Red Meat Consumption and Mortality

Last week, I coincidentally posted on the health benefits of grass-fed beef, with my conclusion basically stating that grass-fed, pastured beef and grain-fed beef are so vastly different in their nutritional values, that they cannot realistically be considered the same food source. Yet after reading the recent article “Red Meat Consumption and Mortality”1 which is all over the news, and seeing my inbox full of about 100 emails, I quickly thought to myself: “Here we go again”.

 

More food-frequency questionnaires, biases, lumping of foods together, and then placing the cross-hairs on red meat.

The gist of this study is that the more meat you eat, the larger your mortality risk. In fact, they state that 9.3% of deaths in men and 7.6% in women could be prevented by consuming a half less per serving of red meat per day. The strength of this study is a large population of participants, and I admire the authors’ honesty in admitting their results may be a little shaky (in the manuscript though, not the abstract). However there are some definite strikes against the study, and therefore basing important dietary decisions based on this data is unwise (to me). Unfortunately every media outlet has blindly accepted their results and conclusions, spreading the claim red meat is harmful.
 
In this study, the authors assessed dietary intake of meat in participants of the Health Professionals Follow-up Study through food frequency questionnaires (FFQ) with:

“In each FFQ, we asked the participants how often, on average, they consumed each food of a standard portion size. There were 9 possible responses, ranging from “never or less than once per month” to “6 or more times per day.” Questionnaire items about unprocessed red meat consumption included “beef, pork, or lamb as main dish” (pork was queried separately beginning in 1990), “hamburger,” and “beef, pork, or lamb as a sandwich or mixed dish.” The standard serving size was 85 g (3 oz) for unprocessed red meat. Processed red meat included “bacon” (2 slices, 13 g), “hot dogs” (one, 45 g), and “sausage, salami, bologna, and other processed red meats” (1 piece, 28 g).”

STRIKE #1: Lumping many food-types together, even though in practice they are very different.

This may be obvious to many of you, but clearly the nutritonal benefit (or detrimental effect) of a hotdog full of processed meat and nitrates is NOT THE SAME as grass-fed beef. Similarly, processed bacon made from chronically-stressed pigs (see my previous posts on the living conditions of animals in industrial farms) is NOT THE SAME as ethically and responsibly raised pork. All meat is not created equal, so to study industrially produced meat products and apply that data to free-range, grass-fed meat is scientifically false.
 
 

STRIKE #2: Basing an entire study on food questionnaires and self-reporting, both recording methods fraught with error and bias.

Shortly after explaining this, the authors state that validity of these questionnaires have been described elsewhere, referring to several papers which analyzed these techniques2,3. One thing they fail to mention is that these same techniques have been intensely criticized for years for inadequacy and inaccuracy4-6, even including an article titled “Is It Time to Abandon the Food Frequency Questionnaire?7, yet they base their strong conclusions on them. Self-reporting is also fraught with biases8 as those in poorer health may attribute their issues to those health hazards which have been preached to them their entire life, i.e. red meat is the cause of all evil…
 
Also, while studies show that the drastic differences in food consumptions throughout the year as seasons and growing periods change requiring data to be collected several times throughout the year for adequacy9,10, yet in this study, not only was information not collected multiple times throughout the year, but it wasn’t even collected on a yearly basis. Data was in fact only collected in 1986, 1990, 1994, 1998, 2002, and 2006. Also, when they asked patients questions regarding their diet and they failed to answer, here is how they solved this problem:

“We replaced missing values in each follow-up FFQ with the cumulative averages before the missing values.”

 

STRIKE #3: Using a poor method of dietary recording and assessment to begin with, and adding insult to injury by further inadequate methods of data collection.

While the food-frequency questionnaires and the averaging of missed values alone will cause significant errors in recording and calculations, lumping all of these meat types together would be the equivalent of grouping leafy green vegetables and bold-colored fruits packed with antioxidants with grain-based bread and pasta then quoting the ill-effects of all plant-based products. Also, while more time-consuming, I like when patients update their diet in real-time via web sites like FitDay, at which point I can discuss the diets with them to see what they are eating, as opposed to a questionnaire where they try to piece together what they ate over the last year, not to mention a site like Fit Day has actual nutritional values for each type of food, and doesn’t lump a large mixture of very different food types as one, per say “red meat”. Finally, they also claim to adjust for nutrients, yet the nutrients in grass-fed and wild animals (all red meat) are several times higher than those of grain-fed beef. This was neither accounted for nor discussed in the article.
 
