By Chris Masterjohn
In the December, 2006 issue of his newsletter1, Loren Cordain, PhD, Professor of Health and Exercise Science at Colorado State University and author of The Paleo Diet, added a new item to his list of the many dangers of consuming dairy: betacellulin. Betacellulin is a growth factor found in the whey fraction of milk that appears to play important roles in the growth and differentiation of tissues in the fetus and newborn. 2,3
The dark side of this essential molecule, according to Cordain, is that it passes into the adult digestive tract intact, where it is taken up by receptors and can then enter circulation and stimulate the growth of cancer cells throughout the body. In support of this theory, he cites 25 studies purportedly linking increased rates of cancer to milk consumption.
Milk, however, contains a wide array of vitamins, minerals, fatty acids and proteins, many of which have anti-carcinogenic activity. As with any other food, the effect of milk on cancer is determined not by any single molecule within it, but by the concerted effect of all its chemical components acting together. The big question, then, is not whether isolated betacellulin can stimulate the growth of cultured tumor cells, but whether milk, particularly whole, raw milk from grass-fed cows consumed in traditional forms, is likely to accelerate the growth of cancer in the people who drink it, or instead to nourish them and protect them against cancer. The epidemiological evidence by and large fails to make a strong case against even commercial milk, especially commercial whole milk. The small amount of existing research on pasture-fed milk suggests that it may, in fact, provide a powerful antidote to a wide variety of cancers.
Betacellulin has been isolated from colostrum, milk and cheese whey,2, 3 but whether other forms of fermentation, such as those used in the production of yogurt or kefir, break it down remains to be tested. Although researchers have not determined how much of the betacellulin in milk will actually survive digestion in the adult gastrointestinal tract, laboratory experiments show that milk contains substances that will inhibit the ability of isolated digestive juices to break down other growth factors,4 so it is possible that much or all of the betacellulin that we consume is biologically active.
Although there are many studies showing hydrolyzed whey protein to protect against cancer in experimental animals, the few studies utilizing non-hydrolyzed whey protein that would contain intact betacellulin have shown conflicting results.5 None, however, has shown whey protein to increase the rate of cancer, despite the presence of betacellulin within it. This suggests either that dietary betacellulin does not promote cancer or other protective factors, such as whey proteins, cancel out its effect.
Conjugated linoleic acid (CLA), found in milk fat, is a powerful anti-carcinogen. Part of its mechanism of action is to decrease the transcription and activation of receptors in the epidermal growth factor (EGF) family. 6,7 These are the very same receptors through which Dr. Cordain hypothesizes betacellulin causes cancer. Even if betacellulin in commercial milk does promote cancer, it may only be able to do so in the absence of the protective factors found in the fat of naturally and traditionally produced milk.
When human breast and colon cancer cells are bathed in high-CLA milk fat from cows raised on pasture, the milk fat decreases the number of cancer cells between 58 and 90 percent.8,9 Unfortunately, such experiments have not utilized whole milk, which would contain betacellulin in addition to CLA, so we cannot conclusively define the type of balance needed between these two components. Grass-fed milk is five times higher in CLA than grain-fed milk,10 and the anti-carcinogenic effect of CLA in animals is enhanced when the experimental diet is high in saturated fat—the kind found in whole milk—and diminished when it is high in unsaturated fat from vegetable oil.11 We can conclude from this finding that insofar as milk is produced commercially and stripped of its natural fat, it is more likely to promote cancer, and that insofar as milk is produced traditionally and retains its natural fat, it is more likely to protect us from cancer.
Epidemiological studies, which observe correlations between variables, cannot be used to prove that one variable causes another, but they can be used to generate new ideas or to evaluate the plausibility of a hypothesis. In this case, if betacellulin contributes to cancer, we would expect milk to be correlated with cancer regardless of its fat content; if CLA is protective, however, we would expect whole milk to carry a lower risk than low-fat milk. On the other hand, such results could be confounded by other variables, especially in populations that consider low-fat milk to be a healthier choice than whole milk, in which case low-fat milk could simply be an indicator of a healthier lifestyle and whole milk an indicator of a less healthy lifestyle.
Of the several types of epidemiological studies, prospective cohort studies are the most reliable because they record dietary data before the diseases they are looking for develop; case-control studies are less reliable because they are retrospective and therefore subject to recall bias; ecological studies are the least reliable because they examine variables that act on individuals, such as milk-drinking, at the level of the population rather than the individual. Meta-analyses pool the results of a large number of published studies together and can be a useful tool for examining the totality of the epidemiological evidence.
Cordain cites 25 studies purportedly associating milk with ovarian, breast, colon, lung, stomach, pancreatic and prostate cancers. Only the results of studies with ovarian cancer, however, are consistent with what we would expect from this hypothesis. The totality of the evidence suggests that whole milk either carries no risk for cancer or may even be protective.
