A paper came out in December reporting mortality rates of different diet groups from the large EPIC-Oxford study containing 60,310 people from the UK (1).
The mortality rate before age 90 was no different between vegetarians (including vegans) and regular meat-eaters (1.02, 0.94-1.10). Vegetarians had lower rates of mortality from pancreatic cancer (0.48, 0.28-0.82) and lymphatic cancer (0.50, 0.32-0.79). Semi-vegetarians had lower rates of death from pancreatic cancer (0.55, 0.36-.86). Pesco-vegetarians had lower death rates from all cancers (0.82, 0.70-0.97) but higher rates of cardiovascular disease (1.22, 1.02-1.46).
In the main analysis (in the paragraph above), some participants were recategorized based on a change in their diet over the course of the study which included over one million person-years of follow-up. The researchers did a second analysis in which participants who changed their diets were removed, and found an 8% reduced risk of early death in vegetarians that was just statistically significant (0.92, 0.84-0.99). Limiting the results further, to deaths before age 75, strengthened the finding (0.86, 0.77-0.97).
When vegans were separated from other vegetarians, there were no statistically significant differences in mortality rates for the six main categories of death. Eliminating participants who had changed diet categories didn’t significantly change the results for vegans. There were only 166 vegan deaths as distinct from 1,929 deaths in the entire cohort; meaning that reaching statistical significance was going to be unlikely.
Results above were not adjusted for differences in body mass index (BMI); such adjustments were performed but they didn’t change the results substantially.
The fact that vegetarians didn’t have lower rates of death from heart disease in this study is surprising given that a 2013 report from EPIC-Oxford showed a highly statistically significant, 31% reduction in heart disease incidence among vegetarians (0.69, 0.58-0.82). This discrepancy as well as the lower death rates for vegetarians before age 75, but not before age 90, might be explained by cases of nonfatal heart disease leading to effective treatment.
I have updated the article Disease Rates of Vegetarians and Vegans with the results from a report on vegetarian mortality rates that was released this week from the Adventist Health Study-2. I have reproduced the highlights below.
In 2013, death rates for the first 5.8 years of Adventist Health Study-2 (AHS2) were released (1). When combining vegans, lacto-ovo-vegetarians, pesco-vegetarians, and semi-vegetarians into one group, vegetarians had a 12% lower risk of mortality. Vegans had a 15% lower risk of death, but it was not quite statistically significant.
The difference in mortality rates can mostly be explained by a lower incidence of cardiovascular disease among vegetarian men. Vegetarian women had about the same rates as non-vegetarian women. This is similar to the findings from the first Adventist Health Study. There was also a benefit for all vegetarians for death from renal and endocrine (mostly diabetes) disease.
The researchers said that having only 5.8 years of follow-up would bias the results towards not finding differences.
In comparing their findings to British vegetarians, they said:
“The lack of similar findings in British vegetarians remains interesting, and this difference deserves careful study. In both cohorts, the non-vegetarians are a relatively healthy reference group. In both studies, the nutrient profiles of vegetarians differ in important ways from those of non-vegetarians, with vegetarians (especially vegans) consuming less saturated fat and more fiber. It appears that British vegetarians and US Adventist vegetarians eat somewhat differently. For instance, the vegetarians in our study consume more fiber and vitamin C than those of the EPIC-Oxford cohort: mean dietary fiber in EPIC-Oxford vegans was 27.7 g/d in men and 26.4 g/d in women compared with 45.6 g/d in men and 47.3 g/d in women in AHS-2 vegans; mean vitamin C in EPIC-Oxford vegans was 125 mg/d in men and 143 mg/d in women compared with 224 mg/d in men and 250 mg/d in women in AHS-2 vegans. Individuals electing vegetarian diets for ethical or environmental reasons may eat differently from those who choose vegetarian diets primarily for reasons of perceived superiority for health promotion. We believe that perceived healthfulness of vegetarian diets may be a major motivator of Adventist vegetarians.”
Make sure you eat your fiber!
