Document 91eQK96rEbzXgxp16z4de9xY6

British Journal of Industrial Medicine 1985;42:750-753 Exposure to vinyl chloride and angiosarcoma of the liver: a report of the register of cases D FORMAN,1 B BENNETT,2 J STAFFORD.2 AND R DOLL1 From the Imperial Cancer Research Fund Epidemiology and Clinical Trials Unit,1 Radcliffe Infirmary, Oxford OX26HE, and Imperial Chemical Industries pic,1 Thornton Cieveleys, Blackpool FY54QD, UK abstract The Association of Plastic Manufacturers in Europe maintains a register of all cases of angiosarcoma of the liver (ASL) resulting from exposure to vinyl chloride monomer (VCM). This register has recorded all known VCM related cases of the disease worldwide that have been histologically confirmed. Although likely to be incomplete, it is the most comprehensive tabulation of such cases available and, by the end of 1984, 118 men had been registered. The information from analysis of the cases currently on the register is used to make an approximate projection of the number of cases to be expected in the future. The conclusion from this projection is that the industry is reaching the halfway stage in the appearance of VCM related ASL. The idea that vinyl chloride might cause cancer was first suggested by Viola in 1970 as a result of experi ments in rats.1 This was subsequently confirmed by Maltoni in 1973 with his comprehensive series of experiments in rats, mice, and hamsters.2 Among the tumours that developed in these animals was the nor mally rare angiosarcoma of the liver (ASL). It was not. however, until three cases of ASL were observed in men who had been heavily exposed to vinyl chloride in a polyvinyl chloride (PVC) polymerisation plant in the United States that the reality of the hazard was generally accepted.'5 ASL is normally so rare that only three or four cases occur each year in the whole of Britain (sec below) so that the occurrence of three cases in a small group of industrial employees in the United States was. in itself, sufficient to justify the belief that the chemical caused the disease. If. how ever, any doubt has persisted it was soon dispelled by the steady accretion of new industrial cases and the finding that angiosarcomas and various other tumours could be produced in rats by levels of exposure that were appreciably less than those that had previously occurred in industry.1 Steps were therefore quickly taken in the mid-1970s to reduce the extent to which individuals could be exposed from levels of the order of several hundred ppm to 5 ppm (13 x I06 ngm/m2) or less.3 This reduction followed Received 25 February I9S5 Accepted 25 March 1985 an earlier lowering of exposure levels in the mid-1960s after concern was expressed about the narcotic effects of vinyl chloride and the risk of it producing acroostcolysis. Before then exposure to some employees was often in the range of several thousand ppm. The rarity of ASL. even under conditions of heavy exposure, has made it difficult to study its epi demiological characteristics11 and it seemed, therefore, that it might be of value to summarise the data that have been recorded in the register of cases that has been maintained by Imperial Chemical Industries pic on behalf of the Association of Plastics Manu facturers in Europe (APME). This tabulates informa tion about all the cases of ASL known to have occurred in men exposed to vinyl chloride that have been reported by individual manufacturing compa nies, trade associations, and occupational health organisations in Europe. North America, and Japan. The register is, doubtless, incomplete. It includes. Table I Numbers of vinyl chloride related cases of angiosarcoma of the liver by date of death Period 1955-9 1960-4 1965-9 1970-4 1975-9 1980-4 1955-8-4 Ho of death 2 A 9 23 46 31 115* - Three men diagnosed in 1980--4 were still alive at I January 1985. 