Lest this blog sink further into its eery two-month silence, I think it’s high time for an update!
First item of business: The Ancestral Health Symposium. Due to some serendipitous events, it turns out I’ll be presenting at this hyperventilation-inducingly-awesome event next week. My lecture is at 10:00 AM on August 6th in the Rolfe 1200 auditorium. If you’re lucky enough to have a ticket, I hope to see you there, and to verify my existence for anyone who still thinks I’m a meat industry puppet. Otherwise, unless PETA pops in and sets fire to UCLA, all the presentations should be available online for free shortly after the symposium is over. Woohoo!
Second item of business: Now that he’s outed the project himself, I feel safe in announcing that Mark Sisson is going to be publishing the book I mentioned working on in an earlier blog post, and that it’ll be released mid-2012. I’m super excited, and couldn’t ask for a better publisher to work with. Or one with more impressive abs (see link above). More details to come in the near future.
Now on to the real point of this post.
One year ago, this happened:
Those are the monthly visits to this blog, according to my top-secret WordPress dashboard. And that crazy jump last July is when I posted my critique of the China Study, which is probably the only reason most of you know this blog exists. Thanks to Richard Nikoley’s smackdown roundup, the critique got passed along by a lot of cyber-hands, eventually reaching Campbell himself.
With the release of the movie Forks over Knives, Campbell’s recent appearance on the Bill Maher show, and continued Wikipedia drama about the peculiar lack of criticism on its “The China Study” page, it seems the China Study is back in the spotlight for awhile.
Since I’m not a vegan anymore, I figure it’s okay for me to beat dead horses. And also to resurrect the ones I buried last year and wallop on their half-rotten, fly-infested carcasses with my fists a few more times. So wallop I will: This post is dedicated to driving a couple more nails into the China Study coffin—and is aimed particularly at the folks out there who would rather listen to peer-reviewed research than some girl with a blog.
So my dear readers, hecklers, and spambots, I present to you a collection of peer-reviewed papers based on the China Study data that contradict or conflict with Campbell’s interpretation in his book, “The China Study.” Some studies you may have seen before; others will be new. Regardless, you can rest assured that these papers—some co-authored by Campbell himself—are by folks generally considered qualified in their field, and that, contrary to the “animal foods are harmful and cholesterol is associated with all Western diseases” message we received in “The China Study,”* other perspectives of the data abound.
*It’s also worth noting that”The China Study”—the one written by Campbell and published by BenBella Books—is not peer reviewed. Shockingly, neither are BenBella’s other books, including “You Don’t Talk About Fight Club,” “Seven Seasons of Buffy,” and “The Psychology of Harry Potter.” Contrary to what’s apparently popular belief, books—even health books—don’t have to pass under the scrutiny of peer review before they hit the stands. Hence why this exists.
But first, a few words on peer review
I want to burst the Peer-Review Bubble of Perfection before we get much further.
Peer review might be the best we’ve got right now—but it’s far from infallible, and its biases are no secret. In an article from 2000, Richard Horton—editor of the uber-peer-reviewed journal “The Lancet”—wrote some rather scathing comments about the peer-review system, stating that it is “biased, unjust, unaccountable, incomplete, easily fixed, often insulting, usually ignorant, occasionally foolish, and frequently wrong.” Even peer-reviewed journals have published papers on the problems with peer review (e.g., this article in “Nature”).
To top that off, history is speckled with some disturbing cases of research fraud that slipped right through the peer-review system. The best example is Scott Reuben, once considered a pioneer in the field of anesthesiology and pain management, who concocted at least 21 “studies” that were pure works of fiction—and managed to get all of them published in peer-reviewed journals. Over the years, he accepted big bucks from pharmaceutical companies to conduct studies on drugs like Celebrex and Effexor, but instead of actually enrolling patients, he made up some numbers and slid his nonexistent findings into major publications. And as it turns out, many of the drugs he convinced the world were beneficial were either ineffective of downright harmful. (You can see a list of his falsified peer-reviewed papers here.)
