The aluminium content of human serum determined by atomic absorption spectroscopy with a graphite furnace

Clinica

Chimicu

Acta,

Elsevier/North-Holland

CCA

53

114 (1981) 53-60

Biomedical

Press

1809

The aluminium content of human serum determined by atomic absorption spectroscopy with a graphite furnace Oskar Oster * Ahteilung

Jiir Kliniwhe

Chemie und L.uhoruroriumsmedrrin, I, 6500 Main;

(Received

Umversitiitsklmrken

Murnz,

Lungenhecksrrusse

(E R. G.)

November

I I th, 1980)

Summary

A method is described for the determination of aluminium in serum by atomic absorption spectroscopy with a graphite furnace. The serum sample is diluted with a diluent of Triton X-100 and HNO,. The dilution enables shorter drying and charring times and a lower charring temperature. The detection limit of the method is 2.5 pg Al/l serum. Precautions for sample handling are discussed and instrument settings are defined. The direct determination of aluminium in serum is compared with the standard addition method. A good correlation between the two methods was found. Serum from 37 healthy persons was investigated. The serum contained less than 4 pg Al/l with a range of < 2.5 pg to 7 pg Al/l. The aluminium content of 50 serum samples from hemodialysis patients was determined. The aluminium content of the serum of the hemodialysis patients was in the range of 8 to 713 pg Al/l. A mean value of 87.7 pg Al/l (S.D. 122.8) was found.

Introduction

Alfrey et al. showed that patients with renal failure absorb orally administered aluminium salts and demonstrated the association between dialysis dementia frequently leading to death and aluminium intoxication [ 11. Meanwhile it is quite clear that there may well be cases of dialysis dementia caused by other agents, but in the main, dialysis dementia is an aluminium-induced disease [2]. Aluminium intoxication caused by renal failure furthermore causes a vitamin D-resistant osteomalacia [3]. In the rat aluminium intoxication is combined with experimental porphyria [4]. Aluminium is an element of ubiquitous distribution and is therefore encountered daily in many forms in different quantities from a variety of sources. For healthy persons aluminium in food and water is probably innocuous [2]. In the case of chronic renal failure, large amounts of aluminium-containing compounds are given to prevent the accumulation of phosphate [5]. It has been shown that patients with renal failure absorb orally administered aluminium salts, and accumulation of aluminium in blood and tissues occurs [1,5]. The aluminium content of the dialysis 0009-898

I /8 1/OOOO-0000/$02.50

0 Elsevier/North-Holland

Biomedical

Press

54 TABLE

I

SERUM ALUMINIUM REPORTED VALUES Serum aluminium

CONTENT

content

DETERMINED

BY ATOMIC

ABSORPTION

No. of samples

References

29 23 21 5

IO

METHODS-

[PLYWI 37 ‘t 25 2x k 9 340 ,190 240 2 55 37 & IO 62 3 7_t 2 6t 3 724 140 2 60 2.1 & 2.2 <4

13 7 6 6 IO I4 31

9 II 12 8 6 6 6 6 I4 13 This paper

bath fluid and the aluminium hydroxide gels taken by dialysis patients are responsible for the hyperaluminemia and the high aluminium body stores [6-81. This paper describes the determination of aluminium in serum by atomic absorption spectroscopy with a graphite furnace. There is a large discrepancy within the reported serum aluminium concentrations as measured by various atomic absorption spectroscopy methods [6,8- 131. The described procedure takes special care to reduce the possibility of contamination and is mainly reported because of the low normal values found in contrast to most other methods (Table I).

Materials and methods Equipment A model 5000 atomic absorption spectrometer was used equipped with a HGA-500 graphite furnace, standard graphite tubes and cones, a deuterium background correction system, a high speed chart recorder, a printer and an aluminium hollow cathode lamp, all from Perkin Elmer Corp. 20+1 aliquots were injected into the graphite tubes by an automatic injection system (AS 40, Perkin Elmer Corp.). Stundurd and reagents Standard. A 1-mg Al/l aluminium working standard solution made from a certified l-g Al/l solution (Merck 9967) by dilution with the diluent was used to prepare the standards. Diluent solution. As diluent 0.01 mol/l HNO, in 0.1% Triton X-100 was used. Cleaning of the bottles, pipettes, blood collection tubes and Teflon cups The pipettes and disposable tips were rinsed before use at least twice with diluent solution. The Teflon cups in which the serum was diluted were soaked in several changes of diluent solution. The tubes for the collection of blood were also cleaned by soaking them in diluent solution.

