Document peb5ZDMoGwg62k0o4O3K0Bvzk

This article was downloaded by: (UOEH- Journal of Occupational and Environmental Hygiene] On: 15 October 2009 Access details: Access Details: [subscription number 768371016] Publisher Taylor & Francis lnforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Journal of Occupational and Environmental Hygiene Publication details, including instructions for authors and subscription information: http ://'<WIW.i nformaworl d. co m/s mppltitl e-content=t713657996 Benzene Exposure in Industries Using or Manufacturing Paint in China-A Literature Review, 1956-2005 Hong Liu ab; Youxin Liang a; Stephen Bowes c; Hongzhi Xu cd; Yimei Zhou c; Thomas W. Armstrong c'; Otto Wong a<gh; A. R. Schnatter c; Jinbin Fang a; Laiming Wang a; Liping Nie '; Hua Fu '; Richard Irons 'i a Fudan University School of Public Health, Shanghai, China' Shanghai Chemical Monitoring Station for Environment Protection, Shanghai, China" Exxon Mobil Biomedical Sciences, Inc., Annandale, New Jersey" Division of Biostatistics, Yale University School of Public Health, New Haven, Connecticut' TWA8HR Occupational Hygiene Consulting, Branchburg, New Jersey' Applied Health Sciences, San Mateo, California g University of North Texas Medical Center, Fort Worth, Texas h Chinese University of Hong Kong, Hong Kong, China' Fudan-Cinpathogen Clinical and Molecular Center, Institutes of Biomedical Sciences, Fudan University, Shanghai, China i University of Colorado at Denver and Health Sciences Center, Denver, Colorado First Published on: 01 November 2009 To cite this Article Liu, Hong, Liang, Youxin, Bowes, Stephen, Xu, Hongzhi, Zhou, Yimei, Armstrong, Thomas W., Wong, Otto, Schnatter, A. R., Fang, Jinbin, Wang, Laiming, Nie, Liping, Fu, Hua and Irons, Richard{2009)'Benzene Exposure in Industries Using or Manufacturing Paint in China-A Literature Review, 1956-2005',Journal of Occupational and Environmental Hygiene,6:11 ,659670 To link to this Article: DOl: 10.1080/15459620903249646 URL: http://dx.doi.org/1 0.1080/15459620903249646 PLEASE SCROLL DOWN FOR ARTICLE Full terrr.s and c:=Jndi ticns of use: h t t p : / /w"'r"'J. informatJcr ld. 2om/t_erms-a~Ld-ccndi t_ic~s-c-a:::c:eE .s .pdf This a~ticle may be used for research, teachirg ~nJ private study purpuses. Any substantial ur systeTatic reproduction, re-d~stribution, re-selling, loan or sub-licensing, syst?matic supply or drst~ibution i r any form to anyone is expressly forbidden. Tht:: puDll.sheL doe.s nul_ give dflY \\/::iLLr:Hll_y expLe.s.s o_r .impl.ied o_r mdk.e dllY _repre.s:::::rtL::iLlun LhciL Llte cunler1L.s v1ill be comp-=._ete or accurate or u~~ to date_ Tl-_e zccu~acy of any ins tructicn3, formulae and drug doses sh::ould be independently verif:'ed .vith primary sources. The publisher shall not be liable for sny l::Jss, actions, cla~ms, proceedings, dem3nd or costs or damages whatsoever o~ ho~soev~r caused arising directly o_r lndlrec Lly _in connec Llon v... r_ Lh ::JL a_r_is_ing oL L o_:::- Lhe use oi Lhls ma l_e_r lal. CGU BEN0000371 ]ou,nur! of()rnquJfirmal and Fm)inHlmPnftJ! FfygiPnP, fl fl.59-fl70 TSSN: 1545-9624 print I 1545-9632 online Copyright 2009 JOEH, LLC DOl: 10.10R0/15459fl209032491i4fl Benzene Exposure in Industries Using or Manufacturing Paint in China-A Literature Review, 1956-2005 Hong Liu,1, 10 Youxin Liang,1 Stephen Bowes,2 Hongzhi Xu,2'3 Yimei Zhou,2 Thomas W. Armstrong, 2-4 Otto Wong,1-5- 7 A.R. Schnatter, 2 Jinbin Fang,1 Laiming Wang,1 Liping Nie,1 Hua Fu,1 and Richard lrons8-9 1Fudan University School of Public Health, Shanghai, China 2 ExxonMobil Biomedical Sciences, Inc., Annandale, New Jersey 3 Yale University School of Public Health, Division of Biostatistics, New Haven, Connecticut 4TWA8HR Occupational Hygiene Consulting, Branchburg, New Jersey 5 Applied Health Sciences, San Mateo, California 6 University of North Texas Medical Center, Fort Worth, Texas 7 Chinese University of Hong Kong, Hong Kong, China 8 Fudan-Cinpathogen Clinical and Molecular Center, Institutes of Biomedical Sciences, Fudan University, Shanghai, China 9 University of Colorado at Denver and Health Sciences Center, Denver, Colorado 10Shanghai Chemical Monitoring Station for Environment Protection, Shanghai, China A systematic review ufthe Chirucse literature was curulucted exposure for painting tasks in the reported industries over from 1956 to 2005. The survey included hath online and time. marmal searching, as well as npert disc!lssions aimed at providing insight intofactors affecting benzene exposure levels in paint/coatings indu,,tries. Data extraetedfrom 204 papers included: (7) year of occurrence. (2) type of paint/coatings J!ruducts, (3) tyr!e of' industries where the [Jruducts were used or produced, (4) job titles and work activities, (5) type of lit- {Supplementctlmaterials are availablefur this article. Go tu the f'Ublisher's online edition of' the Journal of Occupational and Environmental Hygiene for the following fiee supplemental n:suurce: tables detailing benzene concentration for different types ofpainting and other task' in Chinese industry.} erature searched, (6) working conditions whenever datu were rl ,; available, and (7) exposure levels_ }\;fast hen:ene measurements were short-term samples for comparison with the Chinese Keywords henzene poi:-;oning, coatings, e-xposure assc~:-..ment, industrial hygiene, occupational health, organic solvents maximum allowable concentration standard. The accuracy and precision of the ,wmpling and analytical methods were not reported The distribution of benzene concentrations was tested and found to fit neither nonnal nor logrwnnal distributions. Analysis orvariance (comparisonfor more than two groups) and t-rest (comparison for nvu groups) were curulucled Address conespondence to Hua Fu, School of Public Health, Fudan University, PO. Box 248, 138 Ylxueyuan Road, Shangai 200012, China; e-mail hln (ill Indan edn en_ on Blom-transf'onned benzene crmcentulfion data. The over- all median benzene exposure levels were 215, 82, 31, and 6 mg!m3 during the periods 1956-1978, 1979-1989, 19902001, and 2002-2005, re.~pectively. }\dean hen;:ene expnsure was significantly lower for paint mamifacturing tlwn paint INTRODUCTION E xposures ro benzene, toluene, xylene, styrene, and other organic solvents have long been recognized as major spraying. No significant difference was found among paint occupational hazards in industries using or manufacturing types and hen::ene exrwsure for paint applicatinn. Benzene exposure was significanTly higher in workplaces judged to have poor ventilation. No significant differences were found in benzene exposure as a function of industry type. Even though substwltially lower when compared with levels in the paint/coating~. Yin et at<I) studied 528,729 workers exposed to benzene or benzene mixtures in China and showed that 2~0,000 (53%) were paint workers. Dosemeci et aJ.12l estimated historical exposure to henzene in China in multiple [Jil."