Int. J. Cancer: 65, 432-436 (1996)
UICC
Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer
RISK FACTORS FOR ADRENAL CANCER: AN EXPLORATORY STUDY
Ann W. HSING1,4, Jun-Mo NAM1, Harvey T. Co CHIEN2, Joseph K. MCLAUGHLIN3 and Joseph F. FRAUMENI, JR.1
1Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD; 2Westat Inc. and 3International Epidemiology Institute, Rockville, MD, USA.
Adrenal cancer is a heterogeneous group of neoplasms with unknown etiology. In search of risk factors, we conducted a case-control study based on data from the 1986 National Mortality Followback Survey, which included a questionnaire sent to the next of kin of almost 20,000 deceased adults (age ≥25 years) in the United States. Information was obtained on a large number of items, including use of cigarettes, alcohol, oral contraceptives (OCs), height and weight and food consumption patterns. A total of 176 subjects who died of adrenal cancer (88 men and 88 women) and 352 controls (176 men and 176 women) who died of causes unrelated to smoking, drinking or OCs (for female controls) were included in the study. Although information on histologic type was not available, most cases were estimated from incidence surveys to be adrenocortical carcinoma, with a small percentage being malignant pheochro- mocytoma or neuroblastoma. An increased risk was associated with heavy smoking (≥25 cigarettes/day) among men (odds ratio [OR] = 2.0, 95% confidence interval [CI] 1.0-4.4) but not women. No clear association was seen for alcohol use, height and weight or food consumption patterns in either sex. Among women, increased risks were found for ever users of OCs (OR = 1.8, 95% CI 1.0-3.2) and especially those who used them before age 25 (OR = 2.5, 95% CI 1.2-5.5). When the analysis was restricted to subjects with spousal respondents, more pronounced risks were seen for ever users of OCs and for those who used OCs before age 25. Our findings suggest that cigarette smoking and use of OCs may increase the risk of adrenal cancer, but additional studies are needed with more detailed informa- tion on risk factors and histologic type of adrenal cancer. @ 1996 Wiley-Liss, Inc .*
Cancers of the adrenal gland arise from either the adrenal cortex or the medulla (Robbins and Kumar, 1987), and they are extremely rare. Based on data from the Surveillance, Epidemiology and End Results (SEER) program in the United States 1975-1992, the average annual age-adjusted incidence rates were 0.29 per 100,000 for white males, 0.25 for black males, 0.25 for white females and 0.18 for black females (Hsing et al., in press). The annual age-adjusted U.S. mortality rates for adrenal cancer 1985-1992 were 0.24 per 100,000 for white males, 0.23 for black males, 0.21 for white females and 0.18 for black females (Hsing et al., in press), which resembles the incidence rates, suggesting that adrenal cancer is usually fatal. Adrenal cancer is a heterogeneous group of neoplasms, with 34% at all ages classified as adrenocortical carcinoma, 28% as neuroblastoma (including ganglioblastoma), 8% as malignant, pheochromocytoma and 30% as other malignancies (mostly poorly specified; Hsing et al., in press). Adrenocortical carcino- mas occur mainly in adults (58% of cases were over age 50 years), while 90% of adrenal neuroblastomas arise in children 0-4 years of age (Hsing et al., in press).
The role of genetic susceptibility in adrenal neoplasms has been suggested by the relation of adrenocortical carcinoma to Li-Fraumeni syndrome (Garber et al., 1991), congenital hemi- hypertrophy (Fraumeni and Miller, 1967) and Beckwith- Wiedemann syndrome (Henry et al., 1989); pheochromocytoma to multiple endocrine neoplasia type 2 and von Hippel-Lindau syndrome (Neumann et al., 1993); and neuroblastoma to mutations of the gene responsible for neurofibromatosis type 1 (The et al., 1993). Environmental determinants of these tumors are obscure, though geographic variation has been suggested by the elevated risks of adrenocortical cancer in Brazil (Stiller
et al., 1994) and the deficit of neuroblastoma in tropical areas of Africa (Miller, 1977).
