Risk of cancer at sites other than the breast in Swedish families eligible for BRCA1 or BRCA2 mutation testing

J. Lorenzo Bermejo1,* and K. Hemminki1,2

1 Division of Molecular Genetic Epidemiology, German Cancer Research Centre (DKFZ), Heidelberg, Germany; 2 Department of Biosciences at Novum, Karolinska Institute, Huddinge, Sweden

* Correspondence to: Dr J. Lorenzo Bermejo, Division of Molecular Genetic Epidemiology, German Cancer Research Centre (DKFZ), Im Neuenheimer Feld 580, D-69120, Heidelberg, Germany. Tel: +49-6221-421805; Fax: +49-6221-421810; Email: j.lorenzo{at}dkfz.de


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Background: Population-based data on the risk of cancer in families eligible for BRCA1/2 mutation testing may help to reach a consensus about the association of BRCA1/2 mutations with cancer at sites other than the breast and may reveal new, non-BRCA1/2 related components of the familial clustering of cancer in those families.

Patients and methods: The families of the Swedish Family-Cancer Database with at least three generations (n = 944 723) were classified according to the criteria proposed by the German Consortium for Hereditary Breast and Ovarian Cancer. The cancer incidences in the classified families were compared with the incidences in the general population. The percentages of individuals with cancer in families eligible for BRCA1/2 mutation testing were compared with data in the literature to estimate the proportion of malignancies related to BRCA1/2 mutations.

Results: Families with two breast cancers before the age of 50 years showed increased risk of early onset pancreatic, prostate and ovarian cancers; families with ovarian and breast cancers presented increased incidences for ovarian and ocular cancers; families with two breast cancers, at least one of them under the age of 50 years, showed increased risks of prostate and primary liver cancers. Stomach cancer before age 70 years was twice as frequent in families with breast and ovarian cancers as in the general population. BRCA1/2 mutations probably explain most of the aggregation of ovarian cancer in families with male breast cancer, and in families with at least two breast cancers diagnosed before age 50 years.

Conclusions: The association of BRCA1/2 mutations with ovarian, pancreatic, prostate and stomach cancers was confirmed at a population level. However, the clustering of early pancreatic cancer in families with two breast cancers under age 50 years, the aggregation of ovarian cancer in families with breast and ovarian cancers, and the increased incidence of early onset prostate cancer in families with male breast cancer seem to be due to other effects unrelated to BRCA1/2 mutations.

Key words: BRCA1, BRCA2, mutation testing, clinical criteria


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Germline mutations in BRCA1/2 substantially increase the lifetime risk of developing breast and ovarian cancer and have been associated with cancers at other sites. In BRCA1-mutation carriers, the risks for cancers of the pancreas and colon have been found to be significantly increased [1Go–4Go]. BRCA2 mutations have been related to cancers of the pancreas, prostate, gall bladder, bile duct and stomach, and to malignant melanomas [5Go–13Go]. However, data about the risk of cancer in BRCA1/2-mutation carriers for sites different from the breast or the ovary are often contradictory.

Recommendations about genetic testing of BRCA1/2 germline mutations are based on clinical criteria which summarise the personal and/or family histories of breast and ovarian cancers. The German Consortium for Hereditary Breast and Ovarian Cancer (GCHBO) recently classified 989 families according to specific criteria and found the highest frequency of BRCA1 mutations (43%) in families with breast and ovarian cancer and the highest frequency of BRCA2 mutations (23%) in families with male breast cancer [14Go]; similar clinical criteria have been proposed, e.g. by the National Comprehensive Cancer Network [15Go].

