- Email: [email protected]
An Investigation of Factors Contributing to Styrene and Styrene-7,8-oxide Exposures in the Reinforced-Plastics Industry
\
Pergamon PII] S9992Ð3767"87#99979Ð4
Ann[ occup[ Hyg[\ Vol[ 32\ No[ 1\ pp[ 88Ð098\ 0888 Þ 0888 British Occupational Hygiene Society Published by Elsevier Science Ltd[ All rights reserved Printed in Great Britain[ 9992Ð3767:88 ,08[99¦9[99
An Investigation of Factors Contributing to Styrene and Styrene!6\7!oxide Exposures in the Reinforced!Plastics Industry LEENA A[ NYLANDER!FRENCH\ % LAWRENCE L[ KUPPER$ and STEPHEN M[ RAPPAPORT Department of Environmental Sciences and En`ineerin`\ CB (6399 Rosenau Hall\ The University of North Carolina at Chapel Hill\ Chapel Hill\ NC 16488!6399\ USA^ $Department of Biostatistics\ School of Public Health\ The University of North Carolina at Chapel Hill\ Chapel Hill\ NC 16488! 6399\ USA During the manufacturing of reinforced plastics\ large amounts of styrene and trace quantities of styrene!6\7!oxide "SO# are released[ Since previous work suggests that inhalation of even small amounts of SO might be an important health risk\ we investigated several possible factors contributing to styrene and SO exposure during the manufacture of reinforced plastics[ Factors related to job type\ worker and the type and quantity of styrene!containing resins were investigated using mixed!effects multiple linear regression models[ Overall\ SO exposure levels were positively correlated with styrene exposure levels[ However\ this correlation was statistically signi_cant only among hand laminators who had the highest exposures to both styrene and SO[ An important factor for predicting both styrene and SO concentrations was the type of resin used\ while the quantity of resin consumed was predictive of styrene but not of SO exposure[ Since SO exposure appears to be associated with factors other than coexposure to styrene\ more effort should be placed on investigating emissions of SO per se[ The type of mixed!models regression analysis employed in this study can be used for clarifying the underlying patterns for exposures to styrene and SO as well as for evaluating preventive measures[ Þ 0888 British Occupational Hygiene Society[ Published by Elsevier Science Ltd[ All rights reserved[ Keywords] styrene^ styrene!6\7!oxide^ exposure factors^ reinforced plastics
observed among reinforced!plastics workers\ Fleig and Thiess "0867# speculated about 19 years ago that coexposure to airborne styrene!6\7!oxide "SO# might also be important[ SO is a contaminant that is formed at low levels in air "³0 mg:m2# from oxidation of styrene during the manufacture of reinforced plastics "Brighton et al[\ 0868^ Fjelstad et al[\ 0868^ Pfa/i and Saamanen\ 0882^ Pfa/i et al[\ 0868#[ Despite this speculation of Fleig and Thiess "0867# and more recently by Pfa/i and Saamanen "0882# that SO exposure per se may be important\ the potential role of SO exposure in the etiology of adverse health e}ects in the reinforced!plastics industry has been lar! gely ignored[ This may be due to the fact that styrene is present at high concentrations in this industry and is metabolized almost exclusively through SO in humans[ ðAbout 83) of inhaled styrene is absorbed in humans "Petreas et al[\ 0884# and is metabolized to SO via hepatic cytochrome P!349 isozymes "Nakajima et al[\ 0882#Ł[ SO is subsequently metabolized in the
INTRODUCTION
Styrene is one of the most important monomers pro! duced globally for the manufacture of plastics and resins[ In 0881\ the production of styrene was over four million metric tons in the US^ one half to two thirds of the styrene was used to produce polystyrene and the remainder to produce a variety of copolymeric products and resins "IARC\ 0883#[ Although occu! pational exposure to airborne styrene is modest "³09 mg:m2# in most parts of the industry\ it tends to be relatively high "39Ð399 mg:m2# in the reinforced! plastics industry\ especially during the production of large objects such as boats "IARC\ 0883#[ Although exposure to styrene has generally been thought to be responsible for genotoxic e}ects Received 08 June 0887^ in _nal form 07 November 0887[ %Author to whom correspondence should be addressed[ Tel[] ¦0!808!855!2715^ Fax] ¦0!808!855!3600^ E!mail] leena*frenchÝunc[edu 88
099
L[ A[ Nylander!French et al[
liver by epoxide hydrolase to form styrene glycol and its oxidation products "mandelic acid and phe! nylglyoxylic acid#\ which are excreted in the urine ðreviewed by Bond "0878#\ IARC "0883# and Sumner and Fennell "0883#Ł[ Since styrene exposure tends to be 499Ð0999 times greater than SO exposure "Pfa/i and Saamanen\ 0882# and\ since virtually all styrene is metabolized to SO in humans\ we suspect that most investigators tended to discount the possible import! ance of SO exposure per se[ However\ recent evidence of e.cient detoxi_cation of SO in the human liver following metabolism of styrene "Csanady et al[\ 0883^ Filser et al[\ 0882^ Korn et al[\ 0883^ Mendrala et al[\ 0882# suggests that the systemic dose of SO derived from the metabolism of styrene might be small[ This was con_rmed by Rappaport et al[ "0885# who predicted\ from the measurements of SOÐalbumin adducts in the blood of 37 reinforced!plastics workers\ that only about 9[94) of the styrene dose was released to the blood as SO[ Indeed\ Rappaport et al[ "0885# estimated that the typical worker in their study received 60) of the systemic dose of SO from direct inhalation[ The objective of any occupational and epi! demiological investigation is to identify factors that contribute to exposure and the subsequent risks of disease so that interventions can be made to reduce exposures[ In the reinforced!plastics industry\ the con! trol of styrene exposure has been the major focus[ In fact\ special low!styrene!emission "LSE# resins have been developed for this industry[ Studies describing new methods that can be used to reduce the styrene exposures in the reinforced plastics industry have been published "Andersson and Rosen\ 0883^ Hopkins et al[\ 0875a^ Hopkins et al[\ 0875b^ Kalliokoski et al[\ 0873^ Kalliokoski et al[\ 0877^ Lof et al[\ 0882^ Saa! manen and Enbom\ 0883^ Todd\ 0874# and there is some indication of a decreasing trend in occupational exposure to styrene over the past three decades "Sym! anski et al[\ 0887#[ In general\ control measures are problematic in the reinforced!plastics industry where products are often large and di.cult to enclose or to ventilate e}ectively[ Since coexposure to SO may\ in fact\ be a more important risk factor for adverse health e}ects than is exposure to styrene\ strategies aimed at reducing SO emissions could be important[ Unfortunately\ very little information is currently available concerning the generation of SO during the production of reinforced!plastics[ It has been postu! lated\ based upon laboratory experiments\ that SO is formed via] 0[ fragmentation of polymeric styrene peroxide rad! icals resulting from the copolymerization of styrene and oxygen "Brighton et al[\ 0868^ Mark et al[\ 0877#^ 1[ epoxidation of the styrene monomer "Brighton et al[\ 0868# and:or 2[ reaction of styrene with volatile organic peroxides used to initiate the curing of reinforced plastics
"Brighton et al[\ 0868^ Fjelstad et al[\ 0868^ Mayo\ 0847^ Parkyn et al[\ 0856^ Pfa/i and Saamanen\ 0882^ Pfa/i et al[\ 0868^ Walling and Heating\ 0854#[ However\ no _eld investigations have thus far been reported to evaluate these speculations in the reinforced!plastics industry\ or to investigate possible job!related sources of exposure to SO[ In order to investigate the possible factors con! tributing to exposure to both styrene and SO in the reinforced plastics industry\ we used mixed!e}ects multiple linear regression models to analyze exposure data that had been previously collected in Finland ðpublished in part\ "Pekari et al[\ 0882^ Sorsa et al[\ 0880#Ł[ Our goal was to determine how factors related to the job\ worker\ resin formulation and use of styr! ene!containing resins a}ected both styrene and SO exposures in the reinforced!plastics industry[
MATERIALS AND METHODS