1. Orlich MJ, Singh P, Sabaté J, et al. Vegetarian Dietary Patterns and Mortality in Adventist Health Study 2. JAMA Intern Med. 2013;():1-8. doi:10.1001/jamainternmed.2013.6473. | link
Just before going on my break, I wrote about a recent study showing that meat-eaters have bacteria in their digestive tracts that turn carnitine (found in high amounts in red meat) into TMAO which causes atherosclerosis (see Carnitine, Red Meat, TMAO & CVD). That very day, a meta-analysis was released indicating that carnitine supplements can reduce mortality in people who have had heart attacks (1).
Does that meant that red meat actually prevents heart disease via it’s carnitine content?
When someone has a heart attack, their carnitine levels become depleted. The meta-analysis showed that supplementing with large doses of carnitine (an optimal dosage of 6-9 g/day, many times more than one could get from eating red meat) can reduce mortality, particularly in the first 5 days after the heart attack.
Ventricular arrhythmias and angina were also reduced, but heart failure and second heart attacks were not. The paper did not show how long these studies lasted – this information might have been included in their on-line charts, but I could not access them. Many of the studies were not double-blinded and there were some other methodological problems, so it’s not even clear whether carnitine does provide a benefit for all of these parameters, though I would not be surprised if the reduction in 5-day post-heart attack mortality holds true.
In any case, this meta-analysis has basically nothing to do with the study on carnitine and TMAO.
1. Dinicolantonio JJ, Lavie CJ, Fares H, Menezes AR, O’Keefe JH. L-Carnitine in the Secondary Prevention of Cardiovascular Disease: Systematic Review and Meta-analysis. Mayo Clin Proc. 2013 Apr 15. doi:pii: S0025-6196(13)00127-4. 10.1016/j.mayocp.2013.02.007. [Epub ahead of print] | link
Of Meat and Mortality – Part 4 is a review of the results from a study combining the data on 37,698 men from of the Health Professionals Follow-up Study (1986-2008) and 83,644 women from the Nurses’ Health Study (1980-2008). This study was conducted by the Harvard School of Public Health and published one year ago (1).
There were some benefits this study had over the others. The inclusion criteria was a bit more stringent (you could not have had angina or a heart attack compared to just not having had a heart attack in the other studies (2)), the follow-up was long (22-28 years, the other studies either had much shorter follow-up or much less people to follow), and they assessed the diet regularly over time. Since they were all health professionals, it minimized income and lifestyle disparities.
The intake amount categories were divided into fifths, with the largest being about 2 servings per day for both men and women (unprocessed plus processed red meat).
They only reported two models: age-adjusted-only and multivariate. Once again, their multivariate model contained adjustments that likely interacted with red meat’s effect on mortality: history of diabetes, hypertension, or hypercholesterolemia, and energy intake. Even so, unprocessed red meat was associated with increased mortality (1.36, 1.25-1.47) as was processed red meat (1.27, 1.18-1.38). When men and women were separated, the findings held. The results also held for cardiovascular disease (no surprise) and for cancer (not a huge surprise, but the association was weaker).
The authors said:
“Compared with red meat, other dietary components, such as fish, poultry, nuts, legumes, low-fat dairy products, and whole grains, were associated with lower risk. These results indicate that replacement of red meat with alternative healthy dietary components may lower the mortality risk.”
Conclusion Of Meat and Mortality
Let’s sum up the 4 studies:
EPIC (3) – Large study, highest red meat intake category was about 2 servings per day, processed red meat was strongly associated with mortality, unprocessed was associated with mortality in the model that did not adjust for body mass index (BMI), removing first 2 years of follow-up didn’t change results.
NHANES (4) – Small but long study, highest intake category was roughly 1.5 times per day, unprocessed and processed meat was strongly associated with mortality until results were adjusted for many variables which could be influenced by red meat, strong reason to believe people in the lowest intake category had cut down on red meat due to previous disease.
Shanghai (5) – Only “red” meat was pork and very little processed meat, highest intake category was about 1 serving per day, red meat intake was positively associated with total mortality among men but not among women, results were likely adjusted for conditions that could be affected by red meat intake.