\'1~ recent benzene exposure m,easureJnenls ,'\uggested that industries and multiple occupations, involving 672 factories in many facilities in the paint/coatings indMstries in China still have benzene concentrations that are above the current China nccupatinnal expnsure limit fnr hen:;:_ene (6 mg/m3 a~ a rime- weighted average). Benzene concentrations from the present exercise, while not directly supporting quantitative retrospec- 12 cities in China. The highest benzene exposure was observed among rubber workers and painters (spray, electrostatic, drip painters, and paint mixers). Paint manufacturing workers were also highly exposed,( 2) averaging benzene levels greater than tive expo~llre estimnting, pmvide insight rm relative benzene 20 ppm (64 mg/m3 ). Journal of Occupational and Environmental Hygiene November 2009 659 CGU BEN0000372 Early paints and coatings were unique combinations of natural oils and resins that generated relatively little environmental impact. This type of paint was still used in China during the 1950s through early 1970s. Since the late 1980s, the types of paint and coatings used in China have grown and can be categorized into three broad types: (I) oil-based, (2) water-based, and (3) powder coatings. Oil-based coatings include oleoresin, natural resin, phenol resin, and bituminous coatings. More stringent environmental regulation and rapid development of material sciences resulted in increasing use of less toxic, even "zero VOC" products (e.g., powder coatings, UV curing coatings, and some colloidal coatings). The ratio of production of oil-based to synthetic resin coatings in China has been falling since the early 1980s(3 ' This report presents the result of a systematic review of the Chinese literature from 1956 to 2005. The objective was to summarize the Chinese literature on benzene exposure in industries using or manufacturing paint/coatings in China for the period 1956-2005. The temporal trends of benzene exposure levels in the paint/coatings industries in China were analyzed, and some factors affecting worker exposure were identified. These factors provide insight for a subsequent relative ranking of exposures in this industry. Industry group: construction, transport, vehicle/ship/aircraft manufacturing, sport equipment, handicraft/toy manufacturing, machinery equipment manufacturing Workplace ventilation information, if provided by the report author Paint type: nitrocellulose, alkyd resin, amino resin phenolic resin, polyester paint Paint solvent type: benzene, toluene, xylene, gasoline as specified by the report author Type of report: C:ase reports included articles that were initially published following the occurrence of a benzene poisoning case(s); industrial hygiene (IH) surveys including benzene monitoring data and any related benzene poisoning and/or suspected cases found during the survey; other reports included articles of relevance, such as toxicological investigations, summary reports, or short communications Number of benzene poisoning cases reported, if any Location (city) of the study. Analysis The mithmetic means and medians for the reported data were calculated, and the association between average benzene concentration of the exposure and other factors listed above were assessed as follows: METHODS Where only a range of benzene concentrations was pro- Literature Search vided hy the authors, the arithmetic mean of minimum and Computerized scatchcs were can;cd out using the Chinese maximum reported values was taken as the average. If only Biomedical Literature Datahase, China National Knowledge a maximum value was provided, the analysis was based Infrastructure, and the Super Star Reader Database, which on arithmetic mean of the maximum and zero. If only the covered the period between 1978 and 2005. Key words used in minimum was given, then that observation was regarded as these searches included the Chinese equivalents of the follow- missing. ing technical terms: paint/coatings manufacturing, paint/ coat- The distribution of benzene concentrations was tested and rl ,; ings application (spray, brush, immersion), oil-based coatings, found to fit neither normal nor lognormal distribution. Data water-based coatings, paint/coatings thinners, benzene and or- subsets for paint application (spray, bmsh, immersion, mix) ganic solvents exposure, benzene poisoning (aplastic anemia, and for paint making were neither normal nor log-normal. leukemia), industt-ial hygiene, and occupational health. Because the data were highly skewed, the medians instead of Additional searches of print references for pre-1978 data supplemented the online searches, particularly for information describing the development of production technology and chemical compositions. This included reviews of conference proceedings. which partially covered publications before 1978 (a few articles covering years between 1956 and 1978), and professional inquiries regatding the historical development of industrial technologies in the paint/coatings industry with experts at universities and industrial institutions. From each publication, the following information was extracted, as available: the means in each suhgroup were hetter indicators of centTal tendency of original data. For rhe same reason, rank-based transformations (below) were performed so that parametric statistics such as t-test and analysis of variance (ANOVA) could be used to compare different transformed group~. Rank-based nonparametric statistics for original data were used to confirm the results. Because the data were neither normal nor log-normal. parametric analysis techniques such as t-tesl and ANOVA were not applicable without transformation. To transform the data into normal distribution, they were ranked in orders. Then the tank of the i-th data point was changed into a Benzene concentration in mr measurements: number of samples; minimum, average, and maximum benzene concentration; year and type of sampling; analytical method Work task: spray painting, brush painting, immersion painting, paint mixing, paint manufactm-ing, other normal score that corresponded to the i-th data point had the data set been normally distributed. The algorithm for the transformation141 was: _ -1 -Yi- <P (~I"; 0.375) +11 0.25 ("1) 660 Journal of Occupational and Environmental Hygiene November 2009 CGU BEN0000373 20 18 16 14 .1!! 5c. 12 a:(I) 0... 