Annals of Occupational Hygiene Advance Access originally published online on October 29, 2004
Annals of Occupational Hygiene 2004 48(8):683-695; doi:10.1093/annhyg/meh056
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© British Occupational Hygiene Society Published by Oxford University Press;
Comparative Analysis of Passive Dosimetry and Biomonitoring for Assessing Chlorpyrifos Exposure in Pesticide Workers
1 Johns Hopkins Bloomberg School of Public Health, Department of Environmental Health Sciences, 615 N. Wolfe St, Baltimore, MD 21205, USA; 2 US Environmental Protection Agency, National Center for Environmental Assessment (8623D), 1200 Pennsylvania Ave, NW, Washington, DC 20460, USA; 3 US Environmental Protection Agency, Office of Pesticide Programs (7509C), 1200 Pennsylvania Ave, NW, Washington, DC 20460, USA
* Author to whom correspondence should be addressed. e-mail: tbuckley{at}jhsph.edu
Received 7 May 2004; in final form 10 May 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), the US Environmental Protection Agency (EPA) has the authority to regulate the use of pesticides to prevent unreasonable adverse human health effects associated with pesticide exposure. Accordingly, the EPA requires pesticide registrants to perform studies evaluating the potential for pesticide handler exposure. Data from five such studies that included exposure measurements based on both external measurements and biological monitoring were used to examine methods of assessment, routes and determinants of exposure and dose to the pesticide chlorpyrifos. Eighty workers across four job classes were included: mixer/loaders (M/L, n = 24), mixer/loader/applicators (M/L/A, n = 37), applicators (A, n = 9) and re-entry scouts (RS, n = 10). Results showed that doses were highly variable and differed by job class (P < 0.05) with median total (inhalation and dermal combined) exposure-derived absorbed doses (EDADtot) of 129, 88, 85 and 45 µg/application for A, M/L/A, M/L and RS, respectively. Doses derived from the measurement of 3,5,6-trichloro- 2-pyridinol (3,5,6-TCP) in urine were similar in magnitude but differed in rank with median values of 275, 189, 122 and 97 µg/application for A, M/L, RS, and M/L/A, respectively. The relative contribution of dermal to inhalation exposure was examined by their ratio. The median ratios of exposure-derived absorbed dermal dose (EDADderm) (assuming 3% absorption) to exposure-derived absorbed inhalation dose (EDADinh) (assuming 100% absorption) across job classes were 1.7, 1.5, 0.44 and 0.18 for RS, M/L, A and M/L/A, respectively, with an overall median of 0.6. For 34 of 77 workers (44%), this ratio exceeded 1.0, indicating the significance of the dermal exposure pathway. Different dermal absorption factor (DAF) assumptions were examined by comparing EDADtot to the biomarker-derived absorbed dose (BDAD) as a ratio where EDADtot was calculated assuming a DAF of 1, 3 and 10%. Median ratios of 0.45, 0.71 and 1.28, respectively, were determined suggesting the DAF is within the range of 3–10%. A simple linear regression of urinary 3,5,6-TCP against EDADtot indicates a positive association explaining 29% of the variability in the 3,5,6-TCP derived estimate of dose. A multiple linear regression model including the variables EDADderm, EDADinh and application type explained 46% of the variability (R2 = 0.46) in the urinary dose estimate. EDADderm was marginally significant (P = 0.066) while EDADinh was not (P = 0.57). The EDADderm regression coefficient (0.0007) exceeded the coefficient for EDADinh (0.00002) by a factor of 35. This study demonstrates the value of the pesticide registrant database for the purpose of evaluating pesticide worker exposure. It highlights the significance of the dermal exposure pathway, and identifies the need for methods and research to close the gap between external and internal exposure measures.
Keywords: chlorpyrifos • dermal • exposure assessment • whole body dosimetry • biological monitoring • inhalation • pesticides
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Public health concern about occupational exposure to pesticides began after the introduction of the acutely toxic organophosphate insecticides in the late 1940s. Since then, many additional incidents of pesticide poisoning have been observed in agricultural workers [Mehler et al., 1992
; Center for Disease Control and Prevention (CDC, 1998); Das et al., 2001]. Concerns about the potential hazards to farm workers remain today. The Environmental Protection Agency (EPA) estimates that 
10 000–20 000 (0.8%) of the 2.25 million agricultural workers in the United States suffer from acute pesticide poisoning annually (USEPA, 1992a,b). Others have reported as many as 300 000 acute illnesses and injuries of farm workers due to pesticide exposures (USGAO, 1993). According to the US Bureau of Labor Statistics, farm workers experience the highest rate of chemical-related illness of any occupational group (5.5 cases per 1000 workers) (US Bureau of Labor Statistics, 1987). According to the California Pesticide Illness Surveillance Program, 914, 804 and 656 cases of pesticide illness or injury associated with occupational exposure occurred in the years 1998, 1999 and 2000, respectively (California EPA, 2000, 2001, 2002). Pesticide illness and injury surveillance programs in Florida, Texas and New York reported 171, 154 and 16 cases of acute occupational pesticide-related illness in 1999, respectively (NIOSH, 2000).
