We use the population-based Family-Cancer Database from Sweden to stud
y familial breast cancer. The size of the population and the nation-wi
de registration of cancer offer unique possibilities for epidemiologic
al studies of familial cancer, including complete and unbiased identif
ication of cases in the probands and in their relatives, and complete
and unbiased identification of the family relationships. Using the Dat
abase, we wanted to answer the following questions: (i) proportion of
familial breast cancer among all breast cancers; (ii) familial relativ
e risks in breast cancer alone or in combination with another cancer,
defined either through the mother or the daughter; (iii) modification
of familial risk by age; and (iv) effects of paternal breast cancer al
one or in combination with maternal breast cancer. The proportion of f
amilial female breast cancer among all breast cancers before 54 years
of age in Sweden was 8.7%. The familiar relative risk was about 1.8, b
ut is likely to decrease to about 1.5 in the ageing population. The hi
gher familial relative risks were evident in young women, being 4.0 wh
en both the mothers and their daughters were diagnosed at ages <40 yea
rs. Paternal breast cancer, in combination with maternal breast cancer
, caused a large (but not statistically significant) risk in the daugh
ters. In mothers and daughters, ovarian but not colon cancer was incre
ased in combination with breast cancer. (C) 1998 Wiley-Liss, Inc.