549 Mesothelioma, Asbestos, and Reported History of Cancer in First-Degree Relatives Ellen F. Heineman, mo.' Leslie Bernstein, Ph.D.2 Alice D. Stark, D~.P.H? Robert Spirtas, SC.D.' ' Occupational Studies Section, Environmental Epidemiology Branch, Epidemiology and Biostatistics f'rogram, National Cancer Institute, Bethesda. Maryland. Department of Preventive Medicine, University of Southern California School of Medicine, Los Angeles, California. Bureau of Environniental and Occupational Epidemiology, Center for Environmental Health, New Yotk State Department of Health, Albany, New York. BACKGROUND. Although malignant mesothelioma is known to be strongly related to asbestos exposure, its relationship to familial factors is unclear. METHODS. We compared reported histories of cancer in first-degree relatives, obtained from telephone interviews with the next-of-kin of 196 patients who had a pathologic diagnosis of mesothelioma, and with those from 51 1 deceased controls. RESULTS. Among men exposed to asbestos, we found a statistically significant twofold elevation in the risk of mesothelioma for patients reporting cancer in two or more first-degree relatives. We found no significant elevation in women or among the small number of men without asbestos exposure. The next-of-kin of three patients (but no controls) reported a possible mesothelioma in a first-degree relative: asbestos exposure could not be ruled out in those relatives. Associations of asbestos with pleural mesothelioma were stronger among men with a reported family history of cancer than men without, although no statistical evidence of an interaction was detected. CONCLUSIONS. These results provide suggestive, but limited, evidence that a family history of cancer may be a risk factor for mesothelioma, or may indicate an increased susceptibility to mesothelioma given asbestos exposure. Cancer 1996 77:549-54. 0 1996 Aiiierican Caricer Society. KEYWORDS: mesothelioma, occupational exposure, family, asbestos, etiology, casecontrol studies, risk factors, Li-Fraumeni cancer syndrome. M This work was supported by intramural funds of the National Cancer Institute. Current address for Robert Spirtas: Center for Population Research, National Institute of Child Health and Human Development, Bethesda, Maryland. Address for reprints: Dr. Ellen F. Heineman, Occupational Studies Section, EEB, NCI, EPN Room 418, 6130 Eixecutive Blvd., MSC 7364, Bethesda, MD 20892-7364. Received June 19,1995; revision received October 11, 1995; accepted October 11, 1995. CJ 1996 American Cancer Society esothelioma is a rare cancer of the lining of the pleural and peritoneal cavities. Asbestos exposure is the major risk factor for this disease and is responsible for the vast majority of cases among men and perhaps 25% of cases among women.',' Although other substances induce these tumors in experimental animals,"-' few etiologic factors other than asbestos have been suggested in humans, and the evidence is weak for most of them.5 Little is known about the relationship of mesothelioma with familial or genetic factors. Although several series""' of familial cases of mesothelioma have been reported, it has been difficult to rule out carryhome exposure to asbestos.'-g To our knowledge, the role of family history has been explored in only one case-control study in which Vianna and Polan reported a statistically significant threefold risk of mesothelioma among women whose parents were diagnosed with cancer, particularly with gastrointestinal malignancies." We analyzed data from one of the largest case-control studies of mesothelioma conducted to date, to determine if mesothelioma risk was increased among individuals with a family history of cancer and if the association of mesothelioma with asbestos exposure varied according to the presence or absence of a family history of cancer. 550 CANCER February 1,1996 / Volume 77 / Number 3 MATERIALS AND METHODS Patients Details of the design of the study have previously been published.’ Briefly, potential patients were identified from population-based cancer registries in New York State and Los Angeles County, and from the files of 39 large Veterans Administration hospitals. Eligible patients were those diagnosed between January 1, 1975, and December 31, 1980. Of 720 eligible patients, the next-of-kin of 536 (74%)completed a telephone interview, 106 (15%) were not located, 64 (9%) refused an interview, the physician of 8 (1%)refused to allow contact, and 6 (< 1%)had partially completed interviews. Of the 536 patients whose next-of-kin had a completed interview, 208 (39%) were confirmed as having a “definite” or “probable” mesothelioma by a review by 2 pathologists expert in the diagnosis of mesothelioma.’ Details of the pathology review have previously been reported.” Controls were identified from death certificate files for the catchment area of the registries and from deaths listed in the Beneficiary Identification and Records Locator Subsystem system of the Veterans Administration. Eligible causes of death excluded other cancers, respiratory disease, suicide, or violence. Pair-matched controls were selected for 678 of the 720 potentially eligible patients. For these 678 controls, the next-of-kin of 533 (79%) were interviewed. We were unable to locate respondents for 138 controls (20%),4 (< 1%)had partially completed interviews, and 3 (< 1%) refused to be interviewed. We used this larger set of 533 individuals as controls, rather than the controls individually matched to the 208 cases, in order to increase statistical power by taking advantage of the information available from the controls matched to unconfirmed cases. In the analyses, we controlled for age, sex, and geographic location. and reported residence within two miles of an asbestos mine or mill. For the job-exposure matrix, each job held by an individual was classified as having a probability of none, less than lo%, 10% to 19%, 20% to 49%, or 50% or greater “likelihood” of asbestos exposure, based on the results of the National Institute for Occupational Safety and Health’s National Occupational Hazard Survey.13 Each individual was classified into the highest achieved probability category based on the survey’s estimates of proportions of workers exposed to asbestos as determined by walk-through surveys of 5000 U S . workplaces by industrial hygienists in 1972- 1974.13 Asbestos exposure was defined as a nonzero value for at least one of the five measures listed above. The unexposed group for all measures of asbestos was comprised of those individuals who had none of the five asbestos exposures (no reported exposure to asbestos, none of the nine specified activities, no job with a National Occupational Hazard Survey likelihood of exposure greater than zero, no cohabitants with asbestos exposures, and no reported residence within two miles of an asbestos mine or mill). Relative risks (as odds ratios) and attributable risks for asbestos exposure were previously reported.2 The original questionnaire data were examined for all individuals whose next-of-kin reported that the individual had a family history of cancer. We had no means of verifying these family history data reported by nextof-kin. Cancers reported to have occurred in relatives other than the mother, father, sister, brother, daughter, or son were excluded, and the remaining cancers in FDRs were grouped by relative and reported site. Because of an interest in the familial Li-Fraumeni cancer spndrome,I4we grouped reports of leukemia, soft tissue sarcoma, or cancer of the breast, brain, bone, or adrenal cortex in one or more FDRs as “Li-Fraumeni” cancers. Data Collection Telephone interviews were conducted with (in order of preference) the spouse, child, sibling, or other relative or friend of patients and controls. Verbal informed consent was obtained from all respondents. Using a standardized questionnaire, trained interviewers collected information on demographic factors, occupational history, tobacco smoking, medical history, exposure of cohabitants to asbestos or asbestos-related work, and cancer in parents, brothers, sisters, or children (first-degree relatives [FDRs]).Exposure to asbestos was determined by several measures: self-reported exposure to asbestos on or off the job (“ever exposed”); involvement in nine specific activities thought likely to entail asbestos exposure, such as insulation or shipyard work (“9 activities”); a job-exposure matrix based on lifetime work history;z reported exposure of a cohabitant of the individual to asbestos or the cohabitant’s participation in one of the nine activities; Analyses Twenty-five men and 