Change in Plasma Glutamate Concentration During Cardiac Surgery Is a Poor Predictor of Cognitive Outcome James D. Reynolds, PhD, David W. Amory, MD, PhD, Hilary P. Grocott, MD, FRCPC, William D. White, MPH, and Mark F. Newman, MD Objective: To develop a simple and reliable method for quantitating plasma glutamate concentration and apply this method to monitor systemic glutamate levels during coronary artery bypass graft (CABG) surgery, a procedure associated with neurologic deficits. Design: Prospective serial investigation of cardiac surgery patients. Setting: Tertiary-care university teaching hospital. Participants: Patients undergoing CABG surgery (n ⴝ 33). Interventions: Preoperative and postoperative neurologic and neurocognitive testing were done. Intraoperative blood samples for glutamate quantitation were obtained from jugular bulb and pulmonary artery catheters before, during, and after cardiopulmonary bypass. Measurements and Main Results: Glutamate concentrations were determined using a reverse-phase high-pressure liquid chromatography method coupled to precolumn derivatiza-
tion of the analyte with o-phthalaldehyde. The mean prebypass plasma glutamate concentration was 79.4 ⴞ 41.8 mol/L. Plasma glutamate levels fluctuated during surgery with considerable degrees of temporal and quantitative interpatient variability. Neurologic and neurocognitive deficits were observed after CABG surgery. However, neither the occurrence nor the severity of cognitive decline could be predicted by the magnitude of increase in plasma glutamate concentration. Conclusion: Fluctuations in intraoperative systemic glutamate levels do not predict post–CABG surgery neurologic outcome. Copyright 2002, Elsevier Science (USA). All rights reserved.
T
mate levels were unchanged with Parkinson’s disease, whereas Iwasaki et al12 reported a twofold increase in glutamate concentration compared with normal controls. With this background, the purpose of the present study was twofold: (1) to develop a rapid and reliable method for quantitating plasma glutamate concentration; and (2) to apply this method to monitor systemic glutamate levels during a procedure associated with cognitive dysfunction. At this institution, sophisticated neuropsychologic testing has shown that coronary artery bypass graft (CABG) surgery is linked to a significant degree of postprocedural functional and/or neurologic impairment.13,14 Because CABG surgery can produce periods of decreased brain oxygenation, it could be inferred that increases in plasma glutamate concentration during this procedure would correlate with decreases in postoperative mental status. By extension, it was hypothesized that change in intraoperative glutamate concentration could be used to predict neurologic outcome. In this regard, it is believed plasma glutamate levels may hold more promise than previously touted markers of injury after cardiac surgery, specifically S100.15,16 To test this hypothesis, plasma glutamate levels were measured before, during, and after CABG surgery in combination with preoperative and postoperative neurocognitive assessments. Samples were obtained from jugular bulb and pulmonary artery catheters to contrast glutamate levels from (presumed) cerebral and noncerebral sources.
HE NONESSENTIAL AMINO ACID L-glutamate is considered to be the major excitatory neurotransmitter in the cerebral cortex, hippocampus, and other regions of the central nervous system.1 Decreases in cerebral oxygenation can lead to an elevation of glutamate concentration in the synapse.2 If high enough, this glutamate increase results in a cascade of pathophysiologic events within the neuron that eventually can produce excitotoxic neuronal cell death.3 Although most work supporting the role of glutamate in neuronal injury uses experimental animal preparations, the clinical studies that have been conducted suggest similar processes are occurring in the human brain during and after periods of decreased oxygenation.4 Monitoring for increases in intracerebral glutamate concentration can have diagnostic and predictive value.5 However, it is often impractical to employ the invasive instrumentation required to directly determine central nervous system glutamate levels. To that end, some researchers have quantitated systemic glutamate concentration (most often in plasma) as an indirect measure of brain glutamate efflux. Elevated systemic glutamate levels have been reported for many pathophysiologic conditions associated with neurologic injury, including stroke,6 motor neuron disease,7 olivopontocerebellar atrophy,8 and preterm birth asphyxia.9 Various techniques have been used to determine plasma glutamate concentration, but the reliability of some of these sample preparation and quantitative