Annals of Occupational Hygiene Advance Access originally published online on March 21, 2006
Annals of Occupational Hygiene 2006 50(5):517-525; doi:10.1093/annhyg/mel009
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Biomonitoring for Chromium and Arsenic in Timber Treatment Plant Workers Exposed to CCA Wood Preservatives
1 Health and Safety Laboratory Harpur Hill, Buxton, SK17 9JN, UK
2 Health and Safety Executive, Redgrave Court Bootle, L20 7HS, UK
*Author to whom correspondence should be addressed. Tel: +44-1298-218429; fax: +44-1298-218172; e-mail: john.cocker{at}hsl.gov.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This study reports a survey of occupational exposure to copper chrome arsenic (CCA) based wood preservatives during vacuum pressure timber impregnation. The survey involved biological monitoring based on analysis of chromium and arsenic in urine samples collected from UK workers. The aim of the study was to determine the extent of occupational exposure to arsenic and chromium in the UK timber treatment industry. The objectives were to collect and analyse urine samples from as many workers as possible, where CCA wood preservatives might be used, at 6 monthly intervals for 2 years. In addition, to investigate day-to-day variations in urinary excretion of chrome and arsenic by collecting and analysing three samples a week for 3 weeks in subsets of workers and controls (people not occupationally exposed). All urine samples were analysed for chromium and inorganic arsenic. To investigate any residual interference every sample was accompanied by a short questionnaire about recent consumption of seafood and smoking. The analytical methods for arsenic used a hydride generation technique to reduce interference from dietary sources of arsenic and also a technique that would measure total arsenic concentration in urine. The main findings show that workers exposed to CCA wood preservatives have concentrations of inorganic arsenic and chromium in urine that are significantly higher than those from non-occupationally exposed people but below biological monitoring guidance values that would indicate inhalation exposure at UK occupational exposure limits for chromium and arsenic. The effects of consumption of seafood on urinary arsenic were not significant using the hydride generation method for inorganic arsenic but were significant if ‘total’ arsenic was measured. The ‘total’ arsenic method could not distinguish CCA workers from controls and is clearly unsuitable for assessment of occupational exposure to arsenic. There was a significant increase in the urinary concentration of chromium in workers over the four sample collection rounds indicating increasing exposure to chromium during the 2 years of the study. This unexpected finding may be worth further investigation. Overall, the study demonstrated the utility of biological monitoring for assessment of occupational exposure to chromium and arsenic.
Keywords: biological monitoring • dermal exposure • exposure variability
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Copper chrome arsenic (CCA) wood preservatives contain arsenic pentoxide, hexavalent chromium (chromium trioxide or sodium dichromate) and copper (II) oxide or copper (II) sulphate. They are supplied as pastes or water-based concentrates that are diluted to between 1 and 10% w/w total salts and used in the industrial vacuum-pressure impregnation of timber. These products are used as a wood preservative to prevent fungal decay and infestations by wood-boring insects.
Wood preservatives such as CCA are regulated under the Control of Pesticides Regulations (COPR) 1986 and only approved products are allowed to be placed on the market. The Advisory Committee on Pesticides (ACP) considered the review on wood preservation uses of CCA at its meeting in September 1999. Recommendation for the continuing provisional approval of CCA products were agreed dependent on further data gathering including further information on biological monitoring. The purpose of the requirement for monitoring was to satisfy the ACP's concerns that concentration of arsenic in the urine of workers in wood preservative plants using CCA was not significantly elevated above background concentrations.
Operator exposure occurs during the handling of CCA treated timber and associated equipment contaminated with CCA (Garrod et al., 1999). Dermal exposure and ingestion were thought to be the main routes of absorption and exposure by inhalation was considered to be low. Biological monitoring can be used to assess human exposure to substances via all routes (skin, inhalation and ingestion) and offers a relatively simple and inexpensive way to assess occupational exposure. The manufacturers submitted biological monitoring data as part of the ACP review on CCA wood preservative products but comparisons between the data were difficult, because different manufacturers had used different analytical methods to measure different species (forms) of arsenic, and in many cases they had not measured urinary chromium. In 2004 the recommendation of both the American Conference of Governmental Industrial Hygienists (ACGIH, 2004) and the Deutsche Forschungsgemeinschaft (DFG, 2004) was that biological monitoring for inorganic arsenic should be based on inorganic arsenic and methylated metabolites released from that present in urine by hydride generation. This substantially reduces interferences from dietary sources of organic arsenic. In addition, modern analytical techniques now allow the separation and quantification of the different arsenic species (Francesconi and Kuehnelt, 2004).
