Prognosis of Breast Cancer in Carriers ofBRCA1andBRCA2Mutations

To the Editor: Rennert et al. (July 12 issue)1 report that a BRCA1 or BRCA2 mutation was not a significant predictor of mortality from breast cancer in an Israeli population-based cohort. However, 43% of potential cases were not included because tumor samples or medical records could not be retrieved, and excluded cases were more likely than included cases to be fatal. If BRCA1 carriers have more aggressive tumors,2 leading to earlier death, this could result in a spuriously improved prognosis (Neyman bias). In addition, the investigators did not study tumor grade (associated with mutation status),3 nor did they identify prognostic effects of tumor and nodal stage, raising concerns about internal validity. Furthermore, greater chemotherapy use among carriers (37%) than among noncarriers (24%) suggests that decisions about chemotherapy may have been influenced by factors such as family history, leading to a spurious treatment benefit in carriers. Finally, insufficient detail was provided regarding chemotherapy use (BRCA1-associated breast cancer may have increased sensitivity to agents causing doublestrand DNA breaks4 and reduced sensitivity to mitotic spindle poisons) for a full understanding of the effect of chemotherapy in carriers of these mutations.5