The objective of this study was to assess the risk of cancer in high-risk breast/ovarian cancer families at a population level. The families of the Swedish Family-Cancer Database were classified according to the criteria proposed by the GCHBO and the cancer incidences in those families, in which BRCA1/2 mutations were suspected, were compared with the incidences in the general population. Increased cancer risks for sites other than the breast or the ovary may be indicative of additional cancer types, which might warrant clinical follow-up in families that fulfil the clinical criteria. Population-based data on the risk of cancer in families eligible for BRCA1/2 mutation testing may help to reach a consensus about the association of BRCA1/2 mutations with cancer at sites other than the breast and the ovary. Furthermore, increased incidences of cancer at sites that have not been associated with BRCA1/2 mutations may reflect new, non-BRCA1/2 related components of the familial aggregation of cancer in those families. The close-to-100% histological verification of cancers in the Swedish Family-Cancer Database [16Go], the size of the Database and its coverage, offered unique possibilities for this population-based analysis.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
The Swedish Family-Cancer Database was created in the mid-1990s by linking census information, death notifications and the administrative family register at Statistics Sweden to the Swedish Cancer Registry. The Database includes all persons who were born in Sweden after 1931 with their biological parents. The Database was updated at the end of 2002 to include more than 10.2 million individuals. The Swedish Cancer Registry relies on separate compulsory notifications of cases from clinicians who diagnosed a neoplasm and pathologists/cytologists. A 4-digit diagnostic code according to the 7th revision of the International Classification of Diseases (ICD-7) and subsequent ICD classifications are available. In this study, upper aerodigestive tract cancer included lip, mouth, tongue and pharynx; small intestine, colon and rectal cancers included only adenocarcinomas; lung cancer was defined by ICD-7 codes 162–163, and eye cancer included melanoma only. Generic hepatic cancers (ICD-7 code 155) were also analysed separately as primary liver (ICD-7 code 155.0), gall bladder (ICD-7 code 155.1) and extrahepatic bile ductal cancers (ICD-7 code 155.2). In addition to the ICD-7 code 175 (generic ovarian cancer), ICD-7 codes 175.0 (ovary, primary) and 175.1 (fallopian tube) were used. Vaginal (ICD-7 code 176.1) and vulval cancers (ICD-7 code 176.0) were also investigated.

Only families with at least three generations in the Database were considered in this study. If a founder parent of the family was missing or if he/she was married several times, the family was excluded. Families were classified according to the clinical criteria proposed by the GCHBO as described in Table 1. These criteria considered bilateral breast cancers as two independent cases, but the present study also analysed the families with two affected members under the age of 50 years and the families with bilateral breast cancer before the age of 50 years separately.


View this table:
[in this window]
[in a new window]
 
Table 1. Classification criteria used by the German Consortium for Hereditary Breast and Ovarian cancer (GCHBO), number of individuals of the Swedish Family-Cancer Database in families that fulfilled the criteria and BRCA1/2 mutation rates found in the GCHBO study1

 
The standardised incidence ratio (SIR) was used to estimate the relative risks of individuals in the classified families. For male sex-specific cancers, probands were males; for female sex-specific cancers, probands were females and for the rest of the cancer sites, probands were both males and females. Follow-up started at birth, immigration date or 1961, whichever came latest. Follow-up terminated at diagnosis of cancer, death, emigration date, 31 December 2000 or the age of diagnosis as specified in the study. The SIR was calculated as the ratio of observed to expected number of cases. The expected numbers were computed from age- (5-year bands), parity- (six groups from ‘any parturition’ to ‘more than five parturitions’), sex- (male or female), socioeconomic status- (six groups), age at first birth- (five groups, five year bands between ‘before age of 20’ and ‘after age of 35’) and residential area- (four groups) standardised incidence rates.

The risk of developing cancer before the age of 50 years in the absence of other causes of death was estimated by dividing the number of site-specific cancers diagnosed under the age of 50 years by the number of individuals who reached this age. Emigrants before the age of 50 years, persons younger than 50 years by 2000, patients diagnosed with cancer at sites different from the site under consideration before age 50 years and individuals deceased through other causes before this age were excluded from the denominator. Similar ratios were calculated for cancers diagnosed under the age of 70 years and were referred to as the ‘percentage of affected individuals’. Confidence intervals (95% CI) were calculated assuming a Poisson distribution [17Go]; 99% CIs of the SIRs were also presented. Significant associations were explored in further detail by analysing second cancers and other malignancies that occurred in the affected families.

