While cancer frightens us all, breast cancer occupies a special place in the pantheon of human horrors because it attacks not only a woman's body, but also a part of her physiology that -- rightly or wrongly -- helps her define herself.
So women who consume alcohol should look with concern and interest at medical research that indicates that even moderate alcohol consumption may be associated with an increased risk of getting breast cancer. But it may not be.
Breast cancer is such an emotional subject that the news media (who typically have no scientific background with which to make informed decisions on, or evaluate the credibility of, medical research) tend to sensationalize research results and present a fragmented and out-of-context image of the science. In this respect, women are not well served by the American media. As a result, women need to make the extra effort to gather a complete set of information themselves, consult with their physicians and make the necessary informed decision that cannot be made when relying solely on media reports. The chapter in this book on evaluating research will give you more tips on this process.
Even the best scientific minds find breast cancer and alcohol research muddled, flawed, inconclusive and incapable of supporting a definitive public health decision. Unlike studies which can support clear-cut decisions (such as the ones on the cardio-protective effects of moderate drinking), research on the association between alcohol and breast cancer is inconsistent, shows different results from one study to another and is, on the whole, not reproducible in a manner which points to a scientific conclusion.
This is indicated by a review of 23 major studies conducted by Dr. Keith Marton who found that 12 indicated some positive relationship between moderate alcohol consumption and breast cancer, nine indicated no increased risk and two were so flawed that the results could not be reliably interpreted. Dr. Marton has recently updated his review, adding more studies. The majority of the more recent studies fail to show a relationship between breast cancer risk and consumption of two or fewer drinks per day. However, the theme established by the earlier studies remains: no conclusive trend has been found.
Please note that in this context, "positive" does not mean "good" but indicates that a relationship exists; a negative relationship means that no relationship exists. This may be somewhat confusing and this clarification will be repeated to help decrease the confusion.
Some of the nine studies showing no increased risk which Dr. Marton reviewed actually indicated some slight decreased risk of breast cancer in moderate drinkers, but he felt that the decreases might not be statistically significant.
If, to be safe, you choose to ignore the nine studies showing no relationship and focus on the 12 that do, you find that the increased risk for women consuming moderately is about 50 percent.
If these are the correct results (and as you will read below, many scientists believe they are not accurate), this means that a woman's lifetime chance of dying from breast cancer may increase about 4 percent, from 8 to 12 percent. However, Dr. Marton said that even if you accept the worst-case studies as valid, a woman who drinks moderately must balance the 4 percent higher risk of getting breast cancer against the 20 to 30 percent lower risk of dying of heart disease -- a net positive effect of 6 to 14 percent.
"Many of these studies," Dr. Marton said, "do not account for other factors (other than alcohol) that might affect the risk of cancer.
For example, one study conducted by the U.S. Centers for Disease Control and published in the Sept. 24, 1983, issue of Lancet found no increase or a slightly decreased risk of breast cancer for moderate women drinkers (Relative Risk -- as compared to abstainers -- of 1.0 for beer drinkers, 0.9 for spirits drinkers and 0.8 for wine drinkers). In examining why its results varied from earlier research that indicated an increased risk, CDC researchers found that one study used women with ovarian and endometrial cancers as a control group. "However, our data indicate a possible protective [emphasis added] effect of alcohol consumption on the risk of endometrial cancer. We are currently investigating the hypothesis that the increased breast cancer risk observed by [these studies] may be due to alcohol's protective effect on endometrial cancer rather than its causal effect on breast cancer."
In other words, the control group -- the group against which the drinkers were compared -- was not properly chosen and this biased the conclusions. Because epidemiological studies like these involve the mathematical manipulation of hundreds of thousands of pieces of data, the strategies used to collect and analyze the data can introduce hidden biases.
This was pointed out in a 1988 analysis of two reports from the New England Journal of Medicine, conducted by Nathan Mantel of the Mathematics, Statistics and Computer Science Department at American University and published in the book, Preventive Medicine. Mantel found that the two 1987 studies (one by Schatzkin et al and the second by Willett et at) which indicated a positive association between alcohol and breast cancer "did not inquire into other cancers, morbidities, mortalities or total mortality" and did not correct for other physical and lifestyle characteristics, particularly obesity, that could affect the research conclusions.