This study sought to address the concerns of the ACP by generating the necessary biological monitoring data and investigating the usefulness of post-approval biomonitoring surveys in establishing the adequacy of exposure control; particularly where modelled human exposure assessments show some cause for concern. The data collected from the study could also be used to aid the proposal of biological monitoring guidance values (BMGV) for arsenic and chromium.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Study design
This study sought to gather exposure data based on urinary arsenic and chromium from as many firms and workers using CCA wood preservatives as possible. A list of all possible UK users of CCA products, compiled from manufacturer's lists, was supplied by HSE and managers were then contacted and asked for their help to identify and seek the participation of their workers. All the workers identified by their managers were contacted and asked to volunteer.
Because the levels found depend, in part, on the analytical methods used, this project also sought to determine levels of arsenic and chromium in urine of people not occupationally exposed with the same methods used for the exposed group. In addition, to investigate the influence of seafood on levels of arsenic, three samples a week for 3 weeks were collected from a subset of people occupationally exposed and non-occupationally exposed people to establish the variation found within individuals over time. A further subset of samples was also analysed by a method that gave ‘total’ arsenic (inorganic and organic) levels for comparison with historical data.
The study had the approval of HSE's Research Ethics committee (ETHCOM/REG/00/11) and sample collection was carried out entirely by post.
Procedures
Sample collection
Managers of each of the firms thought to be using CCA wood preservatives were contacted, given an explanation of the background to the work and then invited to help. They were asked to tell their workers about the study and supply names and contact addresses of potential volunteers who worked with CCA wood preservatives. The terms of approval from HSE's Research Ethics Committee and Survey Control allowed us to contact managers only twice so if they did not reply to the second letter they were not contacted again.
Volunteers identified by the managers were sent an information sheet, a consent form and a reply paid envelope and asked for their co-operation with the survey. The terms of approval from HSE's Research Ethics Committee allowed only one letter to volunteers.
Those workers who agreed to participate were sent urine sample bottles, instructions, reply-paid packaging and a short questionnaire about smoking and recent consumption of seafood. They were asked to collect a sample of urine at the end of a normal workshift towards the end of their work week and to post the sample back to the Health & Safety Laboratory. Each volunteer was sent their own results of the analysis in a letter with telephone contact points if they wanted to discuss them. To check for possible sample contamination during collection if an inorganic arsenic result exceeded 100 µmol mol–1 creatinine, the results' letter contained a follow-up sampling pack and an invitation to submit a further sample. The sampling round was repeated four times at roughly 6 monthly intervals.
At the end of the first year, some participants were asked if they would agree to take part in an additional study, collecting samples at the end of the work day on Monday, Wednesday and Friday over 3 weeks. Thirty participants took part in this ‘time-series’ aspect of the study from firms across the UK to investigate the variation of urinary values within subjects over the short term.
An unexposed population of HSE staff, from various regional offices, was also asked to provide samples at the end of the work day on Monday, Wednesday and Friday over 3 weeks. Samples arriving at the laboratory were allocated a unique reference number and stored below –18°C until analysed.
Sample analysis
All urine samples were analysed for chromium, inorganic (hydride generated) arsenic and creatinine. A subset of 57 samples was also analysed for ‘total’ urinary arsenic. The subset was chosen from the samples collected three times a week for 3 weeks provided by both the controls and the CCA workers. This was to investigate intra-individual variation and also the relationship between the inorganic values and the ‘total’ values found in older literature reports. Volunteers providing samples for this subset were also asked about seafood consumption.
