Low Incidence of Familial Breast Cancer among Hispanic Women

Low Incidence of Familial Breast Cancer among Hispanic WomenAuthor(s): Melissa L. Bondy, Margaret R. Spitz, Susan Halabi, John J. Fueger and Victor G.VogelSource: Cancer Causes & Control, Vol. 3, No. 4 (Jul., 1992), pp. 377-382Published by: SpringerStable URL: http://www.jstor.org/stable/3553158 .

Accessed: 11/06/2014 02:31

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Cancer Causes &Control.

http://www.jstor.org

This content downloaded from 195.78.108.69 on Wed, 11 Jun 2014 02:31:25 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=springer

http://www.jstor.org/stable/3553158?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Cancer Causes and Control, 3, 377- 382

Low incidence of familial breast

cancer among Hispanic women

Melissa L. Bondy, Margaret R. Spitz, Susan Halabi, John J. Fueger, and Victor G. Vogel

(Received 16 March 1992; accepted in revised form 16 April 1992)

There is a paucity of data on familial patterns of breast cancer among minority populations. This study compared the frequency of cancer in 1,095 first-degree relatives of 50 White, 46 Black, and 49 Hispanic breast- cancer patients referred to The University of Texas M. D. Anderson Cancer Center (United States). Family histories of cancer were derived from a self-administered questionnaire on risk factors. Expected numbers of cancers were calculated from the Connecticut Tumor Registry for White and Black relatives and from the New Mexico Tumor Registry for Hispanic relatives. Family history of a first-degree relative with breast cancer was the most important risk factor for both Black and White patients. Significantly elevated standardized incidence ratios (SIR) for breast cancer were noted among White (SIR = 4.5, 95 percent confidence interval [CI] = 1.2-11.4) and Black (SIR = 4.1, CI = 1.1-10.4) relatives younger than age 45. Despite the small number of Black patients, the combined effect of family history of breast cancer and the relative's age at diagnosis (under 45 years) was associated with an SIR of 7.1 (CI = 1.9-18.1). A deficit of cancer was noted in Hispanic women; only one patient reported having a first-degree relative with breast cancer. These findings, although based on small numbers, suggest that Hispanics have a lower rate of familial breast cancer than Whites and Blacks, and that they may possess protective factors that reduce their risk for breast cancer.

Introduction Familial history of breast cancer has been studied extensively in Whites,"5 but few studies of Blacks 7 and only one study of Hispanics8 have been conducted, despite increasing cancer mortality in both Blacks and

Hispanics. Hispanic women are known to have large families and first births at an early age, and these factors have been presumed to account for their lower breast- cancer risk.' It is important to evaluate familial aggre- gation of breast cancer and other cancers among the

first-degree relatives of White, Black and Hispanic breast-cancer patients to determine if genetic differences may play a role in differential incidence rates across ethnic groups. More information is needed to

determine which risk factor or factors may increase familial risk. Therefore, we also sought to evaluate the relationship of familial cancer risk with breast-cancer risk factors such as age at first birth and age at diagnosis.

Materials and methods

Study population Our subjects were patients with histologically con- firmed breast cancer who registered at The University of Texas M. D. Anderson Cancer Center (Houston,

Dr Bondy, Ms Halabi, and Dr Vogel are with the Department of Medical Oncology and Dr Spitz and Mr Fueger are from the Department of Cancer Prevention and Control of The University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA. Address correspondence to Dr Bondy at the Department of Medical Oncology, Box 501, 1515 Holcombe Blvd., Houston, Texas 77030, USA. This paper was presented at the 1991 meeting of the American Association for Cancer Research in Houston, Texas.

? 1992 Rapid Communications of Oxford Ltd Cancer Causes and Control. Vol 3. 1992 377



M. L. Bondy et al

TX, USA) from 1 October 1985 through 31 December 1989. The subjects were obtained from listings of patients who completed a self-administered, comprehensive, cancer risk-factor questionnaire distributed to all newly registered, adult patients at M. D. Anderson Cancer Center. This instrument is the central compo- nent of our Patient Risk Evaluation Program. It has been evaluated for validity and reliability and is described in detail elsewhere.9,1o The questionnaire eli- cits a detailed family history of cancer, including com-

plete data on all of the patient's first-degree relatives

(parents, siblings, and children) and their medical histories, dates of birth, death, and cancer diagnosis; and

age at diagnosis. Known breast-cancer risk factors such as age at menarche and age at first livebirth also are included in the questionnaire.

