- Email: [email protected]
Cognitive outcome after cardiac operations
Cognitive outcome after cardiac operations Relationship to intraoperative computerized electroencephalographic data Cardiopulmonary bypass frequently causes new postoperative neuropsychologic deficits. To assess whether these deficits could be predicted or limited, we monitored 29 patients receiving bypass intraoperatively with an on-line computerized electroencephalograph. We hypothesized that the 15 patients whose cerebral perfusion pressure was adjusted on the basis of this electroencephalographic data would have fewer postoperative deficits than the 14 patients whose pressure was monitored on the basis of systemic pressure. The results showed that new postoperative cognitive deficits in both groups were less prevalent than in previous studies, but there was not a significant difference in outcomes between the two groups. The intraoperative electroencephalographic records correlated with surgical, but not neuropsychologic, outcome. It is concluded that careful attention to intraoperative cerebral perfusion pressure may decrease the prevalence of postoperative neuropsychologic complications, but that the use of a computerized electroencephalograph does not necessarily contribute to an improved outcome. (J THoRAc CARDIOVASC SURG 1992;104:1405-9)
Christopher L. Grote, PhD,a Patrick T. Shanahan, MD,b Paul Salmon, PhD,c Robert G. Meyer, Phl),? Curt Barrett, PhD,d and Allan Lansing, MD, Phl)," Chicago, Ill., and Louisville, Ky.
Several investigations have shown that new neuropsychologic deficits develop in the immediate postoperative phase in a majority of patients undergoing cardiopulmonary bypass.L' Although many of these deficits are subtle and reverse themselves within a few weeks of operation, approximately 10% to 20% of patients undergoing bypass continue to show a significant decrement in performance on one or more measures of cognitive ability at a 6-month follow-up.v 6-7 Insufficient intraoperative cerebral blood flow, often a result of decreased cerebral perfusion pressure or cardiac From the Department of Psychology and Social Sciences; Rush-Presbyterian-St. Luke's Medical Center, Chicago, Ill.; Medical Center Anesthesiologists," Louisville, Ky.; the Department of Psychology," University of Louisville, Louisville, Ky.; the Department of Psychiatry," Nortons Hospital, Louisville, Ky.; and the Department of Surgery,' Humana Hospital Audubon, Louisville, Ky. Received for publication June
13,1991. 1992.
Accepted for publication Feb. 24,
Address for reprints: Christopher L. Grote, PhD, Department of Psychology and Social Sciences, Rush-Presbyterian-St. Luke's Medical Center, 1653 W. Congress Pwy., Chicago, IL 60612.
12/1/38383
embolization, has been identified as a primary cause of the development of postoperative neurologic and neuropsychologic sequelae.v 8 Although traditional electroencephalographic (EEG) equipment is sensitive to intraoperative cerebral blood f1ow,9-11 it has not been used often because of the complexity of the equipment and the difficulty of making quick clinical interpretations. I I At least one investigation has used a quantified or computerized electroencephalograph in an attempt to overcome these limitations and found that postoperative global neurologic deficits were less prevalent in patients who received intraoperative interventions on the basis of computerized EEG data.l- However, the way in which these deficits were measured and evaluated was not fully described. The present investigation further explored these issues. In addition to examining the effect of cardiopulmonary bypass on postoperative cognitive function, we investigated whether computerized EEG data correlate with postoperative surgical or neuropsychologic outcome. Another goal was to compare the effect of adjusting cerebral perfusion pressure on the basis of computerized EEG data with that of the more standard technique of adjusting on the basis of mean systemic pressure. 1405
I 406
The Journal of Thoracic and Cardiovascular Surgery
Grote et al.
Table I. Outcome of surgical variables Control group (n = 14)
Bypass length (min) Estimatedbloodloss (mI) Postop. hospital stay (dy) Mean PDI score*
Experimental group (n
= 15)
M
SD
M
SD
70.3 345.0
18.2 125.2 2.4 72.6
72.8 434.6 8.0
161.9
9.1 105.5
132.1
21.2 2.3 178.8
Control vs experimental
NS NS NS NS
M, Mean; SD, standard deviation; NS, not significant.
*The calculation of POI scores is referred to in the text.
