Measurement of copper, zinc and magnesium in serum and urine by DC plasma emission spectrometry

Transcription

1 Ann Clin Biochem 1985; 22: Measurement of copper, zinc and magnesium in serum and urine by DC plasma emission spectrometry N B ROB E R T S, D F A I R C LOU G H, S M c LOU G H LINand W H T A Y LOR From the Department of Chemical Pathology, Royal Liverpool Hospital, Liverpool L7 8XP, UK SUMMARY. The application of a DC plasma emission spectrometer to the measurement of copper, zinc and magnesium in serum and urine is reported. Each assay requires only a simple dilution in 1% nitric acid. Comparisons with standard atomic absorption techniques showed good analytical agreement, especially for copper and magnesium. Analysis of quality control preparations for copper and zinc gave results ranging from -1 5% to +3 6% of the stated values for five copper and four zinc experiments, with an unexplained result of +7 6% for one zinc experiment. Reference ranges were compiled for serum zinc and urinary copper which agreed closely with those established by atomic absorption and neutron activation analysis. Assays by DC plasma emission are thus precise and reproducible, and simple enough to authenticate the method for use in a clinical laboratory. Roberts, Fairclough and Taylor! have recently reported the application of a DC plasma emission spectrometer to metabolic balance studies, with particular reference to the analysis of food and faecal calcium, phosphorus and chromium. They emphasised the relative ease of operation of the technique with minimal sample preparation yielding satisfactory accuracy, reproducibility and precision. We have explored further the potential usefulness of this technique in clinical biochemistry by applying it to the measurement of copper, zinc and magnesium in serum and urine. Although elemental analysis in serum and urine does not present the same preparative problems as food and faeces, the atomic absorption and chemical methods for serum copper and zinc are sometimes preceded by a de-proteinisation step, with the risk that the metal ions may be co-precipitated with the protein.? Other difficulties with atomic absorption techniques have also been described; Spranger and Franz' found that viscosity differences between plasma and aqueous standards may cause a substantial measuring error with zinc determination in atomic absorbance spectrometry. They can be minimised by calibrating in 10% glycerol/water solutions, or using propan-l-ol as diluent." For urinary copper determination the limits of detection of flame atomic absorbance procedures have in the past required the adoption of a concentration step, such as complexing with ammonium pyrrolidine dithiocarbamate and then extracting into methyl isobutyl ketone. Plasma emission spectrometry avoids these difficulties by direct aspiration into the instrument of a dilution of serum or urine in 1% (v/v) nitric acid. Materials and methods All reagents were of analytical grade and obtained from British Drug Houses Ltd, Poole, Dorset (BDH). Solutions were prepared in doubly de-ionised water (Spectrum System, Elga Products Ltd, High Wycombe, Bucks). Copper, zinc and magnesium primary standards were atomic absorption grade solutions (BDH). External quality control materials used were Cation Cal, (Dade), QAP control sera, levels 1 and 11 (all supplied by America Hospital Supply, Didcot, Oxon), and normal and abnormal control sera (Wellcome Reagents Ltd, Beckenham). It was found that the siliconised rubber stoppers of commercial control sera containers may contain variable amounts of zinc which can affect the results quite markedly, depending on the degree of liquid contact with the rubber stoppers. Freeze dried QC sera were aliquoted, therefore, into plastic-capped tubes and frozen as soon as reconstituted.