 

STRIKE #4: Finding out there may be several other factors in these groups that could lead to your findings and then using the magic wand of “controlling” for these variables to make them disappear.

They also do state that:

“Men and women with higher intake of red meat were less likely to be physically active and were more likely to be current smokers, to drink alcohol, and to have a higher body mass index”

They note that these participants also ate less vegetables. This is often the case in these studies – several other variables stand in the way of the findings you want. What usually happens in statistics is that you somehow account or control for this, which has been aptly labeled “statistical alchemy”, or making gold out of sub-par results.
 
 

In Conclusion:

Red meat (and not accounting for whether it is grain, hormone, pesticide, and antibiotic-laden), using poor recording mechanisms, inherent biases, and intricate statistical methods, as well as being physically inactive, drinking more alcohol, smoking, and being overweight show that there may be an increase in mortality. Frankly, the techniques in the article above have been criticized a thousand times over in past studies that have used similar methods.
 
 

Final thoughts:

Cutting out red meat, especially healthy, nutrient-dense grass-fed beef and free-ranging animals will only lead to their replacement with more carbohydrate sources, including the basis of the food pyramid: grains. We can use all the poor reporting methods and intricate calculations possible to prove otherwise, but thirty years of skyrocketing rates of obesity, diabetes, and chronic disease have shown us the results of these dietary recommendations, yet we continually cling to them.
 
It’s time for the medical field to admit our mistakes and move on.
 
 

References:

1. Pan A, Sun Q, Bernstein AM, et al: Red Meat Consumption and Mortality: Results From 2 Prospective Cohort Studies. Arch Intern Med:archinternmed.2011.2287, 2012
2. SALVINI S, HUNTER DJ, SAMPSON L, et al: Food-Based Validation of a Dietary Questionnaire: The Effects of Week-to-Week Variation in Food Consumption. International Journal of Epidemiology 18:858-867, 1989
3. Hu FB, Rimm E, Smith-Warner SA, et al: Reproducibility and validity of dietary patterns assessed with a food-frequency questionnaire. The American journal of clinical nutrition 69:243-249, 1999
4. Schaefer EJ, Augustin JL, Schaefer MM, et al: Lack of efficacy of a food-frequency questionnaire in assessing dietary macronutrient intakes in subjects consuming diets of known composition. The American journal of clinical nutrition 71:746-751, 2000
5. Kipnis V, Midthune D, Freedman L, et al: Bias in dietary-report instruments and its implications for nutritional epidemiology. Public Health Nutrition 5:915-923, 2002
6. Kipnis V, Subar AF, Midthune D, et al: Structure of Dietary Measurement Error: Results of the OPEN Biomarker Study. American Journal of Epidemiology 158:14-21, 2003
7. Kristal AR, Peters U, Potter JD: Is It Time to Abandon the Food Frequency Questionnaire? Cancer Epidemiology Biomarkers & Prevention 14:2826-2828, 2005
8. Adams A, Soumerai S, Lomas J, et al: Evidence of self-report bias in assessing adherence to guidelines. International Journal for Quality in Health Care 11:187-192, 1999
9. Capita R, Alonso-Calleja C: Differences in reported winter and summer dietary intakes in young adults in Spain. International Journal of Food Sciences and Nutrition 56:431-43, 2005
10. Joachim G: The influence of time on dietary data: differences in reported summer and winter food consumption. Nutrition and health 12:33-43, 1997
 
© Caveman Doctor 2012. All Rights Reserved

Comments
44 Responses to “Red Meat Consumption and Mortality”
  1. Great and timely response. I’m a huge fan of paleo living but I’m not smart enough to put together a well thought out dismissal of articles/studies like this. Thanks for putting this together.

    • cavemandoctor says:

      Thanks for the comments. Pretty much any headline (even for caveman/paleo type diets) require a decent look at the manuscripts as the media outlets often sensationalize. Glad I could help.

  2. Yoda says:

    Can you post more on why the statistical alchemy is so weak? I think this could be the key weakness, and it deserves at least a few references or a better explanation as to why the “correction” for these factors is not as easy as the research makes it out to be.

    • cavemandoctor says:

      Hi thanks for the comments. This would be a good topic that I will add to the list. A simple exclusion of data here or there, or using different statistical methods until you achieve the results you want is easier than we think. In fact, I recently published a paper where I used somewhat complex statistical methods. I felt as though they were accurate (and my results stated that the findings were insignificant and no conclusions could be made). HOWEVER, I ran the stats a couple different ways afterwards and found that there were actually several ways to make my results appear significant (even very significant), by accounting for different variables, or including or excluding other variables. Hopefully if I would have submitted the research using these methods (which I never would because I didn’t care that my results were inconclusive because I felt that this was correct and stating otherwise would be lying or unethical) the reviewers would have caught it, but it is pretty easy to cover your tracks to get a manuscript through.

