diseases, we measured concentrations of copper, zinc, magnesium, and calcium. Department of Preventive. Medicine and Community. Health,' and. Department.
CLIN. CHEM.23/3, 485-489 (1977)
Copper, Zinc, Magnesium, and Calcium in Plasma and Cerebrospinal Fluidof Patients with Neurological Diseases John D. Bogden,1 Raymond A. Troiano,2 and Morris M. Joselow1
of some trace-metal concentrations in blood plasma or cerebrospinal fluid, or both, could be of value in diagnosis or management of various neurological diseases, and whether concentrations in plasma could serve as a means of estimating the protein or metal concentrations in cerebrospinal fluid. Samples of both from 82 patients were analyzed for copper, zinc, magnesium, and calcium by atomic absorption spectrophotometry. Protein concentrations in cerebrospinal fluid were also determined. Metal and protein concentrations in plasma and in cerebrospinal fluid were not strongly enough correlated to permit the estimation of one from the other. However, the correlation coefficients between calcium in plasma and cerebrospinal fluid (r = 0.41), magnesium and protein in cerebrospinal fluid (r = 0.40), magnesium in plasma and calcium in cerebrospinal fluid (r = 0.36), and magnesium and calcium in cerebrospinal fluid (r = 0.66) were statistically significant (P < .0 1). Patients with cerebral infarctions had abnormally high copper concentrations in their plasma and cerebrospinal fluid. The ratio of plasma copper to plasma
zinc was also significantly higher in cases of cerebral infarction. AdditIonal Keyphrases: cerebral infarct trace elements atomic absorption spectroscopy protein in cerebrospinal fluid interdisease comparisons .
There is a substantial published literature on concentrations of trace metals in the blood of patients with various diseases (1-5), but relatively little information is available on concentrations of metals in blood and cerebrospinal fluid (CSF) of patients with neurologic disease. In an attempt to determine if blood plasma and/or cerebrospinal fluid trace-metal concentrations, or their ratios, are of value in diagnosis, understanding, or management of neurological diseases, we measured concentrations of copper, zinc, magnesium, and calcium Department of Preventive Medicine and Community Health,’ and Department of Neurosciences,2 College of Medicine and Dentistry of New Jersey, New Jersey Medical School, 100 Bergen St., Newark,
N. J. 07103. Presented in part at the 104th annual meeting of the American Public Health Association, Miami Beach, Fla., October 19, 1976. Received Nov. 17, 1976; accepted Dec. 16, 1976.
in the plasma and CSF of patients with neurological disease. We also wished to determine if concentrations in plasma could serve as a means of estimating concentrations in CSF, to see whether it is really useful to measure trace elements in both. The importance of trace metals in the physiology and pathology of the central nervous system is well established, and concentrations of the above four metals in brain tissue have been investigated in a preliminary study by Greiner et al. (6). We chose initially to investigate copper, zinc, magnesium, and calcium because there is some evidence that the concentrations of these metals may deviate from “normal” in either CSF or plasma in some neurological diseases (7-10).
Materials and Methods Patients Eighty-two patients who required a spinal tap on the Neurology Service at Martland Hospital, Newark, N. J. were studied during January through November, 1975, on those days during which one of us (R.T.) was acting as ward attending physician. All were Newark residents of lower socioeconomic status. The median age of the patients was 45 years, the range 15-84 years. Fifty-eight were males, 24 females; 64 were Negroes and 18 Caucasians. We obtained specimens of spinal fluid from patients undergoing routine spinal taps, done for the usual standard indications in patients with neurologic disease. Blood was also sampled, within 5 mm of the CSF collection. All of the patients were hospitalized on a neurology ward in a general hospital. The population base is such that there is a substantial background prevalence of alcoholism, drug addiction, prior head trauma, and chronic infectious disease, such as syphilis, prevailing throughout all of the diagnostic groups. Ultimately, all patients were classified according to the final major neurologic diagnosis of their current admission. General diagnostic groups were formed where possible to facilitate analysis of data. Electroencephalography, computerized axial tomography, brain scan, and CLINICAL
Vol. 23, No. 3, 1977 485