5 women lacking information on a FDR’s history of cancer (including 2 who also lacked information on smoking), and 4 additional individuals without smoking data were excluded from the analyses, leaving 196 patients (172 men; 24 women) and 511 controls (406 men; 105 women). Maximum likelihood estimates of the odds ratio and 95% confidence intervals were calculated using unconditional logistic regression methods,Is separately for men and women, adjusting for age in 4 categories (1-49, 50-59, 60-69, and 70+ years), cigarette smoking (everlnever, because smoking was associated with being a control and confounded the association with asbestos), and geographic area (New York, Louisiana, Virginia). All family history analyses were performed controlling for asbestos exposure or within asbestos-defined subgroups. Individuals with no cancers re- Mesothelioma and Family History of Cancer/Heineman et al. 551 TABLE 1 Relative Risk of Mesothelioma by Reported Number of First-DemeeRelatives with Cancer and Li-Fraumeni Syndrome (Males and Females Separately and Combined, Cont;olling for Age, Smoking History, and Geographic AreaY Males No asbestos exposure Asbestos-exposedh Ca Co Any FDRs with cancer 95 170 No Yes 68 99 Number of FDRs with ca 1 415 81 2 or more 22 18 li 2 Pvalue' .08 20 1 Asbestos-exposed No asbestos exposure All individuals OR 95%CI Ca Co OR 95%CI Ca Co OR 95%CI Ca Co OR 95%CI Ca Co OR 95%CI 1.0 I.? G 3 101 36 1.0 1.4 - 27 27 1.0 0.9 33 18 1.0 1.2 (0.2-8.41 117 79 331 (0.2-3.6) 9 2 - (0.3-6.3) 7 6 - (0.8-1.8) 180 1.0 1.3 1.0 2.2 (0.6-1.6) (1.1-4.4) 3 0 28 8 1.8 (0.4-8.1) - - 4 2 18 9 1.1 (0.2-5.0) 0.7 (0.1-4.7) 1 1 12 6 1.0 1.5 (0.1-12.61 [O.l-19.1) 54 25 139 41 1.1 (0.8-i.;i 1.7 (0.9-3) - Trend test Pvalue' .OR Any Li-Fraumeni cancers in FDRs Yes 19 21 1.7 (0.9-3.5) Number of FDRs with Li-Fraumeni cancers I 2 or more Trend rest Females 1.6 5.1 (0.8-3.3) (0.4-59.7) .99 0 - I 3 4.4 (0.1-226.4) 0 0 - 1 0 3 4.4 (0.1-226.4) 0 - - - - 22 .7i .G3 I 7 1.4 (0.1-16.:) 21 44 1.5 (0.8-2.81 (0.1-16.71 19 2 43 1.4 5.9 (0.8-2.7) (0.4-81.7) 1 7 1.4 n o - - - (0.9-1.8) 1 .I2 Ca: numher of rases; Cn: number of controls: OR: odds ratio: CI: confidence interval: FDR: first-degreerelative 'Also adjusted lor gender and asbestos exposure. "Individualswh3 had at least one positive answer among all the asbestos exposure measures:the direct question "Was the subject ever exposed to asbestos?;nine specified activities,such as shipyardor insulation work: at least one job with a National OccupationalHazard Survey likelihoodof exposure grealer than 0: a cohabitant with asbestos exposure or asbestosdated activities;or individuals reported residence within nvo mile$of an asbestos mine or mill. ' The P value reflects the signifi1:ance of a linear trend for the 3 categories 0. I , and 2 t (more than 1):hvo few individuals reported more than 2 family members with cancer to further subdivide the highest category ported arnong FDRs were the referent group for all family history variables. We tested for effect modification of family history and asbestos exposure by inclusion of interaction terms in the logistic regression model. Linear trends (two-sided test) were evaluated by testing the statistical significance of a continuous variable in the model. RESULTS Men and women whose next-of-kin reported cancer in FDRs had a 30% excess of mesothelioma, which was not statistically significant, after adjusting for asbestos exposure (defined as any exposure to asbestos whether workplace, cohabitant, residential or other reported exposure) (Table 1). Among men exposed to asbestos, we found a statistically significant twofold elevation in mesothelioma risk for report of cancer in two or more FDRs, but there was little evidence of a linear trend in risk with the number of FDRs with cancer (Table 1). Risk estimates were larger in both men and women for the subset of "LiFraumeni" cancers in FDRs, and rose to fivefold among asbestos-exposed men with two or more such relatives, but were not statistically significant. We found no significant associations with the summary family history mea- sures among women, nor among the small number of men without asbestos exposure. Among men with asbestos exposure, only liver cancer in a FDR was significantly associated with a risk of mesothelioma (Table 2). A nearly threefold