      In fact, search google for similar occurrences and criticisms of stats issues and the results are endless…

      Thanks!
      CD

  3. Skeptic says:

    Hey CD, got a question…

    I haven’t seen a copy of the questions asked. But I would imagine that folks who consume red meat versus nut-eaters are far more likely to consume large amounts of high-calorie carbohydrates such as french fries, buns, etc.

    I noticed the study reported an increase risk of diabetes for the meat-eaters which leads me to believe they consume far more carbs than plant-eaters.

    Did the questionnaire take into account carbohydrate consumption??? If not, then I believe this study to be completely worthless.

    • cavemandoctor says:

      Absolutely. Great point. As soon as you see food “in the form of a sandwich” listed as meat, you can pretty much throw the study away – and this study helps us confirm that. They “accounted”, in what was their idea of accounting, for these variables, but I am not even sure if it is even possible to accurately model the effect of carbohydrates when consumed with meat (as meat without bread and other foods is obviously very different). You are 100% correct – the study is worthless from an academic and medical point of view (but great for media headlines), thus why we roll our eyes while the media outlets have a field day.

  4. Skeptic says:

    Thanks for the quick response!

    On second thought, more accurately, if the study didn’t account for caloric intake it’s worthless. French fries are high in calories, obviously. Did the study ask that kind of question? (As opposed to asking whether consumption was in the form of a sandwich) Obviously the side-dishes are of extreme importance in a study like this. In other words, calorie consumption is positively correlated with higher mortality rates (after allowing for exercise, etc.), so the meat eating correlation may be completely spurious (as well as for other reasons you mentioned).

    • cavemandoctor says:

      Yeah, that’s the great part of these studies the accounting you refer to is in the eye of the beholder. If they don’t see any problems with some things (i.e. carbs, etc.) they don’t necessarily account for them. But really you can account for as little or as much as you want, as long as you get your result. Statistical alchemy at its best.

  5. PB says:

    “What usually happens in statistics is that you somehow account or control for this, which has been aptly labeled “statistical alchemy”, or making gold out of sub-par results.”

    As a PhD biostatistician I have to say you raise some good points overall, but this is really off-the-mark. You clearly don’t understand statistical methodology–there are mathematical principles used to derive these methods. To call it “statistical alchemy” (repeatedly) without saying specifically where you take issue with their methodology is simply incendiary (did you miss the biostats week in med school?). You lost credibility right there. I’ve looked over the paper and i think it was sound. The main thing that irks me about all of this is the overinterpretation by the media.

    • cavemandoctor says:

      Hi and thanks for the comments. Your knowledge of stats is clearly superior to mine – that should be said right from the start. In fact, medschool doesn’t offer a week of biostats to miss, as we must learn on our own or during residency, so definitely never missed it! That being said, I know my knowledge of stats is not at a PhD level, which I why I consult with many statisticians (work and friends) on a daily basis that are at a high level and I am lucky enough to be able to run my thoughts and ideas through them.

      My comments in the above article are in no way to discredit statisticians and I’m sorry if I offended you, however, I am very critical of my own profession as well. Also, I have watched statisticians purposely change results right before my very eyes from not significant to very significant to illustrate these points (these statisticains were also very critical of their own profession). As for the article, looking at their methods, there are serious issues (adjusting for a,b,c, but not adjusting for x,y,z, sandwiches were considered a fat – so did they adjust for fats with carbs and the insulin effect, how about interactions?, the list goes on and on). All we see is the final result, not the process of coming to this result. The point being that they make leaps of faith to give very bold conclusions.

      I generally write to simplify (while trying not to overlook) and not to complicate, and specifically addressing this in my post would be neither efficient nor necessary to my time or my audience’s. There is plenty already written on this statistics topic on the internet so a simple google search may answer any remaining questions you may have.

