Document QMbbEk12X49OJn6JnVBBjZE65

RECEIVED JUL ? 3 13/a U. J. BARR MU TO: EP FROM; DATE: JULY 23, 1979 ; : Mutagenesis, ceracogenesis and cancer in the offspring of persons occupationally exposed to petroleum derived hydrocarbons . Report prepared for the American Petroleum Institute. June 1978 Joanna ?. Haas, M.Q and Assistant Professor of Public Health, Cornell University Medical College Assistant Attending Physician, Memarlal-Sloan Kettering Cancer Center David Schottenfeid, M.O. Chief, Epidemiology and Preventive Medicine, Memorial-Sloan Kettering Cancar Cancer Professor of Public Health, Cornell University Medical Collage Tbs. V'7 AP000126I5 tnutatis mutandis AP00012616 TABLE OF COMTr.rrS I. II, III. IV. V. VI. VII. VIII. IX. X. XI. Introduction Relationships between mutagenic and teratogenic characteristics of an agent and potential for carcinogenicity in the offspring of exposed organisms Evidence for cransgeneratlonal carcinogenesis in animals Associations between congenital abnormalities and childhood neoplasia Knudson's model of multistage mutacio s and carcinogenesis Use of peripheral blood lymphocytes for the analysis of chromosomal abnormalities Environmental threats to pregnancy outcome ~" a. Alcohol b. Smoking c. Drug exposure d. Radiation Occupational exposures and pregnancy outcome a. Vinyl chloride . 'f b. Benzene c. Anesthetic gases Cancer in the offspring of hydrocarbon exposed fathers Surveillance of spontaneous abortions as a strategy of environmental monitoring Further studies ! AP00012617 I. Introduction This review evaluates epidetaiologic evidence cor carcinogenic, teratogenic and mutagenic effects of petroleum-derived hydrocarbons in the offspring of occupationally exposed populations. Human studies are con sidered in the light of pertinent background information from the laboratory and from animal studies. Where epidemiologic investigations are reviewed, particular attention is paid to the strengths and limitations of the research strategy employed. The last portion of this paper considers the problems which remain unsolved and suggests realistic approaches to clarify chose Issues. Derived from the Greek work "repas,", meaning a marvel, prodigy or monster, the word ''teratogenic" was used by 1857 to mean the "production of monstrous formations or births" (OED, 1971). Since creation of monsters interests many disciplines, the term has been refined to refer to the bio logical science dealing with "the causes, mechanisms, and manifestations of developmental deviations of either structural or functional nature (Wilson, 1973). Such causes may be chemical, viral, pharmacological or physical. Staples (1975) pointed out that an agent may ace as a teratogen when admin istered in a number of ways, whether to the male or female before macing, to the female during pregnancy, or to the fetus directly. Strictly defined, agents which alter the race of growth of the fetus or are lethal to Che fetus without producing specific anatomic or functional anomalies are better termed developmental toxins chan teratogens. In this discussions of potential adverse effects of environmental agents, particularly polycyclic hydrocarbons and their derivatives, evidence for teratogenicity and for developmental toxicity will be considered. AP000126I8 -2Agencs in che environment may produce alterations in the genome, mutations, and by chat mechanism lead to abnormalities o development in che offspring. Viral, drug, or chemical agents acting through common mutagenic pathways may produce indistinguishable results. The outcome of such mutations .might be fetal defects or predisposition to neoplasi^. Cancer in che offspring of individuals exposed to a carcinogen might theoretically be the result of any one of three types of exposure: pre2ygotlc, by alteration of parental germ cells; transplacental, by passage .of carcinogenic substances across the placental barrier; postnatal, by contamination of che environment with carcinogenic substances, primarily through ingestion, including substances excreted' in human breast milk, and through inhalation. AP00012619 -3- II. Relationships between mutagenic and teratogenic characteristics of an a^enc and potential for carcinogenicity in the offspring of exposed organisms It has been argued that the "same chemical that causes abortion In the early stages may produce malformations during organ development and neoplasia when exposed later in pregnancy. Therefore, screening for transplacental hazards should Include, whenever possible, the entire range of fetal response, including'cancers that may develop sometime after birth" (Fraumeni, 1974). Mutagenesis, teratogenesls and carcinogenesis are related phenomena, but the nature of these relationships is complex and occasionally controversial. While a relationship exists between teratogenicity and carcinogenicity when a multiplicity of systems are considered, they should not be viewed as expressions of the same phenomenon, mediated only by the stage of development of the exposed organism (Di Paolo and Kotin, 1966). Teratogenic, mutagenic and carcinogenic activities are all demonstrable for a number of compounds, but transplacental carcinogenic potential has rarely been demonstrated for known teratogens and mutagens. Teratogenic effects may reflect a variety of actions involving the mother, the placenta or the feto-placental unit, which can be indirect and-may not Involve immediate action of the agent on the target fetal tissue. Rice (1973) Suggests that transplacental carcinogens do act directly on the fetal tissues which ultimately become neoplastic and that ocher effects, such as immunosuppression, play a secondary role. To appreciate potential damage of an environmental exposure to off* spring of the exposed organism, one ought well consider the outcome of all exposed pregnancies. Early fetal wastage commonly results from abnormal fetal development. At the other extreme, transplacencally induced tumors APOOO12620 -4- hardly ever appear in rodent: species until the animal reaches security. In the only documented example of chemical transplacental carcinogenesis in humans, vaginal adenocarcinoma following diethyscllbestrol (DES) exposure, the tumor occurs decades following exposure. Thus, since transplacentally induced tumors, unlike malformations, are rarely present at birth, one must also observe the offspring Into adulthood to see the full spectrum of result ing tumors. Fecal tissue sensitivity to a carcinogen may differ from adult tissue response. Although fecal mice may develop Lung cumors following transplacental exposures to the water soluble anesthetic agent ethyl carbonate (Urethan, II), they are much less prone to develop tumors than are adult alee. By contrast, 1-ethyl-l-nitrosourea (ENU, III) produces neurogenic tumors in mice offspring after a single transplacental exposure, but-Is far less active in the adult mouse (Rice, 1973). The nature of the tumors produced by in utero exposure to a carcinogen, is determined both by the gestational age of the exposed fetus and, most importantly, by the kind of cumors produced by che same agent In the adult of the species. This species or strain specificity doubtless reflects genetic factors. At the same cinrd some agents, which are carcinogenic in adult tissues, are not carcinogenic in fecal tissues, the latter lacking enzymes to metabolite the substance to its active derivative. In ocher circumstances, carcinogenic derivatives may cross che placenta and act directly on the feeus. A question of considerable interest is whether carcinogenic effects of a parental exposure may be visited on offspring subsequently conceived. Such a trans-genecacional effect would necessitate alteration In the parental germ cells involving one of three distinct mechanisms: chromosome breakage, point mutation, or abnormal ities In gamecogenesis or fertilization. AP000I262I -5 III* Evidence for crans-ganeracional carcinogenesis iq animals Trans-generational carcinogenic effaces have been reported in second generation descendants of mice treated with 7,l2-dimethylb*nr(a)antracene and of rats exposed to nitrosomethylurethan. TomatIs and Goodall (1969) exposed pregnane, mice to lncraperitoneal injections of 400 oicrograms of 7,12-dimethylbenz(a)antracene (DMBA) in olive oil and observed a significant excess of tumors in offspring exposed In utero, and also in the offspring of in utero exposed mice, the F2 generation. Oncogenic effects In the ?2 genera tion could, however, reflect persistence of DMBA into adulthood in animals originally exposed while in utero, and its excretion in milk and feces with resultant direct exposure of the F2 offspring. Germ cell alteration would be the alternative explanation, although tumor excesses in the generation and In other Inbred generations, which would be expects 1, were not demonstrated. Untreated offspring of rats which had been axposed in utero to nicrosomethyLurachan (NMU) were said to have a higher incidence of tumors chan expected in a strain with low spontaneous tumor incidence (Tanaka, 1973). The usual spontaneous tumors in this strain are mammary, 2.1%, and lung adenomas', 1.3%. Among the offspring of two rats whose mothers had been exposed In utero the frequency of tumors was 7.2% in males and 9.7% in females, with tumors In four and seven animals, respectively. Control animals had tumors in two males (2.1%) and three females (2.7%). The possibility chat ostensibly unexposed offspring might have had some exposure through NMU retained or excreted by the 'in utero exposed mothers was not specifically excluded. Among the un created offspring of in utero exposed females crossed with uncreated males there was no tumor excess, nor were such effects detected in the third genera tion. Trans-generational carcinogenic effaces of NMU exposure require further APOOOI2622 -6- . documentation and, as a phenomenon In general, crans-geaeracion chemical carcinogenesis requires confirmation. Such doubts were expressed In the discussion and conclusions of the meetlag bn transplacental carcinogenesis ac which Tanaka's work was presented (Tonacis and Mohr, 1973). APOOOf2623 -7- IV. Association between congenital abnormalities and childhood neoplasms Carcinogenic and teratogenic effects of exposure to a