Milk & Breast Cancer
By Mr. Stephen Walsh
In August and September 2001 there were claims in the press that drinking milk could reduce the risk of breast cancer. Dairy Council nutrition manager Anita Wells said, "To discover that life-long milk drinkers have a reduced risk of breast cancer is an exciting step forward." "Good health starts with dairy," said Greg Miller, executive vice president of nutrition and scientific affairs for the National Dairy Council.
These claims were triggered by a prospective study of premenopausal women in Norway, published on 15 September 2001 in the International Journal of Cancer. This study found a significantly reduced risk of breast cancer in women who reported high milk consumption both as adults and as children compared with women who reported low milk consumption both as adults and as children. The risk of breast cancer among the high milk group appeared to be about half the risk in the low milk group.
If this had been one of many studies finding similar results, the claim of a protective effect would have some credibility. However, considering other prospective studies on breast cancer and dairy products shows that this is not so. One study found a significant increase in risk with milk consumption and another found a significant increase in risk with cheese consumption. A Finnish study agreed with the recent Norwegian study that there was a protective association with milk products. Two other studies found no significant association, with one finding a tendency to a beneficial effect and the other finding a tendency to an adverse effect. The authors of the most recent study state, "The contradicting results may indicate that any association between milk and breast cancer is not a strong one." This seems a fair conclusion. Indeed, if a true effect exists there is no reason to assume it will prove to be beneficial. The suggestion that consuming milk reduces breast cancer risk is therefore unjustified.
Some reports latch on to a constituent of dairy fat - conjugated linoleic acid (CLA) - as a likely explanation for the suggested beneficial effect, suggesting that milk fat might benefit health. Neither of the two prospective studies which found a beneficial association of milk with breast cancer showed a link with milk fat. The more likely explanation for any beneficial effect, if one exists, is the calcium and vitamin D content of the milk. Even if CLA showed some benefits, it must be recognised that it is only a small part of dairy fat and that breast cancer is only one aspect of health. Milk fat is highly saturated and promotes elevated cholesterol, which is strongly linked with heart disease. Milk fat is therefore a very unlikely choice as a health food. Calcium and vitamin D can of course be obtained from many sources other than dairy products.
Consuming milk increases levels of a growth hormone, IGF-1, in the body. Increased IGF-1 levels are strongly implicated in prostate cancer, colorectal cancer, premenopausal breast cancer and lung cancer. The effect of milk on IGF-1 may be due to absorption of IGF-1 from the milk or may simply be due to the high protein and zinc content of the milk.
For those concerned with their risk of breast cancer and with their overall health there are useful recommendations to consider. Weight gain in adult life increases breast cancer risk. Physical activity helps in maintaining a steady weight and also has additional benefits in reducing risk. Olive oil provides a healthful alternative to saturated fats. Alcohol consumption increases risk of breast cancer, even at moderate levels, but this increase in risk can be effectively countered by high folate intakes. As moderate alcohol intake is associated with reduced risk of heart disease, increasing folate intake may be a better option than cutting out alcohol altogether. Folate is found in green leafy vegetables and many other plant foods and can also be obtained from folic acid supplements. Adequate B12 may be needed to get the full benefit of folate. Adequate calcium and vitamin D may also be beneficial.
There is a lot we can do to take control of our health for the better, including reducing risk of breast cancer. Green leafy vegetables, olive oil and physical activity can all be expected to be beneficial. Drinking cow's milk doesn't appear on the list.
Good information supports health. Pass it around.
What the dairy industry would like you to believe
Dairy industry representatives sprang into action to promote the idea that drinking milk reduces risk of breast cancer:
Anti-cancer link's all white for milk
South Wales Evening Post, 30 August 2001.
Farmers across South West Wales have welcomed new evidence that women who drink three glasses of milk a day are less likely to develop breast cancer.
Members of the Farmers' Union of Wales say the news is a boost for their White Stuff campaign to give free milk to all Welsh primary school pupils.
The FUW has urged all primary schools in Wales to take advantage of the scheme introduced last February, in which the Assembly and the EC pay the cost of school milk for pupils aged between five and seven. …
The Dairy Council's nutrition manager Anita Wells said: "To discover that life-long milk drinkers have a reduced risk of breast cancer is an exciting step forward."