Harvard (1) – Highest quality study, largest intake category was 2 servings of red meat per day, red and processed meat strongly associated with mortality in both men and women even after adjusting for many factors that could be influenced by red meat intake.
My conclusion is that eating two servings of red meat per day mostly likely increases the risk of early death over eating a half serving or less, and especially if it’s processed. To be more certain, a model is needed that does not adjust for any of the factors that red meat likely influences while adjusting for all those that it doesn’t.
That said, it’s not a slam dunk and I’d rather people were eating cows and pigs than birds (because it takes so many more birds to produce the same amount of meat) so I don’t necessarily consider this good news.
My vote is for people to replace the red meat in their diet with legumes and nuts!
1. Pan A, Sun Q, Bernstein AM, Schulze MB, Manson JE, Stampfer MJ, Willett WC, Hu FB. Red meat consumption and mortality: results from 2 prospective cohort studies. Arch Intern Med. 2012 Apr 9;172(7):555-63. | link
2. van Dam RM, Willett WC, Rimm EB, Stampfer MJ, Hu FB. Dietary fat and meat intake in relation to risk of type 2 diabetes in men. Diabetes Care. 2002 Mar;25(3):417-24. | link
3. Rohrmann S, et al. Meat consumption and mortality – results from the European Prospective Investigation into Cancer and Nutrition. BMC Med. 2013 Mar 7;11:63. | link
4. Kappeler R, Eichholzer M, Rohrmann S. Meat consumption and diet quality and mortality in NHANES III. Eur J Clin Nutr. 2013 Mar 13. | link
5. Takata Y, Shu X-O, Gao Y-T, Li H, Zhang X, et al.Red Meat and Poultry Intakes and Risk of Total and Cause-Specific Mortality: Results from Cohort Studies of Chinese Adults in Shanghai. PLoS ONE. 2013 Feb 18;8(2):e56963. doi:10.1371/journal.pone.0056963 | link
Part 3 of Meat and Mortality is a review of a study from China. Not “The China Study”, but “a China study.” 🙂
If you read Part 2: NHANES when it first came out, I have updated it as of April 10 2013, to reflect that the meat intake categories were times per month not per week (only partially my fault as they listed them as “per week” in Table 2) and some further analysis making a strong argument that there was reverse causation in which people with prior disease had changed their red meat intake habits before the study began.
Now, on to Part 3: A China Study –
This prospective cohort study included 73,162 women taking part in the Shanghai Women’s Health Study and 61,128 men from the Shanghai Men’s Health Study. These numbers make it much larger than NHANES and only about 1/4 as large as EPIC.
The meat intake groups were divided into fifths, with the lowest category eating 17 g for women and 21 g for men (less than 1 oz) and the highest category eating 103 g (3.6 oz) for women and 126 g (4.4 oz) for men. The highest red meat intake categories were even less than what was seen in EPIC.
The authors summarize the results:
“[W]e found that red meat intake was positively associated with total mortality among men, but not among women. This discrepant association was also observed for lung cancer mortality. Further, red meat intake was positively associated with the risk of ischemic heart disease mortality, which was statistically significant among men. In contrast, red meat intake was inversely associated with the risk of hemorrhagic stroke mortality, which was statistically significant among women. Among men, the positive association between red meat intake and total mortality was significant in the low-income group, but no association was observed in the high-income group….”
The finding that red meat prevents hemorrhagic stroke (caused by bleeding in the brain) is not surprising, as higher rates of hemorrhagic stroke have consistently been found with low cholesterol levels. The thinking has always been that the risk of dying of a high-cholesterol-related diseases is so much greater than the risk of dying of hemorrhagic stroke that you’re better off with low cholesterol levels.
No statistically significant associations were found for poultry intake.
In addition to adjusting for the typical variables (age, smoking, alcohol, fruit and vegetable intake, etc), they adjusted the results for total caloric intake and a comorbidity index, thus raising the possiiblity that they wiped out any effect red meat might have had on mortality due to causing a higher caloric intake or previous disease.
