The burden of cancer at work: estimation as the first step to prevention

Transcription

1 1 Imperial College London, Department of Epidemiology and Publi Health, Faulty of Mediine, St Mary s Campus, London, UK; 2 Health and Safety Laboratory, Mathematial Sienes Unit, Buxton, UK Correspondene to: Lesley Rushton, Department of Epidemiology and Publi Health, Imperial College London, Faulty of Mediine, Norfolk Plae, London W2 1PG, UK; l.rushton@imperial.a.uk Aepted 29 November 2007 Published Online First 31 January 2008 The burden of aner at work: estimation as the first step to prevention L Rushton, 1 S Huthings, 1 T Brown 2 ABSTRACT Objetives: Work-related aners are largely preventable. The overall aim of this projet is to estimate the urrent burden of aner in Great Britain attributable to oupational fators, and identify arinogeni agents, industries and oupations for targeting risk prevention. Methods: Attributable frations and numbers were estimated for mortality and inidene for bladder, lung, non-melanoma skin, and sinonasal aners, leukaemia and mesothelioma for agents and oupations lassified as International Ageny for Researh on Caner (IARC) Group 1 and 2A arinogens with strong or suggestive evidene for arinogeniity at the speifi aner site in humans. Risk estimates were obtained from published literature and national data soures used for estimating proportions exposed. Results: In 2004, men and women died from aner in Great Britain. Of these, 7317 (4.9%) deaths (men: 6259 (8%); women: 1058 (1.5%)) were estimated to be attributable to work-related arinogens for the six aners assessed. Inidene estimates were (4.0%) registrations (men: (6.7%); women 2054 (1.2%)). Asbestos ontributed over half the oupational attributable deaths, followed by silia, diesel engine exhaust, radon, work as a painter, mineral oils in metal workers and in the printing industry, environmental tobao smoke (non-smokers), work as a welder and dioxins. Oupational exposure to solar radiation, mineral oils and oal tars/pithes ontributed 2557, 1867 and 550 skin aner registrations, respetively. Industries/oupations with large numbers of deaths and/or registrations inlude onstrution, metal working, personal and household servies, mining (not metals), land transport and servies allied to transport, roofing, road repair/ onstrution, printing, farming, the Armed Fores, some other servie industry setors and manufature of transport equipment, fabriated metal produts, mahinery, non-ferrous metals and metal produts, and hemials. Conlusions: Estimates for all but leukaemia are greater than those urrently used in UK health and safety strategy planning and ontrast with small numbers ( annually) from oupational aidents. Soures of unertainty in the estimates arise prinipally from approximate data and methodologial issues. On balane, the estimates are likely to be a onservative estimate of the true risk. Long lateny means that past high exposures will ontinue to give substantial numbers in the near future. Although levels of many exposures have redued, reent measurements of others, suh as wood dust and respirable quartz, show ontinuing high levels. There is inreasing interest in estimating and omparing burdens of disease generally 1 and for aner. 2 3 Estimates an identify major risk fators Original artile and high-risk populations, support deisions on priority ations for risk redution and provide an understanding of important ontributions to health inequalities. Nearly 30 years ago Doll and Peto (1981), in their report to the US Congress, presented a method of estimating the effets of different fators on aner mortality in the USA 4 ; their estimate for oupational fators was 4% of all US aner deaths with an unertainty range of 2% 8%. The aim of this projet is to produe an updated and detailed estimate of the urrent burden of oupational aner in Great Britain (GB) that will help to inform the development and prioritisation of pratial measures to redue the burden in the future, speifially in GB, but also more generally in the developed world. The estimates of urrent burden of oupational aner are based on exposure levels from up to 50 years ago when exposure levels may have been muh higher than they are at present. Prioritisation for preventive effort requires onsideration of ongoing risks and urrent exposures. The next phase of the projet will inlude preditions of future burden based on urrent exposure levels. In this paper the outomes of the first phase of the projet are presented. Estimates have been made of the urrent burden due to past oupational exposures for six aners, whih are important in terms of both the annual numbers of deaths and aner registrations they produe and their potential to be aused by exposure to oupational arinogens. The six are aner of the bladder; leukaemia; aner of the lung, mesothelioma; nonmelanoma skin aner (NMSC) and sinonasal aner. An overview of the methodology developed and the data used is also given. METHODS Oupational arinogens assessed At two international workshops held as part of the projet to disuss the methodology ( hse.gov.uk/researh/hsl_pdf/2005/hsl0554.pdf; hsl0732.pdf) the partiipants advised that priority should be given initially to International Ageny for Researh on Caner (IARC) Group 1 and 2A oupationally related arinogens. Agents or oupations in these IARC groups were inluded that had either strong or suggestive evidene of arinogeniity in humans for the speifi aner site, as defined by Siemiatyki et al (2004) and subsequent IARC publiations. 5 7 Those with strong evidene were defined as established arinogens for the purposes of this study and those with suggestive evidene of arinogeniity in humans were defined as unertain Oup Environ Med 2008;65: doi: /oem