The widespread health impacts of pesticide use prompted the promulgation of the Worker Protection Standard of the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) in 1992 [Code of Federal Regulations (CFR, 2002)] to protect agricultural workers. Under this Act, the EPA is charged with reducing the risks of illness or injury resulting from occupational exposures to pesticides. To meet this provision of FIFRA, the EPA has implemented a registration process that requires pesticide manufacturers (registrants) to submit studies evaluating occupational exposures to pesticide handlers. Information gathered from these studies such as pesticide formulation, pesticide application rate, and exposure estimates is compiled in the EPA's Pesticide Registrant Database. This database provides a wealth of information by which to explore the reliability and reproducibility of exposure assessment measurements. Registration requirements and recommendations for exposure monitoring can be found in the Office of Prevention, Pesticides, and Toxic Substances' harmonized guidelines 875 Series, Occupational and Residential Exposure Test Guidelines (USEPA, 1998). These guidelines include specific methods for passive dosimetry and biological monitoring to assess worker exposure so that there is general uniformity across studies.
The assessment of dermal exposure to pesticides in workers is both complex and important. A number of studies have established that the dermal route can contribute as much or more to a worker's exposure than inhalation (Wolfe, 1976; Fenske et al., 1989; Fenske and Elkner, 1990; Woollen, 1993; Cattani et al., 2001). Although methods for evaluating inhalation exposure including the use of personal air monitoring are reasonably well standardized, such is not the case for dermal exposure, where there is a lack of consensus regarding a wide range of methods with differing underlying assumptions (Fenske, 1993; Hendersen et al., 1993; Krieger, 1995; USEPA, 1998). The EPA recommends that registrant studies be conducted using whole body dosimetry (WBD) for assessing pesticide worker dermal exposure. Although there are recognized limitations to this methodology, including its adequacy to serve as a skin surrogate and assumed uniform absorption across body regions, it is thought to provide the best external means for assessing dermal exposure (Fenske, 1993; Ness, 1994; USEPA, 1998). In light of the limitations, the EPA recommends that biological monitoring be used in combination with WBD to provide a complimentary assessment of exposure. Biological monitoring captures the chlorpyrifos pesticide penetrating the dosimeter and absorbed through the skin, and that which is inhaled (including the penetration of respiratory protection), as well as non-dietary ingestion. The comparison of exposure with biological monitoring is most informative when parameters of human metabolism and pharmacokinetics are characterized and available (Chester, 1993; Woollen, 1993; Krieger, 1995).
Chlorpyrifos has been selected for a case study because of its prevalent use (USEPA, 1988, 2000; Albers et al., 1999), toxicity (USEPA, 1988, 2000; Albers et al., 1999), and quality and completeness of data among the registrant studies. Chlorpyrifos is extensively used in agriculture and is also applied to pastures and woodlands. An estimated 20–24 million pounds of chlorpyrifos are expected to be applied annually (USEPA, 2000). Chlorpyrifos is a cholinesterase inhibitor associated with many neurophysiological effects such as nausea, dizziness and confusion, and, at high exposures, respiratory paralysis and death (USEPA, 2000). Because of toxicity and exposure concerns with residential use, the EPA negotiated with manufacturers to voluntary phase-out its residential use in the year 2000. The EPA also lowered tolerances on certain crops, such as apples and tomatoes (USEPA, 2000).
A number of pesticide registrant studies have been conducted employing a combination of methodologies including WBD and biological monitoring. These studies may hold valuable data to estimate the importance of specific exposure routes and determinants and to evaluate the performance of assessment methods. Accordingly, the goals of the current study are to: (i) demonstrate the research value and utility of the registrant database for investigating worker exposure; (ii) characterize worker chlorpyrifos exposure and examine the relative contribution of dermal and inhalation routes; (iii) identify exposure determinants; and (iv) provide an estimate of the dermal absorption factor. As a result, this study will provide an improved understanding of worker exposure and methods of assessment as a fundamental basis for health protection.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Among the approximately 100 registrant studies on file with the EPA, five were selected for analysis (Table 1). These studies were subject to EPA's internal review requirements including an approved human studies protocol and informed consent (USEPA, 1991
). Studies were selected based on the following criteria: (i) workers were exposed to the pesticide chlorpyrifos; (ii) exposure was assessed based on the EPA's 1998 Guidelines; (iii) biological monitoring was included; and (iv) results and methods were complete and sufficiently detailed. Individual exposure data as well as information regarding methodology of the studies were extracted, pooled, and entered into ExcelTM spreadsheets. Four different job categories were represented across the five studies including: mixer/loaders (M/L), applicators (A), re-entry scouts (RS) and mixer/loader/applicators (M/L/A). The M/L is responsible for preparing the pesticide and loading it into the application equipment, while the Applicator applies the pesticide. The M/L/A performs both job functions. The RS enters the field following pesticide application to assess the effectiveness of the treatment. EPA's exposure assessment guidelines (USEPA, 1998) were followed for each of the five studies. The number of workers monitored per study ranged from 5 to 28. Each worker was monitored once over a period of time representing a standard work shift or duration of application.