750 R&S 001549 of the arv. Oxford ill cases of CM). This have been tabulation ation from ion of the le industry i mtiil^fcos cotJictT^Hfectts ucing acro employees J ppm. The . of heavy dy its epiI. therefore, e data that es that has dustries pic ics Manues informam to have e that have ng compamal health and Japan. It includes. es of Exposure to vinyl chloride and angiosarcoma of the liver 751 Table 2 Numbers of vinyl chloride related cases of angiosarcoma of liver by age at diagnosis Age t years i 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 All ages So of cases II 24 17 18 :o 17 8 3 113 however, all the cases known to the collaborating organisations and they will probably have been informed, in one way or another, about nearly all cases that have been diagnosed among the relevant groups of employees. Results Until the end ol 1984, 118 men had been recorded as having vinyl chloride related ASL. or whom three were still alive on I January 1985. All the cases had been confirmed histologically: 117 were reported to be hepatoangiosarcomas and one a cholaneiosarcoma. No cases were recorded in women, but few women have been employed in manufacturing PVC. Table I shows the breakdown of the cases by date of death. It should be noted that the register was begun in 1974 when vinyl chloride first became firmly established as a risk factor for ASL. Under-reporting is. therefore, less likely to be a problem after this date when special attention began to be paid to the disease. Table2gives the age distribution of the I I8menat the lime of diagnosis. The mean age at diagnosis is 52. Table 3 gives the number of cases diagnosed and the number of PVC producing factories in each of the 12 countries reporting at least one case of ASL. Some countries have had cases in most of their factories (France. West Germany); others have had cases clus tered in a small number of factories (United States. United Kingdom). Indeed, the United States has had 23 of its 33 cases in just two factories. Other countries that produce PVC but have not reported any cases are shown in jable 4. Table 5 shows the numbers of cases by time since first exposure and maximum exposure time possible. It seems that the incidence of ASL reaches a peak 20-29 years after first exposure but substantial num bers of cases may still occur after more than 30 years. The table also gives hypothetical numbers for expected new cases based on the distribution of cases in men who could have experienced the maximum latency period. These calculations show that the occurrence of vinyl chloride related ASL is not yet at the halfway stage and one might reasonably expect another 150 cases or so with an unknown number of further cases occurring 35 years or more after first exposure. Table 6 shows the principal occupations of the 118 affected men. At least 43% of the men (51/118) had been employed as autoclave cleaners, an occupation that has been thought to have necessitated the highest exposure to vinyl chloride monomer. Discussion The reassuring feature oflhesc results is that the num ber of cases on the register is so small, and. in partic ular, that only 80 new cases have been diagnosed since 1975, despite the large size of the industry. In the decade 1975-84. 19 cases were registered from the United States and seven from the United Kingdom. In men aged over 15 about 3300 deaths a year are attributed to all types of liver cancer in the United States and about 550 a year in the United Kingdom. Many of these cancers arc probably sec- Table 3 Distribution ol vinyl chloride related cases of angiosarcoma of liver by country. (Number offactories with three or more cases in parentheses t Country So of cases .Vo offactories reporting cases Total So offactories making P I'O Pre+1974 Pre-1964 Capacttv* in 1972 United Stales W Germany France Onada United Kingdom Sweden Italy Yugoslavia Czechoslovakia Japan Belgium Norway All countries reporting cases 35 26 18 10 9 5 4 4 T 2 T 1 1 IK 10 (3) 6 (31 5 (3) 1 (I) 2 (1) 1 (I) 4 (--) I (|) 1 (") I <--) 1 (--) i (--) 34 (13) 44 9 7 4 7 M 1 1 31 y 1 122 28 7 5 3 4 1 7 1 16 1 1 76 `Capacity in kilolonncs pvc production a year (information supplied by ICI and by recipients of the register). 