Cancers in Famih'es with AtaxiaTelangiectasia Daphne Morrell, Charles L. Chase, and Michael Swift

ABSTRACT: Cancer incidence was measured retrospectively in 574 close blood relatives of white ataxiatelangiectasia (A-T) patients and 213 spouse controls in 44 previously unreported families. The cancer incidence rate in the adult blood relatives was significant/), elevated over the rote in the spouse controls (rote ratio = 3.9, p < 0.01). For heterozygous carriers of the A-T sane, the relative risk of cancer was estimated to be 6.1 (p < 0.005) as compared with nonheterozygotes. The most frequent cancer site in the blood relatives was the female breast, with nine cancers observed. These Jindings provide further support for the hypothesis that heterozygotes for the A-T gene are predisposed to cancer. INTRODUCTION The incidence of cancer among homozygotes for the autosomal recessive syndrome ataxia-telangiectasia (A-T) is exceptionally high [1, 2]. Accumulated evidence from studies of A-T families suggests that cancer incidence is also elevated among A-T heterozygotes [3-5], who are believed to constitute approximately 1.4 percent of the general U.S. white population [6]. We studied 44 newly identified families of A-T patients for two reasons: 1) to test, in an independent sample of families, the hypothesis that A-T blood relatives have an elevated cancer rate; and 2) to obtain additional information about the specific cancer types associated with the A-T 8ene in heterozy8otes. MATEI~JALS AND MErH,~)DS Between Ausust 1984 and February 1987, 16 families of A-T patients were referred to us by physicians or others familiar with our A-T research program. In June 1987 we began actively searching for additional A-T families by writins to 599 pediatric neurologists and other physicians across the United States and askins them to refer any A-T patients examined in the last 10 years. By October 1, 1988, 28 new A-T families had been found and agreed to participate in the study, bringing the total number of families to 44. For each A-T patient referred, medical records were obtained to verify the diagnosis. The following relatives of the A-T patients were enrolled in the study: all grandparents, parents, aunts, uncles, siblinss, nieces, and nephews who were alive on January 1, 1930 or born since that date. In contrast to our recent study of 128 other A-T From the Biolosical Sciences Research Center, Genetics Division, Department of Medicine, School of Medicine, University of North Carolina at Chapel Hill, North Carolina.

Address reprint requests to: Daphne Metre/l, M.S.P.H., 325 B.S.R.C., CB# 7250, University of North Carolina, Chapel Hill, NC 27599. Received March 23, 1990; accepted April 19, 1990.

119 © 1990 Elsevier Science Publishing Co., Inc. 655 Avenue of the Americas, New York, NY 10010

Cancer Genet Cytogenet 50:119-123 (1990) 0165.4608/90/$03.50

120

N. Sadamori et al. families, first cousins were excluded. The spouses of all the married blood relatives in the sample were enrolled in the study as a comparison group, the "spouse controls." They are the best available comparison group, because they closely resemble the blood relatives regarding age, socioeconomic status, ethnic origins, and adult environmental exposures. The study population consisted of 574 blood relatives and 213 spouse controls. Detailed descriptions of the data collection and analysis procedures are provided elsewhere [4]. Health history questionnaires, death certificates, autopsy reports, and records of all major hospitalizations were obtained. Every data source indicating cancer was examined. Only cancers verified by a hospital or autopsy report or death certificate were included in analyses; five cancers, 9% of the total, were verified by a death certificate alone. The only anecdotal cancer, a melanoma in a male blood relative, was excluded from analyses because confirming records could not be located. For each study member, we calculated the person-years at risk for cancer between January 1, 1930, and the date the person died or returned his or her health questionnaire. Three analyses ware performed to test the hypothesis that the A-T gene may be associated with cancer in heterozygotes. First, a sex- and age-adjusted cancer rate ratio was computed, comparing the incidence rate of cancer from age 20 to 99 years in the A-T blood relatives to the incidence rate in the spouse controls [7, 8]. Although very informative, the rate ratio will underestimate the relative risk of cancer for A-T heterozygotes if the gene does predispose to cancer because the blood relative group contains a mixture of heterozygotes and nonheterozygotes. To circumvent this limitation, a second analysis was performed that estimates the relative risk of developing cancer specifically for A-T heterozygotes as compared with nonheterozygotes. To calculate this heterozygote risk, a maximum likelihood method was used that takes into consideration each relative's prior probability of heterozygosity for the A-T gene based on his or her relationship to the homozygote [4, 9]. Third, to evaluate site-specific cancer incidence in the A-T blood relatives, we calculated the number of cancers that would have been expected in these relatives based on their person-years at risk and the age-, race-, sex-, and calendar year-specific cancer incidence rates for the state of Connecticut [10]. One-tailed p values based on the Mantel-Haenszel Xz test [11] and the likelihood ratio test determined the statistical significance of elevated rate ratios and relative risk estimates, respectively. Differences between observed and expected site-specific cancers were not tested for statistical significance because the population of Connecticut may not be comparable to the A-T blood relatives with respect to risk factors for cancer aside from the A-T gene.