|
Inhalation exposure was evaluated by personal monitoring capturing both gas and particle phase chlorpyrifos in the workers' breathing zone. Samples were collected on a 37 mm mixed cellulose ester filter (0.8 µm pore size) (Gelman Sciences, Ann Arbor, MI) placed upstream to a 99 mg ChromosorbTM 102 adsorbent tube (SKC, Inc., Eighty Four, PA) and connected to the tube by a glass funnel and rubber tubing. The filter cassette and tube assembly were connected by flexible TygonTM tubing to a battery-powered personal sampling pump (SKC, Inc.,) operating at a flow rate of 1–2 l/min that was pre- and post-calibrated with a rotameter. Sample durations ranged from 0.5 to 9 h. Sample volumes ranged from 0.0209 to 0.788 m3. The front and back sections of the tubes were extracted separately. Samples were analyzed by gas chromatography combined with an electron capture detector.
Dermal exposure was assessed using WBD including handwash samples in all five studies. The dosimeter consisted of underclothing in the form of a short-sleeved T-shirt and briefs. All workers wore cotton coveralls. Six wore additional full rain gear, while 10 workers had short-sleeved coveralls rather than long. In all but one study, a ball cap was worn as a dosimeter to estimate exposure to the neck and head. In the study by Contardi et al. (1996), either a cap or two headbands (one on the head, one on the neck) was used to estimate dermal exposure to the head. The coveralls and underclothing were cut into pieces representing potential exposure to body regions such as the arms, legs, and front and back torso. Chlorpyrifos was extracted from the clothing. The amount of chlorpyrifos found on the underclothing relative to the amount on the coveralls was used to estimate individual site-specific penetration factors (PFs), i.e. the percentage of chlorpyrifos passing through the coveralls to the skin. These penetration factors were used to estimate dermal exposure to arm and leg regions where no undergarment was present.
Handwash samples were collected over the monitoring period during times when workers would typically wash their hands (i.e. before meals, before smoking, after using the bathroom and at the end of the monitoring period) to assess surface deposition on the hands. Handwash samples were also collected between tasks to note differences in task contribution. The amount of chlorpyrifos found on the hands was then added to the amount measured with WBD to obtain the total dermal exposure (TDE) (µg chlorpyrifos). A dermal absorption factor (DAF) of 3% was applied to the TDE to obtain an exposure-derived absorbed dermal dose (EDADderm) value (µg chlorpyrifos), as is shown in equation (1). The DAF is 3%, based on human studies by Nolan et al. (1994) and is the value accepted by the EPA.
 | (1) |
The TDE is calculated in equation (2):
 | (2) |
The exposure-derived absorbed inhalation dose (EDADinh) was estimated from equation (3). A 100% absorption factor (AF) was assumed for workers not wearing respiratory protection, whereas a 10% AF was applied to estimates for workers wearing respiratory protection (CFR, 2003). An inhalation rate of 1.5 m3/h was assumed for all workers (USEPA, 1997).
 | (3) |
The total exposure-derived absorbed dose (EDADtot) (µg chlorpyrifos) from passive dosimetry was calculated by summing the EDADderm (µg chlorpyrifos) and the EDADinh (µg chlorpyrifos) per equation (4), where each of the variables is as previously defined.
 | (4) |
Biological monitoring of the primary chlorpyrifos urinary metabolite 3,5,6-trichloro-2-pyridinol (3,5,6-TCP) was also used to measure the absorbed dose of chlorpyrifos. The biological monitoring was conducted concurrently with the passive dosimetry. Workers were asked to avoid exposure to chlorpyrifos approximately seven to ten days prior to the exposure and urine collection period. In several instances, surrogate workers (n = 16) who performed the tasks of the agricultural worker were monitored instead of actual workers due to difficulties with limiting worker exposure for the required seven to ten day pre- sampling clearance period. Urine samples were collected one day prior to exposure (background), the day of exposure, and four days following exposure. Cumulative urine collected over each twelve-h interval was pooled into one sample for analysis. Laboratory analysis was conducted according to methods described by Bartels and Kastl (1992). In brief, this method consisted of acid hydrolysis, liquid extraction (toluene), and derivitization with N-methyl-bis-(tert-butyldimethylsilyl)-trifluoroacetamide (MTBSTFA). The derivitized extract was quantified with negative ion gas chromatography/mass spectrometry (GC/MS) using response factors given from internal standards.