given agent night rarely be manifest in the same individual resulting in the development of both a congenital defect and a neoplasm. For some childhood neoplasms, associations with congenital defects have, in fact, been demonstrated. Con genital heraihypercrophy occurs in excess in three childhood neoplasms: Wilms' tumor, primary liver cancer, and adrenocortical neoplasms. Wilms' tumor also occurs in association with aniridia. Diseases Involving chromosomal aberra tions (Down's syndrome, Bloom's syndrome, and Franconi's aplastic anemia) are associated with leukemia in early childhood. Congenitally impaired cell mediated immunity,as encountered in Wisfcocc-Aldrich syndroae, and ataxiatelangiectasia, is associated with a childhood excess of lyraphoraticular neoplasms (Miller, 1968; Fraumenl, et al., 1963). Chromosomal defects may be the soil for development of leukemia in Down's syndrome, in which leukemia occurs in a three to fifteen-told excess-. Review of sibship histories for 459 children with leukemia and an equal number of controls matched for age, birch order, family size and race revealed chat six leukemic children had Down's syndrome, while chat condition was not pxesenc in any control child. In 'addition to Down's syndrome, ocher major congenital defects, Including pyloric stenosis, urinary tract malformations, neurological defects and musculo-skelatal malformations, occurred with, greater than expected frequency in leukemic children. In the siblings of leukemic children, Down's syndrome and a variety of tumors (among them bladder neoplasm, acute leukemia, and mesothelioma) occurred more frequently than expected. Congenital defects other chan Down!s syndrome were not present in excess in the siblings of AP00012624 -8- leukemic children. Previous pregnancies which terminated* In miscarriages were reported aore frequently by the'mothers'of LeuScemic-children than by Che mothers of controls. Childhood leukemia and Down's syndrome are both relaced to increasing maternal age. The Increase in Down's syndrome probably reflects increased frequency of nondisjunccion with aging, although Down's syndrome with leukemia occurring in children of young mothers may sore ofcen reflect chromosomal damage due to cransLocation. One revealing case reported was chat of a child vich Down's syndrome who developed acuce monocytic leukemia. An abnormal chromosome, reflecting a reciprocal cransformaclon, was present In che proband as well as in the mother, and in two of chree unaffected sib lings (German, ec al., 1962). To summarize, children with leukemia have an excess of Down's- syndrome and congenital abnormalities; their siblings have an excess of ocher cancers and of Down's syndrome; their mothers report an excess of miscarriages. These associations might be tied to the ability of ionizing radiation or ocher agents to produce persisting cytogenetic abnormalities. Chromosomal fragility syndromes also are associated with an excess of leukemia. In identical twins, where one sib develops leukemia under age seven, there exists a 202 risk of the same disease developing in che ocher twin, generally within months. These observations have spurred speculation linking leukeaogenesis, especially In children with genetic damage. AP00012625 V. Knudson's model of multistage mucacions -and carcinogerres is A multistage model of carcinogenesis beginning with two oe more muca tions has been elaborated by Knudson (1973). Such mucacions might be induced by chemical, physical or biologic agents. The proper sequence of mucacions might induce neoplasia independent of the specific mutagenic stimuli. Environmental chemical agents can, under certain situations, initiate cancer through Inducing mutations* In mouse cells, stable chromosomal' altera tions produced by 3-methylcholanchrene, can be transmitted in vitro through 30 cell generations. The cells of such transformed colonies are able to pro duce fibrosarcomas in mice. The mutagenic potential of chemical carcinogens has been exploited through development of in vitro screening systems for poten tial carcinogenic agents (McCann and Ames, 1976). Radiation-induced carcino genesis could readily fit the Knudson model, as vould viral alteration of the host genome. In non-heredicary forms of cancer, mutation would bring a normal somatic cell one step closer to becoming a tumor cell. Other factors, namely the presence of growth promoters or failure of immune surveillance, mighc be required to produce a clinically evident tumor. A related hypothesis vould apply to hereditary tumor syndromes. A mutation occurring in gern'celis might persist and be transmitted to offspring. Predisposition to neoplasia vould be inherited since one of the mutational steps leading to carcinogenesis would have occurred prazygotically. Mot only would there be a familial predisposition to the tumor, but clinical cases would be expected to occur early in life, since part of the required sequence of mutations would be'present at birth.- In recognized hereditary tumor syndromes, tumors characteristically do appear decades earlier than the same tumors in the APOOOI2626 -10- general population. Tor some childhood turners it Is proposed chat, perhaps, as few as two mutations are required. In hereditary patterns of early childhood tumors, one of these mutations, vouLd have been inherited through the parental germ cell. Retinoblastoma appears to fit this model. The relative responsiveness of this tumor to treatment has permitted affected individuals.to reproduce, and the observed pattern of Che disease In the offspring of patients with retinoblastoma is consistent with the hypothesis outlined above. Bilaterally affected individuals with retinoblastoma appear to carry a dominant gene for the disease and SOX of their offspring are carriers and are affected. Among unilaterally affected individuals, only 10% of the offspring are affected, suggesting that a substantial proportion of unilateral cases are not inheritable. Hereditary cases presumably reflect transmission.of a mutation carried in a parental germ cell. Each cell In the retina of the carrier offspring bears the first mutation, predisposing the cell to malignant transformation by further insult. Undefined, but presumably widespread, environmental factors would affect some of the retinal .cells and the carrier of the defect would develop one or multiple tumors in the retina. The frequency distribution of tumors per gene carrier has been -ennuoerated and is regarded as consistent with a Poisson distribution of tumors per carrier having a mean, of three events per person (Xnudson, 1971). Since the susceptible embryonal cell population vanishes in early childhood, persons carrying the gene have a 95% probability of ever developing retinoblastoma. Ic is possible that the predisposicion to neoplasia in carriers of the retinoblastoma gene is not entirely confined co the recina'and that ocher tissues may be more prone than expected to develop AP00012627 -11sarcomas, particularly following cherapeueicj_irradiation (Li, 1977). . Other tumors of childhood sight show patterns of inheritance suggesting prezygocic mutation, if affected individuals were to survive and reproduce. Knudson and Strong (1972a, and 1972b) have proposed similar models for Wilms' tumor and for neuroblastoma. APOOO12628 -12- VI. Use of peripheral blood lymphocytes for the analysis of chromosomal abnornalic ies The human blood lymphocyte can be utilized as a chronicle of genetic damage. Changes in chromosome structure are readily apparent when these cells are stimulated to divide, and this method has been used to measure damage produced by environmental or occupational exposure to chemical or physical agents, including irradiation, benzene and vinyl chloride.* These techniques have a number oC advantages. The cells are readily sampled, long-lived, and distributed throughout the body. Peripheral lympho cytes are synchronized in the same GQ or G^ stage of mitotic interphasa and rarely proliferate in vivo in healthy persons. Stimulation by mitogens causes some cells to undergo division.in a cell culture system and these dividing cells are readily scored for chromosome aberrations. The lymphocytes are not an entirely homogeneous system, and there may be different responses to stimuli in different subpopulacions. It Is possible that the less damaged cells may grow better in culture than chose which have been heavily damaged* Since it is the T-lymphocyte population which, in fact, multiplies and since these cells are closely concerned with the immune response, ocher aneigenic stimuli to,, the organism may affect the growth of these cells in cultura and'alter Che number of'cells with chromosome aberrations. Although scoring for chromosome-type structural aberrations (Figure 1) is generally reliable, scoring more subtle chromatid-type aberrations is much more open to inter-observer variability, and to artifacts resulting from tech nical processes. This muse be taken into account where chromatid and iso chromatid breaks and gaps are included In the. total number of aberrations. AP00012629 13- ationaie foe their inclusion in these counts Is chat certain chesiical its and viruses may cause only chromatid-type damage. An important tech.cal factor which affects the yield of aberrations is the length of time in which cells are maintained in culture. The first division for the majority of transformed lymphocytes occurs after two days of culture at 37* C. After longer periods of culture, many cells will be in their second or subsequent division and readings at this time may reflect'selective reproduction of lymphocytes with lesser degrees of damage. These factors constitute limitations to the use of human lymphocyte culture systems for study of in vivo response to environmental mutagens. It is clear that chromosome aberration rates cannot readily be compared between studies unless procedures have been carefully coordinated and standardized between investigators. Secondly, the importance of matching study and control groups by age. sex and pertinent environmental factors is apparent. Such factors include occupational history, radiacion and drug histories, exposures to toxic substances, and immunologic stimulation (viral infections, recent immunizations). The samples themselves should be processed simultaneously with processing and coding done blindly, and coding best done by someone not. directly involved in- the conduct of the investigation. To eliminate technical and observer bias such precautions are essencial (Evans and O'Rlordan, 1977). AP000I2630 Figure 1 lit; I mvii uf ) c>(ulu^iviUy iUtin^uifiablc at 'nttoiic ni,i:n`h.