Milk may stop breast cancer
The New York Post, 5 September 2001.
The new anti-breast cancer slogan could be: Got milk?
A new study reveals that drinking milk in childhood and as an adult can protect women against the disease.
Women who drank more than three glasses of milk per day had half the risk of breast cancer, compared with women not drinking milk.
Researchers in Norway came to their conclusions after studying the health histories of 48,844 women, ages 34 to 39.
Milk's protective effect is attributed to a cancer-fighting substance called conjugated linoleic acid, which is found in milk fat.
Numerous studies have shown that the acid has a positive effect in fighting off breast cancer. The results of the study were published in the September issue of the International Journal of Cancer.
October is National Breast Cancer Awareness Month; Drinking Milk May Reduce the Risk of Breast Cancer
Business Wire, 27 September 2001
National Dairy Council/Milk Processors Education Program press release
A new study published in the September 15th issue of the International Journal of Cancer found that women who drink milk on a daily basis are less likely to develop breast cancer than those who drink little or no milk….
CLA or conjugated linoleic acid - which is found in milk fat - has been gaining a lot of attention lately as a potential cancer prevention agent. A recent Finnish study found that women with breast cancer had significantly lower levels of CLA in their diets and blood compared to women without cancer….
The results of the Norwegian Women and Cancer Study are in line with earlier research results showing the protective effect of milk on breast cancer. A study published in the British Journal of Cancer in 1996 showed that the women who consumed the most milk had less than half the risk of breast cancer compared to women consuming the least milk.
"Good health starts with dairy," said Greg Miller, Ph.D., F.A.C.N., executive vice president of nutrition and scientific affairs for the National Dairy Council.
In fact, the evidence for an association between milk and breast cancer is weak and conflicting and does not justify a claim for a protective effect. There are many ways of reducing breast cancer risk while promoting overall health which are properly justified and do not involve any harm to other animals. The rest of this paper justifies these observations in detail.
Is there an association between milk and breast cancer?
How are associations between diet and disease established?
Before looking specifically at milk and breast cancer, it is useful to consider some general methods by which researchers have attempted to establish dietary causes for human cancers by studying people. Readers already familiar with nutritional epidemiology can skip this section, as it is largely standard material.
Regional or "ecological" comparison
A common form of study involves comparing variations in disease rates in different countries or regions with variations in diet. This will often show associations that are unlikely to be due simply to chance, but there are many factors that vary between countries and most apparent associations are not causal. It is particularly difficult to differentiate between the effects of the many distinct changes that occur with economic development. These typically include increased consumption of animal products and fat, adequate calorie intake throughout life, less physical activity, access to effective sanitation and medical services and having fewer children at a later age. For this reason, regional comparison is regarded as a weak form of evidence. A strength of inter-country comparisons is that characteristics such as average diet are well defined and vary over a wide range.
Case control studies
Another source of evidence is provided by case control studies. In such studies, a group of people with established disease ("cases") are compared with a group of people who are free from the disease ("controls"). Statistical analysis is then used to identify characteristics that differentiate the cases and the controls and the results are expressed as a relative risk (RR) for the disease associated with such characteristics. The results will usually be given in terms of the value of the relative risk comparing two groups. Typically the overall group will be divided into thirds (tertiles), quarters (quartiles) or fifths (quintiles) after ordering the group members from the lowest to highest values of some characteristic. The relative risk will compare the group with the highest values of that characteristic with the group with the lowest values, e.g. upper quintile (fifth of group with highest values for characteristic) vs lower quintile.
A relative risk of 1 means no difference in risk between the two groups. The probability that the relative risk differed from 1 purely by chance will also be reported. If the probability of a relative risk being different to 1 by chance is less than 0.05 that association is usually described as statistically "significant". The choice of 0.05 as a threshold is a compromise between getting an excessive number of false associations due to chance and missing a real relationship, and is to some extent arbitrary. For example, a relative risk of 1.5 between the upper quintile and the lower quintile with a probability of 0.003 would suggest a moderate and clearly statistically significant association, i.e. very unlikely to be due to chance. The RR of 1.5 means that the upper quintile show 50% greater risk of the disease being studied than the lower quintile. A relative risk of 5 with a probability of 0.04 would indicate a strong effect with modest statistical significance. A relative risk of 10 with a probability of 0.3 means little, as the observed relative risk could easily be a chance observation so the result is not significant.