55 Blood collection Blood was collected for the normal value study in vacutainer tubes. When the first 0.5 to 1 ml blood was in the tube, the tube was removed from the needle and the blood collected in pre-cleaned blood collection tubes directly from the needle. Contamination control of the blood collection tubes and the Teflon cups Blood collection tubes: The pre-cleaned blood collection tubes were filled with 4 ml diluent solution, closed and rotated for 24 h. The aluminium content of the diluent was determined before and after the rotation. Teflon cups: The pre-cleaned Teflon cups were filled with 1 ml diluent and incubated 2 h after exclusion of dust. The aluminium content of the diluent was determined before and after the incubation. Sample preparation The serum was diluted 1: 5 with diluent solution. For the standard addition method the following procedure was used: Three additions of 10, 30 and 50 ~1 of working standard were made directly into the Teflon cups filled with diluent. Pipettes with disposable polypropylene tips were used throughout the procedure. The determinations were made in duplicate. When the aluminium level was determined directly from the standard curve quadruplicate determinations from separate Teflon cups were made. Results and discussion

If trace elements are detected directly in serum by the graphite furnace technique difficulties occur. After several injections (depending on the programme) a residue is produced inside the graphite tube due to a high concentration of organic material and electrolytes, leading to relatively unreproducible results. An additional problem is the high salt concentration of serum, especially of NaCl which produces a high background. The difficulties can be overcome by special treatment of the graphite furnace, by dilution of the serum sample, or both and background correction [13]. Special treatment of the graphite furnace is sometimes difficult in the clinical laboratory because when a large number of_samples have to be measured the heated furnace is not stable enough. Dilution of the serum samples is more convenient but leads to loss of sensitivity and the detection limit may be outside the normal value range. In the case of determination of aluminium in serum, however, a dilution of the serum is possible without disadvantages. In this reported method the serum is diluted 1 : 5 by a diluent of 0.01 mol/l HNO, in 0.1% Triton X-100. In Table II the instrument and furnace settings are summarized. The drying stage of the diluted sample is divided into two parts. This is important because the drop put into the graphite tube has to be dried carefully so that no splashing occurs. The charring stage is a continuous (40 s) heating from 140°C to 1350°C and an additional heating of 10 s at 1350°C. Atomization is done at 2700°C with an extremely fast heating rate (maximum power mode) for 6 s. With the instrumental and furnace settings summarized in Table II a linear standard curve up to 100 pg Al/l was obtained. The sensitivity, defined in atomic absorption spectroscopy as the quantity needed to give a 1% absorption, was found to be 0.5 pg Al/l. Because the serum is diluted 1 : 5 the sensitivity for serum is 2.5 pg Al/l.

56 TABLE

II

INSTRUMENTAL

SETTINGS

Atomic absorption spectrophotometer Aluminium hollow-cathode lamp lamp current wave length slit mode background correction signal

Graphite

25 mA 309.3 nm 0.7 low absorbance deuterium background peak height

correction

furnace

Argon gas flow Drying stage

step

I

step 2

Charring

stage

Atomizing

step 3

stage

step 4

normal IOO”C IO s IO s 140°C IO s IO s 1350°C 40 s IO b 2700°C maximum 6s stop 2O/Jl

temp. ramp time hold time temp. ramp time hold time temp. ramp time hold time temp. ramp time hold time gas now

Sample volume

power mode

In order to check the accuracy, reproducibility and precision of the method 20,40 and 100 pg Al/l serum samples were prepared and the aluminium content determined. In Table III the results are summarized. Additional confirmation of the accuracy of the method was obtained by the comparison (43 serum samples within a concentration range of 8 to 713 pg Al/l) of the direct determination and the standard addition method. In Table IV the results of the regression analysis and the results of the statistical treatment of the data are summarized. A good correlation between the two methods is found (Fig. 1). No

TABLE

III

SUMMARY OF THE ACCURACY, PRECISION WHICH 20,40 AND 100 pg Al/l WERE ADDED Al content of serum before addition

Addition Al

of

Recovery (mean of 15 samples)

(pg Al/B

(pg Al/l)

his Al/I)

3 4 4

20 40 loo

24.3 46.3 107.3

AND

RECOVERY

DATA

OF

SERUM

Coefficient of variation (C.V.)

