Document jgy7Gbvpay2dd68YYzLXL2Jo2

13 April 1987 EFFECTS OF EXPOSU1E TO VINYL CHLOEIDE: AN ASSESSMENT OF THE EVIDENCE Elchard Doll, FU Honorary Cooaultaat, Inparlal Caaear Raaaareh Food Caaear EpldamlolotT and Clinical Trlala Unit, ladcllffa Infirmary, Oxford. URL 13599 COWTPfTS SUMMIT ItfTKBOCriO* OCCUTATIONAl- IAZAKDS Sourest of lafoxwatloa US Study UK Study Canadian Study Italian Study Otter Soureaa Basarda of eanear Anfloaareesa Lunc caoeor Srals caoeor Cantors of hacnatopolotleand lynphatle tlaauao Kalaaoaa Thyroid eanear Caneara of digestion tract Otter caneara Kasards of non-aullgnaat dlaaaaa Clrrboala of 11m Vos-oallgoaot raoylratorydlaaaaa Cardloaaaenlar dlaaaaa Dlaeuaaioo HAZAJtDS TO IKK CEKZKAL NtSUHOI ' CONCLUSIOV iznuROi Pago No. 1 7 8 to 11 13 IS 17 IB 24 25 27 28 29 50 50 50 51 31 32 32 34 34 . 38 41 42 URL 13600 URL 13601 I SUMMARY The results are reported of a review of: (1) th possible offacet of vinyl chloride (VC) oa the personal health of mo exposed to It by virtue of choir occupation, othor thoa tho oorly affoett of tho vary high concentrations to which ana worn exposed whoa tho industry was first developed (uaconoclouoaosa, eardiac arrhythmias, and tho characteristic 'vinyl ehlorlda Illness*) and (11) tho carcinogenic offsets that alghc coneolvably bo obsorvod In tho gonoral population, aa a rosult of tho widespread distribution of VC as a pollutant* The possibility that VC night act aa a toratogon or night causa nutations In tho gore colls has not boon examined, aa tho little evidence that has boon adduced relating to such possible effeccs has boon reviewed elsewhere and the conclusion reached that no such affects have been denonstrsted* Many groups of workers exposed to VC la the nanufaceuro of vinyl chloride nooonor (VQO or polyvinyl chlorldn (FVC) hmve boon studied sines the carclaogsnle potential of VC was first roeognlsod* Tbs results have shown that occupational exposure can eneaa nngloaarcona of tho llvar and seen have suggested that It nay cause eoveral other types of cancer as well* Whether this la to or net can ha decided only by examining nil the evidence end. In particular, the eonblned results of those studios that lnelude n substantial proportion of observations on non nora than 25 years aftsr their first exposure and cover a long enough period for nora than 10 par cant of the employees to have boon expected to die* Tho results of four studios can bo usefully combined in this uey: namely, two national studies la tbs USA and the OK and two studies of employees in one pleat in Cnande end two plente in Itely* Tho results of othor studies In tha Federal tspnblie of Gomany. Koruny, Sweden, Italy, Francs, and Japan can be need only to provide supplementary Information* The easy earlier reporta of expoeed workers la the USA end the UK relate to 2 en covered core eoapleeely la the two reeent national studies and their results serve only no sources of hypotheses. Minor crltIdeas eon be cede of three of the four nose useful studies. They do not seriously effect the value of the results, except thet allowance hes to be eede for the fellure to detsraine the eeues of 6.3X of the deaths recorded in the OS study. Thrss of the studies use national rotes to estimate the makers of deaths thet sight have been expected to occur in the absence of any special occupational hasard and the fourth (Canadian) study uses rates for the province In whleh the plane vas situated. It oust, therefore, he kept in nlnd ehet the rates used aey not here been wholly appropriate for the localities in whleh the planes were situated. This is potentially isportent for the OS study which covered workers in 17 plants, 22 of whleh woro situated In the south of tho country. Tho other lees informative studies ore, for tho uoet part, open to moro serious erltldsu and tho valuo of each sot of rosules ooods to bo asaassod ooparatoly In relation to oeeh diseneo. Tho eoublnod results of tho four principal studios shew ehet the standardised nortallty ratio (SMK), reflecting the ratio between the nlabors of deaths cfcservsd and those expected in the absence of en occupational hessrd nultlplled by 100, bee been: (1) 77 far accidents and other violence, (11) M fur dleeeeee other then cancer, end (ill) 102 for cancers ether then cancer of the liver. All these results ere whet night hi anticipated for en-industry devoid of any specific occupational heserd. The low ratio for diseases other then cancer reflects the 'healthy worker effect', which results frou the selection process thet inevitably excludes seas ef the less healthy sabers of the population (row industrial euploynaut end is eanpetlhle with n higher ratio for cancer as tho nortallty free enneor In not normally subject to ouch an effect, apart from the first two or three years lunedlately following tho start of euployoent. URL 1360: r-o EOStVltifi 3 Th* mortality from cancer of cbe liver we* nearly Hvin tine* ehac expeetad. Host of th 51 icm deaths r known to be duo to angioeareoma, wen though this diagnosis u not recorded on th* death certificate. Th* excess corresponds almost exactly with th* 49 deaths du to u|loMrtM riporttd to th* International Beglster of A&glotarcowa Cases os occurring la employees of the planes concerned during the periods they wore kept under observation. All the men who developed the disease ere likely to hove been exposed to concentrations of VC of several hundred ppm or more* Three other types of cancer, which. It ues suggested soon after the discovery of the carcinogenicity of VC, night he ceueod by VC, ero cancers of tho lung, brain, end lymphatic and haematopoietic systems* The combined data for the mortality from respiratory cancer fail to support the hypothesis regarding lung cancer (gMt 97). tlghar ratios for lung cancer have, however, boon observed consistently in the subgroups of non la whom tho effect of an occupational hasard mould ho moat likely to he seen (that la, man employed for more them 10 years, exposed to higher ehen average concentrations, or observed more them 20 years after first exposure) In two of the supplementary studies it mao also noted thet the mortality from lung cancer mas specifically lucreseed in the most heavily exposed. The combined data ahem smell exeeeeee lx the mortality from concert of the brain and of tho lymphatic end haematopoietic systems; hat the excesses ere - not statistically significant and there la nothing to suggest that they ste occupational la origin* An exception la the observation of an Increased mortality from eaocers of the lymphatic and haematopoietic systems In tho supplementary study from the federal Kepublie of Germany* Two 'types of cancer mare reported to be In excess in the Norwegian studyt namely, thyroid cancer and melanoma* The significance of this finding is difficult to assess because very little information about these URL136M 4 cancers 1* provided by othar studies. SttUMtlou that VC might UUM UMiri of th digestive traet hri failed to tiki account of tho contribution of angiosarcoma of tha liver. When thin disease la excluded, tha mortality from digestive traet eanear la found to hart boon below mni <M for tha four principal atudlas, 82). A email excess mortality from tha hotarogonaoua group of 'other canears' in tha eomblaod raoulto of tha four principal otudiaa la itatlatleally marginally algnlfleant (83 death* against 85.25 expected, p<0.05) torn* of tha axcaaa la likely to hero boon doe to tha alsclassifleatlon of angiosarcoma* aa secondary cancers of tho Uvsr or carelnoastools, alts unknown. Throe nos-aallgnaoc causes of death hare required epeclal examin ation: cirrhosis of tha Hoar baceuae of tha damage to tha User that occurred aa part af tha 'vinyl ehlorlde Illness', non-mallgnant respiratory dlaoaao Wcaeaa of tho changes in lung function and x-ray appaaranea observed in men axpoeod to JVC dust, and myocardial infaretlon from analogy with tho effact of other haloganatad hydrocarbona and bacanao of soms obaarvatlona on Swedlah FTC fabrleatora. Tar from being raised, tha mortality from elrrhoala of tho liver wae leaa than expected la tho thrno principal etodlos and In one of the two aupplementary atndleo which gave aeperate figures for the disease (SHk* of 89 baaed on 48 daaths and of 12 baaed an IS deaths) while in the other \ supplementary atndy the Increase waa trivial. Data for ayocardlal Infarction have not been reported separately, but nyocardial infarction MeMata for moot of the dccthc attributed to lachcemlc heart disease and then la no evidence that either leehaeale heart dleeaae or eardtevanemlar dlaoaao ao a whole waa unduly common (SMLs of 80 emd 92) on minted to occupational expoaure. URL 13605 5 The data for eha third category of non-uallgnant disease (nonuallgnant respiratory dlaaaaa) ara confusing, bacauaa tha two large national studies giro conflicting results. Tha combined data for eha four principal atudlaa show tha low mortality that la commonly found In healthy lnduatrlal populatlona (SMI SO)* This, however, hldaa an lneraaaad mortality from chronic obstructive lung dlaaaaa (SMS 120) which includes emphysema and la due to a groaaly lneraaaad mortality attributed eo anphyaana In eha US etudy (SMI 193). Tha corraapondlng oortallcy in the UK study, which waa preferentially deacrlbed aa bronchitis, la laas than axpactad (SMI S2) aa la tha mortality from pneumonia (SMI SO) and other reaplratory dlaaaaaa (SKI 4$) in tha US study* Thara la no conalataot evidence that tha mortality frou emphysema or bronchitis was spaclfleally occupational and it seems possible that tha reported excess la tha US study Is aa artefact due to nosological diffleultlte In the use of tha 7th revision of the International Classification of Diseases* levlew of tha nasties data oow available relating to tha long tarn health of non oceuaptloaally exposed to VC leads to two clssr conclusions. Flrse ths nan have experienced a specific haserd of tha normally extremely rare angiosarcoma of tbo liver* Approximately ! In 335 of tha nan exposed In tha 49 plants has died of tha dlaaaaa and approximately 21 of observed deaths have bean attributed to it* Za tha course of tine tho tiabtri of cases of angloeareoma oust bo expected to be lneraaaad 2 to 3 tines* Secondly, the mortality from all other causes baa boon typical of that In normally haalthy industrial workara* If say other hassrd has exlstsd its offset has baea Mall* Ths data provide me .reason to think ehmt say hazard other than one of cancer has beam overlooked* It la, however, atill difficult to decide whether VC does produce risks of cancer la elees other than the liver that ara small compared to those due to non-occupational causes, but which URL 13606 6 sight, to total, cause alsost as many deaths as anglosarcosa of the livsr. Thera is too little evidence either to confirm or refute the suggestion that VC sight cause melanoma or cancers of the thyroid, brain, and lymphatic and haematopoietic systems. Hone of the small excesses that have been recorded point specifically to am occupational hatard, apart froa that attrlbutabla to canears of the lymphatic and haematopoietic systems In the German ctudy, and most are likely to he the sort of chance effeet that le certain to be observed when many types of canter ere examined in many different' studies* The leek of eoy Increased mortellty from lung ceneer In the combined results of the four principal studies does net exclude the poselblllty that there may have been a email occupational hatard of developing the disease, as geographical variations la the incidence of the disease throw doubt* on the validity nf using netlonnl rates for estimating the expected numbers of deaths* Tha greater mortality la groups of workers who would be more likely to show aa occupational hatard than la other groups, suggests that a small hasard may havs existed* The evidence