>y. Evans and O'P.iordan, 1977. APOOO12631 -14- VII. Environmental threats to pregnancy outcome Hirschocn and Cohen (1963) have warned of the widespread use of chemicals including pharmacotherapeutic agencs ''capable of producing chromosome breaks and rearrangement in vitro, and in vivo in human material. Such chromosome damage could be produced by a variety of mechanisms and could result .in an increased Incidence of neoplasia in the exposed individual, abortions and mal formations of offspring if exposed ^in utaro and chromosomal rearrangements in future generations, again resulting In fetal wastage and damage. Which of these drugs is, In fact, capable of producing such effects will only be deter mined by che use of large scale epidemiologic studies.'* Freeman ec al (1971) have shown chat extracts of city smog are 60 times as potent in transforming rat and hamster embryo cell culture lines as is benzpyrene. Clearly, agencs able to produce damage to genetic material are widespre- d in the environment. Recognized chceats to pregnancy outcome are not uncommonly encountered in Che environment and must be excluded when new problems are suspected. Threats are posed by biologic, chemical, and physical agents,' and may derive from medical therapeutic interventions, drug and substance abuse, and environ mental or occupational exposures to chemical agents. a, Alcohol - The most widespread documented threats to che fetus come not from the maternal-physical environment but from che.use of tobacco and alcohol during pregnancy. The risk of alcohol ingestion let pregnancy only recencLy has been evaluated and the concept dc the fetal alcohol syndrome refined (Clarren and Smith, 1978). Of infants been to recognized chronic alcoholic mothers, 83.32 had birthveighes under che tenth percentile, compared with 2.342 in. a comparable non-alcoholic population (Ulleland, 1972). In addition to inceruterlne growth retardation, che infants had retarded post-natal growth and AP00012632 15 intellectual development. A characteristic^ Eacies_has_also been described with short palpebral fissures, hypoplastic philtrum, and thin upper vermilion as well as recrognaehla in infancy. In addicion to these features of fecal alcohol syndrome, congenital malforraations of all types occur with excess frequency. Recognition of an associated increase in the occurrence of congenital malformations ironically rediscovers ancient observations and admonitions against alcohol use'-during pregnancy (Jones and Smith, 1973). The adverse effects of maternal alcohol abuse during pregnancy on Che fetus were further quantified by Ouellectee al (1977), who classified women at the first prenatal visit into four categories according to alcohol consumption and Independently evaluated preg nancy outcome. Heavy drinkers drank five or more drinks on occasion and had an average dally intake of more than 45 ml. or absolute alcohol. Women who drank more chan once a month, but did not meet the criteria for heavy drinking, were considered moderate drinkers. Abstinent women and those drinking less than once a month were grouped together in the lowest category. The mean daily dose of absolute alcohol consumed by Che heavy drinkers was 174 ml, an average of 2.2 gm per kilogram per day. The.results showed chat infants born to heavy drinkers had twice the risk of having an abnormality at birch compared to those born to abstinent or moderate drinkers. The frequency of congenital abnormalities in the offspring of heavy drinkers was extremely high, 32% when olnor abnormalities were Included, 17% if only major anomalies were considered. Multiple abnormalities were present in 20% of the offspring in heavy alcohol users. Major, minor and multiple abnormalities occurred significantly Less often among abstinent or moderate drinkers. Aicoholuse was noc the only factor distinguishing the group of mothers. Heavy drinkers were also heavier AP00012633 -16- smokars And aora likely to have a history__of previous .use. of psychoactive drugs. Heavy drinkers as a group were older, and of higher parity chan the ocher groups. Ic Is unlikely Chat these factors entirely explain the sizable differences in pregnancy outcome which were observed. The evidence suggests chac heavy regular maternal use of ethanol during pregnancy produces both, retarded development and a variety of congenital, anomalies in the child. Animal models, Including chicks, rats and guinea pigs, have been developed for evaluating ethanol effects alone on offspring (Kronick, 1976), and support the view that ethanol itself is the harmful agent, and chac its teratogenic potential is common to several species. b. Smoking - Another factor influencing pregnancy outcome, maternal smoking habits, should be addressed In efforts co tie environmental exposures of the parents to the survival and integrity of the fetus. Kline et al (1977a) suggest that the odds of spontaneous abortion among women who smoke are 1.8 times those of nonsmokars. This association remains statistically significant after maternal age, and number and outcome of previous pregnancies are taken into account. How stroking affects fetal survival is controversial. Smoking results la lowered birch weight (Rush4'and Kass,. 1974; Meyer t. ad, 1976), which may be unrelated to the increased frequency of spontaneous abortion. Lowered birth weight might come from impaired maternal nutrition or fecal anoxia, neither of which would produce an excess of chromosomal abnormalities in the conceptus. Alternate explanations for an association between smoking and spontaneous abortion have been proposed. Since spontaneous .abortion, serves as a means of selectively terminating abnormal conceptions, and 95Z of abnormal pregnancies APOOO12634 -17- are believed co terminate this way, an association between an environmental factor and increased spontaneous abortion should always provoke suspicion of a teratogenic phenomenon (Stein e al., 1975). Frequency of karyotypic abnormalities in spontaneously aborted products of conception is yet to be linked to maternal smoking habits (Alberaan at al., 197b), although the problem is complicated by variation in the proportion of chromosomally abnormal fetuses with increasing maternal age. Higher maternal social class and a history of oral contraceptive use appeared to be associated with a higher proportion of chromosotnally abnormal fetuses independent of maternal age* A slightly lower proportion of chromosotnally abnormal fetuses was reported among smokers compared to nonsmokers, but this effect disappeared when maternal age was taken into account because a large proportion of smokers were younger women. Since smoking increases the risk of spontaneous abortions almost two-fold, and the proportion of chromosotnally abnormal fetuses is close to that expected, the possibility persists chat one mechanism by which smoking may impair fecal survival is by damage co chromosomes. An increase in Che race of chromosotnally abnormal conceptions would be masked in a proportionate ratio analysis if there was also an increased loss of normal conceptions. Evidence for increase in congenital malformations in children of smoking mothers is limited and conflicting. In the course of following the develop ment of a cohort of births which had occurred in the first week of March, 1958, in England, Scotland and Vales, Fredrick ec al. (1971) suggested that there was a S0 increased risk of congenital heart disease among children of women who had smoked at least one cigarette 'per day after the fourth month of pregnancy (*7.3 versus 4,7 cases per 1000 live births). This effect was AP00012635 -18 statistically significant afcec adjustment for maternal age, parity and social class. Consenting on a study by Richards <1969), which had railed to detect any association between naternal smoking habits and congenital heart disease, Fredrick and colleagues noted the limited power of the previous study and suggested chat a 1.6-fold excess risk of congenital heart disease in offspring of smokers would likely have gone undetected as a result of the sample size of only 100 cases of congenital heart disease. Richards <1969) did, how ever, examine a variety of congenital defects or all systems which had been registered in South Wales from 1964 to 1966 and could not associate maternal smoking habits with congenital malformations in any system, after accounting for maternal age, parity, social class, area of residence and date of delivery. Given the current status of our understanding, ic is essential that in the search for exogenous causes of teratogenesls and abnormal pregnancy outcomes, the potential confounding impacts of maternal smokiag, drinkingand medication intake be taken into account. c. Drug Exposure - In utero exposure to diethylstilbestrol (DES), with the subsequent development of vaginal adenocarcinoma, is the prototype for trans placental oncogenesis in humans. The <1971) interpretation by Herbst et al. of the causal role played ^in utero exposure to DES followed hard upon the report (Herbst and Scully, 1970) of seven cases of a hichertofore exceedingly rare tumor presenting in individuals much younger chan usual. A carefully executed case-control study of eight cases of vaginal adenocarcinoma, each matched with four controls of the same age and sex, born In the sane hospital, and on Che same type of service, showed chat the- mothers of seven of the eight cancer cases reported having been created during pregnancy with DES, whereas AP00012636 -19- none of the 32 control mothers had been so treated. Subsequent reports (Herbsc et l., 1972; Herbsc al., 1975) served to confirm and extend these observations. The syndrome appears to involve not only cases of vaginal adenocarcinoma, but a spectrum of abnormalities of the vagina and cervix. Only uhen exposure to DES occurred during the first four months of gestation were neoplasms or abnormalities