2090 1155 627 xs 502 105 778 60 4X 1699 195 65 9462 IM R&S 001550 752 Forman. Bennett. Stafford, and Doll Table 4 Countries utth Pl'C factories operational before 1974 with no cases reported to register Country Total t\'o of lactones AW 974 Pn-1964 Western Europe Austria Finland Greece Holland Portugal Spam Switzerland Eastern Europe Bulearia CjsI Germany Hunearv Poland Rumania USSR S`uuf/i Aniermi Meenuna Brazil Chile Colombia Mexico Peru 1 1 l 2 1 4 1 1 2 1 2 3 3 1 5 1 1 - -- 1 1 3 1 _ 3 1 l j 1 ' 2 -- 2 -- Source: Information supplied by ICI and by rccipicnis or the register, ondary liver cancers originanng in olher organs, par* ocularly in men over 60. but if this is ignored n appears that only about 0 06% of liver cancers in adult men in the United States (19/3300 x 10) arc attributable to vinyl chloride and about 013% in Britain (7 '550 x 10). Two pieces of evidence suggest that an extremely high exposure to vinyl chloride may be needed to induce ASL. Firstly, the fact that so many of the cases were among men working as autoclave cleaners in which the highest exposures arc thought to have occurred: it was in these jobs that employees some- times became faint and unconscious before the 1960s. when the main concern about vinyl chloride was its narcotic effect. Secondly, in some countries there is a remarkable clustering of cases into a small number of factories. This suggests that the working conditions and hygiene standards varied in different factories. Fifteen of the 16 Japanese and 19 of the 28 United States factories operating before 1964 have reported no cases, and it would be helpful if it could be shown that the exposures in these factories was appreciably less--unless, of course, the number of employees in these lactones were all very small. It is the rarity of this tumour that has enabled its rapid identification as an effect of exposure to vinyl chloride. Baxter et al have calculated that the back ground incidence of ASL in the United Kingdom is between two and seven cases a year--that is. cases unaitributablc to exposure to vinyl chloride or to thorotrast. the other established aetiological agent.' That being so then the seven industrial cases in the decade 1975-84 represent an increase in the national figures of between 10% and 35%. The mean induction period (time from first exposure to initial diagnosis) for the I IS cases is 22-6 years. Most cases have occurred 15 to 29 years after first exposure as is common with occupationally induced cancers. There are also, however, substantial numbers of cases occurring after longer time intervals. 18 having been diagnosed 30 or more years after first exposure. This means that new cases are likely to be diagnosed well into the 1990s even if it is assumed that exposures were significantly reduced in the mid-1960s when the first major improvements in standards of hygiene were made. The mean induction period of the cases currentlv on the reeisicr <22-6 vears) must thcre- Table 5 Cases of vinyl chloride related angiosarcoma of liver tabulated by years from hrst exposure to vinyl chloride to diagnosis and vears to SI December 1934 fmaximum possible latent period). (Sumber ol cases expected to occur shown in parentheses' I Years from first exposure to SI December !9H4 Years from first exposure lo Jute of diagnosis s- 10 - 15 - S35 30-34 25-29 20-24 15-19 0-14 Total __ _ 1 1 to 1 4 12 3 10 8 --, 4 15 2 ( + 25) _1 _ 2 ( + 7) _0 1 + 17) 1 12 32 ( + 7) 32 ( + 42) 24 - 30-34 12 7 4 (+9) 0 ( + 29) _0(4-18) 23 ( + 56) 7 3<+4) 0( + l0) 0( +22) 0 ( + I4| -- 10 ( + 50) Total within J5 \ears 32 27 ( + 4) 25 (+ 19) 2) ( + 76) 5 ( + 56) 0(+?)t 110 (+ 155) J5 or more 8 ' 8 (+?i: `Number ol`vases expected lo occur calculated by comparing proportion ol'cases m relevant years since tirst exposure category lor previous time periods wuh proportion in other categories--lor example. Tor 25-29 years since nrsl exposure and 25-29 years maximum latency expected: (3 + 10 + 8) <2+I+ 10+1+4+TM (12 + 7) x 19 13.3 As four cases have already been diagnosed, another nine mav be expected. (Number ol new cases to be expected among workers who had less than 15 years since first exposure at 31 December 1984 will be dependent on whether the hazard of vinyl chloride was ehrmna ted tn the late 1960s or m the mid-1970s. In the former instance no cases would be expected, in the latter a number approaching the 61 cases projected for those wuh 15-19 years since hrst exposure by the end ot 1984. Jit is virtually impossible to estimate the number of expected new cases 35 or more years after hrst exposure. 