la, however, weak and the existence of e hatard ia not proven* Clearer answers to some of the questions that have bean posed in this review might be obtained If tha various groups of Investigators could present their results ie more appropriate and comparable ways* As VC la a mutagen in laboratory axparlmeate and a proven htaen carcinogen, ehe minuet doses that have oeeapad into the general anvironment as pollutants must be presumed to have caused comparably mlnuta risks to the general public* Vo such risk could possibly bo detoctod, other than one of angiosarcoma of the liver vhleh le normally an extremely rare disease* Several surveys have sought evidence of tho existence of such an effect, and suggestive evidence that such aa affact may have occurred at a time when anvlromaantal pollution was much greater than It la now has bsea found in one URL 13607 7 imoDGcno* The Inhalation of large amounts of vinyl chloride (VC) hot boon raeognlsed to bo potentially hazardous for nony years* Concentrations of tbo order of 10,000 ppm in nir Induce unconsciousness and cardiac arrhythmia, while prolonged exposure to coneoneratlona an ordor of magnitude lower have boon liable to eaoae a specific pathological syndrome. This "vinyl chloride lllnaaa" haa baao eharaetarlsad by four cardinal algna: neatly, enlargement of the liver and spleen with a apoclflc hiatologleal appearance, patchy infiltration of the akin roaanbllng scleroderma, bony change* in the tip* of the fingers described as acro-osteolysls, and peripheral circulatory ebangee identical with the clasaleal picture of Raynaud's disease. These pathological reactions nay occur singly or together and nay possibly be accoapanied by other leas characteristic effects* They can, however, be completely avoided by ensuring that exposure never exceeds a few hundred ppa, to which level expoeures were generally reduced in the aid 1960s. One other serious effect has, however, been observed that aey not be avoidable in the seat relatively easy weyt naaely, the production of angiosarcoma of the liver* Zt Boat, Indaed, he presided that aoae risk of developing the disease will persist from sxposuro to dosaa that ara avan lower than the current industrial levels of 5 ppa or less, ss VC ha* bass shown to set ss a mutagen (International Agency for Research on Cancer, 1979) end it cannot be aeeuaed that a threshold exist* below which no carcinogenic risk persists* Moreover, the possibility has to be considered that VC may cause some cancers other than angiosarcoma of tha llvar, partly becsoss laboratory studios have shown that It does this In animal exparimaats and partly bacauae tba initial atudlaa, which demonstrated tha production of angiosarcoma of the liver in Man, ware Inadequate is slsa to exelods a material Increase la the risk of cancer la common sites, such ss URL 13608 8 the lung and large bowel. Since no threshold doss eso be postulated It also follows that sows cancers nap have basn produced la the general public by the snail amount* that base escaped Into the general environment. Consideration also needs to be glean to the possibility that exposure to vinyl chloride acting over a long period nay have noxious effects on Man chat eould not easily be seen in animal experiments (by, for example, producing ehroale respiratory disease) and, as It is a autageu, there Is also a possibility that It nay act as a teratogen and eause congenital malformations in the offspring. In this reelev I bare not examined the possibility that VC sets as a teratogen nor that it causes mutations In germ eells, as there Is coo llttla serious evidence to require it* Reviews carried out for sections of ~ the Industry by Downs at al* (1977) aad by HaeMahoa (1977) concluded that tha few reports of positive affects eould not be substantiated aad no additional evidence was found in a similar later review by Barr (1981), apart from a report that embryos ware absorbed aad akalatal ossification produced ehsa pregnant rata ware exposed to doses appreciably less than those that had bean used by other workers wlthoet any such affects being observed. The report of absorbed embryos (Hlrkova at al.. 1978) eould not be evaluated thoroughly, however, as the experiment was inadequately described* l have, therefore, examined only thn possible offsets on tho personal health of mas oeeupatleaally exposed to VC other than on their reproductive capacity, and the carcinogenic affects that eight conceivably be observed in tha goners! population na a result of the widespread distribution of VC na e pollutant* OCCUPATIONAL HAZARDS Many studies of workoro exposed to VC In tho manufacture of vinyl chloride monomer (VQ0 oad polyvinyl chlorldo (PYC) have boon undertaken 9 tinea It wa first found that VC could causa csncar la sniaals (Viola at al, 1971; Maltonl at al., 1974) and in Nan (CraacK and Johnson, 1974). Thaaa hava confirms* chat sxpopurs cauaaa a hasard of angloaareoau of eha llvar and, in ssvaral inataneaa, hava shown sxcass lacldaaca or aortslity rataa froa othar dlaaaaaa, which vara convantlonally statistically significant* Conventional taata of ataelatleal altnlflcanea art, hovavar, daslgnad to halp anawar singla qusations daflnad haforahaad and ssvaral finding* that night ba axpactad to occur by ehaoca aIona onca In (aay) 20 tints must ba axpactad to occur if dosana of rataa ara axaalnad in aach of aavaral aata of indapandant data* Wa hava, tbarafora, baan faead with tha problan of dacldlng whathar tha axeaaa rataa that hava baan obaarwad in individual aeudlaa ara dua to occupational hasarda or to tha vsgariss of chanca Thla problan can ba solvad in part by axaalnlng tba overall raaulta, whan tha data fron conparabla atodlaa ara auaaads that ia, by adding togathar tha numbara of daatha obsarvad and axpactad in aach study, and conparing tha totals* This doss not eaqulrs tba assiaption that tha axpoauraa haws baan tha aaaa In aach study, any aora than It doas for aach individual whan tba raaulta of aach study ara cenaldarad slant* What it doaa raqulrs la that aaeh axpoaad population should bava baan obsarvad evor a parlod whan its nanbora wars at risk of dovaloplng dloaaaa (if a ganulna haaard axiatad) and that in aach study tha control population fron which tba axpactad nunbara of daatha wars -dsrlvsd waa approprlats (that la, at tha sons risk of dovaloplng dlsaaso as tba axpoaad population would bava baan la tha abaanea of oxpoturo)* This doos not lotroduea aay now eoaplsxlty, as both thaaa coadltloaa ara, of conraa, also raqulrsd if corract eonelualona ara to bo drawn fron tba raaulta of tha individual studios Ann tbay ara axaalnad on thalr own* URL 13609 URL 13610 10 Soutcsa of Information four studies bmc these requirementsi neatly, the Co Urge national lumyi reported by Jones (1996) for the OK and by Environmental Health Associates (1966) for tbo USA, ttd studies of individual plant* la Canada raportod by Theriault and Allard (1961) and la lealy reported (aa a part of a national study) by 6*111 et al, (1966)* All four laclud* observation* on men aora than 25 years after their first exposure aad the expected number of deaths lst in eseh case, $reater chan 10 per cent of the eotal nuaber of eaployees Indicating a long average period at risk. Earlier observations oo UK and USA eaployees reported by Monson at al. (1979), Tabershaw and Gaffey (1974), Duck at al. (1975), Wlcholson at al. (1975), Ott at *1. (1975), tfaxveiler et al. (1976), Fox aad Colllar (1977), Equitable tnvlronasncal Health (1976), Huffier at al. (1979), and Cooper (1982) are subsuaad in the national surveys and now serve only as sources of hypotheses and to provide anas detailed information not included la the national reports. Studies of German workers (Gralser at al., 1962), Horveglan workers (Heldaas tl, 1984), Swedish workers (Syren at al.. 1976), French worker* (Flarra at al.. 1979 and Laplancha et al., 1987), Japanese workers (Masoda, 1979 aad Kakaaura, 1981) and some other Italian workers (Belli et al. 1986) provide sons sopplsasntary information bat, in general, tba periods ef observation have net been long enough for useful epldealologleel date to be obtained ebeut diseasee that are unlikely to occur within 20 years of first exposure, or they report only eelected results uhleh ere difficult to interpret aa only excess rotes tend to hove been selected. Studies ef FVC fabricator* are not included ee the workers will hove hod such lee* exposure to WC than those eaployed in the asaufecture of 9CM or FVC and any occupational hassrd to which they may have been exposed la aers likely to hove been produced by FVC dost. URL 13611 11 us study Tha study earrlsd out by Environmental Health Associates (1986) on behalf of eho US Chenleal Manufacturer* Association ts ths tersest snd nose Informative. It covers 10,173 men who had worked In 37 plants owned by 17 companies - 1,214 In 11 planes that produced only VCM, 6,848 in 18 plants that produesd only FVC, 835 In 3 plants that produesd both, and 1,176 In 5 plants that produead homopolymers and eopolynsrs, with or without VCM or PVC. Twsnty-two of ths plants wars In tha Souehsro part of tha country, 24 in tha North aastarn or North central parts, and ona la ths Vast. Man ware Included if they had bean exposed to VC for at least s year bsfors 31 December 1972 end had bean employed la 1962 or subsequently (the flrse yssr depending on tho dato tha plant began Baking or using VC and the earliest dats that personae! rscorda were ds--sd to be conplete, whichever was tha latsr) Individuals who nat tboss criteria wars Identified from conpeay records by company psreonasl. Racial charsctsrlstlcs were known only for 866 wen, 97 por coot of whoa wars white sad It wss prssuBed, for ths purpose of estlnitlng ths aunbor of expected deaths, that all 10,173 m wars whits. Follow-up was obtained fron pleat and Social focurlty Adnlnlstratlon Records aad (for aaa who died after 1979) fron the National Death Index. Five plants did not collaborate in the extension of Cooper's (1961) earlier study and tha 955 employees in these plants who wars known to be alive on 31 December 1972 were not followed any further. For the rest, follow-up wss attesptod to death or 31 December 1962, whichever was the earlier. On this basis 92.72 of the wen ware sueesssfelly traced. Those who were untraeed wore excluded frow the lest date of contact, which was usually ths data when snplsynset ceased. Alneet half the wan (682) wars first swployod before 1955. A large 12 proportion wee,' therefore, observed nor* than 25 poor* after first exposure (and la many casts for aort Chao 30 yeara) whan dlsaasts with s long latsnt parlod sight bo expected eo bo seen. Short torn workers had boos excluded free tho cohort and oote of the men had continued la employment for meny years: two-third* wort employed for 10 years or aero sad tho average duration of employment was 16 years. Fifteen hundred end thirty sin non were found to have died. In 1,659 cases, the cause of death was obtained from the deaeh certificate, but no eauee was obtained for 97 (6.3X) Tho ntabors of deaths expected from each of 39 causes or groups of eauaas were obtained by nuleiplylng the person-years at risk by tha disease-specific national ratao for ahlta aalea, for tho corresponding age groups end quinquennia. This laportant study it opaa to throo minor criticisms, which ere unlikaly to have had any material effects on tha raeulta. First, the Hate of employees were compiled by company personnel from company records without any Independent check. Secondly, tha aseiaptloa that all the employees wore white will have caused the expected deaths to have been very slightly underestimated, as the few black employees are likely to have had higher mortality rates and there la mo reason for supposing that tha small sample, from which tha proportlom of block omployoea