EPA/600/R-93/143
September 1993
FLUORESCENT TRACER EVALUATION OF PROTECTIVE
CLOTHING PERFORMANCE
by
Richard A. Fenske
Department of Environmental Health
School of Public Health and Community Medicine
University of Washington
Seattle, WA 98195
Cooperative Agreement No. CR-814919
Project Officer
Carolyn Esposito
Releases Control Branch
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Edison, NJ 08837
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
Printed on Recycled Paper
-------
DISCLAIMER
This material has been funded wholly or in part by the United States
Environmental Protection Agency under Cooperative Agreement CR 814919. It has been
subject to the Agency's review and it has been approved for publication as an EPA
document Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
11
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial products
frequently carry with them the increased generation of materials that, if improperly dealt
with, can threaten both public health and the environment. The U.S. Environmental
Protection Agency is charged by Congress with protecting the nation's land, air, and
water resources. Under a mandate of national environmental laws, the agency strives to
formulate and implement actions leading to a compatible balance between human
activities and the ability of natural systems to support and nurture life. These laws direct
the EPA to perform research to define our environmental problems, measure the impacts,
and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for planning,
implementing, and managing research, development, and demonstration programs to
provide an authoritative, defensible engineering basis in support of the policies,
programs, and regulations of the EPA with respect to drinking water, wastewater,
pesticides, toxic substances, solid and hazardous wastes, and Superfund-related activities.
This publication is one of the products of that researc h and provides a vital
communication link between the researcher and the user community.
This publication describes a research project that investigated occupational
exposure to pesticides and the role that personal protection can play in reducing such
exposures. Fluorescent tracer analysis of dermal exposure was able to document
significant problems associated with the design of chemical protective clothing currently
available for agricultural workers, and provided guidance for design improvements.
Risk Reduction Engineering Laboratory
E. Timothy Oppelt, Director
m
-------
ABSTRACT
Chemical protective clothing (CPC) is often employed as a primary control option
to reduce occupational exposures during pesticide applications, but field studies evaluating
CPC are limited. This study was designed to evaluate several protective garments and to
determine the ability of specific CPC components to reduce worker exposure. The studies,
conducted in central Florida during citrus applications of Ethion 4 Miscible™, examined
cotton workshirts and workpants, cotton/polyester (C/P) coveralls, SMS coveralls, and
Sontara coveralls. CPC performance was evaluated by fluorescent tracers and video
imaging analysis and by the patch technique. Nonwoven coveralls allowed significantly
greater exposure than did traditionally woven garments primarily because of design factors
(e.g., large sleeve openings). Fabric penetration occurred with high frequency for all test
garments, and none can be considered chemically resistant under these field conditions.
Improved coverall garments would, however, provide only a small further reduction in
exposure. Faceshields would reduce the exposure approximately three times more than
would improved coveralls. Exposure pathways that would probably be undetected or
inaccurately quantified by the patch technique were measured by fluorescent tracers and
imaging analysis. The patch technique, however, was far more sensitive in detecting fabric
penetration.
Workers conducting airblast applications would be better protected by closed cab
systems or any other technology that places a well-defined barrier between the worker and
the pesticide spray. Personal protective equipment (PPE) requirements should consider the
potential for heat stress, and conditions under which PPE is not to be used should be
defined and enforced to reduce the risk of illness related to heat stress. Protective garments
designed and marketed for use by pesticide applicators should be field tested to determine
performance, and users should be provided with accurate information regarding the
chemical resistance of such garments.
This report was submitted in fulfillment of Cooperative Agreement No. CR-814919
under the sponsorship of the U.S. Environmental Protection Agency. This report covers a
period from 1 April 1988 to 11 January 1991, and work was completed as of 1 March 1991.
IV
-------
CONTENTS
Disclaimer ft
Foreword iii
Abstract iv
List of Tables vii
List of Figures viii
Acknowledgments ix
1. Introduction 1
2. Conclusions 3
3. Recommendations 5
4. Protective Clothing Performance Study 7
4.1 Objectives . 7
4.2 Study Design 7
4.3 Methods 8
4.3.1 Field Conditions 8
4.3.2 Sampling '. 10
4.3.3 Analysis 12
4.4 Year 02 Modifications . 14
4.4.1 Field Conditions 15
4.4.2 Sampling 15
4.4.3 Analysis 16
4.5 Results 17
4.5.1 Imaging Analysis 18
4.5.2 Visual Observations 20
4.5.3 Estimated Ethion Exposure to Protected Regions 23
4.5.4 Ethion and Tracer Penetration of Protective Clothing 24
5. Total Exposure Distribution Study 29
5.1 Objectives l. 29
5.2 Study Design 29
5.3 Methods 29
5.3.1 Field Conditions 29
5.3.2 Sampling 30
5.3.3 Analysis 32
5.4 Results 34
5.4.1 Respiratory Exposure 34
5.4.2 Hand Exposure 35
5.4.3 Face and Head Exposure 35
5.4.4 Exposure Beneath Coveralls 37
5.5 Exposure Distribution Assessment 41
5.5.1 Dermal Exposure Scenarios 42
5.5.2 Dermal and Respiratory Dose Scenarios 45
-------
6. Discussion.
7. References
48
50
VI
-------
LIST OF TABLES
Number
4.1 Video Imaging Analysis of Fluorescent Tracer Exposure
4.2 Challenge - Adjusted Tracer Exposure Values by Garment Type
4.3 Nonparametric Analysis of Variance of Garment Types by Body Region
4.4 Ethion Exposure Estimates for Protected Body from Video Imaging
Analysis
4.5 Ethion Challenge and Penetration to the Legs - Year 01
4.6 Tracer Challenge and Penetration to the Legs - Year 01
4.7 Ethion Challenge and Penetration to the Chest and Legs - Year 02
5.1 Respiratory, Hand, Face, and Head Exposure to Ethion
5.2 Comparative Estimates of Face Exposure to Ethion
5.3 Ethion Deposition Rates on Cotton and Sontara Coveralls
5.4 CPC Breakthrough Frequency by Garment
5.5 CPC Breakthrough Frequency - by Body Region
5.6 Ethion Exposure beneath Cotton Coveralls and Sontara Garments
5.7 Dermal Exposure Reduction by Personal Protective Equipment
5.8 Ethion Dose Estimates and Reduction by Personal Protective Equipment
19
21
22
25
26
27
28
35
37
38
39
40
41
43
47
Vll
-------
LIST OF FIGURES
Number
4.1 Video imaging analysis of fluorescent tracer exposure
4.2 Qualitative evaluation of fluorescent tracer exposure
Eage
22
23
vui
-------
ACKNOWLEDGMENTS
This project was conducted in central Florida and at the Department of
Environmental Sciences, New Jersey Agricultural Experiment Station, Rutgers
University, New Brunswick NJ 08904. We thank the workers and management of the
two central Florida citrus cooperatives who participated in the study. Invaluable
assistance in the field portion of the project was provided by Dr. Herb Nigg and his staff
at the Lake Alfred Experiment Station, University of Florida.
This work was accomplished through the combined efforts of students and staff at
Rutgers University. Shari Birnbaum supervised collection of all field data and Mark
Methner played an instrumental role in the success of the Year 01 field study. In
addition, the following individuals played an active role in the collection and analysis
phases of the project: Kathleen Black, Karen Cicero, Kenneth Elkner, Ian Grey, Andy
Hanneman, Der-Jen Hsu, Chorng-Li Lee, Jau-Jang Lui and Eric Zwerling. Diane Sturts at
the University of Washington Department of Environmental Health was of great
assistance in the preparation of this technical report.
This project was funded primarily by the U.S. Environmental Protection Agency's
Office of Research and Development It was supported in part by the NJ Agricultural
Experiment Station and by the University of Washington Department of Environmental
Health.
IX
-------
-------
SECTION 1
INTRODUCTION
Chemical protective clothing (CPC) is a major control option in reducing occupational
exposure to pesticides. In many agricultural settings closed mixing systems, closed cabs and other
engineering control approaches are not feasible. Clothing and other personal protective equipment
(PPE) which can substantially reduce pesticide contact with the skin and which can be worn
comfortably during normal work activities is needed in the workplace (Nielsen and Moraski 1986).
The ability of chemical protective clothing to reduce dermal exposure is dependent both on low
fabric penetration properties and proper design. The U.S. Environmental Protection Agency has
provided substantial support recently for research aimed at identifying appropriate protective
garments for agricultural workers. Evaluation of protective clothing has traditionally been divided
into two phases: laboratory testing and field performance testing. Laboratory testing can provide
information regarding pesticide penetration through fabric, but only field testing under realistic
exposure conditions can determine the overall efficiency of penetration reduction and design.
Penetration characteristics of fabrics may be altered dramatically during field use: 1) worker
movements may affect movement of dusts through fabric weave, 2) direct contact between the
clothing and body may enhance movement of liquids through fabric, 3) sweating may change
penetration rates in unpredictable ways. Design factors which enhance or reduce exposure are only
evident during field use of the clothing.
The methods currently employed to evaluate protective clothing performance in the field are
limited. The patch technique places collection pads above and beneath clothing to estimate garment
penetration. This approach can produce highly variable measurements, since pesticide exposure
during applications is not uniform (Fenske et al. 1985; Fenske 1990). In many cases the patch will
either be "hit" or missed altogether, producing misleading results regarding clothing performance.
Exposure may also occur by means which patch sampling overlooks. Recent work has shown that
a substantial portion of exposure to mixer/applicators wearing coveralls occurs through the garment
openings; i.e., sleeves and collar (Fenske 1988a). Exposure may also occur at seams, and through
secondary contact (e.g., deposition on the skin from a contaminated glove or hand). The use of
fluorescent tracers and video imaging provides an opportunity to conduct realistic field
performance evaluations (Fenske etal. 1986a; b). The fluorescent tracer allows visualization of
exposure patterns on the skin. The imaging system records the pattern of exposure and calculates
-------
the corresponding pesticide deposition on the skin. This approach allows direct evaluation and
comparison of different protective clothing options, and serves to complement both laboratory
testing and field methods such as biological monitoring.
-------
SECTION 2
CONCLUSIONS
1. The nonwoven coveralls tested in Year 01 of this study performed less effectively than
traditional woven garments due to garment design factors. Significantly greater exposure was
measured on the arms of workers wearing the SMS and Sontara garments compared to either a
cotton/polyester coverall or a cotton workshirt. This 'difference was attributable to the large
sleeve openings of the nonwoven garments. When the sleeve openings were reduced in Year
02, no significant difference in exposure beneath coveralls was observed between Sontara and
cotton garments.
2. Fabric penetration occurred with high frequency for all test garments. No significant
differences in percent penetration were found between woven and nonwoven garments. None
of the test garments can be considered chemically resistant under the field conditions evaluated
in this study.
3. Properly designed garments (woven or nonwoven) such as those evaluated in this study
provide a substantial reduction in exposure when compared to a theoretical "unprotected"
worker, but improvement in the chemical resistance of coverall garments will provide only a
small further reduction in exposure. Faceshields could provide approximately three times the
exposure reduction resulting from improved coverall garments. The hands, even when
protected by chemical resistant gloves, contribute a substantial proportion of total dermal
exposure, as does the unprotected face/head region.
4. The inhalation route of exposure was estimated to contribute 20% of the total absorbed dose
under these study conditions if a worker were to wear gloves and coveralls but no respiratory
protection. Proper use of a respirator or use of a faceshield would result in similar reductions
in absorbed dose; a combination of these two PPE options would reduce absorbed dose by
nearly 40%. Either of these measures would be more effective in reducing dose than
improving the chemical resistance of coveralls.
