Study of cross-sectional and longitudinal distribution of some major and minor elements in the hair samples of haemodialysed patients with micro-PIXE† Ja�nos Dombova�ri,*a Lajos Papp,a Imre Uzonyi,b Ildiko� Borbe�ly-Kiss,b Zolta�n Elekes,b Zsuzsa Varga,c Ja�nos Ma�tyusc and Gyo�rgy Kakukc aLajos Kossuth University, Debrecen, P.O.B. 21, H-4010 Hungary bInstitute of Nuclear Research of the Hugarian Academy of Sciences, Debrecen, Bem Square 18/c, H-4026 Hungary c1st Department of Medicine, University Medical School, Debrecen, P.O.B. 19, H-4012 Hungary Received 8th September 1998, Accepted 27th January 1999 The concentration of Zn, K, Ca, Fe and Cl was determined in the hair samples of haemodialysed patients and healthy controls. Cross-sectional and longitudinal measurements were performed using the micro-PIXE method. The concentration changes of the above mentioned elements along the length of the hair and their cross-sectional distribution were studied. The eVect of the washing procedure on the concentration of these elements has also been determined.The ability of the PIXE method was examined using Human Hair certified reference material (NCS DC 73347), and the results show that this method can be used to determine the concentration of the selected elements in the hair samples. We compared the results of the PIXE method with the results obtained from microwave digested reference material with the ICP-OES method. The concentration ranges of Ca, Fe and Zn were similar in the patients and controls while the concentration ranges of Cl and K were diVerent (controls: Ca, 135–1598; Cl, 394– 2382; Fe, 6–51; K, 54–1055; Zn, 102–183 mg g-1; versus patients: Ca, 298–559; Cl, 2476–7821; Fe, 15–58; K, 327– 2031; Zn, 96–163 mg g-1).The washing procedure highly aVected the concentration of K and Cl (unwashed: K, 1164±846; Cl, 6107±1714 mg g-1; versus washed: K, 497±552; Cl, 3470±1446 mg g-1). There were no diVerences between controls and patients in the cross-sectional distribution of the selected elements; the distribution of Cl was homogeneous, the Ca was concentrated in the outer layer of the hair while the distribution of K was nearly homogeneous with a slight increase at the outer layer.(3) to validate the micro-PIXE and ICP-OES methods used Introduction using a certified reference material. Elementary composition of blood and plasma samples involv- It has to be noted that hair analysis does not replace the ing micro- and macro-elements is altered by some diseases conventional clinical methods for elementary analysis.such as chronic renal failure (CRF). The main alterations in However, it gives useful additional information about the laboratory parameters of these patients are the increased levels excretion and long-term distribution of the elements. of K, P and Mg, and the decreased levels of Fe, Ca, and Cl. During the last decade, human hair has been extensively Experimental studied in clinical practice as a noninvasive tool for determination of trace elements in the body.The advantage of the In these studies three haemodialysed (HD) patients (mean age 57±1 a) and three healthy subjects (mean age 47±1 a) were hair analysis is that besides noninvasivity it gives an average concentration of trace elements over a longer period of time. involved. Samples were collected from the same place on the head, In this respect, Al is one of the most interesting ions causing intoxication in CRF patients.It is due to the altered metab- namely from the occipital region. In the case of each person some hair strands were removed from the scalp by pulling olism of Al in these patients.1–4 However, the average concentration of macro-elements such as Ca, Mg, Fe, Cl and P, and them out, preferably together with the hair root (because in this way the identification of the proximal end is easy). the eVect of preparation and washing procedures on the concentration and distribution of these elements in the hair, Hair strands were washed with diethyl ether–acetone (3+1) mixture, nonionic detergent solution (1% Decon-90 in triply is not yet known.Preparation and digestion of hair samples prior to