Method Patients. Patients undergoing aorta-coronary bypass grafting at Humana Hospital Audubon in Louisville,Kentucky, were eligible for participation. Those with previous strokes, with existing conditions such as diabetes, or about to undergo procedures such as carotid endarterectomy were excluded. However, stage of cardiac disease was not an exclusionary criterion. With institutional review board· approval, eligible patients were approached on admission and were invited to participate in the study. Informed consent was obtained. Twenty-nine patients underwent cardiopulmonary bypass and completed both preoperative and postoperative neuropsychologic testing. The average age of the patients was 59.9 years, (standard deviation, 8.8), with a range of 40 to 82. Twenty-one of the patients were male and eight female. All were white. Patients were randomly assigned to the control (n = 14) or experimental (n = 15) groups. Intraoperative management. All patients received a preoperative intramuscular injection of morphine sulfate and scopolamine. Anesthesia was induced and maintained with sufentanil citrate. All patients had electrocardiogram, temperature (esophageal, rectal, and pulmonary artery), radial artery blood pressure, end-tidal carbon dioxide, Sa02 (pulse oximetry), pulmonary artery pressure, and central venous pressure monitoring. A computerized on-line 16-channel electroencephalograph (CNS monitor, CNS, Inc., Minneapolis, Minn.) was used to monitor all patients during the operation. After placement of the electrodes by the standard 10-20 system and induction of anesthesia, a baseline recording of the patient's EEG frequency and total power was made. Subsequent measurements of the intraoperative level of strength of electrical signal were then made and compared with these baseline recordings. The power drop index (PDI) measures the amount of time and the degree to which the intraoperative electrical activity is less than 40% of baseline, a level that has been associated with postoperative complications.l? Differences in management between the control and experimental groups occurred during the bypass period. Patients assigned to the control group were managed with mean systemic pressure used as a guide. Pressure was maintained between 60 and 80 torr during bypass by the addition of volume to the bypass circuit or with an infusion of phenylephrine. Systemic pressure in patients assigned to the experimental group was maintained between 60 and 80 torr if there was no change in the EEG signal below 40% of the baseline output. If the EEG signal declined below 400/0 of baseline, addition of volume, a phenylephrine infusion, or both methods were used to
raise perfusion pressure. In this effort to raise perfusion pressure, a cutoff pressure of 100 ± 5 torr was used. Perfusion was performed with a membrane oxygenator at a full range of 1.6 to 3.0 L/m2 per minute, with an average of 2.4 L/m 2 per minute. On-line arterial filters and prebypass filters were used on all patients. Moderate hypothermia to a core temperature of 28° C (esophageal-rectal) was used on all patients. Arterial blood gases during bypass were managed by alpha-stat pH methods. A decrease in EEG signal of 6% per degree centigrade was accepted as normal in experimental group patients for decreases in temperature. Decreases greater than 60% (40% of original signal remaining) were deemed to exceed the 6% per degree centigrade margin, and attempts were made to increase perfusion pressure in an effort to improve the EEG signal. The arbitrary upper limit cutoff of 100 torr mean arterial pressure was arrived at in consultation with the surgeons who believed that higher pressures might result in possible crossclamp failure and bleeding into the operative site. In all patients studied, no new asymmetric EEG patterns were observed. Several patients had mild hemispheric asymmetry before bypass, which was not thought to be evidence of previous focal injury. No new focal patterns were seen in any patients. Postoperative neurologic examination did not reveal any new deficits in any of the patients. Neuropsychologic tests. The neuropsychologic tests were administered by a trained psychometrician within 48 hours before operation and again at an average of 10 weeks after operation. The neuropsychologic examination consisted of the following measures: Wechsler Memory Scale.!' Rey AuditoryVerbal Learning Test,14 Trails B,15 Symbol Digit Modalities Test,16 and the Shipley Institute of Living Scale Abstraction Subtest.!" These tests were selected to provide measures of attention, verbal and nonverbal memory, psychomotor speed, and abstract thinking. Results Table I shows the outcome of surgical variables for each group. The groups were not different overall (p > 0.14) on these measures, and none of the univariate comparisons were significant. Table II shows the preoperative and postoperative neuropsychologic scores for both groups. The multivariate comparison of groups on the preoperative measures was not significant (p> 0.22), but the experimental group achieved significantly better scores on all but one
Volume 104 Number 5 November 1992
1 4 a7
Cognitive outcome after cardiac operations
Table II. Intergroup and intragroup comparison ofpreoperative and postoperative test results* Preoperative Control
Wechsler Memory Scale Rey AVLT
Trails B SDMT- Written SDMT-Oral Shipley Abstraction
Postoperative Control
Experimental
Mean
SD
Mean
SD
55.9 37.5 110.2 33.4 40.1 8.4
6.6 8.4 32.7 7.3 7.6 3.2
60.3 41.7 103.0 40.2 45.3 10.9
10.0 12.7 69.4 7.4 8.3 4.1
t NS
t
:t
t t
Experimental
Mean
SD
Mean
SD
61.7 43.8 98.3 38.6 44.7 10.3
7.9 10.0 51.4
63.1 43.7 100.2 43.4 49.7 10.8
10.7 10.1 80.8 11.3 9.5 4.8
9.5 9.5 3.7
Preop. vs Postop. Control
Experimental
:j: :j:
NS NS NS
NS NS NS
NS
NS
t
NS
NS
NS
t
NS NS
t
SD. Standard deviation; NS, not significant; AVLT, Auditory-Verbal Learing Test; SDMT, Symbol Digit Modalities Test.