Biomarker-derived absorbed dose (BDAD) was calculated in the same manner in all five studies. Total 3,5,6-TCP excreted was corrected for background exposure, divided by a unitless urinary excretion factor (UEF) of 0.72, and then multiplied by the ratio of the molecular weight of chlorpyrifos (350.6 g/mol) to 3,5,6-TCP (198 g/mol) per equation (5). The UEF was derived from human kinetic data and represents the fraction of an oral dose excreted in urine (Nolan et al., 1984).
 | (5) |
In our analysis, the raw data from the studies were analyzed using Intercooled Stata® v8.0. The dermal, inhalation, and biomonitoring variables were found to be non-normally distributed by the Shapiro-Wilk test. Therefore, to assess the relative contribution of dermal and inhalation routes to total exposure and to compare exposure and dose across job class, non-parametric methods were used. The relative contribution of EDADinh and EDADderm to predicted EDADtot was examined using the Wilcoxon Sign-Rank Test. The Wilcoxon Rank-Sum Test was performed to compare EDADinh, EDADderm and BDAD across job categories. To evaluate the relationship between passive dosimetry results and biomonitoring levels and to identify significant predictors of absorbed dose, simple and multiple linear regression analyses were conducted. BDAD was log-transformed to meet the assumption of normality in the dependent variable.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Across the five registrant studies, there was a total of 80 workers monitored representing a range of formulations, applications, and use of personal protective equipment (PPE) as summarized in Table 1. Inhalation data were missing for three workers, therefore any analyses involving inhalation values included 77 workers.
The distributions of EDADinh (assuming 100% absorption and 10% absorption for workers wearing respirators) (n = 77) and EDADderm (assuming a 3% DAF) (n = 80) by job class are shown in Fig. 1. EDADinh spanned six orders of magnitude, ranging from 0.064 to 32 400 µg/day. Based on a test of medians, inhalation doses among the job classes were found to be different (P = 0.02). Applicators tended to be most highly exposed (median of 74 µg/day, n = 9) followed by M/L (n = 21), M/L/A (n = 37) and RS (n = 10) with median values of 44, 30, and 20 µg/day, respectively. Specific differences between job classes were examined using the Wilcoxon Rank-Sum Test, revealing A to be more highly exposed by inhalation relative to RS (P = 0.03).
|
Based on EPA's guidance and evidence by Nolan et al. (1984)
, we assumed a 3% DAF to calculate potential dermal chlorpyrifos dose (USEPA, 1998). EDADderm estimates assuming 3% DAF ranged from 0.77 to 2340 µg/day. As with inhalation, the difference in EDADderm among job classes was significant (median test; P = 0.001). The Wilcoxon Rank-Sum Test revealed M/L (P = 0.0001, n = 21), A (P = 0.027, n = 9), and RS (P = 0.024, n = 10) were all more highly exposed relative to M/L/A (n = 37). The comparison of EDADderm to EDADinh (assuming 3% dermal and 100% or 10% inhalation absorption) provides an indication of their relative contribution (Fig. 1). The Wilcoxon Sign-Rank Test showed that EDADinh was higher than EDADderm for two job classes: M/L/A (n = 37, P = 0.03), A (n = 9, P = 0.03). In contrast, dermal exceeded inhalation dose for RS (n = 10, P = 0.005). The relative dose contribution was examined as the ratio (EDADderm:EDADinh) by job class (Table 2). Re-entry scouts had the highest median ratio of 1.66 while M/L/A had the lowest ratio of 0.17, with an overall median ratio across the four job categories of 0.49. Ratios were highly skewed (especially M/L) as indicated by a large difference between mean and median values. In 34 of 77 workers (44%) the ratio exceeded one indicating that dermal exposure dominated inhalation.
|
EDADtot (assuming 3% dermal and 100% inhalation absorption if no respiratory protection worn and 10% if a respirator worn) ranged from 5.5 to 32 445 µg/day with A most highly exposed (median 129 µg/day) followed by M/L/A, M/L and RS, with median values of 88, 85 and 45 µg/day, respectively (n = 77). Significant differences in EDADtot by job class were not detected in a global median test (P = 0.18). However, the Wilcoxon Rank-Sum Test revealed statistically significantly higher EDADtot levels in M/L relative to RS (P = 0.007).
The variability in exposure by method of application (planters, handheld sprayers, aerial sprayers and ground boom) was also examined (see Table 3). Workers performing ground boom applications had the highest median EDADderm of 82.7 µg/event (n = 18), followed by aerial applications (39.0 µg/event, n = 15), handheld spraying applications (33.0 µg/event, n = 21) and planter applications (1.9 µg/event, n = 16). A global median test evaluating the differences in EDADderm by application type yielded a P-value of < 0.001. Specifically, planters were less than handheld sprayers (P = 0.0001), aerial sprayers (P < 0.0001) and ground boom applicators (P < 0.0001). The difference between handheld spray and ground boom applicators was marginal (P = 0.055), as was the difference between aerial spray and ground boom (P = 0.06).
|
The EDADinh distribution by method of application differed from EDADderm. For EDADinh, handheld sprayers had the highest median level followed by ground boom, aerial sprayer and planters. As with EDADderm, planters EDADinh was less than handheld sprayers (P = 0.0003), aerial sprayers (P = 0.003) and, marginally, groundboom applications (P = 0.06). Furthermore, EDADinh by handheld spraying was greater than aerial spaying (P = 0.04).
