The above considerations have important implications for understanding how the relationship between moderate alcohol consumption by postmenopausal women and breast cancer incidence can be understood mechanistically. It seems that there are two quite different mechanistic possibilities, which are not mutually exclusive: The first possibility is that alcohol acts on a pre-existing tumor, either by increasing the growth rate of breast cancer cells, or by altering some other aspect of cancer cell biology which increases the likelihood of a clinical diagnosis within the time frame of a given study. We will refer to this as the tumor promoter hypothesis. Another possibility is that alcohol is a weak cumulative carcinogen, and that self-reported alcohol consumption at the time of study entry is a surrogate marker of lifetime exposure. We refer to this as the cumulative carcinogen hypothesis. Because these two explanations have potentially important but fundamentally different implications for disease prevention, we consider these possibilities separately below.