4*<*- *p' R&S 001551 ,/tc; Doll c 1960s, was its rierc is a mber of nditions ictories. United -eported e shown reciablv ovees in ibled its to vinyl ie backgdom is is, cases r to thot.' That decade I figures m first s ts 22-6 trs after tionally istat^^k I'ter first Iv to be ned that d-1960s lards of d of the a there* tde shown in \r mow )r previous y expected: dependent e expected. Exposure to vinyl chloride and angiosarcoma ol the liver 753 Table 6 Recorded occupations among ASL cases on register So* Production workcr/operator Autoclave cleaner General maintenancejritter Other jobs Unknown 69 51 3 11 17 156 *As several employees have more than one job recorded for them, these numbers do not add up to I IS. fore be expected to be materially increased with the passage of time. Any method to assess the number of expected new cases is likely to be imprecise. Our method simply looks at the proportions of cases with different latent periods and makes projections based on the distribu tion of cases that have occurred within different peri ods of maximum possible latency. In the absence of any information about the numbers of exposed employees it is all that can be done. The results suggest that another 155 cases are likely to be diagnosed within 35 years after first exposure. Compared with the 110 cases in this category diagnosed so far. this means that the industry is approaching the halfway stage in the appearance of ASL. The calculation of 155 more cases within 35 years of first exposure assumes that the hazard was essentially eliminated in the late 1960s. If it was not a further 50 or 60 cases (something slightly less than the number expected to have been induced in men who could have been observed for only 15 to 19 years by the end of 1984) must be added to the total. Further cases must also be expected to occur 35 or more years after first exposure but it is. as yet, virtu ally impossible to estimate the number to be expected with any precision. It could, however, easily be around 50. We suggest, therefore, that it should be assumed that there will be another 200-250 deaths from vinyl chloride related ASL over the next 30 years. This estimate is in pronounced contrast with that calculated by Nicholson et a! who suggest that "within a factor of 2 or 3" there will eventually be a total of 350 deaths from vinyl chloride related ASL in the United States and 1200 in Western Europe alone.8 Nicholson el al used a complex model to deduce these figures based on the assumption that the incidence of ASL at any point in time increases in proportion to (i) the first power of the product of the population at risk and the degree of exposure and (ii) a higher power (most probably the third) of the time since exposure first occurred.8 Knowledge of the distribution of cases recorded by the National Institute of Occupational Health and Hygiene (1982) by year of first exposure and year of occurrence, the amount of vinyl chloride produced, estimated average exposures at work at different periods from before 1950 to 1970-4, and normal national mortality rates allowed them to make approximate estimates of the variables in these equa tions and hence to calculate the number of cases that might be expected to occur in the future. Only time will show which of the two methods will give the better estimate of the number of cases that will eventually be observed. We are encouraged, how ever. to think that the lower estimate may be the nearer as the method of Nicholson et al predicted an increase in the annual number of cases after 1979, whereas the number recorded to date has declined from an average of 9-2 in 1975-9 to an average of 6 8 in 1980--4. The latter figures will, however, almost certainly, be revised upwards in the course of the next few years owing to delayed reporting. References | Viola PL. Pathology of vinyl chloride. Med Lav 1970:61:174:80. ! Maltoni C. Occupational carcinogenesis 2nd International Sym posium on Cancer Detection and Prevention, Bologna. 1974. In: Advances in tumour prevention, detection, and characterisation. Vol 2. Amsterdam: Excepla Medica. 1977. J Creech JL, Johnson MW. Angiosarcoma of the liver in the manu facture of PVC. J Occup Med 1974:16-19-26. 150-1. Maltoni C. Lcfcmine G. Carcinogenicity bioassays of vinyl chlo ride. Environ Res 1974:7:387--105 ' Bames AW. Vinyl chloride and the production of PVC. Pro ceedings o/ the Royal Societv ol Medicine 1976:69:277-80, * Fox AJ. Collier PF. Mortality experience of workers exposed to vinyl chlonde monomer in the manufacture of polyvinyl chloride m Great Britain. Sr J Ind Med 1977:34.1-10. Baxter PJ. Anthony PP. MacSween RNM. Schaucr PJ. Angio sarcoma of the liver: incidence and aetiology in Great Britain. Be J Ind Med 1980:37:213-21. * Nicholson WJ, Hennebergcr PK. Tarr D. Trends in cancer mor tality among workers in the synthetic polymers industry. In: Industrial hasards ofplastics and svnthettc elastimers. New York: Lis. 1984:65-78. 