'Raw scores are reported. With the exception of Trails B, higher scores are more desirable. tp < 0.05. :j:p < om.
of the preoperative measures, despite the use of random assignment to groups. The average age of the control group (61.4 years) was higher than that of the experimental group (58.2 years), but the difference did not approach statistical significance (p > 0.28). There were no other apparent preoperative differences between groups such as stage of cardiac illness,so it is not clear why one group outperformed the other on these measures. On postoperative testing, there were no significant differences between groups on any of the measures. Analysis of changes in test results after operation indicate that neither group had any kind of cognitive decline across time. Instead, there were several instances of significantly improved performance. The control group made large and significant gains on the Wechsler Memory Scale, the Rey Auditory-Verbal Learning Test, and on both subtests of the Symbol Digit Modalities Test. Patients in the experimental group made a significant gain on the oral subtest of the Symbol Digits Modalities Test. Repeated measures multivariate analysis showed that there were no significant differences between the groups (p> 0.88). There was a significant change for both groups across time (p < 0.00 1), but the interaction effect of group by time was not significant (p > 0.80). The lack of a significant interaction offers further evidence of the lack of a group effect and negates the importance of the preoperative differences between groups. Many previous studies have referred only to group data, obscuring individual results in the process. Because neitherof the groups in this study showed a decline on any of the postoperative tests, we decided to combine the groupsand to then examine intraindividual changes. The percentage of patients who had a significant change across evaluations was calculated to provide a perhaps moreaccurate reflection of how patients' cognitive scores
Table III. Percentage ofpatients (both groups combined, n = 29) showing significant changes across testings* Variable
Improvement
Decrement
Wechsler Memory Scale ReyAVLT Trails B Shipley Abstraction SDMT-Written SDMT-Oral
28 24 7 21 31 35
0 0 3 7 10 3
AVLT, Auditory-Verbal Learning Test; SDMT, Symbol Digit Modalities Test.
'Changes betweenevaluationsare consideredsignificant whena subject'schange score is greater than the preoperativegroup standard deviation for that test.
changed after operation. In the manner of Shaw and coworkers," a preoperative-postoperative change was classified as significant if the difference in score exceeded that of the preoperative standard deviation for that test. This analysis (Table III) showed that many more patients had a significant increase than had a decrease on these tests. For example, 28% of the sample had a significant postoperative gain on the Wechsler Memory Scale and none had a significant decrease. Similar patterns of change were seen on each of the other tests, with the exception of Trails B. Correlations between demographic, surgical, and cognitive variables were calculated (Table IV). The control and experimental groups were again combined for these analyses. There was a high degree of interrelatedness among the surgical variables, such as longer periods of bypass correlating with increased age, estimated blood loss,and length of postoperative stay. Similarly, high POI scores correlated with undesirable surgical outcome scores. Although cognitive scores generally improved after
14 0 8
The Journal of Thoracic and Cardiovascular Surgery
Grote et al.
Table IV. Correlations between surgical outcome and change in cognitive score* Surgical variables PDf
Length of bypass Estimated blood loss Days to discharge Age Cognitive Change Scores Wechsler Memory Scale Rey AVLT Trails B Shipley Abstraction SDMT-Written SDMT-Oral
Length of bypass
0.45t 0.65* O.44t 0.45t -0.16 -0.37 -0.22 -0.27 -0.17 -0.05
Estimated blood loss
Days to discharge
Age
0.35t
o.su 0.44*
0.37t 0.08
0.49*
-0.05 -0.29 -0.08 -0.39t -0.06 -0.14
-0.05 -0.22 0.01 -0.38t -0.33t -0.17
-0.03 -0.08 -0.47* -0.24 -0.22 -0.38t
0.05 -0.24 -0.31 -0.14 0.03 -0.10
AVLT, Auditory-Verbal Learning Test; SDMT, Symbol Digit Modalities Test.
'With the exception of Trails B, cognitive change scores were calculated as the postoperative score minus the preoperative score. All resulting positive scores reflect an improvement. tp < 0.05. :j:p < om.
operation, these positive changes largely did not correlate significantly with surgical variables. Abstract thinking ability was most closely related to surgical outcome. None of the surgical variables, including PDI score, correlated significantly with more than two of the cognitive change scores. The cognitive change scores ofthose patients with the highest and lowest PDI quartile scores were also compared. On five of the six measures of cognitive outcome, those patients with the highest (and presumably malignant) PDI scores obtained scores superior to those with the lowest PDI scores, which was not what was expected. Discussion Neuropsychologic assessment of all patients undergoing cardiopulmonary bypass in this study showed that very few had any new significant cognitive impairments when tested at an average of 10 weeks after operation. Instead, patients in this study generally showed strong and significant improvements on measures of memory, new learning ability, attention, and psychomotor speed. Noticeable, but not significant, improvement on a measure of abstract thinking was also demonstrated. Intraindividual and group analysis suggested that the postoperative improvements made on these tests exceeded what would be expected from practice effects. It has been demonstrated that among normal persons, large changes between test and retest are not commonl' and that the differences generally do not approach statistical significance.I'' In addition, a review noted that older patients are less likely to show practice effects I 9 and that any changes are usually insignificant, I 5 Thus our sample
of relatively aged subjects (mean age, 59.9 years) appeared to have made improvements that cannot be accounted for by practice effects. Assignment to a control or experimental condition did not lead to a group difference in neuropsychologic or surgical outcome, although control-group patients tended to make greater positive gains on neuropsychologic tests. In part, this may be because the experimental-group patients had achieved significantly better preoperative scores than those in the control group. However, repeated measures multivariate analysis demonstrated a highly nonsignificant interaction effect for time by group, which possibly negated the importance of the preoperative differences. Our finding of improved neuropsychologic function at about 10 weeks after cardiopulmonary bypass contradicts that of others who noted that a sizable number of their patients had postoperative cognitive deficits,"? The mean age of our sample (59.9 years) was comparable to the mean ages (range, 41.4 to 59.0 years) reported in similar studies. I, 6 Another recent study that compared past and present cohorts of patients undergoing coronary artery bypass grafting noted that the current sample's age averaged 61.1 years.P It is possible that our exclusion of patients with preexisting complications positively affected outcome and may explain why our sample seemed to achieved outcomes superior to those in studies that did not exclude these patients. However, our study did not exclude patients on the basis of stage or severity of cardiac disease. It may be that the short bypass times (70 minutes on average) and the use of mean arterial pressures of 60 to 80 torr during bypass in our study contributed favorably to outcome.