increase with familial history of brain cancer, although nonsignificant, was notable because of the apparent consistency across most relatives. Twofold or higher risks of mesothelioma were observed among individuals whose FDR had colon or uterine cancer. Numbers were too small to permit meaningful site- or relative-specific analyses among men without asbestos exposure or among women. Associations of asbestos with pleural mesothelioma were stronger among men with a family history of cancer than men without such a history (Table 3). A similar pattern was seen for women (pleural and peritoneal mesothelioma combined due to small numbers). However, risk estimates had wide, overlapping confidence intervals, and no statistical evidence of interaction was detected. For pleural and peritoneal mesothelioma combined in men, risks were only slightly higher for men with a family history of cancer than men without one (not shown). The next-of-kin of three patients and no controls re- 552 CANCER February 1,1996 / Volume 77 / Number 3 TABLE 2 Men with Anv Asbestos Exposure (Relative Risk of Mesothelioma by. Reported History of Cancer in First-Degree Relatives. Controlling for Age, . Smokine, and Geoesaohic heal Relation Type of cancer Any relative Ca 68 Upper aerodigestiven Digestive and peritoneum 7 18 Stomach 3 Colon 6 Liver 8 Respiratory and intrathoracic, excluding larynxb Lung 10 Bone, connective tissue, etc. 9 Bone 1 Breast 8 Genitourinary 7 Uterus 3 Prostate 2 Lymphatic and hematopoietic 5 Leukemia 3 Hodgkin's disease 2 Other site 9 10 Eye 1 Brain 8 Thyroid 1 Unkriown site 19 OR (95% GI) 1.2 (0.8-1.8) 1.3 (053.51 1.2 (0.6-2.3) 0.3 (0.1-1.3) 2.0 (0.6-6.4) 5.1 (1.3-20.2) 1.6 (0.7-4) 1.5 (0.6-3.8) 0.9 (0.4-2.11 0.8 (0.1-7.7) 1.6 (0.6-4.3) 1.0 (0.4-2.9) 2.4 (0.4-15) 1.2 (0.2-9.5) 1.4 (0.4-4.9) 1.5 (0.3-7.6) 1.9 (0.3-14.4) 2.4 (0.9-7) 1.4 (0.1-25.6) 2.8 (0.8-9.3) 1.9 (0.1-31.4) 1.5 10.8-3) Mother Ca 25 OR (95%CI) 1.2 (0.7-2.1) Father Ca 18 1 L 3 7 1.0 (0.4-2.8) 0.5 (0.1 -2.4) 2.4 (0.5-11.5) 1.5 (0.1-26.2) 4.3 (0.4-50.4) 4.3 (0.4-50.41 0.9 (0.2-3.8) 6 2 4 I 2 2 3 0 3 1 1.3 (0.3-5.6) 0.3 (0.0-2.9) 0 1 OR (95% GI) 1.2 (0.6-2.4) 1.6 (0.3-7.3) 1.7 (0.5-5.4) 0 1 5 3 3 2.6 (0.2-44.4) 4.8 (0.9-25.6) 2.3 (0.4-12.4) 2.3 (0.4- 12.4) 1 1 5 0 4 0.7 10.2-2.7) 1.4 (0.3-6.2) 5.6 (0.6-57.5) L 0 1 1.2 (0.1 - 14.1) ZI 6.9 (0.7-72.4) 1.o (0.5-1.8) 1.1 (0.2-7) 0.4 (0-3.3) 1.0 (0.1- 11.2) 1.9 (0.5-7) 1.6 (0.4-6.4) OR (95% GI) 1.7 (0.3-8.i) Ca OR (95% GI) 2 r (-1 0 0 0 0 0 0 0 0 0 0 0 3 6.5 (0.6-65.4) 0 0 1 ZI 0 0 2 4.3 (0.4-50.41 0 0 1 x I 0 1 0 0 0 1 2.4 (0.4-14.9) 1.5 (0.2- 115) 0 0 0 0 3 1 2.2 (0.1-38) 1 Y, 0 0 2 X 1 0.6 (0.1-5.9) 0 l 0 0 0 l 0 0 0 0 0 0.4 (0.0-3.3) Ca Son 0 2 0 OR (95%GI) 0 5 2 3.1 (0.5-19.5) 1 0 3 2.6 (0.2 -3 1.9) 1.4 (0.1-25.6) 2 0 3 1.3 (0.1-21.2) 1.3 (0.1-21 2) 18 2 0 0.4 (0-3.5) 1.6 (0.9-3.0) 1.4 (0.2-9) 2.2 (0.6-8.1) 0.7 (0.1-7.4) Ca 1 0 1 OR (95% GI) Daughter Brother 0 0 0.7 (0.1-7.9) 1.6 (0.6-4.2) 2 3 0 8 22 0 0 2 Ca 1 n 0 1 Sister 0 1 x 1 1 X 0 7 4.0 (1.0-17) 6 0.6 (0.1-5.91 1.3 (0.4-4.2) x 7- r a La: number of cases; OR adds ratio; CI: confidence interval. 'Includes cancer of oral cavity. pharynx, "throat" and larynx Excluder "throat" and lalynx, which were grouped with upper aerodigestive tract cancers ported a possible mesothelioma among the patient's immediate family. The father of a woman with pleural mesothelioma was reported to have had "asbestos cancer." (We assume this is likely to be mesothelioma or lung cancer.) He was reported to have engaged in demolition, shipyard, and insulation work while living with the patient; she had no other reported history of asbestos exposure. A man with pleural mesothelioma whose brother was reported to have also had mesothelioma worked for 20 years as a heavy equipment mechanic on a railroad Mesothelioma and Family History of Cancer/Heineman et al. 553 TABLE 3 Relative Risk of Mesothelioma with Several Measures of Asbestos Exposure among Study Patients With and Without Self-reported History of Cancer in First-Degree Relatives Men with pleural mesothelioma No family history Family history of cancer ~~ ~~ ~ Ca co OR 95%CI Ca co OR 95%CI 1 1.0 28.5 270.9 33.4 16.5 55.2 89.9 - 5 86 62 73 1 36 99 22 74 39 17 5 24 5 101 170 51 135 60 34 7 1.0 11.0 24.9 11.8 7.0 15.9 19.8 4.3-28.41 9.2-67.4 4.5-30.8 2.4-20.9 5.5-46 3.5-1 13.9 15 26 