Forty-six Black and 49 Hispanic women with breast cancer were identified from the risk-factor database, and 50 White women were selected at random from 585 White breast-cancer patients in the database. Because of the small number of Black (7.3 percent) and His-

panic (10.1 percent) patients referred to M. D. Ander- son Cancer Center, we included all eligible Blacks and

Hispanics. We constructed pedigrees from self-

reported cancer histories about each first-degree relative of the breast cancer patients using the following method. For those family members for whom dates of birth were unknown, the relative's birth year was estimated from the patient's year of birth. The years of birth for each proband's father and mother were

approximated by subtracting 25 and 23, respectively, from the proband's year of birth, on the assumption that this was the average childbearing age for the

parents." If the dates of birth for any of the patient's siblings or children were unknown, we estimated the birthdate from age tables compiled by Sparatley et al." When we estimated the relatives' ages, we took into consideration changes in the intervals between success- ive births according to social and demographic factors such as mother's age and race. The number of relatives at risk was incomplete for one (two percent) Hispanic family and six (12 percent) Black families. These relatives were included in the analysis and considered to be disease-free.

Statistical analysis The analysis included only invasive cancers and excluded nonmelanoma skin cancer and in situ carci- noma. Reported cancers were classified and coded according to the International Classification of Dis- eases for Oncology."2 The expected numbers of cancers for White and Black relatives were derived from age-, sex-, race-, and calendar-year-specific rates from the Connecticut Tumor Registry. Comparable expected

rates for Hispanic relatives were derived from data from the New Mexico Tumor Registry from 1978 through 1982, the earliest period for which such rates are available. For the earlier years, we extrapolated the Hispanic rates by applying the ratio of Connecticut Whites to New Mexico Hispanics by age group. Standardized incidence ratios (SIR) were calculated from the ratio of observed to expected numbers of cancers. Person-years at risk were determined from the date of birth to the date of cancer diagnosis, death, completed interview, or age 75, whichever came first. Risk was truncated at age 75 because cancer rates may be unreliable beyond that age." The calculation of SIRs, 95 percent confidence intervals (CI), and tests of significance were performed using a computer program developed by Monson."4

The data were stratified by known breast-cancer risk factors such as (i) age at menarche, (ii) age at first live birth, and (iii) parity, and family history of breast cancer, to assess whether the proband's risk factors pre- dicted increased risk in the proband's relatives. Tests of homogeneity based on a Poisson distribution" were computed to evaluate variability among patients by known breast-cancer risk factors and to determine whether SIRs differed among the different ethnic groups. Nonparametric correlation tests were used, and their significance levels were computed to determine whether there was any relationship between the variables.16

Results

Analysis of the demographic characteristics of the women with breast cancer showed that a larger proportion of Black (37 percent) and Hispanic (36.7 percent) women than White (24 percent) women were

diagnosed when they were younger than 45 years of age (Table 1), but these differences were not statisti- cally significant. Hispanics exhibited the lowest per- centage (10.2 percent) of nulliparous women, com-

pared with 23.9 percent of Black women and 18 percent of the Whites. Hispanics had the highest average number of children (3.6), and a higher proportion of Hispanics had their first livebirth between ages 20 and 24 (42.9 percent) than Whites (28 percent) and Blacks (28.3 percent). Ten percent of White women (n = 5) and 15.7 percent of Black women (n = 7) reported having a family history of breast cancer, compared with only two percent of Hispanic women (n = 1) reporting a first-degree relative with breast cancer. None of the chi-square values were significant for any of the demographic or reproductive characteristic comparisons.