5. The use of fluorescent tracers and video imaging analysis allows measurement of exposure
which occurs by pathways which would likely be undetected or inaccurately quantified by the
patch technique (e.g., exposure through openings in garments). The patch technique was far
-------
more sensitive in detecting fabric penetration. The techniques appear to play complementary
roles in documenting the performance of chemical protective clothing under realistic field
conditions.
-------
SECTION 3
RECOMMENDATIONS
1. Dermal and respiratory exposures under the work conditions studied were relatively high for
pesticide applicators. Workers conducting airblast applications would be better protected by
closed cab systems or any other technology which places a well-defined barrier between the
worker and the pesticide spray.
2. Personal protective equipment (PPE) options must be considered if engineering controls are not
feasible. PPE requirements or recommendations should take into account the potential for heat
stress and should be designed to strike a balance between protection and comfort. Conditions
under which PPE shall not be used should be defined and enforced to reduce the risk of illness
related to heat stress.
3. Implementation of PPE requirements or recommendations should include procedures whereby
employers and workers receive appropriate and ongoing education and training regarding PPE
use.
4. Important factors to be considered in developing PPE requirements or recommendations
include the following:
a. Woven or nonwoven coveralls similar to those tested in this study provide substantial
protection to most of the body; improvements in the chemical resistance of such garments
will probably not reduce total dermal exposure significantly.
b. The hands, even when chemical resistant gloves are worn, contribute a substantial
proportion of total dermal exposure under the use conditions studied. Further reduction in
hand exposure will be achieved only by more effective employer and worker education and
training.
c. The unprotected head represents a substantial proportion of total dermal exposure; use of a
hood covering the back of the neck and most of the head would reduce exposure
significantly; addition of a faceshield would further reduce exposure.
-------
d. The respiratory route of exposure represents a substantial proportion of total dose under the
use conditions studied. Respirator use will, in theory, eliminate this exposure. Under
current use conditions, further employer and worker education and training will be required
to insure that respirators are properly fitted, tested and used. The use of an air-supplied
respirator covering the entire head would obviate the need for the hood, faceshield and
respirator, and appears to be the best available alternative.
5. Protective garments designed and marketed for use by pesticide applicators should be field
tested to determine performance. Traditional laboratory tests (e.g., permeability testing) cannot
characterize effects of garment design and appear to be inadequate measures of potential
chemical breakthrough.
6. Users should be provided with accurate information regarding garments designed and marketed
for pesticide handlers. Claims regarding the ability of garments to protect workers should be
accurate. In particular, garments should not be referred to as "chemical resistant" or "liquid
proof unless these qualities have been demonstrated under realistic field use conditions.
-------
SECTION 4
PROTECTIVE CLOTHING PERFORMANCE STUDY
4.1 OBJECTIVES
This study involved field testing of protective clothing under controlled, but realistic
pesticide application conditions. The objectives of the study were as follows:
1. Evaluate the performance of test garments in reducing exposure using the fluorescent tracer
technique
a. Compare dermal exposure measurements of workers wearing test garments to the
exposure of workers wearing traditional protective clothing.
b. Compare the results of the fluorescent tracer technique with those of the traditional patch
technique to determine the scientific validity and feasibility of employing the fluorescent
tracer technique as an alternative or complementary evaluation method.
2. Determine the role of garment design in reducing exposure.
a. Establish the relative importance of exposure by clothing penetration versus exposure
through openings in clothing.
b. Determine whether significant reductions in exposure can be obtained by modifying neck,
sleeve and leg closures.
4.2 STUDY DESIGN
Four garment types were selected for study. Two were traditional garments used in
agriculture (workshirt/workpants, woven coveralls) and two were made from nonwoven fabrics
selected by U.S. EPA investigators based on their potential for providing both protection and
comfort in. hot environments. Eight replicate exposures of each garment were proposed based on a
previous studies which indicated statistical differences in garment performance with a similar
sample size (Fenske 1988b). Each applicator in the study wore each of the garments at least once
-------
to minimize potential confounding due personal application procedures. Equipment type, tank size
and amount of fluorescent tracer applied per tank were controlled for all applications in Year 01.
Uncontrolled variables included number of tanks applied, application time and individual work
practices.
4.3 METHODS
4.3.1 Field Conditions
Field studies occurred during the summer application (July through early August) of Ethion
4 Miscible™ [EPA Reg. No. 279-1254]. Ethion 4 Miscible™ is a liquid concentrate formulation
containing 4 Ibs active ingredient (AI)/gal and is 46.5% AI by weight The active ingredient is the
organophosphorus insecticide, ethion [0,0,0',0'-Tetraethyl S,S'-methylene bisphosphorodithioate].
Participants were employed by two citrus cooperatives in central Florida.* All subjects were adult
males who applied pesticides as part of their normal work duties. They read and signed a consent
form prior to participating in the study (Appendix HI), and were paid a nominal sum for each day of
participation. Each morning, a mixer and two pesticide applicators were either met at their citrus
cooperative headquarters and followed to the citrus grove or were met directly in the grove. Each
applicator was given a black, cotton T-shirt and one of the protective garments to wear. The mixers
were not monitored during this study. All participants were provided with chemical resistant gloves
during the study.
Protective Garments: Four types of protective garments were tested. Each participant wore each
type of protective garment at least once during the course of the study. Fabric descriptions and
characteristics are drawn from DeJonge and Easter (1989) and Nigg et al. (1992):
1) SMS coverall (nonwoven): 100% polypropylene composite material with three-layered
construction (thermally point bonded laminate of spunbonded, melt blown, spunbonded
fabric); thickness = 11.8 mils; weight = 62 gm/m2; treated with Kimberly-Clarke RF™, a
repellent finish, exact commercial formulation unknown; 44 cm sleeve circumference.
2) Sontara coverall (nonwoven): 50% polyester, 50% wood pulp material with both point
bonded and spun bonded construction (spunlaced composite); thickness = 12.6 mils;
weight = 72 gm/m2; treated with DuPont RF™, a repellent finish, exact commercial
formulation unknown; 44 cm sleeve circumference.
-------
3) Cotton/Polyester coverall (woven): a 65% cotton 735% polyester twill material (twill woven
construction); thickness = 19.0 mils; weight = 243 gm/m^ weight; untreated; 30 cm sleeve
circumference.
4) Cotton workshirt/workpants (woven): a 100% cotton twill material (twill woven
construction); thickness = 19.0 mils; weight = 243 gm/m^; untreated; 23 cm sleeve
circumference (when buttoned).
Application Procedures: The Ethion 4 Miscible formulation was applied throughout the study
according to label instructions. The amount of formulated ethion added to each 500 gal tank varied
between the two cooperatives. Typically, the mixer measured the ethion formulation into a bucket
and then poured this bucket into the mixing tank. The tank was filled with water, agitated
mechanically, and pumped into the applicators' spray tank. Natural oil and other agricultural
chemicals (e.g., copper, Benlate, Kocide) were frequently added to the spray mixture. In some
cases no ethion was included in the spray mix; i.e., pest control practices dictated that another
insecticide or no insecticide be used on those days.
Cooperative A utilized a 1000 gal mixing tank, allowing one mixer to supply two
applicators. Cooperative B utilized 500 gal mixing tanks, requiring two mixers to supply two
applicators. Both cooperatives utilized airblast sprayers with 500 gal tanks. The sprayers were
pulled by open air tractors with a top canopy for shade. The sides and back of the tractor were
covered by a metal screen to protect the workers from branches. The front of the tractor was open.
Several of the workers covered a portion of the side/back screens with a water-resistant cloth to
block the spray (they found the spray mixture oily and sticky). Some of the workers from
Cooperative A used sprayers which utilized electronic photocells to detect the presence and size of
trees. The sprayer would automatically turn off/on nozzles as needed. Workers from both
cooperatives who used tractors which did not contain the electronic photocells could manually turn
off the nozzles on the left, right or top of the sprayer as needed. Each worker was monitored
during application of four 500 gal tanks. Spraying was occasionally terminated before all four
tanks had been applied due to rain.
Fluorescent Tracer: A commercially available fluorescent whitening agent, Calcofiuor RWP (4-
methyl-7-diethylaminocoumarin), was employed as a tracer of pesticide residue deposition. This
compound has been used previously as a tracer in orchard airblast applications (Fenske etal. 1985;
Fenske 1988a). A pre-measured bag (300 gm) of the fluorescent tracer was mixed into a bucket
containing the ethion formulation. If no ethion was to be applied, the tracer was mixed into a small
amount of natural oil instead. Thus, the tracer concentration in the spray mix was constant
-------
throughout the studies (300 gm per 500 gal H20; 160 ppm), despite changes in ethion application
rates.
4.3.2 Sampling
Video Image Sampling: The mobile laboratory utilized in this project was a 1965 Dodge Travco
recreational vehicle which was provided by the U.S. EPA. Prior to the study, extensive
modifications and repairs were made to the interior and exterior of the vehicle as well as the engine.
In addition, a new generator and air conditioner were installed. The interior of the mobile
laboratory was painted black, a dark carpet was laid on the floor and black curtains were fabricated
to cover all windows, vents and the back passageway. A video imaging system was placed in the
mobile laboratory and secured for transport.
The general design of the second generation instrument used in this study is similar to that
of the original VFTAE system (Fenske et al. 1986a). The major improvements on the original
system include 1) increased resolution and grey level scale, 2) increased processing speed, 3)
replacement of the touch screen with a mouse for image outlining, and 4) replacement of floppy
disks with a tape back-up system for data storage (Fenske et al. 1993). Hardware components
used in this study were a DOS-based microcomputer (Compaq 286), imaging analysis board (Data
Translation DT2851), television camera (RCA T2000), television monitor (Ikegama), data storage
tape unit (Mountain Computer Filesafe Model 7060), optical mouse (Mouse Systems Corporation
serial mouse), UV-A lamps (custom, with 4 F40 BLB bulbs/lamp + UV-passing glass filters), and
subject examination frame (custom, 70 x 70 cm interior dimension). Custom software programs
used in the study were VITAE-PIC (image acquisition), VITAE-MAP (outlining and overlay of
pre- and post-exposure images, and VITAE-CALC (calculation of exposure) The accuracy of the
exposure calculations produced by this system are discussed in detail elsewhere (Fenske et al.
1993). The distance from the UV lights to the frame (subject-light distance) was 90 cm. The
distance from the camera to the frame (subject-camera distance) was either 75 cm (head, hands and
neck images) or 85 cm (all other images). UV-A readings and standard target readings were
collected prior to and following each subject examination.
Pre-exposure video images were acquired of each subject on the first day the subject was
studied. At the end of spraying the applicators were brought to the mobile laboratory where
protective garments and T-shirts were removed by study staff, and post-exposure video images
were acquired. Video images were acquired of the hands, head, neck, forearms, upperarms, upper
torso and lower torso of each worker during each video imaging session. Four views (front, back,
left and right) were acquired of the head, both forearms and the lower torso. Three views (front,
10
-------
back, and outer) were acquired for both upper arms. The inner view of this region was not
collected due to difficulties in positioning subjects and due to previous observations that little or no
exposure occurs on this area during pesticide applications (Fenske 1988b). Two views (front and
back) were acquired for the upper torso and both hands. An image of the front of the neck was
acquired for several, but not all of the workers. Images of the legs were not acquired since some
workers were reluctant to participate in the video imaging procedures for regions below the waist,
and much of the leg surface area was covered both by the inner patch samplers (200 cm2 total) and
by tall rubber boots extending up to the knee All video images were acquired using VTTAE-PIC,
saved onto the microcomputer's hard drive, and transferred to a storage tape at the end of each
session. At least one full set of pre-exposure video images were recorded for each worker.