That’s an extreme example, but it illustrates the point. Even peer-reviewed papers should be taken with a grain of salt instead of held as gospel.
And with that, here are some papers to mull over. The bolded parts within quotes are to highlight the relevant bits.
Erythrocyte fatty acids, plasma lipids, and cardiovascular disease in rural China by Fan Wenxun, Robert Parker, Banoo Parpia, Qu Yinsheng, Patricia Cassano, Michael Crawford, Julius Leyton, Jean Tian, Li Junyao, Chen Junshi, and T. Colin Campbell.
A study involving fat, cholesterol, and cardiovascular disease? Surely they must be talking about how all that nasty saturated fat in animal foods clogs your arteries, right? Oh, snap:
Within China neither plasma total cholesterol nor LDL cholesterol was associated with CVD [cardiovascular disease]. … The results indicate that geographical differences in CVD mortality within China are caused primarily by factors other than dietary or plasma cholesterol.
There were no significant correlations between the various cholesterol fractions and the three mortality rates [coronary heart disease, hypertensive heart disease, and stroke]. In contrast, plasma triglyceride had a significant positive association with CHD and HHD but not with stroke.
We’ve even got a cameo appearance from wheat again:
The consumption of wheat flour and salt (the latter measured by a computed index of salt intake and urinary sodium excretion) was positively correlated with all three diseases [cardiovascular disease, hypertensive heart disease, and stroke].
And for those of your leery of industrial oils and polyunsaturated fats, check this out:
Unlike what might be expected from studies on Western subjects, there was no significant inverse correlation between RBC-PC total PUFAs and CVD mortality; in fact, RBC-PC total PUFAs, especially the n-6 fatty acids, were positively correlated with CHD [coronary heart disease] and HHD [hypertensive heart disease].
That one was a little acronym-y, but basically it says that higher levels of polyunsaturated fats (especially omega-6 fats) in red blood cells was associated with more heart disease.
So if you don’t want to take my word for it, take the word of this peer-reviewed paper: The China Study data showed no correlation between cholesterol and heart disease, but did find wheat and polyunsaturated fats to be mighty suspect.
Association of dietary factors and selected plasma variables with sex hormone-binding globulin in rural Chinese women (PDF) by Jeffrey R. Gates, Banoo Parpia, T. Colin Campbell, and Chen Junshi.
Wheat-fearers, you’ll enjoy this one.
This study focuses on sex hormone-binding globulin (SHBG), a molecule sometimes used to test for insulin resistance. (Higher levels are associated with better insulin sensitivity; lower levels are associated with insulin resistance and diabetes.) After fishing out associations between SHBG, fasting insulin, triglycerides, testosterone, and a number of diet and lifestyle variables, the researchers found:
The principal positive food-SHBG correlates in order of magnitude were rice (0.61, P < 0.0001), green vegetables (0.49, P < 0.001), fish (0.42, P < 0.001), and meat (0.38, P < 0.05). The strongest negative food correlate with SHBG (positively correlated with insulin) was wheat (-0.57, P < 0.0001).
In other words, the foods associated with higher SHBG (and lower insulin) were rice, green veggies, fish, and meat. The main food associated with lower SHBG (and higher insulin) was… dun dun dun… wheat. Not only do we have vindication of some animal foods, but we also have another red flag whipping up over our favorite glutenous grain. Although the link with meat diminished in more sophisticated statistical models, the other foods kept their associations with SHGB—and wheat proved to be the strongest predictor of low SHBG, while rice was the strongest predictor of higher SHBG. In discussing their findings, the researchers note that wheat seemed to accompany a number of markers for poor health:
Significant differences in the diet of rural Chinese populations studied suggest that wheat consumption may promote higher insulin, higher triacylglycerol, and lower SHBG values. Such a profile is consistent with that commonly associated with obesity, dyslipidemia, diabetes, hypertension, and heart disease. On the other hand, the intake of rice, fish, and possibly green vegetables may elevate SHBG concentrations independent of weight or smoking habits.