In search of risk factors, we conducted a case-control study using data from the 1986 National Mortality Followback Survey (NMFS), which included a questionnaire sent to the next of kin of almost 20,000 deceased adults (age ≥ 25 years) in the United States. Although information on histologic type was not available, most cases in this study probably were either adrenocortical carcinoma or poorly specified adrenal malignan- cies since among adult patients (age ≥25 years) with adrenal cancer neuroblastoma and malignant pheochromocytoma ac- count for 12% of the adrenal cancer cases reported in national incidence surveys (Hsing et al., in press).
SUBJECTS AND METHODS
Study subjects were selected from the 18,733 decedents included in the 1986 NMFS, conducted by the National Center for Health Statistics (NCHS). Details of this study have been reported elsewhere (Seeman et al., 1989; Hsing et al., 1992). Briefly, a 10% systematic sample of 1986 U.S. death certifi- cates, excluding Oregon because of the state’s respondent consent requirement, was sent by each of the states to NCHS. From these death certificates, a probability sample comprising approximately 1% of U.S. adult deaths (age 25 years or older) was selected. In addition, among whites aged 25-74 years, all 1985 deaths from several rare cancers, including cancer of the adrenal gland, were ascertained and included in the study.
Questionnaires were sent to next of kin of the selected decedents to obtain information on the subjects’ demographic characteristics, use of cigarettes, alcohol, oral contraceptives (OCs), dietary habits (e.g., frequency of consumption of meat, vegetables and fruits), height and weight and medical history. The response rate for the informant questionnaire was 89%.
A total of 190 deaths from cancer of the adrenal gland (ICD-9 code 194.0) were included in the NMFS (18 from the 10% sample of U.S. deaths in 1986 and 172 from all adrenal cancer deaths in 1985). After exclusion of the few subjects who were non-whites (n = 2) and the non-respondents (n = 12), 176 (88 men and 88 women) cases were available for analysis.
Controls were selected from white decedents dying of causes other than cancer of the adrenal gland whose next of kin completed a questionnaire. Excluded as potential controls were subjects who died of smoking- or alcohol-related causes or of the other 5 rare cancers selected for study (nasopharynx, nasal cavity, small intestine, male breast and primary liver cancer in young women). For female controls, deaths related to OC use were also excluded from potential controls.
Among the 792 eligible male and 317 female controls, 2 controls per case were randomly selected from the matching sex- and age-specific (5-year age groups) stratum. In total, 176
4To whom correspondence and reprint requests should be sent, at Division of Cancer Epidemiology and Genetics, National Cancer Institute, EPN 415, 6130 Executive Blvd., Bethesda, MD 20892-7368, USA. Fax: (301) 402-0081.
Received: August 19, 1995 and in revised form October 24, 1995.
male and 176 female controls were included in the analysis. The major causes of death among male controls were accident and injury (16%), diabetes mellitus (13%), lymphoma (7%), brain cancer (6%) and prostate cancer (4%). Among female controls, major causes of death were accident and injury (15%), lymphoma (10%), brain cancer (7%), skin melanoma (7%) and rectal cancer (5%).
Odds ratios (OR) and corresponding 95% confidence inter- vals (CI) for adrenal cancer in relation to potential risk factors were estimated using the exact method (Gart, 1971; Thomas, 1975). Tests for linear trends in proportions were also per- formed (Cochran, 1954; Armitage, 1955). Potential confound- ing effects of age, income, education and marital status were examined and adjusted for when necessary, using multiple logistic regression (Breslow and Day, 1980).
RESULTS
A total of 176 cases (88 men and 88 women) and 352 controls (176 men and 176 women) were included in the analysis. The median age at death for cases was 54 years for men and 52 years for women. Selected characteristics for cases dying of adrenal cancer and for their controls are shown in Table I. Compared to controls, cases were more likely to be married and to have a higher income. Cases tended to have a slightly higher educational level than controls. For male subjects, spouses were the main respondents, while spouses and parents were the major surrogate respondents for female subjects.
Table II shows that male heavy smokers (≥ 25 cigarettes/ day) had a 2-fold increased risk of adrenal cancer (95% CI 1.0-4.4). Current smokers had a 1.6-fold risk, which was not statistically significant. No increased risk was associated with smoking among women or with alcohol drinking in either sex. Limiting responses to spousal informants did not affect the results for tobacco or alcohol use.