2 Original artile arinogens. In addition there had to be substantial existing exposures in GB and/or ases of aner still ourring due to past exposures. Data soures (i) Caner mortality and registration data Estimation was arried out on a aner by aner basis for 2004 for mortality and 2003 for aner inidene, the most reent years for whih published data were available at the time of estimation. Deaths for 2005 and aner registrations for 2004 are now available but total numbers do not differ substantially from those used. Mortality data were obtained from ONS, Mortality Statistis, Series DH2, for England and Wales and the General Register Offie for Sotland. Caner inidene data were obtained from ONS, Caner Statistis, Registrations, Series MB1 for England, the Sottish Caner Registry, ( and the Welsh Caner Intelligene and Surveillane Unit ( fm?orgid = 242). (ii) Risk estimates Standard searh riteria were used to identify key studies, metaanalyses or pooled studies, taking into aount relevane to GB, large sample size, effetive ontrol for onfounders, adequate exposure assessment, and lear ase definition. Where only a narrative review was available giving a range of risk estimates from several relevant studies a ombined estimate of the relative risks (RRs) was alulated based on a random- (for heterogeneous RRs) or fixed- (for homogeneous RRs) effets model. If no meta-analysis, pooled study or narrative review were available a single key study was seleted using the riteria above. Dose-response risk estimates were generally not available, nor were proportions of those exposed at different levels of exposure over time available for the working population in GB. In our study separate risk estimates were generally extrated relating to an overall higher level and an overall lower level. For one or two speifi agents it was possible to extrat risk estimates for three levels of exposure or for speifi exposure senarios (see table 2 footnotes). Where no estimate ould be identified for very low/bakground/environmental levels of exposure, an RR of one was arbitrarily assigned. (iii) Exposed population estimates If the relative risks were extrated from an industry-based study population, for example a ohort study, a national (external) data soure was used for estimating the proportion of the population exposed. If the relative risks were extrated from a population-based study, for example a ase-ontrol study of aner registry ases, an estimate of the proportion of ases exposed was also obtained from the study, although suh studies were rarely available for GB. The national data soures used were the CARinogen EXposure (CAREX) database, 8 and for exposures not overed by CAREX, the annual Labour Fore Survey (LFS) ( and the Census of Employment (CoE) ( Data from CAREX are not differentiated by sex; 1991 Census data by industry and oupation were used to estimate the relative proportions of men and women exposed ( uk). The industry ategories listed in CAREX were alloated to higher and lower exposure ategories assuming the distributions of levels of exposure and risks assoiated with these broadly mathed those of the studies from whih RRs were extrated. Statistial analysis The attributable fration (AF), that is the proportion of ases that would not have ourred in the absene of exposure, has been estimated. There are two prinipal methods for estimating the AF, both of whih depend on knowledge of the risk of the disease due to the exposure of interest and the proportion of the population exposed. 9 To estimate the AFs for eah aner/oupational arinogen Levin s equation 10 was used if risk estimates ame from an industry-based study, review or meta-analysis together with estimates of the proportion of the population exposed from independent soures of data. Miettinen s equation was used if risk estimates and proportion of ases exposed ame from a population-based study. 11 The equations used are given in the Statistial Appendix. The AFs were applied to total numbers of aner-speifi deaths (2004) and aner registrations (2003) to give attributable numbers. Where AFs were only available for mortality these were used for estimation of attributable registrations and vie versa. Similarly if separate AFs for women ould not be estimated those for men or for men and women ombined were used. To take aount of aner lateny a relevant exposure period (REP) was defined as the period during whih exposure ourred that was relevant to the development of the aner in the target year For solid tumours a lateny of at least 10 years and at most 50 years was assumed giving an REP of For haematopoieti neoplasms 0 20 years lateny was assumed giving an REP of The proportion of the GB population exposed to the oupational arinogens of onern over the REP was estimated taking into aount hanges in numbers employed in the primary and manufaturing industry and servie setors in GB over the REP. Figures from the LFS show, for example, that the numbers of men employed in primary and manufaturing industry were 40% higher in the 1970s than they were in the early 1990s, whereas the numbers employed in the servie industries were 10% lower. Adjustment for employment turnover over the period for grouped main industry setors was also arried out using LFS data on the distribution of length of time with urrent employer (in exess of 1 year) by length of employment. This gave the numbers ever employed for at least 1 year during the REP allowing for normal life expetany to 2004 (see Statistial Appendix equation 3). The adjustment fators for hanging employment levels and perentage annual turnovers used are shown in the table in the Statistial Appendix. The AF for mesothelioma was derived diretly from several studies of UK mesothelioma ases that suggest that between 85% and 90% of male mesothelioma ases are due to oupational exposure (Darnton, personal ommuniation). Studies in whih results were reported separately for females in the UK (Darnton, personal ommuniation) and elsewhere gave estimates of 20% 30%. For the estimate of the AF due to the Established plus Unertain arinogens, ases desribed as due to paraoupational (eg, exposure from living near an asbestos fatory or handling lothes ontaminated due to oupational exposure) or environmental exposure to asbestos are also inluded. A reent analysis of lung aner mortality for the whole of GB between 1980 and 2000 by oupational group in relation to indies of asbestos exposure and smoking habits suggested that the ratio of asbestos-related lung aner to mesothelioma deaths is between two-thirds and one. 16 A ratio of 1:1, mesothelioma to lung aner deaths has been used for the estimation of numbers of lung aners attributable to asbestos. 790 Oup Environ Med 2008;65: doi: /oem

3 For lung aner assoiated with radon exposure from natural soures, estimates of rates of lung aner due to exposure to radon in domesti buildings were applied to estimates of the time employees spend in workplaes where radon exposure ours. AFs for all the relevant arinogeni agents and oupational irumstanes were ombined into a single estimate of AF for eah separate aner. AFs in general annot be summed diretly if there is a possibility that workers will have been exposed to more than one oupational arinogen during their working lifetimes in the relevant exposure period. Where data allowed, the exposed numbers were therefore partitioned between overlapping exposures, for example by exluding steel foundry workers from the CAREX estimates of numbers exposed to other lung arinogens. Alternatively, where exposure to more than one arinogen assoiated with the same aner site ourred, an AF was estimated only for the dominant arinogen with the highest risk estimate. The method of ombining the AFs was then determined by whether there was