- Email: [email protected]
Cerebral blood flow and cognitive dysfunction after coronary surgery
Cerebral Blood Flow and Cognitive Dysfunction After Coronary Surgery Hanne Abildstrom, MD, Peter Høgh, MD, PhD, Bjørn Sperling, MD, Jakob T. Moller, MD, DMSc, Stig Yndgaard, MD, and Lars S. Rasmussen, MD, PhD Departments of Anesthesia and Neurology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
Background. Postoperative cognitive dysfunction after cardiac surgery has been attributed both to embolic events and periods with reduced cerebral perfusion. We investigated whether cognitive dysfunction after coronary surgery is associated with changes in regional cerebral blood flow (CBF) using single photon emission computed tomography. Methods. Before surgery and at discharge, 15 coronary surgery patients were studied. Global and regional CBF were measured using a brain-dedicated single photon emission computed tomography scanner, and neuropsychological testing with seven subtests was performed. Postoperative cognitive dysfunction was defined as a Z score above 2. Normative single photon emission computed tomography data were available from 26 healthy age-matched controls.
Results. Preoperative global CBF was significantly lower in patients compared with controls (53.7 versus 46.1 mL/100 g/min, p ⴝ 0.006). After surgery, global CBF significantly decreased in the patient group (46.1 versus 38.6 mL/100 g/min, p ⴝ 0.0001). No significant differences were detected in regional CBF. Cognitive dysfunction was identified in 4 of the 15 patients (26.7%, 95% CI 7.8% to 55.1%). No correlation was found between the neuropsychological Z score and global or regional CBF. Conclusions. The significant decrease in CBF after coronary surgery was uniformly distributed and was not correlated to postoperative cognitive dysfunction.
A
between postoperative cognitive dysfunction after cardiac surgery and changes in regional cerebral blood flow measured with SPECT 1 week after surgery. To increase the sensitivity and specificity, we included control data for both SPECT and neuropsychological testing [6].
common complication after cardiac surgery is postoperative cognitive dysfunction (POCD) presenting as problems with memory, concentration, and learning. POCD has been attributed to the use of cardiopulmonary bypass (CPB). The pathogenic mechanisms are thought to depend on several factors; First, macroemboli and microemboli [1] entering the circulation from bypass circuit could cause cerebral damage. Second, periods with reduced cerebral perfusion perioperatively, especially during CPB, may cause ischemia predominantly in the watershed areas. No correlation until now has been found between changes in cerebral blood flow (CBF) and neuropsychological test performance [2– 4]. Studies measuring global CBF [2, 3] might have overlooked regional changes in CBF potentially associated with decline in neuropsychological test performance. A well-established method for the measurement of both regional and global CBF, single photon emission computed tomography (SPECT) [5], is available and is currently used as a diagnostic adjunct in patients with stroke and dementia. We hypothesized that there may be a correlation
Accepted for publication Dec 16, 2001. Address reprint requests to Dr Abildstrom, Department of Anesthesia, Section 4132, Center of Head and Orthopedics, Rigshospitalet, Copenhagen University Hospital, DK-2100 Copenhagen, Denmark; e-mail: [email protected].
© 2002 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
(Ann Thorac Surg 2002;73:1174 –9) © 2002 by The Society of Thoracic Surgeons
Material and Methods We included 20 patients aged 60 years and over scheduled for coronary artery surgery with the use of CPB. Approval from the Ethics Committee was obtained on February 10, 1998, and all patients gave written consent. Exclusion criteria were previous heart surgery, preoperative atrial fibrillation, or a daily use of tranquilizers or alcohol abuse. We also excluded patients with known disease of the central nervous system, and a minimum score of 24/30 points was required on the Mini Mental State Examination [7]. Data from two control groups were used. A group of 26 healthy age-matched controls underwent SPECT and served to define normal values for global and regional CBF. The controls were recruited through advertisement and had no known cardiovascular or cerebrovascular disease. Neuropsychological normative data from 176 healthy controls of similar age were obtained as the patients were available [8]. These controls had undergone neuropsychological testing three times with the same intervals as the patients, none were admitted to hospital in the study period. 0003-4975/02/$22.00 PII S0003-4975(01)03618-9
Ann Thorac Surg 2002;73:1174 –9
Single Photon Emission Computed Tomography (SPECT) The patients were scanned under identical conditions preoperatively and before discharge at the fifth to seventh postoperative day in a dark quiet room. The patients were awake and in stable physical condition at the postoperative examination (ie, all drainage and intravenous therapy had been terminated and patients were fully recovered and ready for discharge). Global cerebral blood flow (gCBF) was measured by inhalation of 133Xenon using a brain-dedicated SPECT scanner (Tomotomatic 564; Medimatic Inc, Copenhagen, Denmark). 133 Xenon was inhaled for 1.5 minutes from a 4-L bag filled with atmospheric air and oxygen with a 133Xenon concentration of 740 MBq/L. The 133Xenon distribution was recorded during uptake and the after 3 minutes of washout. After a filtered back-projection technique with a 32 ⫻ 32 matrix, absolute values of flow expressed as mL blood/100 g brain tissue/min were calculated from the slice corresponding to a level 5 cm above the orbitomeatal plane. As changes in hematocrit changes the partition-coefficient (lambda) for 133Xenon, postoperative gCBF values were corrected for individual changes in hematocrit for each patient according to Chen and associates [9]: lambda ⫽ 0.171/(0.094 ⫹ 0.176 ⫻ hematocrit). The regional cerebral blood flow (rCBF) was determined after the intravenous injection of a dose of 800 to 1,000 MBq Technetium-99 m-HMPAO (Amersham International, London, England). During the injection and the after 10 to 15 minutes, the patients had their eyes closed. The radioactivity distribution in the brain was measured, and a total of 27 consecutive slices parallel to the orbitomeatal plane were obtained. The 27 slices were recompressed to nine slices for regional analysis of the images. After normalization of CBF to mean blood flow in the cerebellum [5], semiquantitative values for rCBF were calculated in four cortical regions of interest: frontal, temporal, parietal, and occipital cortex. For each anatomical region, the value of rCBF in each individual patient both pre- and postoperatively was analyzed with reference to mean values in the control group to determine whether rCBF in a given region was abnormal (ie, more than 2 SD from the control group means in the same region).