In order to estimate the cumulative risk of ovarian cancer attributable to BRCA1/2 mutations in families eligible for mutation testing, the BRCA1/2 prevalences from the GCHBO study [16Go] were multiplied by the average risks of ovarian cancer associated with BRCA1/2 mutations reported by Antoniou et al. [18Go]. For example, 24% of the families with two cases of breast cancer under the age of 50 years presented with BRCA1 mutations and 13% of those families showed BRCA2 mutations in the GCHBO study. The average cumulative risks of ovarian cancer by the age of 70 years estimated by Antoniou et al., 39% in BRCA1 and 11% in BRCA2 mutation carriers, would result in (24 x 0.39) + (13 x 0.11) = 10.79% of the women affected with ovarian cancer by the age of 70 years in families with two cases of breast cancer before the age of 50 years. The proportions of pancreatic and prostate cancers related to BRCA1/2 mutations were calculated in the same way by using the cumulative cancer risks reported by the Breast Cancer Linkage Consortium [1Go, 10Go].


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
The number of individuals of the Swedish Family-Cancer Database in families that fulfilled clinical criteria for testing of BRCA1/2 mutations is presented in Table 1. The criterion ‘male breast cancer in the family’ was the most restrictive and permitted to classify 4743 individuals. In contrast, 36 099 individuals belonged to families with two cases of breast cancer after age 50 years. Among the families with two breast cancers before age 50 years, 5199 individuals belonged to families with two patients diagnosed before age 50 years, whereas 7428 individuals were members of families in which bilateral breast cancer was diagnosed under age 50 years.

SIRs of site-specific cancers for members of the classified families are presented in Table 2Table 2. Families with two cases of breast cancer before age 50 years showed significant SIRs (different from 1.00) for ovarian and prostate cancers, the SIR of pancreatic cancer before age 50 years was also increased in those families. Members of families with male breast cancer did not show significant SIRs for cancer at sites different from the breast. Families with at least one member affected by breast cancer and at least one member affected by ovarian cancer presented significant SIRs for ovarian and ocular cancers. The SIR for breast and ovarian cancer in a single woman was also increased in those families. In one of the four families with breast, ovarian and ocular cancers, one woman had leukaemia and another woman was diagnosed with leukaemia after ovarian cancer. The 23 families with breast, ovarian and stomach cancers included 23 men and one woman with stomach cancer (one family had two men with stomach cancer); three of those families included women affected by both breast and ovarian cancers. Families with two breast cancers, at least one of them diagnosed before age 50 years, showed significant SIRs for liver and prostate cancers. Among the 16 families with liver and two breast cancers, two families had two members with hepatic malignancies (gall bladder and ampulla of Vater tumours in one, and gall bladder and liver primary cancers in the other). Families with two cases of breast cancer after age 50 years did not show significant SIRs. Families with breast cancer diagnosed before age 35 years presented significant SIRs for ovarian and pancreatic cancers. Families with two women diagnosed with breast cancer under age 50 years showed significant SIRs for ovarian cancers, whereas families with bilateral breast cancer diagnosed before age 50 years presented increased SIRs for ovarian, pancreatic and prostate cancers.


View this table:
[in this window]
[in a new window]
 
Table 2. Site-specific risk of cancer at sites other than the breast in families eligible for BRCA1/2 mutation testing1

 

View this table:
[in this window]
[in a new window]
 
[Table 2 (Cont)].
 
The GCHBO study found BRCA1/2 mutations in less than half of the classified families. Therefore, the association of BRCA1/2 mutations with site-specific cancers would result in only moderate risks when families eligible for BRCA1/2 mutation testing are analysed. This reason prompted us to calculate the percentage of individuals affected by cancer at those sites that showed increased risks at the 90% confidence level. Results are presented in Table 3 for families with two cases of breast cancer under age 50 years and for families with breast and ovarian cancers, which showed the highest BRCA1-mutation rates in the GCHBO study, as well as for families with male breast cancer, which are eligible for BRCA2-mutation testing. In families with two breast cancers diagnosed before age 50 years, 0.16% of the individuals had pancreatic cancer by age 50 years, a significantly higher percentage than in the general population. The percentage of women affected with ovarian cancer was 1.07% in families with two cases of breast cancer under age 50 years and 15.3% of the women had ovarian cancer before age 50 years in families with breast and ovarian cancers. The percentage of men with prostate cancer by age 50 years was significantly increased in families with male breast cancer. The percentage of individuals affected with stomach cancer by age 70 years in families with breast and ovarian cancers was 1.88%, significantly higher than in the general population. The percentages of women with ovarian cancer by age 70 years were 6.16% in families with two cases of breast cancer before age 50 years, 68.9% in families with breast and ovarian cancer and 2.59% in families with male breast cancer. The cumulative risks of cancer based on published data on BRCA1/2 mutation frequency and the penetrance in mutation carriers are also shown in Table 3. Families with breast and ovarian cancers showed the highest cumulative risks attributable to BRCA1/2 mutations: the risks by age 70 years were 17.5% for ovarian, 1.86% for prostate and 0.74% for pancreatic cancers.