Study population Workers were recruited from 29 di}erent reinforced!plastics facilities in Finland during November 0877 and February 0889[ The plants ranged from small workshops with a few employees to fairly large factories[ Items manufactured included boats\ containers\ pipes and tubes\ small parts\ sheets and vehicle components[ Most of the sites involved boat manufacturing "69)#[ The most common lami! nation process involved hand lamination of open molds[ Other work tasks included spray lamination\ _lament winding\ low!pressure!sack lamination\ cen! trifugal lamination\ robot controlling\ assembly\ glass!_ber cutting\ waxing and cleaning\ carpentry and work supervision[ During data collection\ distinct occupational groups were identi_ed and workers were assigned to di}erent job groups according to the work tasks they performed[ Initially\ six di}erent job groups were established[ Group 0 consisted of laminators who hand laminated objects larger than 0 m1 in surface area[ Group 1 consisted of laminators who hand lami! nated objects smaller than 0 m1 in surface area[ Group 2 consisted of workers who used chopper guns and:or gel coat spraying during lamination[ Group 3 con! sisted of workers who used automated techniques "i[e[\ _lament winding\ low!pressure!sack lamination\ cen! trifugal lamination\ robot controlling#[ Group 4 con! sisted of workers who assembled laminated objects "both large and small in size# and:or prepared molds "i[e[\ glass _ber cutting\ waxing and cleaning\ carpen! try#[ Group 5 consisted of supervisory personnel "fore! men#[ Information was collected about lamination techniques\ the types and amounts of resin used and products[ Resins were categorized according to the company that manufactured them "indicated by numbers 0 through 6# and also according to whether they were
An investigation of factors contributing to styrene and styrene!6\7!oxide exposures
standard "STD# resins\ low!styrene!emission "LSE# resins\ or sprayable "SPR# resins[ A total of twelve types of resins from seven di}erent manufacturers were de_ned by this scheme[ Sample collection Two personal breathing!zone measurements of approximately 3 h each were collected for each worker over one full work shift "approximately 7 h# using di}usive samplers "2M!2499 Organic Vapor Monitor^ 2M\ St[ Paul\ MN#[ Measurements were collected so that they would be representative of the groups and of the exposure situations involved "i[e[\ with due regard for randomness and independence across indi! viduals and operating conditions and over time#[ Within 13 h of collection\ styrene and SO were eluted from the monitor with 1 ml of a mixture of 29) carbon disul_de in dichloromethane containing 9[0 mg:ml 3!tert!butylcatechol as a stabilizer[ Samples were assayed against external standards of styrene and SO "Fluka AG\ Buchs\ Switzerland# by capillary gas chromatography with ~ame!ionization detection[ The detection limits for styrene and SO\ as reported by the laboratory that prepared the analyses\ were 9[1 mg:ml and 9[0 mg:ml\ respectively[ Rates of sample uptake by the passive monitors were assumed to be 17[8 ml:min for styrene and 11[7 ml:min for SO\ based upon information from the manufacturer of the pas! sive monitors and recent laboratory experiments "Anonymous\ 0885^ Tornero!Velez et al[\ submitted#[ Thus\ the corresponding detection limits were 03 mg:m2 and 8 mg:m2\ respectively[ Exposure con! centrations were recalculated using the above men! tioned sample uptake rates[ ðNote] The original exposure concentrations were calculated based upon preliminary laboratory experiments "at the time of the exposure measurements#\ which estimated an uptake rate of 12 ml:min for both styrene and SO "unpub! lished data#[ The concentrations were recalculated with the updated rates\ which are considered to be more accurate than the original values[ This adjust! ment does not a}ect the overall results in the inves! tigationŁ[ Statistical analysis Styrene and SO concentrations were natural log! transformed for all statistical analyses[ Con! centrations that were at or below the detection limits "3) of styrene measurements and 03) of SO measurements# were assigned a value of 1:2 of the detection limit prior to log!transformation[ All statis! tical analyses employed the SAS System Software "SAS Institute\ Cary\ NC# and all di}erences were evaluated at a two!tailed signi_cance level of 9[94[ Spearman correlation coe.cients were used to investigate the strength of univariate relationships between styrene and SO concentrations in di}erent job groups[ Mixed!model multiple linear regression analyses "PROC MIXED# were used to investigate the relative
090
in~uences of _xed e}ects representing job groups\ resin types and resin consumption on styrene and:or SO concentrations\ while modeling between!worker and within!worker variability via the use of random e}ects[ Because of sparse data in many cells\ the mixed!model analyses were restricted to comparisons involving two job groups comprised of hand lami! nators "the combination of Groups 0 and 1# and non! laminators "the combination of Groups 3Ð5\ which served as the reference group# and nine resin cat! egories "Resin 3 STD was the reference resin#[ Two models were employed\ one for styrene exposure ðModel "0#Ł and one for SO exposure ðModel "1#Ł[ Model "0# for styrene exposure is given by the following expression] Yijkl ln "Xijkl# mY¦ai¦bj¦gijk¦xRikl¦eijkl\ "0# for i 0\ 1\ [ [ [ \ G groups\ j 0\ 1\ [ [ [ \ T resin types\ k 0\ 1\ [ [ [ \ Kij workers in group i using resin j\ and l 0\1 measurements:worker\ where Xijkl the styrene concentration of the l!th measurement from the k!th worker exposed to the j!th resin type in the i!th job group\ Yijkl the natural log!transformed value of Xijkl\ mY the overall mean\ ai the _xed e}ect for the i!th job group\ bj the _xed e}ect for the j!th resin type "the ref! erence resin was Resin 3 STD#\ gijk the random e}ect for the k!th worker in the i!th group using j!th resin type\ x the _xed e}ect for the amount of resin con! sumed "for workers who did not use resins\ Rikl 9#\ Rikl the amount of the resin consumed during the l!th measurement for the k!th worker in the i!th job group\ eijkl the random e}ect for the l!th measurement on the k!th worker exposed to the j!th resin type in the i!th job group[ Model "1# for SO exposure is given by the following expression] Yijkl ln "Xijkl# mY¦ai¦bj¦gijk¦dSijkl¦eijkl\ "1# where all the terms are the same as for Model "0# above\ except that Xijkl the SO concentration for the l!th measure! ment on the k!th worker exposed to the j!th resin type in the i!th job group\ d the _xed e}ect for the log!transformed styrene concentration\ Sijkl the value of the log!transformed styrene con! centration for the l!th measurement on the k!th worker exposed to the j!th resin type in the i!th job group[ Since SO is an oxidation product of styrene "i[e[\ for! mation of SO depends on the availability of styrene#
091
L[ A[ Nylander!French et al[
and not vice versa\ the term dSijkl in Model "1# does not appear in Model "0#[ Preliminary analyses showed that the _xed e}ect {resin consumption| ðthe term xRikl in Model "0#Ł was predictive for styrene\ but not for SO\ concentrations^ therefore\ this e}ect was not con! sidered in the _nal model for SO exposure ðModel "1#Ł[ The underlying distribution of concentrations for either styrene or SO was assumed to be log!normal[ It was further assumed that gijk and eijkl are mutually independent and normally distributed with means of 1 zero and respective variances sB1 and sW \ representing the between! and within!worker variance com! ponents\ respectively ði[e[\ gijk½N"9\sB1 # and eijkl 1 ½N"9\sW #Ł[ For Model "0#\ it follows that E"Yijkl# mY¦ai¦bj¦xRikl and\ for Model "1#\ E"Yijkl# mY¦ai¦bj¦dSijkl[ For both models\ Var"Yijkl# 1 sY1 sB1 ¦sW \ Cov"Yijk0\Yijk1# sB1 and Corr"Yijk0\ 1 1 Yijk1# sB:"sB¦s1W# RESULTS
Summary statistics Table 0 provides summary statistics for each of the six job groups\ including the number of workers\ the mean exposure concentrations for styrene and SO and the mean resin consumption[ For all 126 workers com! bined\ the mean 7 h time!weighted average "TWA7h# concentrations of styrene and SO were 011 mg:m2 "range 2[1 to 597 mg:m2^ median 092 mg:m2# and 072 mg:m2 "range 9 to 772 mg:m2^ median 009 mg:m2#\ respectively[ Exposure to styrene varied considerably among the various job groups with a roughly 3!fold range between Group 0 "hand lamination*large objects\ mean TWA7h 045 mg:m2# and Group 5 "fore! men\ mean TWA7h 31[5 mg:m2#[ Exposures to SO also covered about a 4!fold range between Group 2 "mean TWA7h 271 mg:m2# and Group 5 "mean TWA7h 60[1 mg:m2#[ In Table 1\ exposures to styrene and SO and resin consumption are broken down by the type of product manufactured[ Styrene exposures did not vary remarkably among the various product categories[ The highest TWA7h styrene concentrations were mea! sured in boat manufacturing facilities "mean 021 mg:m2^ range 2[1 to 596 mg:m2# and the lowest in factories where sheets\ elements and car parts were made "mean 89[2 mg:m2^ range 6[9 to 595 mg:m2#[ SO