Neuropsychological Testing Neuropsychological testing was performed the same day as the SPECT examination (ie, before the operation and postoperatively on the fifth to seventh day), and in addition, 3 months after surgery. The ISPOCD test battery (Table 1) was applied, which exists in three parallel versions that are given to the patients in random order. Normative data from 176 healthy age-matched controls from Manchester were available [8]. These controls were tested with the same test battery and with the same time intervals as patients, and their test results served as normal material. We used changes in test results between the first test session (baseline) and the postoperative test sessions among the 176 Manchester controls to
ABILDSTROM ET AL CBF AND COGNITIVE DYSFUNCTION
1175
Table 1. The Neuropsychological Test Battery ● Visual Verbal Learning Test [14]: cumulative number of words recalled in three trials and the number of words recalled at delayed trial. ● Concept Shifting Test (adapted from reference 15): the time and number of errors in part C. ● Stroop Colour Word Interference Test [16]: the time and number of errors in part 3. ● Letter-Digit Coding Task [17]: the number of correct answers.
calculate a mean and SD of differences in test performance. The mean changes were taken as estimated learning effects. These values were used to normalize the test results of the patients. The same definition of postoperative cognitive dysfunction as in the ISPOCD study [8], based on seven parameters from the results of four tests, was used (Table 1). For each patient a Z score was obtained at the two postoperative test sessions; baseline scores were subtracted from postoperative test results, and the learning effect calculated from the test results for the Manchester controls was subtracted from these changes. The results were divided by the SD in test results from the Manchester control group to obtain a Z score for each test:
Z⫽
⌬X ⫺ ⌬X (control) SD (control)
Similarly, a combined Z score was defined from the total of Z scores in the Manchester controls, the SD being used to normalize the patient’s combined Z score. A large positive Z score indicates a deterioration in cognitive function from baseline in patients compared with controls. Cognitive dysfunction is present when two Z scores in individual tests or the combined Z score were 1.96 or more (ie, the scores of the patient were more than 2 SD from the mean).
Anesthesia and Cardiopulmonary Bypass The patients received their ordinary antianginal medication plus diazepam 10 mg orally. In addition, morphine 10 mg and scopolamine 0.4 mg was given intramuscularly 1 hour before induction of anesthesia. After a radial artery catheter had been inserted, anesthesia was induced with midazolam 0.1 mg/kg and fentanyl 20 g/kg. Orotracheal intubation was facilitated by pancuronium or cis-atracurium 0.1 mg/kg. Anesthesia was maintained with isoflurane 0.4% to 1% and fentanyle bolus doses of 5 g/kg. Hemodynamic stability with heart rate 45 to 90 beats/min and mean arterial pressure between 60 and 80 mm Hg was maintained by adjusting anesthetic depth, adjusting nitroglycerin infusion, giving intravenous ephedrine bolus doses, and changing volume load. Before connecting the extracorporal bypass machine, the patient was heparinized with heparin 300 U/kg and supplemented if needed to secure an activated clotting time above 480 seconds. The cardiopulmonary bypass circuit was primed with 1,800 mL Ringer-lactate and
1176
ABILDSTROM ET AL CBF AND COGNITIVE DYSFUNCTION
Ann Thorac Surg 2002;73:1174 –9
Table 2. Comparison of Global CBF (mL/100 g/min) in Controls (n ⫽26) and Patients (n ⫽ 15) Preoperatively and Postoperatively
pCO2 (%) Hemoglobin (mmol/L) Left hemisphere Right hemisphere a
Controls Mean (SD)
p Valuea
5.03 (0.41)
0.07
53.8 (8.7) 53.7 (9.4)
0.004 0.006
Preoperatively Mean (SD)
p valueb
Postoperatively Mean (SD)
4.77 (0.36) 8.7 (0.73) 46.1 (5.0) 46.1 (4.6)
0.05 0.0001 0.0001 0.0001
4.47 (0.39) 6.7 (0.71) 38.6 (3.8)c 38.5 (3.7)c
Global CBF in controls and patients preoperatively compared with Student’s t test.
b
Pre- and postoperative global CBF in patients compared with Student’s t test for paired data.
c
Hemoglobin corrected.
CBF ⫽ cerebral blood flow.
10,000 U of heparin. Cannulation was performed in aorta ascendens and a two-stage cannula in vena cava inferior. Hypothermic (32°C), nonpulsatile cardiopulmonary bypass with a membrane oxygenator was performed using ␣-stat technique. During extracorporal circulation, the activated clotting time was kept above 480 seconds with heparin bolus doses of 5,000 U as needed. Pump flow above 2.4 L/min/m2 was maintained at all temperatures. Mean perfusion pressure below 50 mm Hg was treated with bolus doses of phenylepinephrine or increasing pump flow. Perfusion pressure above 80 mm Hg was treated with sodium-nitroprusside infusion and by increasing the anesthetic depth. All peripheral anastomoses were performed with aortic cross-clamping and antegrade crystalloid cardioplegia (St. Thomas II) 10 mL/kg. Cardioplegia 4 mL/kg was supplemented every 20 minutes. The proximal anastomoses were performed with a partially occluding side clamp on beating heart during rewarming. CPB was weaned at a rectal temperature of 36°C. After termination of CPB, cardiac index was kept above 2.4 L/m2/min, mean arterial pressure between 60 and 90 mm Hg, and left atrial pressure below 12 mm Hg. This was achieved by volume therapy, atrial or ventricular pacing, inotropics, and vasodilatation. After surgery, the patients were transferred to the intensive care unit and trachea was extubated after 4 to 12 hours.
Statistical Analysis Data are reported as means with range and proportions with 95% confidence intervals. The sample size calculation was based on the assumption that half of the patients would have POCD enabling us to compare two equally sized groups. If the difference in rCBF in a given region between the groups would correspond to a standardized difference of 2, then 15 patients would suffice accepting a type 1 error of 5% and a type 2 error of 20%. Preoperative values of gCBF and rCBF for the patients were compared with those of the controls using Student’s t test. Pre- and postoperative values of gCBF and rCBF for the patients were compared with Student’s t test for paired data. The difference in gCBF and rCBF in the given regions between pre- and postoperative values for the individual patient was calculated. Using Spearman’s
rank correlation test, these parameters were correlated with the Z score at the first postoperative test session.