Cordain cites four studies associating milk with ovarian cancer. One is a case-control study,12 two are prospective cohort studies,13, 14 and the last is a meta-analysis.15 In the case-control study, the consumption of whole milk, but not low-fat or skim milk, was associated with cancer. By contrast, in the more reliable prospective cohort studies, only low-fat and skim milk, but not whole milk, were associated with an increased risk of cancer. In one cohort study,13 hard cheese was associated with a reduced risk of cancer.
The meta-analysis showed that the results of case-control studies were conflicting, while the results of prospective cohort studies were consistent. The pooled results of case-control studies showed a decreased risk with low-fat milk and an increased risk with whole milk. By contrast, the pooled results of the more reliable cohort studies showed a moderate increase in risk for low-fat milk and no association at all with whole milk. This finding is consistent with the hypothesis that betacellulin contributes to cancer, while the CLA in milk fat cancels out this effect.
Cordain cites seven papers connecting milk to the incidence of breast cancer, but each of them fails to demonstrate any conclusive association. One is a hypothesis paper merely observing that the post-World War II rise in the consumption of animal-based foods in Japan—not dairy in particular—might bear a causal connection to the concomitant rise in breast and ovarian cancer.16 The second is an ecological international comparison showing that the countries with higher per capita milk consumption had higher rates of breast cancer between 1987 and 1994, but not between 1964 and 1986.17 The third is a prospective cohort study that did not even find a statistically significant relationship between dairy and breast cancer at all.18
The first of four case-control studies cited by Cordain found that the consumption of whole milk during childhood was associated with a reduced risk of breast cancer later in life; although it also observed an increased risk of postmenopausal breast cancer associated with the consumption of whole milk during adulthood, this association disappeared after adjusting for confounding factors. The study did not examine the effect of low-fat milk.19 Although total milk was associated with an increased risk in the second study,20 only the final two distinguished between low-fat milk and whole milk. Both of these found an increased risk with whole milk only, and not with low-fat milk. 21,22
If betacellulin, which is contained in the whey fraction, were contributing to the increased risk of breast cancer, we would expect to see at least as great an increased risk with the consumption of low-fat milk as whole milk, so these results do not support Cordain’s hypothesis. Although the last two studies may seem on the surface to suggest that milk fat may cause breast cancer, they observed no association with butter,22 nor with the total amount of milk fat consumed from all sources.21 The studies, therefore, provide neither a convincing association with milk itself nor with milk fat.
Of the four papers Cordain cites associating milk with colon cancer, none shows a convincing correlation. One study merely observes that between 1950 and 1970, Japanese consumption of milk, meat, eggs, fats and oils increased while consumption of rice and potatoes decreased, and then proceeds to analyze the association between colorectal cancer and age, gender, geography and ethnicity; it provides no analysis of the relationship between colorectal cancer and milk consumption at all.23 A second report found a relationship between “bread, milk, butter, margarine, ketchup and beer” and colon cancer when these six foods were analyzed together as a single group, but found no actual relationship with milk itself.24 A third was a case-control study that found an association with whole milk only and not with low-fat milk, yet found no association with total consumption of milk fat, thus incriminating neither the protein fraction nor the fat fraction of milk.21
Finally, although the fourth study found that the regions of Italy that had higher per capita milk consumptions also had higher rates of colon cancer, milk consumption was much more strongly associated with latitude, which itself was the primary predictor of all cancers analyzed.25 Had it been known 20 years ago when this paper was published that vitamin D is a powerful inhibitor of colon cancer,26 the authors would likely have adjusted the results for latitude and concluded that milk consumption was associated with a decreased risk of colon cancer, which was the conclusion of a meta-analysis of ten prospective cohort studies published in 2004.27 A more recent prospective cohort study showed that the consumption of four or more servings of high-fat dairy foods per day was associated with a 41 percent reduced risk of colorectal cancer, while low-fat dairy had no effect.28
Cordain cites four case-control studies associating milk consumption with the risk of lung cancer. Two of these studies found an increased risk associated with total milk consumption,29,30 but the effect was only statistically significant in one of them.30 In the two studies that distinguished between low-fat and full-fat milk, however, only whole milk was associated with increased risk; in fact, low-fat milk was associated with a decreased risk.21,31
Since there is no less betacellulin in low-fat milk than in full-fat milk, these results certainly do not support Cordain’s hypothesis. Although they appear on the surface to suggest that milk fat contributes to cancer, the only one of these studies that analyzed the risk associated with the consumption of increasing amounts of milk fat found that higher intakes of milk fat were associated with a reduced risk, although this effect was not statistically significant.21 Perhaps the consumption of low-fat milk acts as a marker for “healthy” behavior (such as avoidance of sugar or cigarettes) while that of whole milk acts as a marker for general disregard for health advice; or perhaps those who are diagnosed with cancer are more likely to recall drinking more whole milk in the past if they perceive doing so to be unhealthy; regardless, the results do not support a role for either milk protein or milk fat in contributing to lung cancer.