There are two major weaknesses in case control studies looking at diet. Firstly, the controls may not be representative of the general population being studied (selection bias). Often controls show a greater interest in health than most people, making them more likely to volunteer, and they may therefore show an unusually high consumption of foods believed to be healthy. This can lead to a false conclusion that these "healthy" foods make development of the disease less likely, where in fact the controls were simply unrepresentative of the disease-free population. Secondly, cases' current diet may be altered as a consequence of the disease and their recollection of past diet may also be changed (recall bias). A simple example of recall bias is that people with hip fractures tend to underestimate past milk consumption. This underestimation makes case control studies likely to conclude that milk consumption is protective against fractures, regardless of the true effect. Because of these biases, case control studies are viewed with considerable scepticism. However, case control studies have the potential strength that, because relatively few people are involved, it is possible to carry out very thorough evaluations of diet. This can make them more likely to detect a true effect than studies that rely entirely on simple questionnaires to evaluate diet.
Prospective studies overcome the biases of case control studies. In a prospective study a sample of the population (called a cohort) who are free from the diseases of interest at the start of the study is examined. As there are no cases at this stage, selection and recall biases do not come into play. As in case control studies, the volunteers may be healthier than the general population but so long as comparisons are made within the cohort this effect will not introduce a bias. Comparisons between the study cohort and the general population are, however, subject to selection bias. To provide increased confidence that pre-existing disease is not affecting the initial observations it is common practice to eliminate cases arising within a few years of the start of the study from the subsequent analysis. If a significant relative risk is observed in a prospective study that has followed these precautions then it is more likely to be a true association than one obtained from a case control study. A weakness of prospective studies is the difficulty of accurately characterising the population studied without excessive cost as they must include large numbers of people to ensure that a reasonable number of cases will arise during the study. Inaccurate characterisation may lead to true effects being missed.
Ecological, case control and prospective studies are all subject to confounding - where an apparent effect from one characteristic is actually due to a different but related characteristic. As mentioned previously, this problem is particularly severe in comparisons between countries based on a characteristic which changes with economic development. Any such characteristic will be closely associated with many other characteristics, so cause and effect cannot be inferred. Once a behaviour is perceived to be healthy, new possibilities for confounding arise. If people who are at particular risk of a disease adopt a behaviour believed to be protective, then future studies may find the protective behaviour to be associated with increased risk. This may apply, for example, to small boned individuals and individuals with a family history of fracture increasing their calcium intake to reduce fracture risk.
If several characteristics are strongly associated, it can be difficult to separate out the causal characteristic even if all the potentially relevant characteristics are measured. If a relevant characteristic is not measured or not analysed, a false association may readily arise. A key example of this is that intake of most foods is strongly associated with total energy intake, which in turn may be associated with obesity, physical activity and hormone levels. It is therefore standard practice to use statistical methods to adjust for energy intake so that this effect is removed from the analysis. Age must also be adjusted for, as it is an important factor in every disease. Other known risk factors for a particular disease should be measured and adjusted for. Risk of breast cancer rises with earlier age of menarche, later age at first live birth, later age at menopause, increased BMI (weight in kg divided by height in metres squared) in postmenopausal women, and increased height. It would be normal to present results after adjustment for these factors. As some of these factors are themselves dependent on diet it can be useful to present results with adjustment for age and energy only, as well as presenting fully adjusted results. If the partially and fully adjusted risks are very different, particularly careful analysis is required to try to separate the influence of diet.
Randomised intervention studies
The gold standard in investigating diet would be to randomly assign individuals to different diets for many years and compare the outcome. This is generally not practical as people are unlikely to cooperate and large-scale trials are very expensive. However, it is possible to carry out relatively short term or partial dietary modifications. Results from such trials are unlikely to be due to confounding, but the duration may be too short to be definitive in relation to the lifetime effect of a behaviour. If the duration of the study is too short to evaluate the final outcome of interest, such trials may use intermediate risk markers as an outcome measure, e.g. change in bone mineral density or biochemical markers of bone resorption following calcium supplementation.