Results A total of 20 male patients with an age of 66 (range 60 to 74) years and weight of 80.5 (range 60 to 102) kg were included. Their preoperative left ventricular ejection fraction was 57% (45% to 65%) and the duration of CPB was 101 (range 42 to 167) minutes. Twenty-six controls (14 men/12 women) with an age of 66 (range 51 to 79) years were SPECT scanned. Due to a computer error, SPECT data were lost for 1 patient, whereas 4 patients did not complete the postoperative SPECT examination: 2 patients due to respiratory problems, 1 patient refused, and 1 patient suffered a cardiac arrest and died 3 weeks after surgery. The following results are based on the 15 patients with complete data.
Neuropsychological Test Results The 15 patients were assessed with the neuropsychological test battery and scanned 13.5 (range 4 to 77) days before surgery and postoperatively after 5 (range 4 to 8) days. After 106 (range 58 to 153) days, neuropsychological follow-up was performed. At the first postoperative test, 4 of 15 patients had POCD (ie, an incidence of 26.7% [7.8% to 55.1%]), and 3 months after surgery, 3 out of 15 patients had POCD (ie, 20.0% [4.3 to 48.1%]).
SPECT Results Normal values for gCBF (Table 2) were defined from SPECT results in 26 elderly healthy volunteers as mean ⫾ 2 SD. Preoperative gCBF in the patients were significantly lower than in the control group. In the patients, the level of CO2 and hemoglobin in blood was significantly lower at the postoperative SPECT. After correcting for the lower concentration of hemoglobin [9], there was a significant decline in postoperative gCBF. Normal values for rCBF were also defined from the SPECT results in the 26 healthy controls as mean ⫾ 2 SD. There were no significant differences between rCBF values in patients and controls in any of the given regions, nor was there any significant change in pre- and postoperative rCBF for all patients. When dividing patients into
Ann Thorac Surg 2002;73:1174 –9
ABILDSTROM ET AL CBF AND COGNITIVE DYSFUNCTION
Table 3. Patients With (n ⫽ 4) and Without (n ⫽ 11) Postoperative Cognitive Dysfunction (POCD) at Discharge Patients With POCD Mean (SD) Regional CBF Hemisphere Left Right Regional CBF Frontal Cortex Left Right Regional CBF Temporal Cortex Left Right Regional CBF Parietal Cortex Left Right Regional CBF Occipital Cortex Left Right Global CBF Left Right
Patients Without POCD Mean (SD)
p Valuea
⫺5.25 (6.55) ⫺4.75 (5.68)
3.09 (9.67) 2.64 (10.30)
0.19 0.17
⫺6.75 (6.29) ⫺6.50 (6.58)
1.82 (9.30) 1.45 (8.58)
0.15 0.15
⫺4.50 (10.66) ⫺3.50 (7.85)
3.55 (9.88) 2.82 (10.30)
0.29 0.32
⫺6.00 (3.83) ⫺5.25 (3.30)
1.73 (10.74) 1.73 (10.49)
0.24 0.19
⫺6.00 (8.37) ⫺4.25 (7.72)
3.55 (11.98) 3.91 (13.03)
0.14 0.24
⫺7.01 (2.77) ⫺7.30 (3.51)
⫺7.72 (4.80) ⫺7.73 (4.81)
0.47 1.00
Data are change in Regional CBF (in % of mean flow in cerebellum) and Global CBF (mL/100 g/min) a
Wilcoxon Rank Sum Test
CBF ⫽ cerebral blood flow.
two groups whether they had POCD or not, a trend was found. A decrease in rCBF was found in all regions in patients with POCD, and in contrast to this, rCBF increased in all regions in patients without POCD; however, this difference was not significant (Table 3). There was no significant correlation between the neuropsychological Z score at the first postoperative test session and changes in rCBF in any region (p ⬎ 0.40), or the change in gCBF (p ⬎ 0.16).
Comment We found a significant postoperative decrease in gCBF, but these changes did not correlate significantly with neuropsychological test results. There was a nonsignificant trend towards a lower rCBF in patients with POCD. Preoperatively, the patients had significantly reduced gCBF compared with healthy age-matched controls. The incidence of POCD is in accordance with other studies of coronary surgery patients [10, 11]. Smith and associates [2] found a decrease in gCBF in some patients 8 days after cardiac surgery. This change did not correlate with neuropsychological test results. The decrease in gCBF could be related to neuron loss perioperatively as a consequence of either microemboli distributed diffusely
1177
in the cerebral circulation, or irreversible damage caused by cerebral ischemia. However, an additional explanation could be low cerebral metabolism caused by residual effects of sedatives given during hospitalization, or a lower level of psychological stress postoperatively. Hall and associates [4] studied the possible correlation between changes in rCBF during cardiopulmonary bypass using SPECT and neuropsychological decline in cardiac surgery patients. In that study, abnormality in perfusion for a given region was arbitrarily defined as less than 80% of the flow in cerebellum, and cerebral perfusion of the individual patient was defined as having deteriorated if the number of abnormal anatomical regions had increased. According to this definition, 43% of the patients had decreased cerebral perfusion during CPB, but this decrease did not correlate with decline in neuropsychological test performance. In our study, many of the controls had rCBF values less than 80% of the mean flow in cerebellum. If we were to use the same definition of perfusion abnormality as in Hall and associates [4], the majority of the rCBF values for the patients would be abnormal both pre- and postoperatively. Patients scheduled for coronary surgery had significantly lower gCBF than controls. Arteriosclerosis was an exclusion criterion for controls; none of the patients had previous stroke and there was no description of clinically important carotid stenoses, but the patients may have had general arteriosclerosis. Significantly reduced gCBF in nondemented patients with generalized arteriosclerosis has not previously been described. Neuropsychological testing is associated with considerable variability and the definition of POCD varies [6, 12]. Our definition of POCD is more restrictive than in most other studies, and few patients had POCD. But as we used a continuous variable for neuropsychological test performance, the Z score, in the correlation analysis, this did not decrease the ability to find a statistically significant correlation. The SPECT images were parametrically analyzed using standardized region templates defined according to a neuroanatomical atlas. Preferably, structural imaging should have been performed, which enables adjustments for variable neuroanatomy and potential structural lesions. The procedure for normalization of the cortical rCBF values may also be questioned, as we had no imaging data to exclude cerebellar pathology. However, none of the patients had a clinical history or clinical presentation indicating previous lesions or disorders of the central nervous system. The level of CO2 was significantly lower at the postoperative SPECT examination (Table 2). End-expiratory values were measured. If a mismatch between perfusion and ventilation had occurred after surgery, it is possible that these values are not truly reflecting pCO2 in blood. No correction was made for the change in level of pCO2, as the determining factor is pH. As no sudden changes in pCO2 occurred, it is most likely that the individual patient acid-base status was fully adapted. Hemoglobin concentration had declined significantly in patients at
1178
ABILDSTROM ET AL CBF AND COGNITIVE DYSFUNCTION
discharge (Table 2) and we corrected gCBF according to Chen and associates [9]. The average decrease in rCBF in patients with POCD was approximately 7%, in contrast to patients without POCD, who had an increase in rCBF of approximately 2% (Table 3); this difference was not statistically significant. The SD was approximately 10%, and our sample size did not allow us to detect such a small difference similar to the level of interexamination variability in healthy subjects, which may be as high as 10% [13]. Global CBF values in patients before coronary artery surgery were significantly lower than for controls. After surgery, we registered a significant decrease in global CBF, but this did not correlate with neuropsychological test results. In general, patients with cognitive deficits had a decrease in regional CBF, in contrast to patients without cognitive deficits. However, this nonsignificant difference between groups corresponded only to the expected interexamination variability.
