| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2002;74:1576-1580
© 2002 The Society of Thoracic Surgeons
a Department of Anaesthesia, Centre of Head and Orthopaedics, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark
b Department of Neurology, Neuro Centre, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark
Accepted for publication June 26, 2002.
* Address reprint requests to Dr Rasmussen, Department of Anaesthesia, Centre of Head and Orthopaedics, 4132, Copenhagen University Hospital, Rigshospitalet, DK-2100 Copenhagen, Denmark.
e-mail: lsr{at}rh.dk
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
METHODS: We included 15 elderly patients undergoing coronary artery bypass surgery. Neuropsychological testing was performed before surgery and postoperatively at discharge from hospital and after 3 months using a neuropsychological test battery. SPECT was performed before surgery and after 3 months using the iomazenil bolus/infusion technique, and the benzodiazepine receptor density was calculated for the frontal, parietal, temporal, and occipital cortex.
RESULTS: Cognitive dysfunction was found in 46.7% at discharge from hospital and in 6.7% after 3 months. A significant decrease in the estimated density of neurons was found in the frontal cortex, but no significant correlation was found between cognitive dysfunction and SPECT findings.
CONCLUSIONS: Neuron loss was detectable in the frontal cortex, but the decrease did not correlate with neuropsychological test results.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
Anesthesia, surgery, and management of cardiopulmonary bypass
Premedication consisted of oral diazepam, 0.1 to 0.15 mg/kg, supplemented with intramuscular morphine, 0.1 mg/kg, and scopolamine, 0.004 mg/kg. Anesthesia was induced by midazolam, 0.05 to 0.1 mg/kg, and fentanyl, 10 to 20 µg/kg, supplemented with pancuronium, 0.1 mg/kg, for neuromuscular blockade. After intubation, anesthesia was maintained by isoflurane, 0.2 to 1.0%, in oxygen/air and supplemented by bolus doses of fentanyl, 5 to 10 µg/kg, as needed.
During cardiopulmonary bypass (CPB), anesthesia was maintained by isoflurane using a vaporizer in the fresh gas supply. The CPB system consisted of a Polystan or Sarns heart-lung machine, polyvinylchloride tubing, and a hollow-fiber membrane oxygenator (Polystan Safe 2 or Sarns SMO Turbo). No arterial line filter was employed. The system was primed with 1,800 mL Ringer’s lactate and heparin, 100 mg. The prime was circulated through a 0.2-µm filter for a minimum of 5 minutes.
CPB was established after heparin, 3 mg/kg, intravenously and measurement of activated clotting time (ACT). ACT levels above 480 seconds were maintained throughout CPB by additional heparin as needed. Moderate systemic hypothermia (30°C to 32°C) was employed. A systemic flow rate of 2.4 L/m2/min was maintained throughout CPB, and alpha-stat acid-base management was employed. Mean arterial blood pressure values of 50 to 80 mm Hg were considered optimal during CPB. Low pressure was treated by increased pump flow or metaoxedrine; high pressure was treated by increased anesthetic depth or nitroprusside. Hematocrit was between 0.20 and 0.30 during CPB and above 0.25 at termination of CPB. During rewarming, the maximal allowed gradient was 10°C, but it was assured that nasopharyngeral temperature never exceeded 38°C. At conclusion of CPB, heparin was neutralized by protamine sulphate on a milligram-to-milligram basis.
Surgical placement of aortic cannulae and clamps avoided (to the extent possible) manipulation of palpable atheromateous plaques in the aortic wall. An aortic cross-clamp was maintained during all distal coronary anastomoses. Cold, crystalloid cardioplegia (St. Thomas solution) was administered antegrade in the aortic root; 10 to 15 mL/kg was used as an initial dose supplemented with 2 to 4 mL/kg for each 20 minutes of continuous aortic cross-clamp. Proximal aortic anastomoses were done during rewarming with an aortic side-clamp and beating heart.
Normoventilation with PaCO2 of 4.5 to 5.5 kPa and PaO2 values above 12 kPa was maintained throughout the perioperative period. Pre-CPB hemodynamics aimed at keeping a heart rate below 90 beats per minute and diastolic arterial blood pressure above 60 mm Hg. Volume replacement, variations in anesthetic depth, IV nitroglycerin, and occasionally small doses of ephedrine were used to meet these demands.
Intravenous nitroglycerin, 0.1 to 5 µg/kg/min, was started before incision and until aortic cross-clamp application; the dosage was titrated according to hemodynamic response and ST-segment analysis. Nitroglycerine was restarted when the cross-clamp was removed. After conclusion of CPB, a mean arterial blood pressure of 65 to 85 mm Hg was considered optimal. Low values were treated by volume replacement, atrial pacing, or inotropic support. Left atrial pressure was generally kept below 12 mm Hg. If bleeding was not excessive, the patients were weaned from mechanical ventilation and the trachea was extubated after additional rewarming in the intensive care unit.