Individuals who participated more than once in the study waited at least three days before repeating
as subjects in order to insure that tracer from a previous exposure did not remain on the skin. This
waiting period had been found to be adequate previously (Fenske 1988b). No residual tracer was
observed on subjects used on a repeat basis during the study.
Qualitative Scoring Procedure: Fluorescent tracer deposition patterns were evaluated and scored
qualitatively for each body part/view according to a modification of a visual scoring system
(Fenske 1988c). Each view was assigned a score of 0-3 based on the intensity and extent of
deposition on the skin:
0 no visually observed tracer
1 low-level tracer visible; near imaging system limit of detection
2 tracer clearly visible; detectable by imaging system
3 tracer highly visible; easily detectable by imaging system
Scores for views were summed for each body part (i.e., face, forearms, upper arms, torso) to
allow comparisons among workers and garments. These scores were also employed as part of the
quality assurance procedures for video imaging evaluation.
Patch Sampling: Dermal patches were not employed to estimate exposure by traditional
methods (e.g., USEPA 1987). Rather, they were employed to estimate protective clothing
penetration, and to measure the proportion of ethion and tracer reaching the outside of the garment
Six 103.2 cm2 (16 in2) square alpha-cellulose patches were pinned to the protective clothing of
each worker: two were attached at the front of the thighs (one per thigh) on the inside of the
protective garment, and two were attached on the outside immediately adjacent to (but not
overlapping) the inner patches; the fifth patch was attached at the right upper chest (outside the
garment) and the sixth was pinned to the front of the worker's hat in a vertical position. For
workers who did not wear hats the sixth patch was attached at the left upper chest After the
11
-------
worker completed spraying his last tank, the patches were removed, immediately wrapped in foil
and placed in an ice-chest with ice for transport to the University of Florida Lake Alfred
Experiment Station laboratory where they were transferred to a -20°C freezer for storage until
analysis.
4.3.3 Analysis
Video Image Samples: Two custom written C-language software programs (VITAE-MAP
and VTTAE-CALC) which utilize subroutines contained in Data Translation's DT-Iris library were
used to analyze each image. The first program, VITAE-MAP, used a mouse to outline the portion
of each post-exposure image which was to be analyzed. Two reference points in both images were
then identified using the mouse. These two points were used to transpose the outline onto the pre-
exposure image. The outline information was saved onto a floppy diskette. The portions of the
post-exposure images which had been outlined were then quantified using the program VITAE-
CALC (Fenske et al. 1993). Several adjustments are automatically made to the images during
analysis. Each image (pre and post-exposure) is adjusted for distortion due to lens vignetting.
Any change in the lights or camera sensitivity between the pre and post-exposure imaging sessions
are corrected. A histogram of the number of pixels within the outlined area at each gray level (0
through 255) is generated and compared for the pre and post-images. If either the total brightness
or the maximum gray level (gray level containing at least 10 pixels) of the post-image is greater
than the pre-image, the computer considers the post-image to be exposed, and the amount of the
exposure is then quantified.
An anthropometric model is used to adjust both the pre and post-images for the effects of
non-planar surfaces (changes in the brightness and the surface area represented by each pixel).
The unexposed pixels within the image (background skin) are removed by subtracting the
histogram of the pre-image from the histogram of the post-image. All remaining pixels are
considered to be exposed. The total mass of tracer within the outlined image is then quantified
using a standard curve (Fenske et al. 1993). The data for the standard curve was collected in the
laboratory at New Jersey. Several concentrations of tracer in acetone were spotted onto subjects'
skin. Images were acquired of these spots and a line which related the mass of tracer spotted
verses the brightness of the spot was determined. This information is input into the VITAE-CALC
program to quantify the mass of tracer in the post-exposure images.
Patch Samples: The patch samples were extracted and analyzed for ethion at the University of
Florida Lake Alfred Experiment Station laboratory, under the supervision of Dr. Herbert Nigg.
The samples were center cut into a 40.32 cm2 square using a paper cutter. This square was then
12
-------
quartered. All four quarters were placed into a 225 ml jar with 50 ml of a 30% acetone/70%
hexane solvent mixture and shaken on a flat top rack shaker (New Brunswick Scientific, Model R-
2) for 5 min at 350 RPM. The solvent was decanted into a 250 ml round-bottom flask and the
procedure was repeated. The two extracts were combined. An aliquot of the patch extracts
(approximately 10 ml) was transported on dry-ice to New Jersey for analysis of the fluorescent
tracer. The remainder of the extract was evaporated to dryness on a rotaty evaporator at 40°C and
then brought back up in 10 ml hexane. These samples were analyzed for ethion content by gas
chromatography using electron capture Ni^3 detection and a fused silica capillary column.
Extracts received from Florida were analyzed for tracer at the Rutgers University
Department of Environmental Sciences laboratory. A Turner 430 spectrofluorometer was zeroed
on a cuvette containing solvent only (30% acetone/70% hexane). Six standards (calcofluor in 30%
acetone/70% hexane), ranging from 0.0016 to 0.32 ppm, were read. All samples were placed in
the same cuvette and read on the most sensitive scale possible. A blank (cuvette filled with solvent
only) was read approximately every 5 samples to ensure stability of the machine. Linear
regression of the six standards was used to quantify the samples.
4.3.4 Quality Assurance/Control
Quality assurance studies for ethion on patches were conducted under the supervision of
Dr. Herbert Nigg at the University of Florida Lake Alfred Experiment Station laboratories, and
were reported to USEPA previously in his final report. Procedures and results are summarized
here. Extraction efficiency studies were conducted by spiking pads with 10 ug of ethion. Average
recovery was 93 ± 2% (n=12). Field spike samples were prepared by spiking 10 ug of ethion onto
each alpha-cellulose pad. Pads were taken to the field throughout the course of the study. Average
recovery was 90 ± 2% (n=28). Field blank samples (n=30) were prepared at the same time and
exposed to the same environmental conditions. The mean amount of ethion recovered was
substantially below the limit of detection (<0.0025 ug/cm2). All laboratory blank samples were
below the limit of detection. Storage stability of sample pads and extracts were tested and no
losses were noted.
Patch samples were analyzed for the fluorescent tracer, Calcofluor RWP, at the Rutgers
University Department of Environmental Sciences. Extraction efficiency (n=8) was tested by
spiking pads with either 50 ug (n=4) or 5 ug (n=4). The extraction efficiency at the 50 ug level
was 96 ± 3.2%. The extraction efficiency at the 5 ug level was 82 + 2.4%. Field spike samples
(n=15) were prepared by spiking 50 ug of calcofluor onto each alpha-cellulose pad. Three pads
were kept as controls. The remaining 12 were placed in filtered sunlight at midday (11:00-15:00)
13
-------
for times ranging from 1-4 hr. Average temperature was 90° F in the shade and relative humidity
was 55-60%. Mean recovery after 1-2 hr was 85 ± 4%. Mean recovery after 3-4 hr was 75 ±
10%. These results correspond with previous observations that the tracer can be degraded by
sunlight (Fenske et al. 1985). Field blank samples (n=4) were prepared at the same time and
exposed to the same environmental conditions. Three samples were at or below the limit of
detection (0.00016 ppm, equivalent to 0.40 ng/cm2). The fourth sample measured 1.4ng/cm2.
Laboratory blank samples (n=4) were center-cut, quartered, extracted and analyzed in the same
manner as field patches. Three were below the limit of detection. The fourth sample measured
0.69 ng/cm2. Storage stability of sample extracts (n=8) was tested by spiking 500 ug of
calcofluor into 50 ml of solvent (30% acetone / 70% hexane) and storing at -20° C for six months.
Mean recovery was 95 ± 1.8%.
The accuracy and precision of the video imaging system were monitored continuously in
the field throughout sample collection. Images of a standard target were acquired immediately
prior to and immediately following each imaging session for each worker. Since each worker had
two imaging sessions (pre- and post-exposure), four standard target samples were collected for
each worker dataset. The standard target was a 36 cm2 square of standard photocopy paper
covered with 70% shading paper (Letratone), attached to a black board so that the standard target
could be placed in the center of the frame. The percent differences across pre-exposure sessions,
post-exposure sessions and worker evaluation sessions averaged 1.8%, 1.9% and 3.7%,
respectively. Thus, the imaging system performed in a very stable manner (<5% variability)
throughout the entire data collection period. The limit of detection for image samples in this study
was 35 ng of tracer per square centimeter of exposed skin surface.
4.4 YEAR 02 MODIFICATIONS
Preliminary analysis data from Year 01 indicated that the Sontara garment provided greater
protection than the SMS garment. In an effort to narrow the focus of the study and increase
sample sizes, only two protective garments were studied in Year 02: the treated Sontara coverall
and a woven coverall made of 100% cotton denim material (twill woven construction); 0.66 mm
thickness; 274 gm/m2 weight; untreated. This type of woven coverall had been determined in
laboratory studies to be the most effective among available woven garments (E. Easter; personal
communication). The cotton coverall was pre-washed with water softener (most laundry
detergents contain fluorescent whitening agents) to remove any available fluorescing compounds.
14
-------
Field observations and preliminary analysis of Year 01 data indicated that exposure to the
forearms, upper arms and in some cases the torso was due primarily to material being blown up the
sleeves rather than penetrating through the protective garment. The Sontara garment design was
therefore altered from Year 01, with the circumference of the sleeve at the wrist reduced from 44
cm to 32 cm. An effort was also made to tape the sleeves of the Sontara garment with masking
tape in Year 02, although this was not accomplished in aill cases. Additionally, it was decided to
collect images of the worker's legs during the second yezir of this project Since the legs of the
garments were tucked into work boots, the only exposure pathway for the legs, especially the
upper legs, would be penetration through the protective clothing. Applicators were given a pair of
black athletic shorts to wear under the protective garment in addition to the black T-shirt. The
level of fluorescence observed on regions which were covered by protective clothing were low
during Year 01. Even lower levels were anticipated in Year 02 due to alterations in the garment
openings. Thus, the mass of tracer mixed into each spray tank was increased to 400 gm per 500
gal spray mixture.
4.4.1 Field Conditions
The mobile laboratory utilized in Year 02 was a 1985 Chevrolet truck which was provided
by the U.S. EPA. During the first day of the field study, surging of the generator was noticed.
The surging became more pronounced the longer the generator was used. As the video equipment,
especially the camera and lights, are sensitive to electrical voltage, these surges made data
collection unreliable. The field study was discontinued until the problem could be resolved.
Several mechanics were contacted in an attempt to correct the problem. The problem was finally
resolved when a mechanic in Tampa, FL replaced the condenser which had been recalled by the
manufacturer 6 months earlier. After data had been collected from several subjects, a very light,
diffuse exposure (below the detection limits of the imaging system) was noted on the torso of a
worker wearing Sontara (Worker #8). Further inspection of the Sontara garment revealed that it
fluoresced weakly under the UV lights (No fluorescence had been osbserved with this fabric in
Year 01.) A question arose as to whether the observed exposure was due to penetration of tracer
through the protective garment or leaching of a fluorescing agent from the Sontara garment. In
order to determine whether the fluorescence was due to garment breakthrough the tracer
concentration was increased from 400 gm to 1200-1600 gm per tank (Worker #9 -18).
4.4.2 Sampling
A new Cohu Charge Coupled Device (CCD) camera equipped with a 12.5 75 mm zoom
lens was used in place of the RCA camera. The zoom lens was fixed at a 12.5 mm focal length.
The subject-camera distance was changed to 110 cm due to the new lens, and this distance was
15
-------
fixed for all images to simplify the image acquisition process. An additional 12 images were
acquired of the upper and lower legs of each worker. Seven alpha cellulose patches backed with a
polyethylene layer (to prevent sweat from entering the patch) were employed rather than six. The
additional patch was attached to the inside of the protective clothing at the chest Concern that
residue might be transferred from the alpha-cellulose to the aluminum foil wrapping led to the use
of a cover patch (a second patch placed on top of the sample prior to wrapping). The cover patch
was extracted and analyzed with the sample patch.