methods is questionable.10 For example, moderate-speed centrifugation (3,000 to 4,000 ⫻ g) can cause the release of glutamate from platelets, whereas ninhydrin (frequently used to form a ternary amine with glutamate) can react with the anticoagulant ethylenediaminetetraacetic acid (EDTA) to produce a fluorescent compound that elutes at the same rate as the acidic amino acids.10 As a result, there is the potential to generate artificially inflated basal values, while obscuring biologically significant increases in plasma glutamate levels. This potential probably accounts for the wide range in reported basal systemic glutamate concentrations. It may also account for varying conclusions reached by different research groups: Mally et al11 reported serum gluta-
KEY WORDS: high-pressure liquid chromatography (HPLC), coronary artery bypass graft (CABG) surgery, glutamic acid, neurologic manifestations
From the Department of Anesthesiology, Duke University Medical Center, Durham, NC. Supported in part by NIH grants HL 54316 and HD 35236 and the Duke Anesthesiology Research Fund. Address reprint requests to James D. Reynolds, PhD, Department of Anesthesiology, Room 119, Research Park Building 4, Duke University Medical Center, Durham, NC 27710. E-mail:
[email protected] Copyright 2002, Elsevier Science (USA). All rights reserved. 1053-0770/02/1604-0008$35.00/0 doi:10.1053/jcan.2002.125148
Journal of Cardiothoracic and Vascular Anesthesia, Vol 16, No 4 (August), 2002: pp 431-436
431
432
REYNOLDS ET AL
METHODS After receiving Duke University Medical Center institutional review board approval, written informed consent was obtained from each participant before enrollment. Patients were excluded if they had known neurologic disorders or conditions associated with alterations in plasma glutamate metabolism (eg, liver disease). A standard perioperative routine was followed: premedication with diazepam and methadone; surgical anesthesia with midazolam and fentanyl; and intraoperative medication with pancuronium, nitroprusside, or phenylephrine to control blood pressure and crystalloid prime and packed red blood cells to maintain hematocrit at ⱖ18%. For cardiopulmonary bypass (CPB), a Cobe CML membrane oxygenator (Cobe Laboratories, Lakewood, CO), arterial catheter filter (SP3840; Pal Biomedical Products, Glencover, NY) and Sarns 7000 max pump (3M Inc, Ann Arbor, MI) were used. CPB used nonpulsatile perfusion at 2.4 L/min/m2. Nasopharyngeal temperature was maintained at 32°C during CPB; rewarming to 36°C occurred at the conclusion of CPB. Pulmonary artery and jugular bulb blood samples for determination of glutamate concentration were obtained before, during, and after CPB (exact time points are defined in Table 2). Blood samples were collected in glass tubes that contained potassium EDTA and centrifuged at 10,000 ⫻ g for 5 minutes. Separated plasma supernatant was immediately stored at ⫺80°C for later analysis.
Quantitation of Glutamate Concentration in Human Plasma Analytical-grade chemicals were used for all steps of the analysis. All solutions were prepared using deionized water obtained from a Nanopure water purification system (Barnstead Sybron, Boston, MA). Preliminary testing determined that 10 L of human plasma, obtained from either catheter site, contained sufficient glutamate for reliable quantitation. As such, after defrosting, a 10-L aliquot of each plasma sample was mixed with 190 L of 0.14 perchloric acid (pH ⬍1.0), then centrifuged at 14,000 ⫻ g for 2 minutes. Of this supernatant, 10 L was diluted with Nanopure water to a final volume of 200 L and dispensed into a 2-mL glass reaction vial. The glutamate concentration was quantitated using high-pressure liquid chromatography (HPLC). All analyses were conducted using a Dynamax automated HPLC system (Ranin Instrument Company, Woburn, MA), which could analyze 84 samples per 30-hour session. The method involved precolumn derivatization of glutamate with 200 L of o-phthalaldehyde/-mercaptoethanol complete reagent solution (pH 10.4; Sigma Chemical Co, St. Louis, MO). After a 2-minute reaction, 100 l of this mixture was injected onto an octadecylsilane reverse-phase column (Supelcosil LC-18, 4.6 mm ID ⫻ 150 mm length, 5 m particle size; Supelco, Bellefonte, PA). The resultant thio-substituted isoindole glutamate derivative17 was separated by isocratic reverse-phase HPLC. The mobile phase, degassed at the time of preparation, comprised 33.7 mmol/L of aqueous sodium acetate in methanol (67.3/32.7; v/v) at pH 7.5 and was pumped through the HPLC system at a flow rate of 2.0 mL/min. Under these run conditions, glutamate elutes from the column after 1.3 minutes. However, a run time of 20 minutes was used because an unidentified compound, with a retention time of