10 .(cI) E8 z::I 6 4 2 0 ,~~,q~, ,~#,~~~*,~~,~~,~~ ""~' ,~~ "~"' "~"' ,~~ ,~~ ,~~ ,~_q, ,~~ ,~~ ,~_'l) ~,~ ~"q) ~rpi ~rpi ~rpi Publication Year FIGURE 1. The distribution of reported articles by years: CR-ease report initially published following the occurrence of a benzene poisoning case, IH-industrial hygiene survey, Oth-other report including toxicology, summary reports, or short communications. where r, is the rank of the i-th observation and n is the number of nonmissing ohsnvations; <P- 1 is the prohit function, The criterion for statistical significance was set at p < 0.05. A11 analyses were performed using the statistical software SAS or the inverse of the standard cumulative normal function. 9.L3. The transformed data were then normally distributed with mean = 0 and standard deviation = 1. The subset of paint RESULTS application and paint-making data after transformation were rl ,; both normaL As a result, t-test and ANOVA could be ap- Search Results plied to the transformed data in subsequent analysis. However. the numbers in the tables from the subsequent results secLion are presented as observed values for the ease of interpretation. The literature search resulted in 204 papers of benzene poisoning case reports, industrial hygiene surveys, and other reports (toxicological studies and article reviews) published in 51 Chinese refereed medical journals or other sources (includ- ing four conference proceedings) from 1956 to 2005 (Figure ANOVA (comparison for more than two groups) and t- 1). In China, according to the National Diagnostic Criteria of test (comparison for two gmups) were conducted on the Occupational Renzene Poisoning,(SJ ben7,ene poisoning was Blom-transformed14J benzene concentration data. Because classified as observation ca,~e, acute poisoning (mild and se- paint-making activities had a generally lower mean ben- vere), and chmnic poisoning (mild and severe) depending zene concentration than other activities, most analyses were on the exposure history, clinical symptoms and signs, and done for paint-using industries and paint-making industries the feature of hematological ksts. However, the information separately except for the AKOVA for job activities, where reported in the literature was often limited to case classification paint-making and the subcategories of paint-using activities (e.g., observation, acute benzene poisoning, chronic benzene were analyzed together. The pair-wise comparisons within poisoning) without elaboration on clinical signs or exposure groups after ANOVA were adjusted for multiple compar- history isons. ANOVA and t-test results from the transformed data The collected data covered 24 provinces, municipalities, were verified by Kruskal-Wallis (nonparamettic counterpatt and autonomous regions, representing 77% of the 31 adminis- for ANOVA) and \Vilcoxrm (nonparametric counterpart for trative regions in China. About half the reported papns were t-test) tests for untransformed data. Results using these from the most developed industrial areas in the coastal regions, alternate tests were consistent. including Guangdong, Jiangsu, Zhejiang, Shandong, Liaoning Journal of Occupational and Environmental Hygiene November 2009 661 CGU BEN0000374 provinces, and Shanghai municipality; one-quarter from the the authors of the source reports did not provide informa- Central and South China regions; and the remaining quarter tion on task frequency or duration that could have enabled from Northeast, North, Northwest, and Southwest regions_ estimation of full-shift exposure_ Available information was This distribution by geographical location was expected, as tabulated. it was primarily dependent on the status of economic development and industrialization. The distribution ofinformation sources was associated with Benzene Exposure Trend The 204 reports yielded 428 data sets. Five sets of data the economic development during the time period covered. The number of published articles increased with the development showed extremely high benzene concentrations (above 20.000 mg/m3) that the authors reported were based primarily on of economic reform in the late 19SOs and peaked between simulation tests to replicate certain acute poisoning episodes. the mid-1990s and early 2000 when demand for paint and These were excluded from subsequent statistical analysis. Re- coatings was high in the shipbuilding, automobile, machinery, household electric appliances, toys, and house reconstruction industries_ The majority of the literature (n = 180) covered industries where paint and coatings were used, including shipbuilding, ports of accidental or emergency benzene exposures were abo excluded from subsequent analysis_ The final analysis was based on 407 sets ofbenzene concentration values representing more than 5500 reported air samples (Pigure 2). These data showed a skewed distribution as illustrated in Figures 3A automobile. fumiture, household electric appliances, machinery, toys, arts/crafts, stationery, sports equipment, and house decorating. Fewer articles (n = 24) reported benzene exposure and 3B. When Blom-transformed, the data and subsets were normally distributed and could be subjected to parametric statistical analysis. in the paint or coatings manufacturing industry. Most of the air sampling approaches represented relatively To assess the benzene concentration trend over time, 317 data sets that included both workplace air benzene concentra- short-term samples, typically from workers' breathing zones. There were three common sampling and analytical methods used for the measurement of benzene in workplace air in China_ tions and the year of measurement were usable. The pattern of benzene exposure was examined in fourtime periods, based on economic development stages: (1) 1956-1978 (pre-economic The earliest employed an impinger with solvent through which the air was bubbled, followed by a colorimetric analysis_ The reform): (2) 1'J7'J-1'J8'J (open policy and the beginning of economic growth); (3) 1990-200 I (further industt;al develop- next method employed glass syringe grab sampling, with gas chromatographic analysis, which was the method used for most ol lhe dala reported in the reviewed literature. This glass syringe collection method may have continued well into the 1990s. ment and growth); and (4) 2002-2005 (after passage of the Occupational Disease Prevention and Control Act in 2002 and implementation of new occupational exposure limit LOEL] for benzene). Descr-iptive statistics for these periods are listed in Table II. The most recent method used a single-section charcoal Substantial differences were found between the median and tube