4.4.3 Analysis
Analysis of video images revealed several problems. First, images for Workers 1 and 2
were collected while the generator was providing inadequate current to the imaging system.
Fluctuations in the instrument's performance appeared to affect the brightness of the images,
making quantification of the images unreliable. Second, workers 4 and 17 had detectable tracer on
their skin prior to pre-image collection. It is not certain how this exposure occurred. Most likely it
was the result of contacting equipment that had been used the previous day in the study. Third, the
pre-images of workers 5,6 and 7 appeared much darker than expected. This problem is most
likely to have been caused by use of an incorrect f/stop on the camera when the images were
collected or an error in the acquisition of the image used to correct all subsequent images for
system noise. These images could not be used as pre-exposure images, and thus the post-
exposure images could not be quantified. Fourth, worker 11 applied only 0.5 tanks of
ethion/tracer (due to 2 sequential flat tires on his tractor); thus no post-images were collected. In
sum, images from a total of 17 workers (no post-images for worker 11) were collected before the
study was terminated. Elimination of the seven workers listed above left a final data set of 10
subjects. The qualitative examination revealed that exposure was very low for the body parts
covered by the protective clothing. An attempt to increase detectable fluorescence beneath clothing
by adding 1200 or 1600 gm of tracer rather than 400 gm of tracer to each tank indicated that the
observed exposure was due to garment penetration, but at levels so low that they would not be
quantifiable by the video imaging system. In light of these circumstances, video images collected
during Year 02 were not quantified.
All chemical analyses were conducted at the Rutgers University Department of
Environmental Sciences laboratories. Both the sample patch and the cover patch were center cut to
a 25.81 cm2 (4 in2) square. The plastic backing of each patch was removed. Each patch and its
associated cover were placed into a 4 ounce glass jar with 30 ml of toluene and shaken at high
speed on a shaker table (100 cpm) for one hour. The patch extracts were analyzed for ethion
content by gas chromatography (Hewlett-Packard 5890A, electron capture detector) using the
16
-------
following conditions: nitrogen carrier gas flow rate = 83 itnl/min, oven/column temperature 205°
C, injection port temperature 225° C, detector temperature 300° C. The detector was
reconditioned during the analysis; the nitrogen flow rate was changed to 85 ml/min and the
injection post temperature was reduced to 215°C at this time. The glass column (6 ft, 0.635 cm
(0.25 inch) outer diameter, 4 mm inner diameter) was packed with 20 gm 5% 2401 Supelcoport,
100/120 mesh. A 1.0 ul injection was made for each sample. An internal standard (100 pg
Dieldren) was used to quantify the field samples.
4.4.4 Quality Assurance/Control
Patch extraction and analysis in Year 02 were conducted at the Rutgers University
Department of Environmental Sciences. Extraction efficiency (n=6) was determined by spiking
5.08 x 5.08 cm alpha-cellulose pads with either 2.7 ug or 54 ug of ethion in toluene. Each pad was
paired with a cover pad and then extracted and analyzed as described for the field samples. Three
controls were prepared at each level by spiking the same iamounts into toluene. The extraction
efficiency for the two spiking levels was 100.3% ± 4.3 and 99.0% ± 2.4, respectively. Field
spike recovery was determined from patches which had been spiked with either 100 ul or 1 ml of a
150 ppm ethion formulation in toluene standard (70 |ig ethion active ingredient/ml toluene). The
samples were transported and analyzed in the same manner as the field samples. Average recovery
for these samples was 86.7% ± 5.8%. Field blank samples (n=3) were prepared and exposed to
field conditions. All were below the limit of detection. Laboratory blank samples (n=3) were
center-cut, quartered, extracted and analyzed in the same manner as field patches, and were below
the limit of detection. Calcofluor extraction efficiency (n=10) was determined by spiking 5.08 x
5.08 cm alpha-cellulose pads with either 0.25 ug or 2.50 ug of calcofluor in toluene. Samples
were allowed to dry under a fume hood for 1 hour. Each pad was then paired with a cover pad and
extracted and analyzed as described for the field samples. Ten controls were also prepared by
spiking calcofluor directly into 30 ml of toluene. The extraction efficiencies for the low and high
spiking levels was 99.1 ±2% and 94.6 ± 2.5%, respectively. Field blank samples (n=3) were
prepared and exposed to field conditions. All were below the limit of detection. Laboratory blank
samples (n=3) were below the limit of detection.
4.5 RESULTS
In the Year 01 study 33 applications were monitored involving six workers: nine in which
Sontara was worn and eight each in which the other garments were worn. The number of 500 gal
tanks applied varied from 1.5-4, with one worker spraying 1.5 tanks, four workers spraying 3
tanks, and the remaining 28 workers spraying 4 tanks. Tracer concentration was maintained at 300
17
-------
gm/tank for all applications, but ethion concentration varied substantially. In 8 cases no ethion was
used (pest control practices dictated use of another insecticide or no insecticide on these days). In
14 cases the rate was 5 pt (Ethion 4EC)/tank, and in the remaining 11 cases the rate was 12 pt/tank.
The total amount of tracer and ethion AI applied thus varied from 0.4-1.2 and 0-10.9 kg,
respectively. No effect was observed from using the same worker for several applications, so all
applications were treated as independent events for statistical purposes.
In the Year 02 study 18 applications were monitored involving ten workers: nine in which
Sontara coveralls were worn and nine in which 100% cotton coveralls were worn. The number of
500 gal tanks applied varied from 1.5-4, with one worker spraying 1.5 tanks, four workers
spraying 3 tanks, and the remaining 13 workers spraying 4 tanks. Tracer concentration was
purposely increased, as discussed previously. Ethion concentration again varied substantially: in
2 cases no ethion was used, in 7 cases the rate was 5 pt (Ethion 4EC)/tank, and in the remaining 9
cases the rate was 12 pt/tank. The total amount of tracer and ethion AI applied thus varied from
0.8-4.8 and 0-10.9 kg, respectively. No effect was observed from using the same worker for
several applications, so all applications were treated as independent events for statistical purposes.
4.5.1 Imaging Analysis
Fluorescent tracer exposure measurements produced by video imaging analysis were
normalized to reflect a standard application of four tanks. These normalized values were then
divided by 1.113 hr, the average time required to apply the four tanks. This application time value
was derived from data collected in the Total Exposure Distribution Study (Section 5), in which
application time was carefully monitored. Due to staff time constraints total work time rather than
application time was recorded during the Year 01 study and could not be used as a reliable
adjustment factor. Examination of hourly exposure values (Table 4.1) indicated that tracer
exposure beneath protective clothing was greatest for the forearms in all cases. These data also
indicated that forearm exposure was lowest for the workshirt (34 (ig/hr), and that the
cotton/polyester coverall was lower than either of the nonwoven coveralls (64 ng/hr for C/P
coveralls vs. 87 and 93 (ig/hr for SMS and Sontara garments, respectively). A similar exposure
pattern was observed for the upper arms, but was not evident for the torso. Variability within each
garment group was very high for all body regions, with coefficients of variation ranging from 89-
260%. Neither parametric (ANOVA) nor nonparametric (Kruskal-Wallis) tests between garment
types yielded significant differences for any body region.
18
-------
Table 4.1 Video Imaging Analysis of Fluorescent Tracer Exposure (|ig/hr)*
Body Region
N
Mean
Median
Range
C.V. (%)
FOREARM
WS/WP
C/P Coverall
SMS
Sontara
8
8
8
9
33.8
64.4
86.7
92.8
32.8
70.1
39.9
44,2
2-73
2-170
4-214
9-362
74
89
102
124
UPPER ARM
WS/WP 8
C/P Coverall 8
SMS 8
Sontara 9
1.4
12.3
17.9
21.5
0.4
0.3
7.9
7.5
0- 7
0- 92
0-100
1- 96
169
260
88
149
TORSO
WS/WP
C/P Coverall
SMS
Sontara
8
8
8
9
19.7
37.2
22.0
29.5
9.9
12.1
3.8
4.4
0- 82
2-168
1 - 127
0 - 139
140
155
196
163
*Data have been normalized to application of 4 tanks per subject (exposure x 4/# tanks applied) and
by time applied.
A substantial amount of the variability observed across garment types was believed to be
due to differences in garment challenge; i.e., the amount of fluorescent tracer reaching the outside
of the garments and the exposed skin surfaces. Head exposure provides an indication of the
fluorescent tracer challenge which each worker received d uring application, since none of the
workers wore personal protective equipment for this region (Fenske 1988). Exposure data for the
forearms, upper arms and torso were therefore normalized by the average head exposure (96.7
|ig/hr) for the entire study group as follows: a challenge adjustment factor was calculated by
dividing the group mean head exposure by each individual's head exposure; each individual's
forearm, upper arm and torso exposure values were then multiplied by this adjustment factor to
produce normalized exposure data for these body regions. If differences in individual challenge
are contributing to the variability observed within garment groups, then this adjustment should
19
-------
reduce within-group variability and allow a more direct assessment of the effect of garment type on
exposure to protected regions. The adjustment resulted in a decrease in the coefficient of variation
in 10 of 12 cases, with the range of CVs reduced from 89-260% to 64-192% (Table 4.2). The
pattern of exposure between woven and nonwoven garments remained similar to that observed in
the original data set, but the pattern within nonwoven garments was altered such that the SMS
garment exhibited higher adjusted exposure than the Sontara garment for all body regions.
Statistical analysis of the challenge-adjusted data by the Kruskal-Wallis test (KW: non-parametric
analysis of variance) indicated the following (Table 4.3): forearm exposure was significantly
higher for the SMS garment than for the other three garments; forearm exposure was also
significantly higher for the Sontara garment than for the woven garments; upper arm exposure was
significantly higher for the Sontara garment than for the two woven garments; upper arm exposure
was probably higher for the SMS garment than for the workshirt and woven coveralls, but
differences were not statistically significant; no significant differences in torso exposure were
observed. The detection of high levels of tracer on the forearms for the nonwoven garments
suggests that dermal exposure occurred by spray entering through the sleeve opening. The
detection of relatively high levels of tracer on the upper arms for the Sontara garment suggests that
both penetration and deposition through the sleeve opening contributed to exposure.
4.5.2 Visual Observations
Qualitative scores based on visual observations following application corresponded well to
the imaging analysis results (Figures 4.1 and 4.2). Torso exposure was not significantly different
across the garment types (ANOVA: p<.05), but both upper arm and forearm exposures were
different. Qualitative scoring indicated even more pronounced differences between the woven and
nonwoven garments for the arms, and for the forearms in particular. It was also apparent during
visual observation that arm exposure decreased with increasing distance from the wrist, and that
most torso exposure occurred at or near the neck. These observations suggest that: in the majority
of cases the tracer was being deposited on skin by movement under the garment rather than
through the fabric.