Ann Thorac Surg 2002;73:1174 –9
5.
6. 7. 8.
9. 10.
11. The study was supported by grants from the European Commission’s BIOMED2 program (BMH4 –98 –3335), the Danish Medical Research Council, and the Danish Heart Foundation. We thank Gerda Thomsen and Glenna Skouboe for technical assistance.
12. 13.
References 1. Blauth C, Arnold J, Kohner EM, Taylor KM. Retinal microembolism during cardiopulmonary bypass demonstrated by fluorescein angiography. Lancet 1986;2:837–9. 2. Smith PL, Treasure T, Newman SP, et al. Cerebral consequences of cardiopulmonary bypass. Lancet 1986;1:823–5. 3. Venn GE, Patel RL, Chambers DJ. Cardiopulmonary bypass: perioperative cerebral blood flow and postoperative cognitive deficit. Ann Thorac Surg 1995;59:1331–5. 4. Hall RA, Fordyce DJ, Lee ME, et al. Brain SPECT imaging, and neuropsychological testing in coronary artery bypass
14. 15. 16. 17.
patients. Single photon emission computed tomography. Ann Thorac Surg 1999;68:2082– 8. Waldemar G. Functional brain imaging with SPECT in normal aging and dementia. Methodological, pathophysiological, and diagnostic aspects. Cerebrovasc Brain Metab Rev 1995;7:89 –130. Rasmussen LS, Larsen K, Houx P, Skovgaard LT, Hanning CD, Moller JT. The assessment of postoperative cognitive function. Acta Anaesthesiol Scand 2001;45:275– 89. Folstein MF, Folstein SE, McHugh PR. Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189–98. Moller JT, Cluitmans P, Rasmussen LS, et al. Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study. ISPOCD investigators. International Study of PostOperative Cognitive Dysfunction. Lancet 1998;351:857– 61. Chen RY, Fan FC, Kim S, Jan KM, Usami S, Chien S. Tissue-blood partition coefficient for xenon: temperature and hematocrit dependence. J Appl Physiol 1980;49:178– 83. Savageau JA, Stanton BA, Jenkins CD, Klein MD. Neuropsychological dysfunction following elective cardiac operation. I. Early assessment. J Thorac Cardiovasc Surg 1982;84: 585–94. Shaw PJ, Bates D, Cartlidge NE, et al. Early intellectual dysfunction following coronary bypass surgery. Q J Med 1986;58:59– 68. Mahanna EP, Blumenthal JA, White WD, et al. Defining neuropsychological dysfunction after coronary artery bypass grafting. Ann Thorac Surg 1996;61:1342–7. Shirahata N, Henriksen L, Vorstrup S, et al. Regional cerebral blood flow assessed by 133Xe inhalation and emission tomography: normal values. J Comput Assist Tomogr 1985;9: 861– 6. Brand N, Jolles J. Learning and retrieval rate of words presented auditorily and visually. J Gen Psychol 1985;112: 201–10. Reitan RM. Validity of the trail making test as an indicator of organic brain damage. Percept Mot Skills 1958;8:271– 6. Houx PJ, Jolles J, Vreeling FW. Stroop interference: aging effects assessed with the Stroop Color-Word Test. Exp Aging Res 1993;19:209–24. Lezak MD. Neuropsychological assessment, 3rd ed. New York: Oxford University Press; 1995:376–9.
INVITED COMMENTARY I commend the authors for obtaining control groups for this study. The authors choose healthy age matched controls for single photon emission computed tomography (SPECT) scanning, and this is certainly reasonable. We found that at any age, better-educated subjects tended to have better baseline cerebral perfusion as measured by SPECT. I wonder if the test subjects and control group individuals were similar in terms of education level, activity level, and measures of independence, all of which influence global and regional cerebral blood flow (CBF). Regardless, the finding that patients in need of coronary revascularization have diminished cerebral perfusion as compared to healthy volunteers is probably accurate and implies these patients are at risk for cerebral injury. We believe that SPECT may identify a subset of patients with “diminished cortical reserve” and accordingly are vulnerable to cognitive injury from any insult, surgery included. The authors found cerebral blood flow to be signifi© 2002 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
cantly reduced postoperatively as compared to preoperatively. What to account for this? Emboli, ischemic injury, cerebral edema or perhaps altered test environments. Carbon dioxide was lower at time of postoperative SPECT. Was respiratory rate comparable? Were patients compensating for smaller tidal volumes with a relative tachypnea? Similarly, hemodilution with a lower hematocrit postoperatively affects CBF and while a correction factor was applied to global CBF, none exists for regional blood flow evaluations. With significant differences in hematocrit and partial pressure of carbon dioxide before and after surgery, it is difficult to say that cerebral perfusion has indeed changed relative to baseline. It would be fascinating to restudy these patients months down the road, presumably when hematocrit and carbon dioxide levels had normalized, to see if these changes in cerebral perfusion persisted and as this is an important finding, the authors should be encouraged to consider a third SPECT scan in these patients. 0003-4975/02/$22.00 PII S0003-4975(02)03455-0
Background. Postoperative cognitive dysfunction after cardiac surgery has been attributed both to embolic events and periods with reduced cerebral perfusion. We investigated whether cognitive dysfunction after coronary surgery is associated with changes in regional cerebral blood flow (CBF) using single photon emission computed tomography. Methods. Before surgery and at discharge, 15 coronary surgery patients were studied. Global and regional CBF were measured using a brain-dedicated single photon emission computed tomography scanner, and neuropsychological testing with seven subtests was performed. Postoperative cognitive dysfunction was defined as a Z score above 2. Normative single photon emission computed tomography data were available from 26 healthy age-matched controls.