analysis requires special care involving several washing distilled water) and triply distilled water, respectively. After washing the hair samples they were dried at 105 °C for 90 min. procedures5–7 and digestion methods.7–9 Therefore, the aims of our present work were: (1) to study the eVect of the washing For cross-sectional studies samples were embedded into Araldite resin (Eporapid, Budapest, Hungary) as follows: a procedure on the concentration of some elements in the surface and the cross-sectioned part of hair samples using micro- single hair was placed in the axis of a plastic tube (10 mm internal diameter, 30 mm length) and then the tube was filled PIXE; (2) to compare the concentration of these elements in the hair of CRF patients with that of healthy controls; and with resin.After hardening, sections were cut at 5, 10 and 15 mm from the proximal end, respectively. Samples were polished, washed with triply distilled water and dried at room †Presented at the 8th Solid Sampling Spectrometry Colloquium, Budapest, Hungary, September 1–4, 1998. temperature. In longitudinal studies hair strands were streched J. Anal. At. Spectrom., 1999, 14, 553–557 553onto an aluminium frame and their ends were fixed with from the sample, excluding the radiation from the resin.This method was applied in the longitudinal experiments, too. cyanoacrylate glue (Heurel KGaA, Du� sseldorf, Germany). On the frame three parallel single hairs were fixed at a distance of Because of the limited scan length (2.5 mm) the full length of the hair samples (25 mm) was examined in 10 segments and 3 mm from each other. To eliminate the X-ray radiation from the mountings of the equipment, a high purity Teflon plate the total concentration of the selected elements was determined in each segment.(3 mm thickness) was placed behind the frame. The study of the certified reference material (NCS DC 73347 Human Hair, purchased from: National Research Centre for Results and discussion CRMs, No. 7 District 11, Hiepjinge 100013, Beijing, China) was performed by both PIXE and ICP-OES methods. For The applicability of the micro-PIXE method was studied using the certified reference material and the results obtained were PIXE measurement pellets of about 1 mm thickness and 8 mm diameter were made from the powdered reference hair. The compared with the results of the ICP-OES measurement of the same reference hair.These results are shown in Table 3. concentrations of major elements which cannot be determined by PIXE (C, O, H and N) and the concentration of sulfur In the case of Ca, K and Zn the ICP-OES method produced better recoveries, while for Fe 80% (ICP-OES) and 120% which had to be used as an internal standard were determined with independent analytical methods, or the certified values (PIXE) of the certified concentration was measured.The concentration of Cl was not measurable using the direct ICP- were used. In ICP-OES measurements #0.5 g reference hair was digested in a Milestone MLS 1200 Mega high pressure OES method, and the results obtained with PIXE could not be explained with the lack of a certified value.microwave digestion unit (Microwave Laboratory Systems, Italy) in Teflon vessels, by the addition of 3 cm3 of 65% high The comparison of the measured values of HD patients, healthy controls and the concentrations taken from the litera- purity nitric acid (Merck, Darmstadt, Germany) and 0.5 cm3 of 30% high purity hydrogen peroxide (Merck). Samples were ture12 are shown in Table 4. Comparing the results of the HD patients with the controls we found that the concentrations of filled to 10 cm3 and the concentration of the selected elements was determined.ICP-OES experiments were carried out using Zn, Ca and Fe were in the same order of magnitude, if the error of the method and the biological variability was taken a Spectroflame (Spectro GmbH, Germaupled plasma optical emission spectrometer; the experimental param- into account. The concentrations of Cl and K were greater in the case of HD patients than that of the controls.It is also eters are shown in Table 1. In PIXE experiments the samples were