It looks like the post I wrote on wheat and heart disease in the China Study was old news: Campbell and his colleagues already spotted the link back in 1996! But why would wheat have such a vastly different effect than rice? The paper offers a possible explanation:
The effect of rice and wheat on SHBG was remarkable and unexpected. … Nevertheless, there is some evidence to suggest that rice and wheat can have significantly different effects on the biochemical variables we measured. Panlasigui et al (58) found that the high-amylose rice varieties had blood glucose responses that were lower than those of wheat bread. Other varieties, particularly “converted” rice, gave considerably higher values. Miller et al (59) in comparing rice and wheat varieties found that the insulin index (II) was unusually low on the relative scale compared with the glycemic index (GI) of the same foods. For example, Calrose brown rice had a GI = 83 but an II = 51. White bread was used as the reference food (GI = 100, II = 100). Wheat may be unique in its relative capacity to stimulate insulin. Most recently, Behall and Howe (60) reported a significantly lower insulin response curve area in both normal and hyperinsulinemic men consuming a high-amylose diet. The relative differences in the fatty acid proportions and/or amylose content for wheat and rice may thus be responsible for modulating serum SHBG, triacylglycerols, and insulin.
Although it’s still speculative, the amount of amylose (a component of starch) and relative proportion of fatty acids in wheat might make it more problematic than other grains like rice—especially in terms of raising triglycerides and fasting insulin while lowering SHBG. Which is particularly interesting in the context of this paper, because in one of his responses to my critique, Campbell stated:
[The] correlation of wheat flour and heart disease is interesting but I am not aware of any prior and biologically plausible and convincing evidence to support an hypothesis that wheat causes these diseases, as you infer.
Prolonged infection with hepatitis B virus and association between low blood cholesterol concentration and liver cancer (PDF) by Zhengming Chen, Anthony Keech, Rory Collins, Brenda Slavin, Junshi Chen, T. Colin Campbell, and Richard Peto.
This one’s a doozy. But first, some background info to help us understand the full extent of its doozydom.
One of the prevailing themes in “The China Study” is a supposed link between cancer and animal protein (and subsequently, blood cholesterol). Campbell first started chasing this association after finding that casein, a milk protein, raised cholesterol and promoted liver cancer growth in rats exposed to aflatoxin. After reviewing the China Study data, Campbell concluded this link held true in humans: He states that animal protein, as reflected by higher cholesterol levels, promoted liver cancer in folks already at risk for it (namely those who carried the hepatitis B virus). Straight from the book:
In addition to the [hepatitis B] virus being a cause of liver cancer in China, it seems that diet also plays a key role. How do we know? The blood cholesterol levels provided the main clue. Liver cancer is strongly associated with increasing blood cholesterol, and we already know that animal-based foods are responsible for increases in cholesterol. … Individuals who are chronically infected with HBV and who consume animal-based foods have high blood cholesterol and a high rate of liver cancer. The virus provides the gun, and bad nutrition pulls the trigger. (Page 104)
People chronically infected with hepatitis B virus also had an increased risk of liver cancer. But our findings suggested those who were infected with the virus and who were simultaneously eating more animal-based foods had higher cholesterol levels and more liver cancer than those infected with the virus and not consuming animal-based foods. (Page 105)
Seems clear enough. According to Campbell, the data showed that liver cancer went hand-in-hand with high cholesterol in the China Study data.
…Which is what makes this particular study (co-authored by Campbell, nonetheless) so interesting. Campbell et al. set out to investigate “the association between low blood cholesterol concentration and liver disease by studying blood lipid concentrations about middle aged men in rural China.” Already seems fishy, huh? Before even discussing the study results, the paper includes some sections that contradict the notion that high cholesterol is linked with liver cancer:
Several prospective epidemiological studies … have found an inverse relation between cholesterol concentration and the subsequent risk of cancer. … A prospective observational study in a Chinese population … found a significant inverse association between blood concentration of cholesterol and subsequent mortality from non-malignant liver disease or from liver cancer. More recently significant excess risk of death from liver cancer and chronic liver disease has been reported among North Americans with a low blood cholesterol concentration.
