on June 25, 2013
at 10:14 PM
Long post incoming.
The first thing to note is that an updated meta-analysis has been published http://www.ncbi.nlm.nih.gov/pubmed/23001745 Although it's not open access, unfortunately. I like these authors, they do a good job summarizing the stats and they're being critical. They also have a paper solely on processed meats where they discuss the issue http://www.biomedcentral.com/1741-7015/11/136 though I'm not particularly enthused by their overall conclusions, I prefer mine, obviously :)
Basically there is still no epidemiological evidence to suggest that unprocessed red meat is associated with cardiovascular risk, but there is a notable association with unprocessed meat. Also with regards to diabetes, there is a modest association with diabetes risk for unprocessed meat and a larger one for processed meat.
However we really have to analyze this issue well and acknowledge other possible hypotheses to explain the correlations besides "meat increases the risk for diabetes". And we also have to make the distinction between evidence-based medicine and science-based medicine/nutrition. Here's a good article on it http://saveyourself.ca/articles/ebm-vs-sbm.php
Basically the difference is that "evidence-based" approaches tend to automatically defer to whatever the current evidence is and then change if the evidence changes. But "science-based" approaches look at the evidence in the context of all of our scientific understanding and assess plausibility, knowing that most of this evidence doesn't provide us with causation or context. I choose the latter approach because it's oftentimes far more fruitful, illuminating interesting contexts and providing solutions to issues with some foods besides simply not eating them.
So first we have to ask ourselves why there's a significant correlation between processed meats and these diseases, whereas the evidence is less impressive for unprocessed meat. The authors think that a great deal of the CVD risk can be explained by the extra sodium content. High sodium intake can have a modest effect on blood pressure which is more or less negated by a high potassium intake, but considering the fact that potassium intake is generally insufficient in many people, that extra sodium in the processed meat might contribute to the association between processed meat and cardiovascular disease.
Yet it seems unlikely that sodium is the big X-factor here, they seem to be treating associations between blood pressure and CVD as causation, when blood pressure is probably relevant but also accompanied by other factors that contribute. Either way we need to look deeper into this.
Oftentimes when interpreting these epidemiological studies people invoke the idea that there is a healthy-user bias present. The reasoning is that people who are more health-conscious and do more healthy things in general are less likely to consume a food that is perceived to be unhealthy (red meat in this case). And people who don't care about their health and eat all sorts of junk, smoke, drink too much, don't exercise, etc, are more likely to consume the food in question. There are multivariate linear regression analyses that try to control for some of these confounding factors, but they tend to fall short, failing to control for a myriad of junk foods.
Can we can say that processed meat is more maligned and perceived as less healthy than unprocessed red meat? Both are pretty heavily maligned, but there exists a perception that lean red meat might be healthier. Because the reason that most people think that red meat is bad is because it's fatty; this might lend credence to the explanation that the people who eat more unprocessed meat are generally more health-conscious than the people who eat more processed meat, and the association is just an illusion created by confounding factors.
On the other hand the difference in risk is quite different between the two, and hamburgers are considered to be unprocessed meat, yet often consumed at fast food restaurants alongside junk. I personally don't think that the healthy user effect can account for the difference between unprocessed and processed meat.
There is evidence that some of the chemical reactions that happen during high heat cooking can be harmful. There are nitrosamines, various heterocyclic amines, cholesterol oxidation products, lipid oxidation products and AGEs to name a few classes. There is much debate about the actual relevance of any of these; I personally don't think that COPS or AGEs from meat are a big deal in their doses and absorption. What seems less contentious is the idea that cooking intensity (duration, heat, contact with various surfaces) can influence health.
In Stephan Guyenet's epic series on what causes insulin resistance, he talks about inflammation, and he also talks about high heat cooking methods. http://wholehealthsource.blogspot.ca/2012/01/what-causes-insulin-resistance-part-vi.html The difference between high heat cooking and low heat cooking methods can actually make a big difference with regards to inflammation and insulin resistance, which are very relevant to cardiovascular disease and diabetes.