The cancer incidence among relatives of all races was lower than expected; however, the SIR for Hispanics

378 Cancer Causes and Control. Vol 3. 1992



Breast cancer among Hispanic women

Table 1. Characteristics of breast cancer subjects

White Black Hispanic No. (%) No. (%) No. (%)

Age at diagnosis (years) <45 12 (24.0) 14 (30.4) 18 (36.7) 45-54 15 (30.0) 10 (21.8) 17 (34.7) 55 + 23 (46.0) 22 (47.8) 14 (28.6)

Age at first livebirth (years) <20 12 (24.0) 14 (30.4) 12 (24.5) 20-24 14 (28.0) 13 (28.3) 21 (42.9) 25-29 11 (22.0) 5 (10.9) 6 (12.2) 30 + 4 (8.0) 3 (6.5) 5 (10.2) Nulliparous 9 (18.0) 11 (23.9) 5 (10.2)

Family history of breast cancer Yes 5 (10.0) 7 (15.7) 1 (2.0) No 45 (90.0) 39 (84.8) 48 (98.0)

Average number of children 2.0 2.8 3.6

Average number of years of menstruation 31.0 28.7 29.5

Table 2. Observed and expected cancers among first-degree relatives of breast cancer patients by relative's age

Ethnic group and Relatives Observed (0) Expected (E) SIR (O/E) CIs relative's age at risk

White All sites 22 32.3 0.7 0.4-1.0

<45 326 8 4.9 1.6 0.7-3.2 45 + 182 14 27.4 0.5 0.3-0.9

Breast <45 146 4 0.9 4.5 1.2-11.4 45 + 85 2 3.6 0.6 0.1-2.0

Black All sites 22 27.7 0.8 0.5-1.2

<45 327 7 4.8 1.5 0.6-3.0 45 + 158 15 22.9 0.7 0.4-1.1

Breast <45 173 4 1.0 4.1 1.1-10.4 45 + 82 5 3.2 1.6 0.5-3.7

Hispanic All sites 10 28.5 0.4 0.2-0.6

<45 442 2 5.7 0.3 0.0-1.3 45 + 182 8 23.0 0.3 0.1-0.7

Breast <45 225 1 1.1 0.9 0.0-5.3 45 + 90 0 3.1 0.0 0.0-1.2

a CI = 95% confidence interval.

Cancer Causes and Control. Vol 3. 1992 379



M. L. Bondy et al

Table 3. Observed and expected cancers among first-degree relatives of breast cancer patients by family history of breast cancer and race

Number of Observed (0) Expected (E) SIR (O/E) CI' relatives at risk

White Family history (n = 5)

All sites 37 9 2.2 4.0 1.8-7.7 Breast 19 6 0.5 12.1 4.4-26.3

No family history (n = 45) 289 13 29.9 0.4 0.2-0.7

Black Family history (n = 7)

All sites 52 12 4.6 2.6 1.4-4.6 Breast 30 9 0.7 12.2 5.6-23.1

No family history (n = 40) 275 10 23.2 0.4 0.2-0.8

a CI = 95 percent confidence interval. Xldf = 25.6, P <0.001 (comparing Whites with and without family history of breast cancer). X2di = 28.7, P < 0.001 (comparing Blacks with and without family history of breast cancer).

jdf = 44.7, P < 0.001 (comparing family history of Whites with family history of Blacks).

was significantly decreased (Table 2). Among the 326 White relatives at risk, 22 cancers were observed and 32.3 were expected (SIR = 0.7). There were 22 cancers observed among 327 relatives of Black patients (SIR = 0.8). Although Hispanic patients had 32 percent more relatives at risk (n = 442), they reported only 10 cancers among their first-degree relatives and the overall SIR was 0.4 (CI = 0.2-0.6). The SIRs for breast cancer were not significant among the relatives of White (SIR = 1.3) and Black (SIR = 2.2) patients (data not shown).