Arsenic of inorganic origin (hydride generation)
All samples were analysed for inorganic arsenic using hydride-generation coupled with Inductively Coupled Plasma Mass Spectrometry (HG-ICP-MS) (Elan 6100 and FIAS 200, Perkin Elmer, Beaconsfield, UK).
This method reduces pentavalent arsenic (As5+), dimethylarsinate (DMA) and monomethylarsonate (MMA) (organic metabolites of inorganic arsenic species) to trivalent arsenic (As3+), using the reductants L-cysteine and concentrated hydrochloric acid. This trivalent arsenic is then reacted with sodium borohydride in a gas–liquid separator to produce the hydride, arsine gas (AsH3), which is measured directly by the ICP-MS (Morton, 2005; Tolodi, 2005).
Briefly, single arsenic standards were prepared from 100 mg l–1 arsenic standard (Fisher Chemicals, Loughborough, UK) so that the final concentrations were 10, 50, 100 µg l–1. All samples were prepared in 13 ml sample tubes (Sarstedt Numbrecht Germany,) To 1 ml of all samples, QC material and standards 1 ml L-cysteine (BDH Chemicals, Poole, UK) was added. After 15 min 1 ml of concentrated hydrochloric acid (Fisher Primer Grade, Loughborough) was added to these mixtures and the vial shaken on a vortex, to reduce the arsenic species. Samples, QCs and standards were diluted after a further period of 15 min by the addition of 7 ml deionized water. The reduced arsenic species were then mixed with 1.5% m/v sodium borohydride via a Perkin Elmer Flow Injection system (FIAS 200, Perkin Elmer, Beaconsfield, UK). The FIAS 200 gas–liquid separator was connected via a nebulizer gas line to the ICP-MS spray chamber, where all species with m/z 75 were detected. Arsenic compounds from dietary influences, e.g. arsenobetaine and arsenocholine are not detected by this method.
Certified reference material (Bio-Rad) was analysed at the start and end of every analytical run and after every 10 urine samples during the run. In addition samples from the TEQAS proficiency-testing scheme (University of Surrey, Guildford, UK) were also analysed for urinary arsenic. The limit for detection of arsenic by this method was 0.1 µg l–1, which is roughly equivalent to 0.15 µmol mol–1 creatinine (assuming a nominal creatinine concentration of 1 g l–1). The coefficient of variation for day-to-day analysis was 6% at 50 µg l–1.
Arsenic ‘total’
The method for the determination of ‘total’ arsenic used ICP-MS (X7 ICP-MS, Thermo Elemental, Winsford, UK) with collision cell technology and helium gas introduction into the argon plasma to remove argon chloride interference. This method is directly comparable with the older atomic absorption spectrophotometric (AAS) methods because it measures all the arsenic present in the urine sample as one entity. The urine samples, certified reference material (Bio-Rad) and standards (1–100 µg l–1) were simply diluted 1 in 10 with 1% v/v nitric acid and germanium (10 µg l–1) was used as an internal standard. 6 ml min–1 of 8% helium in hydrogen gas was used as the collision cell and the analysis was carried out using direct nebulization with a Meinhard nebulizer. Certified reference material (Bio-Rad) was again analysed in every analytical run at the start, end and after every 10 urine samples. The limit for detection of arsenic by this method was 0.1 µg l–1, which is roughly equivalent to 0.1 µmol mol–1 creatinine. The coefficient of variation for day-to-day analysis was 3% at 75 µg l–1.
Arsenic species
Urine samples (n = 50) that exhibited high inorganic arsenic levels were subjected to arsenic speciation [using liquid chromatography coupled with ICP-MS (LC-ICP-MS)] (Francesconi and Kuehnelt, 2004; Morton, 2006) to investigate the possibility of urine sample contamination during collection. CCA workers were exposed to As5+ in the CCA solution, which can then be metabolized (through methylation pathways hence DMA and MMA). If a urine sample had an As5+ component higher than 90% it was likely that this was a contaminated sample, since no metabolism had occurred. In the 749 CCA samples analysed only one showed an As5+ component that made up 98% of the total arsenic present. This sample was removed from the dataset. In the other samples (n = 49) where separate species were measured the average percentage of As5+ relative to the total arsenic species present was 16%.