with a 20-min sampling period at a normal flow rate of mean of each group. The skewness of the data suggested that 200 mL/min, also with gas chromatographic analysis (Table lhe median was a beller measure lo assess Lmlransformed daLa_ rl ,; I). Information was not available to compare the accuracy, To compare the benzene level of each subgroup, ANOVA precision, limit of quantification, and interferences for the~e for rank-based Blum-transformed concentration and nonpara- different methods. (6) metric Ktu~kal-Wallis test for untransfonned concentrations Data extracted from the search results were tabulated, trans- were conducted. Pair-wise comparisons were conducted for lated. and sorted by task for spray painting, brush painting, immersion painting, paint mixing, and paint making in the transformed benzene concentration (paint using) with adjustment for multiple comparisons using Tukcy's method (same online supplemental Appendix Tables A-I through A-VI, respectively. \Vhcrc the source reports identified air sampling results fm- two tasks, the data were included in both tables. for all other ANOVA unless otherwise specified). The average benzene level was significantly higher in Petiod l ( 1956--1978) than three other later periods; the level in Period 2 (1979- In many cases, the source reports did not include details related to sampling and analysis, sample number, or factors 1989) was significantly higher than Period 3 ( 1990---200 I) and Period 4 (2002-2005)_ Periods 3 and 4 were nul significantly that could potentially influence benzene exposure, such as ventilation, task, paint type, or industry type. In most cases, different. Generally, the mean exposure levels over the 1956-2005 periods suggested a downward trend_ TABLE I. Summary of Sampling and Analytical Methods by Time Period Period Sampler Flow Rate (mL/min) Sampling Time (min) 19"i0s-1960s 1970s-1980s l990s-present Tmpinger Glass ~yringe Charcoal tube 200 100 mL (volume) 200 10 1 15-20 662 Journal of Occupational and Environmental Hygiene November 2009 Analytical Method Colm;metry Gas chromatography Gas chromatography CGU_BEN0000375 0D 1 E rc::i 0 () r~:: l1;l r:: .~c 0.1 D DD 00 D O.GI 1950 1960 1970 1980 1990 Year of Sampling 2000 2010 FIGURE 2. Average benzene concentration measurements in China for period 1956-2005. Average benzene concentrations from 266 paintusing reports (includes spray, brush, immersion painting, and paint mixing) and from 59 paint manufacturing reports. The average benzene level was significantly higher for paint-using industries as a whole than for paint manufacturing as determined by the Wilcoxon test. PERCENT 40 - PERCENT 80_,---------------------------------------------:,-- 70- 30 60- ,-- rl ,; 20 - 5040 30 10 - ,- nnn .nnr~11 0 0 50 100 150 200 250 300 350 400 450 500 550 600 benzene concentration (A) 20 - 10 n~~~~ ,---- 0~~~~~~~~~~~~~~~~~~-=~~- 0 50 100 150 200 250 300 350 400 450 500 550 600 benzene concentration (B) FIGURE 3. (A) Paint-using frequency distribution histogram. Untransformed average benzene concentration (concentration in mg/m3 ). The column corresponding to "0" concentration represents observed concentrations below 25 mg/ml. (B) Paint-making frequency distribution histogram. Untransformed average benzene concentration (concentration in mg/m3). The column corresponding to "0" concentration represents observed concentrations below 25 mg/m3 . Journal of Occupational and Environmental Hygiene November 2009 663 CGU BEN0000376 TABLE II. China Benzene OELs and Mean Benzene Concentration (mg/m3 ) for Different Historical Periods Period Description OEL (mglm3 ) Mean N Std. Dev. Median !VIaximum Minimum 1956-1978 50 (MAC) 336.84 47 503.77 215.00 3212.30 15.00 2 1979-1989 40 (MAC) 191.77 75 309.87 82.4 1827.50 0.00 3 1990-2001 40 (MAC) 125.5 158 344.59 31.33 3373.50 0.00 4 2001-2005 6 (PC-TWA) 10 (PC-STEL) 81.73 37 187.46 6.46 991.50 0.00 The benzene exposure level for the four historical periods "\\'ere -compared using ANOVA. on tran~fonned data_ The. mean concentrations of these periods were 'ignificanlly dirrerenl (p < 0.0001). The resull wa' confirmed by Kru,kal-Walli' Lesl. and pair-wise comparisons belween group' were adjusted ror mLtiLiple comparisons using Tukcy's n1cthod_ The China OEL prior to 2002 was a maxin1um allowable concentration without averaging time, and compliance was assessed by shmt-term sample. The OEL trom 2001-2005 was 6 mg/m3 permissible concentration S-hour TWA; 10 mg/m3 permissible concentration as a shmt-tenn exposure limit. N refers to number of source data sets and does not add to 407 due to missing data from source reports. The average benzene level was significantly higher in Period l (1956-1978) than three other later periods. the level in Period 2 (1979-1989) was significantly higher than Periods 3 (1990-2001) and 4 (2002-2005). Periods 3 and 4 were not significantly different. Report Type and Benzene Poisoning Cases in industrial hygiene surveys and other reports. while Lhe mean Sixty-fom articles reponed 527 cases of benzene poisoning concentrations in the latter two did not differ significantly. Data at vatious stages of sevetity. Of those. 15 articles that were identified as "case reports" covered 120 cases of benzene on mean benzene concentration for reports of benzene poisoning ca~es vs. nonpoisoning cases are sununarized in Table poisoning (23%); 25 articles that were identified as "IH sur- TV. Results were based on a t-test fm the Blum-transformed veys" covered 297 cases of "benzene poisoning" and/or "sus- concentrations and Wilcoxon test for untransfonned concen- pected benzene poisoning" (56%); and 4 articles that were trations. The p-valuc in Table IV is the result of t-tcst. Mean identified as ''others," which described the toxicity, occupa- benzene concentration was significantly higher in reports of tional exposure, and health effects of benzene in paint/coatings benzene poisoning than in other reports. industry covered 110 cases of related "benzene poisoning" (21%). The regional and year distributions of reported benzene Benzene Workplace Air Concentrations by Work Activity for Industries Using or Manufacturing Paint poisoning cases were consistent with the number of papers One hundred eighty articles identified job activities asso- heing reported and the amount of paint and coatings produced ciated with benzene air concentrations in paint use or paint in the specific areas during the same periods. For example, manufacturing. A total of 330 samples of benzene exposure more than 70% of the total cases were reported from East China levels were associated with five main work activities: (1) brush and Central South China; both having similar proportions of painting, (2) spray painting, (3) paint mixing, (4) immersion the total annual production of 1.8-2.0 million tuns of China painting. and (5) work processes in paint manufacturing. As rl ,; manufactured-paint and coatings produced and used in 1999- shown in Table V, for all years, benzene air concenLrations by 2001.