concentrations varied somewhat more for the di}erent products\ being about twice as large in boat manu! facturing "mean 114 mg:m2^ range 9 to 772 mg:m2# com! pared to the manufacturing of containers and tubes "75[9 mg:m2^ range 9 to 522 mg:m2#[ Spearman correlation coe.cients for individual worker|s TWA7h for styrene and SO are given in Table 2 for the six job groups[ When all workers were con! sidered together\ a positive correlation between styr! ene and SO concentrations was observed "rs 9[419\ P³9[9990#[ However\ this relatively strong cor! relation primarily re~ects the large correlation coe.cients in the two largest groups\ consisting of hand laminators "Group 0] N 097\ rs 9[455\ P³9[9990^ Group 1] N 38\ rs 9[558\ P³9[9990# and\ to a lesser extent\ assembly and mold preparation "Group 4] N 17\ rs 9[225\ P 9[970#[ Groups 2\ 3 and 5 displayed more modest levels of correlation "Pr9[277 in all cases#[ Mixed!effects models Mixed!e}ects regression models were applied to investigate the association between the log!trans! formed concentrations of styrene ðModel "0#Ł and SO ðModel "1#Ł and the _xed e}ects representing job group\ resin type\ as well as level of resin consumption for Model "0# and styrene concentration for Model "1#[ As noted above\ these analyses employed pooled job groups ðlamination "Groups 0 and 1# vs[ non! lamination "Groups 3Ð5#Ł to minimize the number of cells with sparse data[ Group 2 was omitted from the analysis based on the small number of workers "00# and resin types used "SPR#\ as well as the distinctly di}erent work tasks and methods which were employed by the other groups considered in the analyses[ For Model "0#\ job group\ resin type and resin con! sumption signi_cantly a}ected styrene concentration "the P!values for the global tests of _xed e}ects for job group\ resin type and resin consumption were all ³9[908#[ As shown in Table 3\ laminators had signi_cantly greater exposure than non!laminators "reference group# and Resins 1 STD\ 2 STD\ 4 LSE and 5 STD all resulted in signi_cantly lower exposures than those of Resin 3 STD "reference resin#[ Regarding Model "1#\ resin type and styrene con! centration signi_cantly a}ected SO concentration "the
Table 0[ Summary statistics for exposures and resin consumption "mean values and standard deviations# for six job groups Group 0[ Hand lamination\ large objects 1[ Hand lamination\ small objects 2[ Spraying and gel coating 3[ Automated lamination 4[ Assembly and mold preparation 5[ Foremen All workers
No[ of workers
Styrene exposure "mg:m2# 2S[D[
SO exposure "mg:m2# 2S[D[
Resin consumption "kg# 2S[D[
097 38 00 12 17 07 126
0452094 0492014 029242[6 44[7221[2 38[6239[2 31[5281[4 0112096
0882078 1022080 2712165 84[2253[1 0392028 60[1266[6 0722073
55249 20220 74239 482043 0[021[1 9[220[0 35255
An investigation of factors contributing to styrene and styrene!6\7!oxide exposures
092
Table 1[ Summary statistics for exposures and resin consumption "mean values and standard deviations# for the various products manufactured Products manufactured Boats Small and form parts Sheets\ elements and car parts Containers and tubes
No[ of workers
Styrene exposure "mg:m2# 2S[D[
SO exposure "mg:m2# 2S[D[
Resin consumption "kg# 2S[D[
026 23 28 17
021285[3 023287[6 89[22016 0952016
1142088 0592058 0152029 75[92015
38230 22221 282019 48274
Table 2[ Spearman correlation coe.cients between log!transformed TWA7h con! centrations for styrene and SO in six groups Group 0[ Hand lamination\ large objects 1[ Hand lamination\ small objects 2[ Spraying and gel coating 3[ Automated lamination 4[ Assembly and mold preparation 5[ Foremen All workers
Table 3[ Point estimates and standard errors for _xed e}ects related to styrene exposure ðModel "0#Ł Group:resin type Intercept "Non!laminators\ Resin 3STD# Laminators Resin 0 STD Resin 1 STD Resin 1 LSE Resin 2 STD Resin 3 LSE Resin 4 STD Resin 4 LSE Resin 5 STD Resin consumption
Standard error
P!value
3[935
9[182
9[9990
0[398 −9[483 −9[752 −9[326 −0[129 −9[346 −9[429 −9[895 −0[072 9[993
9[056 9[279 9[222 9[236 9[296 9[374 9[200 9[202 9[332 9[991
9[9990 9[019 9[909 9[198 9[9990 9[236 9[980 9[993 9[997 9[908
Estimate
P!values for the global tests of _xed e}ects for resin type and styrene concentration were both ³9[9990#\ while the e}ect of job group was not signi_cant "the P!value for the global test of job group was 9[257#[ Focusing upon the individual resins\ as shown in Table 4\ all resins except 3 LSE resulted in exposures that were signi_cantly lower than Resin 3 STD "ref! erence resin#[ Since {resin type| was observed to be a signi_cant predictor of both styrene and SO exposure "the P! value for the global tests of _xed e}ects were both ³9[993#\ this factor was studied in more detail[ Regrouping of the resins into STD and LSE resin types did not reveal any signi_cant di}erence in their e}ects on styrene concentrations[ However\ the use of LSE resins resulted in signi_cantly greater exposures to SO "P³9[9990# than the use of STD resins[
No[ of workers
rs
P!value
097 38 00 12 17 07 126
9[455 9[558 9[153 9[976 9[225 9[106 9[419
³9[9990 ³9[9990 9[322 9[582 9[970 9[277 ³9[9990
Table 4[ Point estimates and standard errors for _xed e}ects related to SO exposure ðModel "1#Ł Group:resin type
Estimate
Standard error
P!value
Intercept "Non!laminators\ Resin 3 STD# Laminators Resin 0 STD Resin 1 STD Resin 1 LSE Resin 2 STD Resin 3 LSE Resin 4 STD Resin 4 LSE Resin 5 STD ln "styrene concentration#
−1[320
9[267
9[9990
9[065 −2[025 −0[705 −9[647 −1[014 −9[165 −0[360 −9[588 −0[385 9[179
9[084 9[307 9[256 9[279 9[230 9[420 9[231 9[235 9[377 9[938
9[257 9[9990 9[9990 9[937 9[9990 9[593 9[9990 9[934 9[992 9[9990
DISCUSSION
During the manufacture of reinforced plastics\ styr! ene and other airborne contaminants\ including SO\ are released[ Although extensive e}ort has been made to study the emissions of styrene during the lamination processes\ very little is known about the formation of SO[ Unsaturated polyesters are dissolved in styrene "39Ð59)# to form a cross!linked structure with the help of a catalyst and an accelerator\ both of which can vary with the type of resin used[ During the cross! linking of polyester polymers\ approximately 09Ð04) of styrene is released to air "Saamanen et al[\ 0880# and\ based upon limited data\ levels of SO range from about 0:499Ð0:0999 of those of styrene "Pfa/i and Saamanen\ 0882^ Rappaport et al[\ 0885#[ The for! mation of SO during this process has been explained in
093
L[ A[ Nylander!French et al[
terms of fragmentation of polymeric styrene peroxide radicals or epoxidation of the styrene monomer\ or both "Brighton et al[\ 0868#[ Rappaport et al[ "0885# recently reported that SO exposure was more highly correlated than styrene exposure with various SO biomarkers in workers| blood\ including SOÐalbumin and SOÐDNA adducts and sister chromatid exchanges[ This observation pro! mpted us to further investigate the possible factors that may a}ect worker exposure to airborne styrene and SO in the reinforced!plastics industry[ Exposure to SO was positively correlated with exposure to styr! ene in selected job groups\ being strongest for hand lamination followed by assembly and mold prep! aration "Table 2#[ However\ the styreneÐSO cor! relations were weak for the other three job groups "Table 2#\ suggesting that factors other than coex! posure to styrene contributed to SO levels[ In mixed model analyses\ an important factor for predicting both styrene and SO concentrations was the type of resin used in the lamination process[ On the other hand\ job group was a strong predictor for styrene exposure level but a nonsigni_cant predictor for SO exposure level[ When the analysis was per! formed after regrouping the resins into STD or LSE resins\ no signi_cant di}erences in their e}ects on styrene concentrations were observed[ This _nding is rather surprising because LSE resins are\ in fact\ thought to reduce styrene exposure "Andersson et al[\ 0883^ Kalliokoski and Jantunen\ 0870^ Saamanen et al[\ 0880^ Skrifvars et al[\ 0881#[ More work is needed to con_rm this _nding[ On the other hand\ LSE resins were observed to signi_cantly increase SO con! centrations[ Since LSE resins were presumably