In the largest study in a Western population (the multiple risk factor intervention trial) 100 deaths from liver cancer were recorded during the follow up period, and a significantly increased risk of death from liver cancer was found among people in the group with the lowest cholesterol concentrations.
In our previous prospective study of another Chinese population the subsequent risk of death from liver cancer was shown to increase significantly with decreasing blood concentrations of cholesterol.
Whether low cholesterol is a cause or a consequence of cancer is a different story—but either way, there’s no mistaking it: Liver cancer looks pretty solidly linked with low cholesterol in other relevant studies. And this study does nothing to refute that:
We have now shown that prolonged infection with hepatitis B virus is an additional factor contributing to the inverse relation between cholesterol concentration and liver cancer. Chronic hepatitis B, which usually starts in early childhood in China, leads not only to liver disease but also to a lower blood concentration of cholesterol in adulthood. This produces, as observed elsewhere, an inverse relation between cholesterol concentration and the risk of death from liver cancer or from other chronic liver disease. This result may also help to explain, at least in part, the inverse association between cholesterol concentration and liver disease observed in Western populations.
So why did Campbell repeatedly state in “The China Study” that liver cancer was associated with higher cholesterol? Probably because it was—but only after the (more reliable) individual data was aggregated at the county level, which made it easy to succumb to a little somethin’ called the “ecological fallacy.”
Let me explain. The publicly available China Study data—the stuff I used for my critique, and that Campbell pulled correlations from in his book—consists of averaged values for 65 counties, instead of the thousands of data points originally collected. But this particular study used the individual data, before any of it was combined and diluted by averaging. And as the paper explains, that makes this study much more reliable than the ones using aggregated data:
In this study there was a negative correlation between chronic infection with hepatitis B virus and blood concentrations of cholesterol (and apolipoprotein B) when people living in the same village were compared with each other, but the correlation was reversed when average values for different villages were compared with each other.
Correlations between populations based on average measures in groups are subject to the “ecological fallacy” (whereby these correlations may not represent the correlations that would be seen among individual subjects). … In general, comparisons within populations are much more reliable than comparisons between populations when assessing association of variables and diseases in individual subjects. So, in the present instance, the negative correlation observed when people living in the same village were compared with each other provides the most reliable evidence as to the real relation between chronic infection with hepatitis B virus and lipid concentrations in individual subjects.
This is kind of fascinating. Here we’ve got a China-Study-based paper—again, co-authored by Campbell—that explicitly describes why aggregated data can be unreliable, and why positive links between liver cancer variables and cholesterol are probably backwards. This also begs the question of how many other associations in the aggregated data would be reversed at the individual level. (Or whether any seemingly neutral relationships are actually correlated—such as liver cancer and the carcinogen aflatoxin, which are paradoxically unassociated in the aggregated data.)
Here’s an example to help illustrate the concept of “ecological fallacy” as it relates to liver cancer in China.
Say we’ve got two counties, each with 1000 residents. Folks in the first county have total cholesterol ranging from 130 to 150, with the average being 140. These lucky ducks are free from liver cancer or hepatitis B infection! The second county has total cholesterol ranging from 120 to 190, with the average being 170. Much more liver cancer and hepatitis B infection in this place, but the afflicted folks all have cholesterol at the bottom end of the spectrum, around 120.
What happens if we look only at the aggregated data? We’d see that the county with the lower average cholesterol (140) had lower rates of liver cancer, and the county with higher average cholesterol (170) had higher rates of liver cancer. Thus, a liver cancer/higher cholesterol relationship is born. But what happens if we look at the individual data instead? We’d see that the people with liver cancer had lower cholesterol than any of the healthy folks—the exact opposite of what the averaged data showed.
See how tricky numbers can be?
At any rate, this (peer-reviewed!) study blatantly contradicts some of the claims made in “The China Study,” especially the concept that cancer goes hand-in-hand with high cholesterol.