(7,&) work activity can be characterized generally as: brush painting Pair-wise comparisons were conducted on transfom"Led data > spray painting> paint making. Pair-wise comparisons were for report type and whether benzene poisoning was reported conducted on transfonned data for job/task type. for the paint-using groups. Paint making was not analyzed due lVIean benzene concentration was significantly lower in to too few observations. JVIcan benzene concentration for dif- paint manufacturing than spray painting, but the rest of the ferent repotttypes was summari7.ed in Table ITT. Mean ben7ene categories did not show a significant difference. The benzene concentration in case reports was significantly higher than thac air concentrations fur the paint-using industry as a whole TABLE Ill. Mean Benzene Concentration (mg/m3 ) for Different Types of Reports (Paint-Using) Description Mean N Std. Dev. Median Maximum l\llinimum Case report IH investigation Other 291.91 163.72 94.11 43 213 73 369.62 387.66 181.56 171.50 50.20 31.58 1758.10 3373.50 1035.00 1.00 000 0.00 The benzene ail" concentrations for different report types were con1pared using ..!\NOVA on transforn1ed data. The n1ean benzene air concentrations of these report types were significantly different (p = 0.01 131. Mean benzene air concentration for case reports was significantly higher than for industrial hygiene survey and other reports as dctcnnincd by pair-vvisc colTiparison, while the mean concentrations in the latter two did nut differ s-ignificantly. The result was cunfinncd by Kruskal-Wallis test, and pair-wise comparisons between groups were adjusted for multiple comparisons using Tukey's method. N refers to number of source data sets and does not add to 407 due to missing data from o;;ouree reports_ 664 Journal of Occupational and Environmental Hygiene November 2009 CGU BEN0000377 TABLE IV. Mean Benzene Concentration (mg/m3 ) for Benzene Poisoning and Nonpoisoning Cases (Paint Using) Benzene poisoning cases reported? IVIean N Std. Dev. Median Maximum Minimum No 110.43 256 257.84 37.05 3212.30 0.00 Yes 356.48 73 535_34 151.10 3373.50 1.00 Thr Iut=an bt=uzene exposure levels fur benzene poisoning and nunbenzene poisoning cases were significantly ilitferent a:i detenuinetl by V\-7ilcuxun test (p < 0.0001'1. The result was confirmed by ttcst on transformed data. N refers to number of source data sets and docs not add to 407 due to missing data from source repuTtS. (including spray painting, brush painting. paint mixing. and immersion painting) were significantly higher than lhose from the paint manufacturing industry using Wilcoxon test (Figure 2). Benzene Workplace Air Concentrations by Type of Paint! Coatings Used or Produced Forty-one articles identified the type of paint/coatings that were used or produced at work, which consisTed of nirrocellulose paint (20 articles), alkyd resin paint (9), amino resin paint (7), phenolic resin paint (2), and polyester paint (3). These articles reported 128 ~ets of air samples (Table Vl). ANOVA for transformed concentrations showed that no significant dif ferences were found among paint Lypes. (Painl making was nol analyzed, and phenolic paint was not included in the analysis, both due to too few observations.) Comparison of Benzene Air Concentrations by Reported Effectiveness of Workplace Ventilation Seventy-seven articles provided information about the sta tus of ventilation aT lhe workplace. "Poor ventilaTion" (P) denoted that adequate mechanical ventilation facility was not present, but included simple electric fans (e.g., floor fans) or "natural" ventilation. These accounted fur must of the articles; TABLE V. Mean Benzene Concentration (mg/m3) for Different Job Activities Job Activity Historical Period Mean N Median Maximum Minimum Spray 1956-1978 509.95 20 280.30 3212.30 1979-1989 201.42 46 88.46 1827.50 1990-2001 79.76 102 24.78 1758.10 2002-2005 101.72 27 8.15 991.50 All years 156.65 195 43.90 3212.30 rl ,; Brush 1956--1978 220.00 3 200.00 285.00 197Y-19tl9 194.87 14 51.70 885.00 1990-2001 319.20 18 119.05 3373.50 2002-2005 29.43 7 5.33 153.20 All years 222.37 42 58.20 3373.50 Mixing 1990--2001 59.45 12 56.44 139.40 2002-2005 23.82 3 9.37 56.70 All years 52.33 15 53.58 139.40 Immersion 1956-1978 106.80 1 106.80 106.80 1979-1989 115.5(i 8 52.70 461.80 1990-2001 129.79 7 17.13 540.00 All years 121.24 16 27.38 540.00 Paint manufacturing 1956--197!) 267.00 2 267.00 344.00 1979-1989 45.78 10 %.21 127.50 1990-2001 14.05 50 12.06 82.00 All years 27.32 62 15.08 344.00 20.30 0.00 0.00 0.00 0.00 175.00 0.00 000 1.25 0.00 1.00 5.38 1.00 106.80 2.15 1.15 1.15 lSJO.OO 1:1.22 1.00 1.00 Benzene exposme levels for different job activities were summarized. The exposme levels were broken down for different historical periods within each job category. Con1parison bet\veen groups \Vas. conducted for .All years data, not by historical period subgroups. The mean benzene exposure levels (all years) for each job activity were significantly different as detennined by ANOVA on transformed data (p = 0.0111). The result was confirmed by Kruskal-Wallis test, and pair-wise con1paris.uns between groups were adjusted for nlltltiplc cu1nparisuns using Tukcy's method. Mean benzene concentration was significantly luwcT in paint manufacturing than spray painting, but the rest of the categories did not show a significant difference. N refers to number of source data sets and does not add to 407 due to missing data fTom source reports_ Journal of Occupational and Environmental Hygiene November 2009 665 CGU BEN0000378 TABLE VI. Mean Benzene Concentration (mg/m3) for Different Types of Paint Paint Type Mean N Std. Dev. Median Maximum Minimum Nitrocellulose Alkyd resin Amino resin Polyester 215.76 25.g!) 28.43 15.82 29 14 4 9 315.64 22.53 11.01 23.35 59.30 21.93 24.55 5.10 965.00 g6.92 44.61 56.80 0.00 0.00 20.00 0.38 The mean benzene exposure levels for different types of paint were not significantly different as determined by ANOVA on transformed data (p = 0.3537). The result was confirmed hy Kntskai-Wallis test. N refen; to numher of source data sets and does not add to 407 due to missing data from source reports. 