"Thus the drinkers and the nondrinkers did not start in at even levels as they might have in an ideal study. And if they were not starting at even levels, they could well have been starting at uneven predilection levels for breast cancer." Mantel called the studies "defective" because the needed data was not collected and said that "it would be irresponsible to lead women to believe that moderate alcohol consumption is dangerous."
Mantel also took a shot at the news media, charging that "journalists seem bound to make. ..unqualified interpretations."
Similar biases can be found in other studies which show positive relationships (meaning that a relationship exists). But biases are not the only problems that plague research into alcohol and breast cancer. Indirect associations may also be to blame for the positive relationships.
This may be true with alcohol and breast cancer. In the same 1988 Preventive Medicine textbook which published Mantel's analysis, a team of Veterans Administration physicians lead by Dr. Randall P. Harris, M.D., Ph.D., of the American Health Foundation, said that the apparent increase in breast cancers associated with alcohol in some studies may be linked not to alcohol, but socio-economics.
Their logic goes as follows:
(1) Alcohol consumption increases linearly with education. Only 39 percent of women with less than nine years of education call themselves drinkers; that rises to about 70 percent for women with 16 years of education.
(2) Well-educated women most often postpone pregnancies, often into their late 2Os and 30s; less-educated women frequently have children in their teens.
(3) Breast cancer risk increases as the onset of the first pregnancy is postponed.
Therefore, the increased risk of breast cancer may be due to the postponement of pregnancy among educated women -- who tend to drink more and postpone pregnancy - and not to alcohol consumption (which is perhaps only indirectly and not causally related).
Supporting this hypothesis is the fact that almost all of the studies which showed a positive relationship between alcohol and breast cancer showed the highest risk among thin drinkers. This is at odds with general data that shows obese women to be at higher risk for breast cancer. The explanation? Highly educated women (who drink more and who postpone their first pregnancy) tend to be thinner than less-educated women. Thinness, then, is also indirectly (but not causally) related to breast cancer.
However, thinness could be related in another way. The same amount of alcohol consumption would produce a higher level of blood alcohol in thin women than in those who are larger, thus accentuating the effects of alcohol.
Science simply has no conclusive answer on these hypotheses.
Because epidemiological studies such as these are subject to biases and indirect associations, one standard which scientists use to assess the credibility of the results is the presence of a logical biological mechanism. In other words, the most believable epidemiological studies are the ones which are backed up by laboratory studies (human or animal) which point to the way in which an agent like alcohol produces good or bad effects.
However, a biological mechanism to account for the possible breast cancer increases among moderate drinkers has not been discovered. In experiments on laboratory animals, there has been no association between alcohol and breast cancer.
Because of this lack of a plausible mechanism, the studies showing a positive association between breast cancer and moderate consumption are all suspect and may all be indirect associations.
Finally, the under-reporting of alcohol consumption which plagues epidemiological studies may mean that significantly larger amounts of alcohol (i.e., abuse) may be required to cause any increase in breast cancers.
How can a woman translate the studies into a personal decision relating to her own health? Decrease the chances of heart disease and increase (to a smaller extent) the risks of breast cancer or vice versa?
The decision must be made not just on the scientific data, but on personal background and preferences. Women with a history of alcohol abuse or breast cancer in their families should be the most cautious since they are already at greater risk from diseases. On the other hand, women without these factors may feel less constrained, especially if they particularly enjoy wine. Women who choose this option, however, should be aware that their smaller size means that the definition of moderate is set at a lower level than for men -- as is the level at which abuse begins.
What this means is that women must accept responsibility for gathering all of their information and making their own informed choices. They must also recognize that their needs are not served well by the news media which tends to sensationalize and make unqualified conclusions. Likewise, women should be skeptical of advocacy groups who may conceal information and distort research results in order to make statements about drinking behavior which support their positions.