Results. Preoperative global CBF was significantly lower in patients compared with controls (53.7 versus 46.1 mL/100 g/min, p ⴝ 0.006). After surgery, global CBF significantly decreased in the patient group (46.1 versus 38.6 mL/100 g/min, p ⴝ 0.0001). No significant differences were detected in regional CBF. Cognitive dysfunction was identified in 4 of the 15 patients (26.7%, 95% CI 7.8% to 55.1%). No correlation was found between the neuropsychological Z score and global or regional CBF. Conclusions. The significant decrease in CBF after coronary surgery was uniformly distributed and was not correlated to postoperative cognitive dysfunction.
A
between postoperative cognitive dysfunction after cardiac surgery and changes in regional cerebral blood flow measured with SPECT 1 week after surgery. To increase the sensitivity and specificity, we included control data for both SPECT and neuropsychological testing [6].
common complication after cardiac surgery is postoperative cognitive dysfunction (POCD) presenting as problems with memory, concentration, and learning. POCD has been attributed to the use of cardiopulmonary bypass (CPB). The pathogenic mechanisms are thought to depend on several factors; First, macroemboli and microemboli [1] entering the circulation from bypass circuit could cause cerebral damage. Second, periods with reduced cerebral perfusion perioperatively, especially during CPB, may cause ischemia predominantly in the watershed areas. No correlation until now has been found between changes in cerebral blood flow (CBF) and neuropsychological test performance [2– 4]. Studies measuring global CBF [2, 3] might have overlooked regional changes in CBF potentially associated with decline in neuropsychological test performance. A well-established method for the measurement of both regional and global CBF, single photon emission computed tomography (SPECT) [5], is available and is currently used as a diagnostic adjunct in patients with stroke and dementia. We hypothesized that there may be a correlation
Accepted for publication Dec 16, 2001. Address reprint requests to Dr Abildstrom, Department of Anesthesia, Section 4132, Center of Head and Orthopedics, Rigshospitalet, Copenhagen University Hospital, DK-2100 Copenhagen, Denmark; e-mail: [email protected].
© 2002 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
(Ann Thorac Surg 2002;73:1174 –9) © 2002 by The Society of Thoracic Surgeons
Material and Methods We included 20 patients aged 60 years and over scheduled for coronary artery surgery with the use of CPB. Approval from the Ethics Committee was obtained on February 10, 1998, and all patients gave written consent. Exclusion criteria were previous heart surgery, preoperative atrial fibrillation, or a daily use of tranquilizers or alcohol abuse. We also excluded patients with known disease of the central nervous system, and a minimum score of 24/30 points was required on the Mini Mental State Examination [7]. Data from two control groups were used. A group of 26 healthy age-matched controls underwent SPECT and served to define normal values for global and regional CBF. The controls were recruited through advertisement and had no known cardiovascular or cerebrovascular disease. Neuropsychological normative data from 176 healthy controls of similar age were obtained as the patients were available [8]. These controls had undergone neuropsychological testing three times with the same intervals as the patients, none were admitted to hospital in the study period. 0003-4975/02/$22.00 PII S0003-4975(01)03618-9
Ann Thorac Surg 2002;73:1174 –9
Single Photon Emission Computed Tomography (SPECT) The patients were scanned under identical conditions preoperatively and before discharge at the fifth to seventh postoperative day in a dark quiet room. The patients were awake and in stable physical condition at the postoperative examination (ie, all drainage and intravenous therapy had been terminated and patients were fully recovered and ready for discharge). Global cerebral blood flow (gCBF) was measured by inhalation of 133Xenon using a brain-dedicated SPECT scanner (Tomotomatic 564; Medimatic Inc, Copenhagen, Denmark). 133 Xenon was inhaled for 1.5 minutes from a 4-L bag filled with atmospheric air and oxygen with a 133Xenon concentration of 740 MBq/L. The 133Xenon distribution was recorded during uptake and the after 3 minutes of washout. After a filtered back-projection technique with a 32 ⫻ 32 matrix, absolute values of flow expressed as mL blood/100 g brain tissue/min were calculated from the slice corresponding to a level 5 cm above the orbitomeatal plane. As changes in hematocrit changes the partition-coefficient (lambda) for 133Xenon, postoperative gCBF values were corrected for individual changes in hematocrit for each patient according to Chen and associates [9]: lambda ⫽ 0.171/(0.094 ⫹ 0.176 ⫻ hematocrit). The regional cerebral blood flow (rCBF) was determined after the intravenous injection of a dose of 800 to 1,000 MBq Technetium-99 m-HMPAO (Amersham International, London, England). During the injection and the after 10 to 15 minutes, the patients had their eyes closed. The radioactivity distribution in the brain was measured, and a total of 27 consecutive slices parallel to the orbitomeatal plane were obtained. The 27 slices were recompressed to nine slices for regional analysis of the images. After normalization of CBF to mean blood flow in the cerebellum [5], semiquantitative values for rCBF were calculated in four cortical regions of interest: frontal, temporal, parietal, and occipital cortex. For each anatomical region, the value of rCBF in each individual patient both pre- and postoperatively was analyzed with reference to mean values in the control group to determine whether rCBF in a given region was abnormal (ie, more than 2 SD from the control group means in the same region).