734 THE LANCET, APRIL 3, 1976 Occasional Survey GENETIC RISKS OF VINYL CHLORIDE I Peter F. Infante Joseph K. Wagoner Anthony J. McMichael Richard J. Waxweiler Henry Falk Division ofSurveillance, Hazard Evaluations and Field Studies, National Institute for Occupational Safety and Health, and Bureau of Epidemiology, Center for Disease Control; and School of Public Health, University ofNorth Carolina Summary A study of pregnancy outcome among wives of workers exposed to vinyl-chlor ide monomer (v.c.M.) indicated that, in comparison with controls, there was a significant excess fetal loss in the group whose husbands had a primary exposure to v.c.M,, whereas no differences between the groups were observed before the husbands' exposures. The difference in fetal death-rates for the post-exposure comparisons was a reflection of a greater fetal loss associated with the wives younger-aged husbands. The significant excess did not seem to be the result of bias from interviewers, re spondents, nor from women who had experienced chronic abortions weighting the results. These findings, in conjunction with the demonstration of a mutagenic response via microbial test systems and with observa tions of significant excesses of chromosomal aberrations among workers exposed to V.C..M., raise scientific 3nd public-health concern for the possible genetic risks of v.c.M. to man. Is the past year, several reports have indicated that vinyl-chloride monomer (v.c.M.) is mutagenic in micro bial test systems.1-3 v.c.M. metabolites also have in duced mutations in mammalian cells.4 Likewise, reports from four countries have shown an excess of chromoso mal aberrations in lymphocytes of workers exposed to v.c.M. compared with controls.5-* However, Purchase et a!,7 have stated (though no animal data were presented), that the mutagenic effects of v.c.M. expressed as chro mosomal aberrations in lymphocytes in humans do not occur in germ cells in mice; they concluded that the potential danger of mutagenic effects on the fetus via sperm seemed unlikely to exist. Tn a study without con trols, Sclikoff observed fetal death-rates among wives of v.c.M. workers that ranged from 7 to 14 per 100 preg nancies.' These rates appear to have been higher than expected.10 To develop further data on this question, pregnancy outcome has been studied among the wives of workers exposed to v.c.M. All current v.c.M. polymerisation and polyvinyl-chloride (p.v.c.) fabrication workers were in cluded for study together with a similar number of cur rent rubber workers (8% of all such workers) selected from work areas relatively free from known toxic materials and matched as a group to the v.c.M. workers by age. Group-participation rates ranged from 62 to 77". Data for the wives of v.c.M. polymerisation workers (primary v.c.M. group) were contrasted with data for the wives of P.v.c. fabrication and rubber workers ("controls"), who were known to have had verylow or no v.c.M, exposure, respectively. A total of 95 v.c.M. polymerisation and 15S rubber and P.v.c. fabri cation workers were interviewed. Paternal age, preg nancy outcome, and estimates for the time of conception of all pregnancies were ascertained by interview in Oct ober, 1974, from males employed at a rubber manufac turing, p.v.c. fabricating, and v.c.M. polymerising facil ity. As part of a larger survey of worker health, date of first employment in the job categories was determined from company records. Mean paternal age, total number of conceptions, total number of fetal deaths (defined as any product of conception not born alive), and fetal deaths per 100 conceptions were then computed for each group prior to and subsequent to the worker's date of employment. No interviews were conducted with workers' wives-and no data were obtained concerning maternal age, except indirectly through paternal age. Since fetal loss is known to increase with ascending parental age, the fetal death-rates for the primary v.c.M. exposure group were age-adjusted to the control group. Table t snows the age-adjusted fetal death-rates for wives of the primary v.c.M. exposure group versus the control group, both prior to and subsequent to each