designed to reduce exposure to styrene without regard for coexposure to SO\ this _nding is perhaps not sur! prising[ Possible explanations for the increase in SO exposure associated with LSE resins relate to the com! position of the resin\ the reduction of evaporation of styrene and:or the amount of catalyst used[ Since the measurements of styrene and SO exposure analyzed in this study were gathered for other pur! poses about 09 years ago it is likely that operating conditions have changed somewhat and that current exposures are probably lower today ðe[g[\ "Symanski et al[\ 0887#Ł[ However\ because the basic types of work "represented by the job groups# and the types of resins are quite similar to those currently observed in the reinforced!plastics industry\ the major _ndings concerning the factors a}ecting exposures to styrene and SO should still be relevant[ We are currently conducting a detailed inves! tigation to determine whether other observable fea! tures of the various production processes contribute to SO exposure and to an increase in SO biomarkers in the blood of reinforced plastics workers[ If SO is con_rmed to be the principal causative agent for the observed biomarkers in these workers\ this type of mixed!models multiple linear regression analysis will assist in clarifying and understanding general patterns
of exposure to styrene and SO and in developing and testing measures to prevent occupational exposures to these compounds[ Acknowled`ements*We thank Pirkko Pfa/i and the Finnish Institute of Occupational Health\ Helsinki\ Finland for pro! viding the data for the analysis[ We are grateful to the com! panies and workers of the Finnish reinforced plastics indus! try for participating in the study[ The collection and analysis of the exposure data was _nanced by the grant from the Finnish Work Environment Fund "TSR 77960#[ This work was supported by grants from National Institute of Environ! mental Health Sciences "T 21 ES96907# and National Cancer Institute "RO0CA58352#[ REFERENCES Andersson\ I[!M[ and Rosen\ G[ "0883# Controlling exposure to styrene during lamination of large moulds[ Evaluation of the IVF booth[ Arbete och Halsa 07\ 178Ð183[ Andersson\ I[!M[\ Rosen\ G[\ Sturm\ M[ and Wielaard\ P[ "0883# Styrene emission and occupational exposure in hand lay!up moulding of glass _bre using two types of polyester[ Occup[ Hy`[ 0\ 196Ð106[ Anonymous "0885# 2M Organic Vapor Monitors "Styrene# 2499:2409:2419:2429[ Technical Data Bulletin (015\ 2M Occupational Health and Environmental Safety\ St[ Paul\ MN[ Bond\ J[ A[ "0878# Review of the toxicology of styrene[ Crit[ Rev[ Toxicol[ 08\ 116Ð138[ Brighton\ C[ A[\ Pritchard\ G[ and Skinner\ G[ A[ "0868# Technology and environmental aspects[ In] Styrene Poly! mers\ pp[ 26Ð39[ Applied Science Publisher\ Ltd[\ London[ Csanady\ G[ A[\ Mendrala\ A[ L[\ Nolan\ R[ J[ and Filser\ J[ G[ "0883# A physiologic pharmacokinetic model for styr! ene and styrene!6\7!oxide in mouse\ rat and man[ Archs[ Toxicol[ 57\ 032Ð046[ Filser\ J[ G[\ Schwegler\ U[\ Csanady\ G[ A[\ Greim\ H[ and Kessler\ W[ "0882# Species!speci_c pharmacokinetics of styrene in rat and mouse[ Archs[ Toxicol[ 56\ 406Ð429[ Fjelstad\ P[ E[\ Thurud\ S[ and Wannag\ A[ "0868# Styrene oxide in the manufacture of reinforced polyester plastics[ Scand[ J[ Work Environ[ Health 4\ 051Ð052[ Fleig\ I[ and Thiess\ A[ M[ "0867# Mutagenicity study of workers employed in the styrene and polystyrene pro! cessing and manufacturing industry[ Scand[ J[ Work Environ[ Health 3\ 143Ð147[ Hopkins\ B[ L[\ Conard\ R[ J[\ Dangel\ R[ F[\ Fitch\ H[ G[\ Smith\ M[ J[ and Anger\ W[ J[ "0875# Behavioral tech! nology for reducing occupational exposures to styrene[ Appl[ Beh[ Anal[ 08\ 2Ð00[ Hopkins\ B[ L[\ Conard\ R[ J[ and Smith\ M[ J[ "0875# E}ec! tive and reliable behavioral control technology[ Am[ Ind[ Hy`[ Assoc[ J[ 36\ 674Ð680[ IARC "0883# Some industrial chemicals[ In] IARC Mono! `raphs on the Evaluation of Carcino`enic Risk to Humans\ Vol[ 59\ pp[ 122Ð235[ International Agency for Research on Cancer\ Lyon[ Kalliokoski\ P[ and Jantunen\ M[ "0870# Control of airborne styrene in reinforced plastics industry[ In] 1nd World Con! `ress of Chemical En`ineerin`\ Montreal\ October 3Ð8[ Kalliokoski\ P[\ Koistinen\ T[ and Jaaskelainen\ M[ "0873# Prevention of styrene hazards*hygienic approaches[ In] Industrial Hazards of Plastics and Synthetic Elastomers\ pp[ 068Ð175[ Alan R[ Liss\ Inc[\ New York[ Kalliokoski\ P[ J[\ Saamanen\ A[ J[\ Ivalo\ L[ M[ and Kokotti\ H[ M[ "0877# Exposure to styrene can be controlled[ Am[ Ind[ Hy`[ Assoc[ J[ 38\ 5Ð8[ Korn\ M[\ Gfrorer\ W[\ Filser\ J[ G[ and Kessler\ W[ "0883# Styrene!6\7!oxide in blood of workers exposed to styrene[ Archs[ Toxicol[ 57\ 413Ð416[ Lof\ A[\ Brohede\ C[ and Gullstrand\ E[ "0882# The e}ec!
An investigation of factors contributing to styrene and styrene!6\7!oxide exposures tiveness of respirators measured during styrene exposure[ Int[ Archs[ Occup[ Environ[ Health 54\ 18Ð23[ Mark\ H[ F[\ Bikales\ N[ M[\ Overberger\ C[ G[\ Menges\ G[ and Kroschwitz\ Z[ J[ "0877# Encyclopedia of Polymer Science and En`ineerin`\ John Wiley and Sons\ New York[ Mayo\ F[ "0847# The oxidation of unsaturated compounds[ V[ The e}ect of oxygen pressure on the oxidation of styr! ene[ J[ Am[ Chem[ Soc[ 79\ 1354Ð1379[ Mendrala\ A[ L[\ Langvardt\ P[ W[\ Nitschke\ K[ D[\ Quast\ J[ F[ and Nolan\ R[ J[ "0882# In vitro kinetics of styrene and styrene oxide metabolism in rat\ mouse and human[ Archs[ Toxicol[ 56\ 07Ð16[ Nakajima\ T[\ Elovaara\ E[\ Gonzales\ F[ J[\ Gelboin\ F[ J[\ Vainio\ H[ and Aoyama\ T[ "0882# Characterization of the human cytochrome P349 isozymes responsible for styrene metabolism[ In Butadiene and Styrene] Assessment of Health Hazards\ IARC Scienti_c Publication No[ 016 "Edited by Sorsa\ M[ et al[#\ pp[ 090Ð097[ International Agency for Research on Cancer\ Lyon[ Parkyn\ B[\ Lamb\ F[ and Clifton\ B[ V[ "0856# In] Unsatu! rated Polyesters and Polyester Plasticisers\ Vol[ 1\ pp[ 46Ð 56[ American Elsevier Publishing Company\ Inc[\ New York[ Pekari\ K[\ Nylander!French\ L[ A[\ Pfa/i\ P[\ Sorsa\ M[ and Aitio\ A[ "0882# Biological monitoring of exposure to styrene*Assessment of di}erent approaches[ J[ Occup[ Med[ Toxicol[ 1\ 004Ð015[ Petreas\ M[ X[\ Woodlee\ J[\ Becker\ C[ E[ and Rappaport\ S[ M[ "0884# Retention of styrene following controlled exposure to constant and ~uctuating air concentrations[ Int[ Archs[ Occup[ Environ[ Health 56\ 16Ð23[ Pfa/i\ P[ and Saamanen\ A[ "0882# The occupational scene of styrene[ In] Butadiene and Styrene] Assessment of Health Hazards\ IARC Scienti_c Publication No[ 016 "Edited by Sorsa\ M[ et al[#\ pp[ 04Ð15[ International Agency for Research on Cancer\ Lyon[
094
Pfa/i\ P[\ Vainio\ H[ and Hesso\ A[ "0868# Styrene and styrene oxide concentrations in the air during the lami! nation process in the reinforced plastics industry[ Scand[ J[ Work Environ[ Health 4\ 047Ð050[ Rappaport\ S[ M[\ Yeowell!O|Connell\ K[\ Bodell\ W[\ Yager\ J[ W[ and Symanski\ E[ "0885# An investigation of multiple biomarkers among workers exposed to styrene and styrene!6\7!oxide[ Cancer Res[ 45\ 4309Ð4305[ Saamanen\ A[ and Enbom\ S[ "0883# A mobile local ven! tilation system equipped with an automatic position control[ Arbete och Halsa 07\ 296Ð201[ Saamanen\ A[ J[\ Niemela\ R[ I[\ Blomqvist\ T[ K[ and Nikan! der\ E[ M[ "0880# Emission of styrene during the hand lay! up molding of reinforced polyester[ Appl[ Occup[ Environ[ Hy`[ 5\ 689Ð682[ Skrifvars\ M[\ Johansson\ M[\ Blomqvist\ T[ and Keinanen\ K[ "0881# Low styrene emission unsaturated polyester resins[ Composites 2\ 31Ð34[ Sorsa\ M[\ Anttila\ A[\ Jarventaus\ H[\ Kubiak\ R[\ Norppa\ H[\ Nylander\ L[\ Pekari\ K[\ Pfa/i\ P[ and Vainio\ H[ "0880# Styrene revisited] exposure assessment and risk esti! mation in reinforced plastics industry[ Pro`[ Clin[ Biol[ Res[ 261\ 076Ð084[ Sumner\ S[ J[ and Fennell\ T[ R[ "0883# Review of the meta! bolic fate of styrene[ Crit[ Rev[ Toxicol[ 13\ S00Ð22[ Symanski\ E[\ Kupper\ L[ L[ and Rappaport\ S[ M[ "0887# Comprehensive evaluation of long term trends in occu! pational exposure] part 0\ Description of the database[ Occup[ Environ[ Med[ 44\ 299Ð298[ Todd\ W[ F[ "0874# Styrene vapor control systems in FRP yacht plants[ Am[ J[ Ind[ Med[ 7\ 108Ð121[ Tornero!Velez\ R[\ Denovan\ L[\ Waidyanatha\ S[ and Rap! paport\ S[ M[ "submitted# Measurement of styrene!6\7! oxide and other oxidation products of styrene in air[ Walling\ C[ and Heating\ L[ "0854# Radical formation in the reactions of t!butyl hydroperoxide with styrene[ J[ Am[ Chem[ Soc[ 76\ 27Ð36[
