johns Hopkins Bloomberg School of Public Health, , USA
e Zanvyl Krieger Mind/Brain Institute, Baltimore, Maryland, USA
* Address reprint requests to Dr Selnes, Department of Neurology, Meyer 100, The Johns Hopkins Hospital, 600 North Wolfe St, Baltimore, MD 21287, USA
e-mail: oselnes{at}jhmi.edu
Presented at the Thirty-ninth Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31–Feb 2, 2003.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
METHODS: Patients undergoing CABG (n = 140) and demographically similar nonsurgical control subjects with CAD (n = 92) completed baseline neuropsychological assessment and were followed prospectively at 3 and 12 months. Cognitive function was evaluated with a battery of neuropsychological tests assessing the cognitive domains of attention, language, verbal and visual memory, visuoconstruction, executive function, and psychomotor and motor speed.
RESULTS: The CABG patients who were tested in their hospital rooms before surgery had lower scores for timed tests; however, after adjustment for demographic variables and testing location there were no statistically significant differences between the CABG and nonsurgical control subjects in baseline neuropsychological test performance. Both groups improved from baseline to 3 months; the only statistically significant group difference was a greater improvement in the CABG group with regard to verbal memory. At 12 months there were no statistically significant differences between the two groups.
CONCLUSIONS: The prospective longitudinal neuropsychological performance of patients with CABG did not differ from that of comparable nonsurgical control subjects with CAD at 3 months or 1 year after base line examination. This suggests that the previously reported cognitive decline during the early postoperative period after CABG is transient and reversible. Continued follow-up will determine whether a specific "late decline" occurs in CABG patients but not in nonsurgical control subjects with similar risk factors for cardiovascular and cerebrovascular disease.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
To further evaluate neurocognitive changes after CABG, we assessed patients both before and after surgery with a battery of standardized neuropsychological tests. In addition we included a group of patients with coronary artery disease diagnosed by cardiac catheterization as an appropriate nonsurgical control group. The control patients were similar to the CABG patients in terms of risk factors for both coronary artery disease and cerebrovascular disease. In this report, we describe the prospective longitudinal changes in neuropsychological test performance from baseline to 3 months and 12 months.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Neuropsychological tests
Study participants were administered a battery of standardized neuropsychological tests at baseline, 3 months, and 12 months. Most of the study participants were tested as outpatients, but approximately one third of the CABG patients were tested in their hospital rooms shortly before their surgery. The following tests were selected to evaluate performance in eight major areas of cognitive functioning (see Lezak [11] for a more complete description of the tests): (1) Verbal Memory: Rey Auditory Verbal Memory Test, a word-list learning task assessing verbal learning, retention, and recognition memory; (2) Visual Memory: Rey Complex Figure–Recall, a measure of the ability to recall a complex visual design previously copied; (3) Language: Boston Naming Test (short form), a measure of visual confrontation naming requiring the subject to name a series of 30 line drawings; (4) Attention: Rey Auditory Verbal Learning test–Trial 1; Attention score from Mini-Mental State Examination; (5) Visuoconstruction: Rey Complex Figure–Copy, a measure of visuospatial abilities requiring the subject to copy a complex visual design; (6) Psychomotor: Trail Making Test–A, a timed task that requires the subject to connect numbered circles in sequence as quickly as possible; Written Alphabet, a timed measure of psychomotor speed in which the subject is asked to write the letters of the alphabet as quickly as possible [12]; (7) Motor Speed: Grooved Pegboard Dominant and Nondominant hand, a test of motor speed measuring how quickly the subject is able to place 25 keyed pegs in an array of 5 x 5 holes with randomly positioned slots; and (8) Executive function: Trail Making test–B, a timed test of psychomotor speed that requires the participant to connect numbered and lettered circles alternately in sequential (numeric and alphabetical) order. Patients were also administered the Mini-Mental State Examination (MMSE). The Center for Epidemiological Studies Depression scale (CES-D) [13] and Functional Status Questionnaire (FSQ) were also administered at baseline and follow-up [14].