20
-------
o\
r-
«n
It
£•
O
OO
CO
3
f
i
-------
Table 4.3 Nonparametric Analysis of Variance of Garment Types by Body Region
KRUSKAL-WALLIS P-VALUES
Garment
Comparison
WS/WP = C/P Coveralls
SMS>WS/WP
SMS > C/P Coveralls
SMS > Sontara
Sontara> WS/WP
Sontara > C/P Coveralls
Forearm
NSD*
.001
.002
.02
.03
.02
Upper Arm
NSD
.09
.16
NSD
.004
.01
Torso
NSD
NSD
NSD
NSD
NSD
NSD
* NSD = no significant difference
DERMAL EXPOSURE fog/hr)
ro *k o> oo c
0 0 0 0 0 C
1 . 1 . 1 • 1 . 1 • 1
•
I
Forearm
Upper Arm
Torso
am
—
i
1
_
— T—
I
1
' i '
I
\
workshirt C/P coverall
SMS
Sontara
GARMENT TYPE
Figure 4.1 Video imaging analysis of fluorescent tracer exposure
22
-------
Forearms
Upper Arms
Torso
workshirt C/P coverall SMS
GARMENT TYPE
Sontara
Figure 4.2 Qualitative evaluation of fluorescent tracer exposure
4.5.3 Estimated Ethion Exposure to Protected Regions ,
Ethion exposure (Table 4.4) was estimated by multiplying the fluorescent tracer exposure
data in Table 4.1 by the average ratio of ethion and tracer deposited on outer patch samplers on the
upper region of the body (chest (chest, shoulder and head). Since workers applied widely varying
amounts of ethion, average ethion/tracer ratios were calculated for applications with 5 pt Ethion 4
Miscible™/tank and 12 pi/tank. These ratios averaged $.90 ± 4.4 and 21.34 ±8.4 for 5 pt and 12
pt tank concentrations, respectively. Despite a broad range of ratio values within each group (4-19
and 9-35, respectively), the proportion of the average ratios was virtually identical to the 2.4
proportion of pt/tank (I2pt/5pt).
23
-------
4.5.4 Ethion and Tracer Penetration of Protective Clothing
Penetration of ethion through protective clothing in Year 01 was measured on the legs with
inner and outer patch samplers (Table 4.5). Mean ethion challenge (deposition on the outer patch
sampler) was similar for the three coverall garments, ranging from 14.0-32.3 |ig/cm2, but was
much lower (1.44 ng/cm2) for the workpants (Kruskal-Wallis; p<.01). Percent penetration was
calculated dividing the inner patch sampler value by the outer patch sampler value and multiplying
by 100. Garment breakthrough occurred in all of the 23 applications for which complete data were
available. Mean penetration values for the four garments were quite similar, ranging from 4.7 -
7.2%, and did not differ significantly. Within each garment type the percent penetration was
highly variable, suggesting factors other than fabric content and construction contributed
substantially to penetration under field conditions.
Penetration of the tracer through the garments was also monitored in Year 01 (Table 4.6).
Mean challenge values followed a pattern similar to that of ethion (nonwoven coveralls>C/P
coveralls>workpants). The difference between the Sontara garment and the workpants was
significant (KW: p<.04), and the challenge to the SMS garment was marginally higher than to the
OP coveralls (p<.07) and to the workpants (p<.09). Garment breakthrough was measured in all
27 applications for which complete data were available. Mean penetration values for the four
garments exhibited a broader range (0.7 - 7.6%) than for ethion. The high SMS penetration value
was marginally greater than both the Sontara (p<.06) and workpants values (p<.09). As with
ethion, percent penetration for each garment type was highly variable.
In Year 02 ethion challenge and penetration was measured at the chest and upper legs for
cotton and Sontara coveralls (Table 4.7). Challenge was consistently higher for the legs when
compared to the chest, but for each region no significant difference was observed between the two
garment types. Penetration at the legs was greater for the cotton coveralls than for the Sontara
coveralls (2.7 vs. 0.8; KW: p<.02). The same pattern was observed for the chest but was not
statistically significant due to high variability within each garment type (5.4 vs. 1.4; KW: p=. 17).
Tracer challenge and penetration values were not calculated for Year 02 due to the high variability
in the amount of tracer applied. Ethion penetration of the Sontara garment was much lower in Year
02 than in Year 01 (0.8% vs. 6.3% penetration at the legs). Comparing the woven coverall
garments across years indicated that the 100% cotton coveralls performed more effectively than the
cotton/polyester coveralls (2.7% vs 4.7% penetration at the legs).
24
-------
Table 4.4 Ethion Exposure Estimates for Protected Body from Video Imaging Analysis (ng/hr)
Body Region
N
5 FT
Mean Range
12 FT
Mean Range
FOREARM
WS/WP
C/P Coverall
SMS
Sontara
8
8
8
9
300.8
573.2
771.6
825.9
18- 650
18 -1513
36 -1905
80 - 3222
721.3
1374.3
1850.2
1980.4
43 -1558
43 - 3628
86 - 4567
192 - 7725
UPPER ARM
WS/WP 8
C/P Coverall 8
SMS 8
Sontara 9
12.5
109.5
159.3
191.4
1- 62
1-819
1-890
9-854
29.9
262.5
382.0
458.8
0- 149
6 -1963
0-2134
21 - 2049
TORSO
WS/WP
C/P Coverall
SMS
Sontara
8
8
8
9
175.3
331.1
195.8
262.6
0- 730
18 -1495
9-1130
0 -1237
420.4
793.8
469.5
629.5
0- 1750
0- 1963
21-27104
0 - 29669
*Fluorescent tracer data (Table 4.1) has been multipled by the ethion / tracer ratio for each Ethion
4 Miscible application rate (8.90 for 5 pt/500 gal; 21.34 for 12pt/500 gal)
25
-------
i
I*
CO
DO
2
§
PM
1
u
«n
•^
i
I
1
O
I
vq
cs
cs
oo
I
es
O
en
vd
v
vq
o\
en
cs
en
o\
en
oo
i
00
00
•i.
o<
t< tj "S
n^2
&S|
§| §
ulu
•< « u
-------
EQ D
/h 00
O §
r M O!
fc H w
8^
D£ Ll^
/"•N. frl »^ •-«
C^ 5T^ ^H S
S ^ g .8
1 S
>— i
o
^ '
i
id Penetration to the Legs
IALLENGEA
Median CV
3 T^
2^ u
oo
1 I
I -•
g' . '
: 1
o
CvJ
o
5
en
o
1—H
VO
o
C5
r-
g
?
oo
oo
ON
00
VO
u
\o
en
en
a
r~
-------
g .
ITj Mft
oo
s
en
I
CO
00
8
1
I
U
ts
1
f-;
£
oo
T-H
cs
en
00
oo
en
1
oo
!
oo
00
s
oo
00
a-s
§
C/3
§
00
il,
28
-------
SECTION 5
TOTAL EXPOSURE DISTRIBUTION STUDY
5.1 OBJECTIVES
This study was designed to examine distribution of dermal exposure across body regions
and the contribution of respiratory exposure relative to dermal exposure during routine pesticide
applications in citrus orchards. The objectives of the study were as follows:
1. determine the relative contributions of hand exposure, face/neck exposure, exposure to
regions protected by coveralls, and respiratory exposure to total exposure.
2. determine the effectiveness of woven and nonwoven coveralls in reducing total exposure.
3. determine appropriate techniques for estimating face exposure.
4. construct scenarios in which the effectiveness of interventions with personal protective
equipment could be evaluated quantitatively.
5.2 STUDY DESIGN
Participants conducted replicate applications of the insecticide, Ethion 4 Miscible™, under
normal field conditions. Two protective coveralls (cotton and Sontara) were assigned to
applicators on a random basis. In one-half of the replicate applications protective gloves were
worn, also assigned on a random basis. All applicators wore plastic face shields and air sampling
equipment
5.3 METHODS
5.3.1 Field Conditions
The study occurred during one week of a summer application of ethion in central Florida
citrus groves. All applicators applied Ethion 4 Miscible™1 at a rate of 5 pts/500 gal. Initial
applications involved two tanks, but the research staff found that this schedule did not allow
sufficient time for proper collection and storage of samples. The remaining applicators applied
29
-------
three 500 gal tanks. -Participants were three adult male employees from Cooperative A who applied
pesticides as part of their normal work activities. Each read and signed a consent form and was
paid a nominal sum for each day of participation.
Protective Clothing Two types of protective coveralls were tested. Fabric descriptions and
characteristics are drawn from DeJonge and Easter (1989) and Nigg et al. (1992):
1) Cotton coverall (woven): a 100% cotton denim material (twill weave construction); 0.66
mm thickness; 274 gm/m^ weight; untreated.
2) Sontara coverall (nonwoven): 50% polyester, 50% wood pulp nonwoven material with
both point bonded and spun bonded construction (spunlaced composite); 0.2 mm
thickness; 43 gm/m^ weight; treated with DuPont RF™, a repeUent finish, exact
commercial formulation unknown.
Twelve replicates of each garment were conducted, with each participant wearing each type of
garment four times.
5.3.2 Sampling
Patch Sampling: The traditional patch technique recommended for applicator exposure assessment
(USEPA1987) was employed with minor modifications. Twenty alpha cellulose patches (see
Section 4.2.2) were positioned on each worker. The outer patches were pinned with two safety
pins so that the polyethylene layer backing the patch was adjacent to the protective garment. The
inner patches were pinned to the inside of the protective garment so that the polyethylene layer
faced the subjects' skin. Patches were positioned so that the outer patches were adjacent to, but
not overlapping the inner patches. A pair of patches (one outer and one inner patch) were attached
in the following locations: upper legs.(4), lower legs (4), upper arms (4), and lower arms (4),
chest (2), and back (2). In addition, a patch was attached to the front of the hat (positioned
vertically) for the!6 cases in which a hat was worn.
On the first day (Workers 1-6) patches were attached to the inside and the outside of the
protective garments while the worker was wearing the garment. On subsequent days the patch
locations were measured from garments used on the first day and the patches were positioned on
the clothing before the garment was given to the applicator. This allowed substantial reduction in
the amount of time required to prepare each volunteer for spraying. Outer patches were removed
immediately after the worker completed spraying. Patches were placed onto aluminum foil with
the plastic side facing down. A blank piece of alpha-cellulose was placed on top as a cover. The
30
-------
patch and cover were then wrapped in the foil and placed in a cooler with dry-ice. The inner
patches were removed after the worker had removed his garment and were stored in the same
manner as the outer patches.
! • - . J
f- - ;
Handwash Sampling: After a worker was suited in a protective garment, both hands were washed
with ethanol (reagent grade, Fisher Scientific) by the following procedure:
1. the hand was placed in a plastic bag containing 250 ml ethanol
2. the mouth of the bag was wrapped tightly around the wrist
3. the participant was instructed to relax the hand
4. a staff member shook the hand in the solution for 30 sec.
This procedure was repeated twice for each hand. The pre-handwash solutions were discarded.
After the worker had completed spraying his tanks, the handwash procedure described above was
repeated. All workers' hands were washed regardless of whether they were bare-handed or
wearing gloves. The handwashes from both hands (1000 ml total volume) were combined
immediately in a large mason jar and mixed well. An aliquot (approximately 100 ml) of this
solution was poured into a glass jar and frozen in the cooler with dry ice.
Faceshield Wipe Sampling: All workers wore face shields (Protecto-Shield, Willson, 17 x 31 em
cellulose acetate, 0.102 cm thick) which extended from their forehead to chin. The shield was
open at the top, bottom and sides and was suspended approximately 5 cm in front of the worker's
face. The shields were pre-wiped with ethanol. When the worker returned from spraying, his
faceshield was removed by the field staff. A gauze pad was moistened with 4 mists from a spray
bottle of ethanol. The entire face of the shield was then wiped with the pad using three horizontal
strokes. This procedure was repeated with a second pad. Both pads were stored in a single
sample jar.and placed in the cooler with dry ice.