78% (60177). "Good ventilation" (G), indicating that adequate mechanical venLilaLion syslem (e.g., air exhausl hood and mechanical fan) was present, accounted for only 22% ( 17/77) of the articles. Summary data are shown in Table VII. At-test was conducted on transformed data for reported workplace ventilation effectiveness. Mean benzene concentration was significantly lower for reports with good ventilation effectiveness for the paint-using groups. Industry Type and Benzene Exposure Ninety articles identified the specific type of industries in which paint or coating~ were used. According to the Classification Standard of National Economy, (9 l the relevant indusuies and/or occupations were grouped into four categories with standard codes as: ( 1) consttuction, including civil engineering and household reconstruction; (2) transport vehicle manufacturing, including shipbuilding, aircraft, automobile, and locomotive vehicles; (3) stationery, sports equipment. handicrafts, and toy manufacturing; and (4) machinery equipment manufacmring, including electric engines, valves, lathes, abrasion wheels, plus paint making (Table VIII). ANOVA and KruskalWallis tests both indicated that no significant differences were found among industry types and benzene concentration for the painL-using groups. (PainL making was nOL analyzed.) DISCUSSION Trend of Benzene Workplace Air Concentrations in Paint/Coatings Industry in China This report summarizes occupational benzene exposure dala from the Chinese literature for the period 1956-2005. Four hundred twenty-eight data sets were extracted from 204 reports in 51 Chinese journals. (The total numbers in each analysis would not add to 428 due to missing values or exclusion.) After data transformation, ANOVA was used to detect associations between benzene exposure and several workplace factors. The published benzene exposure data ranged from undetectable to very high henzene concentrations (over 100 times the China occupational exposure limit for each historical period). In general, the mean benzene concentration for each historical period followed a downward trend, though the last two periods had simila.- mean concentrations. It was neither the intem nor design of this report to assess compliance with the China OEL. However, it was interesting rl ,; TABLE VII. Mean Benzene Concentration (mg/m3) for Different Workplace Ventilation Effectiveness Workplace Ventilation Job Activity Mean N Median Maximum Minimum Poor Good Spray Brush Mixing Immersion Paint manufacture (unspecified) Total Spray Brush Mixing Immersion Paint manufacture Total 226.68 465.57 36.43 203.12 31.51 250.99 192.53 82.82 105.50 9.63 613.62 113.91 ::\0 II 8 3 20 7 79 21 0 4 3 2 30 Tl.51 90.10 26.05 68.20 17.60 130.32 49.50 6.21 111.65 5.60 613.62 18.61 ::\212.::\0 3373.50 120.06 540.00 127.50 800.00 3373.50 991.50 139.40 20.20 1180.00 1180.00 000 1.25 1.00 1.15 12 06 7.00 0.00 0.01 59.30 3.10 47.23 0.01 Mean benzene concentrations for workplaces were sorted by reported ventilization effectiveness and job activity. Con1.parison between groups \Vas conducted for Tnwl data, not by job activity subgroups. Benzene concentrations for poor ventilation (total) were significantly different from those for good ventilation (total) as determined by Wilcoxon test (p = 0.0439). The result was confirmed by t-test on transformed data. N refers to number of somce data <cts and dues not add to 407 due to nllssing data from source reports. VVorkplace ventilation as reported by authors of the source reports: '~Poor ventilation" included simple electric fans (e.g., ffonr fans) or 'natural" ventilation~ "Good ventilation'~ indieated that adequate mechanieal ventilation systetn (e.g., air exhaust hood and mechanical fan). 666 Journal of Occupational and Environmental Hygiene November 2009 CGU BEN0000379 TABLE VIII. Mean Benzene Concentration (mgfm3 ) for Different Types of Industry Industry Historical Period Mean N Median Maximum Minimum Construction Vehicle Craft Machinery Paint making 1956--1978 1979-19S9 1990--2001 2002-2005 All years 1979--1989 1990--2001 2002-2005 All years 1956-1978 1979--1989 1990--2001 2002-2005 All years 1956--1978 1979--1989 1990--2001 2002-2005 All years 1956-1978 1979--1989 1990--2001 All years 151.26 326.53 53.86 195.84 129.92 167.07 18.63 131.18 157.27 49.21 78.74 36.37 7R.99 423.75 184.73 22.55 37.50 142.70 267.00 45.78 14.05 27.32 0 5 18 15 38 9 16 5 30 3 2 19 4 28 11 29 35 76 2 10 50 62 52.40 110.81 5.38 51.00 115.00 22.50 5.33 35.09 165.00 49.21 54.50 26.34 56.90 285.00 97.00 15.73 37.50 37.15 267.00 36.21 12.06 15.08 600.00 3373.50 406.05 3373.50 243.60 1758.10 44.61 1758.10 200.00 68.20 300.00 8Y.61 300.00 1225.00 1104.70 171.66 37.50 1225.00 344.00 127.50 82.00 344.00 6.50 1.35 0.38 0.38 8.57 0.81 0.00 0.00 106.80 30.21 0.98 3.20 0.98 20.30 2.15 0.01 37.50 0.01 190.00 13.22 1.00 1.00 The benzene exp0sure levels frmn ditferent types of industry were sun1n1arized. 'The expu~ure levels broken down for ditferent historical periods "\Vi thin each industry were also shown There was no significant differences among the mean benzene exposure levels (all years) for these indusu-ies (not including paint making) as determined by ANOVA on transformed data (p = 0.9518). Tite re,ult was confirmed by Kruskal-Wallis test.~ refers to number of source data sets and does not add to 407 due tu missing data frotn source reports. Construction includes civil engineering and household reconstruction; Vehicle includes transport vehicle Inanufacturing, including shipbuilding, aircraft, autotnobile, and locotnotive vehtcle~. Crafts include~ stationery~ ~ports equtpment, handicrafts., and toy manufacturing. Mochinc1y includes equipment manufacturing, including electric engines, valveso lathes, abrasion wheels. to observe from Table II that mean benzene air concentrations coatings were derived from these primary categories. Infor- rl ,; were higher than the national OELs in effect at the time for mation about the precise type of paint/coatings reported in the each historical period. Even for the most recent time period, literature was limited. In this report, no significant differences mean exposure was 81.73 mg/m3 exceeding the current OEL (6 mg/m3 time-weighted average, TWA) and the old OEL of were found among paint types and benzene concentration fur the paint-using groups. 50 mg/m3 in effect from the 1950s until the late 1970s. (Until 2002, the OELs in China were defined as maximum allowable Status of Ventilation concenttations in workplace air.)