International journal of cancer 2017 08 26141(11) 2253-2259 doi 10.1002/ijc.30927
Familial risks of breast cancer (BC) are well established but whether BC clusters with other, i.e. discordant, cancers is less certain but of interest for the identification of common genetic and possible environmental factors contributing to a general cancer susceptibility. We apply a novel approach to search for familial associations of BC with other (discordant) cancers based on the Swedish Family-Cancer Database. Relative risks (RRs) were calculated for BC in families with increasing numbers of patients with discordant cancer X, and conversely, familial RRs for cancer X in families with increasing numbers of BC patients. Joint p-values were calculated from independent analyses. The total number of familial BCs was 12,266, 14.9% with one first-degree relative with BC and 1.2% with at least 2 affected relatives. Ovarian and prostate cancers showed the strongest associations with BC (p-value <10-11 ). The p-value for melanoma was <10-6 , for stomach and male colorectal cancer <2.5 × 10-6 , for cancer of unknown primary <2.5 × 10-5 and for lung cancer <5 × 10-5 . Significance level <5 × 10-4 was reached with pancreatic cancer. The remaining associations (p < 0.0025) included thyroid, endometrial, testicular, eye cancers (uveal melanoma), nervous system and endocrine tumors and non-Hodgkin lymphoma. The RR for BC increased by increasing numbers of patients with any cancer in family members and it reached 1.62 when three or more family members were affected. The results suggest that BC shares susceptibility with a number of other cancers. This might alert genetic counselors and challenge approaches for gene and gene-environment identification.