In practice, a combination of randomised intervention studies and prospective studies provides the best evidence. If short duration intervention studies and longer duration prospective studies show conflicting results, then it is difficult to draw firm conclusions.
The particular strengths and weaknesses of each method should be borne in mind in looking at results on diet and health.
What is the evidence on milk and breast cancer?
If we compare breast cancer rates and milk consumption between different countries, we find that countries consuming large amounts of dairy products have higher rates of breast cancer than those consuming low amounts of dairy products . As noted above, this association may well be non-causal and we need to consider other evidence.
Results from individual case control and prospective studies are very diverse so it is essential to look at the studies as a whole. Picking studies to suit a conclusion is easily done but has no value in the search for truth.
Boyd et al. 1993  review studies of dietary fat and breast cancer risk up to February 1993.
7 case control studies considered milk intake.
5 found that high milk intakes were associated with significantly increased risk of breast cancer.
2 found no significant effect.
5 case control studies considered cheese. 3 of the results on cheese came from studies also reporting results on milk (above), and 2 came from separate studies.
3 found a significant adverse effect,
1 found a significant protective effect and
1 found no significant effect.
Only one prospective study considered milk and cheese and found no significant effect for milk (relative risk (RR) 1.03) and a significant adverse effect for cheese.
The summary statistics from this review showed a modest, but significant, adverse effect of both milk and cheese on breast cancer risk (RR =1.17 for each).
The above review did not include the paper by Ursin et al. 1990 . This paper was a prospective study of milk consumption and cancer risk in Norway. Individuals consuming two or more glasses of milk a day vs less than one glass a day showed a relative risk of 1.48 for breast cancer, but this was not statistically significant. Overall cancer rates showed a non-significant relative risk of 0.99. The results were not adjusted for standard breast cancer risk factors.
So up to until 1993, the clear balance of evidence was for a modest adverse effect of milk and cheese on breast cancer risk. The evidence was, however, dominated by case control studies and therefore subject to selection and recall bias.
Toniolo et al. 1994  present a prospective study from New York. This study showed an almost significant protective association of high dairy intake (RR=0.59 upper quintile vs lower quintile) with breast cancer. Milk and other dairy products were not separated. Adjustments were made for standard breast cancer risk factors.
Gaard et al. 1995  present another Norwegian prospective study. Adjustments were made for age, energy, smoking, height and BMI but not for other standard breast cancer risk factors. This study found a significant elevated risk with higher consumption of whole milk (RR=2.91 upper vs lower quintile). The corresponding RR for all forms of milk combined was 1.71, but this was not statistically significant.
Knekt et al. 1996  present a Finnish prospective study. This study had relatively few participants, 4697, but follow-up was over 25 years so the total years at risk were high. Adjustments were made for many risk factors, but not for height or for age at menarche, first childbirth or menopause. The adjusted relative risk for the highest compared with the lowest tertile of the cohort by milk intake was 0.49 and was significant. Adjustments for some other foods did not eliminate the association. Calcium showed an almost identical relative risk to milk, but milk fat showed a weaker and non-significant association. Other nutrients were noted not to show a significant association. It has been suggested that fermented milk may have particular benefits in relation to cancer. There was no association between breast cancer and fermented milk consumption in this study. The association observed was with ordinary milk.
Hjartaker et al. 2001  present a further Norwegian prospective study. This study was limited to premenopausal women. There were 48,844 participants and follow-up was for about 6 years. Results were presented in two forms: with age adjustment only and with adjustment for age and other risk factors, but not for height, age at menopause or energy intake. The dietary questionnaire was insufficiently comprehensive to evaluate energy intake. Women with prior cancer diagnosis were excluded as required to avoid recall bias and elimination of cases arising within one year of the start of the study was shown not to alter the conclusions. Association with adult milk consumption was significant only in the youngest age group (34-39). The negative association tended to be stronger for low-fat milk than for skimmed or whole milk, but was not significant for any of these. Associations with milk consumption as a child were also not significant. Milk fat intake showed a much weaker association with breast cancer risk than overall milk intake did, suggesting that fat was not the key component of any effect of milk. The effect of calcium was not tested.