View this table:
[in this window]
[in a new window]
 
Table 3. Percentage of individuals affected by cancer at sites other than the breast in families eligible for BRCA1/2 mutation testing1

 

    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
A total of 944 723 Swedish families were classified in the present study according to the criteria proposed by the GCHBO for BRCA1/2 mutation testing, which are similar to the criteria used by oncologists in clinical practice to investigate family histories for hereditary breast and ovarian cancer [15Go]. The objective of the present study was to assess, at a population level, the aggregation of malignancies other than breast cancers in high-risk breast/ovarian cancer families. Most studies use interview data to classify families, whereas the present study relied entirely on registered data of complete coverage. The large number of individuals analysed resulted in accurate estimates even for cancer types that are relatively uncommon in families eligible for BRCA1/2 mutation testing. Another important feature of this study was the standardisation of the incidences for age, parity, age of first birth and sex, which aimed at calculating risk estimates for the classified families that were comparable. Only families with at least three generations were included in the study, but the maximum age of individuals in the second generation (68 years) was a limitation of the present study. Ninety-nine per cent confidence intervals were calculated to deal with the issue of multiple comparisons. The SIRs in Table 2 are underlined when the corresponding 99% CI did not include 1.00. In this discussion, significant findings at the 95% CI are examined for consistency with previous literature.

The GCHBO counted bilateral breast cancer as two independent cases, but some differences were found in the present study among families with two women affected by breast cancer before age 50 years and those families that included bilateral breast cancer under age 50 years. The risk of ovarian cancer was higher in families that included two women affected by breast cancer before age 50 years. In contrast, families with bilateral breast cancer diagnosed under age 50 years showed higher risks for pancreatic and prostate cancers. Therefore, the two criteria should be analysed separately in future studies.

The mutation prevalences from the GCHBO study and the cumulative risks estimated by Antoniou et al. [18Go] would result in 10.8% of the women with ovarian cancer by age 70 years in families with both breast and ovarian cancers. In the present study, 68.9% of the women had ovarian cancer by age 70 years in those families. These data suggest that, in addition to BRCA1/2 mutations, other effects are probably involved in the aggregation of ovarian cancer in families with both breast and ovarian cancers. Similar reasoning was applied to estimate the proportion of ovarian malignancies that can be attributed to BRCA1/2 mutations in other families eligible for mutation testing. As shown in Table 3, 6.16% of the women had ovarian cancer by age 70 years in families with two cases of breast cancer before age 50 years, and 2.59% of the women had ovarian cancer by age 70 years in families with male breast cancer. The cumulative risks of ovarian cancer attributable to BRCA1/2 mutations suggest that most of the aggregation of ovarian cancer in families with two breast cancers before age 50 years, and in families with male breast cancer, is probably due to BRCA1/2 mutations.

Stomach cancer before age 70 years was twice as frequent in families with breast and ovarian cancer than in the general population. Both BRCA1 [10Go] and BRCA2 mutations [5Go, 8Go] have been reported to increase the risk of stomach cancer, but the number of tumours observed in previous studies has been small and the excess risk was attributed to some extent to misclassification of ovarian cancer. In the present study, almost all individuals diagnosed with stomach cancer in families with breast and ovarian cancers were males. Moreover, concomitant breast/ovarian cancer in the same woman was frequent in those families and one family included two males with stomach cancer. Since the risk of stomach cancer was not increased in families with male breast cancer, and it was only slightly increased in families with two cases of breast cancer before age 50 years, some association of BRCA1 mutations with stomach cancer is probable.

