Interactions between genetic and reproductive factors in breast cancer risk in a French family sample.

Considerable progress has been made in the characterization of the genetic component of breast cancer (BC). However, BC still remains a complex disease involving a genetic component and many other risk factors essentially linked to reproductive-life factors. To search for interactions between genetic and reproductive-life factors in the etiology of BC, a systematic family study was performed in two French hospitals from December 1987 to January 1990 and led to recruitment of 288 families, the IGRC data ("IGRC" refers to the Institut Gustave Roussy and Institut Curie, where the data were obtained). Detailed information on reproductive factors was recorded for probands and female first-degree relatives. Segregation analysis of BC was conducted by taking into account a variable age at onset of disease, by use of the class D regressive logistic model, as implemented in the REGRESS computer program. Segregation analyses of BC in IGRC data showed evidence for the segregation of a dominant gene and additional sister-sister dependence, both when reproductive factors were ignored and when they were included. A significant interaction was detected between the dominant gene and age when reproductive factors were taken into account. Among the reproductive factors included in segregation analysis, parity was found to interact with the dominant-gene effect, and there was an indication of an interaction, albeit not significant, between the dominant gene and age at menarche. Whereas the usual protective effect conferred on breast-cancer risk by high parity remained in nonsusceptible women, it disappeared in susceptible women. The increased BC risk associated with a late age at menarche was higher in susceptible women than in nonsusceptible women. Interactions between inherited predisposition to BC and reproductive factors were detected here for the first time by segregation analysis. It would be of major interest to confirm these results by family studies in other populations.