SPECT scan
This was used for the estimation of the central benzodiazepine (BZD) receptor density in the cerebral cortex. The examination was performed before surgery and postoperatively after 3 months, on both occasions on the same day as the neuropsychological tests. The patients received a maximal dose of 111 MBq 123I-iomazenil (Mallinkrodt DRN 5381) preceded by 0.2 g potassium perchlorate given orally 1 hour in advance. First, a bolus dose of iomazenil was given, followed by a constant infusion for 5 hours. The intravenous bolus was 3.6 times the hourly infusion rate to achieve a sustained equilibrium binding of iomazenil at the BZD receptor. The equilibrium approach was chosen to avoid influence of variations in cerebral blood flow on the estimation of the density of available BZD receptors [6].
In in vivo experiments, neuroreceptors are quantified by the binding potential (BP), defined as the product of the maximal density of available receptors (Bmax) and the receptor affinity (1/KD, where KD is the dissociation constant):
 | ((1)) |
At equilibrium, the Michaelis-Menten equation expresses the relation between ligands bound to the receptors (B) and the maximal density of available receptors (Bmax), the dissociation constant (KD), and the free ligand concentration (Cfree) in a given compartment as follows:
 | ((2)) |
 | ((3)) |
 | ((4)) |
As the aim of the study was to study within patient changes after operation, no estimation of absolute values of Vt was attempted. Therefore, assuming unchanged protein-bound fractions and metabolism of iomazenil before and after operation, we used the total lipophilic plasma activity (Cp,tot) as the denominator in equation 4 to characterize the density of the BZD receptors. Furthermore, measuring total plasma counts leads to more reproducible values of distribution volumes in repeated studies compared with more sophisticated measures of plasma concentration corrected for protein-bound fraction and iomazenil metabolites, respectively [7]. Thus, our outcome measure was:
 | ((5)) |
Due to national health regulations of radiation hazards and the need for two examinations, we used a dosage of 123I-iomazenil of approximately one-third compared with the dosage applied by Abi-Dargham and coworkers. This would result in lower counts and thus lesser quality of the SPECT scans. Therefore, we decided to scan at an earlier time point than the 360 to 400 minutes post–bolus scan used by the Yale group, thus "spending" our available dose of iomazenil in a shorter time-frame [8]. To ensure the quality of the equilibrium at the earlier time point, we measured total counts in the slice 50 mm above the CM plane in 10-minute acquisitions at three to five time points between 180 and 330 minutes in the first 4 patients. We found a change of, on average, -0.9% (range +3% to -8%) in the period between 180 and 330 minutes using 210 minutes as reference (the start time of whole brain scans), or a change of -0.7%/hour (range +4% to -5%). This is in concordance with the equilibrium criteria of less than 10%/hour applied by the Yale group and their findings of a stable plasma and brain activity from 180 minutes to the end of infusion [6].
Thus, at 210 to 280 minutes after the bolus, 27 transverse slices parallel to the cantho-meatal (CM) plane were obtained by nine consecutive scans with a brain-dedicated SPECT camera (Tomomatic 564; Medimatic Inc., Copenhagen, Denmark). Images were reconstructed by a filtered back projection algorithm with a Butterworth-like filter (standard Medimatic software) in a 64 x 64 matrix with a pixel size of 3.1 x 3.1 mm. The slice thickness was approximately 15 mm. Again due to the test/retest paradigm, we did not perform attenuation correction. Correction for variations in the camera and the 
-counter sensitivity was applied at every study by a calibration factor. This factor was calculated at each study using the solution activity in a phantom solution, and the activity in the -counter was used for measuring plasma activity.
Three venous blood samples were collected at the beginning, in the middle, and at the end of the scanning period. The Cp,tot was estimated by octanol extraction. After separation of the plasma by centrifugation, 1 mL of plasma was stirred with 2 mL of octanol for 2 minutes, then, after centrifugation, 1 mL of the supernatant was counted in a -counter (Cobra 5002/5003; Canberra Packard, Meriden, CT) with subsequent decay correction.
A standardized set of anatomical ROIs covering nine slices of the brain parallel to the CM plane from CM plane + 10 mm to the CM plane + 90 mm were applied on representative slices in each study. As a given anatomical region (eg, frontal cortex) is represented on more than one slice, the average cpm/10 mm2 in the ROI was calculated by volume-weighted averaging of the counts of the ROIs from the slices presenting the region. The BZD receptor density is reported on an arbitrary scale for frontal cortex, parietal cortex, temporal cortex, and occipital cortex.
Neuropsychological testing
Neuropsychological testing was performed before surgery and postoperatively at discharge from hospital (usually after 7 days) and after 3 months. If testing was not possible at the scheduled time, it was performed as early as possible thereafter. The neuropsychological test battery, the control population, and the definition of cognitive dysfunction were the same as in the ISPOCD study [9], where we studied 1,218 elderly patients undergoing general noncardiac surgery. The test battery took approximately 45 minutes to administer, and evaluated memory, sensimotor speed, and cognitive flexibility. All testing was performed by the same examiner, who was supervised by a neuropsychologist.