Because familial cancers often occur at younger ages, we stratified the data by the relatives' ages to determine if cancers occurred more frequently in younger relatives (under 45 years). This age was selected based on the age criteria for the young-onset cancers reported in the Li-Fraumeni cancer-family-syndrome'7 and studies by Ottman et al,2 Claus et al,5 and Hall et al.' The SIR for all cancer sites was elevated in the younger relatives of Whites and Blacks; however, a deficit of cancer was observed in older White and Black relatives

(SIR = 0.5 and 0.7, respectively) (Table 2). Despite the small number of patients in our study, breast cancer was the only cancer site for which an elevated SIR was observed. A significant excess of breast cancer was observed in both White and Black relatives under age 45 but not in relatives over age 45 (Table 2). Only one case of breast cancer was observed among the relatives of the Hispanic breast cancer patients (Table 2). For Blacks and Whites, we evaluated the risk of breast cancer in the relatives by relationship to the patient and found excesses among the children of White patients

(SIR = 2/0.2 = 9.5, CI = 1.1-34.1) and the siblings of Black patients (SIR = 6/1.6 = 3.7, CI = 1.4-8.1) (data not shown).

To determine which breast-cancer risk factors con- tribute to familial cancer risk, we stratified the data by known breast-cancer risk factors such as diagnosis, age at first livebirth, and family history of breast cancer. Family history of breast cancer was the only risk factor

contributing to excess cancer in the patient's first-

degree relatives. We compared the observed and

expected numbers of cancers in the relatives by these risk factors. There were five white patients (10 percent) who had six first-degree relatives with breast cancer (SIR = 12.1, CI = 4.4-26.3) (Table 3). A significant deficit of cancer was observed in the remaining 45 White patients without a first-degree relative with breast cancer (SIR = 0.4, CI = 0.2-0.7). Seven (14.9 percent) Black patients had nine first-degree relatives with breast cancer (Table 3). The same deficit of cancer observed in Whites was seen also among relatives of 40 Black patients without a family history of breast cancer (SIR = 0.4, CI = 0.2-0.8).

Tests of homogeneity were computed to determine whether the SIRs were significantly different from each other. Significant X2 differences were observed between Whites with and without a family history, Blacks with and without a family history, and Whites and Blacks with a family history. For Blacks, the combined effect of family history of breast cancer and the relative's age at diagnosis younger than 45 years was associated with a significant risk of cancer (SIR 4/0.6 = 7.1, CI = 1.9-18.1) (data not shown).




Breast cancer among Hispanic women

Discussion

This study of ethnicity and familial breast-cancer risk showed similar cancer patterns among relatives of White and Black breast-cancer patients, contrasted with a significant deficit of cancer among relatives of Hispanic patients. Breast cancer was the only type of cancer with an excess of familial cancer risk, especially among younger relatives (under 45 years). It is accepted generally that early onset of cancer is associated with familial cancer risk, whereby a person inherits a genetic predisposition (first hit) that normally would result in the development of cancer in early-to-middle adult life.1920 If an individual is genetically predisposed to cancer, several genetic alterations are required before the cancer develops.

The excess cancers in our study occurred in White and Black families of patients with at least one first- degree relative with breast cancer, and a deficit of cancer was observed in the remaining families. The significant deficit of cancer in older White and Black relatives cannot be easily explained; however, a similar obser- vation was noted in older relatives in a familial study of childhood cancer patients.2 Possibly those relatives who are genetically predisposed to cancer contracted it at a younger age, and the relatives who were unaffected have a lower cancer risk, living to older ages without being affected.

There have been two previous epidemiologic studies addressing familial breast cancer in Blacks; the first was a case report,6 and the second7 a genetic segregation analysis. The authors of the latter found familial aggre- gation consistent with Mendelian recessive trans- mission of a single major gene. Amos et al7 found consistent evidence that family history of breast cancer was predictive of an individual's breast-cancer risk,'-' and the relative risks were the same order of magnitude for Blacks as for Whites.25 We were unable to test for genetic models or infer genetic mechanisms because our sample size was too small, and second-degree relatives were not specified.

There has been one study that reported a higher familial risk of breast cancer in Hispanics than in other ethnic groups,8 but our study did not confirm this conclusion. Buchanan et al8 reported that 10 percent of Hispanic breast-cancer patients had a first-degree relative with breast cancer, the rate we observed in the White and Black families. Since we did not observe a familial risk of breast cancer in Hispanics, one could postulate a response bias. Although the response rate to the questionnaire only averaged 50-60 percent, we have compared demographic data over a three-month period of breast-cancer cases who responded to the self-administered questionnaire with those who did

not respond. There were no significant differences with regard to age, ethnicity, and histology between respon- dents and nonrespondents.