analysed with the micro-PIXE technique using the scan- seen that the measured concentration values in HD patients were at the lower part of the reference range in the case of Ca ning proton microprobe facility at the Institute of Nuclear Research of the Hungarian Academy of Sciences in Debrecen. and Zn and higher than the reference for Cl and Ca; and in the case of Fe the two ranges were almost identical.It is Instrumental conditions are shown in Table 2, further details on the experimental setup can be found elsewhere.10 The known in clinical practice13 that the K concentration in the plasma of HD patients is higher than the normal values, while Si(Li) detector was protected from the backscattered protons by a hostaphan absorber of 12 mm thickness. Characteristic Fe, Ca and Cl concentrations are lower than the normal values.In the hair of HD patients this tendency was true only X-ray spectra and the elemental composition of the samples were evaluated by using the PIXYKLM program package.11 for Ca and Zn. The cross-sectional distribution of the selected elements is Because of the lack of a suitable current integrator we used S as an internal standard in the evaluation of Micro PIXE shown in Fig. 1. The concentrations of Ca, K and Cl in the human hair is high enough to obtain adequately evaluable spectra supposing 4.5% for its concentration. Besides S the detection limits were low enough to evaluate the concentration distribution maps.We found that the distribution of Cl was homogeneous in the cross-sections. In the case of K its of Cl, K, Ca, Fe and Zn but they were not satisfactory for P and Cu in many cases. During the cross-sectional study a distribution was also nearly homogeneous. However, the conshort scan was made on the surface of each sample to determine the exact location and shape of the hair section.Table 3 Intercomparison of the certified reference material (NCS DC The long data acquisition (Tm=60 min) was performed only 73347 Human Hair) Concentration/mg g-1±standard deviation Table 1 Instrumental and analytical conditions of ICP-OES Instrument Spectroflame ICP-OES Element Certified ICP-OES PIXE Frequency 27.1 MHz Nebulizer Spectro concentric nebulizer Ca 2900±200 2768±110 3700±200 Cl — — 100±40 Spray chamber Scott-type Flow rate of plasma gas 1.5 L min-1 Fe 54±6 43± 7 65± 5 K 20* 19±2 17± 5 Flow rate of cooler gas 15 L min-1 Flow rate of nebulizer gas 1 L min-1 Zn 190±5 195±8 210±15 Rate of sample 1 cm3 min-1 *Not certified, for reference only.introduction Integration time 25 s on the line, 25 s at the background Applied software Spectro EPC3 for Windows Detection lines/nm Ca, 31.9; Fe, 259.9; K, 766.5; Zn, 213.9 Table 4 Comparison of the measured concentration ranges of diVerent subjects (washed samples, longitudinal experiments, micro-PIXE) Concentration ranges/mg g-1 Table 2 Micro-PIXE experimental setup and analytical conditions Scanning Nuclear Microprobe Haemodialysed Control Element Literature12 (n=3) (n=3) Instrument Oxford, Microbeams type Proton beam energy 2 MeV Ca 146–3190 298–559 135–1598 Cl 950–4805 2746–7821 394–2382 Beam current #50 pA Collected charge #0.2 mC Fe 5–45 15–58 6–51 K 150–663 327–2031 54–1055 Integration time 60 min Beam spot-size #2×2 mm Zn 99–450 96–163 102–183 554 J.Anal. At. Spectrom., 1999, 14, 553–557Ca Cl Fe Zn 50 mm 50 mm 50 mm concentration low high 50 mm 50 mm K Fig. 1 The typical cross sectional distribution of the selected elements (selected maps, haemodialysed patient, micro-PIXE). J. Anal. At. Spectrom., 1999, 14, 553–557 555Fig. 2 Typical results of the longitudinal experiments: 1 segment=2.5 mm (haemodialysed patient, micro-PIXE). centration of K was higher in the outer layer than inside the tendency was observed.The Cl concentration was the lowest in the first segment, and there was a maximum in the second medulla, suggesting that an inadequate washing procedure might aVect its concentration in the outer layer. We found segment. The tendency was the same in the case of K. We found that Fe concentration increased in the cross-sections as small segregations in the case of Ca in the outer layer of the