Only by using a combined measure of childhood and adult milk intake was a statistically significant protective association found and this was only just significant (RR=0.51 "high" milk intake vs "low" milk intake, with full adjustment for known risk factors). 11% of the overall group was in the low milk category. This category had 36 cases of breast cancer against an expected 29, based on the average risk for the whole group. 7% of the group was in the high milk category. This category had 13 cases of breast cancer against an expected 20, based on the average risk. There was little difference between the age adjusted and fully adjusted relative risks, indicating that any interactions between milk consumption and known risk factors, such as age at menarche, did not have a large effect on the observed risk.
The dairy industry sources quoted above also refer to a Finnish case control study, Aro et al. 2000 . This study made use of both dietary questionnaires and evaluation of fatty acid levels in blood samples. Protective associations are found with consumption of cheese and with levels of conjugated linoleic acid, myristic acid and trans-vaccenic acid in the blood. All these fatty acid levels are primarily associated with recent dairy fat intake. There is no particular reason to regard this case control study as carrying more weight than the many other case control studies discussed earlier, which generally indicated an adverse effect of milk and cheese.
The trigger for the dairy industry campaign appears to be a result of modest statistical significance for which there is equally strong contradictory evidence. As Hjartaker et al.  note, "The contradicting results may indicate that any association between milk and breast cancer is not a strong one". Indeed, any effect that may exist could prove to be adverse rather than beneficial.
Of the components of milk suggested as potentially beneficial (calcium, vitamin D, and CLA), only CLA is fairly specific to milk and beef, while calcium and vitamin D have many non-animal sources. CLA forms a small part of dairy fat. Dairy fat is highly saturated and therefore has a strong effect in increasing cholesterol. The ill effect of elevated cholesterol on mortality from heart disease and on overall mortality is well established. The main studies above that were suggestive of a beneficial effect of milk on breast cancer (, ) also suggested that if such an effect existed it was unlikely to be dependent on milk fat. In  calcium showed a clear protective association with breast cancer risk. There is independent evidence that vitamin D (present naturally in milk and also added to milk in many countries) from diet and sunlight has a role to play in breast cancer (, ). As sunlight is insufficient to produce vitamin D for almost 6 months of the year in Scandinavia, this may be a reason why the most positive results on milk and breast cancer have come from Scandinavia. However, calcium and vitamin D can be efficiently obtained by other means than dairy products, so it was therefore convenient for the dairy and beef suppliers to emphasise CLA.
This concludes the direct review of evidence on milk and breast cancer. However, before discussing recommendations for reducing breast cancer risk, it is important to consider other evidence for a role of milk in increasing a major risk factor for premenopausal breast cancer, colorectal cancer, lung cancer and prostate cancer.
Milk, IGF-1 and cancer
CLA, calcium, and vitamin D have been hypothesised to give rise to a beneficial effect of dairy products. Insulin-like growth factor 1 (IGF-1) has been hypothesised to give rise to an adverse effect of dairy products.
High levels of IGF-1 in blood, or more specifically, a high ratio of IGF-1 to IGFBP-3 (insulin-like growth factor binding protein 3), have been associated with notably increased risk of premenopausal breast cancer, prostate cancer, colorectal cancer and lung cancer. The increase in risk observed is about 1-2% per ng/ml of IGF-1, so a 10 ng/ml increase in IGF-1 would correspond to an increase in cancer risk of about 10-20%. All cow's milk contains some IGF-1 and milk from cows treated with bovine growth hormone (sometimes referred to as BST) contains increased amounts of IGF-1. Milk is also high in protein and zinc. Consumption of these nutrients is associated with increased IGF-1 levels.