Pergamon PII] S9992Ð3767"87#99979Ð4
Ann[ occup[ Hyg[\ Vol[ 32\ No[ 1\ pp[ 88Ð098\ 0888 Þ 0888 British Occupational Hygiene Society Published by Elsevier Science Ltd[ All rights reserved Printed in Great Britain[ 9992Ð3767:88 ,08[99¦9[99
An Investigation of Factors Contributing to Styrene and Styrene!6\7!oxide Exposures in the Reinforced!Plastics Industry LEENA A[ NYLANDER!FRENCH\ % LAWRENCE L[ KUPPER$ and STEPHEN M[ RAPPAPORT Department of Environmental Sciences and En`ineerin`\ CB (6399 Rosenau Hall\ The University of North Carolina at Chapel Hill\ Chapel Hill\ NC 16488!6399\ USA^ $Department of Biostatistics\ School of Public Health\ The University of North Carolina at Chapel Hill\ Chapel Hill\ NC 16488! 6399\ USA During the manufacturing of reinforced plastics\ large amounts of styrene and trace quantities of styrene!6\7!oxide "SO# are released[ Since previous work suggests that inhalation of even small amounts of SO might be an important health risk\ we investigated several possible factors contributing to styrene and SO exposure during the manufacture of reinforced plastics[ Factors related to job type\ worker and the type and quantity of styrene!containing resins were investigated using mixed!effects multiple linear regression models[ Overall\ SO exposure levels were positively correlated with styrene exposure levels[ However\ this correlation was statistically signi_cant only among hand laminators who had the highest exposures to both styrene and SO[ An important factor for predicting both styrene and SO concentrations was the type of resin used\ while the quantity of resin consumed was predictive of styrene but not of SO exposure[ Since SO exposure appears to be associated with factors other than coexposure to styrene\ more effort should be placed on investigating emissions of SO per se[ The type of mixed!models regression analysis employed in this study can be used for clarifying the underlying patterns for exposures to styrene and SO as well as for evaluating preventive measures[ Þ 0888 British Occupational Hygiene Society[ Published by Elsevier Science Ltd[ All rights reserved[ Keywords] styrene^ styrene!6\7!oxide^ exposure factors^ reinforced plastics
observed among reinforced!plastics workers\ Fleig and Thiess "0867# speculated about 19 years ago that coexposure to airborne styrene!6\7!oxide "SO# might also be important[ SO is a contaminant that is formed at low levels in air "³0 mg:m2# from oxidation of styrene during the manufacture of reinforced plastics "Brighton et al[\ 0868^ Fjelstad et al[\ 0868^ Pfa/i and Saamanen\ 0882^ Pfa/i et al[\ 0868#[ Despite this speculation of Fleig and Thiess "0867# and more recently by Pfa/i and Saamanen "0882# that SO exposure per se may be important\ the potential role of SO exposure in the etiology of adverse health e}ects in the reinforced!plastics industry has been lar! gely ignored[ This may be due to the fact that styrene is present at high concentrations in this industry and is metabolized almost exclusively through SO in humans[ ðAbout 83) of inhaled styrene is absorbed in humans "Petreas et al[\ 0884# and is metabolized to SO via hepatic cytochrome P!349 isozymes "Nakajima et al[\ 0882#Ł[ SO is subsequently metabolized in the
INTRODUCTION
Styrene is one of the most important monomers pro! duced globally for the manufacture of plastics and resins[ In 0881\ the production of styrene was over four million metric tons in the US^ one half to two thirds of the styrene was used to produce polystyrene and the remainder to produce a variety of copolymeric products and resins "IARC\ 0883#[ Although occu! pational exposure to airborne styrene is modest "³09 mg:m2# in most parts of the industry\ it tends to be relatively high "39Ð399 mg:m2# in the reinforced! plastics industry\ especially during the production of large objects such as boats "IARC\ 0883#[ Although exposure to styrene has generally been thought to be responsible for genotoxic e}ects Received 08 June 0887^ in _nal form 07 November 0887[ %Author to whom correspondence should be addressed[ Tel[] ¦0!808!855!2715^ Fax] ¦0!808!855!3600^ E!mail] leena*frenchÝunc[edu 88
099
L[ A[ Nylander!French et al[
liver by epoxide hydrolase to form styrene glycol and its oxidation products "mandelic acid and phe! nylglyoxylic acid#\ which are excreted in the urine ðreviewed by Bond "0878#\ IARC "0883# and Sumner and Fennell "0883#Ł[ Since styrene exposure tends to be 499Ð0999 times greater than SO exposure "Pfa/i and Saamanen\ 0882# and\ since virtually all styrene is metabolized to SO in humans\ we suspect that most investigators tended to discount the possible import! ance of SO exposure per se[ However\ recent evidence of e.cient detoxi_cation of SO in the human liver following metabolism of styrene "Csanady et al[\ 0883^ Filser et al[\ 0882^ Korn et al[\ 0883^ Mendrala et al[\ 0882# suggests that the systemic dose of SO derived from the metabolism of styrene might be small[ This was con_rmed by Rappaport et al[ "0885# who predicted\ from the measurements of SOÐalbumin adducts in the blood of 37 reinforced!plastics workers\ that only about 9[94) of the styrene dose was released to the blood as SO[ Indeed\ Rappaport et al[ "0885# estimated that the typical worker in their study received 60) of the systemic dose of SO from direct inhalation[ The objective of any occupational and epi! demiological investigation is to identify factors that contribute to exposure and the subsequent risks of disease so that interventions can be made to reduce exposures[ In the reinforced!plastics industry\ the con! trol of styrene exposure has been the major focus[ In fact\ special low!styrene!emission "LSE# resins have been developed for this industry[ Studies describing new methods that can be used to reduce the styrene exposures in the reinforced plastics industry have been published "Andersson and Rosen\ 0883^ Hopkins et al[\ 0875a^ Hopkins et al[\ 0875b^ Kalliokoski et al[\ 0873^ Kalliokoski et al[\ 0877^ Lof et al[\ 0882^ Saa! manen and Enbom\ 0883^ Todd\ 0874# and there is some indication of a decreasing trend in occupational exposure to styrene over the past three decades "Sym! anski et al[\ 0887#[ In general\ control measures are problematic in the reinforced!plastics industry where products are often large and di.cult to enclose or to ventilate e}ectively[ Since coexposure to SO may\ in fact\ be a more important risk factor for adverse health e}ects than is exposure to styrene\ strategies aimed at reducing SO emissions could be important[ Unfortunately\ very little information is currently available concerning the generation of SO during the production of reinforced!plastics[ It has been postu! lated\ based upon laboratory experiments\ that SO is formed via] 0[ fragmentation of polymeric styrene peroxide rad! icals resulting from the copolymerization of styrene and oxygen "Brighton et al[\ 0868^ Mark et al[\ 0877#^ 1[ epoxidation of the styrene monomer "Brighton et al[\ 0868# and:or 2[ reaction of styrene with volatile organic peroxides used to initiate the curing of reinforced plastics
"Brighton et al[\ 0868^ Fjelstad et al[\ 0868^ Mayo\ 0847^ Parkyn et al[\ 0856^ Pfa/i and Saamanen\ 0882^ Pfa/i et al[\ 0868^ Walling and Heating\ 0854#[ However\ no _eld investigations have thus far been reported to evaluate these speculations in the reinforced!plastics industry\ or to investigate possible job!related sources of exposure to SO[ In order to investigate the possible factors con! tributing to exposure to both styrene and SO in the reinforced plastics industry\ we used mixed!e}ects multiple linear regression models to analyze exposure data that had been previously collected in Finland ðpublished in part\ "Pekari et al[\ 0882^ Sorsa et al[\ 0880#Ł[ Our goal was to determine how factors related to the job\ worker\ resin formulation and use of styr! ene!containing resins a}ected both styrene and SO exposures in the reinforced!plastics industry[
MATERIALS AND METHODS