Dietary calcium and bone density among middle-aged and elderly women in China (PDF) by Ji-Fan Hu, Xi-He Zhao, Jian-Bin Jia, Banoo Parpia, and T. Colin Campbell.
True to its title, this paper examines the role of calcium in bone density in the China Study data—with a special focus on the effects of dairy calcium versus plant calcium. Campbell et al. zoomed in on five counties with “distinct lifestyles and diets”: the dairy-and-meat-loving Xianghuangqi, the infamous dairy-full Tuoli, and the rural, nearly-vegan farming towns of Jiexiu, Cangxi, and Changle.
But before we look at the paper itself, let’s see how Campbell summarized its findings in a Cornell Chronicle article in 1994:
Animal protein, including that from dairy products, may leach more calcium from the bones than is ingested, said Campbell, professor of nutritional biochemistry at Cornell and director of the Cornell-China-Oxford Project, the most comprehensive project on diet and disease ever conducted.
Campbell [and other collaborators] analyzed the role of dietary calcium in bone density by following closely the diets of 800 women from five counties that have very different diets in China. … Analyses of these data suggest that increased levels of animal-based proteins, including protein from dairy products, “almost certainly contribute to a significant loss of bone calcium while vegetable-based diets clearly protect against bone loss,” Campbell reported.
Sounds pretty clear: The dairy-eating counties must have had poor bone health due to their animal protein habit, whereas the more plant-based dieters were skeletally superb. In other words, milk does a body bad! But do the summaries above match up with this paper actually found? First, let’s look at what the women in each county were typically eating:
*Lest I get the “you’re trying to justify your dairy addiction” line and/or accusations of dairy industry affiliation, I’d like to remind everyone that dairy hasn’t been part of my diet in over six years, and I believe the dairy most people consume (low-fat, ultra-pasteurized, etc.) is downright nasty stuff. But that doesn’t mean I won’t defend dairy when the science warrants it.
As you can see, Xianghuangqi ate a pretty shabby diet as far as whole-foods veganism is concerned: We’ve got dairy galore, beef, mutton, wheat flour, a mere smattering vegetables, and millet. Their bones should be snapping like peanut brittle! Tuoli’s not much better, what with their milk tea, animal flesh, and decided lack of green leafy veggies. More bone snappage, right?
I’ll let the paper speak for itself:
Analysis by individual for all counties combined showed that [bone mineral content] and [bone mineral density] were correlated positively with total calcium (r = 0.27-0.38, P < 0.0001), dairy calcium (r = 0.34-0.40, P < 0.0001), and to a lesser extent with nondairy calcium (r = 0.06-0.12. P = 0.001-0.100), even after age and/or body weight were adjusted for. The results strongly indicated that dietary calcium, especially from dairy sources, increased bone mass in middle-aged and elderly women by facilitating optimal peak bone mass earlier in life.
Did you catch that? Dairy calcium—far more than plant calcium—was linked with stronger bones. Moreover, the paper notes that “nondairy calcium … showed no association with bone variables after age and/or body weight were adjusted for.”
Comparison of results in Table 7 reveal that calcium from dairy sources was correlated with bone variables to a higher degree than was calcium from the nondairy sources, probably resulting from the higher bioavailability of dairy calcium.
A comparison of the bone mass of women in the five counties revealed that 20% greater bone mass at the distal radius was observed for all age groups of women in county YA [Xianghuangqi], a pastoral county with high consumption of dairy foods, as compared with the nonpastoral areas with lower calcium intakes.
I’ll add my own unsolicited 2¢ and speculate that calcium probably wasn’t the only protective factor in the dairy-eating counties. Aged cheese, likely consumed at least in Xianghuangqi, is high in vitamin K2—a nutritional superstar when it comes to bone health (among other things). K2 isn’t present in plant foods except for a fermented soy product called natto (not everyone’s cup o’ tea). As the paper notes, the dairy-eating counties also had a higher intake of fat (25% of daily calories, opposed to 9.9 – 13.6% for the other counties), potentially increasing the absorption of fat-soluble vitamins necessary for bone health.