observed. The abnormalities ranged from Inconsequential ectropions in otherwise normal genitalia, to gross dis tortions of the cervix or upper vagina fornices caused by mucinous epithelium substituting for normal stratified squamous cells, a condition referred to as vaginal adenosis. Ulfelder (1976) has suggested the use of Che tern "cervico- vaginal epichelial dysgenesis" to describe the spectrum of abnormalities. It is estimated that the total number of females exposed to DES in utero from 1960 to 1970 is most Likely between 10,000 to 16,000 and substantially under 50,000 (Heinonen, 1973). The DES vaginal adenosis-adenocarcinoma relationship is the only docu mented example' in humans of cancer in the offspring attributable to parental- environmental' chemical exposure. The therapeucic agent was given in a high dose, more than 10 grams commonly being given in the first half of pregnancy. DES, a synthetic estrogen analogue, vas likely to have a direct effect on the development of the target tissue at a crucial point in esabryogenests. Since the risk to the offspring appears limited to exposures which occur during the first four months of gestation, direct exposure of the fetus is likely to be essential to the pathogenic mechanism. The long latent period between exposure In utero and emergence' be clinical sequelae in adolescence probably reflects the importance,of pubertal endogenous estrogens as promoting factors, and is APOOOI2637  -20- . of che need for an extended period of. observation following-suspect * cec* ,s. This association, as important as it is, does not serve as a model expc econcepcion parental occupational exposures and cancer in their offfo ng. Teratogenic effects of a multiplicity of drugs is now documented or s ipected and these will not be discussed here further except to reiterate .he need to remove the potentially confounding effects of such agents in investigation of other possible causes of impaired fecal development, d. Radiation - If preconception radiation exposure increases cancer risk in the offspring, it would heighten concern over ocher agents which produce chromosomal damage. Chromosomal aberrations attributed to vinyl chloride and benzene exposure, for example, are remlniscencof chose induced by radiation. That human populations heavily exposed to radiation develop chromosome aberrations has been clearly shown In the Japanese studies conducted by the Atomic Bomb Casualty Commission (Bloom ej: a_i. , 1966; Bloom .ec. a_l. , 1967>. Estimates of doses sustained by atomic bomb survivors considered to have been heavily exposed to radiation are in the range of 204 to 991 rad of mixed gacrna and neutron radiation. Controls, all of whom were more than three kilometers from hypoceneers *c the clips of the bombings, were estimated to have had less than one rad exposure. Controls were matched to cases by sex and by age within five years. Studies of chromosome aberrations were performed some 20 years after the exposure. Heavily exposed survivors who had been under 30 years old at the time of the bombings had an.excess of complex chromosomal abnormalities (rings, dicentrics,` fragments and translocations). StabLe aberrations, including chromatid breaks and gaps, and isochronacid breaks and gaps, occurred with equal frequency in both exposure groups. AP00012638 -21 Results of analogous investigations undertaken in individuals who had been over 30 at the time of the bombings differed from the results in the younger* group in two'ways. First, the overall frequencies of chromosome aberrations were higher than in the younger age groups. This was true within the heavily exposed and within the control survivor groups. The second differ ence was in the pattern of excess chromosomal aberrations. In the older, heavily exposed group, most of this excess was *ue to stable aberrations, rather than the complex types which predominated amor. nose heavily exposed at a younger age. It is interesting that among 33 persons exposad in utero whose mothers received more than 100 rads, 39% had complex chromosome aberrations compared to 4% of controls. All but four of the heavily exposed had been in the second or third trimester of gestation at the time of the bombing (Bloom ec_al\ 1968). Cancers and leukemia did not occur in excess during the first decade of life among 1292 children with prenatal exposure during the atomic bombings (J&blon and Kato, 197d). No abnormality in the sex ratio of offspring of the groupexposed in ucaro was observed, regardless of the sex of the exposed parenc or che trimester of gestation during which the exposure occurred (Ja'olon and Kaco, 1971; Schull ec al. , 1966)'.* While genetic consequences of irradiation of survivors of che atomic bombings might be expected to produce some lethal mucacions in their offspring, this effect has bean difficult to demonstrate. Cohort studies or offspring of survivors classified by radlacion exposure level did not show excess mortality in che children of the high exposure group. Neither were excess congenital malformations, -increased infant mortality, nor impaired survival during the first ten years of life detected regardless of parental exposure level or sex AP000I2639 -22- off exposed parent. Thus, although animal experience strongly suggests chat such effects should be manifest in human offspring of individuals exposed to ionizing 'irradiation, studies of Japanese atomic bomb survivors have not demonstrated a measurable effect (Kaco ec_ a_L. , 1966). Ho substantial increase in Leukemia risk has been detected among off spring of survivors (Hoshino et, al., 1967). Analyses having a 90% chance of detecting a four-fold increase in risk of leukemia among children of heavily exposed survivors were negative. Neither did age-at-onset of observed leukemia cases or type of leukemia in offspring differ by level of parental radiation exposure. Despite the documented persistence of chromosomal abberations in the somatic cells of adults exposed to the atomic bomb, no measurable impact in mortality, malformation or leukemia has been detected in the follow ing generation. In addition to radiation exposures related to the atomic bombings, diagnostic and therapeutic medical radiation of the parents have been investi gated to assess risks to the offspring. In a detailed case-control study o'f the relationships between radiation exposure and childhood leukemia which involved 319 childhood leukemia patients and 334 randomly selected child, con trols, an.excess risk of leukemia in offspring was tied to maternal history of preconception radiation. The magnitude off leukemia risk associated with previous radiation was estimated, after adjustment for maternal age and preg nancy history, to be 1.73 times chat of mothers who had not bee x-rayed. Oddly, only the second and later pregnancies were at increased risk; the first was apparently not. Despite the suggestion of greater risk in women with higher x-ray exposure-, no clear-cut dose-response gradient was demonstrated. APOOO12640 -23- The relative risk of leukemia in the child associated with precon ception diagnostic radiation to the father (1.31) was not statistically different from unity. When both mother and father had histories of pre conception diagnostic x-rays, the order of magnitude of the relative risk, 1.49, was about che same as chat for children whose mothers alone had received radiation (Graham, et_ aJL., 1966). Exposure to medical radiation has been widespread. - 3y che early 1960's, approximately 352 of a group of mothers of healthy children reported having received diagnostic radiation prior to conceiving. Similarly, about 22% of fathers of the same children reported having had somTsore of diagnostic radiation at some time prior to the child's conception (Cranan, ec al., 1966). Self-reporting considerably underestimated the extent of x-ray exposures. When histories of x-ray exposure in 200 subjects were verified, it was esti mated that 79.6% of the' total diagnostic x-ray films documented in physicians' and dentists' records and 66.5% of chose in hospital records had not been men tioned in interviews specifically designed, to elicit such exposures (Graham et al., 1963). The tendency to substantially underreport x-ray exposure points up the difficulty of excluding differences in prior radiation exposure as a basis for observed differences in chromosomal aberrations. Neither Is it reasonable to automatically presume chat medical radiation exposures have been equally present in the study and comparison groups, especially if these groups have not been matched by age or calendar period of observation. AP00012641 expected in persona exposed CO vinyl chloride, especially following intense exposures of long duration. The inpore of such abnormalities for reproductive function remains Ill-defined. The first reports of chromosomal aberrations in vinyl chloride workers from Sweden in 1975 (Funes-Cravioto ec al.) described abnormalities in cells drawn from seven males with occupational exposure histories of nine to 29 years (mean 16.6 years) at vinyl chloride concentrations which had declined to 20-30 p.p.a. by 1974. Calls from three non-exposed individuals working in the same factory provided comparison values. Analysis of chromosomal abnormalities were made from conventional lymphocyte cultures, which were read without knowledge of exposure group. The group exposed to vinyl chloride had aberrations in a total of 9.52% of all cells, compared to 1.94% for cells of the three controls. The frequency of chromatid and isochromacid breaks accounted for most of the disparity. Considerable variation was exhibited in the exposed workers,with cells considered abnormal ranging from 1.5% to 18%. There was no apparent relationship between duration of exposure to. vinyl chloride and the proportion o.f.cells which were abnormal. Age, radia tion exposure or other environmental factors were not evaluated, nor was It clear on what basis individuals had been selected for study. Chromosomal damage in humans exposed to this widely used industrial product was nonetheless a.possibility. In New York State, chromosomal aberrations ware reported in a group of 11 men who had 'worked in a polyvinyl chloride polymerization plant and who t AP00012642 -25- were believed to have experienced incense lntennittene_exposures Co vinyl chloride in concentrations of over 500 p.p.m., as well as chronic, buc unquantitaced exposure for four Co 23 years (mean, 15 years). Control samples were drawn, from four males in che same factory who were not known to have vinyl chloride exposure, and from six males from outside the factory. The mean age of che controls was 27 years, and chae of the subjects, AO years. Chromosome-studies were read blindly. Stable aberrations of che kind which persist in circulating lymphocytes (monosomy, trisomy, deletions, exchanges) were not increase in frequency in che exposed group, although a small increase of borderline statistical significance In the number of chromatid and lsoc.hroma.tid gaps and breaks was observed. Complex chromosomal aberrations (rings, dicentrics, fragments) occurred significantly more frequently in che exposed group than che control, a result which was properly interpreted with caution. The authors pointed out the absence of age-matched controls and che substantial age differ ence between study group and controls. The frequency of chromosome breaks and' gaps was unexpectedly high in both control and study groups (Ducacman ec al,' 1975). As with che Swedish report (Funes-Cravioco ec_ al, 1975), no relation ship could be defined between dose or duration of exposure and frequency of chromosomal- aberrations. Investigators reporting from the Uniced Kingdom (Purchase ec al.. 1975) studied 56 men with chronic exposure to vinyl chloride monomer in che course of manufacturing polyvinyl chloride and compared them with 2A unexposed individuals. Vinyl chloride exposure to che men employed In its manufacture could not be easily quantitated. Individuals with exposures co radiation, recent viral infections, or prolonged drug treatment were excluded. Samples were coded and read blindly. A higher proportion of cells from vinyl chloride AP000I2643 -26- exposed workers had chromosomal aberrations', the majority of which were breakages, compared co the comparison group. Unstable and stable chromosomal aberrations and breaks were all significantly more eoraiaon in cells from exposed workers. Hoc all invescigators have identified such changes. Picciano ec al. (1977) tested 209 employees of one vinyl chloride plant with occupational exposures lasting from 1 to 332 months (average 43.3 months). These were compared with 295 individuals undergoing preenployment physicals, not sur prisingly a group considerably younger than the vinyl chloride workers. The determination of chromosomal aberrations was apparently made with foreknowledge of the exposure group from which the sample was drawn. No differences were found between Che exposed and unexposed samples In frequency of all types of chromosomal aberrations or of any subgroup. No relationship was established between intensity of vinyl chloride exposure and degree of chromosomal damage. The absence of blind evaluation in this study, as well as the inclusion in the exposed group of individuals whose exposure was probably quite low, makes interprets cion difficult. This difficulty Is compounded by the substantial difference in mean age of study and control group members. Infanta, jhc aj^. (1976a, b) inquired whether pregnancy outcome in the wives of men exposed co vinyl chloride monomer (V.C.M.) might be less favor able chan chat experienced by wives of a pooled group of polymerization and polyvinyl chloride fabrication workers (?.V.C.). One comparison group, the rubber workers, were not believed co have been exposed co vinyl chloride monomer (V.C.M.), whereas the P.V.C. workers had low V.C.M. exposure. Data on pregnancies and pregnancy outcomes in the wives were derived from inter views and questionnaires administered to male workers. Participation races `for the groups queried ranged from 62! to 77.!. Fecal death was defined aa AP00012644 -27- any known conception noc born alive and fecal death rates were adjusted by paternal age. The cent "fecal death rate" was not used in the conventional vital statistics sense, but rather referred to the percentage of reported conceptions which did noc result in a live birch. Maternal age was not known, but was assumed to correlate closely with paternal age. Analysis of questionnaire responses suggested that, after beginning work with vinyl chloridemonomer, the number of fecal deaths per 100 conceptions was higher in the vinyl chloride exposed group than in the comparison group. So difference was reported between the two groups for the period prior to vinyl chloride exposure. Adjustments intended to remove women who were chronic aborcers from the study reduced the differences to the extent chat they were no longer statistically significant. While the authors felt chat these observations wera likely to reflect a real difference in pregnancy outcome noc attributable to either interviewer or patient recall bias, it is unfortunate chat the conelusions were based on such indirect sources of information and that there was no possibility of caking into account the multiplicity of maternal factors known to affect pregnancy outcome. Studies of pregnancy outcome (i.e., spontaneous abortion, lace fetal death [stillbirth], low birchveight, neonatal death) ia-other settings have shown the profound and subtle effects of con founding variables such as race, socio-economic status, macernal age, birth order, parity, smoking and alcohol exposure during pregnancy, maternal infec tions, and previous pregnancy outcomes. These effects may readily be sufficient to reverse Che direction of the relationship between a suspect antecedent 'factor and the outcome of pregnancy (Daling and Emanuel, 1977). Moreover, the study design precluded documenting in even the crudest manner, the validity of i | i I i i AP00012645 -28- pregnancy histories. Without such adjustments and validation, the inferences made-by Infante and colleagues cannot be sustained and little light is shed on the possibility of abnormal pregnancy outcome. Future efforts muse validate information on pregnancies and take into account known confounding factors which can affect pregnancy outcome. b. Benzene - 3enzene also may cause chromosomal aberrations in heavily occupa tionally exposed individuals. In one investigation 20 males working in a factory in which benzene had been used as a solvent underwent study for chromosomal aberrations. The exposed group had one co 20 years of benzene exposure and 14 were known co have previously been neutropenic. Control samples were drawn from five non-exposed workers in the same factory and from 33 randomly selec ted males 15 to 64 years of age. Chromosome analysis was performed blindly. Chromosomal abnormalities were reported in 2. SSI of all cells from exposed workers, compared co 1:0 to 1.42 for controls. Most or the excess was due to chromosomal aberrations of the unstable -types. In addition to the excess of abnormal cells in the exposed group, the number of unstable alterations per abnormal cell was higher in the exposed group than in the controls (Tough and Court Brown, 1965>.- In-a separate investigation, ten workers in an Italian rotogravure plant were studied who had had sequential exposures, first to benzene and later to toluene and xylene. These workers had been subjected co concentra tions of benzene ranging from 125 p.p.m. co 532 p.p.m. or higher. The maximal allowable concentration at the time of exposure had been 25 p.p.m. Another group of 24 workers had only toluene exposures except for trace contamination of toluene with xylene. These 34 workers were matched with healthy controls of the similar age and sex, who had been drawn from the general population and who had no history of benzene or radiation exposure (Fornt, Pacifica and * APOOOI2646 -29- Limonta, 1971). Chromesons changes in lymphocytes were analyzed blindly. A significantly higher proportion of abnormal cells were found In the benzene exposed'individuals as compared to their matched controls. In the group with toluene exposure only, no such excess was present. When the benzene and toluene groups were compared .to each other, an excess of chromosome changes was present in the benzene group. Most of the excess abnormalities were attributable to unstable chromosomal alterations. These studies and others (Forni et al., 1971) support the impression that exposure to benzene, particularly exposures intense enough to result in acute toxicity, can be followed by a measurable excess of chromosomal aberra tions which may persist for years. It similar effects occur in germ cells of exposed Individuals, the consequences In fecal wastage, congenital anomalies and even neoplasms might be manifest In their progeny. To dace no excess of such abnormalities has been reported in offspring of benzene exposed individuals, c. Anesthetic Cases - Occupational exposures of female operating room personnel to waste anesthetic gases have been held responsible for decreased fertility, increased race of spontaneous abortion, low birch weight and possibly impaired development in Che offspring. An excess of spontaneous abortions In wives of men exposed to anesthetic 'gases has been suggested but not confirmed. Among operating room nurses, 29.7% of pregnancies terminated with spon taneous abortion, compared to 3.3% among control nurses and a similar excess of spontaneous abortions was observed for female physician anesthesiologists compared to female physicians in ocher specialties (Cohen e_c_ ai.. , 1971). Spon taneous abortions occurred two weeks earlier on the average among the operating room groups than among controls, at an average of eight rather chan ten weeks AP000I2647 -30- of gestation. Another survey suggested that or_eciale nurse-anesthetists who had worked during pregnancy, the frequency or birth defects in the off spring was 16.4% compared to 5.7% for pregnancies in which the mother had not worked (Corbett ec al., 1974). A variety of birch defects occurred. Ten mothers had defects in offspring of two or more pregnancies during which they had worked. In the same study, three neoplasos_occurred in two of the 434 children whose mothers had practiced during pregnancy. One child had a neuro blastoma at birth and developed a second primacy, a thyroid