Study population Workers were recruited from 29 di}erent reinforced!plastics facilities in Finland during November 0877 and February 0889[ The plants ranged from small workshops with a few employees to fairly large factories[ Items manufactured included boats\ containers\ pipes and tubes\ small parts\ sheets and vehicle components[ Most of the sites involved boat manufacturing "69)#[ The most common lami! nation process involved hand lamination of open molds[ Other work tasks included spray lamination\ _lament winding\ low!pressure!sack lamination\ cen! trifugal lamination\ robot controlling\ assembly\ glass!_ber cutting\ waxing and cleaning\ carpentry and work supervision[ During data collection\ distinct occupational groups were identi_ed and workers were assigned to di}erent job groups according to the work tasks they performed[ Initially\ six di}erent job groups were established[ Group 0 consisted of laminators who hand laminated objects larger than 0 m1 in surface area[ Group 1 consisted of laminators who hand lami! nated objects smaller than 0 m1 in surface area[ Group 2 consisted of workers who used chopper guns and:or gel coat spraying during lamination[ Group 3 con! sisted of workers who used automated techniques "i[e[\ _lament winding\ low!pressure!sack lamination\ cen! trifugal lamination\ robot controlling#[ Group 4 con! sisted of workers who assembled laminated objects "both large and small in size# and:or prepared molds "i[e[\ glass _ber cutting\ waxing and cleaning\ carpen! try#[ Group 5 consisted of supervisory personnel "fore! men#[ Information was collected about lamination techniques\ the types and amounts of resin used and products[ Resins were categorized according to the company that manufactured them "indicated by numbers 0 through 6# and also according to whether they were
An investigation of factors contributing to styrene and styrene!6\7!oxide exposures
standard "STD# resins\ low!styrene!emission "LSE# resins\ or sprayable "SPR# resins[ A total of twelve types of resins from seven di}erent manufacturers were de_ned by this scheme[ Sample collection Two personal breathing!zone measurements of approximately 3 h each were collected for each worker over one full work shift "approximately 7 h# using di}usive samplers "2M!2499 Organic Vapor Monitor^ 2M\ St[ Paul\ MN#[ Measurements were collected so that they would be representative of the groups and of the exposure situations involved "i[e[\ with due regard for randomness and independence across indi! viduals and operating conditions and over time#[ Within 13 h of collection\ styrene and SO were eluted from the monitor with 1 ml of a mixture of 29) carbon disul_de in dichloromethane containing 9[0 mg:ml 3!tert!butylcatechol as a stabilizer[ Samples were assayed against external standards of styrene and SO "Fluka AG\ Buchs\ Switzerland# by capillary gas chromatography with ~ame!ionization detection[ The detection limits for styrene and SO\ as reported by the laboratory that prepared the analyses\ were 9[1 mg:ml and 9[0 mg:ml\ respectively[ Rates of sample uptake by the passive monitors were assumed to be 17[8 ml:min for styrene and 11[7 ml:min for SO\ based upon information from the manufacturer of the pas! sive monitors and recent laboratory experiments "Anonymous\ 0885^ Tornero!Velez et al[\ submitted#[ Thus\ the corresponding detection limits were 03 mg:m2 and 8 mg:m2\ respectively[ Exposure con! centrations were recalculated using the above men! tioned sample uptake rates[ ðNote] The original exposure concentrations were calculated based upon preliminary laboratory experiments "at the time of the exposure measurements#\ which estimated an uptake rate of 12 ml:min for both styrene and SO "unpub! lished data#[ The concentrations were recalculated with the updated rates\ which are considered to be more accurate than the original values[ This adjust! ment does not a}ect the overall results in the inves! tigationŁ[ Statistical analysis Styrene and SO concentrations were natural log! transformed for all statistical analyses[ Con! centrations that were at or below the detection limits "3) of styrene measurements and 03) of SO measurements# were assigned a value of 1:2 of the detection limit prior to log!transformation[ All statis! tical analyses employed the SAS System Software "SAS Institute\ Cary\ NC# and all di}erences were evaluated at a two!tailed signi_cance level of 9[94[ Spearman correlation coe.cients were used to investigate the strength of univariate relationships between styrene and SO concentrations in di}erent job groups[ Mixed!model multiple linear regression analyses "PROC MIXED# were used to investigate the relative
090
in~uences of _xed e}ects representing job groups\ resin types and resin consumption on styrene and:or SO concentrations\ while modeling between!worker and within!worker variability via the use of random e}ects[ Because of sparse data in many cells\ the mixed!model analyses were restricted to comparisons involving two job groups comprised of hand lami! nators "the combination of Groups 0 and 1# and non! laminators "the combination of Groups 3Ð5\ which served as the reference group# and nine resin cat! egories "Resin 3 STD was the reference resin#[ Two models were employed\ one for styrene exposure ðModel "0#Ł and one for SO exposure ðModel "1#Ł[ Model "0# for styrene exposure is given by the following expression] Yijkl ln "Xijkl# mY¦ai¦bj¦gijk¦xRikl¦eijkl\ "0# for i 0\ 1\ [ [ [ \ G groups\ j 0\ 1\ [ [ [ \ T resin types\ k 0\ 1\ [ [ [ \ Kij workers in group i using resin j\ and l 0\1 measurements:worker\ where Xijkl the styrene concentration of the l!th measurement from the k!th worker exposed to the j!th resin type in the i!th job group\ Yijkl the natural log!transformed value of Xijkl\ mY the overall mean\ ai the _xed e}ect for the i!th job group\ bj the _xed e}ect for the j!th resin type "the ref! erence resin was Resin 3 STD#\ gijk the random e}ect for the k!th worker in the i!th group using j!th resin type\ x the _xed e}ect for the amount of resin con! sumed "for workers who did not use resins\ Rikl 9#\ Rikl the amount of the resin consumed during the l!th measurement for the k!th worker in the i!th job group\ eijkl the random e}ect for the l!th measurement on the k!th worker exposed to the j!th resin type in the i!th job group[ Model "1# for SO exposure is given by the following expression] Yijkl ln "Xijkl# mY¦ai¦bj¦gijk¦dSijkl¦eijkl\ "1# where all the terms are the same as for Model "0# above\ except that Xijkl the SO concentration for the l!th measure! ment on the k!th worker exposed to the j!th resin type in the i!th job group\ d the _xed e}ect for the log!transformed styrene concentration\ Sijkl the value of the log!transformed styrene con! centration for the l!th measurement on the k!th worker exposed to the j!th resin type in the i!th job group[ Since SO is an oxidation product of styrene "i[e[\ for! mation of SO depends on the availability of styrene#
091
L[ A[ Nylander!French et al[
and not vice versa\ the term dSijkl in Model "1# does not appear in Model "0#[ Preliminary analyses showed that the _xed e}ect {resin consumption| ðthe term xRikl in Model "0#Ł was predictive for styrene\ but not for SO\ concentrations^ therefore\ this e}ect was not con! sidered in the _nal model for SO exposure ðModel "1#Ł[ The underlying distribution of concentrations for either styrene or SO was assumed to be log!normal[ It was further assumed that gijk and eijkl are mutually independent and normally distributed with means of 1 zero and respective variances sB1 and sW \ representing the between! and within!worker variance com! ponents\ respectively ði[e[\ gijk½N"9\sB1 # and eijkl 1 ½N"9\sW #Ł[ For Model "0#\ it follows that E"Yijkl# mY¦ai¦bj¦xRikl and\ for Model "1#\ E"Yijkl# mY¦ai¦bj¦dSijkl[ For both models\ Var"Yijkl# 1 sY1 sB1 ¦sW \ Cov"Yijk0\Yijk1# sB1 and Corr"Yijk0\ 1 1 Yijk1# sB:"sB¦s1W# RESULTS
Summary statistics Table 0 provides summary statistics for each of the six job groups\ including the number of workers\ the mean exposure concentrations for styrene and SO and the mean resin consumption[ For all 126 workers com! bined\ the mean 7 h time!weighted average "TWA7h# concentrations of styrene and SO were 011 mg:m2 "range 2[1 to 597 mg:m2^ median 092 mg:m2# and 072 mg:m2 "range 9 to 772 mg:m2^ median 009 mg:m2#\ respectively[ Exposure to styrene varied considerably among the various job groups with a roughly 3!fold range between Group 0 "hand lamination*large objects\ mean TWA7h 045 mg:m2# and Group 5 "fore! men\ mean TWA7h 31[5 mg:m2#[ Exposures to SO also covered about a 4!fold range between Group 2 "mean TWA7h 271 mg:m2# and Group 5 "mean TWA7h 60[1 mg:m2#[ In Table 1\ exposures to styrene and SO and resin consumption are broken down by the type of product manufactured[ Styrene exposures did not vary remarkably among the various product categories[ The highest TWA7h styrene concentrations were mea! sured in boat manufacturing facilities "mean 021 mg:m2^ range 2[1 to 596 mg:m2# and the lowest in factories where sheets\ elements and car parts were made "mean 89[2 mg:m2^ range 6[9 to 595 mg:m2#[ SO