Adding about 600ml (three cups) of milk to the diet of adults aged 55 to 85 years caused an increase in IGF-1 levels from an average of 125 ng/ml to an average of 137 mg/ml . Reference  provides a correlation of IGF-1 levels with zinc intake for postmenopausal women indicating that IGF-1 increases by about 5ng/ml for each mg per day of zinc intake. 600 ml of milk would provide about 2.4 mg of zinc giving a predicted increase of 12 ng/ml, exactly as observed. This consistency between observation and prediction does not support an effect of milk in increasing IGF-1 beyond that expected from its content of other nutrients. It should be noted that zinc intake is strongly correlated with protein intake and both may contribute to the observed link between diet and IGF-1. A direct effect of the IGF-1 in milk is not indicated by the existing evidence but cannot be entirely ruled out. Regardless of the mechanism, adding milk to the diet increases IGF-1 level. This may lead to an increase in cancer risk.
Until very recently no study had looked at the effect of dairy products on the IGF-1 to IGFBP-3 ratio or the interaction of potential beneficial effects from calcium or vitamin D with potential adverse effects due to elevated IGF-1. A study published on 15 September 2001, does just that . A prospective study was carried out on milk intake and colorectal cancer, including evaluation of the effect of milk intake on IGF-1 and IGFBP-3. In comparing the upper tertile in terms of consumption with the lower tertile, low-fat milk, calcium from milk and calcium from dairy foods all showed a statistically significant 10-16% increase in IGF-1 and a 6-7.5% increase in the IGF-1:IGFBP-3 ratio. Poultry and fish showed similar associations but these did not reach statistical significance. Other forms of dairy products including milk, hard cheese and ice cream also did not show significant associations.
The authors of the study leave as an open question whether the increase in IGF-1 in response to low-fat milk was due to conventional dietary factors, particularly protein, or to absorption of IGF-1 from the milk. It has been suggested to the authors that they analyse the effect of milk consumption adjusted for either protein or zinc intake to clarify this important question.
The study found the following in relation to colorectal cancer:
Intake of low-fat milk showed a moderate, but non-significant, protective association.
Elevated IGF-1/IGFBP-3 showed a statistically significant adverse effect (RR=1.8 upper tertile to lower tertile).
Among the upper tertile of IGF-1/IGFBP-3, intake of low-fat milk showed a significant protective association (RR=0.78 for moderate milk drinkers and 0.39 for frequent milk drinkers).
In men in the lower tertile by IGF-1:IGFBP-3 ratio no effect of low-fat milk consumption was observed.
Among frequent milk drinkers there was no association between IGF-1:IGFBP-3 ratio and risk (RR=1.03).
In non-drinkers of low-fat milk there was a relative risk of 3.05 for those in the highest tertile by IGF-1:IGFBP-3 ratio.
The authors conclude, "this prospective study shows a protective effect of dietary calcium on colorectal cancer among men with a high IGF-1/IGFBP-3 despite the moderate positive influence of milk or dairy food intake on circulating IGF-1 levels." With high calcium intakes it appears that colorectal cancer ceases to be IGF-1 dependent. That is, the adverse effect of elevated IGF-1 on colorectal cancer appears to be conditional on low calcium intake.
It is likely that the net effect of low-fat milk on colorectal cancer is modestly beneficial. For breast cancer, the evidence discussed above suggests that the various effects of milk are in approximate balance. For prostate cancer the net effect of milk appears to be notably adverse. In reference  Willett notes that "one of the most consistent observations has been an association between the consumption of dairy products and prostate cancer incidence or mortality." The net effect on lung cancer is unclear.
Using milk to provide calcium entails accepting an accompanying increase in circulating IGF-1, probably due to increased protein and zinc intake, but possibly due to increased intake of IGF-1. If protein and zinc intake is already sufficient, the addition of milk to the diet therefore poses an unnecessary extra risk. Even if extra protein or zinc would be beneficial, the possibility of a direct effect of IGF-1 in milk on circulating IGF-1 levels suggests that other sources may be preferable.
So, get an adequate calcium intake (at least 800 mg per day) from sources other than dairy and avoid excessive protein or zinc intake (one gram of protein per kilogram of body weight per day and 10 mg of zinc per day is adequate for almost everyone). You have nothing to lose.
Reducing breast cancer risk and promoting health
In contrast to the statements from dairy industry representatives, milk does not seem to be the answer to breast cancer. Calcium and vitamin D may have a beneficial role. However, even these do not rank in the established dietary recommendations for reducing breast cancer incidence.