      Thaks for the constructive part of your comments though as it may be useful for some readers,
      CD

    • Yoda says:

      There’s a difference between mathematically sound and biologically/mechanistically sound. While the mathematics may give 95%CI results, that doesn’t mean that the underlying assumptions are valid. Otherwise, how do you explain the HRT vs. heart disease results that were statistically gleaned from the same cohort as the meat study back in the 1990s? The mathematics said confidently that HRT decreased heart disease risk by 50%. Subsequent direct research has shown the opposite. The main thing that irks me is the over-interpretation of these statistical methods without concern for the very significant assumptions that are made. Even more so, the fact is that many things are not accounted or corrected for because they are simply not known or understood. Certainly a 500% increase in HI would be reasonable for starting to make a case. But 13%? Surely the errors in unknowns must be way larger than that, so it should not be considered statistically significant.

      Not being a biostatician myself, I would appreciate a response as to why you feel that this type of result (13%) is significant.

      • cavemandoctor says:

        Amen. Great post. Stats is important as it tweaks out associations that are hypothesis GENERATING. The HRT example is a great one often quoted as it gives us a strong example of happens when we take associations and falsely make causation.

        Thanks!
        -CD

  6. Steve says:

    Hey, by the way I enjoy reading your website. I’ve never commented before, but I found this great critique of the article by someone with a grass-fed beef point of view that I thought I’d share. http://www.zoeharcombe.com/2012/03/red-meat-mortality-the-usual-bad-science/

  7. BW says:

    Good stuff Doc. I am interested in this statement. Can you tell me exactly what the authors said here or tell me how to find the full manuscript? Thanks.

    “I admire the authors’ honesty in admitting their results may be a little shaky”

    • cavemandoctor says:

      Thanks!

      Here is the article link (not sure if this will work for you).

      Good question. Their comments on why these associations may be present included the typical reasons: burnt meat, sodium, and nitrates (which to me at least approaches the realization that all red meat isn’t the same). I thought this at least was saying there may be other considerations, though as I read this more and more I think I may have been giving them a little too much credit in this area. My original post was much harsher, but I toned it down to not get banished from medicine!

      Thanks!
      CD

      • BW says:

        Thanks again. I’m trying to get better at understanding these things myself instead of relying on smarter people. Is this statement saying that saturated fat and cholesterol may lessen the risk of CVD mortality?

        The association between red meat and CVD mortality was moderately attenuated after further adjustment for saturated fat and cholesterol, suggesting a mediating role for these nutrients.

        • cavemandoctor says:

          BW,

          No problem. They are basically saying the risk is there, but when they adjusted for saturated fat and cholesterol, the risk lowered. This is actually saying the opposite from your question – i.e. that they cause CVD mortality. For instance, if x,y, and z cause an increase in mortality and red meat contains them all, can we adjust for x,y,z to statistically get rid of them to see if red meat alone causes an increase in mortality. Thus when the authors adjusted for cholesterol and sat fat (under the assumption they cause CVD) the mortality of red meat was lowered (but still present). Therefore they conclude that red meat still causes increased mortality. They later discuss a similar method for accounting for iron.

          I hope this makes sense as it’s a simplistic explanation for a somewhat complicated topic.

          -CD

          • BW says:

            Makes perfect sense. Thanks for answering, much appreciated.

            At least they referred to sat fat and cholesterol as nutrients.

  8. Devon says:

    Excellent deconstruction of the “study”
    I know folks are all abuzz about this recent scientific catastrophe, but anyone that is interested in some insight as to why this study was SO bad woukd do well to read Dr. Brian Goldmans book “junk science”, it essentially explains a lot of the points made here as to why this study cant be taken at all seriously. As if it is some sort of miracle discovery that someone who visits McDonalds 3 times a week for a decade is going to be unhealthy. Yes.. It is most certainly the meat causing this, if you can call McDonalds meat.

    • cavemandoctor says:

      Devon,

      Thanks for the insightful comments. You are right on. Also, it’s an unfortunate part of our current food system, but the very definition of “meat” is in the eye of the beholder.

      -CD

  9. Mike says:

    Nicely articulated article covering all the points and thoroughly discrediting this most recent attempt to further promote consumption of refined carbohydrate. Keep flying the flag for a healthy sustained future for all.

    BR

    Mike.