Air Sampling: Aerosol samples were collected on glass fiber filters (Gelman type A/E) using 10M
Personal Inspirable Dust Samplers (Rotheroe and Mitchell) attached to the collar and positioned in
the breathing zone. ORBO-44 tubes containing Supelpak 20 (Supelco, Inc.) were placed in line
behind the samplers to collect volatilized ethion. Gilian pumps were adjusted to maintain an air
flow of 2 L/min through the samplers. The airflow of each sampler was standardized with a
rotameter just prior to placement on the worker. Air sampling began after the pre-handwash and
ended when the spray activities were completed. Air sampling times were affected by the number
of tanks applied, spray equipment problems, and variations in the size and density of the trees
being sprayed. At the end of sampling the glass fiber filter was removed from the sampler with
tweezers and placed in a 4 oz glass jar. The ORBO-44 tube was capped on both ends and wrapped
31
-------
in foil. Samples were placed in a cooler with dry ice.
5.3.3 Analysis
All field samples were stored in a cooler with dry-ice. At the end of the day, samples were
transferred to a freezer (approximately 0° C) for overnight storage, and then transferred to a -10°
C freezer the following morning. At the end of the field study, all samples were packed in a cooler
with dry-ice and transported overnight to New Jersey where they were stored in a -20° C freezer
until analysis.
Patch Samples: The outer patch samples were analyzed in the same manner described in Section
4.3.3, with an injection volume of 1 ul. The inner patches from several workers were analyzed
using the same procedure, but most were below the analytical limit of detection. To improve
detectability the cover portion for each remaining patch was discarded, and the patch was center cut
to 5.08 x 5.08 cm and then quartered. These four pieces were placed into a glass jar with 10 ml of
toluene and were extracted and analyzed according to the procedures for the outer patches.
Handwash Samples: The handwash samples were removed from the freezer and allowed to warm
to room temperature. The ethion content of the samples was analyzed directly by gas
chromatography (Hewlett-Packard 5890A, electron capture detector) under the following
conditions: Alltech RSL-200 capillary column (15m x 0.32mm); oven temperature 215° C,
injection port temperature 230° C, detector temperature 300° C, column flow 2.0 ml/min He,
auxilary gas 35 ml/min N. The retention time for ethion was 3.9 min under these conditions. Two
1.0 ul injections were made for each sample and averaged.
Faceshield Wipe Samples: The gauze faceshield wipes were removed from the freezer, allowed to
warm to room temperature for 2 hr, and then placed in sample jars with 30 ml of toluene. The jars
were placed on a shaker table at high speed (100 cpm) for 1 hr. The gauze wipes were removed
from the jar and discarded. The extracts were stored in the freezer (-20°C) until analysis.
Air Samples: Toluene (25 ml) was added to the jars which contained the glass fiber filters. These
jars were placed onto a mechanical shaking table (100 cpm) for 30 min. The filters were
discarded. The extracts were analyzed by gas chromatography. The contents of the Orbo-44 tubes
were removed from the glass tubes and placed into vials with 10 ml of toluene. These were placed
on the mechanical shaking table for 5 min. The vials were allowed to sit for 30 min after shaking
to allow the granular contents to seperate from the extract The granules were discarded and the
extracts were analyzed by gas chromatography.
32
-------
5.3.4 Quality Assurance/Control
The patch samples collected in this part of the Year 02 study were handled in a manner to
those discussed in Section 4.4.4. Thus, the quality assurance/control information presented there
pertain to these samples. This study included a number of sample media which were not a part of
the Protective Clothing Performance Study; i.e., handwashes, faceshield wipes, and air samples.
Handwash samples were quantified in ethanol, obviating the need for an extraction. However, it
was determined that the ethion analytical standard, which had been prepared in toluene, produced a
higher response factor than ethion in ethanol. Therefore, the relative response of the GC to ethion
in these two solvents was measured. Ethion in ethanol produced a response which was 62.1 +
2.9% of the ethion/toluene standard response. All field saimple values were thus adjusted
(increased) by this factor. Handwash removal efficiency (i.e., the fraction of pesticide on the skin
which is removed by the handwash procedure) was not determined for ethion. Three replicate
handwash field spikes were prepared at two concentrations by adding 20 ug or 100 ug of ethion
formulation to 100 ml of the ethanol handwash solution. The same amounts of ethion formulation
were spiked in duplicate directly into toluene as controls. Mean recoveries of ethion for the low
and high spikes were 113 ± 9.6% and 88 + 1.0%, respectively, with an overall mean value of
100.5 + 15.2%. Field blank samples were prepared by placing ethanol into a plastic handwash
bag, pouring the contents into a mason jar, then pouring the contents into a sample jar for storage
and analysis. None of the handwash field blanks contained detectable ethion.
Faceshield wipe extraction efficiency samples (n=6) were prepared by spiking surgical
gauze pads with either 2.3 ug or 6.9 ug of ethion. Each sample consisted of two gauze pads which
were extracted and analyzed together. Samples were allowed to dry under the fume hood for 1 hr,
and were then extracted and analyzed in the same manner described previously for the field
samples. Control spikes were prepared by the same amo'iints of ethion directly into 30 ml toluene.
Mean recoveries at the low and high spike levels were 104 +6.0% and 109 ±1.6%, with an overall
recovery of 106 ± 4.5%. Laboratory blank samples (n=2) were prepared by placing gauze wipes
under the fume hood for 2 hrs and as above. One sample contained no ethion and the other
contained <0.3 ug. To determine the removal efficiency of the faceshield wipe procedures shields
were spiked with either 100 ng or 300 ^ig of ethion (n=6 in each case). The faceshields were then
wiped according to the sampling procedure described in Section 5.3.2 and analyzed as described in
Section 5.3.3. Average removal efficiency (amount removed from shield by wipe divided by
amount on shield times 100) was 53.0 + 10.4%. This average value was used to adjust the
faceshield wipe data.
33
-------
Air filter extraction efficiency samples (n=6) were prepared by spiking 4.8 ug of ethion
onto each'glass fiber filter. The filters were allowed to dry under the fume hood for 1 hr, and then
each was placed in a glass jar with 25 ml of toluene and extraced. Three controls were prepared by
spiking the same amount of ethion directly into jars with 25 ml of toluene. Mean recovery was 98
± 8.6%. Field spike samples (n=2) were prepared by spiking 0.1 ug of the ethion formulation
directly onto the glass fiber filters. Controls were prepared by spiking the same amount of ethion
into 100 ml of toluene. The recovery for these samples was 107 ± 4.4%. Field blank samples
(n=4) were prepared by using tweezers to place glass fiber filter into sample jars. No ethion was
detected on these samples.
5.4 RESULTS
Twenty-four applications were monitored: 12 in which the cotton coverall was worn and 12
in which the Sontara coverall was worn. The first six applications consisted of two 500 gal tanks,
while the remaining 18 consisted of three tanks. Ethion concentration was maintained at 5 pi/500
gal for all tanks. The total amount of ethion AI applied thus varied from 2.3-3.4 kg. All data have
been normalized to three tanks (ethion AI applied = 3.4 kg). All exposure data are expressed as
hourly rates (jig/hr) based on a measured application rate of 17 min/tank. Worker 17 appeared as
an extreme outlier throughout the data set (very high values), and has therefore been excluded from
statistical calculations.
5.4.1 Respiratory Exposure
Measureable ethion was found in all glass fiber filter samples. No ethion was detected in
the back-up ORBO tubes samplers. Respiratory exposure was calculated by multiplying the air
concentration measured in the worker's breathing zone (|0.g/m3) by a standard factor respiratory
volume rate of 1.5 m^/hr (USEPA 1989). This rate is the maximum rate for light work, and
appears appropriate for workers conducting orchard applications. Respiratory exposure averaged
31.2 ng/hr, ranging from 1-207 jig/hr (Table 5.1).
34
-------
Table 5.1 Respiratory, Hand, Face, and Head Exposure to Ethion (jig/hr)
Body Region
Respiratory
Hands
Gloves
No Gloves
FaceA
Torso estimate
Hat estimate
N
22
12
12
21
13
Mean
32
1/762C
13,812C
965D
228°
Median
18
805
12,213
858
135
Range
1-
193-
2040-
26-
37-
207
9370
23570
2305
684
CV(%)
140
154
43
76
84
Head8
Torso estimate
Hat estimate
21
13
1,752E
413E
1,557
246
47- 4184
67- 1242
76
84
A Mean torso patch rate or hat patch rate x 650 cm2
B Mean torso patch rate or hat patch rate x 1180 cm2
C Significantly different (ANOVA: p < .0001)
D Significantly different (ANOVA: p < .002)
E Significantly different (ANOVA: p < .002)
5.4.2 Hand Exposure
Hand exposure without gloves averaged 13,812 fj.g/hr, ranging from 2000-23,000 |ig/hr
(Table 5.1). When nitrile gloves were worn exposure decreased nearly 8-fold to 1,762 (ig/hr, with
a range of 193-9,370 |J,g/hr (ANOVA: p<.0001). Variability in hand exposure was substantially
higher among workers who wore gloves, indicating that such exposures are probably the result of
intermittent events (e.g., glove removal) rather than chemical breakthrough. Clearly use of gloves
substantially reduced, but did not eliminate hand exposure.
5.4.3 Face and Head Exposure
Face and head exposure have traditionally been calculated by determining the average of
four torso patch samplers (left and right upper arms, chest, back) and multiplying it by an
35
-------
appropirate standard factor for surface area (650 cm2 for face, 1180 cm2 for head; USEPA 1989).
This calculation yielded an average face exposure value of 965 |ig/hr, and an average head
exposure value of 1,752 [ig/hr (Table 5.1). Alternatively, it is considered acceptable to employ a
hat or head patch sampler in lieu of the torso patches (USEPA 1987). Using this approach, face
exposure was 228 Hg/hr and head exposure was 413 jig/hr, significantly lower than the torso patch
extrapolations (ANOVA: p<.002).
The faceshield provided to workers was also sampled for ethion deposition. As discussed
in Section 5.3.4, spike/recovery studies indicated that only 53% of deposited ethion was
removeable by the wipe sampling technique employed. Thus, data from the faceshield wipes have
been divided by a .53 removal efficiency factor. Estimated face exposure based on faceshield
sampling averaged 105 p.g/hr (Table 5.2). When hat and torso patch extrapolations were applied to
the faceshield surface area (527 cm2), the respective face exposure estimates averaged 176 |ig/hr
and 782 |ig/hr. The torso patch estimate was significantly higher than the faceshield estimate
(ANOVA-SNK: p<.0001), while the difference between the head patch estimate and the faceshield
estimate was not statistically significant (ANOVA-SNK: p<.09). The discrepancy between the
faceshield/hat patch estimates and torso patch estimate efface exposure suggests that deposition on
the torso patches was not representative of deposition either at the location of the hat patch (above
forehead) or at the face itself. Thus, face and head exposure estimates derived from torso patch
extrapolation appear to overestimate exposure by a 7-fold or 11-fold factor.
36
-------
Table 5.2 Comparative Estimates of Face Exposure to Ethion
Face Estimate
N
Mean Median Range
CV
FaceshieldA
Hat Estimate8
Torso Estimate0
23
21
13
105DE
176E
782°
84
105
695
2-
29-
29-
321
530
1869
93
84
76
A Faceshield wipe values have been adjusted for 53% removal efficiency (value/.53)
B Hat patch rate x 527 cm2 (faceshield surface area)
C Mean torso patch rate x 527 cm2 (faceshield surface area)
D Significantly different (ANOVA: p <. 0001)
E Marginally different (ANOVA: p < .09)
5.4.4 Exposure beneath Coveralls
Deposition rates on outer patch samplers were similar for the cotton and Sontara coveralls,
with the exception of the upper legs (Table 5.3). The rate at this region for the Sontara coveralls
was nearly twice that for the cotton coveralls (ANOVA: p<.02). Total deposition rates on the
outside of the clothing were calculated by multiplying the outer patch sampler values by the
appropriate standard surface areas for the body regions (USEPA 1989). The total deposition rate
for the Sontara coveralls was 48% greater than the corresponding rate for the cotton coveralls, but
this difference was not statistically significant.