Volume 104 Number 5 November 1992
Another probable difference between this study and others is that both groups of patients may have had their intraoperative cerebral perfusion pressure more closely monitored and regulated than is typical for this procedure. Although the method by which this pressure was monitored and regulated differed between our two groups under investigation, the nature of this study may have prompted the surgical team to more carefully attend and respond to this factor than is typically the case. This scrutiny may have protected against the development of irregular or interrupted cerebral blood flow to the brain during the operation. Although the patients in this sample generally did well after operation, it does not appear that the use of a quantitative electroencephalograph specifically contributed to this outcome. Patients whose cerebral perfusion pressure was monitored and adjusted on the basis of computerized EEG data did not obtain outcomes superior to those receiving standard monitoring. A recent study found that computerized EEG monitoring could lead to a significant decrease in new global neurologic deficits in patients having cardiopulmonary bypass procedures.F The failure of this same equipment to have a demonstrable effect in our study may have been because new neurologic deficits generally did not develop in our sample. However, the PDI scores of even those patients with the worst outcomes were generally not distinguishable from the PDI scores of other patients. Another possible explanation is that the present study made a more careful study of the patients' cognitive status through the use of standardized neuropsychologic tests. This refinement allowed for a more precise measurement of postoperative changes in cognition than do behavioral observations or ratings.
I.
2.
3.
4.
REFERENCES Sotaniemi KA, Juolasmaa A, Hokkanen E. Neuropsychologicoutcome after open-heart surgery. Arch Neuro11981; 38:2-8. Tufo H, Ostfeld A, Shekelle R. Central nervous system dysfunction following open-heart surgery. JAMA 1970; 212:1333-40. Aberg T, Ronquist G, Tyden H, etal. Adverse effectson the brain in cardiac operations as assessed by biochemical, psychometric, and radiologic methods. J THORAC CARDIOVASC SURG 1984;87:99-105. Juolasmaa A, Outakoski J, Hirvenoja R, Tienari P, Sotaniemi K, Takkunen J. Effect of open heart surgery on
Cognitive outcome after cardiac operations
5.
6.
7.
8.
9.
10.
11.
12.
13. 14. 15. 16. 17.
18.
19. 20.
1 409
intellectual performance. J Clin N europsychol 1981;3:18197. Savageau J, Stanton B, Jenkins D, Klein M. Neuropsychological dysfunction following elective cardiac operation. I. Early assessment. J THORAC CARDIOVASC SURG 1982; 84:585-94. Townes BD, Bashein G, Hornbein TF, et al. Neurobehavioral outcomes in cardiac operations. a prospective controlled study. J THORAC CARDIOVASC SURG 1989;98:77482. Shaw P, Bates D, Cartlidge N, et al. Long-term intellectual dysfunction following coronary artery bypass graft surgery: a six month follow-up study. Q J Med 1987;62:25968. Sundt T, Sharbrough F, Anderson R, Michenfelder J. Cerebral blood flow measurements during carotid endarterectomy. J Neurosurg 1974;41:310-20. Ingvar D, Sjolund B, Ardo A. Correlation between dominant EEG frequency, cerebral oxygen uptake and blood flow. Electroencephalogr Clin Neurophysiol 1976;41:26876. Collice M, Arena 0, Fontana R, Mola M, Galbiati N. Role of EEG monitoring and cross-clamping duration in carotid endarterectomy. J Neurosurg 1986;65:815-9. Levy W, Shapiro H, Maruchak G, Meathe E. Automated EEG processing for intraoperative monitoring: a comparison of techniques. Anesthesiology 1980;53:223-36. Arom KY, Cohen DE, Strobl FT. Effect of intraoperative intervention on neurological outcome based on electroencephalographic monitoring during cardiopulmonary bypass. Ann Thorac Surg 1989;48:476-83. Wechsler D. The Wechsler Memory Scale Manual. New York: Psychological Corporation, 1945. Rey A. L'examen clinique en psychologie. Paris: University of France, 1984. Lezak M. Neuropsychological assessment. 2nd ed. New York: Oxford University Press, 1983. Smith A. Symbol Digit Modalities Test. Los Angeles: Western Psychological Services, 1973. Heineman A, Harper R, Friedman L, Whitney J. The relative utility of the Shipley-Hartford Scale prediction of WAIS-R IQ. J Clin PsychoI1985;41:547-51. Matarrazzo JD, Carmody TP, Jacobs LD. Test-retest reliability and stability of the WAIS: a literature review with implications for clinical practice. J Clin PsychoI1980;2:89105. Shatz MW. WAIS practice effects in clinical neuropsychology. J Clin NeuropsychoI1981;3:171-9. Nauheim KS, Fiore AC, Wadley J. The changing profile of the patient undergoing CABG surgery. J Am Coli CardioI1988;11:494-8.