As shown in Table III, women who ever used OCs (OR = 1.8; 95% CI = 1.0-3.2) and those who used them before age 25 years (OR = 2.5; 95% CI = 1.2-5.5) had an increased risk of adrenal cancer. In addition, among subjects with a spousal respondent, risks were significantly elevated for ever users of OCs (OR = 2.4; 95% CI 1.0-5.4) and for those who used them before age 25 years (OR = 2.8; 95% CI 1.0-7.5). A 5-fold risk was found for those who used OCs before age 25 years and for more than 5 years (OR = 5.1; 95% CI 1.5-16.7). No informa- tion was available on current use (prior to death), type or dosage of OCs or on time since last use.
No clear association was found with consumption of meat, vegetables, fruits, cured meat or dairy products or with body mass index in either sex, though a statistically significant association was found for fruit consumption among women and dairy food among men (Table IV).
DISCUSSION
Our exploratory case-control study of adrenal cancer sug- gests that cigarette smoking and use of OCs are potential risk factors. No association was seen with alcohol, food consump- tion patterns, height, weight or body mass index in either sex. Although data were not available to enable an assessment of risk by histologic type, our findings pertain largely to adrenocor- tical carcinoma since we included only subjects between the ages of 25 and 64 years, and in this age group, adrenal neuroblastoma and malignant pheochromocytoma constitute only 12% of adrenal cancer cases in national incidence surveys (Hsing et al., in press). Furthermore, among SEER patients aged 25 years and over who were diagnosed with adrenal cancer and had adrenal cancer mentioned as an underlying cause of death on their death certificates, only 9% were reported with pheochromocytoma and less than 1% with neuroblastoma (data not shown).
| Males | Females | |||||||
|---|---|---|---|---|---|---|---|---|
| Cases | Controls | Cases | Controls | |||||
| N | % | N | % | N | % | N | % | |
| Total | 88 | 100.0 | 176 | 100.0 | 88 | 100.0 | 176 | 100 |
| Age at death (yr) | ||||||||
| 25-34 | 14 | 15.9 | 28 | 15.9 | 12 | 13.6 | 24 | 13.6 |
| 35-44 | 11 | 12.5 | 22 | 12.5 | 17 | 19.3 | 37 | 21.0 |
| 45-54 | 21 | 23.9 | 42 | 23.9 | 22 | 25.0 | 43 | 24.4 |
| 55-64 | 42 | 47.7 | 84 | 47.7 | 37 | 42.1 | 72 | 40.9 |
| Marital status at death | ||||||||
| Never married | 9 | 10.2 | 39 | 22.2 | 3 | 3.4 | 18 | 10.2 |
| Divorced/separated | 10 | 11.4 | 30 | 17.1 | 16 | 18.2 | 32 | 18.2 |
| Widowed | 3 | 3.4 | 5 | 2.8 | 11 | 12.5 | 21 | 11.9 |
| Married | 64 | 72.7 | 96 | 54.6 | 54 | 61.4 | 101 | 57.4 |
| Unknown | 2 | 2.3 | 6 | 3.4 | 4 | 4.5 | 4 | 2.3 |
| Education (yr) | ||||||||
| <9 | 12 | 13.6 | 29 | 16.5 | 7 | 8.0 | 24 | 13.6 |