approximately 18 minutes, was present in many of the extracted plasma samples. This peak was most likely from midazolam, which is fluorescent by virtue of its heteroaromatic and fused imidazo rings. Each experimental sample was analyzed in a single determination. The amount of glutamate was quantitated by fluorescence detection (excitation at 346 nm; emission at 470 nm). Peak heights were integrated using the Dynamax system software, then the glutamate concentration was calculated by interpolation on a standard curve constructed from serial dilutions of an aqueous glutamate stock solution analyzed with each series of HPLC runs. Each 9-point standard curve was linear (r2 ⬎ 0.99) over a wide concentration range (0.009 to 2.36 mol/L). The within-day coefficient of variation of the assay did not
exceed 0.04, and the lower limit of quantitative sensitivity was 9 nmol/L. Select standards were also analyzed at the end of each HPLC run to confirm that assay parameters remained constant. The presence of perchloric acid in the experimental samples quenched the resultant glutamate signal by approximately 15%. To account for this, on each day of analysis, glutamate standards were quantitated in the presence and absence of perchloric acid. The difference between these chromatographic signals was used to obtain a value for percent recovery, which was factored into the calculation of glutamate concentration in the experimental samples.
Neurologic and Neurocognitive Assessments Standardized neurologic examinations (done by the same neurologist) were conducted before and 5 to 7 days after surgery. Testing used the Western Perioperative Neurologic Scale (WPNS)18 and the MiniMental Status Exam (MMSE).19 For longer term evaluation, a neurocognitive assessment was conducted preoperatively and 6 weeks postoperatively by trained psychometrists. The neurocognitive test battery included 5 assessments of cognitive function and psychometric abilities: (1) Trail-Making Test, Part B (perceptual motor speed); (2) Digit Span (short-term auditory memory, forward and backward); (3) Digit Symbols (psychomotor speed, visual memory); (4) Randt Short-Term Memory Test (verbatim and gist, immediate and delayed); and (5) Modified Wechsler Figural Memory (short-term figural memory, immediate and delayed).
Data Analysis Group demographic and surgical data are expressed as mean ⫾ SD except where noted. Neurocognitive results were assessed using previously described methods.14 To assess neurocognitive decline over time, a factor analysis was done on the 10 baseline neurocognitive scores obtained from the 5 neurocognitive instruments for the entire population. This method constructs a smaller set of independent, nonoverlapping factor scores, each representing a separate domain of cognitive function. The independent factor scores were summed to give an overall cognitive function score at each test period. A cognitive change score (Cognitive Impairment Index) was calculated by subtracting the baseline overall from the 6-week overall score, thus representing a continuous measure of change in cognitive function. In addition, a binary cognitive deficit outcome event (defined as a decline in performance of ⱖ1 SD in any of the independent factors or domains) was determined. The amount of glutamate in the J1 and P1 samples (the first samples taken from the jugular and pulmonary catheters) was averaged to obtain a basal plasma glutamate concentration value for each patient. Sampling constraints limited the number of pulmonary blood samples that could be obtained, so comparisons focused on the jugular fractions. Spearman correlation was used to test for associations between either basal or peak percent change in jugular serum glutamate concentration and a number of psychologic variables. The effect of peak jugular glutamate percent change on 6-week cognitive change was analyzed with linear regression models. The baseline summary score was used to control for baseline cognitive function in multivariable models; age, years of education, and CPB time were also tested as covariable effects. Nonsignificant effects were dropped sequentially from the model. Change in MMSE and change in WPNS were similarly investigated as dependent variables in separate linear regression models. RESULTS
There were 33 patients enrolled in this study; demographic information and surgical time data are presented in Table 1. The study group’s makeup is representative of patients under-
PLASMA GLUTAMATE DURING CARDIAC SURGERY
433
Table 1. Patient Demographic and Operative Data Variable
Value
Sample size (M/F) Age (ys) Diabetes (%) Hypertension (%) Aortic cross-clamp time (min) Total CPB time (min)
33 (24/9) 65 ⫾ 10* 14 (41) 17 (53) 55 ⫾ 23* 106 ⫾ 33*
*These values are expressed as mean ⫾ SD.