malignancy, at puberty. A parotid tumor developed at age 22 years in another child. Among 261 pregnancies in which Che mother had not worked, there was one case of acute leukemia occurring in a child three years of age. Studies from Finland and the United kingdom also suggest a deleterious effect on pregnancy outcome from maternal exposure to operating room environ ments. An increase in frequency of early spontaneous abortions, as well as low birch weight, has been reported for Finnish female operating room scarf (Rosenberg and kirves, L975). Among female physicians in England and Wales," congenital defects, low birth weight and higher stillbirth cates, but not higher spontaneous abortion races were reported from pregnancies of women holding anesthesiology appointments, compared to other women doctors (Pharoah et al., 1977). The effects of paternal exposure to anescheclc gases are much more in doubt. Congenital anomalies were reported to be increased by 25% in the off spring of male anesthesiologists compared to offspring of male members of. the American Academy of Pediatrics. No dif Terence .was reported for spontaneous abortion rates among wives of male operating room personnel compared to con trols (Ad Hoc Committee ASA, 1974). Questionnaires completed by 5119 married APOOO12648 male physicians in Che Jnicad Kingdom suggested an-lncrease in minor-buc not major congenital anomalies, and no difference in frequency of infertility, spontaneous abortion or cancer in the offspring of male anesthesiologists compared to control physicians (Knill-Jor.es et. al., 1975). AP000I2649 -32- IX. Cancer in che offspring oi hydrocarbon exposed fathers Two studies offer conflicting evidence with respect to the risk of cancer in children of even whose work might lead them to be exposed to petroleum derived hydrocarbons. Fabia and Thuy (1974) reviewed death certificates of children dying in Quebec during 1965 to 1970, in order to identify children who had died from malignancdisease while under five years of age. For 336 .of 402 such children, birch certificates were found in the Quebec population register. A control group of 772 patients was selected using the birch registration record preced ing and following chat of each case in the official files. This effectively matched for season, calendar period, and province of birch. Occupation of the.father at the time of birch was taken from the birth certificate. An industrial hygienist independently grouped the birth certificates into three levels of probable exposure to petroleum derived hydrocarbons.' Occupation o the father was unknown for 30 cases and 56 controls. The distribution of occupa tion of the facher ac birth differed for cases and controls. Most of chis differ ence was the result of an excess among cases of occupations considered to be hydrocarbon-related*- The relative odds of cancer in che children of men holding hydrocarbon-related jobs at the eiae'or the child's birch were 3.1,.an increase which was unlikely to have occurred by chance. Most of che excess of exposed occupations was accounted for by motor-vehicle mechanics, machinists, miners and painters, When similar analyses were conducted by che type of cancer in che child, che excess of hydrocarbon-related occupations prevailed for che following groupings: leukeraia-lymphoraa, nervous system malignancy, other tumors. For four of five cases of Lecterer-Slwe disease fathers were in che hydrocarbon-related work groups. Mo excess o.f exposed fathers was AP00012650 -33- prasenc for children with`Wilms' tumor. The differences' observed were not attributable to differences in pareneal age or geographic residence at the time of birch. The investigation was based on deaths in children under five, and it is possible that social class bias may have been introduced in that fashion. If children of more affluent parents with a childhood neoplasm are more likely to survive to over age five, and therefore not be included In the study group, compared to children of poorer parents, this would lead to a higher proportion of children with parents of lower socio-economic status in the cancer death group and would bias the distribution of father's occupa tion. Moreover, it was not possible to confirm that fathers in the so-called "exposed" group did, in fact, have contact with hydrocarbon-derived chemicals. While not ideal, the death certificate probably is a valid source of diagnostic information far cancer deaths in young children. The cancer diagnoses employed in the analysis - "leukemias and lymphomas", "nervous system" end "others" -- were quite broad and included many histopathologic entities. Occupational association was not restricted to specific entities within the broader categories. Because the epidemiologic features of the various histopathologic types of leukemias, lymphomas, and nervous system tumors are distinctive, a specific chemical exposure or class of exposures. If assumed to be of causal significance, would more likely be linked with specific histopathologic as well as organ sys tem effects. Although an analysis caking into account the matched pairs de sign would have been more elegant, its use would not have altered the substance of the conclusions. A study with a similar design but different findings was conducted in Finland (Hakulinen ec al.,'1976). Children under 15 years, of age who developed APOOO12651 -34- cancer ware identified from the Finnish Cancer Registry, 1959-1968. Of 1409 cancer cases so identified, the final series~conslsced of 852 pairs for whom birch records which included father's occupation were available. Each case had been matched with a child whose birth dace immediately preceded that of thevcase and who was born in the same maternity welfare district. This procedure matched effectively by calendar period, season and domicile at the time of birch. Father's occupation was drawn not from birch registration records but from records of the free, nationwide antenatal care system in operation in Finland. Father's occupation was classified by likelihood of hydrocarbon exposure in a manner which was shown to be comparable with the study done in Quebec. No excess risk for hydrocarbon exposed fathers was decected for any class of neoplasms either for children whose cancers occurred under five or under 15 years of age. The power of. the study to detect a two fold risk ratio was not determined, but appeared to be insufficient for chil dren under five years of age. This study differed in. design from chat done in Quebec primarily in the source of information on father's occupation. For the Finnish study, chat was drawn from antenatal records compiled for the most part in che first tri mester of gestation. This'may be a better index of exposure at the time of conception chan chat recorded at che time' of birch in che birch certificate itself. Although a large number of pairs were discarded because of unavaila bility of father's occupation, these came from a circumscribed calendar period, and since the matched pairs analysis was maintained throughout the analysis, should in any case not influence che result. The Finnish study used incident cases and was able to look at age groups.up to five as well as up to 15 years at the time of cancer*diagnosis. A further explanation for che discordant findings in che two studies Is possible variation In terminology of occupa tional classification. At present, the contradictory findings oc che two studies are not readily reconciled. Because of che initial Canadian report (Fabla and Thuy, 1974) of an increase in risk of cancer death in children of fathers whose occupation Involves potential exposure to petroleum derived hydrocarbons, a study was undertaken in Karris County, Texas to assess the relationship of parental occupation to childhood neoplasms using a case-control approach. Data col-, lection in this study is near completion but, since data analysis is only beginning, report of its conclusions is some time auay (Heath, 1978). Cases consist of all cases of malignant disease being followed in the Research Hematology Clinic at Texas Children's Hospital during 1975-76, approximately 350-400 cases, ages 0-14 years, with half under five years old. Somewhat more chan half of these patients have leukemia and lymphoma and are residents of Karris County in Texas. It is estimated that chese cases represent 80-90% of all such childhood malignancies in chat county. Three secs of controls are to be used* Clinic controls are children from the Research Hematology Clinic who have-non-m&lignant disease. Clinic controls will not be individually matched to cases, but group characteristics are being, ascertained to insure comparability. Interviews would be conducted with controls in che course of regular clinic visits. The second sec of controls are "self-matched" to che case, with che parents of che case choosing a family fcoa their neighborhood and acquaintance to act as controls. Data would be collected from this group by interview as well as by telephona and mailed questionnaire. This unusual method was pro posed by Deere end Wyschafe In 1971. AP00012653 -36- The final sec of controls are dratm from eh* case's families...Married siblings closest to the patient's parents in age are to be selected so that there would be one maternal and one paternal sibling. Family controls would be Interviewed at the same time as the parents of the case. Occupational groupings by an industrial hygienist are being established according to the 1970 tf.S. census with duration of parental "exposure" esti mated and with account to be taken of socio-economic class. The question naire itself is designed to elicit demographic information about the family, as well as a detailed job history from both of the cases' parents and the control parents. Family residence and school histories are also collected. It is premature to comment extensively on this investigation in the absence of further information about its progress and results. A number of difficulties can be conjectured in its conduct and design -- particularly in the use of unmatched clinic concrols with non-malignane diseases and in Che ''self-selected'1 control group. When Deere and Vyschak (1971) used selfselected controls, the independent variables of interest were