residual exposure to multiple arinogens. If so, it was assumed that the exposures were independent of one another and that their joint arinogeni effets were multipliative. Suh multiple/overlapping exposures were assigned to exposure sets that were judged to be non-overlapping with other exposure sets and single exposure senarios. The AFs within exposure sets were multiplied using equation 5 in the Statistial Appendix. The ombined AFs for eah non-overlapping exposure set were then summed, together with non-overlapping single exposures. An overall AF for oupation for the six aners assessed so far was estimated by summing the attributable numbers for the six aners and dividing by the total number of aners in GB (table 1). Although it is relatively straightforward to estimate a onfidene interval for AFs of single arinogeni agents, 9 the methodology for estimating onfidene intervals for AFs estimated from more than one risk estimate for multiple exposure levels and for ombinations of AFs is more omplex, for example requiring Monte Carlo methods. The methodology for this is urrently under development and onfidene limits are not presented. Separate tehnial reports for eah aner giving full details of data and alulations, and a report expanding on the statistial methodology are aessible at rrhtm/rr595.htm. RESULTS The overall oupational AFs for the six aners investigated so far are summarised in table 1. Six per ent (n = 4693) of aner deaths in 2004 in men and 1.0% (n = 701) in women in GB have been estimated to be due to oupation for arinogens with strong human evidene of arinogeniity, our established arinogens. The estimates were 4.9% (7317 deaths) in total, 8.0% for men (6259 deaths) and 1.5% for women (1058 deaths), for arinogens with strong or suggestive evidene of arinogeniity in humans, our established plus unertain arinogens. The ombined AFs for registrations are 6.7% (n = ) for men in 2003 and 1.2% (n = 2054) for women based on established and unertain arinogens. These are lower than that for deaths beause of the very large numbers for NMSC. Table 2 gives the number of deaths for eah aner attributable to eah of the agents and oupations onsidered (so far) for established and unertain arinogens together. The studies that were used for the risk estimates for eah agent or Original artile oupation and the type of study are also given in table 2. Overall, asbestos exposure ontributes the largest number of deaths (mesothelioma and lung aner), followed by exposure to silia (lung), diesel engine exhaust (DEE) (lung, bladder), radon exposure from natural exposure in workplaes (lung), oupation as a painter (lung, bladder), mineral oils in metal workers (bladder, sinonasal, NMSC), environmental tobao smoke (ETS) in non-smokers (lung), mineral oils in printers (lung), oupation as a welder (lung) and exposure to dioxins (lung). For the six aners, exposures in the onstrution industry are estimated to produe over half of GB s oupational attributable aner deaths in men (n = 3219). Workers in this industry are exposed to 17 of the arinogens onsidered so far (13 resulting in at least one death), shown in fig 1 for men, who aount for an estimated 99% of the onstrution workfore in the REPs for these aners. There are 44 deaths for NMSC attributable to oupational exposure to mineral oils, polyyli aromati hydroarbons (PAHs) and solar radiation. However, estimated numbers of registrations for NMSC assoiated with mineral oils are 1745 males (M), 122 females (F), with PAHs; mainly oal tars and pithes are 547 M (544 in onstrution), 3 F; and with solar radiation are 1824 M (805 in onstrution), 733 F. Table 3 gives for eah aner, numbers of deaths (registrations for NMSC) within industry setors or jobs for whih there were at least 50 estimated attributable aners; the exposures onerned are listed, with those ontributing most (at least 10 aners in men plus women) being shown in bold. Painters and welders are assumed to be exposed to many different arinogens. The importane of single exposures within some industry setors is also highlighted, for example PAHs in oal tar and pithes in roofing and road repair and onstrution, metal working fluids in the metal industries, mineral oils and printing inks in the printing industry. In addition to the onstrution industry table 3 shows that multiple exposures potentially our in several other industries, inluding the manufature of industrial hemials (18, notably asbestos) and other hemial produts (16, also notably asbestos), manufature of transport equipment (15, partiularly asbestos, hromium, obalt, silia, radon and solar radiation), eletriity, gas and steam (15, notably asbestos and solar radiation), non-ferrous metal basi industries (14, notably arseni), the manufature of fabriated metal produts (14, notably obalt, hromium and silia), the manufature of mahinery exept eletrial (13, notably silia, hromium, obalt and radon), servies allied to transport (13, notably DEE and solar radiation), and printing, publishing and allied industries (12, notably solar radiation). More than 10 different exposures were also found in sanitary and similar servies (14, notably solar radiation and asbestos), personal and household servies (11, notably asbestos, diesel engine exhaust, ETS and radon), and land transport (11, notably DEE, asbestos and solar radiation). Table 3 also highlights the range of industry setors where partiular substanes are ourring and ontributing to the burden of oupational aner. These setors are not always those where substantial historial exposures have ourred. For example, the main oupations with substantial historial exposure to inorgani arseni inlude hot opper smelting, manufaturing of arsenial pestiides and sheep-dip ompounds, fur handlers and vineyard workers and some miners In GB the majority of exposure ours in the non-ferrous metal basi industry and the manufature of wood and wood and ork produts (44 and 31 lung aner deaths, respetively). Oup Environ Med 2008;65: doi: /oem