was estimated, wee necessarily representative Thirdly, an element of uncertainty la introduced by the failure to troce as many as 7.3X of cmployoeo. Two othor criticisms are more important. First, tha expected numbers of deaths have been calculated on tha assumption that tha man would havo experienced the same mortality ratao aa tha white mala population of tho whole country ct the corresponding dates. Tha use of national rates la cannon practice in studies of industrial populations and tends to roault in aa over-estimation of tho expected euebero of doatho, so that the employees URL 1361 r-o 13 Appear to ba unusually healthy. Thla "healthy work.tr affect" la wtll known and will be taken into account in the discussion of eha rtaulea. A sort eerloue objection to the uae of national tatee la the way nortallcy varies froa one part of the country to another, due to dlffarencea in the prevalence of environmental and eoclal feetora unrelated to the occupation of internet. It la, therefore, generally preferable to uae State (If not eounty) ratea in plaee of national rates* With 37 plante, however, It eight be thought that their geographical dlatrlbution would be efficiently wide to nake the uae of national ratea appropriate. Unfortunately 22 of the plants wera loeatad In the South end e check would have been desirable to see whether this could have eaueed any material dletoratlon of tha rteulee. Secondly, cause were not obtalnod for 97 of the 1,536 deaths. This wea allowed for in the calculation of tho overall mortality by including deaths duo to unknown eauoea. tt wsa not allowed for, however, in the calculation of tha disease specific mortality rates and will have ceased the etenderdited mortality ration (Stats) to be undereetlmated by, on average, 6.3Z. For tho preeent purpoee, therefore, the numbers of deaths attributed to epoclfle dleoesos have each been multiplied by 1.067* (100/(1-97/1336)) and rounded off to the nearest integer. OK Study The study reported by Jones (1966) on behalf of tha UK Bealtb and Safety Executive covers 5,691 moo who warn employed for et least one year . in John that involved potential exposure to TCK for at least 251 of tho working week and who wera flrat employed in the period 1960*74. Details of the men were compiled from tha personnel reeerda of 9 chemical plants manufacturing or polymerising VC end the vital etatea of tha man were determined et the end of 19S* from the records of the Vatlonal Bealth Service Central kegleeer. Five thousand four hundred end ninety tight man URL 13613 H9enan u ware traced <98.91). Seven hundred and eighty daaeha war* identified and coplaa of cha death cartlflcataa were aanc to tha Investigators coded to the 8th ICD revision of tha Hat of eauaea of death If they occurred before 1979 and to tha tb mlaloa if they occurred later. Several specific points about tha atudy need to be noted. Flrat, national mortality raeaa for England and Velas ware calculated by five year age groups for quinquennial periods for 66 causes of death end these vers used to estimate tha nuabere of deatha that night have bean expeeted In the cohort by multiplying then by the corresponding nuabere of parson years under observation. Sous difficulty was, however, experienced In obtaining suitable rates for all causes of death, which could bo related to the causes of death eodad according to the 8th end 9th revisions of tha ICD Hat, and races for s relatively lata period had to be used for aatinatlng tha nuabere of deaths froa many diseases that night have bean expected to occur in earlier periods. For 2 categories the earliest available rates were 1960-64, for 1 they wore 1965-69, for 24 they were 1970-74, and for 1 they were 1975-79. Secondly, an attanpt was made to classify nan according to wbathar they had high. Intermediate, or lew exposure to VOt or JVC dust end each nan's employment hlstnry was raeorded with advice from tha planes concerned, under 12 Job titles. The nee were thee grouped according to whether exposure to VCM wee likely to have been high (group A), exposure to PVC duet wee likely to have baoe high with exposure to VCM low (group 1), or exposure te VCM end PVC duet wee lnteraedlate sad intermittent (gToup O. All other nan, who would gooerslly hsvu hod low exposure to both VCM and PVC duet, wore classed as group 9. tflthla oil groups, exposure to VOt wee likely to have been higher If It had bagun before 1956. The study nekae me important contribution to tho sun of knowledge concerning tha long tarn affects of VC. Tho mao of naelonal rates to 15 calculate tbe expeeted atabtri of doth* may Vo Justified on the grounds chat the & wtra employed is 9 plant*, which wort presumably distributed about cha country, but no details of Choir location art given. In general, mortality rataa toad to bo higher to cha pert* of brltaln where heavy industry la locaeod than la othor parts, so that cha expeeted aunbara of deaths aro more likely to bo biased downwards than upwards; but this aatdt to ba ahown. The use of recant rates to calculate expected nuabera of deaths fron many specific causes of death was presumably necessary, If the diseases ware to ba studied individually and will have done no ham If the incidence and fatality of the diseases In question have remained stable. It would have bean desirable, however, for the diseases to have been specified so chat the reader would know which were liable to bo distorted. The system used to classify men into four exposure groups Is the sore of system that is comonly used If precise measures of exposure are not available. It ereetes some dlfflcnltlet In tbe statistical analysis if man move from one Job to another and are classed (ea In thia inatanca) aa having had high exposure If they have ever had s particular typa of employment (a.g. over been employed aa am autoclave worker). Vo evidence La provided to show that the person- years at risk before a man entered the category have been ouberaeeed and added to another exposure group before tbe embers of expected deaths are calculated. Movements from one Job to another aro aald to have tended to bo dot of troops A and 1 Into D, but even so it most be presumed that the expected ambers of deaths la the first two exposure categories aro likely to be over-estimated. Canadian Study The Canadian study (Theriault and Allard, 1961) was limited to employees of m single plant la Shswlalgsm, Quebec. The plant, which was situated In aa industrial complex, was opened la 1943. VCM and PTC were URL 1361 URL 13616 16 both asde until -eh* late 1960a, when the production of VQ* cessed, vhllo the production of PVC contlnuod. As attempt wa made to eraea all production worker* who*# nano* appeared on eha unions' llata or the coapwlM' payroll* of the whole induatrlal complex Includlnf the VC plant between 1 January 1968 and 91 December 1972, and contact wa* wade with th worker or hi* next of Ida la 1,611 out of 1,699 lnatancea (97.IX). Detailed occupational and eaoklng hlotorle* were obtained by questionnaire and 156 aen who had been employed by the eonpanlea for leae than 5 year* were excluded* The remaining aen were categorised at (1) exposed to VCM If ehey had worked on the production of VCM or FVC for at least 5 years (451 an), (11) unexpoeed to VCM if they had worked similarly for leea than 6 months (670 aon), and (111) other aen (134 In total)* the last were excluded froa the study* Follow-up wa* dosed oa 31 December 1977* Copies of the death certificates were obtained and the cause* eoded according to the 8th revision of the ICD Hat for all who had died (99 exposed and 233 unexpoeed)* Information was also sought for histological or eytologlcal confirmation of all the diagnoses for all the exposed aon who had died of cancer* The raaulta ware examined in two ways* First, the mortalities of the exposed and unexpoeed ware compared after standardisation for qulnquetwliai and 5 year age group* feeoodly, the aortallty of the exposed aea wee eoapared with that expected If the aen had experienced the sex end ege specific aortallty rata* recorded la Quebec for the year 1971* la both coaparlaona the causae of death used were those #p titled on the deeth certificate and the additional pathological information ana used later only for Interpretation of the raaulta* Moat of the exposed man were exposed for more than 10 years (79Z), the average length of exposure was approximately 17 years, sad 44Z were observed aore them 29 yeere after first exposure* 17 Although sull, tho study uku a useful contribution to ehe overall results* Tha histological revise of tha cancer casta la particularly helpful. This shoved that all 8 cancers diagnosed as liver cancer (Including 2 spaclflad aa hepatoma end ooa apaciflad aa aaglosarcoaa) vara angiosarcomas of tha liver as veil aa oaa that had beta diagnosed as angiosarcoma of tha perltoneta. Two other cases of aaglosareoaa of tha liver vara found to hava beta cartlftad as hepatic cirrhosis. It Is slso helpful to hava a comparison between exposed and "unexposed" employees of tha saaa coapanlss as this shove that tha lev mortality observed for all aon-mallgnant diseases could be attributed to a healthy worker affect and vaa sot due to blee la tha recording of exposure (relative risk for all non-nallgnant eauaaa compared to "unexpossd" nan 0.95). One criticism has to ha modes namely, tha use of provincial races for oaa year (1971) to calculate expected mortality spread over a 30 year period (1948 to 1977 Inclusive). Deaths will hava tended Co bunch up towards tha and of tha period of observation so that tha rates for this particular year may have bean fairly representative, but It must havt caused some distortion of tha expected numbers ef deaths, tha sits (sad svaa tha direction) of vhleh la lapeeeible to estimate. Tor most disease groups tha distorcloa la eallkely to have bean large. Ztellea Study A study of all men employed la the production ef VC and FVC la 9 Italian plants was begun In 1983* All nan uere Included who wars enployad for at least 8 months nt any time from tho start np of tha plant to tha and of 1981* Tha study Is still inesmplsta, but results are now available for nan in throe plants (Sslll at al.. 1988). Two plants (In Psrrsrs and goelgoana) began operatlona in 1933. Vonr hundred aod thirty seven nan ware employed la tha oaa plant and 191 la the other. All but three (from the Ferrara plant) were followed to tho end of 1984* expected deaths wars URL 13617 URL 13618 is td--tad by multiplying tha peraon-yaars at risk by the corresponding national Mortality rata* for each 5 year ago group and qulaquenaiua. total expeeted daatha Id oaeb eaaa aaountad to* ora chan 101 of tho employees (12.41 and 12.81), Clinical Information wae aought about tho nauao of daath of all tha 55 employees of eba Farrars plant who had dlod. larlaod diagnoses, which warn not uaad for comparison with tha expeeted daatha, revealed four daatha fron caocara of tha Hoar In plaea of ona. Tha Savanna plant did not bagln oparatlona until 1959. Sis hundrad and thirty night nan warn anployad. All but 4 warn traead to tha and of 1983 and 17 warn found to have dlad. do nan could hav boon followed for ora than 24 ynaro and only 25.1 daatha (3.91 of tho vorkforca) wara axpactad. Tha data for this plant hava not, tbarafora, boon uaad In tha principal aoalyans. It any bn notad, howovar, that ona doath ona attrlbutad to llvor eancar whan 0.1 ona axpactad. Oehar Sources Tha Norwegian atudy (Baldaaa at el., 1984) provldad obaarvatlona on 454 nan who had boan anployad In a plant In Telemark ohara VCM had hoan manufactured fron 1950 to 1971 and PTC fron 1950 to tha and of tha atudy parlod. All vara incladod ohooa names wara rncardad on tha Company's paraonnal raglatar and health dopartaaat racorda oho had avar haoa anployad ft-- tha start of production to tho and of 1989 and had workad