group's respective exposures. Among pregnan cies occurring prior to exposure, fetal death-rates were 6-9% for the controls versus 6-1% (age-adjusted) for the primary v.c.M, exposure group. These rates were not significantly different by Mantel-Haenszel Chi-square testing.1' Among pregnancies occurring subsequent to the husband's exposure, the difference in frequency of fetal deaths between groups was significant at p<0-05 (y_7=4 00, df=l),17 Although the underlying dis tributions differed, mean paternal ages were virtually the same--30-4 versus 30-2 years. The significant dif ference between the groups subsequent to exposure was a reflection of a relatively greater fetal mortality-rate as sociated with younger-aged husbands in the primary v.c.M. exposure group. Among pregnancies occurring ! subsequent to exposure, the fetal mortality-rates associ ated with husbands 30 years of age and older for the pri mary v.c.M. exposure and control groups were 9/69 (13-0%) and 17/142 (12-0%), respectively; whereas, for R&S 001552 J Ji /a t* {j. fjr TABLE !--MEAN PATERNAL ACE, NUMBER OF PREGNANCIES, ANO FETAL death-rates according to husband's v,c. exposure -- Primary `Controls"4 v.c.M. eaposurej Prior ft hutbond'i exposure Number of families .Mean paternal age at conception (vr.) Number of fetal deaths among wives Number of pregnancies Age*aJ|usied fetal. deathsdlOQ pfcg4 SutKjiulni ru huiband'i exposure; Number of families Mean paternal age at conception (vr.) Number of fetal lh,*jihi among w-jves Number of pregnancies Age-aJiustcd fetal Ocaihs/tOO preg4 95 23-0 11 159 4-9 IIS 30-4 24 273 !$ 70 26-4 15 MS 6-1 62 30-J 23 I 139 j 15 <i 1 ----" I lubber afhj p,v,c. fabrication worker*. \C` paKnicmaiion worker*. jinn ag:-adiustcd to "control" group paternal age distribution, ubxquent to husband'* exposure, the frequency of fetal death* amonf i* significantly greater in the primary v c.M, capoiure group than n the >t* rt.0 05} or n the study group prior to husband's cipoturc (rcO 0* chi-square tceirCt^.** ` JjIV,t ii t: fr, * *i l t * -4 l .-i. ~ TJ.E L.1NCET, APRIL 3, 1976 755 -- HEAS paternal ace, number of pregnancies, and fetal -RAHS ACCORDING to husband's v.c, exposure excluding Pregnancies of uomev with * 5 fetal death's . .h, jtjfrjrij't expirin'*-* ^uvBrtfiton (yr.) amcng wivo Njmbcr of p^^nannes jcjihs'lOO pfeg4 <** ta nusbcnd's exposure: vrJa paternal *5* i*. ivnccpiion (yr.) ^-S:f of fetal among wives Number of p^jti jncio jJ|Uil(\J fetal ij<hi#l00prcg^ R.J^rr anJ r.v.c. fabrication workers. * -wNcrsensatiofl workers. ; ^ ^--jJiusted to ''control" paternal age distribution. huxbands less than 30 years of age, fetal mortality was 1-i TO (20 0%) for the primary v.c.M. exposure group ^mpared with 7/131 (5-3%) for the control group these data are not shown in tables.) Furthermore, intragroup comparisons indicated an ir.erease in age-adjusted rates for the primary v.c.At, o exposure group from 61% before exposure to 15-8% if v.:Scquc-ru to the husband's exposure. This difference u wus significant, P<002 (-/3=5-51, df=l).lJ Similar S* ii'mpjrison for rates in the control group, 6-9% versus ly S-S'I, indicated no significant difference. f- To determine whether women who had chronically as experienced abortions might have weighted the results in is- favour of a higher fetal death-rate in the primary v.c.M. *y pv'-up subsequent to husband's exposure, pregnancies of ng -omen who had more than two abortions were elimci- i.-.jtrJ from the analyses and the data were recalcu in* lated to determine whether or not the trend was main 69 tained. The decision to exclude all pregnancies among for families associated with more than two abortions was made without prior knowledge of how these families ::r distributed among the exposure categories. The data in table It show that the trend was maintained, fttor to exposure, the fetal death-rates in the control j.-J primary v.c.M. exposure groups were 6-9% and ' IT luge-adjusted), respectively, whereas, after expo- the rates were 6-8% and 10-8%, respectively. Sub sequently, data were eliminated for pregnancies of -'men who had experienced, firstly, more than one i'.'