32.5 19 40 12.4 3.9-39.43 ~~ Never expo’ied Any asbestos exposure Ever exposed Any of 9 activities NOHS likelihood 1-19% ‘OtW Cohabitant exposed Cohabitant involved in any of 9 activities 61 41 52 15 23 3.7-221.5 17.2-4262.1 4.2-263.0 2-137.2 6.6-465.3 6-1350.4 18 3.4-307.2 - Women Family history of cancer Never exptrsed Any asbestos exposure Every expoed Any of 9 activities NOHS likelihood 1- 19% 20+% Cohabitant exposed Cohabitant involved in any of 9 activities Ca co 2 6 3 1 1 0 2 18 27 9 I1 3 0 6 3 20 OR 1.0 2.0 6.8 1.2 No family history 95% c1 - 0.3-13.8 0.5-97 0-26.8 Ca 9 I 2 3 co 33 27 6 10 OR 1.o 1.3 1.o 2.3 95% CI - 0.4-4.4 0.1-6.7 0.4-14.8 1.8 - - - - 0 0 5 - 0 - - Zld - 0 5 - - 5 21 4.1 0.2-93.9 1.0 0.2-4 rA: number of caws: Co: number of controls; OR: odds ratio; CI:confidence interval; NOHS: National Occupational Hazard Survey, Patientjsl and controls are in mutally exclusive strata. ” with duties that included wrapping pipes with asbestos insulation: we have no information on the potential exposures of the brother of this patient. The mother of a male patient diagnosed with peritoneal mesothelioma was reported to have had the “same cancer as the [case]”.This man worked for 37 years, beginning at age 19, as a pipefitter in the same small midwestern town. Because of his residential stability, it is possible that his mother could have been exposed to asbestos on his work clothing. These reports should be viewed cautiously because, as noted earlier, only 39% of patients identified from tumor registries were confirmed as having probable or definite mesothelioma. DISCUSSION Asbestos clearly accounts for the vast majority of mesothelioma cases, particularly among men, in whom the attributable risk ranges from 45% to 75% for occupational exposure or 85% for any known exposure to asbestos.’ In this large case-control study, we evaluated whether a family cancer history plays an independent role in meso- thelioma and whether such a history, which might indicate heightened cancer susceptibility, augments the effect of asbestos. We found suggestive but limited evidence of such effects. In men exposed to asbestos, a statistically significant elevation in risk for cancer in two or more FDRs was not accompanied by a significant trend with number of relatives, although numbers of patients with multiple relatives were small. In women, a family history of cancer was not significantly related to the risk of mesothelioma. In men with pleural mesothelioma, associations of asbestos with mesothelioma appeared to be stronger among those with a family history of cancer than those without one. However, the asbestos associations comparing individuals with a family history with those without such a history do not differ statistically, so the appearance of effect modification may be due to chance. It is important to note that our statistical power to identify a modest risk with family history, and particularly assess interaction, was limited by small numbers of patients and controls (especially patients without asbestos exposure). We could not confirm a threefold risk of mesotheli- 554 CANCER February 1, 1996 / Volume 77 / Number 3 oma among individuals whose parents had gastrointestinal malignancies, as reported in the previous epidemiologic study that evaluated family cancer history,” although men whose relatives were reported to have liver cancer did have a significantly increased risk of mesothelioma. Apparent increases in risk with numbers of relatives with cancers that are part of the Li-Fraumeni syndrome (leukemia, soft tissue sarcoma, or cancer of the breast, brain, bone, or adrenal cortex) are intriguing, but could be explained by chance. We are not aware of any reason why liver cancers or the “Li-Fraumeni” cancers might be particularly associated with mesothelioma. Mesothelioma is not considered a typical “Li-Fraumeni” cancer, although at least one mesothelioma has been reported in family members with this cancer syndrome.’