Neither of the two studies Cordain cites associating milk with stomach cancer is convincing. The first is a case-control study that found that drinking whole milk was associated with increased risk, while drinking low-fat milk was associated with decreased risk. As for other cancers, however, there was no relationship between stomach cancer and increasing intake of milk fat per se.21 The second is a more reliable prospective cohort study that found no statistically significant relationship with milk at all; in fact, those who consumed two to three glasses of milk per day had a slightly lower risk of stomach cancer than those that consumed less than one glass per day.32
Cordain cites three ecological studies associating milk with pancreatic cancer. Two of them are the same international analyses conducted by the same authors but published in two different journals, showing that countries that have higher per capita rates of milk consumption have higher rates of pancreatic cancer,33,34 but lower rates of esophageal and stomach cancers.34 The third showed that provinces of Spain with higher rates of milk consumption had higher rates of pancreatic cancer mortality, although milk consumption could only account for about 15 percent of the variation.35 The authors noted that pancreatic cancer mortality rates were higher in the northern provinces than in the southern provinces, but they did not adjust their results for latitude. None of these studies distinguished between low-fat and full-fat milk.
Although these studies could be interpreted to suggest an association between milk and pancreatic cancer, the ecological study design is the least reliable design for evaluating the effects of nutrients and foods, which act on individuals rather than populations. There is therefore very little that can be concluded from these studies.
Cordain cites four studies linking milk to prostate cancer. Rather than lending support to the betacellulin hypothesis, however, these studies lend very convincing support to the hypothesis that calcium, rather than milk protein, contributes to prostate cancer (see sidebar on page 78). They also show that vitamin D and possibly other components of milk fat are protective; for this reason, only the consumption of low-fat milk is associated with prostate cancer. Whole milk has no relationship.
The results of prospective cohort studies for ovarian and prostate cancer both support a model wherein milk protein may contribute to cancer when it is detached from the protective fat; the evidence that calcium, rather than betacellulin, is the active component for prostate cancer, however, is strong. By contrast, epidemiological studies do not convincingly demonstrate a relationship between milk and cancers of the breast, lung, stomach, and pancreas, and appear to show that milk, especially full-fat milk, protects against colon cancer.
Unfortunately, all of these studies have evaluated the effect of commercial milk alone. No epidemiological studies have distinguished between grain-fed and pasture-fed dairy, nor pasteurized milk and raw. Animal and cellular experiments suggest that the CLA in grass-fed, whole milk would powerfully protect us against cancer, especially in conjunction with the rich array of vitamins, minerals and raw whey proteins naturally present in real milk. Given that the epidemiological evidence fails to convincingly indict even commercial whole milk in the causation of any cancer and even suggests it may be protective against some cancers, we can be confident that real milk consumed in its natural state is both nourishing and safe.
A meta-analysis of ten prospective cohort studies published between 1966 and 1998 showed a very small increased risk of prostate cancer associated with milk consumption. The association was even greater, however, especially for advanced prostate cancer, with total calcium intake.36
After 1998, researchers began investigating the hypothesis that calcium contributes to prostate cancer by suppressing the formation of calcitriol, the activated hormone form of vitamin D. Since calcitriol stimulates calcium absorption in the intestines, the kidneys produce less of it when the body gets enough calcium. Because calcitriol has anti-carcinogenic effects on prostate cells, reducing the amount of calcitriol in circulation could increase the risk of prostate cancer.
This hypothesis has gained substantial support from several prospective studies. A recent Harvard study showed that total calcium intake had a clear and strong relationship to the risk of prostate cancer, and that this was associated with reduced levels of activated calcitriol. Dairy intake itself, by contrast, had no relationship, although the researchers did not distinguish between low-fat and full-fat milk.37 Fortunately, dietary intake of vitamin D appears to protect against the calcium-mediated increase in prostate cancer risk.38 This may provide one reason why studies that distinguish between low-fat and full-fat milk show that low-fat milk and skim milk—and especially calcium from these sources—are associated with an increased risk, while whole milk or calcium from whole milk either have no relationship39 or have even a slight, though not significant, inverse association.38
Rather than incriminating milk, these studies simply remind us of the importance of consuming milk in its natural, unprocessed state—the way nature intended it.
This article appeared in the Spring 2007 edition of Wise Traditions, the quarterly journal of the Weston A. Price Foundation.