going CABG surgery, and the durations for cross-clamping and CPB are within the norms for this institution. Mean basal (the average of samples P1 and J1) plasma glutamate concentration was 79.4 ⫾ 41.8 mol/L. Average plasma glutamate levels for each sampling time point are presented in Table 2. The values were not significantly different from each other but did display, as shown by the large SDs, a considerable degree of interpatient variability. This variability is also illustrated in the individual fluctuations in plasma glutamate concentration. Fig 1 presents the individual peak jugular glutamate values, as percent above basal, versus sample period. The magnitude and time of occurrence of peak jugular plasma glutamate concentration varied considerably among the 33 patients studied. For almost a quarter of the patients, it is difficult to report that a peak occurred because plasma glutamate concentration decreased as surgery progressed. In contrast, other patients exhibited dramatic increases in plasma glutamate concentration. At sampling periods J2 (10 minutes after aortic cannulation) through J6 (60 minutes on CPB), at least 1 patient at each time point exhibited plasma glutamate levels ⱖ100% above basal. For patients who peaked at J7 (10 minutes after CPB), glutamate levels were less than these preceding groups. No patient had plasma glutamate concentration peak at the J8 time point (60 minutes after the aortic clamp was removed). In contrast to the jugular bulb data, glutamate concentration in the pulmonary artery samples remained relatively constant (⫾ 35% of basal) for most patients (summary data in Table 2). Neuropsychologic and cognitive function tests were conducted before and after CABG surgery and compared with the intraoperative plasma glutamate concentration data. Fig 2 presents plots of the peak glutamate values versus changes in MMSE score (A) and changes in the Cognitive Impairment
Fig 1. Individual peak jugular glutamate values, expressed as percent above basal, and the jugular bulb sample period at which the peak occurred for the 33 patients studied. Mean peak glutamate levels are demarcated by the line, and n denotes the actual number of patients who peaked at each time point.
Index (B). For each comparison, patients who had the highest increase in plasma glutamate concentration exhibited little or no cognitive deficit. A similar lack of concordance was observed between WPNS scores and glutamate levels. Likewise, comparing the neuropsychologic scores for each test battery to the time of occurrence of the glutamate peak showed no relationship between these variables (data not shown). DISCUSSION
Substantial experimental evidence indicates that excessive increases in extracellular glutamate concentration within the central nervous system are responsible for the neurologic injury produced by various pathophysiologic insults (eg, stroke). Studies have also suggested that this central increase in glutamate concentration may exude to increase circulating blood glutamate levels.6 Within the brain, extracellular glutamate concentration is tightly regulated by transporters located along the presynaptic membranes of glutamate-releasing neurons and on the neighboring astrocytes.20 Central levels are further controlled by the blood-brain barrier, which acts to extrude plasma glutamate entry.21 Under pathophysiologic conditions, there is
Table 2. Plasma Glutamate Values and Corresponding Surgical Time Points Jugular Bulb Sample
1 2 3 4 5 6 7 8
Surgical Time Point
Patient 10 min 10 min 10 min 10 min 60 min 10 min 60 min
anesthetized after aortic cannulation on CPB after aortic clamp on after aortic clamp off on CPB after CPB after aortic clamp off
Pulmonary Artery
Sample
Glutamate (mol/L)
Sample
Glutamate (mol/L)
J-1 J-2 J-3 J-4 J-5 J-6 J-7 J-8
76.5 ⫾ 44.9 75.0 ⫾ 34.8 76.4 ⫾ 61.2 90.1 ⫾ 76.1 77.6 ⫾ 55.1 60.2 ⫾ 29.0 61.2 ⫾ 38.0 55.4 ⫾ 26.9
P-1 —* P-3 —* —* P-6 P-7 —*
82.2 ⫾ 45.9 65.4 ⫾ 40.0
75.0 ⫾ 56.4 74.4 ⫾ 70.5
NOTE. Glutamate data are expressed as mean values ⫾ SD. Samples P3 and P6 were withdrawn from the central venous port of the pulmonary catheter when on CPB. *Could not sample from the pulmonary catheter at these time points.