“ Site-specific analyses are only suggestive because, as noted earlier, fanlily members’ cancers could not be validated, and some common sites of metastases, such as the liver, may have been overreported among both patients and controls. We did not have information on family size or composition, or the age at which the relative was diagnosed with cancer, which would have allowed a more sophisticated analysis. These limitations, along with modest statistical power, might have diminished our ability to detect a true effect. Because we could not validate family histories, it is possible that reporting bias may be responsible in part for some of the apparent increases. Patients’ nextof-kin were slightly more likely than controls’ next-ofkin to volunteer cancers in more distant relatives, about whom the interviewer did not inquire (5% vs. 3%). The exposure of interest, however, was limited to cancers in FDRs to reduce this potential bias. Next-of-kin of both patients and controls are likely to have better knowledge about first-degree relatives than about more distant relatives, tending to equalize their ability to report cancers in these relatives. Although two or three patients (and no controls) reported mesothelioma in other family members, it is not possible to determine from the available data whether they are due to a familial tendency or shared exposures. In two of the patients, it is plausible that mesothelioma (or “asbestos cancer”) in one family member was due to asbestos exposure in the workplace, and asbestos carried home by that person might explain the second mesothelioma. As mentioned earlier, mesothelioma is difficult to diagnose, and these reported familial cases may not have been true niesotheliomas. Although our results provide limited evidence that family history of cancer may be a risk factor for mesothelioma, we cannot rule out a modest effect or specific associations with cancer in particular family members or with certain tumor types. The tendency for risks with asbestos exposure to be higher among individuals with a family history of cancer might indicate an increased susceptibility to mesothelioma, as suggested by Vianna and Polan,” but could also be a chance result. As heavy and prolonged exposure to asbestos becomes increasingly less common, it may be possible to separate the overwhelming impact of asbestos from that of other factors that contribute to this rare disease. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. McDonald AD, McDonald JC. Epidemiology of malignant mesothelioma. In: Antman K, Aisner 7, editors. Asbestosrelated malignancy. New York Grune & Stratton, Inc., 1986. Spirtas R, Heineman EF, Bernstein L, Beebe GW, Keehn RJ, Stark A, et al. Malignant mesothelioma: attributable risk of asbestos exposure. Occup Environ Med 1994;51:804- 11. Stanton MF, Wrench C. Mechanisms of mesothelioma induction with asbestos and fibrous glass. / Nut1 Cancer Inst 1972;48:797-82 1. Warren S, Brown CE, Chute RN, Federman M. Mesothelioma relative to asbestos, radiation, and methylcholanthrene. Arch Padiol Lab Med 1981;105:305-12. Peterson JT, Greenberg SD, Buffler PA. Non-asbestos-related malignant mesothelioma: a review. Cancer 1984;54:951-60. Krousel T, Garcas N, Rothschild H. Familial clustering of mesothelioma: a report on three affected persons in one family. Am / Prev Med 1986;2:186-8. Hammar SP, Bockus D, Remington F, Freidman S, LaZerte G. Familial mesothelioma: a report of two families. Hum Padlo1 1989;20:107- 12. Otte KE, Sigsgaard TI, Kjaerulff 1. Malignant mesothelioma: clustering in a family producing asbestos cement in their home. B r / Ind Med 1990;47:10-3. Dawson A, Gibbs A, Browne K, Pooley F, Griffiths M. Familial mesothelioma. Details of 17 cases with histopathologic findings and mineral analysis. Cancer 1992;70:1183-7. Bianchi C, Brollo A, Zuch C. Asbestos-related familial mesothelioma. Eur J Cancer Prev 1993;2:247-50. Vianna NJ, Polan AK. Nonoccupational exposure to asbestos and malignant mesothelioma in females. Lancet 1978;I: 1061-3. Spirtas R, Keehn RJ, Beebe GW, Wagner IC, Hochholzer L, Davies JNP, et al. Results of a pathology review of recent US mesothelioma cases. Accompl Oncol 1986;1:144-52. National Occupational Hazard Survey, Vol 111: survey analysis and supplemental tables. Cincinnati, (OH): National Institute for Occupational Safety and Health; 1977 X DHEW Publication No. (NIOSH) 78-114. Li FP, Fraumeni JF Jr., Mulvihill JJ, Blattner WA, Dryfus MG, Tucker MA, et al. A cancer family syndrome in twenty-four kindreds. Cancer Res 1988;48:5358-62. Kleinbaum DK, Kupper LL, Morgenstern H. Epidemiologic research. Belmont, CA: Lifetime Learning Publications, 1982.