4 Original artile Table 1 Estimated attributable frations, deaths and registrations by aner site in 2004 (2003 for registrations) Attributable numbers Attributable fration (%) Deaths Registrations Caner site Male Female Total Male Female Male Female (a) Established arinogens only (IARC Group 1, strong human evidene) Bladder Leukaemia Lung Mesothelioma 85 90{ 20 30{ { 75{ NMSC Sinonasal Total Based on deaths Based on registrations (b) Established + unertain arinogens (IARC Group 1 and 2A, strong + suggestive human evidene) Bladder Leukaemia Lung Mesothelioma 98* 90* 97* { 270{ NMSC Sinonasal Total: Based on deaths Based on registrations Total aners in GB *Inludes ases desribed as due to paraoupational or environmental exposure to asbestos. {Taken as equal to attributable deaths for this short-survival aner. {Mid-points of ranges used when estimating attributable numbers and ombining results for mesothelioma with the other aners. GB, Great Britain; IARC, International Ageny for Researh on Caner; NMSC, non-melanoma skin aner. Although potential asbestos exposure ourred in large numbers of workers (over in the REP) in the mining industry (exluding oal mining) giving 305 deaths eah from lung aner and mesothelioma in men, the industry with the greatest potential for asbestos exposure was the onstrution industry, ourring for example in asbestos removal or stripping, giving 1012 deaths eah from lung aner and mesothelioma in men. In personal and household servies, 362 deaths eah from lung aner and mesothelioma ourred (221 eah of these in women). Other industry groups where asbestos exposure ontributed to fairly large numbers of deaths in men for both lung aner and mesothelioma were work in land transport 45 and manufature of transport equipment. 39 Other substanes ourring aross several industry setors that ontributed substantially to the burden of aner are listed below. Diesel engine exhaust In addition to 21 male bladder aner deaths and a total of 268 lung aner deaths attributed to exposure to DEE in the land transport industry where over workers were estimated to be potentially exposed over the REP, an additional half a million workers were exposed to DEE over the REP in the onstrution industry giving 20 male bladder aner deaths and 238 male lung aner deaths. ETS (non-smokers) Signifiant numbers of workers were exposed to ETS in the wholesale and retail trade, restaurants and hotels, onstrution, and finaning, insurane, real estate and business servies giving 104, 35 and 29 lung aner deaths, respetively. Radon There are now very few workers in metal ore mining in the UK exposed to radioative radon and its progeny. High levels of radon in the workplae our in similar areas to those of onern in residential dwellings in the UK suh as Cornwall, Devon Northamptonshire and parts of Derbyshire, Somerset, Wales, Grampian and the Highlands of Sotland. Approximately 2000 lung aner deaths a year have been estimated to be due to radon exposure of whih between about 90 and 275 are attributable to exposure ourring in the workplae. 55 Although any workplae in the affeted areas is potentially at risk of exposure, the large numbers of workers employed in the wholesale and retail trade, restaurants and hotels, and in finane, insurane, real estate and business servies gave relatively high estimated numbers of lung aner deaths 75 and 47, respetively. Silia In GB the majority of workers exposed to silia work in the onstrution industry, manufature of other non-metalli mineral produts and manufature of pottery, hina and earthenware giving an estimated 667, 28 and 25 deaths of men in these industries, respetively. Solar radiation The risk for NMSC aused by oupational exposure to solar radiation is diffiult to estimate beause everyone is exposed to sunlight to a greater or lesser degree depending on residential loation and leisure time ativities. Risk estimates from a USbased ase-ontrol study of 6565 ases of NMSC were used that estimated separate risks for work that ombined indoor and outdoor work, outdoor work by non-farmers and farming Oup Environ Med 2008;65: doi: /oem

5 Table 2 Caner deaths in 2004 attributable to oupation, by exposure and aner site Caner site Bladder Leukaemia Lung Mesothelioma NMSC Sinonasal Total Referene (numbers in the referene list)1 Type of study M F M F M F M F M F M F M F Exposure{ 22{{{,111 10{{{, Aromati amines Sorahan et al (1998) 17 West Midlands hospitalbased ase-ontrol study Arseni Lee-Feldstein (1986) 18 US opper smelter ohort Asbestos Darnton et al (2006) 16 " Ratio of lung aner to mesothelioma deaths in asbestos-exposed jobs Proportion of mesotheliomas judged due to oupational Darnton (personal ommuniation)"" asbestos exposure Benzene Collins et al (2004) 19 Industry ohort study 2{{{ 4{{{ 2 4 Lewis et al (2000) 20 Industry ohort study Bloemen (2004) 21 Industry ohort study Beryllium Ward et al (1992) 22 US industry ohort 5{{{ 2{{{ 5 2 Cadmium Verougstraete et al Review of industry 13{{{ 5{{{ 13 5 (2003) 23 studies Chromium Cole & Rodu (2005) 24 " Meta-analysis 56{{{ 18{{{ 1{{{ 0{{{ Rosenbaum & Stanbury US industry ohort (1996) 25 {{ Cobalt{{{{ Moulin et al (1998) 26 Frenh industry ohort Diesel engine Lipsett & Campleman Meta-analysis exhaust (1999) 27 " Coggon et al (1984) 28 " Case-ontrol death ertifiate study Boffetta & Silverman Meta-analysis of industry (2001) 29** ohorts""" Dioxins Kogevinas et al (1997) 30 IARC multi-national ohort""" Kheifets et al (1997) 31 Meta-analysis Eletromagneti fields{{{{ Zhong et al. (2000) 32 Meta-analysis Environmental tobao smoke for non-smokers Ethylene oxide Coggon et al (2003) 33 Industry ohort study 0{{{,{{{ 0{{{,{{{ 0 0 Teta et al (1999) 34 Meta-analysis 4***,{{{ 2***,{{{ Formaldehyde Mannetje et al (1999) 35 {{ Pool of population-based ase-ontrol studies Original artile Coggon et al (2003) 36 {{ UK industry ohort Collins & Lineker Meta-analysis of industry (2004) ohorts Czene et al (2003) 37 Swedish ohort study Hairdressers and barbers (oupation) Continued Oup Environ Med 2008;65: doi: /oem

6 Original artile Table 2 Continued Caner site Bladder Leukaemia Lung Mesothelioma NMSC Sinonasal Total Referene (numbers in the referene list)1 Type of study M F M F M F M F M F M F M F Exposure{ Ionising radiation Blettner et al (2003) 38 Multi-national oupation group ohort Lead Steenland & Boffetta Meta-analysis 38{{{ 7{{{ 38 7 (2000) 39 Leather dust1111 Fu et al (1996) 40 English shoe manufaturing workers ohort Tolbert (1997) 41 ** Review""" 243{{{ 13{{{ 17{{{ 1{{{ 20{{{ 2{{{ Eisen et al (2001) 42 {{ US automobile industry Mineral oils (metalworkers) ohort Roush et al (1980) 43 {{ US ase-ontrol study Leon et al (1994) 44 Industry ohorts""" Mineral oils (printers) Nikel Sorohan & Williams Clydah refinery ohort 6***,{{{ 2{{{ 3***,{{{ 1{{{ 8 3 (2005) 45 " Seilkop & Oller (2003) 46 " Review of industry studies Grimsrud & Peto Clydah refinery ohort (2006) 47 {{ Non-arsenial Aquavella (1998) 48 Meta-analysis of industry pestiides ohorts Painters Chen & Seaton (1998) 49 " Meta-analysis of ohort (oupation) studies Bosetti et al (2005) 50 ** Quantitative review of 18{{{ 0{{{ industry-based studies PAHs (general) Armstrong et al (2004) 51 " Meta-analysis of industry ohorts**** Unwin et al (2006) 52 " Narrative review of Boffetta et al (1997) 53 ** industry ohorts""" Partanen & Boffetta Meta-analysis of ohort (1994) 54 studies in asphalt workers PAHs (oal tars and pithes) Radon NRPB (2000) 55 Attributable domesti death rates applied to employees time at work Silia Kurihara & Wada (2004) 56 Meta-analysis Steenland et al (2001) 57 Cohort pool Solar radiation Freedman et al (2002) 58 US death ertifiate- 17{{{ 5{{{ 17 5 based ase-ontrol study Sorahan et al. (1994) 59 UK industry ohort Steel foundry workers Textile dust{{{{ Lue et al (2002) 60 Pool of population-/ hospital-based aseontrol studies""" Continued 794 Oup Environ Med 2008;65: doi: /oem

