The Two Sister Study

National Institute Of Environmental Health Sciences

Investigators

Linked publications, trials & patents

Abstract

As discussed in our published overview of the study, we enrolled 50,884 US and Puerto Rican women who were between the ages of 35 and 74 and had a sister with breast cancer but had never had breast cancer themselves when they joined the study between 2003 and 2009. At enrollment, data on potential risk factors and current health status were collected using computer assisted telephone interviews and mailed questionnaires. Blood, urine, and environmental samples were collected in a home visit and banked for future use in nested studies. More than 4,000 Sister Study participants have reported a diagnosis of invasive or in situ breast cancer. The cohort is tracked annually for changes in vital status and major health outcomes. Detailed follow-up questionnaires on health outcomes, environmental and lifestyle exposures, and special topics are completed every 2-3 years. We retrieve medical records and tumor tissue for those who develop cancer or other conditions of interest. Breast cancer and ovarian cancer cases through 2014 and a random sample of the cohort have been genotyped as part of the multi-study "Oncoarray" project. Through this project, Sister Study and Two Sister Study data have been included in many collaborative analyses, including transcriptome-wide association studies. One consortial project was completed by my post-doc, Ann Von Holle, who considered the role of BMI in relation to risk of breast cancer as women transition through the peri-menopausal years. In work with O'Brien we have recently assessed the association between use of genital talcum powder and douching and risk of ovarian and endometrial cancer. We also previously generated data on 450,000 CpGs for the non-Hispanic white women in the genotyping sample, with plans to evaluate methylation patterns in relation to risk factors of interest. In work with a postdoc in the Epidemiology Branch (Kresovich, et al.,published in Molecular Oncology) we developed a methylation-based risk score for breast cancer. This score performed very well in test data and in an independent case-control data set from a substudy of the European EPIC study. It provides a novel risk factor that is evidently as strong as predictor for breast cancer as any known risk factor (excepting age and sex) and is independent of the known risk factors. In other work related to the methylome we recently found that women who experience a breast cancer diagnosis and subsequent treatment show accelerated methylation-assessed biologic aging, particularly if they had been treated with radiation. In work with Diaz Santana we are using the Sister Study data to study both depression and use of anti-depressant medications in relation to risk and found that those medications, which are widely prescribed in the US, appear to be associated with the risk of breast cancer. We are also evaluating methods of risk prediction as applied to Hispanic populations, in particular to women from Puerto Rico, in hopes that more personalized and better-calibrated risk assessment methods can be developed for them. We have assessed levels of metals in toenail samples and found that unfortunately the iron levels in toenails correlate poorly with biomarkers based on serum. In measures involving 17 trace elements in toenails the time of year at which the clipping was done was quite consistently related to the levels measured, with many showing peaks in late summer. Thus, despite their convenience, season of sampling needs to be adjusted for if toenails are used in an epidemiologic study. We also evaluated predictive models for iron status in a paper now under review. In work with a new postdoc, Wanyu Huang, we are developing a mathematical way to summarize risk based on pregnancy history, and will be using the Sister Study cohort to assess its performance for gestational diabetes, type 2 diabetes, and other outcomes.