37
-------
^
>
u
1 a
w w
c* £
3 o
{3 K ^— x
^ <*><£: i
O ^"N Kh »rt
.• T; I J W*J tJ
1 S* 1
3 Q
a
0
00
T3 J
9 g
o O a
o *S
r i |^*|
V««^ r~~1
a
0
s
3
a
o
'•S
•*»<
t § t
§ -sill
•3 *3 52 *3 S
" ^5*1
«n
42
3
w
H hi-
a
o
•a
es ^H vd
vo r~> oo oo o ON ON
f- ON -- ON r- vo
es i> en es' ri o £g
,,,,..
T-J vc, "*' —5 21
CN «o oo i— 1 1— i oo en
•<* en >r> o o o •*
en vo Tf es t— i vo Tf
es WT
r- r-- ON vo --I vp ON
ON •<* O oo en O >n
es ON vo es o oo o
CN >n
Q
O en I-H o f*' ON ,
,O
di "^ "*~* C^
p o ^ .
^3 Siv'IS^
> 73 •§ 2
nil
•< » U Q
38
-------
Inner patch samplers were categorized as either 1) quantifiable (>0.84 )j.g/sample), 2) trace
(0.24-0.84 jig/sample), or 3) unexposed (<0.24 |j.g/sample). In the majority of cases garment
breakthrough occurred leading to exposure to the body regions protected by coveralls (Table 5.4).
For cotton coveralls 34% of the inner patch samplers had quantifiable ethion and an additional 29%
had trace levels, resulting in a breakthrough frequency of 63%. For the Sontara coveralls 26% of
the inner patch samplers had quantifiable ethion and an additional 43% had trace levels, resulting in
a breakthrough frequency of 69%. Since breakthrough frequencies for the two garments were
similar, these data were pooled to examine breakthrough frequency by body region (Table 5.5).
Quantifiable ethion was measured most frequently on the lower arm (forearm) samplers (43%) and
the upper leg samplers (43%). For all body regions breakthrough frequency (quantifiable and
trace) was >50%.
Table 5.4 CPC Breakthrough Frequency by Garment
Garment
Cotton
Coverall
SontaraD
Total
Patches
114
96
QuantifiableA
Ethion
39
25
Percent
34.2
26.0
TraceB
Ethion
331
41
Percent
28.9
42.7
Q + TC
Ethion
72
66
Percent
63.2
68.8
Coverall
A Quantifiable = > 28 pg/jjl; > 0.84 jig/sample
B Trace = < 28 pg/^1 and >8 pg/pl; .24 - .84 ng/sample
C Frequency of quantifiable + trace breakthrough
D One subject excluded (W17) due to very high deposition rates
39
-------
Table 5.5 CPC Breakthrough Frequency - by Body Region (both Garments)
Body Total QuantifiableA Trace8
Region Patches Ethion Percent Ethion Percent Ethion Percent
Chest
Back
Up Arm
LoArm
Up Leg
LoLeg
20
23
47
47
46
46
5
3
10
20
20
13
25.0
13.0
21.3
42.6
43.5
28.3
6
11
17
14
19
17
30.0
47.8
36.2
29.8
41.3
37.0
11
14
27
34
39
30
55.0
60.9
57.4
72.3
84.8
65.2
A Quantifiable = > 28 pg/|il; > 0.84 p.g/sample
B Trace = < 28 pg/jil and >8 pg/^il; .24 - .84 jig/sample
C Frequency of quantifiable + trace breakthrough
Exposure to regions beneath protective garments was calculated by multiplying the inner
patch sampler deposition rate by the appropriate standard surface area (Table 5.6). Only
quantifiable ethion and trace values were used, with trace values being assigned one-half the limit
of detection (.007 p.g/cm2); unexposed samples were assigned values of zero. Total exposure to
these regions was then determined for each worker and average "protected body" exposure
determined (protected body is defined here as all regions beneath coveralls.) Exposure beneath
cotton coveralls appeared to be lower than that beneath Sontara coveralls, based on inspection of
both the mean and median protected body exposure values. However, this difference was not
significant statistically due to high variability within each garment group. If the difference
observed in these mean values were significant, it would in part be attributable to the greater
challenge (deposition rates) received by the Sontara garment.
40
-------
Table 5.6 Ethion Exposure beneath Cotton Coveralls and Sontara Garments (|ig/hr)
Garment/Region
COTTON (N= 12)
Upper Arm
Lower Arm
Chest
Back
Upper leg
Lower leg
Mean
11.4
37.6
20.5
13.2
31.6
112.4
Median
6.6
12.1
0
13.2
26.1
12.0
Range
0 -
0 -
0 -
0 -
0 -
0 -
61
243
121
26
85
1,214
CV (%)
150
184
173
104
74
309
Protected Body
SONTARA (N= 11)
226.7
104.0
5 - 1,540
187
Upper arm
Lower arm
Chest
Back
Upper leg
Lower leg
Protected Body
13.9
80.2
44.4
29.2
44.7
45.7
258.1
13.3
10.6
26.4
26.4
28.8
19.2
152.4
0 -
0 -
0 -
0 -
9 -
0 -
9 -
40
571
147
147
154
308
1,272
92
212
111
143
98
192
136
5.5 EXPOSURE DISTRIBUTION ASSESSMENT
The distributional characteristics of exposure are important in that they indicate the
effectiveness of specific interventions for reducing exposure, and provide data for recommending
additional interventions. Numerous applicator exposure studies have reported the distribution of
dermal exposure across body regions, and the relative contribution of respiratory exposure to total
exposure, but most often these studies have lacked specificity regarding methods of calculations,
use of personal protective equipment (PPE), and underlying assumptions. Furthermore, traditional
sampling techniques may have underestimated exposure beneath protective clothing due to
deposition through garment openings, as documented in Section 4. As a result, generalizations
which are sometimes cited regarding exposure distribution may be inaccurate. Based on the data
collected in this study a series of exposure scenarios has been developed to identify the role of PPE
in exposure reduction. It is believed that these data are representative of airblast applicator
exposure in citrus orchards, and may be representative of orchard airblast exposure in general.
41
-------
They are not, however, applicable to other types of pesticide applications (e.g., groundboom,
backpack), nor do they reflect exposure patterns of pesticide mixers or mixer/applicators.
Label requirements for Ethion 4 Miscible™ (January 1991 label) require that a worker wear
the following personal protective equipment (PPE) during application:
• Protective suit of one or two pieces covering all parts of the body except the head, hands
and feet;
• Chemical resistant gloves and shoes
• NIOSH or MESA approved respirator
In practice, these requirements are not followed consistently during summer spraying of citrus in
Central Florida. Indeed, there is substantial evidence to suggest that such requirements place an
undue burden on workers and may contribute to physiological conditions related to heat stress
(Nigg et al. 1992). It is not uncommon for workers applying under high temperature and high
humidity conditions to forego the use of a respirator, and to alter protective suits in a manner that
allows greater air circulation to the body.
5.5.1 Dermal Exposure Scenarios
In light of the realities of actual field use of PPE cited above the following scenarios have
been constructed to assess the role of specific PPE combinations in reducing dermal exposure.
Exposure estimates generated by these scenarios are presented in Table 5.7. Since this study did
not measure exposure to the feet, the use of chemical resistant shoes or boots is not discussed, and
exposure to this body region is assumed to be zero in subsequent calculations. Unfortunately, one
PPE option « chemical resistant hoods - was not investigated in this study. Hoods would appear
to provide substantial protection for all portions of the head except the face; however, no published
studies are available to demonstrate the effect of hoods on head exposure.
42
-------
s
Si'
s
<0
a
ex
•3
a*
^.
TJ
O
p
£
1
1
>>
.0
e
o
1
•a
o
Cu
^5
w
13
CD
P
C-
w-j
1
H
13
1
P
i-g
it
tS
iDUCHON
s
K
TrT
1
2
fc
JZJ
fi-1
y
^
CM
w
tD'i'
1*
3H
So5
X
H
0,
^ *2
S §
a,K
1^
PI;JQ
X ej>
rVl — ^
vs
Scenario 3
f-O
.2
>|
_o
o3
> §
oo
E Scenario
D-i
ffi
•
en
o
es
^o
«n
en
ON
T \
Cotton or Sontara Coverall1
+ Gloves
en
es
oo
vo
es
en
es
ON
£.
ON
f\f
oo
r— I
OO
ON
^j.'
ON
Cotton or Sontara Coverall
+ Gloves
+ Faceshield
'*
o
ON
ON
^
|
^5
v>
3
**n
en
vd
oo
r-
ON
en'
ON
i
Chem-Resistant CoverallsE
+ Gloves
">
o
>o
en
>o
T^-'
vo
en
ts"
en
f-
es
00
en
ON
Chem-Resistant Coveralls
+ Gloves
+ Faceshield
VO
y.B^
i
s
1
03
Q
*^
1
1
I
"2
§
_J_
«5
1
rt"\
Deposition to outside of cov<
<
fli
l_l
S3
§
s*
1
•o
3 1— <
•C i/^
"^ u
--2>3
S *e3
it
3 CN
Si2
Q ^X
^ *~^
o °
*g es
— °°
§ en*
v§g
c
o
• 1
a
Vs
8
I
i
H
.^•^ -2
11
si
••.•J ^5
^^ -i3
S«
*O c^
CTj »-S
O O-i
<<_, ^
° "S
&- g
^ 1
Deposition beneath coveralls
Gloves reduced exposure fro
Assumes faceshield protects
Assumes chemical-resistant <
» u
O H
43
-------
SCENARIO 1: The unprotected worker. This scenario assumes that workers use virtually no PPE
or that PPE is used in a manner which provides little protection. Thus, the hands, face and
Protected Body regions (regions beneath coveralls) are considered unprotected. Deposition rates
measured on the outside of coveralls have been used to estimate exposure to the Protected Body;
the mean exposure to the outside of coveralls has been used, since no significant difference was
noted between cotton and Sontara coveralls. The mean torso patch value (Table 5.1) has been used
to estimate head exposure here and throughout the scenarios to provide consistency with the
extrapolation procedures used for other body regions. However, it should be noted that data
presented previously (Section 5.4.3) suggest that use of this value may overestimate the
contribution of head exposure to total dermal exposure.
SCENARIO 2: Cotton or Sontara Coveralls only. Use of a protective coverall is added to
Scenario 1. The mean exposure value for cotton and Sontara coveralls extrapolated from inner
patch samplers was used to estimate exposure to the Protected Body (Table 5.6). Hand and Head
estimates remain unchanged. This scenario assesses the effect of the coveralls used in this study,
but assumes that the worker does not follow label requirements regarding gloves.
SCENARIO 3: Cotton/Sontara Coveralls + Gloves. Use of chemical resistant gloves has been
added to Scenario 2. Measured exposure beneath gloves was used to estimate Hand exposure
(Table 5.1). Head and Protected Body estimates remain unchanged. This scenario assesses the
effect of chemical resistant gloves on hand exposure, and is consistent with label requirements.
SCENARIO 4: Cotton/Sontara Coveralls + Gloves + Faceshield. Use of a 527 cm2 faceshield
has been added to Scenario 3. The exposure calculation assumes that the faceshield protects
44.7% of the head surface area (527/1180 cm2). Hand and Protected Body estimates remain
unchanged. This scenario assesses the effect of the faceshield when a worker is following label
requirements.