So how did Campbell conclude from this study that “increased levels of animal-based proteins, including protein from dairy products, almost certainly contribute to a significant loss of bone calcium”? The dairy part is unfounded no matter which way you spin it, but the rest of his statement probably stemmed from this:
The associations between bone mass and other nutrients, like dietary protein and phosphorous, were also examined. However, none of these nutrients showed an association with bone mass as significantly as did dietary calcium, although an inverse correlation was observed consistently for nondairy animal protein.
Unfortunately, that’s the only blurb in the entire paper that mentions animal protein in relation to bone mass, so we can’t see the data behind the “consistent inverse correlation.” In the context of this study, though, it makes sense: Protein has a complex relationship with bone formation, serving as a synergist when calcium intake is adequate, but as a potential antagonist when calcium intake is low. In other words, the effects of protein on bone health depend on how much calcium you’re taking in.
So for the counties in this study that ate more animal protein but sparse calcium—such as Changle, which had the highest non-dairy animal food consumption and also the lowest calcium intake (averaging a mere 230 mg per day)—I wouldn’t be surprised if an animal protein/weaker bones connection showed up. Whether that trend would hold at higher calcium intakes is a different story. And either way, this finding doesn’t jive with most other research done on this topic: Most studies show a protective association between animal protein and bone density, formation, and retention:
- Prospective study of dietary protein intake and risk of hip fracture in postmenopausal women. “Protein from animal sources was the nutrient variable with the strongest negative association with risk of hip fracture in this prospective study of Iowa women. Protein from vegetable sources did not appear to protect against hip fractures.”
- Effect of Dietary Protein on Bone Loss in Elderly Men and Women: The Framingham Osteoporosis Study. “Contrary to expectations, elders with animal protein intake up to several-fold greater than the RDA also had the least bone loss after controlling for known confounders. Nonanimal sources of protein were not related to BMD. These results suggest that typical population intakes of animal protein, within the range commonly consumed, do not result in bone loss. Rather animal protein intake appears important in maintaining bone or minimizing bone loss in elderly persons.”
- Protein Consumption and Bone Mineral Density in the Elderly. “Multiple linear regression analyses … showed a positive association between animal protein consumption … and BMD. Conversely, a negative association between vegetable protein and BMD was observed in both sexes. … This study supports a protective role for dietary animal protein in the skeletal health of elderly women.”
- Controlled High Meat Diets Do Not Affect Calcium Retention or Indices of Bone Status in Healthy Postmenopausal Women. “Calcium retention is not reduced when subjects consume a high protein diet from common dietary sources such as meat.”
In addition, if animal protein was such a bone-killer and plant protein was bone protective, we’d see vegetarians or vegans having the best outcomes in the bone department. But this just ain’t the case. At best, non-meat-eaters are equally matched with their omnivorous counterparts; at worst, they’re more prone to fracture:
- Veganism and osteoporosis: A review of the current literature. “The findings gathered consistently support the hypothesis that vegans do have lower bone mineral density than their non-vegan counterparts.”
- A Comparison of Bone Mass Measurements of Vegetarians and Omnivores. “In this review of 9 cross-sectional and 1 longitudinal study, little statistical significance between bone density and bone content was found between vegetarians and omnivores.”
- Effect of vegetarian diets on bone mineral density: a Bayesian meta-analysis. “The results suggest that vegetarian diets, particularly vegan diets, are associated with lower BMD, but the magnitude of the association is clinically insignificant.”
- Long-Term Vegetarian Diet and Bone Mineral Density in Postmenopausal Taiwanese Women. “Long-term practitioners of vegan vegetarian were found to be at a higher risk of exceeding lumbar spine fracture threshold … and of being classified as having osteopenia of the femoral neck.”
So, although the “calcium-leeching” properties of animal protein is a common battle cry in the vegan world, the research just doesn’t support it. There are even some interesting (and peer-reviewed!) papers out there looking at how belief systems influence the interpretation and misrepresentation of bone/protein studies. Read that link because it’s awesome.