The chlorpyrifos BDAD estimated from urinary 3,5,6-TCP ranged from –87 to 2681 µg/day (negative values indicate higher pre-application levels relative to post-application), with a median of 163 µg/day (n = 80). Applicators had the highest median BDAD (275 µg/application), followed by M/L, RS and M/L/A, with median dose values of 189, 122 and 97 µg/application, respectively (Fig. 2). Differences in BDAD by job class were not statistically significant in a global median test (P = 0.24). However, results of the Wilcoxon Rank-Sum Test revealed that M/L had statistically significantly higher absorbed dose levels than M/L/A (P = 0.04), and A had higher dose levels than M/L/A (P = 0.02). Applicator BDAD levels were marginally significantly higher relative to RS (P = 0.06).
|
EDADtot estimated from external passive dosimetry assuming three DAFs (1, 3 and 10%) was compared with BDAD. Ratios of the external estimates to the biomarker values by job classification revealed that a 3% dermal absorption factor (DAF), the EPA's accepted DAF, resulted in a median ratio of 0.71, as compared with a 1% DAF (median ratio = 0.45) and 10% DAF (median ratio = 1.28) (Fig. 3).
|
Regression analysis was used to examine determinants of BDAD considering independent variables, including: (i) EDADinh (assuming 100% and 10% AF for workers without and with respirators, respectively); (ii) EDADderm (assuming 3% DAF, including and excluding hands and hands by themselves); (iii) EDADtot; (iv) formulation type; (v) pounds of pesticide handled; (vi) application type; (vii) job class; (viii) personal protective equipment (PPE) type; (ix) sample duration; (x) bodyweight; and (xi) % formulation. BDAD was log-transformed due to its skewed distribution. EDADtot explained 29% of the variability in BDAD from simple linear regression analysis (Fig. 4). A more complex multivariate regression model was considered to explain additional variability. Both the inhalation and dermal doses were included a priori. In order to be considered for inclusion in the multivariate model, the additional variables needed to achieve a P-value of <0.2 in the simple linear regression model (Table 4). Variables satisfying this criterion included: pounds chlorpyrifos handled, application type, % chlorpyrifos formulation, PPE type, job type, formulation type and sample duration.
|
|
|
The multiple linear regression model was built in a stepwise fashion using the variables identified as potentially important from the simple linear regression results. Due to the suspected correlation among some of the independent variables (e.g. application type and PPE use), several different models were developed with different combinations and ordering of variables. Interaction among variables was also explored. We selected the variables for the final model based on the following criteria: statistical significance, strength of association, parsimony, and the amount of variability explained by the model (R2). This model (equation 6) includes the variables EDADderm, EDADinh, and application type (i.e. handwand, groundboom, aerial or corn planter). The application type variables were treated as indicator variables in our model, with corn planters serving as the reference group. In this final model, EDADderm was marginally statistically significant (P = 0.066) whereas EDADinh was not (P = 0.59). The regression coefficient for EDADderm (0.0007) exceeded EDADinh (0.00002) by a factor of 35. Overall, the final model included a total of three independent variables (including the two dose variables) and explained 46% of the variability in the log-transformed urinary 3,5, 6-TCP (BDAD) (n = 62) (Table 4). The RS were dropped from this regression model because they could not be assigned to an application type, as well as five workers missing 3,5,6-TCP data, and three workers missing inhalation data.
 | (6) |
i is the model error. |
DISCUSSION AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Exposure assessment is crucial to evaluating the hazard posed to agricultural pesticide handlers. The current study provides such an assessment across a range of studies and job classes. Exposure has been assessed based both on external measurements and biological monitoring, affording the opportunity to not only evaluate the relative routes of exposure but also to examine the predictive association between the two methods of assessment. Exposure was examined by job class as suggested by Henderson et al. (1993)
.
Among the four job classes investigated, A were consistently ranked as the most highly exposed based on median values of both EDADtot and BDAD. This finding agrees with Cattani et al. (2001), who attributed higher exposures among pesticide applicators to frequent spills on the body and improper use and poor condition of PPE. Applicators also had the highest median EDADinh consistent with the aerosolization required for application. This finding is in contrast with Krieger (1995) who suggested that M/L/A have the highest exposure potential. However, in the current study, six of the 33 M/L/A wore respirators thereby decreasing their exposure potential, whereas none of the A wore respirators.
The BDAD rank across the remaining job classes was M/L > RS > M/L/A whereas the EDADtot was M/LA > ML > RS. Median EDADinh and EDADtot ranked the lowest for RS consistent with their job function that excludes working with the aerosolized pesticide. For this group, both BDAD and EDADderm were similarly ranked as third highest, suggesting that the dermal methods underlying EDADderm were effective in classifying these workers whose exposure was likely dermal dominated. M/L/A were inconsistently ranked as second and fourth by BDAD and EDADtot, although EDADderm was consistent with BDAD (fourth), suggesting that inhalation exposure was potentially overestimated for this class.
Dermal exposure by itself tended to be highest among M/L, where there is risk for hand exposure to concentrated solutions. The high dermal exposure in this group was driven by extreme values among a small number of workers (shown in Fig. 1 and Table 2) implicating work practices as an exposure determinant. Machado-Neto and Matuo (1998) and Chester (1993) observed that 86% of exposure to M/L was to the hands. The significance of dermal exposure across all job classes was substantiated based on exposure measurements (i.e. EDADderm assuming 3% absorption and EDADinh assuming 100% absorption) identifying a median dermal to inhalation ratio of 0.6. This ratio exceeded unity for 34 of 77 workers (44%) indicating that the dermal exposure pathway was dominant for nearly half of all workers.