Chromium
Chromium was determined by graphite furnace AAS (GF-AAS, Perkin-Elmer 5100). Calibration is by matrix matched additions calibration using spiked urine samples, and the urine samples and standards were diluted 1 + 1 with deionized water before analysis by GF-AAS. The chromium hollow cathode lamp was used at 357.9 nm and the graphite sample tubes were pyrolytically coated, 15 µl of sample was injected into the tubes. Peak area was used to determine the chromium absorption.
Certified reference material (Bio-Rad) was analysed in every analytical run at the start, end and after every ten urine samples. In addition HSL participates in the TEQAS proficiency-testing scheme (University of Surrey, Guildford, UK).
The limit for detection of chromium by this method was 0.1 µg l–1, which is roughly equivalent to 0.2 µmol mol–1 creatinine (assuming a nominal creatinine concentration of 1 g l–1). The coefficient of variation for day-to-day analysis was 5% at 4 µg l–1.
Creatinine
Creatinine was determined in all urine samples using an automated alkaline picrate method (Jaffe, 1886; Bonsnes and Toussky, 1945). Internal quality assurance samples were run after every 10 samples and external proficiency samples for creatinine from the Finnish Institute of Occupational Health were run at regular intervals. The concentrations of chromium and arsenic we divided by the concentration of creatinine to compensate for urine dilution effects.
Statistical analysis
The data were analysed using linear mixed-effects models (Pinheiro et al., 2000) in Splus. Such models are appropriate for describing the relationship between a response variable and covariates where there are repeated measurements on the same individual. In the context of this study fixed effects have been used to model the differences between controls and exposed workers, the possible contributions from smoking or dietary consumption of seafood, as well as possible temporal changes over the study duration. Random ‘worker’ effects model other systematic differences between individuals and yield separate estimates of within and between-person variability. As the raw urinary data are highly skewed the data were first log transformed to obtain an approximate normal distribution. The models can be represented in the following form:
 |
1, 2, 3 are fixed effects for no occupational exposure, smoking and seafood, respectively. The random effect for the ith individual µI is assumed to follow a normal distribution with zero mean and standard deviation 
B. The within person errors 
i,j are normally distributed with zero mean and standard deviation W. For the 3 x 3 time series data autocorrelation in the within-person errors was modelled using a continuous time AR(1) structure. Comparing models with and without autocorrelation using the likelihood ratio test assessed the statistical significance of autocorrelation. For the 6 monthly occupational data within-person errors were assumed to be independent and autocorrelation was not modelled.
Urinary results below the limit of detection (LOD) were set to one-half of the analytical LOD before applying a nominal creatinine correction. This gives a substituted value for chromium of 0.1 µmol mol–1. The robustness of the main findings was verified by alternatively substituting LOD/
and LOD (0.14 and 0.2 µmol mol–1, respectively). No arsenic values were below the LOD.
In the study protocol workers who submitted urine samples that gave unusually high arsenic levels (>100 µmol mol–1) were asked to submit a second sample. However, to use the second samples in place of the originals would risk introducing a negative bias into the dataset. Similarly, to include both samples in the analysis would risk a positive bias because of the tendency to request a second sample from workers with higher than average arsenic exposures. Instead, as noted above, only one abnormally high exposure was excluded on the basis of sound scientific evidence of sample contamination from arsenic speciation analysis and in all cases the first value was used.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Sample collection
Four hundred and fourteen site managers on the HSE list of companies potentially using CCA preservatives were asked for their help to identify and recruit volunteers from potentially exposed workers. Roughly half (212) did not respond even after a second letter and this may be owing, in part, to them no longer using CCA wood preservatives. Replies were received from 202 site managers and 163 provided the names of 440 potential volunteers. Thirty-nine firms did not wish to take part because; they no longer used CCA (25); were already doing biological monitoring (11); had taken part in previous studies (2); or exported CCA rather than using it on site (1).