<10l The overall benzene levels at workplaces subjectively char- acterized as having good ventilation effectiveness were as- Factors Affecting Benzene Exposure Levels Job Activities The job activity seemed to be an important determinant affecting benzene exposme level in the paint/coatings industry. As shown in Table V, benzene exposure levels by work activity were characterized as: brush painting> spray painting> paint making. Paint-making exposures were significantly lower than spray painting exposures. sociated with lower benzene air concentrations than those with poor ventilation effectiveness. As suggested from Table VII, exposures above the OEL were commonly reported even for facilities with ventilation effectiveness judged "good" by the authors of the reports. Reliable prevention of overexposure to benzene in the painting industry can require costly ar1d carefully designed and maintained ventilation systems,\12! which was not commonly implemented until the passage of the Occupational Disease Prevention and Control Act in 2002. Type of Paint/Coatings As shown in Table IX, 17 categories of paint/coatings have been commonly used in China nationwide since the early 1980s.Cl I) Meanwhile, a variety of new products of paint/ Type of Industries using Paint/Coatings There were no significant differences in mean benzene con- centration by industry type or type of paint or coming. However, there was large dispersion in the data, and all industries Journal of Occupational and Environmental Hygiene November 2009 667 CGU BEN0000380 TABLE IX. Categories of Coatings in China Coating Type Benzene Potential? Solventsffhinners Commonly Used(ll,lli-IY) Oil-based coatings Yes 200 solvency gasoline (petroleum distillaLe with a boiling range of 140--200C) containing certain amount of aromatic hydrocarbons; mainly but not purely xylene and toluene Natural resin coatings No Combination of natural oils and resins such as shellac Phenolic resin coatings Yes Solid phenolic resin is soluble in most polar organic solvents and aromatic-oxygenated solvent mixtures Bituminous coatings Yes (1) 200 solvency gasoline containing certain amount of aromatic hydrocarbons, mainly but not purdy xylene and toluene; and (2) "heavy benzene"+ kerosene Alkyd resin coatings Yes 200 solvency gasoline containing certain amount of aromatic hydrocarbons, mainly but not purely xylene and toluene Amino resin coatings Yes Xylene, butyl alcohol, ethyl acetate in proportion of 80: I 0: I 0 Nitro-cellulose coatings Yes "Xiangiiaoshui," a solvent may contain as high as 20-40% benzene Cellulosic coating No Vinyl per-chloride coatings Yes Esters, ketones, and solvents containing benzene, toluene, m- xylene Alkene-based resin coatings Uncertain Epoxy resin coatings No Cyclohexanone, xylene, butyl alcohol Polyester coatings Yes Likely to be naphtha or mineral spirits Acrylic resin coatings Yes 1\1ost acrylics require strong ~olvent~; esters, alcohol~, and benzene, toluene, or xylene are commonly used Polyurethane coatings No xylene, butyl acetate, cyclohexanone Elemental organic coatings No Rubber coatings Yes Other types of coatings Varies had reported exposure measurements substantially higher than The computer and manual literature searches yielded 204 the OEL. papers of benzene poisoning case repmts, industrial hygiene surveys, and other reports (toxicological studies and article Solvent Composition reviews) published in 51 Chinese medical journals or other The decreasing OELs for benzene during the period of this sources from 1956 to 2005_ rl ,; review, and recognition of the health impact of benzene, led to The collected data covered 24 provinces, municipalities, the current controls on benzene in solvents in China. Benzene and autonomous regions representing 77% of the '11 admin- content in adhesives and coatings have been regulated as less istrative regions in China. than 0.5% (i.e., <5g/kg) since 2001.<131 Available exposure data were translated from Chinese, tab- ulated, and summarized, and additional information on the Outlook paint/coatings industry in China (online Appendix Tables Several actions in China offer hope for future reduction in A-VII through A-IX) were identified to improve the under- benzene exposure in the painting industry including: standing of histm;cal exposures involving paint/coatings in More stringent environmental regulations set for paint/ coat- China. ings products. such as VOCs ::::200 giL and benzene _::::0.5%(!3) Adoption of an indoor air quality limit for benzene of 0_11 mg/m3(14l Some of the limitations associated with the published information and the analyses currently presented include: Improvement of occupational health services after the passage of the Occupational Disease Prevention and Control Act in 2002 and the 1educed occupational exposure limits for benzene (personal exposure ::::6 mg/m3 T\VA and ::::10 rng/rn 3 STRT ,).<lUJ Strengths and Limitations of This Review ~trengths of this review included: Consistent with industrial hygiene practice in China, most of the literature data represented short-term (1-min grab) personal breathing zone or area (often unspecified which) samples from the work environment (Table I)_ Construction of a task TWA lO estimate full-shift exposure was not undertaken as part of this review, in part, because the literature reports generally did not provide task duration and task frequency information_ 668 Journal of Occupational and Environmental Hygiene November 2009 CGU BEN0000381  The relative accuracy and precision of the sampling and measurements by charcoal tube against the STEL, rather that1 analytical methods used over the years in China has not TWA measurements. Even though substantially lower when been established_ cmnpared with levels in the past, recent benzene concen- The current analysis relied predominately on two types of tration measurements suggested that many facilities in the data; industrial hygiene surveys, and case reports. Case paint/coatings industries in China still have benzene task con- reports had higher average concentrations than III surveys centrations above the current China OEL for benzene (STEL (Table III). It was uncommon for the purpose of the indus- or TWA). trial hygiene surveys to be reported-they were commis- The information presented here is not adequate by itself for sioned for a vatiety of reasons and were not necessarily quantitative estimation oflong-term average benzene exposure targeted at the ~ituations with hen7ene poisoning or high hecause most nf the data were short-tenn samples, and infor- benzene exposures. mation on task duration and frequency was not provided by the authors of the source reports. Without information on the SUMMARY AND CONCLUSIONS accuracy and precision of the Chinese sampling and analytical methods, comparison with Western data is problematic. With T he objective of this review was to summarize the Chinese literature on benzene exposure in industries using or these