SCENARIO 5: Chemical Resistant Coveralls + Gloves. Chemical resistant coveralls have been
substituted for the cotton or Sontara coveralls used in the study, and the faceshield has been
removed. It has been assumed that these coveralls are 100% effective and that no exposure occurs
on Protected Body regions. Head exposure is that used in Scenarios 1-3. Hand exposure remains
unchanged from Scenario 4. This scenario assesses the effect of a truly chemical resistant coverall
on total exposure when a worker is wearing label-required protective clothing. It should be noted
44
-------
that no field studies to date have documented that commercially available coveralls perform in this
manner.
SCENARIO 6: Chemical Resistant Coveralls + Gloves + Faceshield. Faceshields have been
added to the PPE in Scenario 5 to create a scenario in which all PPE options are combined.
Dermal exposure to the unprotected worker (S-l) was primarily to the Protected Body
regions (73%), with hand exposure contributing nearly one-quarter (24%) of total exposure. The
use of cotton or Sontara coveralls (S-2) reduced total dermal exposure by 73%, and exposure to
unprotected hands became the primary contributor to total dermal exposure (87%). Thus, coveralls
play the most important role of any PPE in reducing exposure during citrus airblast applications.
The addition of chemical resistant gloves (S-3) further reduced dermal exposure to 94% of that
received by the unprotected worker. When compared with workers wearing coveralls the use of
gloves reduced total dermal exposure by 76%. Under this scenario the contributions of protected
hands and unprotected head were equal, accounting for more than 90% of total dermal exposure.
The addition of faceshields (S-4) produced further, but slight decreases in exposure (to 95%
compared to the unprotected worker; to 81% compared to workers with coveralls), and hands
again became the predominant source of exposure. When compared to Scenario 3, however, in
which workers followed label requirements, exposure was reduced by 21%.
In light of the partial failure of the coveralls evaluated in this study to prevent exposure, it
seems reasonable to ask whether improved coveralls would provide substantially greater
protection. If 100% effective coveralls had been worn with gloves (S-5), only a slight decrease in
exposure (to 94% compared to the unprotected worker; to 78% compared to workers with
coveralls; only 6% compared to coveralls + gloves) would have occurred, with remaining dermal
exposure distributed equally between the protected hands amd unprotected head. Thus, use of
faceshields would provide greater exposure reduction under these conditions than further efforts to
provide truly chemical resistant coveralls. The final scenario (S-6) indicates use of faceshields and
improved coveralls would reduce exposure by 27% when compared with the label-required PPE
used in this study.
5.5.2 Dermal and Respiratory Dose Scenarios
Estimation of absorbed dose from dermal and respiratory exposure values requires
quantitative absorption factors for each exposure route. The dose estimate produced here assumed
that oral exposure was negligible. While this assumption is probably correct under these study
conditions, oral exposure may occur when workers who ha.ve not washed their hands thoroughly
45
-------
handle food and cigarettes or place other materials in their mouths. An absorption factor of 1.0
(100%) was assumed for respiratory exposure. This factor likely overestimated respiratory dose
since some fraction of the inhaled aerosol may have been exhaled, and an additional fraction may
have been deposited in the airways, bound to mucous, and swallowed, thereby contributing to oral
exposure. Notwithstanding these considerations, the lack of an experimentally generated
absorption factor for ethion inhalation makes the assumption of 100% absorption the valid estimate
of dose at present.
An absorption factor of 0.033 (3.3%) was assumed for dermal exposure, based on
controlled percutaneous absorption experiments in humans (Feldmann and Maibach 1974). The
use of a single percent absorption value for dermal exposure is probably not accurate for several
reasons: 1) percent absorption varies with skin loading and can be affected by the vehicle in which
the toxicant is administered; in the experiment cited above loading was 4 (.ig/cm^ in acetone,
whereas in the field study ethion loading in a water vehicle was much higher in some cases (e.g.,
hands, unprotected head) and much lower in others (e.g., regions beneath coveralls); 2)
percutaneous absorption varies across body regions and across individuals; 3) workers' skin may
not have stratum corneum barrier properties intact relative to volunteers in controlled experiments.
Despite these concerns the 3.3% absorption value is the best available information regarding
dermal absorption of ethion.
Absorbed dose estimates (|ig/hr) have been generated by multiplying these absorption
factors by the mean exposure estimates calculated in this study (Table 5.8). Four scenarios have
been created to estimate the impact of PPE options on total absorbed dose. In all scenarios
workers wear the cotton or Sontara coveralls and nitrile gloves. This PPE combination serves as
Scenario 1. In Scenario 2 a respirator only is added. In Scenario 3 a faceshield only is added. In
Scenario 4 an idealized chemical resistant coverall only is added.
In Scenario 1 the total aborbed dose was 156 |J.g/hr, with the following distribution:
hands, 37%; head, 37%; inhalation, 20%; protected body, 5%. Thus, addition of a respirator (S-
2) reduced dose by 20%, roughly equivalent to the 17% reduction produced by use of a faceshield
(S-3). Use of completely effective coveralls reduced dose by only 5%. Label requirements for
respirator use appear justified by these calculations, since the respiratory route was responsible for
a substantial proportion of absorbed dose despite an extremely low exposure estimate relative to
total dermal exposure. Gains in reduction of total absorbed dose could also be made by providing
greater protection for the face and head.
46
-------
i
w
I
I
a
o
Q
I
oo
•H
I
c/5
g«
•s
O
UU +
U
o
^
« o a
47
-------
SECTION 6.
DISCUSSION
These studies have demonstrated that coverall garments similar to those used routinely by
pesticide applicators did not provide the levels of protection expected. Dermal exposures occurred
both due to garment design and chemical breakthrough of fabric. No significant improvement in
protection occurred when nonwoven garments were substituted for traditional woven garments.
Indeed the nonwoven garments suffered from the most serious flaws in design, and provided little
if any increased resistance to chemical penetration.
The use of fluorescent tracers and video imaging analysis clearly documented substantial
exposure to the arms of workers wearing garments with large sleeve openings. When this design
failure was rectified, little tracer exposure could be detected on the protected body. It appears that
the tracer/imaging analysis is most useful for measuring exposures occurring under rather than
through the garments, and in detection of exposures that otherwise would have been
undocumented by the patch technique. The use of patches to detect fabric penetration was far more
sensitive than tracer/imaging analysis. Low levels of tracer on skin were difficult to quantify by
video imaging, whereas chemical analysis of patch extracts detected <10 ng/cm2. The techniques
thus served complementary functions in documenting the limitations of chemical protective clothing
performance.
Analysis of exposure distribution revealed that further improvements in protective coveralls
would do little to reduce total dermal exposure or total absorbed dose of applicators under the field
conditions tested. Proper use of such personal protective equipment as gloves, respirators and
faceshields could provide greater exposure reduction than more chemically resistant coveralls. It
should be noted that hand exposure may have been even higher than the values reported here.
Recent studies in our laboratory indicated that only about 30% of the organophosphorous
insecticide, chlorpyrifos, in a liquid formuation, was removed from hands by the ethanol
handwash procedure used in this study (Fenske and Lu 1993). Further efforts should be directed
at establishing accurate hand exposure assessments methods.
The findings of this study are consistent with those of an earlier study of protective clothing
performance during airblast applications (Fenske 1998a; b). In that study exposure through
48
-------
garment openings (neck and sleeves) was determined to be the major pathway for dermal exposure
beneath coveralls. Findings regarding fabric penetration were also consistent with those of a recent
study in which similar garments were tested with a similar applicator population (Nigg et al. 1992).
In that study no significant differences (p<05) were demonstrated between woven and nonwoven
garments. The further finding by Nigg et al. that shoulder and hat patch deposition rates were
similar was not supported by data in this study. Indeed use of patch values from these different
locations produced dramatically different estimates of head exposure. Evidence from faceshield
sampling indicated that torso patches may have overestimated head exposure by an order of
magnitude.
The most important finding of Nigg et al. (1992) concerned the role of chemical protective
clothing in exacerbating heat stress, and was confirmed by our observations. Use of such
garments during high temperature, high humidity conditions places an excessive and potentially
dangerous burden on workers. Label requirements for CPC must be qualified by limits on
environmental parameters related to heat stress in order to strike a proper balance between
protection and comfort
49
-------
SECTION?.
REFERENCES
DeJonge JO, Easter E (1989) Laboratory Evaluation of Pesticide Spray Penetration and Thermal
Comfort of Protective Apparel for Pesticide Application. Final Report, U.S. Environmental
Protection Agency Cooperative Agreement 812486-01-0, Risk Reduction Engineering
Laboratory, Office of Research and Development, U.S. Environmental Protection Agency,
Cincinnati, OH 45268. ^
Feldmann RJ, Maibach ffl (1974) Percutaneous Penetration of some Pesticides and Herbicides in
Man. ToxicolApplPharmacol, 28:126-132.
Fenske RA, Leffingwell JT, Spear RC (1985) Evaluation of Fluorescent Tracer Methodology for
Dermal Exposure Assessment. In Dermal Exposure Related to Pesticide Use, ACS Symposium
Series Vol. 273, American Chemical Society, Washington, D.C., pp. 377-393.
Fenske RA, Leffingwell JT, Spear RC (1986a) A Video Imaging Technique for Assessing Dermal
Exposure -1. Instrument Design and Testing. Am Ind Hyg Assoc J, 47:764-770.
Fenske RA, Wong SM, Leffingwell JT, Spear RC (1986b) A Video Imaging Technique for
Assessing Dermal Exposure - H.. Fluorescent Tracer Testing. Am Ind Hyg Assoc J, 47:771-
775.
Fenske RA (1988a) Use of Fluorescent Tracers and Video Imaging to Evaluate Chemical
Protective Clothing During Pesticide Applications. In Performance of Protective Clothing:
Second Symposium, ASTM Special Technical Publication 989, SZ Mansdorf, R Sager, AP
Nielsen, Eds., pp. 630-639, American Society for Testing and Materials, Philadelphia.
Fenske RA (1988b) Comparative Assessment of Protective Clothing Performance by Measurement
of Dermal Exposure during Pesticide Applications. ApplIndHyg, 3:207-213.
Fenske RA (1988c) Visual Scoring System for Fluorescent Tracer Evaluation of Dermal Exposure
to Pesticides. Bull Environ Contam Toxicol, 41:727-736.
50
-------
Fenske RA (1990) Nonuniform Dermal Deposition Patterns during Occupational Exposure to
Pesticides. Arch Environ Contam Toxicol, 19:332-337.
Fenske RA, Lu C (1993) Determination of handwash removal efficiency: incomplete removal of
the pesticide, chlorpyrifos, from skin by standard handwash techniques. Am Ind Hyg Assoc J(
submitted for publication)
Fenske RA, Birnbaum, SG, Cho, K (1993) Second generation video imaging technique for
assessing dermal exposure. Am Ind Hyg Assoc J (submitted for publication).
Nielsen AP, Moraski RV (1986) Protective Clothing and the Agricultural Worker. In Performance
of Protective Clothing, ASTM Special Technical Publication 900, R.L. Barker and G.C.
Coletta, Eds., pp. 95-102, ASTM, Philadelphia.
Nigg HN, Stamper JH, Easter E, DeJonge JO (1992) Field Evaluation of Coverall Fabrics: Heat
Stress and Pesticide Penetration. Arch Environ Contamin Toxicol, 23:281-288.
USEPA (1987) Pesticide Assessment Guidelines. Subdivision U: Applicator Exposure
Monitoring. U.S. Environmental Protection Agency Office of Pesticides Programs. Document
No. PB87-133286, National Technical Information Service, Springfield, VA.
USEPA (1989) Exposure Factors Handbook. Exposure Assessment Group, Office of Health and
Environmental Assessment, U.S. Environmental Protection Agency. EPA/600/8-89/043.
51
•&U.S. GOVERNMENT PRINTING OFFICE: 1993 - 750-002/80258
-------
------- |