2 534 Roberts et al. ATOMIC ABSORPTION PROCEDURES The atomic absorbance spectrometer was the SP 1900 (Pye-Unicam, Cambridge). For copper, the settings adopted were: wavelength nm; slit width 0 08 nm; burner: 10 em acetylene; burner height 1 2 cm; fuel: acetylene 300 mumin (10 psi); oxidant: air 5 Umin (30 psi); integration time 1 s. Serum copper was measured after de-proteinisation with 25% w/v trichloroacetic acid according to the method of Piper and Higgins.? Zinc settings were as for copper, except the wavelength was set at nm. Serum zinc was measured directly after dilution (1 in 10) in doubly de-ionised water. Calibration and standard dilution was carried out in aqueous solution containing sodium chloride (14 mmoj/l final concentration). This solution was used to simulate the emission and scatter from the diluted serum sample since zinc is measured in the ultraviolet region of the spectrum where sodium interference is greatest. Magnesium settings were as for copper except that the wavelength was set at nm, (and burner head rotated 35 to reduce sensitivity), Magnesium (serum and urine) was measured after diluting 1 in 20 in strontium chloride diluent (19 g strontium chloride, 45 ml perchloric acid in 5 L of water). PLASMA EMISSION PROCEDURES The DC plasma emission spectrometer (Beckman-RIIC, High Wycombe) was set up as reported by Roberts et al. 1 using an argon (high purity grade, British Oxygen Company, Prescot, Merseyside) flow of 2 Umin across the -tungsten and carbon electrodes. The samples were nebulised through a crossed-flow nebuliser with argon carrier gas at 7 Umin. The sample was aspirated at approximately 2 mumin and analysed over 2 x 10 s or 2 x 5 s integration periods. The exit slits of the spectrometer were set at 200 urn vertically and 500 um horizontally, and the entrance slits were usually set at 50 or 100 urn vertically and 200or 500 urn horizontally. The instrument was calibrated for each element under the conditions given in Table 1 and the concentration of the unknown element was computed directly by the instrument. Copper and zinc assay For serum analysis, the samples were diluted 0 2 ml serum plus 1 8 ml 1%(v/v) nitric acid. Standard addition procedures were carried out by adding known amounts to each diluted sample. For urine analysis, the following standard addition procedure was used: to four tubes (plastic) 1 0 ml of urine was added; then to one tube, the 'blank', 1 ml of 1% HN0 3 was added and to the others 1 0 ml of diluted standard added to achieve the required final concentration (Fig. 1). The extrapolated values could then be compared with the direct assay of the 'blank'. Magnesium assay Serum samples were diluted 1 in 100, (0 05 ml serum with 4 95 ml 1% nitric acid v/v containing 0 1 % Triton X-1(0), and urine 1 in 20 in 1 0 M lithium nitrate in 1% nitric acid v/v. The lithium diluent can also be used for food and faecal analysis but was not used with serum due to the protein-denaturing effect of the high lithium concentration. The addition of Triton X-100 for serum magnesium assay was found to confer enhanced signal-nebulisation stability. TABLE 1. Analytical conditions for copper, zinc and magnesium assay by plasma emission spectrometry. Calibration was carried out for serum and urine analysis between low and high standards of the BDH spectroscopy grade solutions diluted in the respective diluents, and was linear throughout the range. The elemental emission lines were chosen to give the highest sensitivity and the least spectral and matrix interference by reference to the 'Spectraspan Spectral Line Handbook' (Beckman-RIle) Values of low and high standards Emission line Element (nm) Standard solution Serum Urine Concentration Copper Copper nitrate I1moUL Zinc Zinc nitrate IJ.moUL Magnesium Magnesium chloride mrnol/l

3 Cu, Zn and Mg by DC plasma emission spectrometry 535, 8 FIG :; "0 E "0.. " 16 '" c. Ė Analyte added (pmol/l) The assay, bydcplasma emission, of copper and zinc in urine and serum with increasing amounts of analyte added. a Urine copper; serum copper; o urine zinc; serum zinc. The intercept values for copper in urine and serum were 2 4 and 3 9 compared with extrapolated values of 2 35 and 3 9 respectively. The intercept values for zinc in urine and serum were 5 6 and 7 6 compared with extrapolated values of 5 7 and 7 5, respectively. Blood samples were collected in to plastic bottles, with no additive, and 24 h urine samples in to de-ionised water-washed plastic containers with no preservative added. Results COPPER AND ZINC ASSAY Figure 1 shows that the' method of simple dilution in 1% HN0 3, and calibration of the instrument without any 'matrix matching', gives the same results with or without standard addition, thus confirming the absence of any significant matrix effect. Analysis of commercially available quality control sera gave good agreement with expected values for copper and zinc as measured by atomic absorption procedures (Table 2). Further, the precision for within- and betweenbatch analyses indicates satisfactorily reproducible assays (Table 2). Overall, in the five experiments with copper (Table 2) the mean values ranged from -1 3% to +3 6% of the stated values; for four of the zinc experiments the mean values ranged from -1 5% to +2 9% of the stated values, but for monitrol 1 there was a difference, acceptable but unexplained, of +7 6%. COPPER IN SERUM AND URINE The assay of serum copper by plasma emission was compared with the stated atomic absorption procedure and gave a highly significant degree of correlation (Fig. 2) over the entire pathological range. National quality control returns (over 18 months) for serum copper gave a similar highly significant degree of correlation with no signifi- TABLE 2. Precision analysis, within- and between-batch, for the measurement of copper and zinc by plasma emission. The stated values for Cation Cal copper and zinc were 30 7 and 45 3!Lmol/L respectively. For monitrollevel 1, copper and zinc values were 17 3±1 8 (SD), and 16 22±1 79 respectively, and for level 11, 13 64± 1 43 and 13 63±1 59 umol/l, The values in parentheses indicate the number of observations Cation Cal Monitrol Within-batch Between-batch Between-batch Copper 30 3±0 2 (5) CV 0 66% 31 8±1 6 (11) CV 5 0% Levell 17 58±I l (3) CV 6 3% 31 2±O 6 (6) CV 1 9% Level ±0 85 (3) CV 6 2% Zinc 44 64±1 7 (10) CV 3 8% 46 6±1 2 (14) CV 2 6% Levell 17 46±1 1 (5) CV 6 3% 45 4±2 0 (16) CV 4 4% Level ±0 5 (5) CV 3 6% The Cation Cal solution was diluted two-fold for this analysis.