We also have evidence that some processed meats (fried bacon! I know I know...) contain very large amounts of various inflammatory and reactive heterocyclic amines compared with various unprocessed meats http://www.ncbi.nlm.nih.gov/pubmed/22129588 And there is quite a bit of variance between different cooking methods for unprocessed meats.
So I find it plausible that processed meat consumption is bad when it contains large amounts of cooking toxins which can increase the risk of diseases. But this wouldn't necessarily implicated processed meats as a whole, and it wouldn't necessarily exonerate unprocessed meats cooked in a manner that produce a lot of HCAs. It's just that in practice maybe processed meats are worse in this regard, but there's the possibility that they could be cooked in a way that minimizes cooking toxins; though I'm not entirely sure on this like I am that moderate cooking in the oven and water-based methods are generally benign for unprocessed meats. There are also marinades and spice rubs that reduce their formation. This is an excellent article on it http://naturalmedicinejournal.com/article_content.asp?article=88
There's also the possibility that endogenous antioxidant defenses like glutathione can reduce the toxicity of HCAs and someone eating a highly nutritious diet would be less affected, though it's not extremely well-studied. Maybe a Paleo scientist could do a study with identical diets, lots of fruits, vegetables and nutritious foods, and make the only difference the cooking method used to cook meat. That would be a great experiment.
Nitrosamine formation is one of the more popular reasons for the difference between processed and unprocessed meat. It's not that nitrates are bad, it's that they react with iron or under heat and produce harmful compounds. Could be, though endogenous production is the main source, and we're not sure if that really differs significantly between the two. Unprocessed has less nitrate, but still some nitrate, but it has more heme iron which is a primary reactant. Is the rate-limiting factor for endogenous nitrosation nitrates or is it heme iron?
This process can be suppressed by eating significant amounts of calcium and chlorophyll with antioxidants in the same meal as the red meat. The calcium and chlorophyll chelate the excess heme iron that doesn't get absorbed and prevent it from reacting with nitrates to form nitrosamines.
In summary, if there is anything legitimate in these associations, it's probably cooking toxin content. But we still can't judge the magnitude of the effect because of potential confounding factors. I think that reducing the formation of cooking toxins when cooking is beneficial and as a rule it's good to try to cook gently, though I admit that the true relevance to health in all contexts is less clear. Processed meats deserve scrutiny, and so do unprocessed meats, but the argument against the latter seems very weak. We don't even know if unprocessed meat being associated with processed meat increases the association with diabetes due to the association between the two (I would assume that there's an association there).
There's an interesting question that I asked myself: Why if unprocessed red meat is still a significant source of HCAs in the diet isn't it associated with cardiovascular risk at all? And if it's associated with diabetes risk, and diabetes causes heart disease, what gives?
The best answer I can think of is that unprocessed red meat is a beneficial food with numerous nutrients that help prevent heart disease. It has carnitine, carnosine, taurine, creatine, and others that are lost to some extent with processing and high heat cooking, and more prevalent in beef and lamb than pork, which is the main processed meat. All of those nutrients are beneficial to diabetics. Health Correlator found that red meat was associated with less risk for other diseases in diabetics http://healthcorrelator.blogspot.ca/2012/03/2012-red-meat-mortality-study-arch.html
If cooked gently most of the time and eaten in the context of a healthy diet, unprocessed red meat might be a health food. Speculation? Yes but nutritional epidemiology is highly speculative. All we have is the evidence and our analyses of it at the end of the day.
I predict that if they ever do diabetes/cardiovascular epidemiology for unprocessed red meat cooked at high temperatures vs. low temperatures, they'll find that they're different with low heat being associated with lower rates. That is unless these views catch on and health-conscious people start cooking meat at lower temperatures!
The best studies would try to control for nutrient density and baseline glutathione levels but this is probably unrealistic.
Questions and criticism are welcome. Especially if you know your chemistry!
on May 28, 2013
at 04:17 PM
That is the type of article that supports the paleo perspective. I like that it's a meta-analysis, seems solid enough.