Day-to-day precision (10 days)

6.8% 5.4% 4.5%

6.6% 5.7%

TO

57 TABLE IV SUMMARY OF THE CORRELATION SIS

STATfSTICS AND THE LINEAR REGRESSION ANALY-

Compared is the direct determination with the standard addition method. The standard addition method is used as the reference. Number of samples compared Mean value standard addition (pg Al/l) Mean value direct determination (pg Al/l) Concentration range investigated (pg Al/l) Correlation coefficient R Linear regression Y= A + XB Paired t statistic Pairedfstatistic Degrees of freedom

43 87.7-k 122.8 83.01: 118.2 8-713 0.986 A= -0.18 B= 0.95 t= I.498 f= I.080 42

significant statistical difference between the data from the two methods was present as the t test and the f test for paired observations produced a P > 0.05. The good correlation between the two methods was expected because the slope of the standard curve is strictly parallel to the curve obtained by the standard addition method. Another reason for the good result is that from each serum quadruplicate determinations from separate sample cups were made, which reduced the possibility of outliers produced by contamination which is difficult to control. The good co~elation between the two methods also confirms that the matrix effect of different serum samples has no or negligible influence on the results; this is probably due to dilution of the serum. Another important factor is that the background correction is very good. After confir~ng that the accuracy and precision of the direct deter~nation of aluminium in serum was satisfactory, a normal value study was done. The normal values reported for aluminium in serum are very different as is seen from Table I in which the normal values for aluminium in serum, determined by atomic absorption spectroscopy, are summarized. If the normal values determined by other methods,

z .-f D

Stondord

B=

0.95

R=

0.99

Addition [pg At/l]

Fig. 1. Linear regression analysis comparing the direct determination with the standard addition method (reference method).

58 e.g. neutron activation [ 12,151 or spectrographic methods [ 16- 181,are considered the picture becomes still more confusing. 37 serum samples from 23 women and 14 men were analyzed. Their ages were between 18 and 60 years, with an average of 36 years. In Fig. 2 the results of the normal value study in form of a histogram are summarized. 95% of the samples contain less than 7 pg Al/l serum, with a range of < 2.5 to 10 pg Al/l. This result agrees with that of Kaehny et al. [6] and Alderman and Gitelman [ 131. Because of the low normal value of less than 7 pg Al/l the contribution made by contaminations has to be checked carefully. As mentioned before aluminium is an element of ubiquitous distribution and therefore contaminations of the serum samples by the blood collection tubes, the serum storage tubes and the Teflon cups as well as the pipettes, must be considered. Contamination of the pre-cleaned blood collection tubes and serum storage tubes is on average, if the lo-ml tubes are filled with 4 ml blood or serum, 0.63 pg Al/l (S.D. 0.47). The contamination by the Teflon cups filled with 1 ml on average is 0.44 pg Al/l (S.D. 0.28). For the contamination study 20 tubes and cups were examined. Because of the 1: 5 dilution of the serum sample a total average contamination per serum sample of 2.83 pg Al/l (S.D. 1.87) was calculated. If the average contamination is subtracted, 95% of the serum samples contain less than 4 pg Al/l. The normal value of aluminium in serum of less than 4 pg Al/l agrees very well with the lowest normal value so far reported by Alderman and Gitelman, for the alurninium content in serum: 2.1 pg Al/l (S.D. 2.2). In Fig. 2B a histogram of the normal value study after correction for contamination is shown. An average contamination of 2.84 pg Al/l of the serum sample between the patient and the graphite furnace is an

[w Al/l] 100%

50 %

B

L-_ 0334455667

[w

AI/I]

Fig. 2. Histogram of the normal contamination; B, with correction

values of 37 healthy persons; of the average contamination.

A, without

correction

of the average

59

Fig. 3. Histogram

of the swum

aluminium

content

in 50 hemdialysis

patients

acceptable error. To reduce contamination below an average of 2.84 ~18Al/l, major steps are necessary. In the ordinary clinical laboratory these normally are not possible for financial reasons, e.g. installation of a dust-free working bench. However, if serum of dialysis patients is investigated, the degree of error in this method including pre-cleaning steps of the tubes and cups is tolerable. The aluminium content of 50 serum samples from hemodialysis patients was determined. From these 50 serum samples, 43 were measured by the direct method and by the standard addition method. The aluminium content of the serum of the hemodialysis patients was in the range 8-713 pg Al/l. A mean value of 87.7 pg Al/l (S.D. 122.8) was found. In Fig. 3 the distribution of the serum aluminium content of hemodialysis patients is shown in the form of a histogram. Conclusion

A simple method for the determination of aluminium in serum is described. Interfering effects due to the complex matrix of serum are minimized by dilution of the sample. The dilution enables shorter drying and charring times as well as a lower charring temperature. No residue remains in the tube, which is important when larger batches of serum are to be measured. The reproducibility is not very much affected and remains acceptable over a large number of samples measured. The loss of sensitivity by dilution is compensated for by adequate instrumental settings using a shorter drying time, a lower charring temperature and an atomization step in the maximum power mode. With the detection limit for the method described of 2.5 pg Al/l it is possible to measure within the normal range of aluminium in serum from healthy persons. The contamination problem involved with any aluminium determination is minimized by a direct short route from the patient to the serum storage tube and from there to the graphite furnace. The avoidance of any contact of the serum sample with glass is very important. Thus it is possible to work with acceptable accuracy without expensive laboratory equipment. For avoidance of aluminium contamination see also Smeyers-Verbeke et al. [ 191. The normal value found is lower than most other normal values so far reported. A normal value of less than 4pg Al/l serum supports that found by Alderman and Gitelman [ 131.

