Biomonitoring of occupational exposure: Neutron activation ...

Journal of Radioanalytical and Nuclear Chemistry, Vol. 259, No. 1 (2004) 7–11

Biomonitoring of occupational exposure: Neutron activation determination of selected metals in the body tissues and fluids of workers manufacturing stainless steel vessels J. Ku era,1* V. Bencko,2 J. Tejral,3 L. Borská,4 L. Soukal,1 Z. -anda1 1 Nuclear Physics Institute, CZ-250 68 8e9 near Prague, Czech Republic University of Prague, 1st Faculty of Medicine, Institute of Hygiene and Epidemiology, Czech Republic 3 Institute of Hygiene and Preventive Medicine, Faculty of Medicine, Charles University, Hradec Králové, Czech Republic 4 Institute of Pathological Physiology, Faculty of Medicine, Charles University, Hradec Králové, Czech Republic 2 Charles

(Received July 21, 2003)

Occupational exposure was examined for 20 workers dealing with welding, polishing, and assembling of stainless steel vessels. Instrumental neutron activation analysis was used for determination of selected elements in hair and nail, whereas urinary Cr and Mn, blood Mn and serum Cr were determined by radiochemical neutron activation analysis. Increased levels of Cr in hair, nails, serum and urine, Mo in hair, and Mn in blood were found in the exposed group compared to controls. Accuracy of the results was proven by analysis of reference materials and by comparison of element levels in controls with reference values for non-exposed persons.

Introduction Workers in the stainless steel construction industry are exposed to numerous substances with known neurotoxicity, genotoxicity, carcinogenic, allergenic and immunological effects.1 The main health risk is associated with inhalation of welding fumes and airborne particulate matter (APM), which contain elevated levels of steel-alloying elements, such as Cr, Ni, Mo, Mn, V. There is a growing need to harmonize activities in the field of occupational health methodology and approaches to risk assessment2,3 in which direct and/or biological monitoring can be used. Results of direct monitoring, i.e., the assessment of exposure from concentrations of toxic substances in the workplace may be misleading. Various mechanisms of intake and/or absorption may lead to a different body burden, when multiple mechanisms of intake are to be considered, and especially when short-term and long-term effects are to be distinguished.4 Therefore, in the assessment of the health risk arising from environmental, occupational and accidental exposure to toxic metals and other substances, the use of biological monitoring is steadily increasing,5–7 because our knowledge of various factors that influence levels of biological indicators of exposure8–10 and those that affect reference values for occupationally non-exposed populations has considerably increased in recent years, especially for toxic trace elements in the body tissues and fluids.11 Both approaches were used to examine occupational exposure of workers of a plant manufacturing stainless steel storage and production vessels for the pharmaceutical, food and chemical industries. Results of

multi-elemental analysis of workplace APM performed by INAA, which showed exceeding of maximum admissible limit for Cr in the workplace air, have already been published.12 In this work the biological monitoring involved the determination of levels of Cr, Mn, Mo, Ni and V in hair and nails by instrumental neutron activation analysis (INAA), and levels of Cr in serum and urine and Mn in blood and urine by radiochemical neutron activation analysis (RNAA) in exposed and control persons.

Experimental Exposed and control persons Details on both groups have already been given elsewhere.12 Briefly, the group of exposed workers (average exposure 16 y) consisted of 18 men and 2 women (average age 33.1 y), of which 52 % were smokers. They were occupationally exposed to welding fumes and/or APM originating from shaving, polishing and assembling of vessels made of austenitic stainless steel containing on average 18% of Cr, 9–10% of Ni, 2– 2.5% of Mo, 1–2% of Mn and traces of V (~0.01%). Most of the workers were not specialized in one type of work, but were rather changing their activities during a week and/or during a shift. Thus no subgroup of welders, polishers, etc. exposed exclusively to welding fumes or APM could be distinguished. The control group was formed by 20 men (average age of 45.5 y, 47% of smokers) employed in an agricultural enterprise located about 5 km from the plant producing stainless steel vessels.

* E-mail: [email protected] 0236–5731/2004/USD 20.00 © 2004 Akadémiai Kiadó, Budapest