But back on topic. This paper, with Campbell’s own name on it, suggests a strongly bone-protective role for dairy in the diet. Not quite the message we heard in “The China Study.”
Reply to TC Campbell by Frank B. Hu and Walter Willett.
Short n’ sweet, this one speaks for itself. Frank Hu and Walter Willett responded to Campbell’s criticism of their findings in the Nurses’ Health Study*, and in so doing, offered their take on the China Study.
Campbell questioned the validity of our findings because they contradict the results of international correlation studies on animal product consumption and disease rates. … Correlational studies conducted within a country can usually provide more credible data than international comparisons because of relatively homogeneous populations and the possibility of collecting data on potential confounding variables at individual levels. A survey of 65 counties in rural China, however, did not find a clear association between animal product consumption and risk of heart disease or major cancers.
*The Nurses’ Health Study has its share of problems, and I actually agree with Campbell’s assessment in some cases, but that’s a story for a different day.
Correlation of Cervical Cancer Mortality with Reproductive and Dietary Factors, and Serum Markers in China by Wan-De Guo, Ann W. Hsing, Jun-Yao Li, Jun-Shi Chen, Wong-Ho Chow, and William J. Blot:
As this paper notes, cervical cancer is the second leading cause of cancer-related deaths among Chinese females. There are a few known risk factors (especially herpes infection), but other than that, the reason for its variation across China is a big mystery. Can we blame animal foods for this one?
After identifying the variables that had the strongest correlations with cervical cancer in the China Study data—including herpes infection, serum ferritin, body mass index, cigarette smoking, age at first birth, green vegetable intake, and animal food consumption—the researchers ran multiple regressions to see which correlations were legit. The results?
When these variables were considered in the multiple regression analysis, early age at first birth and higher BMI were positively associated with cervical cancer mortality, while consumption of green vegetables and animal foods were negatively correlated.
Simply put, animal foods were inversely associated with death from cervical cancer—meaning the folks eating more animal products had fewer deaths from this disease. If it were true that animal protein promotes the growth of cancer cells, as Campbell theorized based on his research with casein and liver cancer, then we’d expect to see the opposite. What an anomaly.
Risk Factors for Stomach Cancer in Sixty-Five Chinese Counties (PDF) by Robert W. Kneller, Wan-De Guo, Ann W. Hsing, Jun-Shi Chen, William J. Blot, Jun-Yao Li, David Forman, and Joseph F. Fraumeni, Jr.
In this paper, the authors pulled out variables that had strong correlations with stomach cancer in the China Study data, and then analyzed them in greater depth using some regression models. The most significant associations involved plants—but since the focus of this post is animal foods, let’s see what the researchers uncovered on that front:
Consumption of green vegetables, rice, meat, and fish was associated with reduced mortality. … On the other hand, salt-preserved vegetables, potatoes, wheat, and millet, plus combinations of wheat, corn, and millet, were correlated with significantly increased mortality.
Our finding of a significant inverse association for meat is consistent with a recent case-control report from Turkey. Meat is a common source of selenium, which showed the strongest protective effect among all the plasma micronutrients.
Say what? Not only did meat not seem to increase stomach cancer rates (which we might expect if Campbell’s “animal protein spurs cancer growth” hypothesis held true), but it actually showed the opposite trend. Perhaps the selenium was enough to counteract meat’s general evilness. Also interesting is that the foods associated with increased stomach cancer mortality were mainly of plant origin. Including wheat.
However, this paper did uncover a relationship between stomach cancer and one animal food: eggs. Given the known associations between stomach cancer and salt-preserved foods (and lack of association with any other animal food), I’d wager this link has something to do with the way eggs are eaten in China rather than anything inherent in the eggs themselves. China is known for dishes like salted duck eggs, which are preserved with salt or charcoal, and century eggs, which sit for weeks or months in a mixture of ash, salt, clay, and other ingredients, gradually becoming glossy balls of ammonia stench. It seems likely that preserved eggs could increase stomach cancer risk for the same reason preserved vegetables do.