for at laaat oaa yaar. Maa wara folloond from 1953 to 1979 lncloalva. Daatha and eaaaa of eancar wara identified fron tha racorda of tha Control Duraau of ftatlatlea and tha national eancar registry, lb rafaranea la nods to any an haing lost to follow-up, bat It can ba use--ad that tha a--bar (If not taro) waa --all aa all cltisa-- have -- Idantlty nuabar which la uaad by both aaployara tad tha control aganelaa. Fifty --n wara fonad to hava dlad agalaat 59.34 axpactad if the sex- age- and quinquennium-specific national URL 13619 19 mortality rates had oparatad. Twenty-one nan vara found to have developed 23 caneers against 20>16 cancers expected from tha eonparabla national incldanc* rates, tha naa of which was justified by tha finding that tha Incidence In tha county In which tha plant waa altuatad waa between 90 and 95 per cant of tha rata In tha country aa a whole* One man who had been employed on PVC production developed angtoaarcoma of tha liver. Observed end expected numbers of cases ere given for cancers of the lung, colon, and thyroid, for melanomas, and for all eaaeara, but no axpectad numbers ere given for other types of cancer. It ia evident that aaveral other types of eancar must havt bean in deficit, aa there were I casee in all agalnat 14.93 expected, and It la difficult to know what weight to give tha axeaaaaa observed for tho reported type# of cancer aa they seam likely to havo bean reported specifically baeaosa the numbers wart In excess of those expected. The authors ants that one further case of melanoma had occurred after tha closure of tha study and that one "incipient ease" waa alee known to them. Tha Carman study (Cralaar et al., 1982) Included three groups: (1) 7,021 man who had bean exposed to TC In the coorso of their employment in any of tha 11 plants la which TC and PTC had been produced In the Federal % hepublic of Germany, (11) 4,120 man who had been employed la 7 chemical plants withowt hawing had any exposure to TC, and (ill) 4,007 man employed In 2 other plants where PTC was processed* Employees ware Included only If they were of German er Austrian nationality and were regarded aa exposed to TC if they were production wnrkara or other skilled workers or labourers assigned regularly to tbs plants, hot not if they vara employed In than only occasionally. All mao ware Included from tha time of opening of tha plants to tha end of 1974 sad were followed to the and of 1974. Many of the maa ware, in consequence, observed for only a few years after first employment, and only 142, 36Z, and 19X respectively of tha three groups URL 13620 20 wri first employed bfow 1934 and so capable of contributing person-yaart at risk sore than 20 years after first employment t whan an occupational haaard of cancer could be expected'to be observed* 93.21 of the exposed group were successfully followed up and causes of death were discovered for 92*82 of the 414 wen discovered to have died. The proportions for the ocher two groups wars respectively 89.82 and 88.72 for the uaexpoaed end 92*12 end 86*92 for the FVC proeese workers* The failure to obteln causes of death for all the nen who had died was allowed for in the subsequent analysis by weighting the numbers ettrlbueed to each cause by a system whleh took account of the age group end calender period in which death with an unknown cause occurred* The expected makers of deaths froa each eauee was calculated by multiplying the person-years at risk by the sex, age, and cause specific eortality retee for the nc* National data befora 1968 uead an ldioayscratle classification systes and the 1968 rates had to be need to nultlply ell the person-years at risk up to the and of 1966* For subsequent peers (1969 to 1974) tbo person years at risk were nultlplltd by the correepoodlng rates dor the sane calendar pent. Epidemiological studies are note difficult to earrp out In Germany than In N. iaerlea, the UK* or Seandlnavla because tbs medical cause of death la not recorded publlelp end there Is no central spates which ean be used for checking whether an indvlldual Is alive or deed* to the circueseamees* the Cernaa authors have made valiant efforts to obtain reliable date end the proportions of son in the exposed groups who were not euceesefullp followed up (6*82) and the proportions of deaths for whleh tho cause wee not obtained (7*22) were alnllar to those in the Environmental Vealth Associates (1986) study* Two defects* however, sake the date lees useful* Plrat* ee national mortality rates were available before 1968 and the use of the 1968 retee to 21 estimate eh* numbers of death* In and before 1968 will have overestimated cha numbers attributable to diseases that vara boeomlng nor* prevalent or diagnosed nor* often and will have ondaraatlnatad thoaa dua to dlaaasaa chat wara becoming lee* prevalent. Secondly, and nor* Importantly, a large proportion of tha nan had been first enployed lee* than 10 year* before the follow up was ended* It follow* that cha observation* on tha few non who had bean first exposed long enough before to have had nuch chance of developing an occupational dieoaea with a long latent period u*t have been tvaaped by observations that had llttla to eontrlbuta. Tha expeeted death* anouneed to only 6.21 of tha exposed nan and there is, therefore, little to be gained, and something to be lost, by including the German data in tha ovarviav. It aay bn noted, however, that 12 deaths vara attributed to eanear of tha llvar anong the VC exposed workers against 0*9 axpectad and that snallar excesses ware also observed la tha uaaxposad chemical workers (4 observed against 1*1 axpectad) and is tha FVC process workers (2 deaths against 0*8 axpectad). Two Svtdlsh pleats hava produced 9QI sad PTC alaaa 19*5 and 1971 respectively sad employee* of tha first hava baaa studied by Byrsn at si. (1976). All persona who had aver bean employed when exposure to 9Qt could oeeur were Hated from the pereoeael files of tha factory. Twenty-one were excluded because they ware forelgaars who left the country after a short period of employment. Tha remaining 750 wara followed to October 1974. Expeeted otabere of deaths wara estimated by multiplying tha person-years at risk by tha corresponding ag*-apeclfle mortality rates for tha whole country and tha expected numbers of cancer cases from 1958 to 1971 Inclusive (during which period all cancer eases had bean registered nationally) wara estimated by nultiplylag by tha national age-speelfle eanear incidence rates. Xa both lastaacoa, the rate* used were those recorded la 1989. Fifty-eight deaths wara found hot no figure is given for URL 13621 URL 13622 22 the expected number. Detailed figures ere given only for th nuabers of deaths and eaaaa observed aad expected for eaacor of eh* lung and for eiDCtri of eho liver aad fucrui combined and for eho numbers of daaeha from brain eanear and ehraa categories of cardiovascular dlsaasa. Two a*n known to hava angiosarcoma of cha llvar vara cartlflad an dying of llvar eanear or pancreatic eanear respectively, and a third nan diad of eh* dlsaasa 17 aonths after the follow up closed. Two French studies provide the results of a long tan follow up of u*n employed In one plant (Pierre at el., 1979-) and of a shore tarn follow up of sen employed In 12 plants (Laplaneha at al.. 1987). Th* first provided observations on 1,311 nan exposed to VC In the production of VCH and FVC aad in aeleeted ancillary operations fron the opening of tbs Tevsux plane in 1953 to the and of 1976 (Flarra at al., 1979). Six other anployeas ware excluded fron the study because of lack of occupational histories and 160 nan because tbelr vital etatua at the aad of the study period was undetermined. Twenty-five nan ware found to hava died against *8.75 sxpactsd fron contsnporsnoous sax and aga-spsclfle national nortsllty rates (3.7S of the non at rlak). One death was attributed to anglosarcona of tbs llvar in a nan who had been exposed fer aore than 15 years. The reported data are eo Incomplete aad cover oaeh e relatively short period fron the opening of the plant that they add nothing of epldaalologlesl value to the results of tbs other studies, sport fron the addition of a further case of aaglooarcona. The second provides observations on 1100 nan aged 40 to 55 years who, in 1980* were exposed or had bean exposed to VC in 12 pleats, which contstltuted "most of the French VQI polymerisation plants" (Laplaneha at al.. 1987). Many of tbs aan wars, or had boon anployad at Tavaux and ware presumably survivors of the cohort atudlsd by Flarra at al (1979)* The men vara followed for 3 years and their morbidity and mortality coopered URL 13623 23 with Chat observed la 1100 non of the iiu ages (+2 years) eaployed in ch aaa plants, bat who had never baaa exposed to VQi. Man in both groups art interviewed personally and lafonutlon obtalaad about chair aaohlng and drinking habits which warn fooad to ba similar la cha two groups. Man la tha exposed group had baan first exposed oa average about id yaars and had first baan employed la tha plant about 11 yaars previously. Morbidity and aortallty data wars recorded annually by tha plant physician who successfully traced 9IX of tha exposed nan and 96X of tha controls* One of the exposed nan but noae of tha ueexposed developed an aaglosareoaa of the liver. Onto are not given aaparataly for different periods afear first employment and It la lapoaslbla ta assaaa tha significance of the finding that I of the exposed sen developed lung cancer against 2 of tha controls, vbleh My wall roflaet ehsnea occurrence of unusually few eases la the control groups as the proportion of oil cancsrs in the'control group that arose frow the lung (2 out of 15) is unusually low. One exposed su developed e cancer of tho lympho-haauatopolat1e system against none of the controls, but no mo la either group were known to hove developed uelenouee or cancer of tho brsla or thyroid. A Japanese study reports the uortalley experience of 4524 non employed for at least om year baforo 191) la om or other of 25 Japanese plants which began producing vinyl chloride monomer (TOO or polyvinyl chloride (FTC) between 1949 and 1914 Inclusive (Rshaaura, 1913). The an were followed to 31.10.75. Twenty eight per seat of tho ms were observed Mrs then 20 years after first employment, but com wee observed nore then 21 years. Only 0.12 of the mo were entroeod end espies of the death certificates were obtained for ell the 209 men Mo hod died (4.4X of the Initial cohort). Individuals mere classified according to the lob on which they hod boon longest employed at tho terninstloe of their follow up end date ere given separately for 2544 mb classed as saployed on WC URL A362' 24 production sad 1971 others (Including 900 classed at VCM production worker*) If this Italy 1* continued for further to peers It should provide useful additional Information, but the present data include coo few observation* os nee nor* than 20 pears afcer first exposure to be of any material use* They confirm the evidence of e heterd of liver eencer with 6 deathe In PVC production workers eg*last 2.54 expsetsd from national rates while only 1 such death wee observed la the other workers against 1*82 expected. One of the 8 deaths from liver cancer le PVC production workars was certified as due to enfloeercooa of the liver end et leeee one of the other liver eencer deaths was due to the same eausa, lung cancer deaths were not In excess (2 observed in PVC workers against 2.33 expected). Bo data are gives for caacers of the lymphatic aod haematopoietic epatema, for cancer of the bralo or thyroid, or for molanomaa. The mortality of 30S Japanese VC workers reported bp Heauda (1979) Is presumably subsumed lo tfakamura's (1983) later end larger study, Batarde of cancer The results of the four most useful studies ere -listed individually In Tablet 