Christopher L. Grote, PhD,a Patrick T. Shanahan, MD,b Paul Salmon, PhD,c Robert G. Meyer, Phl),? Curt Barrett, PhD,d and Allan Lansing, MD, Phl)," Chicago, Ill., and Louisville, Ky.
Several investigations have shown that new neuropsychologic deficits develop in the immediate postoperative phase in a majority of patients undergoing cardiopulmonary bypass.L' Although many of these deficits are subtle and reverse themselves within a few weeks of operation, approximately 10% to 20% of patients undergoing bypass continue to show a significant decrement in performance on one or more measures of cognitive ability at a 6-month follow-up.v 6-7 Insufficient intraoperative cerebral blood flow, often a result of decreased cerebral perfusion pressure or cardiac From the Department of Psychology and Social Sciences; Rush-Presbyterian-St. Luke's Medical Center, Chicago, Ill.; Medical Center Anesthesiologists," Louisville, Ky.; the Department of Psychology," University of Louisville, Louisville, Ky.; the Department of Psychiatry," Nortons Hospital, Louisville, Ky.; and the Department of Surgery,' Humana Hospital Audubon, Louisville, Ky. Received for publication June
13,1991. 1992.
Accepted for publication Feb. 24,
Address for reprints: Christopher L. Grote, PhD, Department of Psychology and Social Sciences, Rush-Presbyterian-St. Luke's Medical Center, 1653 W. Congress Pwy., Chicago, IL 60612.
12/1/38383
embolization, has been identified as a primary cause of the development of postoperative neurologic and neuropsychologic sequelae.v 8 Although traditional electroencephalographic (EEG) equipment is sensitive to intraoperative cerebral blood f1ow,9-11 it has not been used often because of the complexity of the equipment and the difficulty of making quick clinical interpretations. I I At least one investigation has used a quantified or computerized electroencephalograph in an attempt to overcome these limitations and found that postoperative global neurologic deficits were less prevalent in patients who received intraoperative interventions on the basis of computerized EEG data.l- However, the way in which these deficits were measured and evaluated was not fully described. The present investigation further explored these issues. In addition to examining the effect of cardiopulmonary bypass on postoperative cognitive function, we investigated whether computerized EEG data correlate with postoperative surgical or neuropsychologic outcome. Another goal was to compare the effect of adjusting cerebral perfusion pressure on the basis of computerized EEG data with that of the more standard technique of adjusting on the basis of mean systemic pressure. 1405
I 406
The Journal of Thoracic and Cardiovascular Surgery
Grote et al.
Table I. Outcome of surgical variables Control group (n = 14)
Bypass length (min) Estimatedbloodloss (mI) Postop. hospital stay (dy) Mean PDI score*
Experimental group (n
= 15)
M
SD
M
SD
70.3 345.0
18.2 125.2 2.4 72.6
72.8 434.6 8.0
161.9
9.1 105.5
132.1
21.2 2.3 178.8
Control vs experimental
NS NS NS NS
M, Mean; SD, standard deviation; NS, not significant.
*The calculation of POI scores is referred to in the text.