| 9-11 | 13 | 14.8 | 38 | 21.6 | 10 | 11.4 | 30 | 17.0 |
| 12 | 28 | 31.8 | 49 | 27.8 | 40 | 45.4 | 70 | 39.8 |
| >12 | 29 | 33.0 | 50 | 28.4 | 23 | 26.1 | 48 | 27.3 |
| Unknown | 6 | 6.8 | 10 | 5.7 | 8 | 9.1 | 4 | 2.3 |
| Total annual family income | ||||||||
| <$11,000 | 19 | 21.6 | 42 | 23.9 | 17 | 19.3 | 51 | 29.0 |
| $11,000-$24,999 | 21 | 23.9 | 36 | 20.4 | 24 | 27.3 | 39 | 22.2 |
| ≥$25,000 | 33 | 37.5 | 45 | 25.6 | 31 | 35.2 | 46 | 26.1 |
| Unknown | 15 | 17.0 | 53 | 30.1 | 16 | 18.2 | 40 | 22.7 |
| Type of respondent | ||||||||
| Spouse | 52 | 59.1 | 85 | 48.3 | 43 | 48.9 | 79 | 44.9 |
| Parent | 6 | 6.8 | 17 | 9.7 | 18 | 20.4 | 28 | 15.9 |
| Child | 10 | 11.4 | 30 | 17.0 | 13 | 14.8 | 27 | 15.3 |
| Sibling | 10 | 11.4 | 21 | 11.9 | 4 | 4.5 | 19 | 10.8 |
| Other | 10 | 11.4 | 23 | 13.1 | 10 | 11.4 | 23 | 13.1 |
| Males | Females | |||||||
|---|---|---|---|---|---|---|---|---|
| Cases | Controls | OR' | 95% CI | Cases | Controls | OR' | 95% CI | |
| Cigarette use | ||||||||
| Non-smoker | 17 | 46 | 1.0 | - | 35 | 73 | 1.0 | - |
| Current smoker | 35 | 62 | 1.6 | 0.8-3.3 | 26 | 60 | 1.0 | 0.6-1.9 |
| Ex-smoker | 34 | 56 | 1.4 | 0.7-2.9 | 25 | 39 | 1.2 | 0.6-2.4 |
| Ever smoker | 69 | 121 | 1.4 | 0.8-2.8 | 52 | 100 | 1.1 | 0.6-1.9 |
| 1-14 cigarettes/day | 13 | 33 | 1.0 | 0.4-2.4 | 13 | 24 | 1.0 | 0.5-2.4 |
| 15-24 cigarettes/day | 22 | 39 | 1.4 | 0.6-3.1 | 20 | 40 | 1.1 | 0.5-2.1 |
| ≥ 25 cigarettes/day | 32 | 41 | 2.0 | 1.0-4.4 | 16 | 31 | 1.2 | 0.6-2.5 |
| Alcohol use | ||||||||
| Non-drinker | 10 | 22 | 1.0 | - | 15 | 42 | 1.0 | - |
| Drinker | 75 | 145 | 0.9 | 0.4-2.1 | 72 | 129 | 1.4 | 0.7-2.1 |
| <1 time/week | 25 | 54 | 0.7 | 0.3-1.9 | 47 | 73 | 1.5 | 0.7-3.1 |
| 1-2 times/week | 17 | 26 | 1.1 | 0.4-3.1 | 7 | 29 | 0.6 | 0.2-1.6 |
| ≥ 3 times/week | 27 | 60 | 0.8 | 0.3-2.1 | 15 | 26 | 1.3 | 0.5-3.3 |
1Adjusted for marital status and income.
| All subjects | Spouse respondents | |||||||
|---|---|---|---|---|---|---|---|---|
| Cases (n = 88) | Controls (n = 176) | OR1 | 95% CI | Cases (n = 43) | Controls (n = 79) | OR2 | 95% CI | |
| OC use | ||||||||
| Non-users | 52 | 118 | 1.0 | - | 22 | 57 | 1.0 | - |
| Ever users | 29 | 37 | 1.8 | 1.0-3.2 | 18 | 19 | 2.4 | 1.0-5.4 |
| Years of use | ||||||||
| <5 | 13 | 17 | 1.6 | 0.7-3.6 | 10 | 12 | 2.1 | 0.8-5.6 |
| ≥5 | 14 | 17 | 1.9 | 0.8-4.2 | 7 | 7 | 2.5 | 0.8-8.3 |
| Unknown | 9 | 24 | 1.0 | 0.4-2.5 | 4 | 3 | 3.6 | 0.7-17.8 |
| Age at first use (yr) | ||||||||
| <25 | 18 | 16 | 2.5 | 1.2-5.5 | 11 | 10 | 2.8 | 1.0-7.5 |
| ≥25 | 9 | 16 | 1.2 | 0.5-2.9 | 6 | 7 | 2.2 | 0.7-7.5 |
| Unknown | 9 | 26 | 0.9 | 0.4-2.2 | 4 | 5 | 2.0 | 0.5-8.6 |
1Adjusted for marital status and income .- 2Adjusted for income.