7 Table 2 Continued Caner site Bladder Leukaemia Lung Mesothelioma NMSC Sinonasal Total Referene (numbers in the referene list)1 Type of study M F M F M F M F M F M F M F Exposure{ Ambroise et al (2006) 61 Meta-analysis Welders (oupation){{{{ 21*** 0*** 21 0 Wood dust Demers et al (1995) 62 Pool of population-based ase-ontrol studies Established exposures only{ Established plus unertain exposures{ *Totals do not always sum aross rows due to rounding error. {Numbers for the separate exposures do not sum to the ombined exposure totals due to allowane made for overlapping exposures. {Estimates have not been made for some IARC Group 1 and 2A arinogens. Reasons inlude: relevant exposures had eased in GB by 1950 (rubber industry/for bladder aner); very small or unknown numbers of workers exposed (BCME and CME, act&bc, epihlorohydrin, haematite mining, 1,3-butadiene,ethylene oxide in men); no relative risk (RR) estimates were available (4,49methyl bis(2-hloroaniline) and styrene-7,8-oxide for bladder aner, benzo[a]pyrene, benzo[a]anthraene & dibenz[a,h]anthraene for NMSC, isopropanol manufature, strong aid proess for sinonasal aner); workers were also exposed to another dominant arinogen (boot and shoe manufature and repair inluded under benzene for leukaemia, and under exposure to aromati amines before 1962 for bladder aner, rubber industry exposure inluded under exposure to hromium, admium, silia and PAHs for lung aner). 1Where two referenes are given, the first was used for a higher exposure risk estimate and the seond for a lower/bakground exposure risk estimate, unless otherwise stated. "Lung aner. **Bladder aner. {{Sinonasal aner. {{Non-melanoma skin aner. 11Leukaemia. ""Mesothelioma. ***Based on three exposure levels. {{{Based on separate exposure senario ategories. {{{RR for bakground exposure level was set to 1, giving AF = RRs from inidene studies used. For all other estimates RRs from mortality studies or meta-analyses ombining mortality and inidene studies were used. """Inverse variane weighted average RR estimated by study team using RRs given in the referene. ****A unit relative risk estimate was used to derive exposure level-speifi RRs. {{{{Exposure lassified as IARC 2B; inluded in the unertain group. For obalt, estimated lung aner deaths were based on total numbers exposed to obalt, with or without exposure to tungsten arbide. Cobalt with tungsten arbide is lassified as IARC 2A. {{{{Low exposed RR estimated by the study team as 1+(RR high -1)/ Boot and shoe manufature and repair. act&bc, a-hlorinated toluenes & benzoyl hloride; BCME, bis(hloromethyl)ether; CMF, hloromethyl methyl ether; IARC, International Ageny for Researh on Caner; NMSC, non-melanoma skin aner; PAH, polyyli aromati hydroarbon; RR, relative risk. Original artile Oup Environ Med 2008;65: doi: /oem

8 Original artile Figure 1 Lung aner deaths for men in 2004 attributable to work in the onstrution industry. ETS, environmental tobao smoke; PAH, polyyli aromati hydroarbon. Large numbers of registrations were estimated for the onstrution industry (860), publi administration and defene (armed fores) (232), wholesale and retail trade, restaurants and hotels (168), land transport (166), manufature of transport equipment (154), agriulture and hunting (143) and ommuniation (132). A table giving industry setors and oupations with at least 10 attributable deaths and/or registrations for eah of the six aner sites assessed so far (50 registrations for NMSC) by oupational exposure is given in a supplementary table online. DISCUSSION All oupational aners are potentially avoidable. Our estimate of the urrent burden in 2004 of six aners due to past oupational exposures of 8% for men and 1.5% for women translates to over 7300 aner deaths in GB. This is in ontrast to the 223 deaths due to oupational injuries that ourred that year ( Burden estimates from other studies range between 3% and 10% With the exeption of leukaemia, all our updated estimates are greater than those of Doll and Peto (1981). 4 The steep rise in asbestos-related deaths from lung aner and mesothelioma sine 1981 has made a major ontribution to the inrease. Our methodology and the data available in GB have allowed a more detailed investigation of the arinogeni agents, oupational irumstanes and industry setors than has been possible in other burden estimation studies. We have also addressed the potential to be exposed to several arinogens onurrently and the impat on total burden. The results must be onsidered taking aount of several unertainties and limitations. These are disussed below and the potential impat on the estimates is indiated in table 4. Agents lassified by IARC as Group 1 and 2A arinogens were assessed. Other substanes suh as IARC Group 2B arinogens, many of whih may be treated as if they were human arinogens in regulatory settings have not yet been evaluated; our estimates ould thus be too low. Unertainty or bias may have been introdued in the hoie of the study for obtaining data for the risk estimates, for example if the exposures in the soure study did not reflet those experiened in GB or distributions of onfounders differed between the soure population and GB. A major gap in available information was a lak of separate risk estimates for women and/or aner inidene. The use of risk estimates derived from studies of men for women and mortality risk estimates for inidene may have biased the AFs. Epidemiologial studies of oupational groups often result in a healthy worker effet, that is a redued overall risk estimate ompared to the general population. This together with potential mislassifiation of exposure in epidemiologial studies ould lead to an underestimation of the true effet and thus an underestimation of the burden. Most of the risk estimates from the published literature were related to some estimate of umulative exposure. In assigning higher and lower ategories to the CAREX industry groups impliit assumptions were made regarding the similarity of durations and intensities of exposure between the soure and target (national) populations. National data are not generally available on the proportions of those exposed at different levels of exposure. Where no risk estimate ould be identified for very low/ bakground/environmental levels of exposure, a risk estimate of one was arbitrarily assigned to the lower group, giving a zero attributable