Akadémiai Kiadó, Budapest Kluwer Academic Publishers, Dordrecht

J. KUBERA et al.: BIOMONITORING OF OCCUPATIONAL EXPOSURE

Sampling and sample handling Both groups of workers were sampled in the middle of the week, prior to the beginning of a work shift, after all persons took a shower. Hair and nail samples were obtained by clipping with stainless steel scissors. The IAEA recommended procedure was employed for hair washing,13 whereas an adapted procedure described elsewhere3 was used for nail cleaning. Spot samples of urine of a volume 50 to 100 ml were collected in acid leached polystyrene vials. Blood was obtained from a cubital vein using a Teflon cannula, which was first flushed with 15 ml of blood collected for immunological tests. Then 5 ml of blood was collected in polyethylene (PE) cryogenic vials (Nalgene) which were pre-cleaned as given below. From this volume, 1 to 2 ml were immediately transferred to another Nalgene vial for determination of manganese, while the remaining blood was centrifuged to obtain serum for chromium determination. Serum aliquots of 1 to 1.5 ml were placed into pre-cleaned vials made of synthetic quartz (Suprasil AN, Heraeus). The quartz vials were closed with an acid-leached Teflon stopper. All manipulations on blood sampling and handling were performed in the stream of air obtained from a mobile filtration unit with ULPA filters (Holten) providing a Class 10 environment. The urine, blood and serum samples were deep-frozen prior to further treatment. All collection vials used for sampling and sample handling were pre-cleaned by leaching in dilute subboiled nitric acid for 24 hours and washing with de-ionized water in a clean laboratory providing a Class 100 environment. Irradiation For INAA, samples and standards were prepared for irradiation as follows. Hair (75 to 150 mg) and nail (25 to 75 mg) samples were packed in disk shaped polyethylene PE capsules of a 20-mm diameter made by heat sealing of PE acid-leached foils of 0.15 mm thick. Aliquots of 0.5 ml of urine for the determination of Mn and Cr by RNAA were placed in acid-cleaned PE and quartz vials, respectively. Frozen blood samples for the determination of Mn by RNAA were irradiated in the collection PE vials, whereas the serum samples for the determination of Cr by RNAA were freeze-dried in the quartz vials prior to irradiation. Multi- and single element standards for INAA and RNAA, respectively, were prepared in the same geometrical shape as the samples by weighing out 20 to 50 µl aliquots of solutions containing known concentrations of elements. For INAA, this has been achieved by depositing the solutions onto disks of chromatographic paper Whatman, air drying and subsequent heat sealing in the

8

PE disk capsules, whereas for RNAA the standards were made to the sample volume by de-ionized water. The samples and standards were irradiated in VVR15 nuclear reactor of the Nuclear Research Institute, JeK, plc at fluence rates of 1.1014 and 3.1013 n.cm–2.s–1 for thermal and fast neutrons, respectively. Short-time irradiation (1 min) in PE capsules and/or vials was carried out in a pneumatic facility with the transport time of 4 seconds. The samples and standards were irradiated in PE “rabbits” individually, together with neutron flux monitors (5 µg of gold prepared as the standards). For long-time irradiation (20 h) in quartz vials, the samples and standards were packed together in an Al irradiation can. Radiochemical separation Chromium in urine and serum: After 1 to 2 weeks of decay, the outer surface of quartz vials were cleaned by leaching in hot aqua regia and cooled in liquid nitrogen. After opening of the vials, urine was washed out with 5 to 10 ml of distilled water, while serum was solubilized in 5 ml of fuming nitric acid in a beaker. Radiochemical separation of 51Cr was carried out using an adapted procedure according to GREENBERG and ZEISLER.14 To both sample types, 1 ml of a K2CrO4 solution containing 4 mg of Cr per ml (inactive carrier) and 3 ml of concentrated H2SO4 were added. Then, the samples were mineralized in a Kjeldahl decomposition flask by repeated additions of 1 ml aliquots of concentrated HNO3 during heating the flask over a gas burner until a clear solution was obtained and white fumes of H2SO4 appeared. Then, Cr(III) was oxidized to Cr(VI) by addition of 1 ml of concentrated HClO4 and heated till white fumes of HClO4 and subsequent change of the color of the solution from green to orange. After cooling down to laboratory temperature, 1 ml of a 0.05 mol.l–1 KMnO4 solution and 25 ml of a 2.5 mol.l–1 HCl were added and Cr(VI) was extracted in a separatory funnel for 1 minute by two 10 ml-portions of a 5% (w/v) solution of tribenzylamine (TBA) in chloroform. Combined chloroform fractions were scrubbed with 5 ml of water for 30 seconds (to remove co-extracted 65Zn) and Cr was stripped into water phase with 5 ml of 3 mol.l–1 NH4OH for 90 seconds. After washing the separatory funnel with 1 ml of water, the 6-ml fractions with separated 51Cr were placed in plastic vials for counting. The chemical yield of separation that varied in the range of 90–95% was determined by reactivation of 25-µl aliquots of the water phase, which contained 16.66 µg of Cr inactive carrier, i.e., an amount in sufficient excess to Cr present in the samples. Manganese in urine and blood. An adapted procedure published earlier15 was employed, which is briefly described as follows. After 1 to 2 minutes of