In this study, we could not determine if the edu- cation level of the patient reflected the accuracy of reporting breast cancer in her relatives because we did not validate the reports. However, 38 percent of the Blacks in our study did not complete high school and 60 percent had a family income under $13,000. Half of the Hispanics did not complete high school and half had an annual family income below $13,000.

The behavioral and sociologic literature is replete with studies showing that Hispanics' social-support network is closely linked to their relatives. Marital status is an indicator of social support and, in our population, 70 percent of the Hispanic women were married compared with 47 percent of the Black women. Several studies2224 have reported that extended family ties are stronger in Hispanic families. Keefe2s found that ethnicity was related to type of social contact, with Anglos more likely to interact with friends and Hispanics more likely to interact with their relatives. In addition, Hispanics are more likely than Anglos to call on nonspouse relatives if they need a caretaker.26 Therefore, it is likely that the Hispanic women in our study would be knowledgeable about the occurrence of cancer in their first-degree relatives. The patients were asked to complete a comprehensive family history including dates of birth and death, and questions about the occurrence of cancer for each relative. One would expect that if they completed the detailed information about each relative they also would have completed the question about whether that relative had cancer. Another possible reason why our results did not concur with Buchanan et al8 was because our risk-factor questionnaire was not trans- lated into Spanish. Thus, nonbilingual Hispanics are underrepresented. However, it is most improbable that Hispanics who are not bilingual would have differing familial patterns of breast cancer.

We did not contact the patients or their relatives to validate their reports of cancer; however, both we27 and Love et al28 have noted that patients accurately report cancer in their first-degree relatives. In addition, both studies reported that breast cancer was one of the most accurately reported cancers. Therefore, we do not believe we introduced a bias by not validating the cancer reports.

Since multiple comparisons were made, there was an increased likelihood of observing significant results. Some of the observed associations with high cancer risk may have been due to chance alone, and the results presented should be interpreted with caution, especially because of the small numbers of relatives

Cancer Causes and Control. Vol 3. 1992 381



M. L. Bondy et al

included in this study. Nevertheless, there are strong indications from other studies'-7 that family history of cancer contributes to increased cancer risk in Whites and Blacks, and only one study8 has reported this in

Hispanics. The negative association of familial risk in

Hispanics forms the basis for hypotheses that should be tested using other data sets, or in specifically designed studies in the future.

Hispanics have lower rates of breast cancer29,30 than Whites and Blacks and could possess protective factors

against breast cancer. Buchanan et al" evaluated reproductive risk factors such as age at first livebirth and parity, and found an association with age at first livebirth, but not parity. We did not find an association with either factor, but the Hispanic women in our study had a higher average number of children, a lower proportion were nulliparous, and a higher proportion had first livebirths at younger ages (20-24 years) than the White and Black patients. Higher parity and younger age at first livebirth suggest that Hispanics may have

hormonally related protective factors against familial breast cancer. A larger study of Hispanic women is

necessary to address the relationship of breast-cancer risk factors. If protective factors are found in Hispanic women, this may lead to understanding other risk factors in both White and Black populations.

Acknowledgements-We thank Ms Alcena Doxley for her secretarial support.

References

1. Anderson DE, Badzioch MD. Risk of familial breast cancer. Cancer 1984; 56: 383-7.

2. Ottman R, Pike MC, King MC, et al. Familial breast cancer in a population-based series. Am J Epidemiol 1986; 123:15-21.

3. Schwartz AG, King MC, Belle SH, et al. Risk of breast cancer to relatives of young breast cancer patients.JNCI 1985; 75: 665-8.

4. Sattin RW, Rubin GL, Webster LA, et al. Family history and the risk of breast cancer.JAMA 1985; 253:1908-13.

5. Claus EB, Risch NJ, Thompson WD. Age of onset as an indicator of familial risk of breast cancer. AmJ Epidemiol 1990; 131: 961-72.