434
Fig 2. Individual peak jugular glutamate values, expressed as percent above basal, plotted against post-CABG change in MMSE score (A) and change in the Cognitive Impairment (CI) Index (B). There was no significant correlation between peak glutamate levels and the results from either functional test.
widespread depolarization. This depolarization can impair the normal functioning of the neuronal and astrocyte sodium-dependent transporters, but the transporters present along the blood-brain barrier can act to remove excess glutamate from the extracellular fluid to the bloodstream.20 Although the kinetics of human glutamate transport across the blood-brain barrier are not well understood, animal studies indicate efflux can occur relatively fast: In rats, almost 40% of intracerebrally injected radiolabeled glutamate is excreted into blood within 20 minutes. Thus, changes in plasma glutamate could be temporally associated with central nervous system events that elevate glutamate concentration. By extension, monitoring plasma glutamate concentration could have some diagnostic value in clinical situations in which the potential for neurologic injury exists. The procedures previously used to quantitate plasma glutamate levels have some analytical limitations, starting with sample preparation. Care must be taken when initially processing a blood sample to avoid release of intracellular glutamate stores. This is an important consideration because the glutamate concentration within erythrocytes and other cells in the blood is at least 10-fold higher than that found in plasma.22 Failure to account for this pool when preparing blood samples for analysis results in spuriously high plasma glutamate values. To avoid this contamination, the extracted blood should be centri-
REYNOLDS ET AL
fuged at high speeds (ⱖ10,000 ⫻ g) or passed through a micropore filter at low speed (1,000 ⫻ g).10 Samples prepared by simply centrifuging at 3,000 to 4,000 ⫻ g produce ideal conditions to inflate glutamate values. Studies using this latter method of moderate-speed centrifugation have reported some of the highest basal plasma glutamate concentration values.23 By extension, quantitation of whole-blood glutamate concentration also produces variable results.24 For the glutamate assay, a method was sought that allowed for relatively fast quantitation with a minimal amount of sample preparation. It was also necessary to avoid derivatization with ninhydrin because it can react with the anticoagulant EDTA to produce a fluorescent compound that elutes at the same rate as the acidic amino acids.10 It was decided to use precolumn derivatization with o-phthalaldehyde, a ringed compound that reacts with the primary amine of the amino acid backbone.17 The isoindole derivatives can be readily separated on an octadecylsilane reverse-phase column. Isocratic (single mobile phase) HPLC was used, which, in contrast to gradient systems that employ ⱖ2 mobile phases, does not require extra time to recharge the column between samples. It was recognized that a gradient system might have been useful in the present setup to flush out immaterial compounds quickly after glutamate had eluted from the column (eg, midazolam). As it was, a run time of 20 minutes was not burdensome, although it could become so if, in subsequent studies, the number of samples to be analyzed increased exponentially. Separation was followed by fluorescence detection. Fluorescence, when part of an automated system, proves to be more reliable than other methods, such as electrochemical detection, which can be easily affected by extraneous factors (eg, electrical interference). As a caveat, although a similar method was used to measure other neuroactive amines (eg, ␥-aminobutyric acid),25 the reliability of this method to quantitate the entire range of amino acids that are present in blood has not been evaluated. A direct comparison with other published methods was not made because most previous studies do not include data on the sensitivity or reproducibility of the assay, and previous studies do not report information on through-put (ie, the length of time it takes to run each sample). Nonetheless, by avoiding the technical limitations listed previously, a method was developed that provides reliable quantitation of plasma glutamate concentration with good sensitivity (9 nmol/L). Basal plasma glutamate concentration in this patient population was 79.4 ⫾ 41.8 mol/L. Although within the range of previously reported data, this level is slightly higher than that seen in young adults. Two (probably interrelated) factors seem to account for this increase. The first is that the population was older, and plasma glutamate concentration has been reported to increase with age.26 The second is the propensity of this age group to have arthritis, which can produce elevations in plasma glutamate concentration.27 The study did not include arthritis in the exclusion criteria because of its expected high incidence in patients undergoing CABG surgery and because the source of glutamate seems to be due to altered metabolism within the synovial membrane,28 making it unlikely to impede the ability to detect an efflux of glutamate from the central nervous system into the systemic circulation.