Method Patients. Patients undergoing aorta-coronary bypass grafting at Humana Hospital Audubon in Louisville,Kentucky, were eligible for participation. Those with previous strokes, with existing conditions such as diabetes, or about to undergo procedures such as carotid endarterectomy were excluded. However, stage of cardiac disease was not an exclusionary criterion. With institutional review board· approval, eligible patients were approached on admission and were invited to participate in the study. Informed consent was obtained. Twenty-nine patients underwent cardiopulmonary bypass and completed both preoperative and postoperative neuropsychologic testing. The average age of the patients was 59.9 years, (standard deviation, 8.8), with a range of 40 to 82. Twenty-one of the patients were male and eight female. All were white. Patients were randomly assigned to the control (n = 14) or experimental (n = 15) groups. Intraoperative management. All patients received a preoperative intramuscular injection of morphine sulfate and scopolamine. Anesthesia was induced and maintained with sufentanil citrate. All patients had electrocardiogram, temperature (esophageal, rectal, and pulmonary artery), radial artery blood pressure, end-tidal carbon dioxide, Sa02 (pulse oximetry), pulmonary artery pressure, and central venous pressure monitoring. A computerized on-line 16-channel electroencephalograph (CNS monitor, CNS, Inc., Minneapolis, Minn.) was used to monitor all patients during the operation. After placement of the electrodes by the standard 10-20 system and induction of anesthesia, a baseline recording of the patient's EEG frequency and total power was made. Subsequent measurements of the intraoperative level of strength of electrical signal were then made and compared with these baseline recordings. The power drop index (PDI) measures the amount of time and the degree to which the intraoperative electrical activity is less than 40% of baseline, a level that has been associated with postoperative complications.l? Differences in management between the control and experimental groups occurred during the bypass period. Patients assigned to the control group were managed with mean systemic pressure used as a guide. Pressure was maintained between 60 and 80 torr during bypass by the addition of volume to the bypass circuit or with an infusion of phenylephrine. Systemic pressure in patients assigned to the experimental group was maintained between 60 and 80 torr if there was no change in the EEG signal below 40% of the baseline output. If the EEG signal declined below 400/0 of baseline, addition of volume, a phenylephrine infusion, or both methods were used to
raise perfusion pressure. In this effort to raise perfusion pressure, a cutoff pressure of 100 ± 5 torr was used. Perfusion was performed with a membrane oxygenator at a full range of 1.6 to 3.0 L/m2 per minute, with an average of 2.4 L/m 2 per minute. On-line arterial filters and prebypass filters were used on all patients. Moderate hypothermia to a core temperature of 28° C (esophageal-rectal) was used on all patients. Arterial blood gases during bypass were managed by alpha-stat pH methods. A decrease in EEG signal of 6% per degree centigrade was accepted as normal in experimental group patients for decreases in temperature. Decreases greater than 60% (40% of original signal remaining) were deemed to exceed the 6% per degree centigrade margin, and attempts were made to increase perfusion pressure in an effort to improve the EEG signal. The arbitrary upper limit cutoff of 100 torr mean arterial pressure was arrived at in consultation with the surgeons who believed that higher pressures might result in possible crossclamp failure and bleeding into the operative site. In all patients studied, no new asymmetric EEG patterns were observed. Several patients had mild hemispheric asymmetry before bypass, which was not thought to be evidence of previous focal injury. No new focal patterns were seen in any patients. Postoperative neurologic examination did not reveal any new deficits in any of the patients. Neuropsychologic tests. The neuropsychologic tests were administered by a trained psychometrician within 48 hours before operation and again at an average of 10 weeks after operation. The neuropsychologic examination consisted of the following measures: Wechsler Memory Scale.!' Rey AuditoryVerbal Learning Test,14 Trails B,15 Symbol Digit Modalities Test,16 and the Shipley Institute of Living Scale Abstraction Subtest.!" These tests were selected to provide measures of attention, verbal and nonverbal memory, psychomotor speed, and abstract thinking. Results Table I shows the outcome of surgical variables for each group. The groups were not different overall (p > 0.14) on these measures, and none of the univariate comparisons were significant. Table II shows the preoperative and postoperative neuropsychologic scores for both groups. The multivariate comparison of groups on the preoperative measures was not significant (p> 0.22), but the experimental group achieved significantly better scores on all but one
Volume 104 Number 5 November 1992
1 4 a7
Cognitive outcome after cardiac operations
Table II. Intergroup and intragroup comparison ofpreoperative and postoperative test results* Preoperative Control
Wechsler Memory Scale Rey AVLT
Trails B SDMT- Written SDMT-Oral Shipley Abstraction
Postoperative Control
Experimental
Mean
SD
Mean
SD
55.9 37.5 110.2 33.4 40.1 8.4
6.6 8.4 32.7 7.3 7.6 3.2
60.3 41.7 103.0 40.2 45.3 10.9
10.0 12.7 69.4 7.4 8.3 4.1
t NS
t
:t
t t
Experimental
Mean
SD
Mean
SD
61.7 43.8 98.3 38.6 44.7 10.3
7.9 10.0 51.4
63.1 43.7 100.2 43.4 49.7 10.8
10.7 10.1 80.8 11.3 9.5 4.8
9.5 9.5 3.7
Preop. vs Postop. Control
Experimental
:j: :j:
NS NS NS
NS NS NS
NS
NS
t
NS
NS
NS
t
NS NS
t
SD. Standard deviation; NS, not significant; AVLT, Auditory-Verbal Learing Test; SDMT, Symbol Digit Modalities Test.
'Raw scores are reported. With the exception of Trails B, higher scores are more desirable. tp < 0.05. :j:p < om.
of the preoperative measures, despite the use of random assignment to groups. The average age of the control group (61.4 years) was higher than that of the experimental group (58.2 years), but the difference did not approach statistical significance (p > 0.28). There were no other apparent preoperative differences between groups such as stage of cardiac illness,so it is not clear why one group outperformed the other on these measures. On postoperative testing, there were no significant differences between groups on any of the measures. Analysis of changes in test results after operation indicate that neither group had any kind of cognitive decline across time. Instead, there were several instances of significantly improved performance. The control group made large and significant gains on the Wechsler Memory Scale, the Rey Auditory-Verbal Learning Test, and on both subtests of the Symbol Digit Modalities Test. Patients in the experimental group made a significant gain on the oral subtest of the Symbol Digits Modalities Test. Repeated measures multivariate analysis showed that there were no significant differences between the groups (p> 0.88). There was a significant change for both groups across time (p < 0.00 1), but the interaction effect of group by time was not significant (p > 0.80). The lack of a significant interaction offers further evidence of the lack of a group effect and negates the importance of the preoperative differences between groups. Many previous studies have referred only to group data, obscuring individual results in the process. Because neitherof the groups in this study showed a decline on any of the postoperative tests, we decided to combine the groupsand to then examine intraindividual changes. The percentage of patients who had a significant change across evaluations was calculated to provide a perhaps moreaccurate reflection of how patients' cognitive scores
Table III. Percentage ofpatients (both groups combined, n = 29) showing significant changes across testings* Variable
Improvement
Decrement
Wechsler Memory Scale ReyAVLT Trails B Shipley Abstraction SDMT-Written SDMT-Oral
28 24 7 21 31 35
0 0 3 7 10 3
AVLT, Auditory-Verbal Learning Test; SDMT, Symbol Digit Modalities Test.