Variability occurring within and among studies could be due in part to differences in the application device, PPE use, spray technique, and quantity and formulation of pesticide applied. For example, the current study showed that among application types, workers applying chlorpyrifos by ground boom received higher dermal exposures relative to aerial, handheld sprayers or planters. The highest median inhalation exposure occurred among handheld sprayers (76 µg/application) although ground boom applicators were a close second (74 µg/application). Total exposure was greatest among handheld sprayers with a median of 418 µg/application, exceeding ground boom (202 µg/application), aerial sprayer (78 µg/application) and planter (11 µg/application) by factors of 2.1, 5.4 and 38. Rutz and Krieger (1992) examined exposure levels in 19 groups of pesticide M/L and A in California and found that application with a hand wand accounted for the highest exposure levels (1040 µg/h) in comparison to application with a ground boom or aerial application. They also found that M/L involved in open dumping of powder formulations received the highest mean exposure (24 274 µg/h). Attempts at lowering worker exposure potential in packaging have included the use of dripless containers, integrated handling systems and other mechanical transfer systems, as well as enclosure of work processes. Thus formulation, application type and product packaging may serve as useful indicators of worker exposure.
The DAF was examined by comparing (as a ratio) the absorbed dose estimated from exposure (EDADtot) and biomarker measurements (BDAD). This ratio was evaluated across three assumptions of DAF (i.e. 1, 3 and 10%). A study by Nolan et al. (1984) supports the use of a 3% DAF, the EPA's accepted value, whereas Krieger (1995) supports a 10% DAF and Griffin et al. (1999) report a 1% DAF. Our findings suggest that the DAF lies in the range of 3–10%.
Using a multiple linear regression model, 46% of the variability in the urinary 3,5,6-TCP levels could be explained. The unexplained variability may stem from measurement error in the dependent or independent variables or misalignment in timing in collection of biomarker samples relative to exposure. The dermal exposure estimate is especially vulnerable to error because of the assumptions required including that the dosimeter has retention similar to the skin and that dermal absorption is a constant. Cross- contamination can also occur during dosimeter removal. Thus, dermal exposure estimates may be more prone to uncertainty due to difficulty in standardization of the method and error associated with assumptions of the method. Woollen (1993) suggests that a nonlinear relationship between external exposure estimates and biomonitoring exists due to factors such as differences in personal hygiene habits (i.e. washing) and individual differences in metabolism. The rate of pesticide metabolism in the worker is influenced by duration and rate of exposure. The absorption process is thought to be saturable, not conforming to a linear dose–response model (Zendzian, 2000). Others suggest that the variability seen in passive dosimetry estimates might be attributable to lack of standardization of methods and exclusion of the ingestion pathway.
Fenske and Elkner (1990) identified similar positive associations of the 3,5,6-TCP urinary metabolite and external exposure measures (R2 = 0.86) in a study of residential pesticide applicators. Hines and Deddens (2001) found positively and highly correlated weekly means of ln(TCP) and ln(chlorpyrifos air concentrations) (R2 = 0.73, P < 0.0001) in 41 termiticide applicators in North Carolina. In an environmental in-home exposure assessment, Buckley et al. (1997) found that 3,5,6-TCP levels in the urine were moderately correlated with indoor air (R2 = 0.55; P 
0.01) and dust (R2 = 0.46; P 0.01) concentrations of chlorpyrifos during two seasons of sampling.
The fact that dermal exposure represents a substantial portion of pesticide workers' exposure and yet there are no standard methods of assessment, highlights a significant gap in the protection of worker health and safety. For some chemicals, this gap is addressed through biological-based exposure indices (e.g. ACGIH Biological Exposure Indices or BEIs®), however, such an index is not available for chlorpyrifos exposure other than nonspecific cholinesterase inhibition [American Conference of Governmental Industrial Hygienists (ACGIH, 2001)], which was unmeasured in the current study. The inhalation 8 h Threshold Limit Value (TLV®) (ACGIH, 2001) of 0.2 mg/m3 was exceeded for two M/L who had 8 h time-weighted averages of 0.37 mg/m3 (assuming no exposure for the balance of the workday). However, both workers wore respirators so that the measured air concentration was not representative of their inhalation exposure. Study results are compared with the 2001 rather than the current TLV® because the current guideline is based on inhalable sampling whereas the earlier TLV® was consistent with methods used in the registrant studies, i.e. 37 mm filter cassettes.
There are study caveats and/or limitations associated with the use of registrant studies to be considered in interpreting the current results. The use of WBD to measure dermal exposure alters (reduces) the amount of chlorpyrifos absorbed thereby affecting urinary TCP levels. Therefore, BDAD will underestimate exposure relative to what would occur without WBD monitoring. To the extent that BDAD is underestimated, the DAF derived from the ratio of EDADtotal to BDAD will be overestimated and the slope of BDAD regressed EDAD will be underestimated. Another limitation relates to the reliance on secondary data. It is significant to note that despite the added protection afforded by the WBD, 75 out of 80 workers had measurable levels of urinary 3,5,6-TCP associated with chlorpyrifos application. Although all studies used a common set of methodological guidelines designed to minimize measurement differences across studies (USEPA, 1998), each study was conducted independently and it is possible that there were subtle methodological differences in sampling and/or analysis that resulted in an unknown bias.