RESULTS Fifty cancers occurred in the A-T blood relatives, whereas three occurred in spouse controls. For relatives aged 20 years or more, the age-adjusted cancer rate ratio comparing blood relatives with spouse controls was 3.9 (p < 0.01). When each blood relative's prior probability of carrying the A-T gene was taken into account, the relative risk that cancer would develop in an adult heterozygous for A-T was estimated to be 6.1 (p < o.oos). The most frequent cancer site in the blood relatives was the female breast, with nine cancers observed in the 303 female blood relatives (Table 1). Three of these women had bilateral breast cancer, including one mother of an A-T homozygote. Age at diagnosis ranged from 36 to 71 years. For breast, as well as the five next most frequent cancer types shown in Table 1 (lung, prostate, stomach, hematologic and lymphoid, and melanoma), the observed number of cancer cases exceeded the number expected based on general population

121

Appearance Time of t(8;21) Table I

Observed and expected numbers of cancer cases among 574 blood relatives in 44 white families with ataxia-telangiectasia No. of cases Cancer type

Observed

Expected

Breast (female)

9

5.6

Lung

8

5.0

Prostate Stomt~ch Hematologic and lymphoid

7 6 5

2.3 1.2 3.0

Melanoma Pan~reas Bladder Colon and rectum Other

4 2 2 2 5

0.9 0.9 1.8 5.5 --

50

37.6

Total

Co_m__m__ents Bilateral cancers were diagnosed at ages 36 and 56 in one mother of an A-T homozygote In fathers of A-T homozygotes, two cases developed, at ages 45 and 68 years, with 0.2 cases expected Two lymphomas, two chronic lymphocytic leukemias, and one acute lymphocytic leukemia.

One each: salivary gland, uterus, testis, rhabdomyosarcoma, and unknown primary site.

Abbreviation: AT, ataxia-telangiectasia. The expected numberof cancer cases was calculatedby multiplyingthe A-T relatives'person-yearsat risk of developingcancerby the age,race,sex,calendaryear,and cause-specificcancerincidenceratesforConnecticut for the years 1935 through 1979. Two cancersdiagnosedbeforethe patients were aged 20 years are included: one rhabdomyosarcoma and one acute lymphocytic leukemia, Double primary cancers occurred in seven blood relatives, including three femaleswith bilateral breast cancer.

incidence rates from Connecticut. Rate ratios for these specific cancer types could not be calculated because only three cancers occurred in spouse controls. Four of the 44 families learned of our A-T research program from television and newspaper reports that followed publication of our 1987 A-T study [4], and were "self referred." They contacted us directly to obtain more information about A-T and to volunteer participation in our research. Ten cancers had occurred in these four families, nine in blood relatives (nine in 50 or 18% of the total) and one in a spouse control (one in three or 33% of the total).

DISCUSSION Earlier reports have documented excess cancer in the close blood relatives of A-T homozygotes [3-5]. The elevated cancer rate observed in the blood relatives in this study of 44 previously undescribed families provides further support for the hypothesis that heterozygotes for the A-T gene are predisposed to cancer. The study also provides new information about which specific cancer types might be associated with heterozygosity for the A-T gene. As in the largest study published of A-T families [4], breast cancer was the most frequently occurring cancer in this group of A-T families. This high incidence of breast cancer, together with the occurrence of bilateral breast cancer in an obligate A-T heterozygote, the mother of a homozygote, supports the hypothesis that the A-T gene predisposes to breast cancer. It is also consistent with data from Britain,