hair, in contrast with the more homogeneous distribution of a function of distance from the proximal end.The Zn concentration in the cross-sections does not change considerably at K. Because of their lower concentration, the distribution maps of Fe and Zn are less valuable, but it is observable that the diVerent distances. Fig. 2 shows the results of the longitudinal studies. The distribution of Zn was nearly homogeneous and the Fe was segregated in the outer layer. concentration of Zn was practically constant on the surface along the length of the hair sample.The concentration maxi- The diVerence in the concentration before and after the washing procedure in the case of K was the highest among mum at the distal end of the hair sample was probably of exogenous origin, which was proved by the fact that the the examined elements (see Table 5), which suggests that K can easily be extracted from the outer layer of the hair by the washing procedure removed this maximum, and the Zn level of the hair became nearly constant. After the washing pro- washing procedure because of its mobility.14 The identification of the exogenous or endogenous origin of the extracted K cedure the Zn concentration of the hair decreased #15%, except for the proximal end. In the case of unwashed hair requires further experiments.The change in concentration of Cl during washing was also high, although its distribution was samples K and Cl concentrations increased from the proximal to distal end.After washing this tendency was eliminated, and homogeneous in the cross section of the hair. The concentration of Ca, Fe and Zn remained nearly unchanged during the K concentration became constant along the length of the hair. In the middle of the hair the K level is higher than at washing. In Table 6 the typical concentrations of the selected elements the ends. The possible reason for this might be that the washing procedure did not completely remove the exogenous measured in the cross-sections of a single hair taken from diVerent distances from the proximal end (5 mm, 10 mm, K or the endogenous K is distributed inhomogeneously in the hair.It is discernible that the concentration of K was almost 15 mm, respectively) are shown. In the case of Ca decreasing the same at the proximal end of the single hair before and Table 5 Changes in concentration during the washing procedure (average of all measured values in the longitudinal experiments, micro- Table 6 Typical results of cross-sectional experiments (washed samples, haemodialysed patient, micro-PIXE) PIXE, n=6) Mean concentration/mg g-1±standard Concentration at diVerent distances (l) from the hair root/mg g-1±standard deviation deviation Element Before washing After washing Element l=5 mm l=10 mm l=15 mm Ca 1130±14 629±7 713±8 Ca 408±150 384±112 Cl 6107±1714 3470±1446 Cl 1474±72 3362±40 2162±49 Fe 57±11 6±4 49± 6 Fe 32±20 31±22 K 1164±846 497±552 K 179±9 420±6 239±6 Zn 202±30 175±16 186±20 Zn 142±24 121±42 556 J.Anal. At. Spectrom., 1999, 14, 553–557after washing. At this region the exogenous K concentration In clinical practice the monitoring of the trace element content of the body during a longer life period could be which can be removed by washing was smaller than at the other segments of the hair. The K concentration increased important (except for acute problems like poisoning). Such examinations require analytical methods in which small from the proximal to the distal end of the hair when unwashed, while it was nearly constant after washing, because the exogen- amounts (0.5–0.01 g) of hair are used, and prior to analysis homogenisation, dissolution or digestion is applied to avoid ous contamination was proportional to the exposure time, and this period was the shortest in the case of the proximal hair the sample inhomogenity and concentration fluctuations.However, the distribution of the elements inside the hair, segment.The concentration of Ca slightly increased along the hair before washing because of the accumulation of exogenous along the hair strands, can be determined by using PIXE or micro-PIXE methods. These are nondestructive and