decay the irradiated samples were transferred to the quartz Kjeldahl decomposition flask to which 200 µl of a MnSO4 solution containing 5 mg of Mn per ml (1 mg of Mn inactive carrier), 25 µl of a radioactive 54Mn tracer solution (approximately 500 Bq) and 3 ml of concentrated H2SO4 were added. Sample decomposition was carried out by repeated addition of 1 ml aliquots of concentrated HNO3 (in total 5 to 6 ml and 1 to 2 ml for blood and urine, respectively) and heating the flask over a gas burner until a clear solution was obtained. The solution was then cooled down, transferred to a beaker, diluted with 35 ml of water, 13 ml of concentrated NH4OH were added to reach pH~8, then 5 ml of a 5% (NH4)2S2O8 solution were added and hydrated MnO2 and Fe(OH)3 were precipitated by a short boiling on a hot plate. The resulting precipitate was filtered off after cooling using a membrane filter Synpor (Pragochem, Czech Republic) with a pore diameter of 0.4 µm. The separation procedure took approximately 20 minutes. The chemical yield of Mn separation as determined by measuring the 54Mn tracer was in the range of 95–99%. Counting and interference correction The irradiated hair and nail samples were counted using a coaxial HPGe detector with a relative efficiency of 23% and a FWHM resolution of 1.8 keV at 1332.4 keV in conditions similar to those described earlier.12 The 320.1-keV photons of 51Cr in the 6-ml separated fractions were counted using a 150 cm3 well type HPGe detector for 4 to 10 hours. The activities of 56Mn (846.8 keV and 1810.9 keV) and 54Mn (834.8 keV) in the membrane filters were counted on the top of a coaxial HPGe detector with relative efficiency 53% and a 1.8 keV FWHM resolution. After counting, the filters were irradiated for 2 hours to determine the content of Fe to be able to correct for the interfering reaction 56Fe(n,p)56Mn with fast neutrons.

The interference contribution yielded 16.4 ng of “apparent” Mn per mg of Fe as determined by INAA of JSS 003-4 High Grade Pure Iron 3 reference material16 with a certified content of Mn (32.2±1.65 mg.kg–1). The contribution was negligible in the analysis of urine, but in the case of blood about a half of 56Mn measured originated from the above interfering reaction. Another interfering reaction 59Co(n, )56Mn with fast neutrons was negligible for both matrices due to a very low Co content. For the Cr determination in urine and serum, the interference reaction 54Fe(n, )51Cr was also negligible due to a low content of Fe in both matrices.

Results and discussion Quality assessment The accuracy of INAA results was proven in an earlier stage of this study in which membrane filters with air borne workplace APM were analyzed together with the NIST SRM-1648 Urban Particulate Matter. Excellent agreement was obtained with NIST-certified, information (non-certified) and/or literature values.12 Due to the matrix independence of INAA, this proof of accuracy applies also for INAA results for hair and nail samples reported in this work. The accuracy of the Cr and Mn determination by RNAA was proven by analyses of low element-level biological reference materials. Table 1 shows that our results compared with the reference values within the uncertainty margins. Although the element levels in the reference materials were about one order of magnitude higher compared to those in urine, blood or serum of control persons, another proof of accuracy was obtained by comparison of our results with reference values for occupationally non-exposed population as pointed out below.

Table 1. Quality assessment of Cr and Mn determination by RNAA Material NIST SRM-1549 Milk Powder IAEA RM A-13 Animal Blood IAEA RM H-4 Animal Muscle

Element

This work (N),* µg kg–1

Reference value [Ref.], µg kg–1

Cr

2.9 ± 0.4 (5)

2.6 ± 0.7 [17]

Mn

30.8 ± 3.9 (3)

31 ± 2.6 [15]

Mn

465 ± 13 (3)

446 ± 42 [18]

* Combined uncertainty (coverage factor k = 1), N: number of determinations.

9


Biological monitoring of exposure Table 2 shows results of the determination of Cr, Mn, Mo, Ni and V in hair and nail samples by INAA, and of Cr in serum and urine, and Mn in blood and urine by RNAA for the exposed and control persons together with their statistical evaluation. The data sets obtained were examined for normality using the KolmogorovSmirnoff-Liliefors and Shapiro-Wilk tests. For normally distributed data, the arithmetic mean and standard deviation were calculated, for those deviating from a normal distribution the median, lower and upper quartiles were evaluated. Based on the data distribution, differences between exposed and controls were evaluated by the parametric t-test for normally distributed data sets, whereas for those with a nonnormal distribution the non-parametric Mann-Whitney U-test was employed. Significantly increased hair-Cr, nail-Cr and hair-Mo values were found in the exposed workers compared to controls. Although element contents in hair and also in nails, as another ectoderm derivative, are regarded suitable bioindicators of environmental and occupational exposure4,7,21,22 the increased element levels should be interpreted as a proof of occupational exposure with caution. The reason is that there is no hair-cleaning procedure, which would completely remove exogenous contamination (caused by, e.g., elevated element levels in the workplace air) without influencing the endogenous element content of hair.21 The above statement is even much more critical for nail analysis, because removing of the exogenous contamination from this tissue is much more difficult compared to hair.

Moreover, unlike the IAEA hair washing procedure,13 no generally accepted procedure for nail cleaning has been recognized until now. Urinary and blood element levels are therefore regarded much more straightforward biological indicators of occupational exposure.22 Table 2 shows that a highly significant increase (p