burden. This implies an assumption that a threshold existed in the dose relationship between exposure and effet ontrary to usual risk assessment guidelines for arinogens; this may have ontributed to underestimation of the burden; a large number of people exposed at low levels assoiated with a low risk of disease may ontribute more to the burden than a small number exposed at high levels assoiated with a high risk. In most oupational epidemiologial studies very short-term workers, for example those employed for less than 1 year, are exluded. Our turnover fator was thus alulated exluding workers with less than 1 year s employment. Inlusion of these would have inreased the numbers ever exposed onsiderably. For example, for the onstrution industry, the annual turnover would inrease from 13% exluding workers with under 1 year of employment to 22% when they are inluded. The overall effet of inluding these short-term workers would be to inrease the AFs and attributable numbers. However, when these short-term workers are exluded the turnover fator 796 Oup Environ Med 2008;65: doi: /oem

9 Table 3 Industry setors and oupations with an estimate of a total of at least 50 attributable deaths (registrations for NMSC) by aner site and oupational exposure Industry/job ategories Attributable deaths (registrations for NMSC) Bladder Leukaemia Lung Mesothelioma NMSC Sinonasal Total* M F M F M F M F M F M F M F Exposures Constrution inluding: Ar, Asb, Ch, Co, DEE, ETS, Pb, PAH, R, Si, Sr, W, Ca, N, Fo, PAH, painting Roofers, glaziers, PAH road surfaers, onreters, roadman, paviours, kerb layers and their foremen Painters & deorators painting Metal workers MWF Personal and household servies Asb, Ca, Bz, DEE, ETS, PAH, R, Sr, Ch, Pb, Fo Mining (not metals) Asb, DEE, PAH, Si, Sr Land transport Asb, Bz, DEE, ETS, PAH, R, Sr, Ch, Pb, Si, W Wholesale and retail trade and restaurants and hotels Asb, Bz, DEE, ETS, PAH, R, Sr, Pb Printers and printing mahine minders and their foremen mineral oils + printing ink Printing, publishing and allied industries Farming, hortiulture, gardening, forestry and related Manufature of transport equipment Publi administration and defene (Armed Fores) Ca, Co, DEE, Pb, PAH, R, Sr, Ch, ETS, Ni, Si, W D, NAP, Sr, R, ETS Ar, Asb, Be, Ch, Co, DEE, N, PAH, R, Si, Sr, Ca, ETS, Pb, W R, ETS, PAH, DEE, Sr, Pb Servies allied to transport Bz, Co, DEE, ETS, Pb, PAH, R, Sr, Ca, Ch, N, Si, W Welders Welding fumes Finaning, insurane, real R, ETS, Sr estate and business servies Communiation DEE, ETS, R, Sr, Pb Manufature of fabriated metal produts, exept mahinery and equipment Be, Ca, Co, Ch, Fo, DEE, Pb, N, PAH, R, Si, Sr, ETS, W Sanitary and similar servies Ar, Asb, Bz, Co, DEE, ETS, PAH, R, Sr, Ca, Ch, Pb, Si, W Eletriity, gas and steam Ar, Asb, Be, Ch, Co, DEE, PAH, R, Si, Sr, Ca, ETS, Pb, N, W Manufature of mahinery exept eletrial Be, Ca, Ch, Co, DEE, PAH, R, Si, Fo, ETS, Pb, N, W Non-ferrous metal basi industries Ar, Bz, Ca, Ch, Co, DEE, Pb, N, PAH, R, Si, Sr, Fo, W Manufature of other hemial produts Coah and other spray painters and painting assembling and related oupations Rereational and ultural servies Manufature of industrial hemials Original artile Ar, Bz, EO, 1 3B, Asb, Ch, Co, DEE, Pb, R, Si, Fo, Ca, ETS, N, W Spray painting Ar, ETS, R, Sr Ar, AA, Asb, Bz, Fo, 1 3B, Ca, Ch, Co, DEE, Pb, PAH, R, Si, Be, ETS, N, W Table 3 gives for eah aner, numbers of deaths (registrations for NMSC) within industry setors or jobs for whih there were at least 50 estimated attributable aners; the exposures onerned are listed, with those ontributing most (at least 10 aners in men plus women) being shown in bold. *Totals are for lung, bladder, leukaemia, mesothelioma and nasal aners plus attributable registrations for NMSC. 0 =,0.5; blank ell = aner not represented. AA, aromati amine (bladder); Ar, arseni (lung); Asb, asbestos (lung, mesothelioma); Be, beryllium (lung); Bz, benzene (leukaemia); Ca, admium (lung); Ch, hromium IV (lung, sinonasal); Co, obalt (lung); D, dioxins (lung); DEE, diesel engine exhaust (lung, bladder); EO, ethylene oxide (leukaemia); ETS, environmental tobao smoke (lung); F, female; Fo, formaldehyde (sinonasal, leukaemia); M, male; MWF, metal working fluids (bladder, NMSC, sinonasal); N, nikel (lung, sinonasal); NAP, non-arsenial pestiide (leukaemia); NMSC, non-melanoma skin aner; PAH, polyyli aromati hydroarbon (lung, bladder); PAH, oal tar and pith (NMSC); Pb, lead (lung); R, radon (lung); Si, silia (lung); Sr, solar radiation (NMSC); W, wood dust (Sinonasal); 1 3B 1 3 butadiene (leukaemia). Oup Environ Med 2008;65: doi: /oem

10 Original artile estimates are similar to those used in the Global Burden of Disease projet. 66 There was a general lak of information on the lateny of the aners, partiularly in relation to speifi oupational exposures. The assumptions made in the study have influened the numbers ever exposed giving high estimates in some ases. In partiular a uniform distribution of aner indution between the maximum and minimum lateny was assumed, although reality may be a distribution that peaks in the early 1970s and tails off towards more reent periods. In ombining the AFs for different risk fators, multiple exposures and other non-oupational risk fators were onsidered. Caner is a multifatorial and multistage disease that may not be due to any single suffiient ause but rather a sequene of hits over a life ourse. For example, smoking alone may not be suffiient to ause lung aner and those who get it are likely to have been exposed to several lung arinogens and possess other harateristis suh as some form of inherited suseptibility. The mathematial impliation of this is that the sum of attributable frations for several exposures may be greater than 100%, with the amount exeeding 100% being partly due to synergisti interations among the risk fators. 74 We have avoided this problem of double ounting of the interations to some extent by partitioning exposed worker populations between overlapping arinogeni exposures before estimating AFs. In other ases where overlap remains we have assumed risks are multipliative, so that the ombined AF inorporates the interation between exposures. Many past exposures will have been at muh higher levels than those existing today. However, trends vary depending on the substane and soure of data. For example, analyses of exposure measurement data held in the National Exposure Database (NEDB) and from Health and Safety Exeutive (HSE) inspetion surveys and other surveys showed downward trends of 11% per year for toluene between 1985 and 2002 based on inspetion surveys but follow-up surveys of eight ompanies using toluene-ontaining ompounds show an average derease of only 1% per year in toluene onentrations. 