Neuropsychological Testing Neuropsychological testing was performed the same day as the SPECT examination (ie, before the operation and postoperatively on the fifth to seventh day), and in addition, 3 months after surgery. The ISPOCD test battery (Table 1) was applied, which exists in three parallel versions that are given to the patients in random order. Normative data from 176 healthy age-matched controls from Manchester were available [8]. These controls were tested with the same test battery and with the same time intervals as patients, and their test results served as normal material. We used changes in test results between the first test session (baseline) and the postoperative test sessions among the 176 Manchester controls to
ABILDSTROM ET AL CBF AND COGNITIVE DYSFUNCTION
1175
Table 1. The Neuropsychological Test Battery ● Visual Verbal Learning Test [14]: cumulative number of words recalled in three trials and the number of words recalled at delayed trial. ● Concept Shifting Test (adapted from reference 15): the time and number of errors in part C. ● Stroop Colour Word Interference Test [16]: the time and number of errors in part 3. ● Letter-Digit Coding Task [17]: the number of correct answers.
calculate a mean and SD of differences in test performance. The mean changes were taken as estimated learning effects. These values were used to normalize the test results of the patients. The same definition of postoperative cognitive dysfunction as in the ISPOCD study [8], based on seven parameters from the results of four tests, was used (Table 1). For each patient a Z score was obtained at the two postoperative test sessions; baseline scores were subtracted from postoperative test results, and the learning effect calculated from the test results for the Manchester controls was subtracted from these changes. The results were divided by the SD in test results from the Manchester control group to obtain a Z score for each test:
Z⫽
⌬X ⫺ ⌬X (control) SD (control)
Similarly, a combined Z score was defined from the total of Z scores in the Manchester controls, the SD being used to normalize the patient’s combined Z score. A large positive Z score indicates a deterioration in cognitive function from baseline in patients compared with controls. Cognitive dysfunction is present when two Z scores in individual tests or the combined Z score were 1.96 or more (ie, the scores of the patient were more than 2 SD from the mean).
Anesthesia and Cardiopulmonary Bypass The patients received their ordinary antianginal medication plus diazepam 10 mg orally. In addition, morphine 10 mg and scopolamine 0.4 mg was given intramuscularly 1 hour before induction of anesthesia. After a radial artery catheter had been inserted, anesthesia was induced with midazolam 0.1 mg/kg and fentanyl 20 g/kg. Orotracheal intubation was facilitated by pancuronium or cis-atracurium 0.1 mg/kg. Anesthesia was maintained with isoflurane 0.4% to 1% and fentanyle bolus doses of 5 g/kg. Hemodynamic stability with heart rate 45 to 90 beats/min and mean arterial pressure between 60 and 80 mm Hg was maintained by adjusting anesthetic depth, adjusting nitroglycerin infusion, giving intravenous ephedrine bolus doses, and changing volume load. Before connecting the extracorporal bypass machine, the patient was heparinized with heparin 300 U/kg and supplemented if needed to secure an activated clotting time above 480 seconds. The cardiopulmonary bypass circuit was primed with 1,800 mL Ringer-lactate and
1176
ABILDSTROM ET AL CBF AND COGNITIVE DYSFUNCTION
Ann Thorac Surg 2002;73:1174 –9
Table 2. Comparison of Global CBF (mL/100 g/min) in Controls (n ⫽26) and Patients (n ⫽ 15) Preoperatively and Postoperatively
pCO2 (%) Hemoglobin (mmol/L) Left hemisphere Right hemisphere a
Controls Mean (SD)
p Valuea
5.03 (0.41)
0.07
53.8 (8.7) 53.7 (9.4)
0.004 0.006
Preoperatively Mean (SD)
p valueb
Postoperatively Mean (SD)
4.77 (0.36) 8.7 (0.73) 46.1 (5.0) 46.1 (4.6)
0.05 0.0001 0.0001 0.0001
4.47 (0.39) 6.7 (0.71) 38.6 (3.8)c 38.5 (3.7)c
Global CBF in controls and patients preoperatively compared with Student’s t test.
b
Pre- and postoperative global CBF in patients compared with Student’s t test for paired data.
c
Hemoglobin corrected.
CBF ⫽ cerebral blood flow.
10,000 U of heparin. Cannulation was performed in aorta ascendens and a two-stage cannula in vena cava inferior. Hypothermic (32°C), nonpulsatile cardiopulmonary bypass with a membrane oxygenator was performed using ␣-stat technique. During extracorporal circulation, the activated clotting time was kept above 480 seconds with heparin bolus doses of 5,000 U as needed. Pump flow above 2.4 L/min/m2 was maintained at all temperatures. Mean perfusion pressure below 50 mm Hg was treated with bolus doses of phenylepinephrine or increasing pump flow. Perfusion pressure above 80 mm Hg was treated with sodium-nitroprusside infusion and by increasing the anesthetic depth. All peripheral anastomoses were performed with aortic cross-clamping and antegrade crystalloid cardioplegia (St. Thomas II) 10 mL/kg. Cardioplegia 4 mL/kg was supplemented every 20 minutes. The proximal anastomoses were performed with a partially occluding side clamp on beating heart during rewarming. CPB was weaned at a rectal temperature of 36°C. After termination of CPB, cardiac index was kept above 2.4 L/m2/min, mean arterial pressure between 60 and 90 mm Hg, and left atrial pressure below 12 mm Hg. This was achieved by volume therapy, atrial or ventricular pacing, inotropics, and vasodilatation. After surgery, the patients were transferred to the intensive care unit and trachea was extubated after 4 to 12 hours.
Statistical Analysis Data are reported as means with range and proportions with 95% confidence intervals. The sample size calculation was based on the assumption that half of the patients would have POCD enabling us to compare two equally sized groups. If the difference in rCBF in a given region between the groups would correspond to a standardized difference of 2, then 15 patients would suffice accepting a type 1 error of 5% and a type 2 error of 20%. Preoperative values of gCBF and rCBF for the patients were compared with those of the controls using Student’s t test. Pre- and postoperative values of gCBF and rCBF for the patients were compared with Student’s t test for paired data. The difference in gCBF and rCBF in the given regions between pre- and postoperative values for the individual patient was calculated. Using Spearman’s
rank correlation test, these parameters were correlated with the Z score at the first postoperative test session.