limitations, the data do allow identification of primary correlates of exposure that can be potentially useful in relative manufacturing paint/coatings in China for the period 1956- ranking of exposure in this industry 2005. The literature survey included both online and manual searching and 1csultcd in 204 papers of benzene poisoning case reports, industrial hygiene surveys, and other reports ACKNOWLEDGMENTS (toxicological studies and article reviews) published in 51 Chinese medical journals or other sources. The collected data W e are indebted to the Benzene Health Research Consortium for sponsoring the Shanghai Health Studies, covered 24 provinces, municipalities, and autonomous re- as well as Jerry Rice, chair of Scientific Review Panel, and gions, representing 77% of the 31 administrative regions in his associates for their reviews and comments on the draft of China. the manuscript. The authors are grateful to Xipeng lin and Most of the reports included air sampling approaches that Xueshat1 Feng at Fudan University School of Public Health, represented relatively short-term samples, typically from work- at1d Shixing Yang at Shanghai Municipal Institute of Public ers' brealhing zones, for comparison wilh lhe Chinese max- Health Supervision for advising on indust1ial exposure assess- imum allowable concentration standard. The accuracy and rnenL and/or slalisLi~.:al analysis_ \Ve also lhank Limin Wu al the precision of the sampling and analytical methods were not Department of Macromolecular Sciences ofFudan University, reported. The final analysis was based on 407 sets of benzene Guangli Xiu at the College of Resource & Environmental concentration values representing more than 5500 reported Engineering, East China University of Sciences and Tech- air samples. These data showed a skewed distribution and nology, and experts at China Coatings Industry Association rl ,; was transformed for subsequent analysis. The published ben- for their information sharing on technological developments zene exposure data ranged from undete~.:table to very high and environment protection relevant to the coatings indust1-y benzene concentrations (over 100 times the China occupa- in China_ tional exposure limit for each of four historical periods). In generaL the mean benzene concentration for each historical period followed a downward trend, though the last two periods REFERENCES had similar mean concentrations. Mean benzene exposure was significantly lower for paint manufacturing than paint spraying. No significant differences were found among paint types and mean benzene concentration for the paint-using groups. Benzene exposure was significantly higher in workplaces judged to have poor ventilation. No significant differ- I_ S.N. Yin, Q. Li, Y. Lin, F. Tian, C. Du, and C Jin: Occupational exposure to henzene in China_ Br: -'- Jnd Med 44:192-llJ:'i (1987)_ 2. M. Dosemeci, G.L. Li, R.B. Hayes, et al.: Cohort study among worker'> exposed to benzene in China: TL Exposure assessn1ent. Am_ 1. Tnd. Med. 26:4()1-411 (1994)_ en~.:e~ were fuund in benzene expo~tne as a funt:tiun of industry type. From 1956 to 200 I, the China OEH for benzene was expressed as a maximum allowable concentration (without a specified averaging time) a11d complia11ce was assessed by 3. Zhao, Q.S.: The development and prospects of paint industry in China (llJ7l>-llJlJl). China Coatings 2:7-11 (llJlJ3). [In Chinese] 4 Blom, G_ Staliaical E.,'linulfe.'' an.d Transformed Belll 1kJri11hles_ New Ymk: John "Wiley & Sons, Tnc., 1958. 5. Ministry of Health, PR China: National Diagnostic Criteria a( OccupLtfun-ull Benzene Pmson1ng (GBZ6~-2002). Beijing: Ministry of short-term (1-10 min) samples. With the passage of the Oc- cupational Disease Prevention and Control Act of 2002, the permissible concentration for benzene was reduced to 6 mg/m3 as an 8-hr TWA exposure; 10 mg/m3 as a 15-min STEL Most of the recent data summarized here were short-term Health, 2002. [Tn Chinese] 6. Wang, L.M., Y.l\1. Zhou, Y.X. Liang, et a!.: Benzene exposure in the shoetnaking lndustry in China, a literature survey, 1978-2004. Regul. Toxicol. Phannacol. 46148-156 {2006)_ 7 _ Yang, H., and T.Y. Huang: The development of China coating' industry. Tn Yearbooks of China National Coatings lndustry Association, 200()_ Journal of Occupational and Environmental Hygiene November 2009 669 CGU BEN0000382 PP- 1-10. Beijing_ Chine"e Library Cla"ification (CLC) No_ F426.7-54. 13_ State General Administration for Quality Supervision and Inspec- [In Chinesel tion: Restriction of Hazardous Materials in Coalings (GB I 8581-2001 ). 1\. Sun, S.H., P.Q. Sun, and D.Z. Liu: The development and prmpects of Beijing, 200L [In Chinese] coatings industry in China_ Shanghai Coati11g> 41 (4): 19-22 (2003). [In 14_ State General Administration for Quality Supervision and Inspec- Chinesel tion: State Envirumnental Prute<...:tion Ad1ninistratiun and Ministry of 9. National Statistics Bureau, National Standard Bureau-General, and Health. Indoor Air Quality Standards (GB/T 18883-2002), 2002_ [In National Cenus Oflice of State Counst:l: The Classification Standard Chim"e] of' National Economy. Beijing: National Statistics Bureau Office, 1982. 15. Caldwell, D.J., T.,V. Armstrong, N.J. Barone, J.A. Suder, and M.J. [lnChmese] Evans: Hydrocarbon solvent exposure data: Comparison and analysis o 10. Liang, Y.X., 0. Wong, L. Yang, L. Tao, and S. Zhi: The development the literature_ A THAI 61(6):881-894 (2000). and regulation of occupational exposure liinits in China. Regul. To:ticol. 1G. Lin, G.J., and Y.Z. Geng: A.pplied Sciences and Tcclmolor:ies o( Plzarmacol. 46:107-113 (2006)- Coatings. Beijing: China Light Indu"tty, 1994. [In Chinesej 1I. National Standard Bureau, PR China: Classification, nomenclature 17 _ Sheng, C.L.: Handbook of Coatings Formula. Beijing: China Petroleum and specification of coatings (GR2705-R I). In Handhnnk nf' t:naring,, Chemi<try, 2000_ [Tn Chinese] Fnmnila, C.L Sheng (ed.). Beiji11g: China Petroleum Chemistry Pub- lR_ \Vang, Y.H.: Formulation and AppliL-alion ofCoating.i. Chengdu, China: lisher, 2000. [In Chinesel Sichuan s~iences ami Technology, I 987. [Tn Chinesel 0 12. ACGIH Committee on Industrial Ventilation: Industrial Ventilation 19_ Xu, B.G., and B.G. Zhang: Handbook o( Coatings Application. co - A Manual of Recommended Practice for Design, 26th eeL Cincinnati, NanJing, China: Jiangsu St:i-ences and Technolugy, 2000. p.l4. H .;0; Ohio: ACGIH, 2007. [In Chinesel +' 0 0 ::, ,, "' c0 ~ r",'l'", +' ~ g -~ > """",c'; rl g' ~ ,'~", 0 ~ 0 rl ,; c H 5 e0 ""0'' D ,, P1 ] 1 0 rl c ~ n 670 Journal of Occupational and Environmental Hygiene November 2009 CGU BEN0000383