angiography were used to support the diagnoses in all groups when indicated. The patients were categorized into one of eight diagnostic categories: cerebral infarction (21), subarachnoid hemorrhage (9), alcohol withdrawal syndrome (13), meningoencephalitis (5), schizophrenia (5), dementia (4), seizures (8), or miscellaneous (17). The “miscellaneous” category consisted of patients who had various other diagnoses: multiple sclerosis, amyotrophic lateral sclerosis, myelogenous leukemia, myopathy, spinocerebellar degeneration, or severe chronic headache. Assignment to a group was based on the following criteria. Cerebral infarction was diagnosed on the basis of sudden, nontraumatic onset of focal neurologic deficit, with clear spinal fluid. The “subarachnoid hemorrhage” group is heterogenous. They were classified together based on the common denominator of blood in the CSF. There were three cases of ruptured aneurysms, five of hypertensive intracerebral hemorrhage, and one of subacute subdural hematoma. The “seizure” group is also quite heterogenous. It consists of patients admitted to the hospital because of convulsions, in whom complete neurologic evaluation failed to reveal any evidence of currently active neurologic pathology. All patients in the “alcoholic” group were admitted with a history of current heavy drinking and recent onset of one or usually a combination of the neurologic complications of alcoholism. Seizures, Wernicke’s encephalopathy, cerebellar degeneration, delirium tremens, and alcoholic hallucinosis are all represented in this group. All patients in the “meningoencephalitis” group were diagnosed as having acute viral meningitis based on their benign clinical course, negative results for cultures of blood and spinal fluid, and predominantly lymphocytic pleocytosis in the CSF. The diagnoses of the remaining cases were based on the currently accepted criteria of neurologic diagnosis after thorough neurologic evaluations. Most of the schizophrenic patients were catatonic. The “cerebral infarction” group was the only diagnostic group in which the number of cases was sufficient to allow statistically meaningful analysis of the data. The other groups were considered together for evaluation of the relationship between data for blood and CSF. Sample
From each patient, about 15 ml of whole blood and 5 ml of CSF were obtained, during the day between 0900 and 1630 hours. The blood was collected by venepuncture into plastic syringes and prerinsed (see below) polypropylene tubes to which 20 tl of a sodium heparin solution (10 000 USP units/ml) was added. The CSF was collected by lumbar puncture into prerinsed plastic tubes. The entire system (including the heparin solution) was checked for contamination with copper, zinc, 480
Vol. 23, No. 3. 1977
calcium, and magnesium. Only negligible ( .01) between the protein concentration in CSF and any metal concentration in CSF, except for magnesium (r = 0.40). This is in agreement with a previous report that there is no significant correlation between CSF protein and CSF calcium or zinc (7), but our results differ in that we found a significant correlation between protein and magnesium in CSF. Normal values (per liter) for the analytical techniques used are as follows: plasma copper, 650-1450 pg (11, 15); plasma zinc, 550-1200 tg (11, 15); plasma magnesium, 18-30 mg (11); and plasma calcium, 85-105 mg (11). The mean (±SE) concentrations we found, per liter, were: copper, 1350 ± 30 sg; zinc, .830 ± 30 gig; calcium, 89 ± 1 mg; and magnesium, 20±0.5 mg. These values are for the entire population studied, and are within the normal ranges for each of the four metals. Because of the variety of diseases represented in this patient population, we cannot necessarily conclude that certain specific neurologic diseases may not be associated with abnor-
Table 1. Significant Correlations between Pairs of Metal Concentrations, for 82 PaIrs Examined a Speerman corraL cooff.
CalcIum plasma-magnesium plasma Calcium CSF-magnesium CSF
Zinc plasma-magnesium plasma
0.46 0.41 0.40
Calcium CSF-calcium plasma Magnesium
Mean ± SE
0.84 ± 0.03 2.20 ± 0.07 15.4± 1.0
32.8 ± 2.6
Zinc CSF-calcium CSF Calcium CSF-magnesium plasma 8 Correlations between .05) metal or protein concentration differences between the 21 cerebral infarction cases and the 61 other cases. The normal range of protein concentrations in CSF for the method of analysis used is 200-450 mg/liter (18), and the mean value we found (414 ± 40 mg/liter) is within these limits, but it is relatively high because in-
creased values were found for all nine subarachnoid hemorrhage patients; their mean (±SE) value was 1020 ±80 mg/liter. The mean CSF protein concentration for all cases except the nine in the subarachnoid hemorrhage group is 340 ± 20 mg/liter. The relative paucity of data available on metal concentrations in CSF prevents our defining a normal range, but the concentrations we report here are similar to previously reported values for magnesium, calcium, copper, and zinc (Table 2). Table 3 lists the mean plasma/CSF concentration ratios.