Walter Willett, a nutritional epidemiologist with the Harvard School of Public Health and the American Institute for Cancer Research makes the following recommendations for avoiding breast cancer .
Avoid weight gain in adult life.
Limit alcohol intake to moderate amounts.
Ensure sufficient folate (or supplementary folic acid) intake to avoid the excess risk associated with even moderate amounts of alcohol.
Replace saturated fat with olive oil.
Willett does not mention milk, though high-fat dairy products are implicitly rejected in favour of olive oil as they are a major source of saturated fat. Willett does not suggest a direct adverse role for saturated fat in breast cancer, but rather a benefit to health from replacing it with olive oil, a particularly healthful source of monounsaturated fat. The recommendation of olive oil as opposed to rapeseed/canola oil or other oils high in monounsaturated fats is deliberate: there is specific evidence for a benefit from olive oil over and above its major constituent fats. The recommendation to substitute olive oil for saturated fat appears to be based on evidence for positive benefits of olive oil in relation to breast cancer and the fact that substituting olive oil for saturated fat will improve blood cholesterol levels and thus reduce risk of heart disease.
Willett does not recommend completely eliminating alcohol, as moderate intakes of alcohol appear to have beneficial effects on heart disease. He notes that the adverse effects of alcohol in promoting breast cancer appear to be completely eliminated by high folate intake, allowing the benefit to be obtained without the damage. Folate is particularly well supplied by green leafy vegetables but is present in many other plant foods as well. In supplements it usually takes the form of folic acid. Some of the effects of folate are dependent on the presence of adequate vitamin B12, so vegans should ensure an adequate B12 intake (about 3 micrograms per day) to ensure that they get the full benefit of folate.
Willett takes a sceptical stance towards soy products due to recent evidence that they can stimulate proliferation of breast tissue, potentially increasing the risk of breast cancer. The overall effect of increased soy intake is more likely to be adverse in postmenopausal women than in premenopausal women due to interactions with oestrogen. He therefore suggests that "soy products should be used in moderation if started during midlife, perhaps no more than a few times a week, until further data are available". Typical Japanese and Chinese consumption is about 10g of soy protein (about 300ml (half a pint) of soya milk or 100g (4oz) of tofu) per day, consumed with apparent lack of adverse effects. Indeed, both these populations show low levels of breast cancer. However, their consumption of soy is from an early age and this is likely to be relevant. Overall it seems prudent not to exceed Japanese levels of consumption significantly, particularly if soy consumption is started late in life.
One of the more surprising observations is the comment that in the largest prospective study on dietary fat and breast cancer to date (pooling results from multiple studies) women consuming less than 15% fat showed double the risk of breast cancer . No details of the particular diet characteristics of this group are given, but this observation adds to reasons for caution about very low fat diets.
A review paper by Timothy Key and Naomi Allen  provides an alternative summary of risk factors for breast cancer, concluding that obesity in postmenopausal women and alcohol consumption are the only well established diet-related risk factors for breast cancer. They conclude, "Current dietary advice should be to avoid obesity, limit alcohol intake and to maintain a varied diet."
Willett makes some more general observations on diet and cancer in . He emphasises the need to make choices that are beneficial for both cancer and heart disease - we can only eat one diet.
Willett does not regard the evidence against meat as clear enough to make a strong recommendation regarding breast cancer. However, he notes that the evidence against red meat in relation to colorectal cancer is much stronger.
While the evidence for fibre and whole grains in relation to cancer is weak, Willett comments that there is stronger evidence in relation to heart disease, diabetes and diverticular disease and this justifies a recommendation that increased cereal fibre consumption is likely to be beneficial for health.
Willett also emphasises that physical activity reduces risk of colorectal and breast cancer directly, as well as by reducing weight.
Finally, Willett estimates that remaining lean (BMI < 25), taking at least 30 minutes a day of moderate physical exercise, avoiding excessive alcohol consumption, getting plenty of folate, not smoking, and consuming red meat less than three times a week could reduce colorectal cancer by 70%.
There is a lot we can do to take control of our health for the better, including reducing risk of breast cancer. Green leafy vegetables, olive oil and physical activity can all be expected to be beneficial. Drinking cow's milk doesn't appear on the list.