    PaleoWorks Ltd
    Ilkley,
    UK

  10. solaro says:

    Hello ! excuse me for my bad english = about the “study” about “red meat consumption and mortality” =- with such a title, it is rather easy to be on the web.. and how to be sure that people can remember well what they ate four years earlier ?’(‘the food frequency questionnaires were sent every four years only) and this FFQ was so long, with so many many questions, how people can answer well all the FFQ long ? and after “study” about meat, we have now study about white rice, based upon the same datas (for a half of the cohort) – what can we expect now ? french wine, bread, caesar salad, love ? these studies are more and more numerous, and my freedom seems to be smaller and smaller

  11. Blake says:

    Hi, I read most of the original article, which can be found if you Google for it. I work in the technical field, and the statistics are done correctly, the premise of the research clearly explained, therefore the conclusion is significant and absolutely “need-to-know”. I am surprised you do not focus on the simple fact that by publishing science like this, the authors are on your side. This study is part of your argument. It shows that meat quality, as most people buy it, is poor and leads to poorer health and increased mortality. This study is the beginning of proof that we need better legislation enforcing the quality of meat. To address a remark made on your forum comments, I do not see how this study pushes consumers to eat non-meat products, rather it soundly trashes the food production methods our legislators have allowed and calls for an improvement. It incites people to seek out better quality meat, to protest the use of nitrates, hormones and steroids in meat production. Which is exactly what you want as well. I second the complaints made about the media treatment of this subject. But today’s newspapers wrap tomorrow’s fish (or meat…)

    • cavemandoctor says:

      Blake,

      Thanks for the comments. I agree with a lot of what you said, and while I would love to think the authors are on my (or our) side, their conclusions suggest avoiding ALL red meat and make no comment of type of red meat. Also, the icing on the cake was the invited review by Dean Ornish. After the article, we have a vegetarian invited to tell us about global warming and whatever other nonsense they are blaming on meat eaters. The article in itself was tough to swallow, but the follow-up raised the level of my issues with it.

      Your comment “It incites people to seek out better quality meat, to protest the use of nitrates, hormones and steroids in meat production.” is a great one and I hope this is the case. I also hope people push for more humane treatment of animals in general.

      Thanks,
      CD

  12. Jacques Duranceau says:

    Thank you for your cogent analysis. I have type 2 diabetes and I have spent the last 20 years researching all the cholesterol, low-fat, and high grain nonsense that is passing as science. One of the hardest things is dealing with the barrage of studies that show results that support the status quo – especially meta-studies of flawed studies.

Trackbacks
Check out what others are saying...
  1. [...] to fry.  The Blog, Constantly Varried had the aptly titled Red Meat: Here we go again while CavemanDoctor had a more sciency [...]

  2. [...] Red Meat Consumption and Mortality by Caveman [...]

  3. [...] The Caveman Doctor also published a nice piece on how the study was fraught with error and bias. Bookmark/Share | Suggest a Link « Previous: Why Not Pardon Drug Users Instead of Murderers? | LRC Home | LRC Blog [...]

  4. [...] to fry. The Blog, Constantly Varried had the aptly titled Red Meat: Here we go again while CavemanDoctor had a more sciency [...]

  5. [...] Harvard looked at 18,000 women with food questionnaires (which are a horrendous method as discussed here), but it’s still at least worth giving it a look.  Their results showed that as we move from low [...]

  6. [...] to fry.  The Blog, Constantly Varried had the aptly titled Red Meat: Here we go again while CavemanDoctor had a more sciency [...]

  7. [...] lower triglycerides. Modern vegetarians have tried to discredit the benefits of animal products and red meat through statistics and flawed modeling and by following cohorts, but cannot show this in any trials [...]

  8. [...] Varied escogió un nombre apropiadamente titulado La Carne Roja: Aqui vamos nuevamente mientras que CavemanDoctor le puso un título más [...]

  9. [...] Varied escogió un nombre apropiadamente titulado La Carne Roja: Aqui vamos nuevamente mientras que CavemanDoctor le puso un título más [...]

  10. [...] Loss Difficult10 Tips For Making Healthier Food Choices — The Picky Eater: A Healthy Food BlogRed Meat Consumption and Mortality/* —————————————————————— ———- BASE LAYOUT [...]

  11. […] Também tem outros bons posts (todos em inglês) no site do Robb Wolf, no Constantly Varied, e no Caveman Doctor,  falando sobre o assunto. Quem tiver curiosidade de ler, vale o clique. Como podem ver, a […]

  12. […] are some great critiques of overblown claims about the ill effects of red meat consumption, like this, or this, in response to sensational news headlines in early 2012 that red meat consumption […]

  13. […] links red meat consumption with an increased risk of total, CVD, and cancer mortality. Check out this critique by Dr. Colin Champ of ‘the Caveman […]



Leave A Comment

Calendar

March 2012
M T W T F S S
« Feb   Apr »
 1234
567891011
12131415161718
19202122232425
262728293031  

Contact Caveman Doctor