Discussion As mentioned, the mean plasma copper concentration (1510 ± 80 pg/liter) for the 21 cerebral infarction patients is abnormally high, and their mean CSF copper concentration is threefold higher than in the other patients. The relatively high copper concentrations in the plasma of the cerebral infarction patients are of interest in light of previously reported high plasma copper concentrations after myocardial infarction (19). The mechanism for this increase may be common to both. Khandekar et al. (20) have speculated that the increase in serum copper in myocardial infarction may be the result of injury to and subsequent necrosis of myocardial cells. Alternatively, increased hepatic synthesis or decreased breakdown (or both) of the copper-binding protein, ceruloplasmin, may explain the increase (20). Other possible explanations are: altered intestinal absorption, altered excretion patterns, changes in the distribution among body tissues, or some combination of the above factors. The fact that plasma copper concentrations are significantly higher in the cerebral infarction patients than in.the subarachnoid hemorrhage patients is of potential diagnostic value because of the difficulty in distinguishing cerebral infarction from brain hemorrhage.
Table 2. Mean Concentrations of Four Trace Elements in CSF Metal
Zinc Magnesium Calcium S
95± 16(82) 74±5(82) 24.6 ± 0.7 (82)
Woodbury at al. (7) 1968
42.7 ± 1.0 (82)
± standard error (n in parentheses).
Sthl a Agarwal(24)
Gooddy (9) 1974
240 (27) 240 (27) 47.8 (27) 39.0(27)
Calcium and magnesium concentration
Meret & Henkln(13) 1971
Decker at ci. (12) 1964
McCall (30) 1971
30.2 (68) 44.4 (68) units are mg/liter. Copper and zinc concentration
23, No. 3, 1977 487
Taylor et al. (21) suggested that the relatively high copper concentrations found in the cerebral vessels of American Negroes and Caucasians, as compared with those of Nigerians with substantially less atherosclerosis, may be of significance in the pathogenesis of cerebral atherosclerosis. However, there is no proof that increased copper concentrations in cerebral vessels are related to increased copper concentrations in CSF or plasma of patients with cerebral infarction. The high plasma copper/plasma zinc ratio in the cerebral infarction patients is not unique; above-normal ratios are found in leukemia (16) and pulmonary tuberculosis (15) and during pregnancy (1, 2, 4, 5) or the use of oral contraceptives (2). Use of this ratio also serves to correct for diurnal variation in plasma copper and zinc concentrations, because the pattern of diurnal variation of these two concentrations is very similar (22). Monitoring copper in plasma and CSF or zinc in plasma, or both, may be of value in the management of cerebral infarction cases, and studies are in progress to evaluate the relationship between the clinical state of cerebral infarction patients and these trace-metal concentrations. The demographic composition of the cerebral infarction group is somewhat different from that of the remaining cases. For this group the median age was 50 years. Negroes comprised 90% of the group, and a third of the group were women. The median age for the 61 other patients was 44 years; Negroes comprised 74% of the group and 27% were women. It is unlikely that this difference in composition could contribute to the differences we found in plasma or CSF copper concentrations, because age, race, and sex reportedly have little or no effect on copper concentrations (15, 23). Caution is required in interpreting increased plasma copper concentrations, because various conditionssuch as tuberculosis, pregnancy, and acute infectionsalso produce increased plasma copper values (1,2,4, 15). Although an increased plasma copper is nonspecific, it may be of value in the differential diagnosis of neurological disease in the absence of other complicating conditions that are known to produce such increases. There could be no “control” group in this study, because spinal taps are not done except in cases in which there is a legitimate medical need for them. Thus we compared the cerebral infarction cases to a combined group consisting of all other patients in the study, to determine whether information on plasma and CSF trace elements would be of value in the differential diagnosis of cerebral infarction, as appears to be the case for copper concentrations. The plasma/CSF ratios we found (Table 3) indicate that only in the case of magnesium is the value for CSF consistently higher than for plasma. In fact, for copper, zinc, and calcium, the concentration in plasma is higher than the CSF concentration for every sample analyzed. The higher average CSF magnesium concentrations have been previously noted (7,24). This study verifies the poorly understood fact that the magnesium concentration in CSF is maintained at a value about 25% 488 CLINICALCHEMISTRY,Vol. 23, No. 3, 1977