Earlier studies have found some association of BRCA1 mutations with colon cancer, but the excess risk of colon cancer has been attributed to misdiagnosed ovarian or rectal cancers [1Go]. Although cervical cancer has been previously associated with both BRCA1 and BRCA2 mutations [10Go, 13Go], the strength of this association remains unclear. In the present study, the families that fulfilled the clinical criteria for BRCA1/2 mutation testing did not show increased risks for colon or cervical cancers. The risk of cancer of the peritoneum was high in families with breast cancer before age 35 years, but only two cases were observed and the corresponding SIR was not statistically significant.

The risk of liver cancer has been found to be increased among BRCA1 mutations carriers in earlier studies, but the excess risk has been attributed to misreported metastases from other sites [1Go]. In the present study, the SIR for primary liver cancer was increased in families with two breast cancers, at least one of them under age 50 years, but the percentage of individuals affected by liver cancer in families with breast and ovarian cancers was only slightly increased, thus limiting the strength of the association of BRCA1 mutations and hepatic malignancies. BRCA2 mutations have been associated with hepatocelullar carcinomas, gall bladder and bile duct cancers [8Go]. In the present study, the risk of liver cancer was not increased in families with male breast cancer but gall bladder cancer was the only hepatic malignancy by age 70 years in those families, thus suggesting some histological specificity of the hepatic carcinomas found in families with male breast cancer.

The percentage of patients with pancreatic cancer by age 50 years was higher in families with two cases of breast cancer under age 50 years than in the general population. The cumulative risks of pancreatic cancer by age 50 years have been estimated to be 0.12% for BRCA1 mutation carriers [1Go] and 0.2% for BRCA2 mutation carriers [8Go]. These average estimates, together with the mutation frequencies found by the GCHBO in families with two breast cancers under age 50 years [14Go], would result in a cumulative risk of pancreatic cancer by age 50 years of 0.05% (Table 3). The percentage of individuals with pancreatic cancer by age 50 years found in the present study among those families was 0.16%. We did not find any pancreatic cancers before age 50 years in families with breast and ovarian cancers but the risk of pancreatic cancer was significantly increased in families with breast cancer before age 35 years. These data suggest some association of early onset breast and pancreatic cancers through causes unrelated to BRCA1 mutations.

The percentage of men affected with prostate cancer by age 50 years was significantly increased in families with male breast cancer, probably due to the association of BRCA2 mutations and prostate cancer. However, the average cumulative risk of prostate cancer by age 50 years estimated in an earlier study [8Go] and the BRCA2 mutation prevalences found by the GCHBO, would result in a cumulative risk of 0.02% by age 50 years (Table 3). In the present study, 0.27% of the men had prostate cancer by age 50 years in families with male breast cancer. In contrast, the percentage of men affected by prostate cancer in families with suspected BRCA1 mutations was only slightly increased in the present study, which probably reflects the moderate risk of prostate cancer conferred by BRCA1 mutations [1Go]. The percentage of individuals affected by ocular melanomas was slightly increased in families with breast and ovarian cancers and in families with male breast cancer, which is in agreement with the association of BRCA2 mutations and ocular malignancies found in a previous study [19Go].

In conclusion, breast and ovarian cancers are the major worry in the clinical counselling of families that fulfill the clinical criteria for BRCA1/2 mutation testing. In families with bilateral breast cancer or two breast cancers before age 50 years, there is concern about early onset pancreatic cancers. Prostate cancers are also in excess in these families but the risk is only moderate. Most cases of ovarian cancer in families with male breast cancer, and in families with at least two breast cancers diagnosed before age 50 years, are probably attributable to BRCA1/2 mutations. Other, non-BRCA1/2 related effects are probably involved in the clustering of early onset pancreatic cancer in families with two breast cancers under age 50 years, in the aggregation of ovarian cancer in families with breast and ovarian cancers, and in the increased incidence of early prostate cancer in families with male breast cancer. Remarkably, in this large study no entirely novel associations were found at previously unreported sites.


    Acknowledgements
 
The Study was supported by Deutsche Krebshilfe and the Swedish Cancer Society. The Family-Cancer Database was created by linking registers maintained by Statistics Sweden and the Swedish Cancer Registry.