Letters were sent to the 440 volunteers, identified by their managers, giving them information about the study and asking for their help. Two hundred and seventy-seven workers agreed to take part (63%): instructions, bottles and pre-paid packaging were sent to them. Samples were received from 217 volunteers in the first round (78% of those agreeing to take part or 49% of workers known to be potentially exposed to CCA). As the study progressed the number of samples in each round reduced, with 164, 124 and 93 in rounds 2, 3 and 4, respectively. Some companies stopped participating because they had ceased trading or were switching to alternative preservatives or the worker had left the company.
The selection of workers invited to send three samples a week for 3 weeks was based on asking volunteers from different companies spread across the UK who had sent in samples for the first two collection rounds. Samples were received from 19 workers in Scotland, Wales, Northern Ireland and over 12 counties in England. Eleven of these workers sent in a full set of 3 x 3 and a further six sent in over six samples each.
Urinary inorganic arsenic and chromium 3 x 3 series
The results of urinary inorganic arsenic and chromium analysis are summarized in Table 1.
|
A linear mixed-effects model (Cocker et al., 2006; supplementary material, Table 1) has been fitted for both workers and controls with additional fixed effects describing the possible contributions from smoking or consumption of seafood and random ‘worker’ effects. This model shows no statistically significant effects on urinary chromium levels from smoking or dietary consumption of seafood. Substantial differences (P-value < 0.0001) were observed between urinary levels of chromium in workers (geometric mean 1.48 µmol mol–1) and controls (geometric mean 0.31 µmol mol–1). There is considerable variation in urinary chromium levels, both within-person, (
W = 0.94) and between-person (B = 0.60). The estimated autocorrelation for urinary chromium samples taken from the same individual 24 h apart is 0.61 P-value < 0.0001, Supplementary Table 1).
Separate models can be fitted for workers and controls allowing each to have their own variance structure (Cocker et al., 2006; supplementary material, Table 1). These models show that between-person variability is larger for workers (B = 0.99) than controls (B = 0.22). Within-person variability is similar for each group (W = 0.88 and 0.98, respectively). Autocorrelation was the same for both groups: workers 0.60, P-value = 0.0003; controls 0.60, P-value = 0.0003.
In the case of urinary arsenic the mixed effects analysis again found a statistically significant difference (P-value = 0.0005) between urinary arsenic levels in controls (geometric mean 8.92 µmol mol–1) and workers (geometric mean 16.74 µmol mol–1). This analysis again allows for possible effects from smoking or seafood although neither had effects (P-values 0.12 and 0.54, respectively; Cocker et al., 2006; supplementary material, Table 1). Again there is considerable variation in urinary exposure levels with B = 0.54 and W = 0.51 but no statistically significant autocorrelation was uncovered. Supplementary Table 2 also contains separate models fitted for workers and controls that allow each to have their own variance structure. The exposed population exhibits more between-person variability (B = 0.63) than the non-exposed control population (B = 0.47).
Urinary arsenic and chromium in workers: 6 monthly sampling rounds
The results for all data in the main study are shown in Table 2. Because the data in the 3 x 3 time series showed no differences due to smoking or recent consumption of seafood the data are summarized without these distinctions.
|
There is highly significant (P < 0.0001) upward trend in chromium exposures over the four rounds of the study with geometric means of 0.59, 0.83, 1.37 and 1.64 µmol mol–1. The data are shown as a box and whisker plot in Fig. 1.
|
This trend is not mirrored in urinary arsenic exposures, which show no clear trend with time (Fig. 2).
|
In the 6 monthly survey the median urinary arsenic level in non-smoking workers is 17.1 µmol mol–1 and for smokers it is slightly higher at 21.4 µmol mol–1. Although modest, this is a statistically significant (P = 0.04) result and can be established because of the large dataset (598 samples). Within-person and between-person variation are of equal importance with
B = 0.54 and W = 0.51, a within-person geometric standard deviation of 1.80 and a between-person geometric standard deviation of 1.73.