greater than in plasma, while the calcium concentration in CSF is about half the plasma concentration (7), and CSF copper and zinc concentrations are usually much lower than in plasma. The widely different plasmalCSF concentration ratios for the four metals we studied suggest that there are several different mechanisms for establishing and maintaining specific concentrations for each metal in CSF. The comparative analysis of paired specimens of simultaneously obtained blood and lumbar spinal fluid must be analyzed and interpreted with due regard to the total complexity of blood-cerebrospinal fluid relationships. In general, it is currently accepted that under normal circumstances the brain extracellular fluid and CSF are in continuity and are formed by both simple diffusion and active transport between and through the cell membranes of cerebral vascular endothelial cells, which are in apposition to choroid plexus cells, cerebral glial cells, and ependymal cells (25-27). The rate at which various substances penetrate these barriers depends on their molecular weight, polarity, lipid solubility, and metabolic demand. Normally, the CSF concentrations of (e.g.) K+, H+, and Mg2+ are maintained within fairly narrow limits despite wider fluctuations in blood concentrations. On the other hand, the concentration of other substances in CSF is a function of the concentration in blood, and fairly stable blood/CSF concentration gradients may be maintained. It is well known that when concentrations of these substances in blood change there is a variable latency period before equilibrium is established with CSF. For example, it takes several hours for simple molecules such as bicarbonate ion, glucose, urea, or penicillin to so equilibrate (27, 28). If the concentration of trace elements in the CSF were a simple function of their concentrations in blood, then differences in blood/CSF ratios at any given time could be due to rapid fluctuations in blood concentrations in the face of delayed or impaired equilibration between blood and CSF. However, concentrations of trace elements in CSF are probably not simply a passive reflection of those in blood. The blood/CSF ratios, absolute CSF concentrations, and equilibration latencies will be a function of the total amount and of the relative percentage of the trace elements which are free in plasma or bound (e.g., to amino acids, albumin, or metalloenzymes). Under pathological circumstances selective changes in trace element concentrations in CSF could be hypothesized to occur independently of their concentrations in blood. The integrity of the blood-brain-CSF barriers will be another crucial factor. Microligands such as amino acids may be more likely to penetrate a slightly impaired blood-brain-CSF barrier than would larger molecules, and so could be a sensitive indicator of subtle brain damage. Trace-element concentrations in CSF could also change if metalloproteins or cations leaked into CSF from damaged brain tissue; this may be the source of the increased CSF copper concentrations that we found in cerebral infarction patients. Because there
normally are significant concentration gradients between ventricular, cisternal, and lumbar spinal fluid for proteins and electrolytes (28) it is possible that regional changes in CSF constituents could reflect local cerebral pathology and might not be detected by analyses of lumbar spinal fluid. For all these reasons it may be anticipated that, for trace elements, analysis of lumbar CSF concentrations and blood/CSF ratios will be complex, and that very strong correlations, with predictive value, between plasma and CSF concentrations will generally not be found. Our data suggest this to be the case. However, in cerebral infarction cases we found altered (increased) copper concentrations in plasma and CSF. We thank Mr. M. Feuerman (Scientific Data Processing Center, New Jersey Medical School) for his help with data reduction and analysis.
References 1. Underwood, E. J., Trace Elements in Human and Animal Nutrition. Academic Press, New York, N. Y., 1971, pp 68-73 and 214216. 2. Halsted, J. A., Hackley, B. M., and Smith, J. C., Plasma-zinc and copper in pregnancy and after oral contraceptives. Lancet ii, 278 (1968).