75 Although respirable dust exposure in the quarry industry delined by 6% eah year from 1984 to 2003 there was no lear hange in exposure over time for respirable quartz exposure. Other exposures have all but disappeared due to the deline of the industry or the substitution of hazardous substanes by Table 4 Unertainties and limitations of the methodology and their potential impat on the estimate of the burden of disease due to oupation Potential impat on burden Soure of unertainty estimate Exlusion of IARC Group 2B and unknown arinogens Q Inappropriate hoie of soure study for risk estimate qq Impreision in soure risk estimate qq Soure risk estimate from study of highly exposed workers applied q to lower exposed target population Risk estimate biased down by healthy worker effet, exposure Q mislassifiation in both study and referene population Use of RR = 1 for very low/bakground/environmental levels of Q exposure where no value available from literature Inaurate risk-exposure period qq Unknown proportion exposed at different levels of exposure qq Effet of unmeasured onfounders qq Use of Levin s formula when RR adjusted for onfounders Q IARC, International Ageny for Researh on Caner; RR, relative risk. Main messages Overall, 4.9% (8% men, 1.5% women) of all aner deaths in Great Britain in 2004 were attributable to work-related arinogens (based on the assessment of six aners and International Ageny for Researh on Caner Group 1 and 2A arinogens with strong or suggestive human evidene). Asbestos ontributed over half the oupational attributable deaths, followed by silia, diesel engine exhaust, radon, work as a painter, mineral oils in metal workers and in the printing industry, environmental tobao smoke (non-smokers), work as a welder and dioxins. Oupational exposure to solar radiation, mineral oils and oal tars/pithes ontributed large numbers of skin aner registrations. Industries/oupations with large numbers of aner deaths and registrations inlude onstrution, metal working, mining, land transport, roofing and road repair/onstrution, printing, farming, some servie industry setors in partiular personal and household servies and wholesale and retail trades, restaurants and hotels and manufature of mahinery, transport equipment, non-ferrous metals and metal produts, and hemials. Poliy impliations Estimates for all but leukaemia are greater than those urrently used in UK health and safety strategy planning and ontrast with small numbers from oupational aidents. Carinogeni agents, oupations and industrial areas are highlighted for prioritisation of risk redution strategies. Past high exposures will ontinue to give substantial numbers in the near future and, although levels of many exposures have redued, reent measurements of others show ontinuing high levels whih must be addressed. other non-arinogeni agents. Other arinogens suh as naturally ourring radon ould also easily be eliminated from workplaes. However, the long lateny of some aners means that numbers of deaths and registrations due to past high exposures will ontinue to be substantial in the near future (partiularly asbestos-related aners). For some arinogeni agents exposures remain high. For example, reent wood dust measurements have shown ontinuing high exposures. 76 Although some hazards, suh as ertain solvents in paints, may have been removed from oupations with multiple exposures, the potential for exposure to other hazards remains, for example, silia exposure in the onstrution industry, in whih the number of employees is inreasing. Studies in the Duth onstrution industry suggested that over half of the full-shift respirable quartz dust measurements were above the Duth Oupational Exposure Limit with exposure being highly variable from day to day and between jobs and tasks. 77 In addition there will be onsiderable numbers of workers exposed at low levels and risk to some arinogens that may ontribute substantially to both high AFs and numbers. Future work will address estimation of the urrent burden due to oupational exposures for the remaining aners, the use 798 Oup Environ Med 2008;65: doi: /oem

11 of other measures suh as years of life lost and Disability- Adjusted Life Years, together with development of appropriate methodology for prediting future estimates of the oupational aners due to more reent exposures and for exploring the sensitivity of the estimates to soures of unertainty and bias. In summary, this projet is the first to quantify in detail the burden of aner due to oupation speifially for GB. An upto-date estimation of the urrent burden of six aners due to past oupational exposures has been arried out using a robust and transparent methodology. On balane the estimates are likely to be onservative estimates of the true burden. The results highlight speifi arinogeni agents and the oupational irumstanes and industrial areas where exposures to these agents our, and should failitate prioritisation of risk redution strategies. Aknowledgements: Funding was obtained from the HSE and managed through the Health and Safety Laboratory. We would like to thank Damien MElvenny for initiating the projet and for his role as HSE projet offier and Gareth Evans for his management role. Andy Darnton from the HSE provided the data for the work on mesothelioma. The ontributions to the projet and advie reeived from many other HSE and HSL staff is gratefully aknowledged. Two workshops were held during the projet bringing together experts from the UK and around the world. We would like to thank all those who partiipated and have ontinued to give advie and omment on the projet. Funding: Health and Safety Exeutive. Competing interests: None. REFERENCES 1. Lopez AD, Mathers CD, Ezzati M, et al. Global and regional burden of disease and risk fators, 2001: systemati analysis of population health data. Lanet 2006;367: Brown ML, Lipsomb J, Snyder C. The burden of illness of aner: eonomi ost and quality of life. Ann Rev Publi Health 2001;22: Danaeii G, Vander Hoorn S, Lopez AD, et al. Causes of aner in the world: omparative risk assessment of nine behavioural and environmental risk fators. Lanet 2005;366: Doll R, Peto R. The ause of aner. Oxford: Oxford University Press, Siemiatyki J, Rihardson L, Straif K, et al. Listing oupational arinogens. Environ Health Perspet 2004;112: Rousseau M-C, Straif K, Siemiatyki J. IARC Carinogen update. Environ Health Perspet 2005;113:A Straif K, Baan R, Grosse Y, et al. Carinogeniity of polyyli aromati hydroarbons. Lanet Onol 2005;6: Pannett B, Kauppinen T, Toikkanen J, et al. Oupational Exposure to Carinogens in Great Britain in : Preliminary Results. In: Carex: international information system on oupational exposure to arinogens. Helsinki: Finnish Institute of Oupational Health, Steenland