concentrations varied somewhat more for the di}erent products\ being about twice as large in boat manu! facturing "mean 114 mg:m2^ range 9 to 772 mg:m2# com! pared to the manufacturing of containers and tubes "75[9 mg:m2^ range 9 to 522 mg:m2#[ Spearman correlation coe.cients for individual worker|s TWA7h for styrene and SO are given in Table 2 for the six job groups[ When all workers were con! sidered together\ a positive correlation between styr! ene and SO concentrations was observed "rs 9[419\ P³9[9990#[ However\ this relatively strong cor! relation primarily re~ects the large correlation coe.cients in the two largest groups\ consisting of hand laminators "Group 0] N 097\ rs 9[455\ P³9[9990^ Group 1] N 38\ rs 9[558\ P³9[9990# and\ to a lesser extent\ assembly and mold preparation "Group 4] N 17\ rs 9[225\ P 9[970#[ Groups 2\ 3 and 5 displayed more modest levels of correlation "Pr9[277 in all cases#[ Mixed!effects models Mixed!e}ects regression models were applied to investigate the association between the log!trans! formed concentrations of styrene ðModel "0#Ł and SO ðModel "1#Ł and the _xed e}ects representing job group\ resin type\ as well as level of resin consumption for Model "0# and styrene concentration for Model "1#[ As noted above\ these analyses employed pooled job groups ðlamination "Groups 0 and 1# vs[ non! lamination "Groups 3Ð5#Ł to minimize the number of cells with sparse data[ Group 2 was omitted from the analysis based on the small number of workers "00# and resin types used "SPR#\ as well as the distinctly di}erent work tasks and methods which were employed by the other groups considered in the analyses[ For Model "0#\ job group\ resin type and resin con! sumption signi_cantly a}ected styrene concentration "the P!values for the global tests of _xed e}ects for job group\ resin type and resin consumption were all ³9[908#[ As shown in Table 3\ laminators had signi_cantly greater exposure than non!laminators "reference group# and Resins 1 STD\ 2 STD\ 4 LSE and 5 STD all resulted in signi_cantly lower exposures than those of Resin 3 STD "reference resin#[ Regarding Model "1#\ resin type and styrene con! centration signi_cantly a}ected SO concentration "the
Table 0[ Summary statistics for exposures and resin consumption "mean values and standard deviations# for six job groups Group 0[ Hand lamination\ large objects 1[ Hand lamination\ small objects 2[ Spraying and gel coating 3[ Automated lamination 4[ Assembly and mold preparation 5[ Foremen All workers
No[ of workers
Styrene exposure "mg:m2# 2S[D[
SO exposure "mg:m2# 2S[D[
Resin consumption "kg# 2S[D[
097 38 00 12 17 07 126
0452094 0492014 029242[6 44[7221[2 38[6239[2 31[5281[4 0112096
0882078 1022080 2712165 84[2253[1 0392028 60[1266[6 0722073
55249 20220 74239 482043 0[021[1 9[220[0 35255
An investigation of factors contributing to styrene and styrene!6\7!oxide exposures
092
Table 1[ Summary statistics for exposures and resin consumption "mean values and standard deviations# for the various products manufactured Products manufactured Boats Small and form parts Sheets\ elements and car parts Containers and tubes
No[ of workers
Styrene exposure "mg:m2# 2S[D[
SO exposure "mg:m2# 2S[D[
Resin consumption "kg# 2S[D[
026 23 28 17
021285[3 023287[6 89[22016 0952016
1142088 0592058 0152029 75[92015
38230 22221 282019 48274
Table 2[ Spearman correlation coe.cients between log!transformed TWA7h con! centrations for styrene and SO in six groups Group 0[ Hand lamination\ large objects 1[ Hand lamination\ small objects 2[ Spraying and gel coating 3[ Automated lamination 4[ Assembly and mold preparation 5[ Foremen All workers
Table 3[ Point estimates and standard errors for _xed e}ects related to styrene exposure ðModel "0#Ł Group:resin type Intercept "Non!laminators\ Resin 3STD# Laminators Resin 0 STD Resin 1 STD Resin 1 LSE Resin 2 STD Resin 3 LSE Resin 4 STD Resin 4 LSE Resin 5 STD Resin consumption
Standard error
P!value
3[935
9[182
9[9990
0[398 −9[483 −9[752 −9[326 −0[129 −9[346 −9[429 −9[895 −0[072 9[993
9[056 9[279 9[222 9[236 9[296 9[374 9[200 9[202 9[332 9[991
9[9990 9[019 9[909 9[198 9[9990 9[236 9[980 9[993 9[997 9[908
Estimate
P!values for the global tests of _xed e}ects for resin type and styrene concentration were both ³9[9990#\ while the e}ect of job group was not signi_cant "the P!value for the global test of job group was 9[257#[ Focusing upon the individual resins\ as shown in Table 4\ all resins except 3 LSE resulted in exposures that were signi_cantly lower than Resin 3 STD "ref! erence resin#[ Since {resin type| was observed to be a signi_cant predictor of both styrene and SO exposure "the P! value for the global tests of _xed e}ects were both ³9[993#\ this factor was studied in more detail[ Regrouping of the resins into STD and LSE resin types did not reveal any signi_cant di}erence in their e}ects on styrene concentrations[ However\ the use of LSE resins resulted in signi_cantly greater exposures to SO "P³9[9990# than the use of STD resins[
No[ of workers
rs
P!value
097 38 00 12 17 07 126
9[455 9[558 9[153 9[976 9[225 9[106 9[419
³9[9990 ³9[9990 9[322 9[582 9[970 9[277 ³9[9990
Table 4[ Point estimates and standard errors for _xed e}ects related to SO exposure ðModel "1#Ł Group:resin type
Estimate
Standard error
P!value
Intercept "Non!laminators\ Resin 3 STD# Laminators Resin 0 STD Resin 1 STD Resin 1 LSE Resin 2 STD Resin 3 LSE Resin 4 STD Resin 4 LSE Resin 5 STD ln "styrene concentration#
−1[320
9[267
9[9990
9[065 −2[025 −0[705 −9[647 −1[014 −9[165 −0[360 −9[588 −0[385 9[179
9[084 9[307 9[256 9[279 9[230 9[420 9[231 9[235 9[377 9[938
9[257 9[9990 9[9990 9[937 9[9990 9[593 9[9990 9[934 9[992 9[9990
DISCUSSION
During the manufacture of reinforced plastics\ styr! ene and other airborne contaminants\ including SO\ are released[ Although extensive e}ort has been made to study the emissions of styrene during the lamination processes\ very little is known about the formation of SO[ Unsaturated polyesters are dissolved in styrene "39Ð59)# to form a cross!linked structure with the help of a catalyst and an accelerator\ both of which can vary with the type of resin used[ During the cross! linking of polyester polymers\ approximately 09Ð04) of styrene is released to air "Saamanen et al[\ 0880# and\ based upon limited data\ levels of SO range from about 0:499Ð0:0999 of those of styrene "Pfa/i and Saamanen\ 0882^ Rappaport et al[\ 0885#[ The for! mation of SO during this process has been explained in
093
L[ A[ Nylander!French et al[
terms of fragmentation of polymeric styrene peroxide radicals or epoxidation of the styrene monomer\ or both "Brighton et al[\ 0868#[ Rappaport et al[ "0885# recently reported that SO exposure was more highly correlated than styrene exposure with various SO biomarkers in workers| blood\ including SOÐalbumin and SOÐDNA adducts and sister chromatid exchanges[ This observation pro! mpted us to further investigate the possible factors that may a}ect worker exposure to airborne styrene and SO in the reinforced!plastics industry[ Exposure to SO was positively correlated with exposure to styr! ene in selected job groups\ being strongest for hand lamination followed by assembly and mold prep! aration "Table 2#[ However\ the styreneÐSO cor! relations were weak for the other three job groups "Table 2#\ suggesting that factors other than coex! posure to styrene contributed to SO levels[ In mixed model analyses\ an important factor for predicting both styrene and SO concentrations was the type of resin used in the lamination process[ On the other hand\ job group was a strong predictor for styrene exposure level but a nonsigni_cant predictor for SO exposure level[ When the analysis was per! formed after regrouping the resins into STD or LSE resins\ no signi_cant di}erences in their e}ects on styrene concentrations were observed[ This _nding is rather surprising because LSE resins are\ in fact\ thought to reduce styrene exposure "Andersson et al[\ 0883^ Kalliokoski and Jantunen\ 0870^ Saamanen et al[\ 0880^ Skrifvars et al[\ 0881#[ More work is needed to con_rm this _nding[ On the other hand\ LSE resins were observed to signi_cantly increase SO con! centrations[ Since LSE resins were presumably designed to reduce exposure to styrene without regard for coexposure to SO\ this _nding is perhaps not sur! prising[ Possible explanations for the increase in SO exposure associated with LSE resins relate to the com! position of the resin\ the reduction of evaporation of styrene and:or the amount of catalyst used[ Since the measurements of styrene and SO exposure analyzed in this study were gathered for other pur! poses about 09 years ago it is likely that operating conditions have changed somewhat and that current exposures are probably lower today ðe[g[\ "Symanski et al[\ 0887#Ł[ However\ because the basic types of work "represented by the job groups# and the types of resins are quite similar to those currently observed in the reinforced!plastics industry\ the major _ndings concerning the factors a}ecting exposures to styrene and SO should still be relevant[ We are currently conducting a detailed inves! tigation to determine whether other observable fea! tures of the various production processes contribute to SO exposure and to an increase in SO biomarkers in the blood of reinforced plastics workers[ If SO is con_rmed to be the principal causative agent for the observed biomarkers in these workers\ this type of mixed!models multiple linear regression analysis will assist in clarifying and understanding general patterns