related to menstrual history and were unlikely to be biased by the fact that che control was selected by the case. The application attempted in che Harris County study is quite different and it' is likely chat for the self-selected controls, and also for che family concrols, there is a tendency cor control parents to have an occupation similar to the case's parents. This would, however, tend to minimize any association between occupational exposure to petroleum derived hydrocarbons and childhood neoplasms. AP0001Z654 -37- X. Surveillance of spontaneous abortions as a strategy of environmental monitoring . ------ * As che previous discussions have suggested, the relationship between mutagenic, teratogenic and carcinogenic effects in the offspring of persons exposed to noxious environmental agents is a complex one. Surveillance of spontaneous abortions may have a number of advantages as a means of prospective monitoring for such effects since defective conceptions are aborted selectively. As a result, studies of teracogenesis In spontaneous abortions nay be consider ably more efficient than those conducted in newborns. Because the frequency of abnormalities is higher in spontaneous abortions, the sample size needed to demonstrate a change in risk is ouch smaller chan for a parallel query addressed to chromosomal defects or developmental abnormalities recognized at birth. The magnitude of this difference can be dramatic. If chromosomal defects diagnosable on the appearance of the newborn are considered alone, the sample size required may be hundreds of times chat of an investigation examining prevalence of chromosomal anomalies in early abortions. This is illustrated in Table 1, adapted from Kline et al. (1977). In addition to sample size considerations, study of spontaneously aborted conceptions .offers a lead time of at lease six months over studies of live births. The abortion specimen can be studied with care and thoroughness, permitting detection of anomalies lethal to the fetus which would not have been detected as readily in studies of live births. [Insert Tabla 1 about here] The problem of power and sample size is only one of the issues which beleaguers Investigators of associations between exposures to parents end outcomes of pregnancy. Related factors are the timing of the exposure, maternal AP00012655 -38or paternal, in relation to conception and/or gestation, and the specific measures of pregnancy outcome (Scrobino jet al., 1978). Paternal exposures may act in two ways -- by contamination of the maternal environment resulting in secondary maternal exposures, or directly by affecting paternal germ cells. Since spermatogenesis is a continuous process, paternal exposures pre-conception, rather than cumulative or delayed effects, are those most likely to have been responsible for deleterious effects. By focusing on early pregnancy wastage, particularly If the frequency of chromosomal abnormalities in the aborted product of conceptions can be deter mined, che objectives of study are better focused and achieved with reduced sample size, and the confounding effects of the maternal in utero environmental factors are minimized. AP00012656 -39- XI# Further studies ______ __________________ . .. ____ Based on Che evidence summarized in che preceding discussion, risks to Che fetus which result from paternal exposure to hydrocarbons, if they exist ac all, are likely co be small in magnitude. The confliccing scientific evidence and the-social Implications of an occupational exposure which may affect the fetus, however, urge further investigation. "Are. the pregnancies of women employees and wives of men who are occupa tionally exposed co petroleum-related hydrocarbons likely co come co a less favorable*conclusion than those of unexposed petroleum industry employees?" This question can be addressed using the developing resource of the ongoing study of morbidity and mortality among petroleum industry employees. .In the families of the more than 33,500 mostly male employees whose census data is now recorded, there will be some 2050 conceptions per year. The background for this estimate is recorded in Table 2, which givas che age and sex break down of pare of that study population, along with the corresponding age-specific birth rates for females in the corresponding age group. On che assumption that che age of the workers1 wives corresponds closely to that of male workers, it can be estimated that there will be approximately 1375 live births in this group annually. Since 314 Induced abortions cake place for every 1000 live births nation-wide', there would, in addition, be some 400 conceptions termina ted by induced' abortions in the same year. On che assumption that of all conceptions, approximately 152 will terminate in spontaneous abortion, che total annual number of conceptions is estimated at about 20S0 (Kline et_ al,, 1977). About 50 per 1000 recognized conceptions are chcomosomally abnormal, and approximately 92% of all chromospnally abnormal conceptions abort AP00012657 -40- sponcaneously. Changes in races of chromosomal abnormality can be scud led direccly through examination of spontaneously aborted'products of conception or indirectly through Che scudy of frequency of spontaneously occurring abortions. The number of pregnancies which must be studied to detect a given Increase in the spontaneous abortion race was shown in Table 1. To detect a relative risk of spontaneous abortions of two (*d*0.05 /3*0.2), there would have to be 146 pregnancies each in the "exposed*' and "unexposed** group. To generate this number of pregnancies in one year (assuming United States birth rates) would require surveillance of a total population of about 6500 workers given che age-sex distribution of employees in the MSKCC-API study. These estimates suggest the suitability of following che reproductive experience of a portion of the population in the MSKCC-API scudy. This could be achieved by monitoring a sample of male employees and their wives as well as all female employees, under age 49 years. Such surveillance would require che cooperation or the employees under study, since only through voluntary reporting would all conceptions be identified and their outcome determined. It would be necessary to document previous obstetrical history, past and present smoking history and medical history, and spouse's occupational history prior to and post conception. While area-wide surveys have provided information on the expected frequency of various pregnancy outcomes (Stain ec al., 1.975; Creasy _ec <al. , 1976), a more appropriate comparison group of "unexposed" preg nancies can be drawn from employees (and their wives) of the petroleum industry who have little or no petroleum derived hydrocarbon exposure. Given che personal nature of such inquiries, considerable attention would have to be given to developing approaches which would permit the scudy to proceed success fully. The advantage of such an approach is chat, if tactical problem can be rt AP00012658 -41surmouated, it directly addresses the question of pre3nancy outcome as it pertains to employees In the petroleum industry and the occupational exposures which they experience. AP00012659 Prevalence^ .20 .25 .30 .35 .40 .45 .50 .55 .60 TA3LE 1 Sample Sires co Dececc Increased Sponcaneous Abortion Slices Number of exposed pregnancies required co reach Indicated power .80 .90 984 1290 289 371 146 186 91 116 64 60 48 60 37 ' 46 30 37 25 30 Number of exposed pregnancies needed co dececc an increase in sponcaneous aboccions co Prevalence2 given Prevalence^ equal co 0.15 and =0.05. The same number of pregnancies is required in che unexposed group. AP00012660 TABLE 2 Preliminary Enolove s Population in the MSKCC-API study by Aee and Sex'* Age-specific birth rates Age Male Female oer 1000* No. of births in one year by sex and aae erouo of employee parent Male Female - 19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65 6 1358 4202 4441 2813 2136 2960 4591 5196 . 3302 462 31,472 . 5 243 543 404 276 201 178 168 173 104 10 2320 55.5 158.3 141.7 70.3 29.5 7.3 0.4 0 214 595 312 83 16 1 1221 0 40 77 23 8 1 0 154 Total births: 1375 * U.S. white, 1969-71. **This is a portion of the MSKCC-API study of Mortality and Morbidity in Petroleum Industry Employees which represents about half of the final population. APOOO12661 REFERENCES Abortion Surveillance Branch, Family Planning Evaluation Division, Bureau of Epidemiology, Center for Disease Control. Abortion SurveillanceU.S., 1976. Morbidicy and Mortality; Weakly Report. Vol. 27, May 26, 1978. Alberrnan, E.f Creasy, M. , Elliott, M. et al. Maternal factors associated with fetal chromosomal anomalies in spontaneous abortions. Sr. J. Obstet. Gynaecol.. 83:621-627, 1976. Atkin, N.B. Cytogenetic Aspects of Malignant Transformation. Texas: S. Karger, 1976. Bibbo, M. , Kaensxel, W.M., Wted, G.L., Hubby, M., Herbst, A.L. A twenty-five year follow-up study of women exposed to ^iethylscilbescrol during pregnancy. N. Engl. J. Med., 293:763-767, 1973. Bloom, A.D., Neriishi, S., Archer, P.G. Cytogenetics of the in utaro exposed of Hiroshima and Nagasaki. Lancet, 2:10-12, 1968. Bloom, A.D., Neriishi, S., Awa, A.A., Honda, T., Archer, P.G. Chromosome aberrations in leukocytes of older survivors of the atomic bombings of Hiroshima and Nagasaki. Lancet, 2:802-80.5,- 1967. Bloom, A.D., Neriishi, S., Kamada, if. , Iseki, T., Keehn, R.J. Cycogenecic investigation of survivors of the atomic bombings of Hiroshima and Nagasaki, Lancet, 2:672-674, 1966. Clarren, S.K. and Smith', D.W. The cecal alcohol syndrome. M. Engl. J. Mad. .298:1063-1067, 1978. Cohen, E.N., Ballville, J.W. and Brown, B.W. Anesthesia, pregnancy and mis.carrlage: A study of operating room nurses and anesthetists. Anas theslology. 