Overall, the EPA's Pesticide Registrant Database provided useful information for evaluating pesticide worker exposure and worker exposure assessment methods. These data help demonstrate the significance of the dermal route of exposure in pesticide workers. This analysis also served to demonstrate the public health utility of the EPA's Pesticide Registrant Database. This untapped resource should be made publicly available for the purpose of evaluating exposure assessment methods to better protect worker health.
This study highlights the need for continued development of methods to assess worker dermal exposure in the workplace to close the unexplained variability gap between measures of exposure and biological monitoring.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Disclaimer—This paper has been reviewed in accordance with the US Environmental Protection Agency's peer and administrative review policies and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
ACGIH. (2001) 2001 TLVs® and BEIs® threshold limit values for chemical substances and physical agents. Cincinnati, OH: ACGIH.
Albers JW, Cole P, Greenberg RS. (1999) Analysis of chlorpyrifos exposure and human health: expert panel report. J Toxicol Environ Health B Crit Rev; 2: 301–24.[CrossRef][Web of Science][Medline]
Barnekow DE. (1998) Evaluation of workers' exposures to chlorpyrifos during the use of Dursban TC termiticide concentrate for post-construction termiticide applications. The Dow Chemical Company. Washington, DC: Public Docket and Freedom of Information Office, Field and External Affairs Division (7506C), Office of Pesticide Programs, US Environmental Protection Agency.
Bartels M, Kastl P. (1992) Analysis of 3,5,6-trichloropyridinol in human urine using negative-ion chemical ionization gas chromatography-mass spectrometry. J Chromatogr Biomed Appl; 575: 69–74.[CrossRef]
Buckley TJ, Liddle J, Ashley DC, Paschal VW, Needham LL, Ackland G. (1997) Environmental and biomarker measurements in nine homes in the lower Rio Grande valley: multimedia results for pesticides, metals, PAHs, and VOCs. Environ Int; 23: 705–32.[CrossRef]
California EPA. (2000) Summary of results from the California Pesticide Illness Surveillance Program 1998. Sacremento, CA: California EPA.
California EPA. (2001) Summary of results from the California Pesticide Illness Surveillance Program 1999. Sacremento, CA: California EPA.
California EPA. (2002) Summary of results from the California Pesticide Illness Surveillance Program 2000. Sacremento, CA: California EPA.
Cattani M, Krzysztof C, Edwards J, Pisaniello D. (2001) Potential dermal and inhalation exposure to chlorpyrifos in Australian pesticide workers. Ann Occup Hyg; 45: 299–308.
CDC. (1998) Farm worker illness following exposure to carbofuran and other pesticides—Fresno County, CA. J Am Med Assoc; 281: 981–2.
Chester G. (1993) Evaluation of agricultural worker exposure to, and absorption of, pesticides. Ann Occup Hyg; 37: 509–23.
CFR. (2002) 40CFR170.1. Worker Protection Standard.
CFR. (2003) 40 CFR158.340. Toxicological Data Requirement.
Contardi JS, Gilbert JR, Lambesis DA, Murphy PG, Nolan RJ. (1993) Evaluation of chlorpyrifos exposures during mixing/loading and application of EMPIRE 20 insecticide to ornamental plants in commercial greenhouses. The Dow Chemical Company. Washington, DC: Public Docket and Freedom of Information Office, Field and External Affairs Division (7506C), Office of Pesticide Programs, US Environmental Protection Agency.
Das R, Steege A, Baron S, Beckman J, Harrison R. (2001) Pesticide-related illness among migrant farm workers in the United States. Int J Occup Environ Health; 7: 303–12.[Medline]
Fenske RA, Elkner KP. (1990) Multi-route exposure assessment and biological monitoring of urban pesticide applicators during structural control treatments with chlorpyrifos. Toxicol Ind Health; 6: 349–71.[Web of Science][Medline]
Fenske R, Birnaum S, Methner M, Soto R. (1989) Methods for assessing fieldworker hand exposure to pesticides during peach harvesting. Bull Environ Contam Toxicol; 43: 805–13.[CrossRef][Web of Science][Medline]
Fenske RA. (1993) Dermal exposure assessment techniques. Ann Occup Hyg; 37: 687–95.
Griffin P, Mason H, Heywood K, Cocker J. (1999) Oral and dermal absorption of chlorpyrifos: a human volunteer study. Occup Environ Med; 56: 10–13.
Henderson PT, Brouwer DH, Opdam JJG, Stevenson H, Stouten JTJ. (1993) Risk assessment for worker exposure to agricultural pesticides: review of a workshop. Ann Occup Hyg; 37: 499–507.