1 end 2, The overall results for ell causes, liver cancer, end three broad groups of conditions are shown to Table 3, end those for 10 types, or daesoo, of censor in Table 4, Dot# have not been reported for each typo of cancer In each study and the sources of the deco are, therefore, specified separately for each type* Additional Information le given la Table 3 for 7 types or classes af eencer that baa bean obtained from four other lees lafomstlvs studies (the Carman study reported by Croleer et el., 1982, the Borweglan study reported by Beldass at el., 1984, the Swedish study reported by Byron at si., 1978, end the pert of tbo Italian study roletiog to the Bavonaa plant reported by Belli et al.. '9 t in n 25 1986). Table 3 show* that, apart frow cancer of chs liver, the overall mortality it what would bo anticipated for on industry without any major hazard of aeeldent or disease. la particular tho otoadordlsod mortality ratio (8101) of 84 for diseases othor than caoear is typical of tho ratios that aro cp--only observed la groups of employed een. A low SMK of this ordor reflects tho "healthy workar effect", which results frow tha slaetloa procsss that inevitably excludes some of tha laast healthy aabere of tha population froa laduatrlal aaployaaat. This offset doss not, however, normally offset tha mortality from eaaear bsyond that observed ia tha first faw yaara after tha start of aaploywaac aod aa 8HK of 102 for eaaeara other thas cancsr of tho liver is eoapatlbla both with the absence of hasard sad with SHKs of 14 for other diseases aod 77 for accidents, poisonings, and violence. AaglosareoBs heath eartlflcatas era an unreliable source of information about tha histology of cancers that causa death, but there lo no reason to suppose that the exeess mortality attributed to liver cancer (or, in tha 08 series, liver and gallbladder caoear) la not entirely accounted for by tha known % hazard of angiosarcoma. Fifteen of tha 37 deaths* attributed ta eancers ef tha liver and gallbladder In tha 08 series are known to have bean due to angiosarcoma of tha liver (Invlroomentel Bealth Associates, 1988). In tha British oerlee, 7 of tha 11 deatho attributed to liver cancer, net specified as secondary, wort known to bo englosercoaaa sod oil occurrad lo autoclave workers against 0.38 expectad liver cancers of oil typos (F<10 \ `Increased to 39 In Table 1, to taka account of the 97 extra daatha of unknown eaeee. URL 13626 26 Joaea, 1966) la the Canadian series, histological review showed that 7 of tha 9 ao called liver caaeara vara anglosareomae (l had haaa daserlbad aa an angiosarcoma on eha death certificate, 2 aa hapatoaaa and 5 aa uaspeclflad liver caaeara) One ao called liver eancar daath was found to hav haaa duo to eaaear of tha sigmoid colon, vhlla 1 daath attributed to angio sarcoma of tha peritonea vao foosd to haaa haaa doa to angiosarcoma of tha liver (Theriault and Allard, 1991) la tha Italian study, furthar evidence raaaalad that 3 furthar daatha should haaa haaa attrlbntad to caacar of tha llvtr (making 6 la all), hut only 1 of tha 4 was daserlbad aa aa angio sarcoma (Ball! at al., 1996)* Furthar evidence that the eueeee mortality fron llaar eaaear <or liver cad gallbladder eaaear la tho US earlea) eaa ba attributed wholly to tha knows hasard of angloaareooa la obtalnad by comparing tha axeaaa daatha with tha ntabera of daatho froo angiosarcoma befora tha aada of tha followup period that art racordad la tha Register of Liver Angiosarcoma Caaaa salntalned oa bahalf of tho AssocIstioa of Plasties Kaaufacturars la Europe by Imperial Chemical Industry PLC (Beamott, 1996). Fifty oae eaeesa liver cancers ere recorded la the combined data aad 49 aagloaareomaa are recorded la tho to|liur for tho relovaat parloda for tho throe countries and 'the two Italian plants* that ara eovarad by tha survey. Boms of the 120 eases yet recorded oa tho Be9lcter were la men who wore first exposed altar 1969 and mona of tho 45 man affected la Borth America were first exposed altar 1994, All may, cberafora, hava base exposed to concentrations of several hundred ppm and many may haws bean Twenty alma ara registered aa occurring la the OTA against aa catlnatad axceat of 33, 9 are registered la the OK against aa axeaaa of 9, 10 la Canada cgalaat an excess of 9, cod 1 la a aaa employed la one of the two Italian plants studied against an excess of loss than 1. Too of tho Canadian caaaa ara racordad la tha Register ao occurring la men who had base employed for S years and it la possible that tho actual duration of employment had beta slightly lata than 3 years, whan tha nan would hava been excluded from tha study. URL 13627 27 expoeed to concentration# appreciably higher (Barr, 19S1). Lung Canesr The lBoo that ixponrc to VC night cause cancer'of tho lung was suggested by Mwiob at al. la 1974, when they noted 13 eaaaa la a study of proportional mortality against aa expeeted maber of 7.9. Tha conblntd data that art shown la Tabla 4 do not provide any support for tha hypoebesle atlthar for respiratory eanear as a whola (SMB 97) nor la eha apaelflad data for lung eanear la tha USA, UK, and Italy (SMB 99). Thors ara, however, consistently hlghar rlaka la tha subgroups of nan la tha US and UK aarlaa la when occupational haaarda would bo uora likely to bo soon than la other groups. This la aaaa fron Table 4, which shows that tha SMBs ara slightly higher la nan observed 20 years or nore after first asposura than in mea observed earlier, la nan employed before 1954 la the UK than In nan first employed after 193S (whoa exposure levels are believed to have been lower), la non employed for longer than for a shorter parlodo la the USA, and la autoclave workers in the UK, la whoa the aaglosarcoaa cases have neatly occurred, than la other workers. The differences ere all snail or vary snail, but they are all la the sane direction and tha probability that the rates should all be higher la the groups la which an occupational basard la eora llkaly to ho seen la each of the four pelre ef groups le 1 la 16. Addltloael lafometloe froa ether sources le glvan in Table 5. Thirty deaths (or cases) were observed la total and this la reload to 31.5, whan allownaee la node for the eonber of deaths due to ualdentlflod causes la tha Gemaa study, giving an Ml of 103. la tha Geroaa study tha SMB was hlghar la nea who hod boon exposod for 10 years or wore thaa la eea who had been exposed far lass (111 against 79) and la tha Porweglsn study, 4 of the 5 eaaaa observed occurred la ana whose occupations warn rogerdnd ea Involving high exposure egalaat 1.82 nut of the 2.04 expected, both the URL 13628 28 CtiUB cad cbc Swedish studies derived the expected ounbsr* of death* frou actional mortality rate* for a elngle year toward* the end of the study period. Tb# expected numbers of death* are likely, therefore, to have been overeatlnated and the Sit* correspondingly umdernstlwated ae the aortallty fro* loot cancer had beea rising throughout the period of obeervatlon. train Cancer The Idea that VC night eanee brala canear wae euggeeted by won*on et el. (1974), when they reported S caeee agalaet 1,2 expeeted, In the ease paper la which they euggeeted that VC night eanee long cancer* The combined data that are shown la Table 4 provide none eupport for thla hypothesis* Moeaen et al.*u caeee were, however, observed la US workers end ones he preauned to be laeladed la the total raportad by Eavlronaental Bealth Associates (1986), la which ernes they will have contributed e euheteatlal proportion of tbo total la Tabla 4* To tost the hypothesis we ought, therefore, to subtract Moneon et ajU's data* Their study wae not a cohort study and their expected death* do not correspond exactly to those in Table 4; w* can, however, only subtract both the observed end the expected cases fran the present total# Thla leaves 24 observed against epproxlaetely 11*3 expected, a difference which night easily occur by chance (P, one-tailed, * 0*1)** Additional laforaetion froa two other sources la given In Table 5* The snail excess reported provides little farther evidence of aa occu pational hasard, aa ana of the two deaths observed la the Swedish study occurred la a young nan who had been eaployed for less thaa a year when the diagnosis was made, Ohlle the excess death rata for brain cancer observed If we regard Vaxwellar at al*'s (1978) study as the origin of the hypothesis ww should ha Telt~wlth 28 deaths against 18*94 expected (P, one-tailed, 0*07)* URL 13629 29 In tho Gruo study was less than that observed la ehealeal workars not exposed to vinyl chloride (2.9 deaths after allovanct for desehs froe unknow eeueoe against i.6 expected) end In workers la the FVC fabrication Industry (5.9 deaths after allowance for deaths from unknown eauses against 1*1 expected). Cancers of lyapheele end Haematopoietic Tisanes The ides that VC alghc cause eancev of the lynphatle end haeaato- poietlc tleeuss - and wore specifically the lyaphstlc tissue - was suggested by Tabershew end Caffey (1974) and by Vaxwoller et_ al. (1976) whan they found, respectively. In two cohort etudles (i) 5 deaths froa lymphomas la the aoet heavily exposed workers against 2*54 expected, end (11) 4 deaths froa eeecers of the lyaphstlc end heaaetopoletlc tissues egelnst 2.5 expected. Thess sasll sxessses alght have bean Ignored If the laboratory findings had not been Interpreted as suggesting chat lymphomas were produced experlaeatally la aalaals exposed to vinyl chloride by inhalation (Maltonl and lafaalna, 1975) On this heels, the idee that similar exposura might also cause lyaphoass la Man needed serious consideratIon. Ths data froa the four principal studies that ere suaasrlxed in Table 4 provide little enppert for tbo hypothesis when all esaesrs of ths lyaphatle sad hssaatopolotle tissues art considered together (57 deaths egeiast 50.97 expected, SMI 112) end very little aors is obtained froa the separate date for cancers of the lyaphstlc syStsa (using Tabershew end Caffey*s definition of ICO List numbers, 9th revision, 20CK203 end 205) that are shoos in Table 1 (35 deaths against 29.40 expectsd). Ths position is, aoreovsr, hardly altered by subtracting tbe data lo Tsbershsw and Caffey's Initial report (29 deaths egelnst 23.39 expected, Dflt 124). goae little eddltlonel Information is provided by tbe results of tbs Ceraaa study, teas of which aro shown in Teblo 5. This obtalasd an SMI of 214 for expossd workers (bssod on 15 oboorvod dosths, inersasod to 16.5 URL 13630 30 whan tUoiraaei li ia4 for tha nuabar of daatha of unknown causa) against SKRs of 77 aad 3d for an unexposod group of chaaical workars and a group of PVC fabricators and abowad that tha excess la tha exposed workars was prasanc only for nan who had baan axposad for aora than 1 year and was nose aarkad for aan who had baan aspaaad for 5 years ar aora (10.7 daaths aftar allowance for tha noahar of daaehs of unknown eaasa against 4.0 expected, SHR 268, P ooa-tallad <0.01). Mslanoas An exeese of aalaneaa was raportod In Vorwaglan workars bp Baldass at al. (1914), who ralsad tho poaelblllty that VC night have producad tha dlaaaaa. Pour aaaaa wnra obaarrad whan 0.79 vara axpected, thraa of which wara in nan whoaa occupations involved tha highest axpoauraa (against 0.51 sxpactad). At tha tlaa of writing thslr report oaa further ease had bean dataetad with onset 3 pears after the closure af the study. Subsequent studies in other countries haws, so far, reported only 2 daaths against 2.0 sxpaetad (sea Tables 1 and $). Thyroid Cancer An excess of thyroid cancer was also reported In tha Korvaglan study, in which 2 caeca wara obaarrad against 0.16 sxpactad. The investigators wara not aware af any ether studies indicating an exeeaa af this type of cancer and they draw no conclusion froa their observation. Two of tha thraa asjor studies that hare been