-.'tinn, and, secondly, more than three abortions, and <j-h time the trend was maintained. No changes in rates cuntrols were observed, whereas a 2--3-fold increase "4I observed in the primary v.c.M. group subsequent to rT*.n-jrc, To determine whether differences in fetal loss might -J't been the result of one or two interviewers weight- { :-.e results, the data were analysed by individual in r rt *cr. The results demonstrated a general trend for mterviewer to report a higher ascertainment ' V-C.M. polymerisation workers than among the "'-'ol group. ! `^cr, the possibility was entertained that the inter- o o: t *- .e-.ueen the date of interview and the date of fetal M m.-gnt have influenced the results through dif- ferences in recall. The interval, however, was estimated to have been about two years less for controls, suggest ing that, if a bias did exist, it would have been towards a greater ascertainment in the control group. In some cases, the worker failed to indicate the ages of his children and in other cases he was unable to recall the approximate time of his wife's abortion; therefore, the data were analysed to determine the distribution of fetal death-rates among the respondents in each occupational group who did not complete the interview properly. The difference in fetal death-rates between groups was very slight. Finally, the workers may have been subject to bias resulting from prior knowledge of known hazards of vinyl chloride. However, the workers themselves did not always know into which of our employment categories they were being allocated: For example, several p.V.c. fabrication workers who were included in the control group thought that they had a primary v.c.M. exposure as a fabrication worker. In addition, the questions regarding pregnancy outcome were contained in a much larger interview-questionnaire, the results of which demonstrated very few significant differences with no consistent bias for the parameters ascertained between the workers with a primary v.c.M. exposure, and the other groups. This observation as well as several others presented above tend to support the validity of the study. In summary, a significant excess of fetal loss was observed among wives of workers following exposure to v.c.M. The excess did not appear to be the result of bias from interviewers or respondents, nor from women who experienced chronic abortions weighting the results. Several mechanisms by which such fetal loss may arise are suggested. Either fetal or maternal toxicity or germcell mutagenesis in the mother through indirect v.c.M. exposure from the father might be considered, although these mechanisms seem highly unlikely in view of the highly volatile nature of v.c.M.1J When the findings of the present study are taken in conjunction with the prior demonstration of a mutagenic response via microbial test systems and observations of significant excesses of chromosomal aberrations among workers exposed to v.c.M., the leading possibility is germ-cell damage in the father through direct v.c.M. exposure. The increased fetal mortality among wives of workers subsequent to v.c.M. exposure now raises serious scientific and publichealth concern for the possible genetic risks of vinyl chloride to man. Request* for reprints should be addressed to P.F.I., N.I.O.S.H.* Post Office Building, Room 515. Cincinnati, Ohio 45202, U.S.A, REFERENCES 1. Hartseh, H.. Malavielte, C., Montcsano, R. fnt,}. Cancer, 1975, 15, a29. 2. I^oprieno, N\, Karate, R., Varoncelli, S.. cl al. Afuurron Rei. (in the press). 3. Rannug, U., Johansson, A., Ramel, C., Waehmeistcr, C. A, Ambur. 1974, 3, 194. 4. Huberman, E., Bansch, H., Sachs, U Int.J. Cancer, 1973, 16, 639. 5. Ducatmin, A., Hirschhom, K.. Sehkoff, I. J, Mutation Ret. 1973, 31, 163. 6. Funes-Cranoto, F., Lambert, R,, Lindnen, J,, Ehrenberg, L., Natarajan, A. T., Osterrrsan-Colkar, S. Lancet, 1975, i, 459. 7. purchase, 1. F. M.f Richardson, C, R., Anderson, D. ibid, 1975, il, 4 10. S. Millesiad, L., Thiis-Eventen, E. Unpublished, 9, Schkoff, I. J.. N.l.E.H.S. Conference on Public Health Implications of Com ponents of Plastics Manufacture, Ptnehurit, North Carolina, July, 1Q74. 10, Infante, P. F. ,4#wi, iV.F. Acad, Sci. (in the press). 11, Shapiro, S.. Jones. E. W., Deruen. P. M, Milbank Q 1962, 40. 7. 12, Mantel, N,, Hienucl, .Vdis, Cancer Jntt 1959, 22, 719. I 3. United States Environmental protection Agency, sampling and analysis of select tone substances, task tit vmyl chloride. Contract no. 68-01-2646. Jan. 20, 1976. R&S 001553 i ' 1\ *41 M' V 9 l t ; i 4 s t- Uh i !