In contrast for chromium neither smoking nor seafood has a statistically significant effect on urinary levels. The median urinary chromium level in workers is 1 µmol mol–1. Within-worker and between-worker variation are of equal importance with B = 1.01 and W = 0.97 (Table 3).
|
The urinary arsenic and chromium levels are consistent between the main study and the 3 x 3 time series study. Moreover, although the time intervals between repeat measurements are very different between the main study (repeats every 6 months) and the 3 x 3 studies (repeats only days apart) the estimates of within-person variability are reasonably consistent [see Table 5, Cocker et al. (2006) and supplementary material, Tables 1 and 2] and so it is reasonable to pool the data (Table 2).
|
Total arsenic analysis
The mixed effects model for total arsenic analysis (Table 5) shows no difference between controls and workers but a very large contribution from dietary exposure to arsenic in seafood (P-value 0.0002).
A comparison of total arsenic results with their corresponding values from the inorganic (hydride generation method) analysis shows that the ratio of total arsenic to inorganic arsenic ranged from 0.9 to 465 (median 5.0, mean 15.6 µmol/mol creatinine).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The mailing list used to seek the help of site managers was based on customer lists supplied by manufacturers. Given that these were somewhat dated the study had a surprisingly good response. Roughly half the managers replied with the names of their workers. There was also a good response from the workers with 63% initially agreeing to take part and 49% actually sending samples for the first round. Although the 93 people sending samples for the last round of collection are only 21% of the population potentially exposed at the start of the study, they probably represent a larger percentage of those actually exposed at the end of the study, as firms stopped using CCA based preservatives. Given that no financial inducements were made to workers to participate in the study this level of participation is remarkable and reflects both the altruism of the workers and the ease of use of biological monitoring as a survey tool.
Although the levels of urinary chromium and arsenic are significantly higher in occupationally exposed workers compared with controls, the levels found in this study are generally within the BMGV produced by the American Conference of Governmental Industrial Hygienists (ACGIH, 2004) and German MAK Commission (DFG, 2004).
Both the ACGIH and MAK guidance values are based on published studies to describe the relationship between the concentrations of arsenic and chromium in urine and the concentrations of arsenic and chromium in air samples collected near the workers breathing zone.
In the case of arsenic the ACGIH biological exposure index (BEI) value is based on the level of arsenic likely to be found in workers' urine after an inhalation exposure for 8 h at the US threshold limit value of 0.01 mg m–3. The value is 
50 µmol mol–1 and is also based on association with a standard mortality ratio value of 100 (background level) for lung cancer. The German Expositionsaquivalente fur krebserzeugende Arbeitsstoffe (EKA) value is derived from the relationship between inhalation exposure and urinary arsenic. The level of arsenic expected in urine after 8 h inhalation at the UK Workplace Exposure Limit (WEL) of 0.1 mg m–3 is 200 µmol mol–1 creatinine.
The time series data showed a statistically significant difference between levels of arsenic and chromium in workers compared with non-exposed controls. Workers and controls have different variance structures with the workers exhibiting larger between-person variability where as for non-exposed controls day-to-day variability is the larger component. Intuitively this makes sense, as differences in long-term average exposures for controls are only due to smoking and dietary differences while differences in working practices, exposure controls and frequency of work with CCA are all additional sources of between-person variability for occupationally exposed persons. Both workers and controls exhibited autocorrelation in their urinary chromium values but not their arsenic results. The reasons for this are unclear but a possible explanation might be offered by their differing biological half-lives.
Knowledge of within-person and between-person variability influences the interpretation of biological monitoring data. For example, for an ‘average worker’ with a personal median arsenic level of 17.1 µmol mol–1, 95% of their individual samples will lie in the range 5.9–49.9 µmol mol–1. The probability of this individual providing a sample greater than the BEI (53 µmol mol–1) is 0.02. For this individual it is estimated that only four samples in a million would exceed the EKA (196 µmol mol–1). In contrast, a highly exposed worker at the 95th percentile of the distribution of median exposures has a personal median exposure to arsenic of 45.0 µmol mol–1. For this worker 95% of individual samples will lie in the range 15.4–131.1 µmol mol–1 with the probability of being greater than the BEI equalling 0.38 (probability > EKA = 0.004). For a non-smoking control at the 95th percentile of the normal population (personal median = 14.8 µmol mol–1) the probability of a urinary arsenic result above the BEI is only 0.01 with only one in a million samples >EKA. A figure showing the probability of exceeding the BEI for an individual in the CCA workforce is given in Cocker et al. (2006), supplementary material, Fig. 3. Knowledge of the likelihood of obtaining a high urinary arsenic exposure can be informative in making risk management decisions: a worker providing a single urine result in excess of the BEI would not be unusual; two consecutive samples above the BEI probably indicate a worker who is habitually over exposed.