have high Ca. After performing a washing procedure the Ca level fluctuated. In the middle of the washed hair sample, similarly to K, resolution, but are slow and expensive methods, which cannot be applied routinely in clinical practice. the Ca level was high, possibly because of the increased K and Ca uptake at that period.The concentration of Cl increased on the surface of the hair, although the tendency is not as Acknowledgements clear as in the case of Ca or K. After washing, the Cl level This research was granted by the National Research greatly decreased, the change in concentration was higher at Foundation (OTKA) under project number T22739. the distal end where the exogenous contamination was higher.We observed that the concentration of Fe on the surface References fluctuated before washing. After washing we found that the Fe concentration decreased from the proximal to the distal 1 T. D. Lyon, C. Cunningham, D. J. Halls, J. Gibbons, A. Keating end. The iron concentrations at the distal and proximal ends and G. S. Fell, Ann. Clin. Biochem., 1995, 32, (2), 160. of the washed sample were higher than in the case of unwashed 2 L. Scarpino, A. Confessore, A. Bruno, A.Bonifati, M. Gatti, I. sample. This decrease is in good agreement with former results Maimone, M. P. Minella, C. Sapio, F. Smilari and G. Attademo, Minerva Urol. Nefrol., 1994, 46, 89. found in the literature,15 which show that the iron is not 3 U. Vogelsang, Schweiz Rundsch. Med. Prax., 1994, 83, 738. incorporated into the hair at the hair bulb but it is incorporated 4 J. E. Jervis, B. T. Kua and G. Hercz, Biol. Trace Elem. Res., 1994, from other sources close to the skin surface.Hence its concen- 43–45, 335. tration is high in the outer layer, and along the hair the highest 5 J.S¡ tupar and F. Dolnis¡ek, Spectrochim. Acta, Part B, 1996, 51, values are measured not in the hair bulb, but close to the skin 665. at the proximal end. Therefore the decrease of the Fe level 6 J. Bacso� , P. Kova�cs and S. Horva�th, Radiochem. Radioanal. Lett., 1978, 33(4), 273. can be explained by the dissolution of the iron from the hair 7 D. E. Ryan, J. Holzbecker and D.C. Stuart, Clin. Chem., 1978, by environmental eVects. 24, 1996. 8 J. Dombova�ri and L. Papp, Microchem. J., 1998, 59, 187. 9 P. Senofonte, N. Violante, L. Fornarelli, E. Beccaloni, A. Powar Conclusions and S. Caroli, Ann. Ist. Super. Sanita, 1989, 25, 385. 10 I. Rajta, I. Borbe�ly-Kiss, Gy. Mo� rik, L. Bartha, E. Koltay and Our results show that there are considerable diVerences A. Z. Kiss, Nucl. Instrum. Methods Phys. Res., Sect. B, 1996, 109– between the concentrations of Ca, Cl and K in the hair of HD 110, 148. 11 Gy. Szabo� and I. Borbe�ly-Kiss, Nucl. Instrum.Methods Phys. Res., patients, healthy controls and the results found in the literature. Sect. B, 1993, 75, 123. However, sample treatment has a major eVect on the measured 12 G. V. Iyengar, W. E. Kollmer and H. J. M. Bowen, The Elemental values; the applied washing procedure influences the element Composition of Human Tissues and Body Fluids, Verlag Chemie, content of the sample. Our general observation is that the Weinheim, New York, 1978, pp. 7–13, 51–54. measured values and the tendencies along the length of a 13 F. Fischbach, A Manual of Laboratory Diagnostic Tests, third single hair were more variable prior to the washing procedure. edition, ed. D. Inteno, J.B. Lippincott Co., Philadelphia, 3rd edn., 1988, pp. 41, 46, 66, 97, 159. A possible reason is the inhomogeneous endogenous trace 14 F. Watt and J. P. Landsberg, Nucl. Instrum. Methods Phys. Res., element distribution (which is due to the irregular uptake). Sect. B, 1993, 77, 249. This endogenous distribution is covered by the more fluctuat- 15 A. J. J. Bos, C. C. A. H. Van Der Stap, V. Valkovic, R. D. Vis and ing exogenous contamination of the external layer. Another H. Verkenl, Nucl. Instrum. Methods in Phys. Res., Sect. B, 1984, reason could be the insuYciency of the applied washing 3, 654. procedure which cannot completely remove the contamination from the surface of the sample. Paper 8/07030J J. Anal. At. Spectrom., 1999, 14, 5