Operative technique
All patients underwent median sternotomy and received at least one arterial graft. Anesthetic technique was standardized and consisted of low–intermediate dose narcotics, inhalation agents, and paralytics. Cardiopulmonary bypass was carried out using a Sarns roller head pump, nonpulsatile flow, membrane oxygenator, 
-stat pH blood gas management, antegrade crystalloid cardioplegia and topical hypothermia, moderate systemic hypothermia (28° to 32°C), and pump flow rates to achieve a mean arterial pressure of 60 to 80 mm Hg. There was no routine use of intraoperative ultrasonic aortic scanning. Cardiotomy suction was returned to the CPB circuit for all patients. All surgeons but 1 used the double-clamp technique. The aortic cross clamp was applied and distal anastomoses were made. The aortic crossclamp was then released and a side-biting clamp was applied once or twice, after which the proximal anastomoses were made.
Statistical methods
The primary data analyses examined within-patient changes in neuropsychological test scores from baseline to 3 months and 12 months. This approach allows both improvement and decline in performance to be taken into account. All analyses were performed using z-scores based on the mean and standard deviation of the base line performance of the nonsurgical control patients (N = 92). For cognitive domains with more than 1 test, each patient received a composite score consisting of the mean of the z-scores for the individual tests. These were then renormalized so that the nonsurgical control group had a mean of 0 and standard deviation of 1 at baseline. Group differences in base line scores were examined using linear regression models adjusting for demographic variables that are known to be important predictors of neuropsychological test performance, including age, education, sex, and ethnicity. In addition, we also adjusted for testing location, as being tested in a hospital room might be associated with lower performance on some tests because of inadvertent distractions or background noise. To examine how the changes in neuropsychological test z-scores over time might be related to subject-specific covariates, we also used linear regression analysis. The response variable was the difference in an individual subject’s test scores on two occasions (later score minus earlier score). For timed tests, the sign of the z-score was reversed so that improved performance resulted in a higher score for all variables. To minimize the effect of outliers, z-score changes for the timed tests were truncated at 3 standard deviations. The covariates included were demographic variables and medical history variables.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Demographic characteristics for the CABG patients and the nonsurgical control subjects are shown in Table 1. There were no statistically significant differences in demographic characteristics between the two study groups other than age (a mean of 2.4 years higher in the nonsurgical controls). The CABG group did, however, differ from the nonsurgical controls with regard to several medical history variables, including a higher prevalence of peripheral vascular disease, history of recent myocardial infarction, and number of diseased vessels. The frequency of previous percutaneous transluminal coronary angioplasty procedures was statistically significantly higher among the nonsurgical controls. There was no meaningful difference in the prevalence of history of diabetes, hypertension, family history of Alzheimer’s disease, or apolipoprotein E distribution in the two groups. The baseline Functional Status Questionnaire score was lower in the CABG group, and the mean CES-D score was higher, suggesting higher frequency of self-reported symptoms of depression among the CABG patients at baseline.
|
|
|
|
Although there were no statistically significant within-subject changes for some cognitive domains, multivariate analysis nonetheless identified some factors that were significantly associated with decline. At 3 months the presence of carotid bruit was associated with less improvement in the domains of visual memory (p < 0.02) and executive function (p < 0.007), whereas increasing age was associated with less improvement in language (p < 0.001). At 1 year increasing age was associated with worse performance in the domains of visuoconstruction (p < 0.02), language (p < 0.01), motor speed (p < 0.03), and executive function (p < 0.01). A history of past cerebrovascular accident was associated with worse performance in the domains of attention (p < 0.01), visual memory (p < 0.007), psychomotor speed (p < 0.0004), and executive function (p < 0.002). A history of diabetes was associated with worse performance in visual memory (p < 0.009).
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
In our previous studies [8], we and others have observed a significant decline in verbal memory and other cognitive domains 1 month after surgery. In contrast, in the present study, we found no cognitive decline 3 months after CABG; rather, there was statistically significant improvement in five of the eight domains. Moreover, the degree of improvement from baseline to 3 months was similar to that observed for the nonsurgical control group except that the CABG group showed greater improvement in the domain of verbal memory. These findings strongly suggest that the cognitive changes that have been reported to occur up to 1 month after CABG are transient and reversible. On the other hand, coronary artery disease, even at subclinical levels, influences cognitive performance in older individuals [15], and it has been shown that symptomatic coronary artery disease may be associated with significantly lower performance in some cognitive domains [9].