K, Armstrong B. An overview of methods for alulating the burden of disease due to speifi risk fators. Epidemiology 2006;17: Levin M. The ourrene of lung aner in man. Ata Unio Internationalis Contra Canrum 1953;9: Miettinen O. Proportion of disease aused or prevented by a given exposure, trait or intervention. Am J Epidemiol 1974;99: Yates D, Corrin B, Stidolph P, et al. Malignant mesothelioma in south east England: liniopathologial experiene of 272 ases. Thorax 1997;52: Howel D, Arblaster L, Swinburne L, et al. Routes of asbestos exposure and the development of mesothelioma in an English region. Oup Environ Med 1997;54: Goldberg M, Imbernon E, Rolland P, et al. The Frenh national mesothelioma surveillane program. Oup Environ Med 2006;63: Spirtas R, Heineman EF, Bernstein L, et al. Malignant mesothelioma: attributable risk of asbestos exposure. Oup Environ Med 1994:51: Darnton AJ, MElvenny DM, Hodgson JT. Estimating the number of asbestosrelated lung aner deaths in Great Britain from 1980 to Ann Oup Hyg 2006;50: Sorahan T, Hamilton L, Wallae DM, et al. Oupational urothelial tumours: a regional ase-ontrol study. Br J Urol 1998;82: Lee-Feldstein A. Cumulative exposure to arseni and its relationship to respiratory aner among opper smelter employees. J Oup Environ Med 1986;28: Collins JJ, Lineker GA. A review and meta-analysis of formaldehyde exposure and leukaemia. Regul Toxiol Pharmaol 2004;40: Lewis RJ, Shnatter AR, Katz AM, et al. Updated mortality amongst diverse segments of a petroleum ompany. Oup Environ Med 2000;57: Bloemen LJN, Youk AO, Bradley TD, et al. Lymphohaematopoeiti aner risk amongst hemial workers exposed to benzene. Oup Environ Med 2004;61: Original artile 22. Ward E, Okun A, Ruder A, et al. A mortality study of workers at seven beryllium proessing plants. Am J Ind Med 1992;22: Verougstraete V, Lison D, Hotz P. Cadmium, lung and prostate aner: a systemati review of reent epidemiologial data. J Toxiol Environ Health Part B: Critial Reviews 2003;6: Cole P, Rodu B. Epidemiologi studies of hrome and aner mortality: a series of meta-analyses. Regul Toxiol Pharmaol 2005;43: Rosenman KD, Stanbury M. Risk of lung aner among former hromium smelter workers. Am J Ind Med 1996;29: Moulin JJ, Wild P, Romazini S, et al. Lung aner risk in hard-metal workers. Am J Epidemiol 1998;148: Lipsett M, Campleman S. Oupational exposure to diesel exhaust and lung aner: a meta-analysis. Am J Publi Health 1999;89: Coggon D, Pannett B, Aheson ED. Use of job-exposure matrix in an oupational analysis of lung and bladder aners on the basis of death ertifiates. J Natl Caner Inst 1984;72: Boffetta P, Silverman D. A meta-analysis of bladder aner and diesel exhaust exposure. Epidemiology 2001;12: Kogevinas M, Beher H, Benn T. Caner mortality in workers exposed to phenoxy herbiides, hlorophenols, and dioxins: an expanded and updated international ohort study. Am J Epidemiol 1997;145: Kheifets L, Afifi A, Buffler PA, et al. Oupational eletri and magneti field exposure and leukemia: a meta-analysis. J Oup Environ Med 1997;39: Zhong L, Goldberg M, Parent ME, et al. Exposure to environmental tobao smoke and the risk of lung aner: a meta-analysis. Lung Caner 2000;27: Coggon D, Harris EC, Poole J, et al. Extended follow-up of a ohort of British hemial workers exposed to formaldehyde. J Natl Caner Inst 2003;95: Teta MJ, Sielken RJJ, Valdez-Flores C. Ethylene oxide aner risk assessment based on epidemiologial data: appliation of revised regulatory guidelines. Risk Analysis 1999;19: Mannetje A, Kogevinas M, Lue D, et al. Sinonasal aner, oupation, and tobao smoking in European women and men. Am J Ind Med 1999;36: Coggon D, Harris EC, Poole J, et al. Extended follow-up of a ohort of British hemial workers exposed to formaldehyde. J Natl Caner Inst 2003;95: Czene K, Tiikkaja S, Hemminki K. Caner risks in hairdressers: assessment of arinogeniity of hair dyes and gels. Int J Caner 2003;105: Blettner M, Zeeb H, Auvinen A, et al. Mortality from aner and other auses among male airline okpit rew in Europe. Int J Caner 2003;106: Steenland K, Boffetta P. Lead and aner in humans:where are we now? Am J Ind Med 2000;38: Fu H, Demers PA, Costantini AS, et al. Caner mortality among shoe manufaturing workers: an analysis of two ohorts. Oup Environ Med 1996;53: Tolbert PE. Oils and aner. Caner Causes Control 1997;8: Eisen EA, Bardin J, Gore R, et al. Exposure-response models based on extended follow-up of a ohort mortality study in the automobile industry. Sand J Work Environ Health 2001;27: Roush GC, Meigs JA, Kelly JA, et al. Sinonasal aner and oupation: a aseontrol study. Am J Epidemiol 1980;111: Leon DA, Thomas P, Huthings SJ. Lung aner among newspaper printers exposed to ink mist: a study of trade union members in Manhester, England. Oup Environ Med 1994;51: Sorahan T, Williams SP. Mortality of workers at a nikel arbonyl refinery, Oup Environ Med 2005;62: Seilkop SK, Oller AR. Respiratory aner risks assoiated with low-level nikel exposure: an integrated assessment based on animal, epidemiologial, and mehanisti data. Regul Toxiol Pharm 2003;37: Grimsrud TK, Peto J. Persisting risk of nikel related lung aner and nasal aner among Clydah refiners. Oup Environ Med 2006;63: Aquavella JF, Olsen G, Cole P, et al. Caner among farmers: a meta-analysis. Ann Epidemiol 1998;8: Chen R, Seaton A. A meta-analysis of painting exposure and aner mortality. Caner Detet Prev 1998;22: Bosetti C, Pira E, La Vehia C. Bladder aner risk in painters: a review of the epidemiologial evidene, Caner Causes Control 2005;16: Armstrong B, Huthinson E, Unwin J, et al. Lung aner risk after exposure to polyyli aromati hydroarbons: a review and meta-analysis. Environ Health Perspet 2004;112: Unwin J, Coker J, Sobbie E, et al. An assessment of oupational exposure to polyyli aromati hydroarbons in the UK. Ann Oup Hyg 2006;50: Boffetta P, Jourenkova N, Gustavsson P. Caner risk from oupational and environmental exposure to polyyli aromati hydroarbons. Caner Causes Control 1997;8: Partanen T, Boffetta P. Caner risk in asphalt workers and roofers: review and metaanalysis of epidemiologi studies. Am J Ind Med 1994;26: NRPB. Health risks of radon. Chilton: National Radiologial Protetion Board, Kurihara N, Wada O. Siliosis and smoking strongly inrease lung aner risk in silia-exposed workers. Ind Health 2004;42: Steenland K, Mannetje A, Boffetta P, et al. Pooled exposure-response analyses and risk assessment for lung aner in 10 ohorts of silia-exposed workers: an IARC multientre study. Caner Causes Control 2001;12: Freedman DM, Dosemei M, MGlynn K. Sunlight and mortality from breast, ovarian, olon, prostate, and non-melanoma skin aner: a omposite death ertifiate based ase-ontrol study. Oup Environ Med 2002;59: Oup Environ Med 2008;65: doi: /oem