Received for publication April 19, 2004. Revision received July 14, 2004. Accepted for publication July 15, 2004.


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1. Thompson D, Easton DF. Cancer incidence in BRCA1 mutation carriers. J Natl Cancer Inst 2002; 94: 1358–1365.[Abstract/Free Full Text]

2. Arason A, Barkardottir RB, Egilsson V. Linkage analysis of chromosome 17q markers and breast-ovarian cancer in Icelandic families, and possible relationship to prostatic cancer. Am J Hum Genet 1993; 52: 711–717.[ISI][Medline]

3. Ford D, Easton DF, Bishop DT et al. Risks of cancer in BRCA1-mutation carriers. Breast Cancer Linkage Consortium. Lancet 1994; 343: 692–695.[ISI][Medline]

4. Brose MS, Rebbeck TR, Calzone KA et al. Cancer risk estimates for BRCA1 mutation carriers identified in a risk evaluation program. J Natl Cancer Inst 2002; 94: 1365–1372.[Abstract/Free Full Text]

5. Risch HA, McLaughlin JR, Cole DE et al. Prevalence and penetrance of germline BRCA1 and BRCA2 mutations in a population series of 649 women with ovarian cancer. Am J Hum Genet 2001; 68: 700–710.[CrossRef][ISI][Medline]

6. Goggins M, Schutte M, Lu J et al. Germline BRCA2 gene mutations in patients with apparently sporadic pancreatic carcinomas. Cancer Res 1996; 56: 5360–5364.[Abstract]

7. Thorlacius S, Olafsdottir G, Tryggvadottir L et al. A single BRCA2 mutation in male and female breast cancer families from Iceland with varied cancer phenotypes. Nat Genet 1996; 13: 117–119.[ISI][Medline]

8. Cancer risks in BRCA2 mutation carriers. The Breast Cancer Linkage Consortium. J Natl Cancer Inst 1999; 91: 1310–1316.[Abstract/Free Full Text]

9. Giusti RM, Rutter JL, Duray PH et al. A twofold increase in BRCA mutation related prostate cancer among Ashkenazi Israelis is not associated with distinctive histopathology. J Med Genet 2003; 40: 787–792.[Free Full Text]

10. Johannsson O, Loman N, Moller T et al. Incidence of malignant tumours in relatives of BRCA1 and BRCA2 germline mutation carriers. Eur J Cancer 1999; 35: 1248–1257.[CrossRef][ISI][Medline]

11. Tonin P, Ghadirian P, Phelan C et al. A large multisite cancer family is linked to BRCA2. J Med Genet 1995; 32: 982–984.[Abstract]

12. Struewing JP, Hartge P, Wacholder S et al. The risk of cancer associated with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. N Engl J Med 1997; 336: 1401–1408.[Abstract/Free Full Text]

13. Gudmundsson J, Johannesdottir G, Bergthorsson JT et al. Different tumor types from BRCA2 carriers show wild-type chromosome deletions on 13q12-q13. Cancer Res 1995; 55: 4830–4832.[Abstract]

14. Meindl A, Scherneck S, Schmutzler RK et al. Comprehensive analysis of 989 patients with breast or ovarian cancer provides BRCA1 and BRCA2 mutation profiles and frequencies for the German population. Int J Cancer 2002; 97: 472–480.[CrossRef][ISI][Medline]

15. National Comprehensive Cancer Network. www.nccn.com.

16. Cancer Incidence in Sweden, 2000. Stockholm: National Board of Health and Welfare 2002.

17. Esteve J, Benhamou E, Raymond L. Statistical methods in cancer research. Descriptive Epidemiology, Volume IV. Lyon, France; IARC Sci Publishers 1994.

18. Antoniou A, Pharoah PD, Narod S et al. Average risks of breast and ovarian cancer associated with BRCA1 or BRCA2 mutations detected in case series unselected for family history: a combined analysis of 22 studies. Am J Hum Genet 2003; 72: 1117–1130.[CrossRef][ISI][Medline]

19. Easton DF, Steele L, Fields P et al. Cancer risks in two large breast cancer families linked to BRCA2 on chromosome 13q12-13. Am J Hum Genet 1997; 61: 120–128.[ISI][Medline]