122

N. Sadamori et al. where excess breast cancer mortality among A-T mothers was observed in a series of 61 A-T families [5]. The lung cancers in two obligate A-T heterozygotes and the increased number of observed cases over the expected number provides new evidence that lung cancer may be one of the specific cancers associated with A-T heterozygosity. For several additional cancersmprostate, stomach, pancreatic, melanoma, and lymphoma and lymphocytic leukemia--more cancers were observed than were expected from the Connecticut incidence rates. Because each of these cancers occurred in excess in one or more earlier A-T family studies [3-5], they may also be associated with A-T hetarozygosity. In contrast, the observed deficit of cancers of the colon and rectum is consistent with an earlier finding [4] and suggests that these particular cancers are not positively associated with A-T heterozygosity. This report adds to the considerable body of evidence that A-T heterozygotes are predisposed to cancer. The important question now is: Which specific cancers are involved? It is clear that using the present family study method, the number of cancers in spouse controls will never be sufficient to answer this question definitively. The A-T gene has recently been localized to chromosome 11q22-23, however [12]. The application of a newly developed definitive method to test which specific cancers are associated with the A-T gene in heterozygotes, which relies on newly available molecular tests to identify the A-T heterozygote [13, 14], may conclusively answer this important question. This study was supported by NIH Grants No. CA14235 and HD03110. We thank Debra Sadler and Ruby Massey for locating the A-T families and collecting the data.

REFERENCES 1. Spector BD, Filipovich AH, Perry GS, Ill, KerseyJH (1982):Epidemiolosy of cancer in ataxia. telansiectasia. In: Ataxia-telangiectasia: A Cellular and Molecular Link Between Cancer, Neuropatholosy, and Immune Deficiency,BA Bridses, DG Hamden, eds. John Wiley &Sons, Chichestar, England, pp. 103-138. 2. MorrellD, Cromartie E, Swift M (1986): Mortality and cancer incidence in 263 patients with ataxia-telangiectasia. J Natl Cancer lnst 77:89-92. 3. Swift M, Sholman L, Perry M, Chase C (1976): Malignant neoplasms in the families of patients with ataxia-telangiectasia. Cancer Res 36:209-215. 4. SwiftM, Reit~duer PJ, Morrell D, Chase CL (1987): Breast and other cancers in families with ataxia-telant~iectasia. N Engl J Med 316:1289-1294. 5. Pippard EC, Hall AJ, BarkerDJ, Bridges BA (1988):Cancer in homozygotesand heterozysotes of ataxia-telangiectasia and xeroderma pigmentosum in Britain. Cancer Re8 48:2929-2932. 6. Swift M, Morrell D, Cromartie E, Chambarlin AR, Skolnick MH, Bishop DT (1986): The incidence and gene frequency of ataxia-telangiectasiain the United States. Am J Hum Genet 39:573-583. 7. Kleinbaum DG, Kupper LL, Morsenstern H (1982): Epidemiologic Research: Principles and Quantitative Methods. Lifetime Learning Publications, Belmont, CA. 8. Rothman KJ, Boice JD, Jr (1979): Epidemiologic analysis with a prosrammable calculator. National Institutes of Health DHEW publication No. 79-1649. 9. Swift M, Cohen J, Pinkham R (1974): A maximum-likelihood method for estimating the disease predisposition of hetero~ygotes. Am J Hum Genet 26:304-317. 10. Monson RR (1974):Analysis of relative survival and proportional mortality. Comput Biomed Res 7:325-332. 11. Mantel N, Haenszel W (1959): Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer lnst 22:719-748. 12. Gatti RA, Berkel I, Boder E, Braedt G, Charmley P, Concannon P, Ersoy F, Foroud T, Jaspers

Appearance Time of t(8;21)

123

NGJ, Lense K, Lathrop GM, Leppert M, Nakamura Y, O'Connell P, Paterson M, Salser W, Sanal O, Silver J, Spsrkes RS, Susi E, Weeks DE, Wei S, White R, Yoder F (1988):Localization of an ataxia-telan~tectasia 8ene to chromosome 11q22-23. Nature 336:577-580. 13. Swift M, Chase CL, Morrell D (1990):Cancer predisposition of ataxia-telansiectasia heterozyMotes. Cancer Genet Cytopnet 46:21-27. 14. Swift M, Kupper LL, Chase CL (1990): Effective tastinMof Mane-disease associations. Am J Hum Genet (in press).

Cancers in 44 families with ataxia-telangiectasia.

Cancer incidence was measured retrospectively in 574 close blood relatives of white ataxia-telangiectasia (A-T) patients and 213 spouse controls in 44...
395KB Sizes 0 Downloads 0 Views