Hines CJ, Deddens JA. (2001) Determinants of chlorpyrifos exposures and urinary 3,5,6-trichloro-2-pyridinol levels among termiticide applicators. Ann Occup Hyg; 45: 309–21.
Knuteson JA, Barnekow DE, Cook WL, Meitl TJ. (1999) Evaluation of potential exposures to workers mixing and loading lorsban-4E insecticide products for aerial application. Dow AgroSciences LLC. Washington, DC: Public Docket and Freedom of Information Office, Field and External Affairs Division (7506C), Office of Pesticide Programs, US Environmental Protection Agency.
Krieger RI. (1995) Pesticide exposure assessment. Toxicol Lett; 82/83: 65–72.[CrossRef]
Machado-Neto JG, Matuo MT. (1998) Efficiency of safety measures applied to a manual knapsack sprayer for paraquat application to maize (Zea mays L.). Arch Environ Contam Toxicol; 35: 698–701.[CrossRef][Web of Science][Medline]
Mehler LN, O'Malley MA, Krieger RI. (1992) Acute pesticide morbidity and mortality: California. Rev Environ Contam Toxicol; 129: 51–66.[Web of Science][Medline]
Murphy PG, Beard KK, Marino TA, Myers CR, Ormand JR, Timchalk C. (1998) Evaluation of chlorpyrifos exposures to workers during loading and application of lorsban 15G granular insecticide during corn planting. The Dow Chemical Company. Washington, DC: Public Docket and Freedom of Information Office, Field and External Affairs Division (7506C), Office of Pesticide Programs, US Environmental Protection Agency.
Ness SA. (1994) Surface and dermal monitoring for toxic exposures. New York: Van Nostrand Reinhold.
NIOSH. (2000) SENSOR pesticides. Annual report, FY00. Available at http://www.cdc.gov/niosh/pestsurv/pdfs/pest-ar2000.pdf
Nolan RJ, Rick DL, Freshour NL, Saunders JH. (1984) Chlorpyrifos: pharmacokinetics in human volunteers. Toxicol Appl Pharmacol; 73: 8–15.[CrossRef][Web of Science][Medline]
Rutz R, Krieger RI. (1992) Exposure to pesticide mixer/loaders and applicators in California. Rev Environ Contam Toxicol; 129: 121–39.[Web of Science][Medline]
Shurdut BA, Murphy PG, Nolan RJ, McNett DA. (1993) Lorsban 4E and 50W Insecticides: assessment of chlorpyrifos exposures to applicators, mixer/loaders, and re-entry personnel during and following application to low crops. DowElanco. Washington, DC: Public Docket and Freedom of Information Office, Field and External Affairs Division (7506C), Office of Pesticide Programs, US Environmental Protection Agency.
US Bureau of Labor Statistics. (1987) 52 Federal Register 16050, 16059.
USEPA. (1991) Federal policy for the protection of human subjects; notice and rules. (40 CFR 26). Federal Register 56 FR 28002. June 18. Washington, DC: USEPA.
USEPA. (1992a) Exposure Assessment Group. Dermal exposure assessment: principles and applications. Interim report. January. Washington, DC: USEPA.
USEPA. (1992b) Regulatory impact analysis of worker protection standard for agricultural pesticides. Washington, DC: USEPA Biological and Economic Analysis Division.
USEPA. (1996) Federal Insecticide, Fungicide, and Rodenticide Act. 7 USC. s/s 136 et seq. Available at http://www.epa.gov/region5/defs/html/fifra.htm
USEPA. (2000) Revised human health risk assessment for chlorpyrifos and agreement with registrants. Available at http://www.epa.gov/pesticides/op/chlorpyrifos.htm
USEPA. (1988). Full record for chlorpyrifos. Integrated risk information system. Available at http://www.epa.gov/iris/subst/0026.htm
USEPA. (1998) Harmonized OPPTS test guidelines series 875. Occupational and residential exposure test guidelines, group B post application exposure monitoring test guidelines. Washington, DC: USEPA Office of Prevention, Pesticides, and Toxic Substances.
USEPA. (1997) Exposure factors handbook. National Center for Environmental Assessment. EPA/600/C-99/001. Washington, DC: USEPA.
USGAO. (1993) Pesticides on farms: limited capability exists to monitor occupational illnesses and injuries. GAO/PEMD-94-6. Washington, DC: USGAO.
Wolfe HR. (1976) Field exposures to airborne pesticides. In Lee RE, editor. Air pollution from pesticides and agricultural processes. Boca Raton, FL: CRC Press. pp. 137–61.
Woollen BH. (1993) Biological monitoring for pesticide absorption. Ann Occup Hyg; 37: 525–6.
Zendzian RP. (2000) Dermal absorption of pesticides in the rat. Am Ind Hyg Assoc J; 61: 473–83.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. GEER and T. J. BUCKLEY Reply Ann. Hyg., August 1, 2006; 50(6): 639 - 639. [Full Text] [PDF]
|
|
|
|
|
|
D. T. MAGE Dermal Absorption of Chlorpyrifos Ann. Hyg., August 1, 2006; 50(6): 638 - 639. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||