The battery comprised the following tests: Mini-Mental State Examination [10], as a screening test for dementia, all patients should obtain a preoperative score above 23 points to be included; Visual Verbal Learning Test [11], immediate and delayed recall after 15 to 25 minutes of 15 words presented on a computer screen; Concept Shifting Test, a speed test that is developed on the basis of the Trail Making Test [12]. There are three subtests that measure the simple and complex cognitive speed and cognitive flexibility; Stroop Color Word Interference Test [13], a speed test that gives a measure of attention and cognitive speed in simple and complex conditions; Letter-Digit Coding, a substitution task based upon the Symbol Digit Substitution task in the Wechsler Adult Intelligence Scale [14].
We used seven parameters for the evaluation of cognitive function: from the Visual Verbal Learning Test, cumulative learning in three trials and delayed verbal recall; from the Concept Shifting Test, time and errors for part C; from the Stroop Colour Word Interference Test, time and errors for part 3, and finally the score from the Letter-Digit Coding. Normative data were available for an elderly population and were used for the calculation of z-scores that express how many standard deviations the patient’s performance deviates from the expected in the control population. These z-scores were calculated as differences in the seven test parameters between the pre- and postoperative test sessions for the individual patients with appropriate sign (+ or -). After subtracting the mean change between sessions in the control population, the result is divided by the standard deviation for this change in the control population. Finally, all z-scores were added for the individual patients. This z-score sum was also calculated for the control individuals, and the standard deviation of these were again used to normalize the patient’s z-score sum into a composite z-score. Patients were defined as having cognitive dysfunction when two z-scores in individual test parameters or the composite z-score were above 1.96. This deterioration was encountered in less than 3.5% of the control population of 176 healthy volunteers with a median age of 67 years (range 60 to 93 years). A detailed description of the control population has been published [9]. The neuropsychological test results have been reported previously in a paper addressing blood markers for brain damage [15].
Statistical analysis
Demographic data are reported as medians with range and proportions with 95% confidence interval. Differences between SPECT estimations of neuron density at the preoperative and postoperative examination were analysed by Wilcoxon’s paired rank-sum test. Differences between patient groups were analyzed using the Mann-Whitney test. Correlation analyses were performed using Spearman’s rank sum correlation. A p < 5% was considered statistically significant.
The sample size calculation was based on a decision that we should be able to detect a decrease in neuron density corresponding to 1.5 SD, significance level of 5%, and power of 80% [16].
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
Estimation of benzodiazepine receptor density in cerebral cortex by SPECT was only possible in 12 patients. In 2 patients, the calculations showed unacceptable differences in calibration results and, accordingly, the SPECT results were rejected. In 1 patient, the data from the postoperative SPECT could not be used at all because of major contamination of tracer during the calibration process. In the remaining 12 patients, the estimated benzodiazepine receptor density decreased significantly 3 months after surgery bilaterally in the frontal cortex and in the left temporal cortex (p between 0.01 and 0.05; Table 1). No significant difference was found in frontal neuron density between patients having cognitive dysfunction and those who did not have cognitive dysfunction, and no significant correlation was found between the change in benzodiazepine receptor density and the composite z-score at any of the postoperative test sessions (Table 2, Fig 1).
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
Brain damage after cardiac surgery has been attributed to embolism during CPB [3, 4]. The emboli may consist of atheromatous plaques, thrombi, lipid, and also air. If air embolism was important, the extent of embolism might be most pronounced in the uppermost regions of the brain. In a supine patient undergoing cardiac surgery, this corresponds to the frontal lobe, where a significant loss of BZD receptors was actually detected bilaterally in our study. We did not use an arterial filter, and the use of side-biting clamps could also contribute to the embolic load.
In previous studies using brain imaging techniques, cerebral infarction has been an uncommon finding, but cerebral edema is a common phenomenon [18–20]. The estimation of BZD receptor density is much more sensitive and has enabled the detection of neuron loss in incomplete cerebral infarction where computed tomography scan examination shows no pathology [5]. Interference with ingested BZD was avoided because no patients received any BZD before or after surgery, but the common use of BZD in the population is an important limitation in the use of this method. The very time-consuming examination and the difficulties associated with the calibrations limit the use of iomazenil SPECT scan to scientific studies only. Preferably, a control group undergoing two SPECT scan examinations with the same interval should have been included as well, but radiation hazards are of major concern.
The BZD receptors are located on neurons, and no evaluation of glial cells is thus possible. Glial cell dysfunction has been detected in several studies after cardiac and noncardiac surgery and could be an important factor in cognitive dysfunction, thus explaining the lack of significant correlation.
In conclusion, we found that a neuron loss was detectable in the cerebral cortex but the decrease did not correlate with neuropsychological test results.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. Gao, R. Taha, D. Gauvin, L. B. Othmen, Y. Wang, and G. Blaise Postoperative Cognitive Dysfunction After Cardiac Surgery Chest, November 1, 2005; 128(5): 3664 - 3670. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