| Males | Females | |||||||
|---|---|---|---|---|---|---|---|---|
| Cases | Controls | OR' | 95% CI | Cases | Controls | OR2 | 95% CI | |
| Meat | ||||||||
| <3 times/week | 10 | 29 | 1.0 | - | 24 | 40 | 1.0 | - |
| 3-6 times/week | 43 | 72 | 1.2 | 0.5-2.9 | 38 | 76 | 0.7 | 0.4-1.4 |
| ≥ 7 times/week | 29 | 57 | 1.2 | 0.5-2.8 | 24 | 51 | 0.8 | 0.4-1.6 |
| Fruits | ||||||||
| <3 times/week | 17 | 49 | 1.0 | - | 12 | 54 | 1.0 | - |
| 3-6 times/week | 29 | 41 | 1.9 | 0.9-4.1 | 37 | 39 | 4.0 | 1.8-8.9 |
| ≥ 7 times/week | 34 | 67 | 1.7 | 0.8-3.6 | 35 | 74 | 2.2 | 1.0-4.8 |
| Vegetables | ||||||||
| <3 times/week | 2 | 8 | 1.0 | - | 4 | 15 | 1.0 | - |
| 3-6 times/week | 22 | 30 | 2.6 | 0.3-14.1 | 16 | 40 | 1.3 | 0.4-4.7 |
| ≥ 7 times/week | 58 | 119 | 1.7 | 0.3-9.0 | 66 | 112 | 2.2 | 0.7-7.0 |
| Cured meat | ||||||||
| <1 time/week | 21 | 53 | 1.0 | - | 29 | 56 | 1.0 | - |
| 1-2 times/week | 37 | 48 | 1.7 | 0.9-3.5 | 35 | 70 | 0.9 | 0.5-1.7 |
| ≥ 3 times/week | 23 | 50 | 1.0 | 0.5-2.2 | 20 | 41 | 1.0 | 0.5-2.0 |
| Dairy food | ||||||||
| <3 times/week | 3 | 21 | 1.0 | - | 20 | 29 | 1.0 | - |
| 3-6 times/week | 23 | 49 | 2.7 | 0.9-10.4 | 23 | 48 | 0.7 | 0.3-1.5 |
| ≥ 7 times/week | 57 | 88 | 4.3 | 1.2-15.4 | 42 | 90 | 0.7 | 0.3-1.4 |
| Body mass index | ||||||||
| Quartile 1 | 22 | 39 | 1.0 | - | 25 | 40 | 1.0 | - |
| Quartile 2 | 21 | 42 | 1.2 | 0.6-2.6 | 25 | 42 | 0.9 | 0.4-1.8 |
| Quartile 3 | 20 | 41 | 0.9 | 0.4-1.9 | 18 | 41 | 0.7 | 0.3-1.5 |
| Quartile 4 | 20 | 40 | 0.8 | 0.4-1.7 | 17 | 42 | 0.6 | 0.3-1.4 |
1Adjusted for marital status, income and smoking .- 2Adjusted for marital status, income and OC use.
Although the next-of-kin informant may have limited knowl- edge about the deceased subject’s exposure history, it has been shown that for broad categories of exposure, such as smoking, drinking and use of OCs, reliable information can be obtained from surrogate respondents, particularly a spouse (Glass et al., 1974; Thorogood and Vessey, 1989; Mclaughlin et al., 1990). Differential recall between surrogate respondents for cases and controls is unlikely since controls were also deceased. A large percentage of them had other cancers as well and respondents were probably not sensitized to any potential relationship of smoking and OC use with adrenal cancer. Due to anticipated recall problems with surrogate interviews, the NMFS questionnaire sought only limited information on particular exposures.
Although combination OCs have been reported to increase the risk of breast, cervical and hepatic cancers (IARC, 1987), an association with adrenal cancer has not been previously investigated, perhaps due to its low incidence and the absence of case-control studies. Experimental studies, however, have indicated a high risk of adrenal tumors in ovariectomized mice (Strickland et al., 1980) and in rats given exogenous estrogens (Noble et al., 1975). Since the OC associations we observed were borderline significant and the trend with duration of use was not strong, these findings need to be confirmed in future studies. We had no information on current use of OCs. In future studies it will be of interest to evaluate the risks among current and past users and in relation to cessation of use.
