This is an epidemiological project designed to address a number of hypotheses related to environmental risks for breast cancer:

Specific Aim 1. Are DDE (bis[4-dichlorophenyl]-1,1-dichloroethane)/PCB (polychlorinated biphenyl compounds) levels related to breast cancer in a nested case-control study?

Specific Aim 2. Are regional differences in breast cancer rates explained by known breast cancer risk factors, or variations in diet or alcohol consumption?

Specific Aim 3. Does regular overnight use of an electric blanket increase risk of breast cancer?

Specific Aim 4. Are levels of 25-hydroxy vitamin D related to breast cancer in a nested case- control study?

To accomplish the specific aims we will utilize prospective data from the Nurses cohort, and perform nested case-control studies based in the cohort of 33,000 women from whom we obtained a blood sample in 1989-90. We will analyze levels of DDE and PCB’s among 430 women expected to develop breast cancer after giving us a blood sample and before June 1, 1996, and compare these with levels among 430 controls matched on year of birth and month of blood return. Similarly, we will analyze these specimens for 25-hydroxy vitamin D, the major circulating metabolite of vitamin D.

In 1992 we enquired about electric blanket us. Use of electric blankets is an important source of exposure to EMF and it has been implicated in breast cancer etiology in a previous case-control study. We will study this exposure prospectively. In four years of follow-up we expect 1,652 cases to occur which will give us ample power to examine this issue.

We will study regional variations in the incidence and mortality of breast cancer prospectively from the start of the study in 1976. We will examine the prevalence of specific risk factors, including screening behavior, in each region, and control for these risk factors to assess the extent to which they explain regional variations in incidence and mortality.

DDE and PCB determinations: In year one, Mary Wolf of the Mt. Sinai Hospital, New York City, developed methods to analyze organochlorine pesticide residues in 0.5 ml samples of plasma rather than 1 ml samples. We then tested these procedures in a blinded quality control test. We submitted 25 blinded split specimens from 12 individuals under code to her laboratory for analysis. The results indicated excellent assay precision. For DDE the within-to-between variance ratio was 0.04, for total PCBs this was 0.02. The Spearman correlation between samples was 0.96 for DDE, 0.98 for PCBs. The mean coefficient of variation between specimens (s.d./mean) percent was 6.3 for DDE, 7.5 for PCB. For DDE the concentration was not significantly correlated with the CV percent (r = 0.07, p = 0.74). For PCB the mean was inversely correlated with CV percent (r = -0.54, p = 0.01), implied that the assay was less precise at lower PCB concentrations. Levels of DDE and PCBs were not significantly correlated (r = 0.10, p = 0.65).

Regional differences in breast cancer rates and known breast cancer risk factor: We are continuing our analyses of regional differences in breast cancer incidence in the Nurses’ Health Study. To date, we do not see clear, large differences in breast cancer incidence, although age-adjusted incidence is somewhat higher in California and lower in Texas than average incidence. Controlling for known breast cancer risk factors further attenuates these differences. Multi variate analysis is ongoing.

Regular use of electric blankets and breast cancer: We have completed our sub-study to analyze the reproducibility of the questions we asked prospectively on the Nurses’ Health Study Questionnaire on electric blanket and waterbed use. We mailed a much more detailed instrument to 500 cohort members, inquiring about lifetime patterns of electric blanket and waterbed use. We received questionnaires back from 421 women. Comparison between the detailed instrument and the original questionnaire suggests that concordance is high for a history of never or very infrequent electric blanket use, as well as a history of long duration and frequent electric blanket use, suggesting that we can discriminate these two extremes of exposure. For example, comparing the supplemental questionnaire with the 1992 Nurses’ Health Study Questionnaire information, and asking about electric blanket use between 1989 and 1991 categorized as 0 years or 3+ years, the Nurses’ Health Study Questionnaire information had sensitivity of 0.94, and specificity of 0.95 compared with the supplemental questionnaire.

Levels of 25-hydroxy vitamin D and breast cancer: We have continued to send specimens to Dr. Michael Holick, at Boston University, and vitamin D analyses are ongoing.

Findings are limited to the preparatory and validation studies. In year 3 we will start to test the main hypotheses.