'Changes betweenevaluationsare consideredsignificant whena subject'schange score is greater than the preoperativegroup standard deviation for that test.
changed after operation. In the manner of Shaw and coworkers," a preoperative-postoperative change was classified as significant if the difference in score exceeded that of the preoperative standard deviation for that test. This analysis (Table III) showed that many more patients had a significant increase than had a decrease on these tests. For example, 28% of the sample had a significant postoperative gain on the Wechsler Memory Scale and none had a significant decrease. Similar patterns of change were seen on each of the other tests, with the exception of Trails B. Correlations between demographic, surgical, and cognitive variables were calculated (Table IV). The control and experimental groups were again combined for these analyses. There was a high degree of interrelatedness among the surgical variables, such as longer periods of bypass correlating with increased age, estimated blood loss,and length of postoperative stay. Similarly, high POI scores correlated with undesirable surgical outcome scores. Although cognitive scores generally improved after
14 0 8
The Journal of Thoracic and Cardiovascular Surgery
Grote et al.
Table IV. Correlations between surgical outcome and change in cognitive score* Surgical variables PDf
Length of bypass Estimated blood loss Days to discharge Age Cognitive Change Scores Wechsler Memory Scale Rey AVLT Trails B Shipley Abstraction SDMT-Written SDMT-Oral
Length of bypass
0.45t 0.65* O.44t 0.45t -0.16 -0.37 -0.22 -0.27 -0.17 -0.05
Estimated blood loss
Days to discharge
Age
0.35t
o.su 0.44*
0.37t 0.08
0.49*
-0.05 -0.29 -0.08 -0.39t -0.06 -0.14
-0.05 -0.22 0.01 -0.38t -0.33t -0.17
-0.03 -0.08 -0.47* -0.24 -0.22 -0.38t
0.05 -0.24 -0.31 -0.14 0.03 -0.10
AVLT, Auditory-Verbal Learning Test; SDMT, Symbol Digit Modalities Test.
'With the exception of Trails B, cognitive change scores were calculated as the postoperative score minus the preoperative score. All resulting positive scores reflect an improvement. tp < 0.05. :j:p < om.
operation, these positive changes largely did not correlate significantly with surgical variables. Abstract thinking ability was most closely related to surgical outcome. None of the surgical variables, including PDI score, correlated significantly with more than two of the cognitive change scores. The cognitive change scores ofthose patients with the highest and lowest PDI quartile scores were also compared. On five of the six measures of cognitive outcome, those patients with the highest (and presumably malignant) PDI scores obtained scores superior to those with the lowest PDI scores, which was not what was expected. Discussion Neuropsychologic assessment of all patients undergoing cardiopulmonary bypass in this study showed that very few had any new significant cognitive impairments when tested at an average of 10 weeks after operation. Instead, patients in this study generally showed strong and significant improvements on measures of memory, new learning ability, attention, and psychomotor speed. Noticeable, but not significant, improvement on a measure of abstract thinking was also demonstrated. Intraindividual and group analysis suggested that the postoperative improvements made on these tests exceeded what would be expected from practice effects. It has been demonstrated that among normal persons, large changes between test and retest are not commonl' and that the differences generally do not approach statistical significance.I'' In addition, a review noted that older patients are less likely to show practice effects I 9 and that any changes are usually insignificant, I 5 Thus our sample
of relatively aged subjects (mean age, 59.9 years) appeared to have made improvements that cannot be accounted for by practice effects. Assignment to a control or experimental condition did not lead to a group difference in neuropsychologic or surgical outcome, although control-group patients tended to make greater positive gains on neuropsychologic tests. In part, this may be because the experimental-group patients had achieved significantly better preoperative scores than those in the control group. However, repeated measures multivariate analysis demonstrated a highly nonsignificant interaction effect for time by group, which possibly negated the importance of the preoperative differences. Our finding of improved neuropsychologic function at about 10 weeks after cardiopulmonary bypass contradicts that of others who noted that a sizable number of their patients had postoperative cognitive deficits,"? The mean age of our sample (59.9 years) was comparable to the mean ages (range, 41.4 to 59.0 years) reported in similar studies. I, 6 Another recent study that compared past and present cohorts of patients undergoing coronary artery bypass grafting noted that the current sample's age averaged 61.1 years.P It is possible that our exclusion of patients with preexisting complications positively affected outcome and may explain why our sample seemed to achieved outcomes superior to those in studies that did not exclude these patients. However, our study did not exclude patients on the basis of stage or severity of cardiac disease. It may be that the short bypass times (70 minutes on average) and the use of mean arterial pressures of 60 to 80 torr during bypass in our study contributed favorably to outcome.