In the case of chromium, the guidance value from the ACGIH BEI (25 µg l–1 or 54 µmol mol–1) has been derived mostly from data on manual metal arc welding and the relationship between urinary chromium and inhalation of hexavalent chromium in water-soluble fume. The German EKA value for urinary chromium after exposure for 8 h at the UK WEL for hexavalent chromium of 0.05 mg m– 3 is 20 µg l–1 (or 43 µmol mol–1) and is not a health-based limit. The value adopted as guidance in the UK is 10 µmol mol–1 creatinine and is based on a survey of occupational exposure in workplaces with good control of exposure to chromium (excluding the CCA data from this study) and is thus a value associated with good occupational hygiene practice rather than a health-based value.
Considering the within-person and between-person variability for chromium the probability of urinary chromium exceeding the UK BMGV can be estimated. A ‘highly’ exposed worker at the 95th percentile of the distribution of median exposures has a personal median exposure to chromium of 4.4 µmol mol–1. For this worker, the probability of being greater than the BMGV is 0.21. A non-smoking control at the 95th percentile of the normal population (personal median = 0.47 µmol mol–1) has the probability of a urinary chromium result above the BMGV of only 0.0005. A figure showing the probability of exceeding the BMGV within the CCA workforce is given in Cocker et al. (2006), supplementary material, Fig. 4.
The analysis for the four sampling rounds showed a surprising and statistically significant upward trend in levels of urinary chromium but not in urinary arsenic. Although this trend has been identified through an analysis of the complete 6 monthly data an almost identical trend exists in the 91 workers who provided samples in all four rounds. This suggests that this trend is not an artefact of the sample collection process. Furthermore, this trend has been identified in the median exposure levels and, therefore, is unlikely to be due to a small number of high chromium results in the later rounds. Table 2 reveals a dramatic reduction in the number of samples below the LOD between the first and second years of the study. There was no change in the analytical method during this period and the increase in number of samples with detectable chromium suggests the trend might be more pronounced at the lower range of the exposure distribution. The differing trends of the two analytes suggest that the increasing levels of urinary chromium may not be due to exposure to CCA wood preservatives. Some site managers said they were no longer participating because they had switched to other wood preservatives. If other sites had also switched to preservatives containing chromium (but not arsenic) and continued to send samples this might explain the increase in urinary chromium in the later rounds. It is not possible to confirm this or whether the increased exposure is due to higher levels of chromium or workers taking less care about occupational hygiene because of a perceived lower toxicity of the new preservative. In view of the carcinogenic properties of chromium VI solutions further surveillance may be prudent. However, it should be noted that chromium has not been supported under the Biocidal Products Directive as an active substance and so any biocidal product containing hexavalent chromium will have to be removed from the European Union market by 1 September 2006. This may change if companies can prove that chromium is not an active substance but acts as a fixative in the formulations.
Within the time series data, consumption of seafood and smoking had no statistically significant effects. But within the larger 6 monthly occupational study there was a small but significant effect on arsenic results due to smoking with median values of 17.1 and 21.4 µmol mol–1 for non-smokers and smokers, respectively. There was a slight increase in urinary chromium from a median of 0.84 µmol mol–1 in non-smokers to 0.9 µmol mol–1 in smokers but this was not significant. One disadvantage of a linear model for the log transformed urinary concentrations model is that it implies a common multiplicative effect for smoking and seafood for both workers and controls. Intuitively, a common additive effect for smoking and dietary consumption of seafood seems more realistic. As their effects are either statistically insignificant or not of a meaningful magnitude this is a minor consideration.