6. Siraganian PA, Levine PH, Madigan P, Mulvihill JJ. Familial breast cancer in black Americans. Cancer 1987; 60: 1657-60.

7. Amos CI, Goldstein AM, Harris EL. Familiality of breast cancer and socioeconomic status in blacks. Cancer Res 1991; 51: 1793-7.

8. Buchanan AV, Weiss KM, Anderson DE, et al. Epidemi- ology of breast cancer in a Mexican-American population. JNCI 1985; 74: 1199-206.

9. Spitz MR, Fueger JJ, Newell GR. The development of a

comprehensive, institution-based patient risk evaluation program: I. Development, content and data manage- ment. Am J Prey Med 1988; 4:183-7.

10. Spitz MR, Fueger JJ, Newell GR. The development of a comprehensive, institution-based patient risk evaluation program: II. Validity and reliability of questionnaire data. Am JPrev Med 1988; 4: 188-93.

11. Sparatley E, Taffel S. Interval between births: United States, 1970-1977. Hyattsville, MD: National Center for Health Statistics 1981; Series 21, No. 39.

12. World Health Organization. International Classification for Diseases-Oncology. Geneva: WHO, 1976.

13. Cutler SJ, Young JL Jr (eds). Third national cancer sur- vey: incidence data. NCI Monogr 1975; 41: 1-454.

14. Monson RR. Analysis of relative survival and pro- portional mortality. Comput Biomed Res 1974; 7: 325-32.

15. Breslow NE, LubinJH, Marek P, Langholtz B. Multipli- cative models and cohort analysis.] Am Stat Assoc 1983; 78: 1-12.

16. Snedecor GW, Cochran WG. Statistical Methods. Ames, IA: The Iowa State University Press, 1980.

17. Li FP, Fraumeni JF Jr, Mulvihill JJ, et al. A cancer family syndrome in twenty-four kindreds. Cancer Res 1988; 48: 5358-62.

18. Hall JM, Lee MK, Newman B, et al. Linkage to early- onset familial breast cancer to chromosome 17q21. Sci- ence 1990; 250: 1684-8.

19. Knudson AG Jr. Retinoblastoma: a prototypic heredi- tary neoplasm. Semin Oncol 1978; 5: 57-60.

20. Vogelstein B. A deadly inheritance. Nature 1990; 348: 681-2.

21. Strong LC, Stine M, Norsted TL. Cancer in survivors of childhood soft tissue sarcoma and their relatives. JNCI 1987; 79: 1213-20.

22. Alvirez D, Bean FD, Williams D. The Mexican American family. In: Mindel CH, Habenstein RW, eds. Ethnic Families in America: Patterns and Variations 2nd edn. New York: Elsevier, 1981: 269-92.

23. Farris BE, Glenn ND. Fatalism and familism among Anglos and Mexican Americans in San Antonio. Sociol Soc Res 1976; 60: 393-402.

24. Mindel CH. Extended familism among urban Mexican Americans, Anglos and Blacks. Hispanic J Behav Sci 1980; 2: 21-34.

25. Keefe SE. Personal communities in the city: Support net- works among Mexican Americans and Anglo Americans. Urban Anthropol 1980; 9: 51-74.

26. Vernon SW, Roberts RE. A comparison of Anglos and Mexican Americans on selected measures of social support. HispanicJ Behav Sci 1985; 7: 381-99.

27. Bondy ML, Strom SS, Colopy MW, Brown BW, Strong LC. Comparison of accuracy information in a family study of childhood soft tissue sarcoma (Abstract). Am] Epidemiol 1991; 134: 766.

28. Love RR, Evans AM, Josten DM. The accuracy of patient reports of a family history of cancer. J Chronic Dis 1985; 38: 289-93.

29. McDonald EJ, Heinze EB. Epidemiology of Cancer in Texas: Incidence Analyzed by Type, Ethnic Group, and Geographic Location. New York: Raven Press, 1978.

30. Menck HR, Henderson BE, Pike MC, et al. Cancer incidence in the Mexican American. JNCI 1975; 55: 531-6.




Documents

Low Incidence of Familial Breast Cancer among Hispanic Women