Results A total of 20 male patients with an age of 66 (range 60 to 74) years and weight of 80.5 (range 60 to 102) kg were included. Their preoperative left ventricular ejection fraction was 57% (45% to 65%) and the duration of CPB was 101 (range 42 to 167) minutes. Twenty-six controls (14 men/12 women) with an age of 66 (range 51 to 79) years were SPECT scanned. Due to a computer error, SPECT data were lost for 1 patient, whereas 4 patients did not complete the postoperative SPECT examination: 2 patients due to respiratory problems, 1 patient refused, and 1 patient suffered a cardiac arrest and died 3 weeks after surgery. The following results are based on the 15 patients with complete data.
Neuropsychological Test Results The 15 patients were assessed with the neuropsychological test battery and scanned 13.5 (range 4 to 77) days before surgery and postoperatively after 5 (range 4 to 8) days. After 106 (range 58 to 153) days, neuropsychological follow-up was performed. At the first postoperative test, 4 of 15 patients had POCD (ie, an incidence of 26.7% [7.8% to 55.1%]), and 3 months after surgery, 3 out of 15 patients had POCD (ie, 20.0% [4.3 to 48.1%]).
SPECT Results Normal values for gCBF (Table 2) were defined from SPECT results in 26 elderly healthy volunteers as mean ⫾ 2 SD. Preoperative gCBF in the patients were significantly lower than in the control group. In the patients, the level of CO2 and hemoglobin in blood was significantly lower at the postoperative SPECT. After correcting for the lower concentration of hemoglobin [9], there was a significant decline in postoperative gCBF. Normal values for rCBF were also defined from the SPECT results in the 26 healthy controls as mean ⫾ 2 SD. There were no significant differences between rCBF values in patients and controls in any of the given regions, nor was there any significant change in pre- and postoperative rCBF for all patients. When dividing patients into
Ann Thorac Surg 2002;73:1174 –9
ABILDSTROM ET AL CBF AND COGNITIVE DYSFUNCTION
Table 3. Patients With (n ⫽ 4) and Without (n ⫽ 11) Postoperative Cognitive Dysfunction (POCD) at Discharge Patients With POCD Mean (SD) Regional CBF Hemisphere Left Right Regional CBF Frontal Cortex Left Right Regional CBF Temporal Cortex Left Right Regional CBF Parietal Cortex Left Right Regional CBF Occipital Cortex Left Right Global CBF Left Right
Patients Without POCD Mean (SD)
p Valuea
⫺5.25 (6.55) ⫺4.75 (5.68)
3.09 (9.67) 2.64 (10.30)
0.19 0.17
⫺6.75 (6.29) ⫺6.50 (6.58)
1.82 (9.30) 1.45 (8.58)
0.15 0.15
⫺4.50 (10.66) ⫺3.50 (7.85)
3.55 (9.88) 2.82 (10.30)
0.29 0.32
⫺6.00 (3.83) ⫺5.25 (3.30)
1.73 (10.74) 1.73 (10.49)
0.24 0.19
⫺6.00 (8.37) ⫺4.25 (7.72)
3.55 (11.98) 3.91 (13.03)
0.14 0.24
⫺7.01 (2.77) ⫺7.30 (3.51)
⫺7.72 (4.80) ⫺7.73 (4.81)
0.47 1.00
Data are change in Regional CBF (in % of mean flow in cerebellum) and Global CBF (mL/100 g/min) a
Wilcoxon Rank Sum Test
CBF ⫽ cerebral blood flow.
two groups whether they had POCD or not, a trend was found. A decrease in rCBF was found in all regions in patients with POCD, and in contrast to this, rCBF increased in all regions in patients without POCD; however, this difference was not significant (Table 3). There was no significant correlation between the neuropsychological Z score at the first postoperative test session and changes in rCBF in any region (p ⬎ 0.40), or the change in gCBF (p ⬎ 0.16).
Comment We found a significant postoperative decrease in gCBF, but these changes did not correlate significantly with neuropsychological test results. There was a nonsignificant trend towards a lower rCBF in patients with POCD. Preoperatively, the patients had significantly reduced gCBF compared with healthy age-matched controls. The incidence of POCD is in accordance with other studies of coronary surgery patients [10, 11]. Smith and associates [2] found a decrease in gCBF in some patients 8 days after cardiac surgery. This change did not correlate with neuropsychological test results. The decrease in gCBF could be related to neuron loss perioperatively as a consequence of either microemboli distributed diffusely
1177
in the cerebral circulation, or irreversible damage caused by cerebral ischemia. However, an additional explanation could be low cerebral metabolism caused by residual effects of sedatives given during hospitalization, or a lower level of psychological stress postoperatively. Hall and associates [4] studied the possible correlation between changes in rCBF during cardiopulmonary bypass using SPECT and neuropsychological decline in cardiac surgery patients. In that study, abnormality in perfusion for a given region was arbitrarily defined as less than 80% of the flow in cerebellum, and cerebral perfusion of the individual patient was defined as having deteriorated if the number of abnormal anatomical regions had increased. According to this definition, 43% of the patients had decreased cerebral perfusion during CPB, but this decrease did not correlate with decline in neuropsychological test performance. In our study, many of the controls had rCBF values less than 80% of the mean flow in cerebellum. If we were to use the same definition of perfusion abnormality as in Hall and associates [4], the majority of the rCBF values for the patients would be abnormal both pre- and postoperatively. Patients scheduled for coronary surgery had significantly lower gCBF than controls. Arteriosclerosis was an exclusion criterion for controls; none of the patients had previous stroke and there was no description of clinically important carotid stenoses, but the patients may have had general arteriosclerosis. Significantly reduced gCBF in nondemented patients with generalized arteriosclerosis has not previously been described. Neuropsychological testing is associated with considerable variability and the definition of POCD varies [6, 12]. Our definition of POCD is more restrictive than in most other studies, and few patients had POCD. But as we used a continuous variable for neuropsychological test performance, the Z score, in the correlation analysis, this did not decrease the ability to find a statistically significant correlation. The SPECT images were parametrically analyzed using standardized region templates defined according to a neuroanatomical atlas. Preferably, structural imaging should have been performed, which enables adjustments for variable neuroanatomy and potential structural lesions. The procedure for normalization of the cortical rCBF values may also be questioned, as we had no imaging data to exclude cerebellar pathology. However, none of the patients had a clinical history or clinical presentation indicating previous lesions or disorders of the central nervous system. The level of CO2 was significantly lower at the postoperative SPECT examination (Table 2). End-expiratory values were measured. If a mismatch between perfusion and ventilation had occurred after surgery, it is possible that these values are not truly reflecting pCO2 in blood. No correction was made for the change in level of pCO2, as the determining factor is pH. As no sudden changes in pCO2 occurred, it is most likely that the individual patient acid-base status was fully adapted. Hemoglobin concentration had declined significantly in patients at
1178
ABILDSTROM ET AL CBF AND COGNITIVE DYSFUNCTION
discharge (Table 2) and we corrected gCBF according to Chen and associates [9]. The average decrease in rCBF in patients with POCD was approximately 7%, in contrast to patients without POCD, who had an increase in rCBF of approximately 2% (Table 3); this difference was not statistically significant. The SD was approximately 10%, and our sample size did not allow us to detect such a small difference similar to the level of interexamination variability in healthy subjects, which may be as high as 10% [13]. Global CBF values in patients before coronary artery surgery were significantly lower than for controls. After surgery, we registered a significant decrease in global CBF, but this did not correlate with neuropsychological test results. In general, patients with cognitive deficits had a decrease in regional CBF, in contrast to patients without cognitive deficits. However, this nonsignificant difference between groups corresponded only to the expected interexamination variability.