PLASMA GLUTAMATE DURING CARDIAC SURGERY
435
Mean plasma glutamate levels fluctuated during CABG surgery, but no significant increase in glutamate concentration was associated with a single surgical time point. The observed degree of fluctuation was not unexpected. Previous studies involving repetitive blood sampling of awake humans reported circadian variations in plasma glutamate concentrations of 30 to 50 mol/L.29 This variability did lead the authors to focus, however, on intrapatient changes in glutamate levels as a predictor of cognitive outcome after CABG surgery. Consistent with previous investigations,13,14 CABG surgery was associated with a significant degree of postoperative neurologic deficit and cognitive decline. Studies have estimated that the acute incidence of neurocognitive dysfunction after CABG surgery may reach 75%.30 This decline in intellectual function can significantly increase the morbidity and mortality of CABG surgery as well as hinder the general surgical recovery process. An intrasurgical marker that could be used to identify patients at risk for these cognitive deficits would be useful because such patients could be targeted for more aggressive postoperative interventions. Unfortunately, this marker does not appear to be plasma glutamate. Comparison of peak jugular bulb glutamate levels with change in either MMSE or WPNS scores or change in the Cognitive Impairment Index revealed no significant correlations. For all comparisons, the highest glutamate values were associated with minimal changes in mental function, whereas patients who showed the most severe decreases had plasma glutamate levels that varied little from basal levels during the procedure.
Despite no overt correlation in the intraoperative systemic circulation, this by no means excludes glutamate within the central nervous system from being involved in the generation of post–CABG surgery cognitive deficits. In this regard, cerebrospinal fluid sampling of such patients might provide definitive information, although the risks of conducting such a procedure in individuals expected to be at least mildly coagulopathic would make such sampling more complex than the simple intraoperative monitoring the authors had envisioned. Likewise, a role for systemic glutamate levels in this process is not being excluded, especially because the sample size used in the present study is not sufficient to make such an inference. There are many innate (eg, age, genetics) and external (eg, arterial catheter filtration, choice of anesthetic) factors that interact to produce the cognitive decline observed after CABG surgery. By virtue of being the major excitatory transmitter in the central nervous system and its potential to produce neural injury, glutamate probably plays several roles in the cognitive decline, although a complete delineation of these roles and when they are important (possibly postoperatively) remain to be achieved. In summary, a chromatographic method was developed that allows for relatively rapid and reproducible quantitation of plasma glutamate concentration with only a modest amount of presample preparation. This simple analytic procedure may be useful in the early detection of various neurologic conditions. However, application of this method to CABG surgery leads to the conclusion that fluctuations in intraoperative plasma glutamate levels are not a useful marker for predicting post–CABG surgery cognitive outcome.
REFERENCES 1. Ottersen OP, Storm-Mathisen J: Excitatory and inhibitory amino acids in the hippocampus, in Chan-Palay V, Kohler C (eds): The Hippocampus—New Vistas. New York, NY, Alan R. Liss, 1989, pp 97-117 2. Benveniste H: The excitotoxin hypothesis in relation to cerebral ischemia. Cerebrovasc Brain Metab Rev 3:213-245, 1991 3. Choi DW, Lobner D, Dugan LL: Glutamate receptor-mediated neuronal death in the ischemic brain. Mongr Clin Neurosci 16:2-13, 1998 4. Berg-Johnsen J, Haugstad TS, Langmoen IA: Glutamate in the human brain: Possible roles in synaptic transmission and ischemia. Prog Brain Res 116:287-302, 1998 5. Nilsson OG, Brandt