Accordingly, some recent studies have addressed the question of preoperative cognitive impairment in candidates for cardiac surgery [16, 17]. Vingerhoets and colleagues [16] reported significantly lower performance on measures of verbal memory, executive function, and motor speed in a group of 77 patients scheduled for CABG. In their multivariate analysis, duration of heart disease was one of the variables significantly associated with worse baseline performance. In the present study we found that, at baseline, CABG patients tested in the hospital the night before surgery performed worse on measures of verbal memory, attention, and psychomotor speed; however, after adjustment for testing location, there were no significant differences in baseline scores between CABG patients and controls. This suggests that testing location may explain why CABG patients perform worse at baseline in some cognitive domains, although other factors such as duration or degree of cardiovascular disease may still play a role.
The findings in this study emphasize the need for comparable control groups in investigations of cognitive outcomes after surgical procedures, and raise questions about interpretation of previous data on cognitive decline. Most of these prior studies, including our own, prospectively examined only the surgical population. Thus, "abnormality" or decline was defined in a number of ways, including a negative change from baseline of 0.5 SD in one or more cognitive domains, a 1 SD decline in any three cognitive tests, or a 20% drop in performance on any test [18–20]. Applying any of these criteria to the two groups in the present study, we find that a similar proportion of patients declined in both the surgical and nonsurgical groups.
Some previous studies have included a surgical control group to control for the effects of anesthesia and nonspecific effects of major surgery. For example, Treasure and colleagues [21] found that a larger percentage of CABG patients than controls with major vascular or non-CPB thoracic surgery had moderate or severe cognitive deficits 8 days after surgery. Nevertheless, 2 months after surgery, the incidence of cognitive impairment for the CABG patients (37%) was not statistically different from that of the surgical controls (50%) [21]. Similar findings were reported by Murkin and colleagues [4], who included a control group of 40 patients with major vascular thoracic surgery; although a greater proportion of the CABG patients had cognitive impairment 7 days postoperatively, the incidence of impairment at 2 months was similar for the two groups.
In our study, the available control group data do not allow us to take into account possible adverse effects of the surgery itself, anesthesia, or other factors associated with the surgery. Because the performance of the CABG group was comparable to that of the nonsurgical controls, there is no evidence that surgical factors by themselves resulted in cognitive changes that persisted beyond the first few weeks after surgery. Several recent studies have compared the incidence of cognitive decline after CABG with and without CPB [22, 23]. These studies have reported no significant group differences in the frequency of either early postoperative or later (ie, 3 months and 12 months) cognitive decline, thus providing additional support for the hypothesis that CPB may not be the sole source of cognitive impairment after cardiac surgery.
Our study has several possible limitations. 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 neuropsychological test battery almost inevitably implies a certain selection bias. The patients did not differ significantly in age (which has previously been found to predict worse outcome after CABG) or in risk factors for stroke. The incidence of some type of neurologic symptoms during the immediate postoperative period was greater in the patients who had CABG during the study enrollment period but who did not participate in our study (n = 989), and their length of stay was also significantly longer (data not shown).
Moreover, as with any prospective study, it is possible that the more impaired patients may have been lost to follow-up. For most cognitive domains, the baseline scores of both CABG and nonsurgical control patients who did not complete the 3- or 12-month follow-up testing were significantly lower than the scores of those who completed the testing. This suggests that the attrition was not likely to be due to random factors. The dropout rate for the CABG patients was similar to that of the nonsurgical controls, however; and overall, the dropout rate (<20%) was lower than reported for other prospective studies of cognitive outcomes after CABG.
Another possibility to account for the absence of decline at 3 months and for the similar results in the CABG and nonsurgical groups is that the neuropsychological measures were not sensitive enough to detect small changes in cognitive performance. We think that this is less likely, as we used essentially the same test battery for our previous study in which we documented decline in several cognitive domains 1 month after surgery [8]. The analytical methods in our previous study were, however, different in that cognitive decline was defined arbitrarily as a drop of 0.5 SD or more in any cognitive domain. Nonetheless, we cannot rule out the possibility that some patients in our current study may have had cognitive decline that was not detected by our test battery or our use of summary scores for each cognitive domain. For example, certain cognitive domains (such as, working memory) may not have been adequately assessed by our test battery.