Volume 104 Number 5 November 1992
Another probable difference between this study and others is that both groups of patients may have had their intraoperative cerebral perfusion pressure more closely monitored and regulated than is typical for this procedure. Although the method by which this pressure was monitored and regulated differed between our two groups under investigation, the nature of this study may have prompted the surgical team to more carefully attend and respond to this factor than is typically the case. This scrutiny may have protected against the development of irregular or interrupted cerebral blood flow to the brain during the operation. Although the patients in this sample generally did well after operation, it does not appear that the use of a quantitative electroencephalograph specifically contributed to this outcome. Patients whose cerebral perfusion pressure was monitored and adjusted on the basis of computerized EEG data did not obtain outcomes superior to those receiving standard monitoring. A recent study found that computerized EEG monitoring could lead to a significant decrease in new global neurologic deficits in patients having cardiopulmonary bypass procedures.F The failure of this same equipment to have a demonstrable effect in our study may have been because new neurologic deficits generally did not develop in our sample. However, the PDI scores of even those patients with the worst outcomes were generally not distinguishable from the PDI scores of other patients. Another possible explanation is that the present study made a more careful study of the patients' cognitive status through the use of standardized neuropsychologic tests. This refinement allowed for a more precise measurement of postoperative changes in cognition than do behavioral observations or ratings.
I.
2.
3.
4.
REFERENCES Sotaniemi KA, Juolasmaa A, Hokkanen E. Neuropsychologicoutcome after open-heart surgery. Arch Neuro11981; 38:2-8. Tufo H, Ostfeld A, Shekelle R. Central nervous system dysfunction following open-heart surgery. JAMA 1970; 212:1333-40. Aberg T, Ronquist G, Tyden H, etal. Adverse effectson the brain in cardiac operations as assessed by biochemical, psychometric, and radiologic methods. J THORAC CARDIOVASC SURG 1984;87:99-105. Juolasmaa A, Outakoski J, Hirvenoja R, Tienari P, Sotaniemi K, Takkunen J. Effect of open heart surgery on
Cognitive outcome after cardiac operations
5.
6.
7.
8.
9.
10.
11.
12.
13. 14. 15. 16. 17.
18.
19. 20.
1 409
intellectual performance. J Clin N europsychol 1981;3:18197. Savageau J, Stanton B, Jenkins D, Klein M. Neuropsychological dysfunction following elective cardiac operation. I. Early assessment. J THORAC CARDIOVASC SURG 1982; 84:585-94. Townes BD, Bashein G, Hornbein TF, et al. Neurobehavioral outcomes in cardiac operations. a prospective controlled study. J THORAC CARDIOVASC SURG 1989;98:77482. Shaw P, Bates D, Cartlidge N, et al. Long-term intellectual dysfunction following coronary artery bypass graft surgery: a six month follow-up study. Q J Med 1987;62:25968. Sundt T, Sharbrough F, Anderson R, Michenfelder J. Cerebral blood flow measurements during carotid endarterectomy. J Neurosurg 1974;41:310-20. Ingvar D, Sjolund B, Ardo A. Correlation between dominant EEG frequency, cerebral oxygen uptake and blood flow. Electroencephalogr Clin Neurophysiol 1976;41:26876. Collice M, Arena 0, Fontana R, Mola M, Galbiati N. Role of EEG monitoring and cross-clamping duration in carotid endarterectomy. J Neurosurg 1986;65:815-9. Levy W, Shapiro H, Maruchak G, Meathe E. Automated EEG processing for intraoperative monitoring: a comparison of techniques. Anesthesiology 1980;53:223-36. Arom KY, Cohen DE, Strobl FT. Effect of intraoperative intervention on neurological outcome based on electroencephalographic monitoring during cardiopulmonary bypass. Ann Thorac Surg 1989;48:476-83. Wechsler D. The Wechsler Memory Scale Manual. New York: Psychological Corporation, 1945. Rey A. L'examen clinique en psychologie. Paris: University of France, 1984. Lezak M. Neuropsychological assessment. 2nd ed. New York: Oxford University Press, 1983. Smith A. Symbol Digit Modalities Test. Los Angeles: Western Psychological Services, 1973. Heineman A, Harper R, Friedman L, Whitney J. The relative utility of the Shipley-Hartford Scale prediction of WAIS-R IQ. J Clin PsychoI1985;41:547-51. Matarrazzo JD, Carmody TP, Jacobs LD. Test-retest reliability and stability of the WAIS: a literature review with implications for clinical practice. J Clin PsychoI1980;2:89105. Shatz MW. WAIS practice effects in clinical neuropsychology. J Clin NeuropsychoI1981;3:171-9. Nauheim KS, Fiore AC, Wadley J. The changing profile of the patient undergoing CABG surgery. J Am Coli CardioI1988;11:494-8.