4 536 Roberts et al. 50.Y 40 :::l./. ::::. 0 :::l! z;:- :::: ;: e 20 c.;; -=e': c./. 0 ytll..,..;; " 20 '" '/ OJ.~. s s " 10 OJ '" OJ ~ e -e 10 '" OJ ~ 30./ Atomic obsorbonce (umol/l) FIG.2. Comparison of serum copper assay by plasma emission and atomic absorbance. Mean values (n=51) obtained were 25 9 umol/l (range ) for plasma emission and 26 11!mo1lL(range ) for atomic absorption. The correlation coefficient was +0 98, and the regression line was given by y=i 07x-l 67. The theoretical slope 1 00 is drawn Atomic absorbance (pmol/l) 30 FIG. 3. Comparison of serum zinc analysis by plasma emission and atomic absorbance. Mean values (n=28) were umol/l (range ) for plasma emission and urnol/l (range ) for atomic absorption. The correlation coefficient was +0 97, and the regression line given by y=i 25 x-2 8. The theoretical slope 1 00 is drawn. cant difference between the results. An overall assay bias of +2'17% was computed for 9 returns. To investigate urine copper assay by plasma emission a comparison was made of values determined directly upon diluted urine with those obtained after standard addition. Table 3 shows that there was no significant difference between the two procedures. Twenty-four hour urine copper outputs were determined by plasma emission on 12 healthy volunteers and gave a mean value of 0 38±0 15 (SD) I!moV24 h, range This compares well with the mean of 0 43±0 28 I!moVL 6 obtained from 16 volunteers. ZINC IN SERUM AND URINE Twenty-eight blood samples analysed for zinc by both plasma emission and atomic absorption, gave a highly significant degree of correlation (Fig. 3). It should be noted that 10 of the 28 analytical data points did not agree to within ±10%. Comparison of results from commercially available quality control material (assayed by atomic absorption) gave very good agreement (Table 2). A reference range was determined by plasma emission on 45 healthy volunteers, (29 male and 16 female). The overall mean was 14 9±1 8 I!moVL with individual ranges (normally distributed) of and I!moVL respectively. Fifty-five samples of urine, mainly from patients in intensive care, were assayed for urinary zinc, and the method of standard addition was compared with direct aspiration of the diluted urine. The excellent correlation obtained (Table 3) showed no evidence of any significant differences between the two procedures. TABLE 3. Urine copper and zinc measurement by plasma emission. Comparison of direct and extrapolated standard addition results. Values as ILmoVL Copper Zinc Number Direct Mean (y) Range Extrapolated Mean (x) Range G Correlation Coefficient P<O OOI P<O'OOI Slope Intercept