L, Ungerstedt U, Soveland H: Bedside detection of brain ischemia using intracerebral microdialysis: Subarachnoid hemorrhage and delayed ischemic deterioration. Neurosurgery 45:1176-1184, 1999 6. Castillo J, Davalos A, Noya M: Progression of ischaemic stroke and excitotoxic amino acids. Lancet 349:79-83, 1997 7. Babu GN, Bawari M, Mathur VN, et al: Blood glutamate levels in patients with motor neuron disease. Clin Chim Acta 273:195-200, 1998 8. Plaitakis A, Berl S, Yahr MD: Abnormal glutamate metabolism in an adult-onset degenerative disorder. Science 216:193-196, 1982 9. Schultz K, Mestyan J, Soltesz G: The effect of birth asphyxia on plasma free amino acids in preterm newborn infants. Acta Paediatr Acad Scient Hungar 18:123-130, 1977 10. Hancock WS (ed): CRC Handbook of HPLC for the Separation of Amino Acids, Peptides, and Proteins. Boca Raton, FL, CRC Press, 1984
11. Mally J, Szalai G, Stone TW: Changes in the concentration of amino acids in serum and cerebrospinal fluid of patients with Parkinson’s disease. J Neurol Sci 151:159-162, 1997 12. Iwasaki Y, Ikeda K, Shiojima T, Kinoshita M: Decreased plasma concentration of aspartate, glutamate and glycine in Parkinson’s disease. Neurosci Lett 145:175-177, 1992 13. Tardiff BE, Newman MF, Saunders AM, et al: Preliminary report of a genetic basis for cognitive decline after cardiac operations. Ann Thorac Surg 64:715-720, 1997 14. Newman MF, Kirchner JL, Phillips-Bute B, et al: Longitudinal assessment of neurocognitive function after cardiac surgery: Perioperative decline predicts long-term (5-year) neurocognitive deterioration. N Engl J Med 344:395-402, 2001 15. Vaage J, Anderson R: Biochemical markers of neurologic injury in cardiac surgery: The rise and fall of S100. J Thorac Cardiovasc Surg 122:853-855, 2001 16. Grocott HP, Arrowsmith JE: Serum S100 protein as a marker of cerebral damage during cardiac surgery. Br J Anaesth 86:289, 2001 17. Jones BN, Paabo S, Stein S: Amino-acid-analysis and enzymatic sequence determination of peptides by an improved ortho-phthaldialdehyde pre-column labeling procedure. J Liq Chromatog 4:565-586, 1981 18. Murkin JM, Martzke JS, Buchan AM, et al: A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery. J Thorac Cardiovasc Surg 110:349-362, 1995 19. 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 12:189-198, 1975
436
20. O’Kane RL, Martinez-Lopez M, DeJoseph MR, et al: Na⫹dependent glutamate transporters (EAAT1, EAAT2, and EAAT3) of the blood-brain barrier. J Biol Chem 274:31891-31895, 1999 21. Smith QR: Transport of glutamate and other amino acids at the blood-brain barrier. J Nutr 130:1016S-1022S, 2000 22. Hagenfeldt L, Arvidsson A: The distribution of amino acids between plasma and erythrocytes. Clin Chim Acta 100:133-141, 1980 23. Castillo J, Davalos A, Naveiro J, Noya M: Neuroexcitatory amino acids and their relation to infarct size and neurological deficit in ischemic stroke. Stroke 27:1060-1065, 1996 24. Thomassen AR, Nielson TT, Bagger JP, Henningsen P: Whole blood versus plasma measurements of glutamate and alanine across the myocardium. Scand J Clin Lab Invest 47:293-300, 1987 25. Reynolds JD, Brien JF: Ontogeny of glutamate and gammaaminobutyric acid release in the hippocampus of the guinea pig. J Dev Physiol 18:243-252, 1992
REYNOLDS ET AL
26. Hack V, Stutz O, Kinscherf R, et al: Elevated venous glutamate levels in (pre) catabolic conditions result at least partly from a decreased glutamate transport activity. J Mol Med 74:337-343, 1996 27. Trang LE, Furst P, Odeback AC, Lovgren O: Plasma amino acids and rheumatoid arthritis. Scand J Rheumatol 14:393-402, 1985 28. McNearney T, Speegle D, Lawand N, et al: Excitatory amino acid profiles of synovial fluid from patients with arthritis. J Rheumatol 27:739-745, 2000 29. Tsai PJ, Huang PC: Circadian variations in plasma and erythrocyte concentrations of glutamate, glutamine, and alanine in men on a diet without and with added monosodium glutamate. Metabolism 48: 1455-1460, 1999 30. Newman M, Frasco P, Kern F, et al: Central nervous system dysfunction after cardiac surgery. Adv Card Surg 3:243-284, 1992