The researchers also note that they “know of no previous reports linking egg consumption to increased [stomach cancer] risk” and that the counties with high egg consumption had other confounders not accounted for—including a tendency to be in coastal areas, have a higher percent of industry-employed workers, and have higher indexes of socioeconomic status. In addition, the range of egg intake in China may be too narrow to determine anything meaningful: The eggiest county ate the equivalent of only two or three chicken eggs per week, and the average for all counties was about 1/15th of an egg per day.
Nonetheless, no other animal foods, nor animal protein as a whole, contributed to stomach cancer risk in this analysis. Which is pretty interesting, because Campbell still links stomach cancer with animal foods via blood cholesterol in his book (pages 78 – 79):
What a surprise we got! As blood cholesterol levels decreased from 170 mg/dL to 90 mg/dL, cancers of the liver, rectum, colon, male lung, female lung, breast, childhood leukemia, adult leukemia, childhood brain, adult brain, stomach and esophagus (throat) decreased.
Yet as meat consumption increased, stomach cancer decreased. How curious!
Fish consumption, blood docosahexaenoic acid and chronic diseases in Chinese rural populations by Yiqun Wang, Michael A. Crawford, Junshi Chenb, Junyao Li, Kebreab Ghebremeskel, T. Colin Campbell, Wenxun Fan, Robert Parker, and Julius Leyton.
As an animal food rich in protein, fish should top the list of health villains—at least according to the animal protein/disease theory. Was that the case in this study? From the full text (not linked above):
Our finding that the highest blood cholesterol levels in the Chinese were associated with DHA and fish consumption but with the lowest risk [of heart disease], is also a contradiction of what might be expected.
The higher blood LDL cholesterol levels associated with the marine coastal and lacustrine communities in China as compared with their inland neighbours, needs to be seen as starting from very low levels. In this context, it is the largely vegetarian, inland communities who have the greatest all risk mortalities and morbidities and who have the lowest LDL cholesterols. It could well be that there is a minimum level of LDL cholesterol below which cell membranes are adversely affected.
Translations: In the China Study data, fish-eaters (with higher cholesterol, to boot) were generally healthier than the more vegetarian populations that didn’t consume seafood.
And here’s a nice table showing the associations between red blood cell concentration of DHA (which the researchers determined was mostly due to fish intake) and chronic diseases. (The bars to the left of the center line indicate a negative or “protective” correlation; the bars to the right are positive.)
The researchers note that the liver cancer correlation has a likely explanation:
[It] is not difficult to visualise the reason for the link with liver cancer. The coastal, estuarine and lacustrine regions with the high fish and sea food intakes are also those with the highest humidities. Storage of food in regions of high humidity is known to encourage the spread and growth of hepatitis B virus and Aspergillus flavus which produces aflatoxin, both are major causes of primary carcinoma of the liver.
Diet and Blood Nutrient Correlations with Ischemic Heart, Hypertensive Heart, and Stroke Mortality in China by Wande Guo, J.Y. Li, H. King, and F.B. Locke.
Here we’ve got a paper reporting some of the correlations between blood markers, food intake, and cardiovascular disease in the China Study data. Yep, cardiovascular disease! We should see animal products splattered all over this one, right?
Five variables were positively correlated: triglycerides and herpes antibodies with ischemic heart disease; salt and phosphorus (females) with hypertensive heart disease; and only albumin (males) with stroke. … Some findings confirm those observed in the West (salt, triglycerides, herpes, legumes, oleic acid, and liquor), but molybdenum and age at first pregnancy have not been emphasized previously. Still others significant in the West have not been observed here, such as cholesterol and smoking.
This bears repeating: This correlative study (the kind Campbell drew heavily from to link animal products and disease in the China Study data) found no association between cardiovascular diseases and cholesterol. Nada. Or smoking, which is also pretty interesting. How peculiar that one of the most monstrous “diseases of affluence” was unrelated to blood cholesterol.
Okay, folks: That should be enough to chew on for now. I hope to see some of you at the Ancestral Health Symposium in a few days!