The effects of dietary consumption of seafood on urinary ‘total’ arsenic were so pronounced that it was not possible to separate exposed workers from controls using ‘total’ arsenic in urine. There was no correlation between ‘total’ arsenic and inorganic arsenic. The ratio of ‘total’ to inorganic varies from 0.9 to 465 µmol mol–1 with a mean of 15.6 µmol mol–1 and a median of 5 µmol mol–1. This means that a simple factor to compare historical ‘total’ results with the present study is unfeasible.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Overall, the study demonstrated the utility of biological monitoring for assessment of occupational exposure to chromium and arsenic. This study has clearly shown that workers exposed to CCA wood preservatives have urinary levels of inorganic arsenic and chromium that are significantly higher than those in urine samples from people not occupationally exposed. The effect of consumption of seafood and smoking on urinary levels of inorganic arsenic (using the hydride generation method) and chromium are small and do not interfere with the assessment of occupational exposure. However, the ‘total’ arsenic method could not distinguish CCA workers from controls and is clearly unsuitable for assessment of occupational exposure to arsenic.
Workers exposed to CCA wood preservatives have concentrations of inorganic arsenic and chromium in urine that are significantly higher than those from non-occupationally exposed people but below BMGV that would indicate inhalation exposure at UK occupational exposure limits for hexavalent chromium and arsenic.
There was a significant increase in the urinary concentration of chromium but not inorganic arsenic in workers over the four sample collection rounds potentially indicating increasing exposure to chromium during the 2 years of the study. This unexpected finding may merit further investigation.
There were considerable within-worker and between-worker variability in urinary chromium and arsenic levels but this is comparable with variability in between-worker measured dermal and inhalation exposure. Multiple samples are required to fully determine an individual's exposure levels (by air, dermal or biological monitoring).
|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The authors would like to thank the site managers and volunteers who helped and participated in this study.
Received October 18, 2005; in final form January 26, 2006
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
ACGIH. (2004) Threshold limit values for chemical substances and physical agents and Biological exposure indices. (American Conference of Governmental Industrial Hygienists, Cincinnati, OH) ISBN 1-882417-54-2.
Bonsnes RW and Toussky HH. (1945) On the colourimetric determination of creatinine by the Jaffe reaction. J Biol Chem 158:581–7.
Cocker J, Morton J, Warren N, et al. (2006) Biomonitoring for chromium and arsenic in timber treatment plant workers exposed to CCA wood preservatives. Ann Occup Hyg supplementary material available online at www.annhyg.oxfordjournals.org.
DFG. (2004) Deutsche Forschungsgemeinschaft List of MAK and BAT values 2003 published by the Commission for the investigation of health hazards of chemicals in the work area 39. (Wiley-VCH, Weinheim, FRG) ISBN 3-527-27517-7.
Francesconi K and Kuehnelt D. (2004) Determination of arsenic species: A critical review of methods and applications, 2000–2003. Analyst 129:373–95.[CrossRef][Medline]
Garrod ANI, Martinez M, Pearson J, et al. (1999) Exposure to preservatives used in the industrial pre-treatment of timber. Ann Occup Hyg 43:543–55.
Jaffe M. (1886) Uber den niederschlag, welchen pikriksaure in normalem harn erzeugt und urber eine neue reaction des kreatinines Z. Physiol Chem 10:391–400.
Morton J. (2005) Chapter 9.1 ‘Clinical applications of ICP-MS’. In Nelms Simon (Ed.). ICP Mass Spectrometry handbook (Blackwell Publishing Ltd, Oxford, UK).
Morton J and Mason HJ. (2006) Speciation of arsenic compounds in urine from occupationally unexposed and exposed persons in the UK using a routine LC-ICP-MS method. J Anal Toxicol in press.
Pinheiro JC and Bates DM. (2000) Mixed effects models in S and S-plus. (Springer-Verlag, New York).
Todoli JL. (2005) Chapter 5.2 ‘Flow injection sample introduction for ICP-MS’. In Nelms Simon (Ed.). ICP-MS spectrometry handbook (Blackwell Publishing Ltd, Oxford, UK).
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||