In a recent Opinion article by Gail Risbridger and colleagues, Breast and prostate cancer: more similar than different (Nature Rev. Cancer 10, 205–212 (2010)), the extensive biological similarities of the hormone-dependant pathways of breast and prostate cancers were discussed1. They also pointed out that mutation of BRCA2 predisposes to both breast and prostate cancer, and that mutation carriers with prostate cancer have a poorer survival rate than patients with prostate cancer who do not have BRCA2 mutations2. However, whether the biological similarities would lead to a familial aggregation of these cancers was not discussed. Familial associations between breast and prostate cancer would imply a shared genetic susceptibility that might go beyond mutations in BRCA2. The recently characterized low-penetrance genes for these cancers provide few clues about a possible shared susceptibility, with the exception of the 8q24 locus. However, the associated single nucleotide polymorphisms found at this locus are not concordant3,4,5,6.

We have used the nationwide Swedish Family Cancer Database, the world's largest data set of its kind7, to study shared familial clustering of various cancers, including breast cancer with prostate cancer8. This study identified 10,553 sons (0 to 70 years old) and 107,518 fathers (unlimited age) with prostate cancer, among a total of 170,000 cancer patients in the offspring generation and more than 800,000 cancer patients in the parental generation. We calculated familial standardized incidence ratios (SIRs) and confidence intervals (CIs) for prostate cancers and other cancers in family members8,9. Separate proband groups were used: affected parents only, affected siblings only, and affected parents and siblings, to suggest genetic modes of inheritance and level of penetrance10. The methods of calculation have been described.

For breast cancer, we found that the SIRs were 1.64, when the mother was diagnosed with breast cancer, 1.91 when the sister was diagnosed and 3.83 when both mother and sister were diagnosed. In discordant analysis when probands were diagnosed with prostate cancer, the SIRs for breast cancer were 1.11, 0.96 and 1.62. For prostate cancer, each of the concordant risks was higher, particularly for brothers (3.40), and fathers and brothers (6.71); the corresponding SIRs for prostate cancer in breast cancer families were 1.19 when the mother was diagnosed with breast cancer, 1.07 when the sister was diagnosed and 1.55 when both mother and sister were diagnosed. All these discordant SIRs were significant at a 5% level (Table 1), except those among siblings; many were significant even at a 1% level. When prostate cancer was diagnosed before the age of 60 years in sons of a mother with breast cancer the SIR was 1.47 (n = 41, 95% CI 1.17–1.82)8.

Table 1 Familial risk of breast and prostate cancer
Full size table

Considering that BRCA2 mutations account for only a small proportion of familial breast cancers in Sweden it is unlikely that the common parent–offspring aggregation of breast and prostate cancers would be explained by these mutations, except in the high-risk families of parent and sibling probands11. However, the median diagnostic age of breast cancer was 54 years in families with both father and brother diagnosed with prostate cancer, and this was higher than that found in BRCA2 mutation carriers (<50 years)12. Therefore, these data suggest that only a small proportion of the prostate–breast aggregation in families was explained by BRCA2 mutation and that new shared susceptibility genes are likely to exist.