It is also important to clarify the smoking-related risk of adrenal cancer that was seen primarily among men in our study. There are some experimental data consistent with a smoking effect on the adrenal glands. In an inhalation study of
cigarette smoke in rats, a low but statistically significant incidence of adrenocortical carcinomas and adenomas was noted (Dalbey et al., 1980). In hamsters, the intratracheal administration of benzo(a)pyrene also resulted in a significant yield of adrenocortical adenomas (Beems and Beck, 1984; Beems, 1986). To our knowledge, adrenal tumors have not been linked to tobacco-specific nitrosamines (Hoffmann et al., 1984), though other N-nitroso compounds have induced adre- nocortical tumors in rats (Moore et al., 1989). In interpreting our findings, it is noteworthy that smoking and drinking are usually over-represented in dead controls (McLaughlin et al., 1985a,b). Although we excluded persons who died of alcohol- and smoking-related causes of death as potential controls, the prevalence of smoking among the male controls (32%) was still higher than that in the U.S. population (25%) during the time period of this study (US Surgeon General, 1989). This high frequency may have resulted in an under-estimate of the real association between smoking and adrenal cancer.
In summary, despite its limitations, this nationwide case- control study represents a systematic attempt to examine risk factors for adrenal cancer. Further investigations with more direct and detailed exposure information and specific histo- logic types of adrenal cancer are needed to clarify the risks that may be associated with cigarette smoking and use of OCs.
ACKNOWLEDGEMENT
We thank Dr. S.S. Devesa for providing the relevant SEER data regarding the proportion of adult patients with neuroblas- toma and pheochromocytoma who had adrenal cancer cited as the cause of death on their death certificates.
REFERENCES
ARMITAGE, P., Tests for linear trends in proportions and frequencies. Biometrics, 11, 375-386 (1955).
BEEMS, R.B., Dietary selenium- and benzo(a)pyrene-induced respira- tory tract tumors in hamsters. Carcinogenesis, 7, 485-489 (1986).
BEEMS, R.B. and BECK, L., Modifying effect of dietary fat on benzo(a)pyrene-induced respiratory tract tumors in hamsters. Carcino- genesis, 5, 413-417 (1984).
BRESLOW, N.E. and DAY, N.E., Statistical methods in cancer research, Vol. 1, The analysis of case-control studies. IARC Scientific Publ. 32. International Agency for Research on Cancer, Lyon (1980).
COCHRAN, W.G., Some methods for strengthening the common x2 tests. Biometrics, 10, 417-451 (1954).
DALBEY, W.E., NETTESHEIM, P., GRIESEMER, R., CATON, J.E. and GUÉRIN, M.R., Chronic inhalation of cigarette smoke by F344 rats. J. nat. Cancer Inst., 64, 383-390 (1980).
FRAUMENI, J.F., JR. and MILLER, R.W., Adrenocortical neoplasms with hemihypertrophy, brain tumors, and other disorders. J. Pediatr., 70, 129-138 (1967).
GARBER, J.E., GOLDSTEIN, A.M., KANTOR, A.F., DREYFUS, M.G., FRAUMENI, J.F., JR. and LI, F.P., Follow-up study of twenty-four families with Li-Fraumeni syndrome. Cancer Res., 51, 6094-6097 (1991).
GART, J.J., The comparison of proportions: a review of significance tests, confidence intervals and adjustments for stratification. Rev. Int. Statist. Inst., 39, 148-169 (1971).
GLASS, R., JOHNSON, B. and VESSEY, M., Accuracy of recall of histories of oral contraceptive use. Brit. J. prev. Soc. Med., 28, 273-275 (1974).
HENRY, I., JEANPIERRE, M., COUILLIN, P., BARICHARD, F., SERRE, J.L., JOUERNEL, H., LAMOUROUX, A., TURLEAU, C., DE GROUCHY, J. and JUNIEN, C., Molecular definition of the 11p15.5 region involved in Beckwith-Wiedemann syndrome and probably in predisposition to adrenocortical carcinoma. Hum. Genet., 81, 273-277 (1989).
HOFFMANN, D., RIVENSON, A., AMIN, S. and HECHT, S.S., Dose- response study of the carcinogenicity of tobacco-specific N-nitrosa- mines in F344 rats. J. Cancer Res. clin. Oncol., 108, 81-86 (1984).