35:343-347, 1971. Corbett, T.H., Cornell, R.G., Endres, J.L. and Liading, K. Birch defects among children of nutse-anesthetiscs. Anesthesiology, 41:341-344, 1974. Creasy, M.'R., Crolla, J.R. and Alberrnan, E.D. A cytogenetic study of human spontaneous abortions using banding techniques. Human Genet. , 31: 177-196, 1975. Dal'lng, J.R. and Emanuel, I. Induced abortion and subsequent outcome of pregnancy in a series of American women. N. Engl. J. Med., 297:1241-L245, 1977. Deere, K. and Wyshak, C. A matching procedure for epidemiologic studies. J. Chronic DLs., 24:33-92, 1971. DIPaolo, J.A. and Koci'n, P. Teratogenesis-oncogenesis: A study of possible relationships. Arch. Path., 31:3-23, 1966. Ducacman, A., Hirschorn, K. and Selikoff, I.J. Vinyl chloride exposure and human chromosome aberrations. Mutation Res. , 31:163-163 , 1975. AP00012662 References 2. Evans, H.J. and O'Riordan, M.L. Human peripheral iood lymphocytes for Che analysis of chromosome aberrations in mutagen tests. In, 3.J. Kilbey (ed.), Handbook of Mutagenicity Test Procedures, M.Y.: Elsevier Scientific Publishing Co., 1977, pp. 261-274. Fabia,*J. and Thuy, T.D. Occupation of father at time of birch of children dying of malignant diseases. Be. J. Prev. Soc. Med.. 28:96-100', 1974. Fleiss, J.L. Staeiscical Mechods for Races and Proportions. New York: John Wiley and Sons, 1973. Forni, A.M., Cappellicii, A., Pacifico, E. and Vigiiani, E.C. Chromosome changes and their evolution in subjects with past exposure to benzene. Arch.. Environ. Health. 23:385-391, 1971 Forni, A., Pacifico, E. , Limonta, A. Chromosome studies in workers exposed to benzene or coluene or both. Arch. Environ. Health, 22:373-378, 1971. Frauraeni, J.F., Jr. The susceptibility of the fetus and child to chemical pollutants. Chemicals in human ceratogenesis and transplacental carcino genesis. Pediatrics, 53:307-812, 1974. Frauoeni, J.F., Miller, R.V. and Hill, J.A, Primary carcinoma of the liver in. childhood: An epidemiologic study. J. Nat. Cancer Inst., 40:1037-1099, 1968. Fredrick, J., Alberman, E.D. and Goldstein, H. Possible teratogenic effect of cigarette smoking. Nature, 231:529-530, 1971. Freeman, A.E., Price, P.J., Bryan, R.J., al. Transformation of rat and hamster embryo cells by extracts of city smog. Proc. Natl. Acad. 3ci. " 68:445-449, .1971. Funes-Cravioco, F., , Lambert, B., Lindsten, J. , Ehrer.berg, L, Natarajan, A.T. and Oscerman-Colkar, S. (Letter) Chromosome aberrations in workers exposed to vinyl chloride. Lancet, 1:459, 1975. German, J.L., De.Ma.yo, A.?., and Bearn, A.G. Inheritance of an abnormal chromosome in Down's syndrome (mongolism) with leukemia, Aa J. Human Genet. , 14:31-43, '1962. Graham, S., Levin, M.L., Lilienfeld, A.M., Schuaan, L.M., Gibson, R., Dowd, J.E. and Hempelmann, L. Preconception, intrauterine, and postnatal irradiation as related to leukemia. Nat. Cancer Inst. Honoec.. 19:347-371, 1966. Crahatn, S., Levin, M.L... Lilienfeld, A.M., Dowd, J.E., Schunan, L.M., Gibson, R. , Herapelmann, L.L. and Gerhardt, D, Methodological problems and design of the cristate leukemia purvey. Ann. N.Y.'Acad'. Sci., 107:557-569, 1963. Hakulinan, T., Salonen, T. and Teppo, L. hydrocarbon-related occupations. 1976. Cancer in the offspring of fathers in Br. j. prev. Soc. Med., 30:133-140, s : AP00012663 References 3. Heach, C-, Jr. Personal communication, 1973. Heinonen, 0.?. DiethvlscilbestroL in. pregnancy: Frequency of exposure and usage patterns. Cancer, 31:573-577, 1973. Herbst, A.L., Kureraan, R.J., Scully* R.E. and Paskanzer, D.C. Clear cell adenocarcinoma of the genital trace in young females. N. Engl. J. Med.. 287:1259-1264, 1572, Herbst, A.L., Poskanzer, D.C., Robboy, S.J., Frudlander, L. and Scully, R.J. Prenatal exposure to scilbesterol; A prospective comparison of exposed female offspring with unexposed controls. N. Engl. J. Med. 292:334-339, 1975. Herbst, A.L. and Scully, R.E. Adenocarcinoma of the vagina in adolescence: report of 7 cases including 6 clear-cell carcinomas (so called mesonephromas). Cancer, 25:745-757, 1970. A Herbst, A.L., Ulfelder, H. and Poskanzer, D.C. Adenocarcinoma of the vagina: Association of maternal stllbescrol therapy with tumor appearance in young women. N. Engl. J. Med. , 284:573-381, 1-971. Hirschhorn, K. and Cohen, M.M. Drug-induced chromosomal aberrations. Ann N.Y. ' Acad. 5cl.. 151:977-937, 1968. Hoshino, T., Kato, H. , Finch, S.C. and Hru'oec, 2. Leukemia in offspring of atomic boob survivors. Blood, 30:719-729, 1967. Infante, P.F., Maaoner, J.K., McMichael, A.J., Waxveiler, R.J. and Falk, H. Genetic risks of vinyl chloride. Lancet, 1:734-735, 1976a, 1:12S9-1290, 1976b. . Jablon, S. and Kato, K. Childhood cancer in relation to prenatal exposure to atomic bomb radiation. Lancet, 2:1000-1003, 1970. Jablon, S. and kato, H. Sex ratio in offspring of survivors exposed prenatal!*/ .to the atomic bombs,in Hiroshima and Nagasaki. Am. j. Epidemiol., '93:253-258, 1971. ' Jacobs, P.A., Price, W.H. and Law, P. Human Population Cytogenetics. Baltimore: Williams and Wilkins Co., 1970. Jones, K.L. and Smith, D.W. Recognition of the fetal alcohol syndrome la early infancy. Lancet, 2:999-1001, 1973. ato, H., Schull, W.J. and Meel, J.V, A cohort-tvpe study of survival in the children of parents exposed to atomic bombings. Am, J. Human Genet.. 18:339-373, 1966. Kline, J., Stein, Z.A.. Susser, M. and Warburton, D. Smoking: A risk factor for spontaneous abortion. M. Engl. J. Med., 297:793-396, 1977a. AP00012664 References 4. Kline, J., Stain, Z., Strobino, 3., Susser.'M. and Warburcon, D. Surveillance of spontaneous abortions: Power in environmental monitoring. Am. J. Eoidemiol.. 106:345-350, 1977b. Knill-Jones, R.P., Neuman, 3.J. and Spence, A.A. Anaesthetic practice and pregnancy: Controlled survey of male anesthetists in the United Kingdom. Lancet. 2:807-809, 1975. Knudson, A.G., Jr. Mutation and Cancer: Statistical study of retinoblastoma. Proc. Mael. Acad. Scl.. U.S.A.. 68:320-823, 1971. Knudson, A.C., Jr. Mutation and human cancer. * Adv. Cancer Res.. 17:317-352. 1973. Knudson, A.G. and Strong,L.C. Mutation and cancer: A model for Wilms' tumor of the kidney. J. Natl. Cancer Inst.. 48:313-324, 1972a. Knudson, A.G. and Strong, L.C. Mutation and cancer: Neuroblastoma and pheochromocytocna. Am. J. Human Genet., 24:514-532, 1972b. Kronick, J.3. Teratogenic effects of ethyl alcohol administered to pregnant mice. Am. J. Obstet. Gynecol., 124:676*r630, 1976. Li, F.P. Second malignant tumors after cancer in childhood. Cancer. (Suopl.) 40:1899-1902, 1977. McCann, J. and Ames, 8.N. A simpLe method for detecting environmental carcino gens as.mutagens! Ann. H.Y. Acad. Sci. . 271:5-13, 1976'. Meyer, M.3., Jonas, B.S. and Tonascla, J.N. Perinatal events associated with maternal smoking during pregnancy. Am. J. Epidemiol., 103:464-476, 1976. Miller, R.W. Down's syndrome (mongolism), other congenital malformations and cancers among the sibs of leukemic children. ' N. Engl. J. Med., 268:393-401, 1963. Miller, R.W. Relation between cancer and congenital defects: An epidemiologic evaluation. J. Nat. Cancer Inst., 40:1079-1085, 1963. Mond&l, S. and Heidelberger, C. In vitro malignant transformation by mechylcholanchrene of the progeny of single cells derived from C3h mouse prostate. Proc. Nat. Acad. Sci. U.S.A.. 85:219-225, 1970. Oppenheim, 3.S., Criem, M.L. and Meier, P. The effects of diagnostic x-ray exposure on the human fetus: An examination of the evidence. Radiology, 114:529-534, 1975. Ouellette,* S.M., Rosect, H.L., Rosman, ?.N. and Weiner, L. Adverse effects or. offspring of maternal alcohol abuse during pregnancy. N. Engl. J. Med.. 297:528-530, 1977. AP00012665 References 5. O.E.D. The Compact Edition of the Oxford English Dictionary. Oxford: Oxford University Press, 1971, cites Dunglison Medical Lexicon, 1857. Pharoah, P.O.D., Alberman, E. and Doyle, P. Outcome of pregnancy among women in anesthetic'practice. Lancet. 1:34-36, 1977. Picciano, D.J., Flake, R.E., Cay, P.C. and Kilian, D.J. Vinyl chloride cytogenetics. JOM. 19:527-530, 1977. Purchase, I.F.H., Richardson, C.R. and Anderson, D. Chromosomal and dominant lethal effects of vinyl chloride. Lancet, 2:410-411, 1975. Rice, J.M- An overview of transplacental chemical carcinogenesis. Teratology, 8:113-125, 1973. Richards,I.D.C. Congenital malformations and environmental influences in pregnancy. Sr. J. Prev. Soc. Med., 23:218-225, 1969. Rosenberg, P. and :<irve$, A. Miscarriages among operating theatre staff. Acta Anaesthesiol. Scand., (Suppl.) 53:37-42, 1973. Rush, D. and Kass, E. Maternal smoking: A reassessment of the association with perinatal mortality. An. J. Epidemiol., 96:183-195, 1972.' Schull, W.J., Neel, J.V. and Hashi2ume, A. Some further observations on the sex ratio among infants born to survivors of the atonic bombings of . Hiroshima and Nagasaki. Am. J. Hunan Genet.. L3:328--338, 1966. Shepard, T.H., Miller, J.R. and Marois, M. (eds.) Methods for detection of environmenca! agents chat produce congenital defects. Proceedings of the Guadeloupe Conference. Sponsored by L'Institute de La Vie. Amsterdam: North Holland Publishing Co.; New York: American Elsevier, 1975. Staples, fc.E. Definition of taratoaenesis and teratogen. In_ T.H. Shepard et al. (ed),Methods for detection of environmental agents that produce congenital defects. Proceedings of the Guadeloupe Conference. Amsterdam: North Holland Publishing Company, 1975. Stain, Z.t Susser, M. and Warburcon, D. Spontaneous abortion as a screening device: The effect of fetal survival on the incidence of birth defects. Am, J. Epidemiol.. 102:275-290, 1975. Strobino, B.R., Kline, J. and Stein, Z. Chemical and physical exposures of parents: Effects on human reproduction and offspring. Early Human Oevel., 1:371-399, 1973. Tanaka, T. Transplacental induction of- tumours and malformations in rats created with some chemical carcinogens.-' I_n L. Tomatis and U. Mohr (eds.), Transplacer.cal Carcinogenesis, IARC Scientific Publications ,4 :100-111, 1973. APOOOt2666 References Tomatis, L. and Coodall, C.M. The occurrence of tumors in T^r^and descendancs of pregnant mice injected with 7,12-dimcchylbeftz(a)anthtacene. Inc. J. Cancer. 4:219-225, 1969. Toroaeis, L. and Mohr, U. Transplacental carcinogenesis. Proceedings of a meeting held at the Medisinische Hochschule, Hannover Federal Republic of Germany, October 1971. International Agency for Cancer Research, Lyon, 1973. Tough, I.M. and Court Brown, V.M. Chromosome aberrations and exposure to ambient benzene. Lancet, 1:634, 1963. Ulfalder; H. The scilbescerol-adenosis-carcinoma syndrome. Cancer (Suppl.), 38:426-431, 1976. Ulleland, C.N. The offspring of alcoholic mothers. Ann. N.Y. Acad. Sci,, 197;167-169, 1972. Wilson, J.G. Environment and Birth Defects. New York: Academic Press, 1973. AP00012667