reported eloca tha Norwegian observation vaa aada give no data for thyroid cancer; the third reported two deaths agnlnot 0.43 sxpaetad (aaa Table 1). One death froa thyroid.cancer, it may be noted, was reported In tha USA by Hanson at cl. (1974). Cancers af Digestive Tract uggcstlona that VC aight causa cancers of tho digestive tract in general have toaeclaas baan aada but ebay haws not taken adequate account 31 of the contribution that cancers of the liver nake to the total nunber of uaettt of the digestive particularly whoa It la bora# la mind that none llr cancers are likely to be nladiagnoaed at eaaeart of other organa, The combined data froa the four principal atudlaa shown la Table i weigh heavily agalaac tha Idea that aay eucb affect haa been produced. Other Cancers One of eho remaining types, or elaaeea, of eancer Hated In Table 4 ehove a statistically slgnlflcaat exceeet n--ilj~. tha hataroganaoua group of "other cancer*" (83 observed deaths against 85.24 expected, ?, twosided, <0.05). This excess is only narglnally significant and nay he a ebancs observation. Tha nest likely explanation la, however, that a few angleeareowas of tho liver were sot recognised end were diagnosed as saeondary livar cancsr or earclnooatoels, alta unknown, and ae boosted the tuabtr of deaths la this category. URL 13631 tasards of non-nalianant disease Ho previous study has suggastod that any nen-nallgnant eauat of death other than clrrheala of tho liver would be likely to be Ineraaaad aa a result of exposure ts VC, and that la only propound to bo Increased because of tho Host changes that ware observed as part of tho 'vinyl chloride Illness1 (Meretailor at al.. 1973 and 1975; Jonas and Sulth, 1982). Two othor possibilities have, however, bean reloads namely tha production of aon-aallgaaat respiratory dlaansa because of tho changea la lung function and x-ray appearances that have bean recorded for nan exposed to PVC dust (Lilia at al. 1978; laser at ol., 1985; footer, 1980; Lloyd at al., 1985) and acuta cardiac death, froo analogy with the effect of other halogeaated hydrocarbons (Jonas, 1988) sad the observation of an Increased nortallty froa ayoeardlal infarction in tho few years following the cassation of exposure lx tho Swedish PTC processing industry (Melon ae al., 1981). URL ^ 3632 32 hslevsat figures for tha mabers of deaths fro* tbooo ond other ncnmallgnaat causes Chat art obtainable fro* the four principal acudlat vara given la Table 2 and are euMarlsed la Table 7. Clrrhoelo of Um Three of the four studies save eeparate figures for elrrhoaia of the liver, aoae of which shooed an lacreaaed mortality (Table 2); la combin ation they gave an SMS of 69 baaed oa 46 deaths* The fourth study, vhleh did not give separate data for elrrhoaia of the liver, reported 4 deaths fro* all diseases of the digestive system combined against 3.8S expeeted aad noted that the 4 included 2 that were certified as due to elrrhosio of the liver, but were actually due to aaglosarcooa (Theriault aad Allard, 1961)* In the two supplementary studies In vhleh data are given for this disease, the Ml was S2 la one, based on 15*1 deaths after making allowance for the number with unknown cense (Greleer e el., 1982) end 133 la tbe other, based on 7 deaths (gakamura, 1983)* goa-mellgnant tasplratovy Disease The date for non-mellgaeat respiratory dlseeae ere confusing la that the total SMX from fcha eomblnad data for tba four principal studies Is 80 a^ la tba sort of figure that la coMonly found lo healthy Industrial populations, yet the 08 study recorded a substantially increased mortality from emphysema (41 deaths aad Ml of 180 baforo any sllovaaca ts uads for deaths of imbnon cause)* go such oncost was found in tba OX vhsrs 36 deaths from bronchitis gave ea Ml of 82* International comparisons of chronic noo-nsligaant respiratory disease ere complicated by the usage of . different tame to describe abet It la nov agreed. Is beet called chronic obstructive lung (or pulnoaery) disease, but vhleh la the peat tended to be celled emphysema in the QSA and chronic bronchitis In the OK* It muse, therefore, be presumed that ths two categories nf 'emphysema' and E9enun 33 'bronchitis' used respectively in tho two Largs national studies were unc to describe the mm thing. On this basis, we should here to asstae thse the experiences in the two countries were very different, despite the fset thet both relsted to cohorts Chet hod Tory slailsr experiences of angiossreoM of the liver sod so, presumably, fairly slailsr exposures to VC. Separate figures are shown, where available, for the aortallty observed la mo with different durations and intensities of exposure in Table S. Unlike the data for cancer of the lung that were shown in Table 6 they provide no consistent evidence of greater risk in the groups in which an occupational basard would be expected to be concentrated.* The authors of the Bnvlronaental health Associates (1946) report were unable to give any explanation for the increased aortallty fron aaphyseaa and polne out that It could haTdly be due to excess cigarsece smoking as thsrs was no overall exeeee for cancer of tho lung. It lo striking, however, that the exease is Mrs than compensated for by deficiencies in tbs ether categories of non-Mllgnant respiratory disease (pneMonla 19 deaths,* SKA 47.0; other respiratory disease 14 deaths, M 42.6) end the question arises whether the enphyssM excess could ho s claselfleatory artefact. environmental health Associates (1966) list all the 41 deache which ehew that they were coded wader 1CD List no. 927 which, in tho out of dote 7th revision thee was used for tho coding of all deaths in the study, wee the code for 'ocher respiratory disease not otherwise classified' and included emphysema. Under that revision, however, emphyssM thet was associated with bronchitis should bo elaeoiflod with bronehltls under List snobsra 900 to 902 and tho *The deaths attributed to different groups of respiratory disease* and the corresponding ms that are eitad in this section for tho OS study are as given by SovlroMontal Health Associates (1986) sad have not been adjusted to taka account of the makers of deaths of unknown ennso. To take account of tho deaths the observed deaths and SMHs can both bo multiplied by 1.0674. URL 13634 34 possibility aay b considered thst sons of ths eaphyseaa deaths should have been classified in hm category of rosptrstory dlssatt othor than List no. 527. Zf this were so. It could sccouat for both ths excess aortsllty Iron o^hyssM and tho grossly deficient aortsllty froa other noo-aaUgnant respiratory diseases. Mo excess aortsllty froa "bronchitis, saphyssas, sad ssehaa" was observed to ths Gsrasa study (Crslser at el,, 1982; SM8 44 with 6.5 desths observed after aaklng elloweoce for the auaber of deaths of unknown esuss). Cardiovascular Dlsesso Dees for lscheealc (or srteriosclerotlc) hesrt disease (which Bay be preefed to include the vest aajorlty of all deethe certified as due to scuts csrdise disease) are glvsa ooly by the two big national studies sad they provide no evidence of an locroassd aortsllty. Ths SMI of SO for this grosp of diseases sad of Si for all cardiovascular dlsssss rseordsd in ths four principal studios are typical of tho Ml thst are rseordsd la healthy Industrial populations sad chars is as suggestion of say occupational hacerd In the subsidiary analyses provided by the two oetlooel studios. In particular, chars is as svidsaes of so laeraesed aortsllty within one aonth of Issvlag saployasot la ths OK study slthsr for all wbrksrs (52 dsaths, M 61) or for the aost heavily expoeed autoclave workers (9 dsaths, M 42). A slight lacrosse la lscheealc heart disease aortallty was rseordsd for tho exposed workers la tho Coraaa study (Cretsor e sl, 1962), but It was loss thso thst rseordsd for tho uoaxposed chsatesl workers sad ths FVC fabricators (tMLs of 127, 151, sad 15$ based oa 97.2, 126.7, sad 109.7 dsaths sftsr allowing for tho aiarfiar of deaths of unknown esuss). Discussion Ths Information thst has now bean obtained about ths long term health 35 of mb occupationally xp0ad to vinyl chloride la aaaslve and compares favourably with chat available for any othar occupational group. Two face* ara outstanding. First tha nan have experienced apaclfle hasard of a typo of canear chat ta normally extraaely rare: namely, angloaarcoma of tha liver. Tha rarity of chit dlaaaaa under othar eoodlclona mada th dataetlon of tha haiard aaay; but tha long latant parlod bafora tha dlaaaaa appaara aftar flrat exposure (alaoat always mora than 10 yaara and usually mart than IS yaara) maant that a largo nuabar of man had bttn exposed bafora tha haaard waa datactod and that It will atlll ba many yaara bafora we can ba esrtain about tho axeant of tha protection provided by tha reduction in exposure in tha 1960s and tha furthar reduction that followed tha racognltlea of tha haaard la 1W4. Thera la, unfortunately, no effective traataant for tha dlaaaaa and tha nuabar of eaaas la raflaeted in tha nuabar of dastbs. Soma 50 daatha havt occurred la tha 16,710 ara oho hava baan followed up la tha four principal atodlea that have been reviewed in thin report, no that appraalmataly 1 la 33S ara hava baan affaetad, while 22 of ebe deaths hava bean das to this oao causa, Eventually aany aora ara must ba expected to davalop the dlaaaaa. Ona astlaata (Meholeon at a., 1986) auggeata that the total may be increased 10 timet, but e more realistic astlaata ta 2 to 5 tlaae (Forman c^cl,, 1985). Tha ascend outstanding observation le that tha nortallty of the axposad men, other then that due to angloaareoaa of the liver. It typical of tha normally healthy Industrial worker. That it not to tty that no other hasard exists, but that the affect of any othar hasard Is small. Tha aaaslva data that art now available provide no reason for thinking that any hasard othar than ona of cancer baa bean verlookad. it Is, howavar, still dlfflealt to daeldt whether fC does produce e risk ef developing cancer, other then aagleeereoan of the liver, which might be mall compared to the rlska produced by non**occupational causes, but yat segenan URL 13636 36 absolutely almost as large as the risk of developing tha normally wary rare anglosareoma* On* of tha many Hazard* ehae have boon suggested earn be dismissed, at ehora ta mo evidence to support Its namely, Cha Ida* that VC causae any caacar of Cha digestive trace oehar than anioaarcama of eha llvar. two hazards Cof melanoma and eamcar of tha thyroid) harm haam suggested only vary recently and few of eha available stadias have provided Information about Cham* Thara la mo good thoorotleel ramson or laboratory avldtne* to auggasc that althar ahould ba produead by VC and, on prasant evidence, tha slmplaat explanation Is that eha raportad excesses arm eha chanca affacta ehae suae ba expected whan many dlffarane typas of cancar arm atudlad In aavaral dlffaraat populations. go far as melanoma Is eonesnad It has co ba ramambarad ehae eha disease baa bacoma much more eoamoa in recant years la Scandinavia (vhars tha exeeas has been raportad) due, It la believed, to tha popularity of sunbathing and tha ineraaaad opportunities for Scandinavians to travel to tha wsrmar parta of Southsrn Europe and North Africa. Tha extent to which this may hava affoetod tha obtarvatlon in Norway needs to ba examined. Two other hazards (of lymphoma and of brain eamcar) ware suggested by the early results of soma of tha American studios* That VC might produce a hazard of lymphoma was Initially supported by tha preliminary rasulta of animal studies, but tha complete results of the many studies that hava now bean undertaken (saw Melton! and Lefemloe, 19$4) do mot smggaat that lymphoma or any other caacar of eha hscnatepoletle system la liable to ba produced* Thara Is, however, soma avidanca that brain