of exposure to styrene and SO and in developing and testing measures to prevent occupational exposures to these compounds[ Acknowled`ements*We thank Pirkko Pfa/i and the Finnish Institute of Occupational Health\ Helsinki\ Finland for pro! viding the data for the analysis[ We are grateful to the com! panies and workers of the Finnish reinforced plastics indus! try for participating in the study[ The collection and analysis of the exposure data was _nanced by the grant from the Finnish Work Environment Fund "TSR 77960#[ This work was supported by grants from National Institute of Environ! mental Health Sciences "T 21 ES96907# and National Cancer Institute "RO0CA58352#[ REFERENCES Andersson\ I[!M[ and Rosen\ G[ "0883# Controlling exposure to styrene during lamination of large moulds[ Evaluation of the IVF booth[ Arbete och Halsa 07\ 178Ð183[ Andersson\ I[!M[\ Rosen\ G[\ Sturm\ M[ and Wielaard\ P[ "0883# Styrene emission and occupational exposure in hand lay!up moulding of glass _bre using two types of polyester[ Occup[ Hy`[ 0\ 196Ð106[ Anonymous "0885# 2M Organic Vapor Monitors "Styrene# 2499:2409:2419:2429[ Technical Data Bulletin (015\ 2M Occupational Health and Environmental Safety\ St[ Paul\ MN[ Bond\ J[ A[ "0878# Review of the toxicology of styrene[ Crit[ Rev[ Toxicol[ 08\ 116Ð138[ Brighton\ C[ A[\ Pritchard\ G[ and Skinner\ G[ A[ "0868# Technology and environmental aspects[ In] Styrene Poly! mers\ pp[ 26Ð39[ Applied Science Publisher\ Ltd[\ London[ Csanady\ G[ A[\ Mendrala\ A[ L[\ Nolan\ R[ J[ and Filser\ J[ G[ "0883# A physiologic pharmacokinetic model for styr! ene and styrene!6\7!oxide in mouse\ rat and man[ Archs[ Toxicol[ 57\ 032Ð046[ Filser\ J[ G[\ Schwegler\ U[\ Csanady\ G[ A[\ Greim\ H[ and Kessler\ W[ "0882# Species!speci_c pharmacokinetics of styrene in rat and mouse[ Archs[ Toxicol[ 56\ 406Ð429[ Fjelstad\ P[ E[\ Thurud\ S[ and Wannag\ A[ "0868# Styrene oxide in the manufacture of reinforced polyester plastics[ Scand[ J[ Work Environ[ Health 4\ 051Ð052[ Fleig\ I[ and Thiess\ A[ M[ "0867# Mutagenicity study of workers employed in the styrene and polystyrene pro! cessing and manufacturing industry[ Scand[ J[ Work Environ[ Health 3\ 143Ð147[ Hopkins\ B[ L[\ Conard\ R[ J[\ Dangel\ R[ F[\ Fitch\ H[ G[\ Smith\ M[ J[ and Anger\ W[ J[ "0875# Behavioral tech! nology for reducing occupational exposures to styrene[ Appl[ Beh[ Anal[ 08\ 2Ð00[ Hopkins\ B[ L[\ Conard\ R[ J[ and Smith\ M[ J[ "0875# E}ec! tive and reliable behavioral control technology[ Am[ Ind[ Hy`[ Assoc[ J[ 36\ 674Ð680[ IARC "0883# Some industrial chemicals[ In] IARC Mono! `raphs on the Evaluation of Carcino`enic Risk to Humans\ Vol[ 59\ pp[ 122Ð235[ International Agency for Research on Cancer\ Lyon[ Kalliokoski\ P[ and Jantunen\ M[ "0870# Control of airborne styrene in reinforced plastics industry[ In] 1nd World Con! `ress of Chemical En`ineerin`\ Montreal\ October 3Ð8[ Kalliokoski\ P[\ Koistinen\ T[ and Jaaskelainen\ M[ "0873# Prevention of styrene hazards*hygienic approaches[ In] Industrial Hazards of Plastics and Synthetic Elastomers\ pp[ 068Ð175[ Alan R[ Liss\ Inc[\ New York[ Kalliokoski\ P[ J[\ Saamanen\ A[ J[\ Ivalo\ L[ M[ and Kokotti\ H[ M[ "0877# Exposure to styrene can be controlled[ Am[ Ind[ Hy`[ Assoc[ J[ 38\ 5Ð8[ Korn\ M[\ Gfrorer\ W[\ Filser\ J[ G[ and Kessler\ W[ "0883# Styrene!6\7!oxide in blood of workers exposed to styrene[ Archs[ Toxicol[ 57\ 413Ð416[ Lof\ A[\ Brohede\ C[ and Gullstrand\ E[ "0882# The e}ec!
An investigation of factors contributing to styrene and styrene!6\7!oxide exposures tiveness of respirators measured during styrene exposure[ Int[ Archs[ Occup[ Environ[ Health 54\ 18Ð23[ Mark\ H[ F[\ Bikales\ N[ M[\ Overberger\ C[ G[\ Menges\ G[ and Kroschwitz\ Z[ J[ "0877# Encyclopedia of Polymer Science and En`ineerin`\ John Wiley and Sons\ New York[ Mayo\ F[ "0847# The oxidation of unsaturated compounds[ V[ The e}ect of oxygen pressure on the oxidation of styr! ene[ J[ Am[ Chem[ Soc[ 79\ 1354Ð1379[ Mendrala\ A[ L[\ Langvardt\ P[ W[\ Nitschke\ K[ D[\ Quast\ J[ F[ and Nolan\ R[ J[ "0882# In vitro kinetics of styrene and styrene oxide metabolism in rat\ mouse and human[ Archs[ Toxicol[ 56\ 07Ð16[ Nakajima\ T[\ Elovaara\ E[\ Gonzales\ F[ J[\ Gelboin\ F[ J[\ Vainio\ H[ and Aoyama\ T[ "0882# Characterization of the human cytochrome P349 isozymes responsible for styrene metabolism[ In Butadiene and Styrene] Assessment of Health Hazards\ IARC Scienti_c Publication No[ 016 "Edited by Sorsa\ M[ et al[#\ pp[ 090Ð097[ International Agency for Research on Cancer\ Lyon[ Parkyn\ B[\ Lamb\ F[ and Clifton\ B[ V[ "0856# In] Unsatu! rated Polyesters and Polyester Plasticisers\ Vol[ 1\ pp[ 46Ð 56[ American Elsevier Publishing Company\ Inc[\ New York[ Pekari\ K[\ Nylander!French\ L[ A[\ Pfa/i\ P[\ Sorsa\ M[ and Aitio\ A[ "0882# Biological monitoring of exposure to styrene*Assessment of di}erent approaches[ J[ Occup[ Med[ Toxicol[ 1\ 004Ð015[ Petreas\ M[ X[\ Woodlee\ J[\ Becker\ C[ E[ and Rappaport\ S[ M[ "0884# Retention of styrene following controlled exposure to constant and ~uctuating air concentrations[ Int[ Archs[ Occup[ Environ[ Health 56\ 16Ð23[ Pfa/i\ P[ and Saamanen\ A[ "0882# The occupational scene of styrene[ In] Butadiene and Styrene] Assessment of Health Hazards\ IARC Scienti_c Publication No[ 016 "Edited by Sorsa\ M[ et al[#\ pp[ 04Ð15[ International Agency for Research on Cancer\ Lyon[
094
Pfa/i\ P[\ Vainio\ H[ and Hesso\ A[ "0868# Styrene and styrene oxide concentrations in the air during the lami! nation process in the reinforced plastics industry[ Scand[ J[ Work Environ[ Health 4\ 047Ð050[ Rappaport\ S[ M[\ Yeowell!O|Connell\ K[\ Bodell\ W[\ Yager\ J[ W[ and Symanski\ E[ "0885# An investigation of multiple biomarkers among workers exposed to styrene and styrene!6\7!oxide[ Cancer Res[ 45\ 4309Ð4305[ Saamanen\ A[ and Enbom\ S[ "0883# A mobile local ven! tilation system equipped with an automatic position control[ Arbete och Halsa 07\ 296Ð201[ Saamanen\ A[ J[\ Niemela\ R[ I[\ Blomqvist\ T[ K[ and Nikan! der\ E[ M[ "0880# Emission of styrene during the hand lay! up molding of reinforced polyester[ Appl[ Occup[ Environ[ Hy`[ 5\ 689Ð682[ Skrifvars\ M[\ Johansson\ M[\ Blomqvist\ T[ and Keinanen\ K[ "0881# Low styrene emission unsaturated polyester resins[ Composites 2\ 31Ð34[ Sorsa\ M[\ Anttila\ A[\ Jarventaus\ H[\ Kubiak\ R[\ Norppa\ H[\ Nylander\ L[\ Pekari\ K[\ Pfa/i\ P[ and Vainio\ H[ "0880# Styrene revisited] exposure assessment and risk esti! mation in reinforced plastics industry[ Pro`[ Clin[ Biol[ Res[ 261\ 076Ð084[ Sumner\ S[ J[ and Fennell\ T[ R[ "0883# Review of the meta! bolic fate of styrene[ Crit[ Rev[ Toxicol[ 13\ S00Ð22[ Symanski\ E[\ Kupper\ L[ L[ and Rappaport\ S[ M[ "0887# Comprehensive evaluation of long term trends in occu! pational exposure] part 0\ Description of the database[ Occup[ Environ[ Med[ 44\ 299Ð298[ Todd\ W[ F[ "0874# Styrene vapor control systems in FRP yacht plants[ Am[ J[ Ind[ Med[ 7\ 108Ð121[ Tornero!Velez\ R[\ Denovan\ L[\ Waidyanatha\ S[ and Rap! paport\ S[ M[ "submitted# Measurement of styrene!6\7! oxide and other oxidation products of styrene in air[ Walling\ C[ and Heating\ L[ "0854# Radical formation in the reactions of t!butyl hydroperoxide with styrene[ J[ Am[ Chem[ Soc[ 76\ 27Ð36[