Ann Thorac Surg 2002;73:1174 –9
5.
6. 7. 8.
9. 10.
11. The study was supported by grants from the European Commission’s BIOMED2 program (BMH4 –98 –3335), the Danish Medical Research Council, and the Danish Heart Foundation. We thank Gerda Thomsen and Glenna Skouboe for technical assistance.
12. 13.
References 1. Blauth C, Arnold J, Kohner EM, Taylor KM. Retinal microembolism during cardiopulmonary bypass demonstrated by fluorescein angiography. Lancet 1986;2:837–9. 2. Smith PL, Treasure T, Newman SP, et al. Cerebral consequences of cardiopulmonary bypass. Lancet 1986;1:823–5. 3. Venn GE, Patel RL, Chambers DJ. Cardiopulmonary bypass: perioperative cerebral blood flow and postoperative cognitive deficit. Ann Thorac Surg 1995;59:1331–5. 4. Hall RA, Fordyce DJ, Lee ME, et al. Brain SPECT imaging, and neuropsychological testing in coronary artery bypass
14. 15. 16. 17.
patients. Single photon emission computed tomography. Ann Thorac Surg 1999;68:2082– 8. Waldemar G. Functional brain imaging with SPECT in normal aging and dementia. Methodological, pathophysiological, and diagnostic aspects. Cerebrovasc Brain Metab Rev 1995;7:89 –130. Rasmussen LS, Larsen K, Houx P, Skovgaard LT, Hanning CD, Moller JT. The assessment of postoperative cognitive function. Acta Anaesthesiol Scand 2001;45:275– 89. Folstein MF, Folstein SE, McHugh PR. Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189–98. Moller JT, Cluitmans P, Rasmussen LS, et al. Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study. ISPOCD investigators. International Study of PostOperative Cognitive Dysfunction. Lancet 1998;351:857– 61. Chen RY, Fan FC, Kim S, Jan KM, Usami S, Chien S. Tissue-blood partition coefficient for xenon: temperature and hematocrit dependence. J Appl Physiol 1980;49:178– 83. Savageau JA, Stanton BA, Jenkins CD, Klein MD. Neuropsychological dysfunction following elective cardiac operation. I. Early assessment. J Thorac Cardiovasc Surg 1982;84: 585–94. Shaw PJ, Bates D, Cartlidge NE, et al. Early intellectual dysfunction following coronary bypass surgery. Q J Med 1986;58:59– 68. Mahanna EP, Blumenthal JA, White WD, et al. Defining neuropsychological dysfunction after coronary artery bypass grafting. Ann Thorac Surg 1996;61:1342–7. Shirahata N, Henriksen L, Vorstrup S, et al. Regional cerebral blood flow assessed by 133Xe inhalation and emission tomography: normal values. J Comput Assist Tomogr 1985;9: 861– 6. Brand N, Jolles J. Learning and retrieval rate of words presented auditorily and visually. J Gen Psychol 1985;112: 201–10. Reitan RM. Validity of the trail making test as an indicator of organic brain damage. Percept Mot Skills 1958;8:271– 6. Houx PJ, Jolles J, Vreeling FW. Stroop interference: aging effects assessed with the Stroop Color-Word Test. Exp Aging Res 1993;19:209–24. Lezak MD. Neuropsychological assessment, 3rd ed. New York: Oxford University Press; 1995:376–9.
INVITED COMMENTARY I commend the authors for obtaining control groups for this study. The authors choose healthy age matched controls for single photon emission computed tomography (SPECT) scanning, and this is certainly reasonable. We found that at any age, better-educated subjects tended to have better baseline cerebral perfusion as measured by SPECT. I wonder if the test subjects and control group individuals were similar in terms of education level, activity level, and measures of independence, all of which influence global and regional cerebral blood flow (CBF). Regardless, the finding that patients in need of coronary revascularization have diminished cerebral perfusion as compared to healthy volunteers is probably accurate and implies these patients are at risk for cerebral injury. We believe that SPECT may identify a subset of patients with “diminished cortical reserve” and accordingly are vulnerable to cognitive injury from any insult, surgery included. The authors found cerebral blood flow to be signifi© 2002 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
cantly reduced postoperatively as compared to preoperatively. What to account for this? Emboli, ischemic injury, cerebral edema or perhaps altered test environments. Carbon dioxide was lower at time of postoperative SPECT. Was respiratory rate comparable? Were patients compensating for smaller tidal volumes with a relative tachypnea? Similarly, hemodilution with a lower hematocrit postoperatively affects CBF and while a correction factor was applied to global CBF, none exists for regional blood flow evaluations. With significant differences in hematocrit and partial pressure of carbon dioxide before and after surgery, it is difficult to say that cerebral perfusion has indeed changed relative to baseline. It would be fascinating to restudy these patients months down the road, presumably when hematocrit and carbon dioxide levels had normalized, to see if these changes in cerebral perfusion persisted and as this is an important finding, the authors should be encouraged to consider a third SPECT scan in these patients. 0003-4975/02/$22.00 PII S0003-4975(02)03455-0