tumours can ba produead In rata (aaa Ksltonl and lefamlne, 1M4) Tha hypotheses that lynphemas and brain cancers might ba produced by VC haws bean supported by tha observation that both Chase types of cancar hava caused death mere often than might ba expected from national mortality rates, but tbs /fm iH n 37 excesses observed Is ths combined data from the four principal studies or* small end sot statistically significant sad th* hypotheses remain unproven. Ths noil excess of brsla cancer to particularly difficult to evaluate as mortality rataa froa this disease her* cheated rapidly over time la company with laprovad methods of diagnosis and the suspicion of aa occupational hazard (which was raised la 1975) could have Influenced the findings* What excess has occurred has been Halted to the USA sad Germany and the Geman findings carry little weight* ae the excess was found la each of th* thre* occupational groups studied* Irrespective of the ehealeale to vhleh they were exposed. The supplementary data froa the German study showing sn lncreeoed mortality froa lymphatic and haematopoietic cancers ere more Impressive* pertlculerly ae the eseees was most marked for men who hsd been employed for at laeet 5 years* la these elreueetancee* vs must still suspend Judgment until the data for each study are analysed for each specific type ef cancer* by Intensity and duration of expoauro* aod time since exposure began* Thara renalna the suggestion that VC might causa lung cancer* At first sight* this is ruled out by the SML of 97 for the combinod data for respiratory cancer for tho four principal atudlaa* Lung cancer la* however* normally so common (accounting for about Ml of the oxpected deaths) that aa increase in mortality that was half aa important (uaerleally) so the mortality from angiosarcoma of the liver might eaally be overlooked (95X confidence limit of the SML, S5-U2). The Incidence of the disease la moreover very variable within e country and there mutt be doubts whether tho national experience provides a suitable control far non employed In plants that ate mot evenly dletrlbutad about the country. In these el retaintanesa wo cannot exclude aa occupational haserd unless it can bo ehown that the mortality within tho exposed men is independent of the factors that might be expected ca influence it* if some of it were 38 occupational In origin: namely, the Intensity end duration of exposure and tho time aloco exposure began. It In not possible to exsnine thoto relationships in detail, at tho roportt do not provide all tho neeessary Information* Seek Information aa they do provide, vhleh was smasrlsed in Table 6, support! tho ldaa that VC produeaa a small haaard of lung eanetr. Taken In conjunction with tho knowledge that long toeours bars botn produead in several spociaa of aninsls axpoaod to 7C by Inhalation (taa Maltonl and Lefanise, 1984) it would seen that a snail hasard of lung canear probably did occur* Tho evidence la not* however, strong anough to conclude that It daflnltaly did. If it did, tha hasard was evident only in nan who had baan anployad for nany yaaro at a ttna whan exposures of several hundrad ppn or nora vara coanon and any paralaelng rlak can ba only mlnuta and incapabla of dataction* Tha questions that Hava baan laft unansvorad by thia discussion night wall ba antvarad daflnltaly if (1) all tha exposed nan could bo follovad to (say) tha and of 1984, (11) tha Investigators eould praaant tbalr data In conparsbla nays, taking account of duration of anploynant and tins aloe* flrat anploynant, and proaantlng data aaparataly for nan flrat anployad bofora (nay) 1985 and batman 1985 and 1974, and (111) aatlnataa muld ba nada of tha affact of eorracting tha resulta for oach group of anployaas for tho locality In which they lived and worked* BAZAMDS TO TBE GBTOAL POPULATIOH Aa VC has baan proved te causa canear in Han and la a nutagan la laboratory experiments* it nuat be praamed that even the minute doaaa that aaeapad into tha genaral anvlronneat from production plants or (in tha aarly days of manufacture) fron PTC materials* will havn caused sens risk of cancer to the general publle* These risks nuat* however, have beeo vary snail, aa air concent rations of 1C, even within a kilometre of YC handling 39 plant*, wad to be (to or around 1975) of tho ordor of 10 to 40 ppb (Eavlronaontal Protection Ageae?, 1973; Air Products and Chemicals Inc., 1976; Baxter at al., 1977) and this la about oh tan thouaandth of tha concentration that haa eauaad an occupational hasard, it la obvioui, therefore, that it would ba impossible to datace tha rlak of any eancar that night ba produced bp VC other than a rlak of anglosareona of the liver, aa it haa proved to difficult to detect an? other rlak in ueo who vara exposed occupational!?. Tha pooltlon with regard to aagloaarcona of tha liver la different* Thle dlaaaaa la normally eo rare that, in the abaanca of specific exposure to one or other of tha known cauaaa (VC, thorlm dioxide, and areenle In paatlcldaa and medicines), tha annual lacldanca la of the order of 1*2 x 10~7 (Byren and lolnbarg, 1973; Brad? at al,, 1977),* In thaaa elrcmstsneee tha discovery of ovon ono caao in a man living clean to a factor? la whleh VC wee need In tho da?a before axpoouro waa tlghtl? controlled may ba regarded aa preemptive evidence of tho affoct of environmental pollution. Several surveys have bean undertaken to nee-If an? eueh caaaa have occurred, Sarle at al* (1976) and Blinder and Parabagan (1978) sought for caaaa la tha vicinity of plants handling VC la Yugoslavia and Swadan and found none, Dalderup at al* (1976) found I eoafiraod caaaa la Holland aot attributable to thorotraae or araaala and could trace "no contact with vlnpl chloride" but the? node no epeelfle Motion of the patient's place of rasldanca, Baxter at al* (1977) found 14 eoafiraod cases In Great Britain over a 12 past period, one of ohleb waa In a nan who had lived half kllonetre fron a FVC naaufaeturlag plant, and Brad? at al. (1977) found 19 caaaa la Raw York State over an IS pear period that could not bo The figure of 1*4 x 10*5 eltod b? loath o^ al. (19S5) aeons to have bean a nlaprlnt for 1.4 x 10* URL 13639 40 Attributed to eny knew* causa, S of which wore to people living within o lie of pleats manufacturing or using VC* The oworsll incidence rotes In these lest two studies ere not unduly high, but the occurrence of as isoy so 6 coses in people living so close to manefecturing pleats Is surprising, tredy snd Ms colleagues,.moreover, coopered their series of patients with oetehed controls end found thee non# of the control subjects lived equally close to e plant. Two of those 6 neighbourhood esses (2 In England end 5 In Pew York Seats) eennoe be attributed to environmental pollution with VC, ns the man who developed the disease had lived near eh# planes for 6 end 8 years respectively before developing the disease and this le too abort a tloa to allow for tha nacessary latant parlod. Tha other 4 eases, however, all oeeurred after IS or more years local rasldanee and tha dlseovary of thasa casss strongly suggests that pollution of tbo environment round pleats manufacturing VCM or PVC may have caused e minute hasard to tha geoaral public. Current concentrations round VC handling plants art certainly much lower than those reported previously by the CPA, heeent brltlsh measure ments made within a fsw hundrsd mstres of tha VCM arsss have givan average values below the dally limit of dotoetlon (5 ppb) for throe of flva plants, tha readings at tho two others being 20 ppb (100 m outside the boundary fence) end II ppb just inside It (Turner, Payee, Buehby, 1904) though substantially higher values were recorded os two occasions associated with cqamtsslonlag of one plant and am accident at the other. On any reasonable erleerloa the hasard to tbo gansrml public (if tbora la any at all) oust bo negligible (loyal Society, 1913). Ho otbar hasard to the general population, other than a hasard of cancer, can reassembly be postulated. 41 CONCLUSION Occupational exposure to concentration# of VC of the order of 1,000 ppe ho* caueed th classical *vlnyl chloride illness'* Exposure to such levels, or perhaps lower levele of tho ordor of several hundred ppa, haa eauaad angloeareona of tho Hoar ofeor a latent ported that la aaldoo laaa than IS years* Tho evidtnee now available from following up neny thousand* of oan occupationally exposed to vinyl ehlorldo suggests that a small rlak of lung eancor may alao have boon produced, which must, hovewar, have bean substantially loaa (In abaoluto terms) than tho rlak of angiosarcoma of the liver* Tho balance of evidence doe* not support the Idea that VC ha* caused cancer In any other organ, but It la impossible to atate categoric ally that It haa not* Clearer evidence, one way or the other, could be obtained If the varloua Invaatlgatora would prevent their reaulta In eooperable forma that would allow them to be combined end subjected to ehe atandard taata for an occupational hasard* There la no evidence that exposure to VC has eauaad any other disease* Suggestive evidence haa baan obtained that pact level# of environmental pollution with vinyl chloride may have eaueed a few eases of cancer la the general public. URL 13641 umoiczs All PEODCCTS AMD CHEMICALS, I9C. (1976) Co--wf op the propped standard for vinyl chloride. Letter to 0.1. Goodwin, CPA, 23 Sapteebar 1976 (elted by Barr, 1976)* BttI JT. (1981) Hasard evaluation and rlk aaaeeenent for vinyl chloride. Eeport to Air Produces sod Cfceeieals* loc. (unpublished) BASE! ME, TOOCMA! MS, KENHEDY TP. (1985) Puluoury function end rctplrstory syeptoee la polyvinyl chloride fabrication workers. Aner . lev leap Pis. 131. 203-208. 1AXTEE PJ, AHTHOHY FP, MecSWEEM Ml, SCHEDEE PJ. (1977) Aniiosarcoaa of the liver la Orest Britain. 1963*73. Brit --d J. 2,. 919-921. BTBEE 0 6 H0LMBE1C B. (1975) Two possible esses of antlosareoes of the liver in *roup of Svodlsh vinyl chloride workers. Ann MY Acad $cl. 2*6. 2*9-250. BELLI S, BEETAZZI PA* 00MBA P, POA V, KALTONI C, HASXKA A, PZBASTD 1. EEGIAlfln A, VIGOTTI MA. (1*86) Indaflne aulla wortallta del produttorl dl PTC la Italia: dlsefne dello studio e prlai rlsultotl. Cancer Letters (in press). BENNETT B. (1986) Beslstsv of esses of sntlosereooa of the liver due to ssposure to vinyl chloride, (unpublished) BEADY J. L1BEBAT0R1 F* HAEPEE P, GBZENVALD P, BCTUUTT V* DAVIES TN. BISHOP M* POLAV A* VIAHEA - (1977) Anclosarc-- of the liver: an spldeolologle survey. J Met Center Inst. 59. 1383-13B5. BOFFLEB FA* HOOO t* IXFUE C* SUAEZZ L, EXLXAME DJ. (1979) Mortality experience of workers in e vinyl chloride nononer production pleat. J. Occuo Med. 195-203. Tin D* EMCH0LM C. EMCL0MD A, NESTEBHOLM F. (19B*) Mortality and cancer Morbidity In a (Poop of Swedish VOf end PTC productive workers. Environ 11th Fersoect. 17. 167-170. References eontd: 43 COOPER WC. (1901) Epidemiologic study of vinyl chloride workers: mortality thre*h Mir 31, 1*72. Environ Kth Persnect. 1_, 101-106. CREECS JL 6 JOHNSON Ml. (1*74) Aagloureou of liver to tho aanufseeure of polyvinyl chloride. J Oceuo Hod. 16. 150-131. DALDERDF LM, PEENI SC, IRAS C, 1R0NCKHURST PR. (1976) Anflosarcoea of the liver. Loncot. L, 246. DOWNS TD, STALLONES RA, FEAHE0V5KI RP, LABARTHE DR. (1977) Vlnrl chloride, birth defects. and fetal witmt > critical ritlw, Report to tha Society of Plastic Industries, Ioc. (unpublished) DOCK SW, CARTER JT, C00MU9 EJ. (1*75) Mortality study of workers in a polyvinylchloride production pleat. Lancet. 