In most studies of mutation carriers, recruitment of subjects is largely carried out by testing an affected family member from a multiple-case family. Considering the first identified carrier in a family as the "proband," only probands are selected directly on the basis of disease outcome (consistent with standard case-control designs), whereas nonprobands may fall into two other categories with regard to selection. First, those who are selected independently of disease outcome (consistent with standard retrospective cohort designs). Second, those who have decided whether or not to be tested for the identified familial mutation based on whether or not they have cancer (which is not consistent with either design). Many of the affected carriers are prevalent cases, while virtually all unaffected carriers are relatives of cases. Members of the same family are likely to have correlated exposures (both environmental and genetic), and are likely to have correlated disease risks that may be independent of the mutation carried in BRCA1 or BRCA2. In addition, environmental exposures are typically assessed retrospectively, the recall of which may vary by affected status as well. These conditions mean that standard analytical methods cannot necessarily be applied, and it is not immediately obvious how these potential biases may influence estimates of relative risk (RR) obtained from these methods.
However, until data are available from large, prospective cohort studies of unaffected carriers (keeping in mind that such studies may be limited by subjects avoiding much of the disease risk by choosing to undergo prophylactic surgical intervention), we must endeavor to analyze the available data and make appropriate inferences in order to inform mutation carriers, clinicians, and genetic counselors about the possible effect of environmental factors on cancer risk. It is therefore important that multiple, independent studies are conducted, using different analytical approaches that address different potential biases. Indeed, there are several collaborations of researchers currently studying potential environmental modifiers of breast cancer risk in mutation carriers. These include at least four international consortia, each applying different analytical models: the studies led by Narod involving approximately 55 centers from around the world (11,12); the International BRCA1/2 Carrier Cohort Study (IBCCS) (13); the PROSE and MAGIC consortia involving approximately 24 centers in Europe and North America (14); and another, led by the Breast Cancer Family Registry (BCFR) (15), which has combined its carrier data with those of the Kathleen Cuningham Foundation Consortium for Research into Familial Breast Cancer (KConFab) (16) and the Ontario Cancer Genetics Network (OCGN) (17).
The analytical approach adopted by the Narod et al. group has matched affected carriers to unaffected carriers on potential confounders such as age, year of birth, and gene in which the mutation occurs (ignoring familial relationships), and applied conditional logistic regression to matched sets. Apart from being subject to the potential biases mentioned above, this matching approach has resulted in up to 40% of carriers being excluded from the analyses because of the lack of an appropriate match (12), and this may introduce an additional bias. Nevertheless, theirs is currently the largest sample set and represents the largest geographic spectrum. The BCFR-led consortium has not used this matching approach, but has instead applied unconditional logistic regression, adjusting for family history and other potential confounders. The degree to which this approach adequately deals with ascertainment bias depends on how well the family history variable used acts as a surrogate for the familial phenotypes that led to inclusion in the study. In an effort to minimize recall and other biases due to the inclusion of prevalent cases, the BCFR group has also restricted inclusion to cases diagnosed within five years of interview, which has resulted in a reduced dataset in their analyses. The IBCCS and PROSE/MAGIC studies have taken a cohort approach, including all carriers for whom exposure information was available and applying Cox regression models to time to breast cancer diagnosis from birth. The IBCCS study has used a weighted regression approach, with weights determined to account for ascertainment biases due to age-specific preferential sampling of cases relative to controls. This weighting results in reduced power to detect associations (18). All groups have used robust estimates of variance to account for correlations within families. It is clear that each analytic approach adopted has its advantages and disadvantages, and so the most informative and clinically applicable results will be those that are consistent across studies.
A review of the studies assessing environmental modifiers of breast cancer risk among BRCA1 and BRCA2 mutation carriers published to date is presented, by exposure, in the following subsections and summarized in Table 1.
Was this article helpful?
Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.