In December 2005, Deborah Winn's Science and Society article reviewed research findings from the Long Island Breast Cancer Study Project (LIBCSP)1. Her conclusion that the LIBCSP provided “no evidence that environmental exposures were responsible” for breast cancer patterns on Long Island is misleading and understates the complexity of the data. In fact, Winn reviews three reports that show a clear association between biomarkers of environmental exposure and increased risk of breast cancer on Long Island, including: 1) women with the highest exposure to polycyclic aromatic hydrocarbons had a 50% increase in breast cancer risk2; 2) those with elevated levels of a specific polychlorinated biphenyl (PCB) variant in breast fat had a fourfold increased risk of recurrence of the disease3; 3) women who lived within 1 mile of hazardous waste sites that contained organochlorines in Nassau and Suffolk counties had a threefold increased risk of breast cancer4. In addition, her review also notes that women living on Cape Cod for 5 or more years had a 20% increased risk of breast cancer compared with those in the rest of the state. Interestingly, for the Cape Cod cohort, women who had been living there for 25–30 years — since 1948, the year that the use of DDT began in that area — had the greatest odds ratio for developing cancer5.'

Experimental research has clearly shown that many environmental exposures that cannot easily be studied in humans can damage cells and cause disease, including cancer. It is very difficult to reconstruct real-life exposures for a multifactoral disease like breast cancer, where the timing of exposure is uncertain and the ability to reconstruct a lifetime exposure assessment using biomarker analysis or Geographical Information System (GIS) mapping techniques are limited. Despite all of the epidemiological research that has been performed, the causes of regional 'hot spots' for breast cancer in the United States, are not well understood at this time.

The use of case–control studies to elucidate geographic risks of breast cancer, as was done in the LIBCSP studies, raises the question of what are the appropriate controls. In an area of high risk, it is possible that cases might be the most sensitive or hypersusceptible individuals, whereas local, regional controls are those who are delayed in developing breast cancer. This idea is supported by a recent study of sisters that were discordant for breast cancer, which found that DNA repair capacity was lower in patients with breast cancer compared with their disease-free sisters. Women in the lowest quartile for DNA repair had nearly 3 times the risk of breast cancer compared with their individual sisters (2.99 CI = 1.45–6.17, P = .02) (Ref. 6).

In truth, a number of studies conducted as part of the LIBCSP did not find associations between the environment and breast cancer. The absence of evidence connecting specific environmental factors and patterns of breast cancer in some studies should not be confused with proof that such a connection does not exist. Rather, it could reflect the methodological complexities of epidemiological studies, the difficulties of obtaining appropriate surrogates or direct measures of relevant exposures that can take place over a lifetime, and the dearth of reliable biomarkers for accurate correlative historical studies.