2, 1197-11*9 BLENDER CG, PERS8A6EN 6. (1*76) Pilot study constraint tha oortallty In Sjurunda Cowinltv. Svadlata latora Conssrveney Board, (eltad by Barr, 1*61) ENVIRONMENTAL HEALTH ASSOCIATES (1*66) An update of an aoldanloloilcal atody of vinyl chloride workers 1*42-62. Pinal report to the Chemical Manufacturers Association. Envlrooaantal Health Associates, Oakland, California. ENVIRONMENTAL PROTECTION ACENCT (1975) Standard aopoort docuaant and environmental impact ststaaantr oalaslon standard for vlnvl chloride. IPA 450/2-75-009. Environmental Protection Agency, Vcehlnston. EQUITABLE ENVIRONMENTAL HEALTH (1976) Epldoaloloalcal atody of vinyl chloride workers: final report to Manufacturlns Chaaiscs Association. Equitable Environmental Health, Rockville, Marylead. PALE H, 1SLLE6 NC, ISHAK EC, THOMAS LB, POPPER 1. (1979) Epldamloloty of thorotraat-inducad hepatic oaRioaarcomaa. Environ Res. IS. 65-73. cr> s* , > inuin i t V Refereneee contd: 44 FIECHTNER J, UilS C 1ENTMERSTER K. at el. (1976) Epldealoloslc notes and reports: aoslosareooa of the liver - Wisconsin. Morbidity end Mortality Weekly Report. 25, 57-56. FORMAN D, BENNETT B, STAFFORD J, DOLL R. (1965) Expoeure to vinyl chlorlda and angloaareoM of tha liver: a rapore of the register of eases. Brit J Industr Mad. 42, 750-753. FOE AJ 4 COLLIER FF. (1977) Mortality experience of workers exposed to rlnyl chlorlda aoooaar la tha aaaufacture of polyvinyl chlorlda In Croat Brltala. Brit J Industr Mad. 34, 1-10. GREISER von E, REDIL V, WEBER S. (1982) Vlayl-ehlorld exposition und oortalitlt deutscher ebaaloaTbaitar tm vergleicb >ur aortalitlt alcbtaxponlartar chealearbelter uad FVC-verarbeltar. Zbl Arbattened. 32, 44-62. REAIR.or, FALX H, CREECH JL. (1975) Charactarlatlea of eases of sagloearcooa of tha liver anon* vinyl chloride workers la the United States. Ann WT Acad Sci. 246. 231-236. RELDAAS SS, LAHCABD SL, ANDERSEN A. (1964) Incidence of cancer saosg vinyl chloride and polyvinyl chloride workers. Irlt J Industr Mad. 41. 25-40. INTERNATIONAL AGEHCT FOR RESEARCH OH CANCER (1979) IARC nonotraohe on the \ evaluation of tha cerclaoaaalc risk of chaailcala to hunana. 71. 19, pp.377-436. Xatarnaeioaal A|aey for Research on Cancer, Lyon. JONES DF 4 SMITH FM. (1982) frosreasloa of vinyl chloride induced hepatic flbroele to engloearcoM of tho llvor. Brit J Industr Mod. 39. 306-307* JONES ID. (1986) A aortellty etody of vinyl chlorlda aonooer workers employed la the United Kiagdoa 1940-1964. . (la preee) *wror Inference* eontd: 45 LAPLABCHZ A, CUVEL P, COKTASSOT J-C, LANOOZIEXE C. (1907) Exposure to vinyl chloride aoaoMr: report on e cohere study. Xnstltue Gustave losuy (to to published). LIUS t, AHOEXSQB B, MItLEI A, SELIKOrr I. (1976) Pulmonary changes won* vinyl chloride polynerleatloo workers. Chot. 2,(Suppl.), 299-505. LLOYD KHt GA0LD S, COPLAKD L, SO0TAI CA. (1905) Epldsetologlcsl study of lung fuuctlou of worker* *t a factory manufacturing polyvinyl chloride. Brit J Industr Hod. 41. 320-533. MscMABOM, 0. (1977) Vinyl chloride end bieta reproduction. taport to Chauical Manufacturers Association, (uapuhllshod) MALTOBI C, CtLIBEITI A, GIAHHZ L, CHIXCO P. (1974) Vinyl chloride earclnogaassls: current reaelts end perspectives. Mod Lev. 63. 421-444. 33 ^ ^ MALT0K1 C. A LEFEK1BE G. (1975) Carcinogenicity bloaooaya of vinyl chloride: current rosnlts. Ann Iff Acad Set. 246. 195-210. KALTOHX C, LEPEMIBE C, CILIIEITI A, COTTI G, CABIETH 0. (1904) Experimental raaaarch on vinyl chloride carcinogenesis.* Archives of Keeearch on Iodustrial Carclnoneneeia. Vol. 2. Eds. Naltonl C. 6 Hehlaea HA. Princeton Scientific Publishers, Princeton. MAXSTELLSE BJ, PfLIAflB VC, MULES t, GEDIGK P. (1975) Unusual eplenomagallc liver disease as evidenced by peritoneoscopy end guided liver biopsy aamg polyvinyl chloride production workers. Ann iff Acad Eel. 246. 95-134, MAESTELLEE HJ, PCLSAtt WE, W1LEX X, JUHE S, LAKE CEt KOBREE BG VELTKAR G. (1973) Chronisctotoxlseto letorschlden tol Arbeleetn in der PVC-Produktion. Ptoch Med Bochenochr. 90. 2311-2314. ftefereocoa eoned: 46 UHL 13646 masuda T. (1979) Long term mortality studj of vinyl chloride and polyvinyl chloride workara la Japaoeee plant. Arch Induetr art Toricol. 30 (Soppi), 403-409. HtKKOTA ft, MIHAILOVA A, H0SK0 H* (1978) ttlt Uttnow. 21., 460. (cited by larrt 1986) NOLAN I, NOLAN G, BOLMBERC 9, CLOPSSON 8, HOLMBND L, NOOSING ft, WESTERHOLM F. (1981) Mortality aad cancer rataa among workara in tha Swedish pvc proceeding Indoeery. Environ Hlth fsraoect. 41. 145-151. MONSON U, PETEIS JM, JOHNSON MR. (1974) Proportional mortality iwb| vlayl-ehlorida vorkara. Lancet. , 397-398. MAKAMOftA ft. (1983) A mortality etody of vinyl ehlorldo workara la Japan. SaniTo Ike Plaawka Zaaahl. (Suppl), 49-37. NICHOLSON HJ, HAM10MD 1C, SEIWtAi H, StLXKOFF U. (1973) Mortality experience of a cohort of vlayl ehlorld^polyvlayl chloride workara. Ana NT Acad Sci, 246 . 223-230. NICHOLSON VB, KENNZUftGCR PK, TAB! D. (1984) Tranda In cancer wrtallty among workara la tha ayathatlc polymara Industry. las Industrial haaarda of plaatlca aad avnthetlc alaatlmara. pp. 65-78. Lleei Now York. OTT KJ, LANCNEft ftft, HOLDER PS. (1975) Vinyl chloride aapoauva la a controlled Industrial eavlrooment. Arch ftavlron Hlth. 30. 333-339. PIESIE C, TASSICNON JP, PEINZN R, SPCLKENS J. (1979) Itnda da U mortmllte ebas dec travallleora exposes an chlorara do vlnyla. Arch naladlaa orofaaalonallaa da medicine da travail at da la Securlts Soclala. 40. 1131-1145. ftOTAL SOCIETY STUDY CftODP (1983) ftlak Aaaaaamanti Report by a Study Croon* Royal Society, London. Inferences contds 47 SAKIC K KVLCAR 2, Z01XCA M, GELlC J. (1976) Malignant tutors of the liver and lungs lo an area with o PTC Industry. Environ Hlth Perspsct. 17. - I09-92* 90QTAK CA> (1960) tpldenlologleal study of respiratory diseases in worker* exposed Co polyvinyl chloride dune* Thotii. >3. 644*632* TABEKSBAtf n. 4 arm n. (1974) Mortality study of workers In the enufeeture of vinyl chloride end Its polynere. J Occup Med. 16, 309-316. THEJUA01T c. 4 *tT Ar> p. (1981) Csncer uortsllty of group of Cenedlen workers exposed to vlnylehlorlde sower. J Occuo Med, 23. 671-676. TOURt CA, PATHS' A? 6 MUST Hi. (1964) Determination of Ambient Levels of Vtnvl Chloride Monomer (TOO round Mennfecturers In the OS. Pert 7, 1964. Warren Sprint Laboratory, Department of Trade sod Industry, Stevenage. VIOLA PL, IXGOTTX A, CAP0T0 A. (1971) Oneotenle response of rat skin, lungs, sod bones to vinyl chloride. Cancer tee. 31, 516-319. UAXWIILE* S3, STUMCEK tf, UACOMXX JK, JOKES J, FALK I, CAATEt C. (1976) Keopleselc risk among workers exposed to vinyl chloride. Ann HT Acad 6cl. 271. 40-46. 48 Tabla 1 KUMBEES OF DEATHS ROM DITTEttN? CANCEXS 0B5E1VED AND EXPECTED: FOCI STUDIES Cauaa of daath USA Ob**/Exp UK Ob*/Exp Canada Oba/Exp Canear of buecal cavity 4 pharynx oaaophagua atomaeh largo lntoatlna raetun livar Hoar 4 gallbladdor paneraaa othar dlgaatlwa larynx lung othar raaplratory bona akin (non-nolanona) nalanoma prostata taaeia bladdar kidnay othar & unapaeiflad urinary brain aya 4 CHS thyroid lynpho- 4 ratleuloaareoaa Hodgkin*a dlaaaaa laukaaaia ultlpla ayaleaa othar lynphatie othar eaocara 13/11.35 7/8.07 11/16.01 21/28.79 * - 39/5.77 17/18.40 118/115.87 S/6.38 2/1.81 6/7.36 16/15.20 - 5/8.46 12/9.06 25/12.76 12/7.98 3/5.45 14/13.94 ) ) 11/8.37 46/40.50 4/3.58 6/14.34 26/23.91 9/13.94 11/10.49 11/1.94 - 7/9.88 4/2.21 81/92.12 * - 2/1.74 12/9.59 2/1.38 14/8.00 3/4.10 3/4.29 4/6.18 2/0.43 4/2.35 3/2.50 7/5.16 2/2.35 18/8.12 ) > ) ) ) ) ) ) ) ) ) ) ) ) 0/0.64 - - '8/0.14 6/5.26 2/5.78 - 1/1.33 0/0.60 - 1/1.67 2/0.95 All eaneara 383/341.73 235/228.60 20/16.37 Italy Ob*/Exp 1/0.8 1/0.6 3/3.0 0/1.2 _ 1/0.6 - 0/0.7 - 1/0.9 12/6.1 - 0/0.2 1/0.7 - 1/0.7 - 0/0.4 0/0.7 9/4.5 30/21.1 *Obsarv#d daatha wultlpllad by i.0874 and roundad off to naaraat iatagar to allow / for duths without diieovtnd cun URL 13648 Table 2 hithbeis or deaths non hoh-malicnaitt and all caoses: FOCI STUDIES Cause of death USA Ob**/Exp Benito 4 other unspecified tUBOurs 4/3*08 Cerebrovascular disease 75/91*93 Isehaeaic heart disease 321/597.73 Other circulatory disease 157/123*55 Bronchitis** Pneunoni* Ocher respiratory disease 44/22.83 16/31*941 15/32*64] Cirrhosis of liver 37/54*06 Other dlgeetlve dlseeoe 27/39*52 Dlsesse of gsnlto-nrlnary systen 11/20*41 Other diseases Suicide Accidents 4 other violence 67/115*68 49/52.781 130/173.19/ All non-aslignant ceusee All ceases 1153/1363.54 1336/1705.27 UK Obs/Enp . 1 276/268 10S/14I J 36/44 40/62 3/3 - 43/76 40/50 543/663 780/894 Ohs/Exp 25/31.67 6/3.21 +4/3.83 * ^2/3.40 2/10.58 39/54.76 39/71.07 Italy Obs/Exp ' 0/0.5 - 19/27.4 3/5.0 4/5.2 3/2.7 2/7*0 1/0.9 4/7.7 36/36.4 66/77.5 * See footnote to Table 1 **Eapbyeeae la US data * Include* 2 certified ee cirrhosis of the liver which proved to be anglossrcoaa of liver Include* 1 eeuee unknown 50 Tabla 3 MOETALITT TKM CAHCEt OF THE LIVE* AND OTHE1 CAUSES IN VINYL OaORlDt HOUSES IB 49 FUNTS HI 4 STUDIES Ho. of daatha Obaarvad Expactad Staadardiaad mortality ratio (SHE) Cancor of llaor* " of othor altaa Othor dlaoaaoa Aeeldaae*. polaonlaga, and aiolanea All eauaaa 39 609 1547 226 2441 a.45 $99*35 1044.89 295.15 2747.84 696 102 $4 77 89 *XncIudlBS eaocara of tho gallbladdar to tho US tri* URL 13650 51 Tabla 4 MORTALITY FROM VARIOUS CANCERS IN VINYL CHLORIDE WOKERS IN 49 PLANTS IN 4 STUDIES Soure of Information Typo of caocar No, of daatha Obaaraad Expaetad SMR 1,2.3,* 1,2*3,4 1,2,3,4 1,2,4 1,2,3.4 2,4 1,2,3 2 1*2,3,4 1.2,3,4 1,2,3.4 Ctaetr of south and pharynx m " diROOtlVO OROtOB (othor than Uvar) " " raapiratory ayataa * * lu&t " " fanlto-urlnary ayataa Halaooaa Caoear of brain " " thyroid " " lymphatic and haanatopoiotie ayataa Othor eaneora All easeara otbar than llw IS 16.37 109 123 154.39 81 223 229.36 97 211 214.09 99 70 62.81 111 2 1.94 - 29 19.34 148 2 0.43 - 57 30.87 112 83 63*24 131 409 599.35 102 1 US atttdy 2 UR atody 3 CiBailao atody 4 Italian itudy 52 I c l4ic* nd e t n a in N o rw |la a study 53 Table 6 MORTALITY FROM LUNG CARCEE 1R THE OS AMD OK SERIES It CHARACTERISTICS RELEVANT TO AR OCCUPATIONAL HAZARD Data eheracterlatle Ho* of deaths la category 1 *Obs. Exp* 2 Oba. Exp. SMR category 12 Observed 20 yrs or nor* after flret esployeent (1), othere (2) Employed 10 yrs or sore ta USA (1), others la USA (2) Employed before 1954 la UK (l) others la UK (2) Ever eaployed as autoclave worker la UK (l) others la UK (2) 114 113.96 85 93.83 100 91 S3 52.45 63 63.44 105 99 52 51.39 29 40.50 101 72 16 17.08 65 74.82 94 87 *See footnote to Toblo 1 for obientd deaths in USA* r 54 TabU 7 NOETALITT FIOK SELECTED NOR-MALIGNANT CAUSES AND ALL CAUSES IN FOCI FKINCIFAL STUDIES COMBINED Sourea of Information Type of eaacsr 1.2 1.2 1**2*,3,A 1.2 1.2 1.2.3,4 1.2.4 1,2,3,4 1*2,3,4 1,2,3,4 1.2,3,4 1,2,3,4 ^Bronchitis, aaphysaua Other respiratory dlsaasa All respiratory dlsaasa Ischaaulc heart dlsaasa Other circulatory All circulatory dlsaasa Cirrhosis of Hoar Other disease All aoa-uellsaant disease All asternal causes All uou-uallsnant causes All eauses No. of deaths Observed Expected SMR SO 71 ISO 7*7 252 1103 44 23S 44.83 125.78 200.82 885.73 244.55 1209.35 44.24 348.07 120 54 80 90 95 91 49 45 1547 224 1844.50 295.15 84 77 1773 2441 2139.45 2747.85 85 89 1 OS study 2 UK study 3 Csoadlss study A Italian study + Bronchitis la UK study, caphyscua la os study **Ineludea cerebrovascular dlsaasa ia UK and Italiaa stadias t>$9v 55 Tiblt 8 MORTALITY FROM CHRONIC OBSTRUCTIVE LONG DISEASE* n CHARACTERISTICS RELEVANT TO AN OCCUPATIONAL HAZARD Dot* characteristic No. of deaths in category 12 Ohs. Exp. Ohs. Exp. SMR category 12 Obaantd 20 yrs or wta afttr first employment In USA (1), others In USA (2) Employed 10 years or noro in USA (1), others In USA (2) Employed before 1956 In UK (1), others in UK (2) +Evr eaployed os sutoclsre worker In UK (1), others In UK (2) 30 13.8 11 7.0 190 157 18 10.8 25 12.0 147 20828 30.17 10 13.60 86 74 3 8.53 33 37.22 46 89 *D*ierlb*4 u ph7HU is US itudy and bronthltli la Of itudy. +H*n ever enployed as a bagger or drier, which wonld hare caused the grtiteit oeevpetieul wpeture to PVC dtut, nptrineed 1 dooth froo 'bronchitis' etelMt 4.88 npeettd* Ss9 iia n