5 Cu, Zn and Mg by DC plasma emission spectrometry 537 TABLE 4. Comparison of serum and urinary magnesium values as measured by plasma emission and atomic absorbance MAGNESIUM IN SERUM AND URINE Comparisons of magnesium assay in serum and urine by plasma emission and atomic absorbance are shown in Table 4. Both procedures gave essentially the same results over the quoted serum range and thus indicated that serum magnesium assay by plasma emission is free from interference. A within-batch precision analysis for serum magnesium on an external quality control serum (Wellcome) gave 2 9% coefficient of variation with a mean of 0 68 mrnol/l ±0'02 (n=7); the stated mean value was 0 67 (range ) mmol/l, Individual data points for urine magnesium showed that plasma emission agreed closely with atomic absorption at values less than 6 0 mmov24 h, of which there were 21. With higher values in three urine samples, plasma emission gave slightly higher levels than atomic absorption and this has caused a slightly higher mean (3,86 as against 3 66 mmov24 h) and accounts for the intercept of Discussion Serum (mrnol/l) Number 26 Plasma emission (x) Mean 0 80 Range Atomic Absorbance (y) Mean 0 81 Range 0' Correlation coefficient (r)+0'99 Slope Intercept P<O OOI Urine (mmov24 h) P<O OOI The methods described for serum and urinary copper and zinc have now been in routine use in our department for at least 12 months. Serum and urinary magnesium are still determined by atomic absorption. The optical system and plasma, when set up according to the manufacturer's instructions, have shown no major instability, ie, no tendency to drift off peak or to cause fluctuating emissions, provided that sufficient air extraction through the plasma chimney is used to prevent any distortion of the plasma. Maintenance of constant instrument base-plate 24 temperature has also proved unnecessary. It was important to ensure adequate sample equilibration for at least 5-10 s before taking readings when using plasma emission. The use of plasma emission procedures for the assay of trace elements has been reported to be subject to signal enhancement," particularly with respect to trace element analysis in sea water or brine solutions. We have found that this effect is not of any major consequence in. physiological samples: for example, in urine analysis for copper and zinc the sample is aspirated with minimal dilution and the results obtained by reference to aqueous nitric acid standard solutions. Further, dilution of standards and sera with water give similar results to those obtained using nitric acid as diluent. The latter was in fact simply used to maintain element solubility at low ph. Littlejohn (personal communication) has suggested that nitric acid diluent will suppress alkali metal dependent signal enhancement. The use of a lithium nitrate diluent (to minimise any alkali metal effect) has similarly been found unnecessary for the assay of magnesium in urine. Thus the effect of alkali metal dependent signal enhancement on the elements studied may be a minor phenomenon in physiological matrices, particularly at the dilutions reported. The values for urinary copper output determined by plasma emission agree closely with those reported by Halls et al. 6 using flameless atomic absorbance procedure, in comparison with the earlier high results of Cowell.f For serum zinc there may well be unexplained analytical differences between the atomic absorbance procedure used and plasma emission, but the reference range now established by plasma emission in our laboratory agrees closely with the mean value of 13 8±1 9 SD (n=50),' the normal range of obtained by atomic absorption and the range of urnol/l obtained by neutron activation analysis." The principal advantage of plasma emission in the determination of copper and zinc is that it is not subject to some of the analytical restrictions associated with atomic absorption, such as viscosity effects, sodium interference and the necessity for protein precipitation. Further advantages are that lamps are not required because anyone of the various spectral lines can be selected, and that sample preparation is minimal. The multi-element version of the instrument would enable several trace elements

6 538 Roberts et al. to be determined simultaneously once the analytical conditions for each have been optimised. It can be concluded then that the methods validated and now adopted in our laboratory show that the DC plasma emission spectrometer is a versatile analytical instrument, simple to use, and worthy of consideration for analysis of a range elements in a routine clinical laboratory. References 1 Roberts NB, Fairclough D, Taylor WHo Plasma emission spectrometry: measurement of calcium, phosphorus, and chromium in metabolic balance studies. Ann Clin Biochem 1984; 21: Weinstock N, Uhlemann M. Automated determination of copper in undiluted serum by atomic absorption spectroscopy. Clin Chem 1981; 27: Spranger KBG, franz HE. Viscosity adaptation for an automated micro-method of flame atomic absorption spectrometry and intracellular traceelement analysis after pressure decomposition: zinc determination in plasma and erythrocytes. Clin Chem 1983; 29: {j. 4 Peaston RT. The determination of zinc and copper in plasma and urine. Med Lab Technol 1973; 30: Piper KG, Higgins G. Estimation of trace metals in biological material by atomic absorption spectrometry. Proc Assoc Clin Biochem 1967; 4: Halls DJ, Fell GS, Danbar PM. Determination of copper in urine by graphite furnace atomic absorption spectrometry. Clin Chim Acta 1981; 114: Nygaard SD. Plasma emission determination of trace heavy metals in salt water matrices. Anal Chem 1979; 51: Cowell DC. Determination of urine and serum copper levels using a single beam atomic absorption spectrophotometer. Med Lab Technol 1973; 30: Olehy DA, Schmitt RA, Berthard WF. Neutron activation analysis of magnesium, calcium, strontium, barium, manganese, cobalt, copper, zinc, sodium and potassium in human erythocytes and plasma. J Nucl Med 1966; 7: Accepted for publication 27 March 1985