In studies that have relied on self-report measures to quantify cognitive changes after bypass surgery, a surprisingly high incidence of such subjective memory complaints has been observed at different time points after surgery [24, 25]. However, self-report measures have not always been in accordance with standardized neuropsychological tests when both have been used in a study. Although the ultimate significance of the discrepancy between self-reported cognitive symptoms and objective test findings remains unclear [26], it is possible that some of our patients may have had subclinical cognitive changes that were not detected by the cognitive tests.
By using the performance of the nonsurgical control subjects as the reference point, we can only draw conclusions with respect to how the performance of the CABG group differs from this group. A healthy control group with no history of coronary artery disease would most likely be expected to have higher baseline performance and greater improvement from practice effects than both of our groups with coronary artery disease.
Finally, although it is possible that a subgroup of the CABG patients did experience cognitive decline, even though the means of the change scores for the two groups were similar, our post hoc analyses did not reveal any such subgroups for any of the cognitive domains.
In summary, our observations indicate that the changes in neuropsychological test scores from baseline to follow-up are very similar for surgical and nonsurgical control patients. There is little or no evidence that the means of the change scores, standard deviations, or shapes of the distributions of changes are different between the two groups. Our inability to detect a difference between the two groups does not preclude that a difference may still exist. Given the sample size and variability as observed in the change scores, however, the study had 80% or more power to detect a difference on the order of 0.3 baseline standard deviation for most cognitive domains. Therefore, the observed data are not consistent with a sizable difference between the surgical and control groups. These data indicate that early postoperative cognitive changes after CABG may be transient and reversible by 3 months. Continued follow-up of our cohort will help to determine the extent of any late cognitive decline in the CABG group, and whether this decline is different from or similar to that observed for our nonsurgical control group.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
For many years, the message that cardiac surgery is inherently and frequently associated with brain injury and cognitive decline has been reported in scientific and lay publications. In February 2001, the influential Duke University Neurological Outcomes Research Group reported in a widely quoted article in the New England Journal of Medicine that they had observed a 53% incidence of measurable neurological injury immediately after CABG surgery. Although some of these patients recovered cognitive function during the first year after surgery, nearly 80% of those with early measurable injury demonstrated a significant loss of cognitive function from their preoperative status when reexamined at 5 years.
This particular study from Duke and several other scientific papers reported in the lay press have contributed to the widely held belief that heart surgery in general and coronary bypass grafting with the use of a heart-lung machine in particular lead to brain damage and neurocognitive decline in many patients.
This illustration depicting the embolization of debris from surgical manipulation of the ascending aorta appeared in Time magazine shortly after the Duke report was published in February 2001. Even a sophisticated reader would conclude from this article in Time that the incidence of significant neurological injury 5 years after CABG surgery is more than 40%; and, as many of us have experienced, patients referred to us in need of CABG surgery have been very concerned about these reports. The corollary assumption is that this decline in function is a result of the original bypass operation, as shown graphically on this particular illustration, and is not a predictable and understandable consequence of aging or progression of vascular disease.
This present report now provides some early important additional information, namely, that when compared to a nonoperated group of patients with coronary artery disease, there was no greater deterioration in cognitive function after 3 months, or, as I read the manuscript, after 1-year follow-up. In other words, despite the detection of neurocognitive changes immediately after CABG surgery, when compared to similar patients not having surgery, the CABG patients have equally good neurological function a year later.
Cardiac surgeons and our clinical partners, especially cardiac anesthesiologists, have learned much about how to avoid or minimize central nervous system injury during cardiac surgery. Equally important is our ability to use predictive models to identify patients at higher risk and then to modify the patient’s perioperative management. Although cardiac surgery continues to pose a real and probably irreducible risk of neurological injury, both permanent and transient, this present study puts the scope of the problem in much clearer focus. It does so because it was designed to follow conventional investigational principles such as inclusion of a control population, it employed the best possible testing instruments for neuropsychological evaluation, and it followed strict statistical analytic methods.
The most important early message from the study, I believe, is the implication that exposure to cardiopulmonary bypass does not, a priori, lead to a predictable and substantial occurrence of sustained neurocognitive dysfunction. Even with the capability of beating heart surgery techniques for some CABG procedures, the heart–lung machine cannot be dispensed with at any time in the foreseeable future for the management of many patients requiring heart surgery.
In addition to asking Dr Selnes and colleagues about their reluctance to present their 12-month data, I would ask Dr Selnes this, and from the perspective of his long-term involvement in this area: Do you foresee the feasibility and near-term availability of neuroprotective agents that might result in improved central nervous system tolerance to episodes of ischemia or pump-related metabolic changes that are now believed to be a consequence of cardiac surgery?