3. Delves, H. T., Clayton, B. E., and Bicknell, J., Concentration of trace metals in the blood of children. Br. J. Prey. Soc. Med. 27, 100 (1973). 4. Reinhold, J. G., Trace elements-a selective survey. Clin. Chem. 21,476
5. Burch, R. E., Hahn, H. K., and Sullivan,
J. F., Newer aspects of the roles of zinc, manganese, and copper in human nutrition. Clin. Chem. 21, 501 (1975). 6. Greiner, A. C., Chan, S. L., and Nicolson, G. A., Human brain contents of calcium, copper, magnesium, and zinc in some neurological pathologies. Clin. Chim. Acta 64, 211 (1975). 7. Woodbury, J., Lyons, K., Corretta, R. et al., Cerebrospinal fluid and serum levels of magnesium, zinc, and calcium in man. Neurology 18, 700 (1968). 8. Chhaparwal, B. C., Singh, S. D., Mehta, S., and Pohnalla, J. N., Magnesium levels in serum and in C.S.F. in “meningoencephalitic syndrome.” Indian J. Pediatr. 38, 331 (1971). 9. Gooddy, W., Williams, T. R., and Nicholas, D., Spark-source mass
11. Perkin.Elmer Corp., Clinical Methods for Atomic Absorption Spectroscopy. Norwalk, Conn., 1971, pp Cu 1.1, Zn 1.1, Mg 1.1, and Ca 1.1. 12. Decker, C. F., Aras, A., and Decker, L. E., Determination of magnesium and calcium in cerebrospinal fluid by atomic absorption spectroscopy. Anal. Biochem. 8, 344 (1964). 13. Meret, S., and Henkin, R. I., Simultaneous direct estimation by atomic absorption spectrophotometry of copper and zinc in serum, urine, and cerebrospinal fluid. Clin. Chem. 17, 369 (1971). 14. Ayer, J. B., Dailey, M. E., and Fremont-Smith, F., Denis-Ayer method
for the quantitative
estimation of protein 26, 1038 (1931).
fluid. Arch. Neurol. Psychiatry, 15. Bogden,
J. D., Lintz,
M. M., et al., Effect
nary tuberculosis on blood concentrations of copper and zinc. Am. J. Clin. Pathol., in press (March, 1977). 16. Delves, H. T., Alexander, F. W., and Lay, H., Copper and zinc concentration in the plasma of leukemic children. Br. J. Haematol. 24, 525 (1973). 17. Fisher, G. L., Byers, V. S., Shifrine, M., and Levin, A. S., Copper and zinc levels in serum from human patients with sarcomas. Cancer 37,356 (1976). IS. Annino, J. S., Clinical Chemistry-Principles and Procedures, 3rd. ed., Little, Brown and Co., Boston, Mass., 1964, pp 355-358. 19. Versieck, J., Barber, F., Speecke, A., and Hoste, J., Influence of myocardial centrations.
and zinc con-
Clin. Chem. 21, 578 (1975). 20. Khandekar, J. D., Makurji, D. P., and Sepaha, G. C., Serum copper and iron in ischemic heart disease. Indian J. Med. Sci., 26,813 (1976).
21. Taylor, G. 0., Williams, A. 0., Resch, J. A., et al., Trace metal content of cerebral vessels in American blacks, Caucasians, and Nigerian Africans. Stroke 6,684 (1975). 22. Henkin, R. I., Newer aspects of copper and zinc metabolism. Newer Trace Elements in Nutrition. Mertz, W., and Cornatzer,
E., Eds., Dekker, New York, N. Y., 1971, pp 282-289. 23. Yunice, A. A., Lindeman,
R D., Czerwinski,
A. W., and Clark,
spectrometry in the investigation of neurological disease. I Multi element analysis in blood and cerebrospinal fluid. Brain 97, 327
Influence of age and sex on serum copper and ceruloplasmin levels. J. Gerontol. 29, 277 (1974). 24. Sethi, V. K., and Agarwal, K. C., Magnesium levels in serum and cerebrospinal fluid of normal children. Indian J. Pediatr. 36, 120 (1969). 25. Milhorat, T., The third circulation revisited. J. Neurosurg. 42, 628 (1975). 26. Davson, H., Dynamic aspects of cerebrospinal fluid. Dev. Med. Neurology. 14 Suppl. 27, 1 (1972). 27. Fishman, R. A., Cerebrospinal fluid. In Clinical Neurology, chap. 5, Baker, A. B., and Baker, L. H., Eds., Harper & Row, New York, N. Y., 1975, pp 1-40. 28. Davson, H., Physiology of the Cerebrospirial Fluid. Little, Brown and Co., Boston, Mass., 1967, p 187-191.
10. Gooddy, W., Hamilton, E. L., and Williams, T. R., Spark-source mass spectrometry in the investigation of neurological disease. II Element levels in brain, cerebrospinal fluid and blood: Some observations on their abundance and significance. Brain 98,65 (1975).
human cerebrospinal fluid: Copper and manganese. J. Nucl. Med. 5, 643 (1964). 30. McCall, J. T., Goldstein,.N. P., and Smith, L. H., Implications of trace metals in human diseases. Fed. Proc. 30, 1011 (1971).
E. L., Case, L. F., Miller,
E. B., et al., A study
CLINICALCHEMISTRY,Vol. 23, No. 3, 1977 489