HSING, A.W., HOOVER, R.N., MCLAUGHLIN, J.K., CO CHIEN, H.T., WACHOLDER, S., BLOT, W.J. and FRAUMENI, J.F., JR., Oral contracep- tives and primary liver cancer among young women. Cancer Causes Control, 3, 43-48 (1992).
HSING, A.W., NAM, J.M., DEVESA, S.S. and FRAUMENI, J.F., JR., Adrenal cancer rates in the United States: 1975-1992. Cancer (1996) (In press).
INTERNATIONAL AGENCY FOR RESEARCH ON CANCER, Overall evalua- tion of carcinogenicity: an updating of IARC monographs volumes 1 to 42. IARČ Monogr., Suppl. 7, 297-308 (1987).
MCLAUGHLIN, J.K., BLOT, W.J., MEHL, E.S. and MANDEL, J.S., Problems in the use of dead controls in case-control studies: I. General results. Amer. J. Epidemiol., 121, 131-139 (1985a).
MCLAUGHLIN, J.K., BLOT, W.J., MEHL, E.S. and MANDEL, J.S., Problems in the use of dead controls in case-control studies: II. Effect of excluding certain causes of death. Amer. J. Epidemiol., 122, 485-494 (1985b).
MCLAUGHLIN, J.K., MANDEL, J.S., MEHL, E.S. and BLOT, W.J., Comparison of next-of-kin with self-respondents regarding questions on cigarette, coffee, and alcohol consumption. Epidemiology, 1, 408- 412 (1990).
MILLER, R.W., Ethnic differences in cancer occurrence: genetic and environmental influences with particular reference to neuroblastoma. In: J.J. Mulvihill, R.W. Miller and J.F. Fraumeni, Jr., (eds.) Genetics of human cancer. pp. 1-14 Raven Press, New York (1977).
MOORE, M.A., WEBER, E., MAYER, D. and BANNASCH, P., Adrenal cortical foci and nodules in Sprague-Dawley rats treated with N- nitrosomorpholine: light and electron microscopic findings. Virchows Arch. B. Cell Pathol., 57, 91-97 (1989).
NEUMANN, H.P.H., BERGER, D.P., SIGMUND, G., BLUM, U., SCHMIDT, D., PARMER, R.J., VOLK, B. and KIRSTE, G., Pheochromocytomas, multiple endocrine neoplasia type 2, and von Hippel-Lindau disease. New Engl. J. Med., 329, 1531-1538 (1993).
NOBLE, E.L., HOCHACHKA, B.C. and KING, D., Spontaneous and estrogen-produced tumors in Nb rats and their behavior after transplan- tation. Cancer Res., 35, 766-780 (1975).
ROBBINS and KUMAR (eds.), Basic pathology (4th ed.), W.B. Saunders, Philadelphia (1987).
SEEMAN, I., POE, G.S. and MCLAUGHLIN, J.K., Design of the 1986 National Mortality Followback Survey: considerations on collecting data on decedents. Public Hlth. Rep., 104, 183-188 (1989).
STILLER, C.A., International variations in the incidence of childhood carcinomas. Cancer Epidemiol. Biomarkers Prev., 3, 305-310 (1994).
STRICKLAND, J.E., SAVIOLAKIS, G.A., WEISLOW, O.S., ALLEN, P.T., HELLMAN, A. and FOWLER, A.K., Spontaneous adrenal tumors in the aged, ovariectomized NIH Swiss mouse without enhanced retrovirus expression. Cancer Res., 40, 3570-3575 (1980).
THE, I., MURTHY, A.E., HANNIGAN, G.E., JACOBY, L.B., MENON, A.G., GUSELLA, J.F. and BERNARDS, A., Neurofibromatosis type 1 gene mutations in neuroblastoma. Nature (Genet.), 3, 62-66 (1993).
THOMAS, D.G., Exact and asymptotic methods for the combination of 2 × 2 tables. Comput. Biomed. Res., 8, 423-446 (1975).
THOROGOOD, M. and VESSEY, M., The reliability of surrogate informa- tion about oral contraceptive use, smoking, height and weight col- lected from men about their wives. Contraception, 39, 401-408 (1989). US SURGEON GENERAL, Reducing the health consequences of smok- ing: 25 years of progress. Department of Health and Human Services, Public Health Service, Office on Smoking and Health. DHHS Publ. 89-8411 (1989).