Drs Baumgartner, McKhann, and Selnes and their skilled neuropsych research team at Hopkins deserve congratulations for this excellent and timely study. Thank you.
DR RICHARD J. NOVICK (London, Ontario, Canada): Doctor Selnes, as a surgeon with additional training in epidemiology, I congratulate you on your study, which included three separate control groups. One note of caution, however, is that in one group, 16% of patients were lost to follow-up at 3 months and in another group 17%. It is quite possible that the patients lost to follow-up had a worse neurocognitive outcome than those who in fact were studied at the 3-month interval. Before espousing the message that on-pump coronary artery bypass does not impact postoperative neurocognitive outcomes, it is important that we strive to attain as complete a follow-up as possible in these patient subsets and in the controls.
DR THORALF SUNDT (Rochester, MN): My question was similar. I am just concerned about our ability to get meaningful data with incomplete follow-up, and I wonder if you could address whether or not you think that that is a fatal flaw in the study and how you plan to address that criticism of your degree of follow-up. Thanks.
DR CARY W. AKINS (Boston, MA): I really enjoyed your presentation. I have one question about your control group. Knowing that they had coronary disease but not bypass surgery, did any of these people have catheter interventions during the study period? Because we know that cholesterol embolization and other complications are potentially associated with cardiac catheterization, is it possible that they suffered catheter-induced neurologic dysfunction consistent with what some have claimed happens with cardiopulmonary bypass?
DR BRIAN D. MOTT (Clarks Summit, PA): In those patients that you had who did have some decline in cognitive function at 3 months, were there any pre-, intra-, or postoperative variables that you identified that contributed to that decline? Thank you.
DR KENNETH OBERHEU (Dayton, OH): As we know, nicotine does increase memory, and the reason baseball players chew tobacco is it increases their hand–eye coordination. Have you had any studies where people have been smokers and the reason why their memory has been decreased is because they have stopped smoking after surgery?
DR SELNES: Thank you, Dr Gardner, for your helpful comments. The reason that I chose not to include the 1-year follow-up data is that, interestingly, no one has reported any decline at 1 year. This is true for the long-term data from Duke that you just showed us: they found no evidence of decline at 1 year. We have found the same thing in our own studies. Therefore, at 1 year, patients appear to be back to normal. So I thought it would be more interesting to focus on the short-term findings and to emphasize that even at 3 months we do not see any evidence of decline or any difference between our CABG group and the nonsurgical controls.
In terms of the use of neuroprotective agents, I think there is some consensus that coronary artery bypass surgery represents a potentially very useful model to test some of these new agents. As far as I know, there are several products in the pipeline, but at this point, I am not aware of any study outcomes to tell you about.
The loss to follow-up is a difficult problem for all studies that have investigated cognitive decline after bypass surgery. I think our follow-up rate is actually somewhat higher than that in some published studies, but at this point we don’t have any obvious solution for how to avoid the problem that some patients will not come back for follow-up testing; however, we are somewhat reassured by the fact that the dropout rate in our controls is not different from that among the surgery patients. So while I agree that incomplete follow-up will always remain a problem for studies like these, I would not consider it a fatal flaw.
With regard to the question about catheter interventions, all of the nonsurgical controls had cardiac catheterization at some point prior to enrollment in the study. In addition, approximately half of the nonsurgical controls had also had percutaneous transluminal coronary angioplasty, so there is a possibility that this may have been associated with embolization that could have influenced the cognitive performance of the control group.
The last question was about the effect of nicotine on memory. The neuropsychopharmacology of nicotine and smoking is a very controversial topic. There is at least one study that has reported a relationship between stopping smoking after myocardial infarction and subjective memory symptoms. We have not yet looked for any similar associations in our study. More than half of both our CABG and nonsurgical controls had a history of smoking, and the frequency of patients and controls who had stopped smoking was similar in the two groups. Several years ago we did a trial of the use of Nicorette gum in patients with early Alzheimer’s disease, but we were not able to demonstrate any clinically significant improvement in the memory performance of these patients. Whether nicotine helps non-Alzheimer’s patients who have stopped smoking is a very complex question, and I am not sure there is any consensus on this issue yet.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