References Alfrey, A.C., Le Gendre, G.R. and Kaehny, W.D. (1976) The dialysis encephalopathy syndrome. Possible aluminum intoxication. New Engl. J. Med. 294. 184- 188 Berlyne, G.M. (1980) Aluminum toxicity in renal failure. Int. J. Artif. Org. 3, 60-61 Ward, M.K., Fiest, T.G., Ellis, H.A., Parkinson, I.S., Kerr, D.N.S., Herrington, J. and Goode, G.T. (1978) Osteomalacia dialysis osteodystrophy: evidence for a water based agent, probably aluminum. Lancet 1, 841-843 4 Ellis, A.A., McCarthy, J.H. and Herrington, J. (1979) Bone aluminium in haemodialysed patients and in rats injected with aluminiumchloride: relationship to impaired bone mineralisation. J. Clin. Pathol. 32, 832-834 Thurston, H., Gilmore, G.R. and Swales, J.D. (1972) Aluminum retention and toxicity in renal failure. Lancet I, 881-883 Kaehny, W.D., Hegg, A.P. and Alfrey, A.C. (1977) Gastrointestinal absorption of aluminum from aluminum containing antacids. New Engl. J. Med. 296, 1389- 1390 Flendrig, J.A., Kruis, H. and Das, H.A. (1976) Aluminum intoxication procedure. Eur. Dial. Transpl. Assoc. 13, 355-356 Wolf, A., Graf, H., Pingger, W.F., Stummvoll, H.K. and Meisinger, V. (1980) Serum aluminum and continuous ambulatory peritoneal dialysis. Ann. Intern. Med. 92, 130- 13 1 Gorsky, E. and Dietz, A. (1978) Determination of aluminum in biological samples by atomic absorption spectrophotometry with a graphite furnace. Clin. Chem. 24, 1485- 1490 10 Fuchs, C., Brasche, M., Paschen, K., Nordbeck, H. and Quelhorst, E. (1974) Aluminium Bestimmung im Serum mit fiammenloser Atomabsorption. Chn. Chim. Acta 52, II-80 I1 Waldron-Edward, D., Chan, P. and Skoryna, S.C. (1971) Increased prothrombin time and metabolic changes with high serum aluminum levels following long-term exposure to Bayer-process alumina. Can. Med. Assoc. J. 105, 1297-1301 Berlyne, G.M., Ben-An, J., Pest, D., Weisberger, J., Stem, M., Gilmore, G.R. and Levine, R. (1980) Hyperaluminemia from ahuninum resins in renal failure. Lancet 494-497 Alderman, F.R. and Gitelman, H.J. (1980) Improved electrothermal determination of aluminum in serum by atomic absorption spectroscopy. Clin. Chem. 26, 258-260 Anderson, K.J., Noregaard, K., Julsham, K. and Schjoensby, H. (1979) Increased serum aluminum in patients with jaundice. New Engl. J. Med. 301, 728-729 Clarkson, E.M., Luck, V.A. and Hynson, W.V. (1972) The effect of alummum hydroxide on calcium, phosphorus, and aluminum balances, the serum parathyroid hormone concentration and the aluminum content of bone in patients with chronic renal failure. Clin. Sci. 43, 519-523 Kehoe, R.A., ChJoak, J. and Story, R.V. (1980) A spectrochemical study of the normal ranges of concentration of certain trace metals in biological materials. J. Nutr. 19, 579-584 Butt, E.M., Nusbaum, R.E. and Gilmour, T.C. (1964) Trace metal levels in human serum and blood. Arch. Environ. Health 8, 52-56 Wolff, H. (1948) Spektrochemische Untersuchungen tiber den Aluminiumgehah des Blutes. B&hem. Zeitschr. 3 19, 1- 5 Smeyers-Verbeke, J., Verbeelen, D. and Massart, D.L. (1980) The determination of aluminum in biological fluids by means of graphite furnace atomic absorption spectrometry. Clin. Chim. Acta 108, 67-13