The reason for the difference between the men and women might be that a much higher percentage of men than women smoke in China, 70% vs. 3%, which could cause confounding despite the attempt to adjust for smoking. The authors also thought that perhaps iron-deficiency in the women protected them from iron overload that could be caused by red meat.
The Chinese eat very little processed meat and the red meat they eat comes mostly from pigs (very few cows). I’m not sure if this matters, but thought it worth noting.
In summary, it’s too bad that these studies do not start out with a healthy population and do not provide a model not adjusted for conditions that red meat might effect.
Part 4 coming up…
1. Takata Y, Shu X-O, Gao Y-T, Li H, Zhang X, et al.Red Meat and Poultry Intakes and Risk of Total and Cause-Specific Mortality: Results from Cohort Studies of Chinese Adults in Shanghai. PLoS ONE. 2013 Feb 18;8(2):e56963. doi:10.1371/journal.pone.0056963 | link
The second study comes from the National Health and Nutrition Examination Survey (NHANES) III. NHANES is a nutrition survey that is done every few years in the USA to see how the nation is eating (1). For this arm of the study, about 17,500 people were included, a number dwarfed by the over 448,000 people in the EPIC study reviewed in Part 1.
I was particularly interested in this study because I thought the meat intake categories would be much higher than in EPIC. That is not the case, as the highest red meat intake category was ≥ 45 times per month, which comes to about 1.5 times per day, as compared to about 2 servings per day in EPIC.
In the model that adjusted only for age, sex and ethnicity, high red meat and processed meat consumption was significantly associated with increased total mortality (red meat: 1.57, 1.07–2.30; processed meat: 1.27, 1.06–1.54). However, in the fully adjusted model there were no statistically significant associations for red meat, processed meat, or fish consumption with total mortality.
In the fully adjusted model, they adjusted for smoking, alcohol, physical activity, socioeconomic status, marital status, fruit and vegetables, use of aspirin and ibuprofen, hormone replacement therapy, oral contraceptives, mineral and vitamin supplements, family history of diabetes and high cholesterol, and – get this – body mass index (BMI), history of hypertension, diabetes, and hypercholesterolemia (high cholesterol).
Adjusting for these last few items would remove any association with mortality for meat if meat causes death through increasing BMI, raising blood pressure, increasing diabetes risk, or increasing cholesterol – the very things considered to be how meat would cause an increase in mortality.
But even more odd is that had they run such a model, it might have shown that more red meat decreases mortality. As red meat intake went from ≤ 6 times per month to over 45 times per month, BMI went down in men and women, the percentage with a history of high blood pressure went way down in men and women, the percentage with diabetes went way down in men and slightly up in women, and the percentage with high cholesterol went way down in men and women!
In other words, according to this study, adding red meat an extra 40 times per month lowers cholesterol, blood pressure, BMI, and a man’s chance of getting diabetes.
The lowest category of red meat had a fairly large number of participants (18% for men, 28% for women), whereas eating red meat ≥ 45 times per month had very few (3.3% for men, 1.4% for women). And except for BMI, that lowest category is where there is a high level of those diseases mentioned above.
It seems to me that there must be some sort of reverse causation going on here where the people with those diseases were eating less red meat in response to having those diseases and then dying sooner. It’s interesting to note that the low red meat eaters were not likely to eat more poultry.
This underlines the benefit of starting with a population that is free of disease for a prospective study.
I have tried to steer clear of writing about papers comparing high meat intakes to low meat intakes and the association with various diseases because there is no end to the flow of these studies and I don’t find them particularly relevant to vegan nutrition. But there have been quite a few papers released on meat and mortality recently and people, including myself, are feeling a bit confused.
So I have decided to review some of these papers in blog posts, one at a time, until it no longer seems useful to continue. Here is the first one…
The European Prospective Investigation into Cancer and Nutrition (EPIC) released a study in March looking at the association with mortality of red meat, processed meat, and poultry (1). There were 448,568 participants from 10 European countries, followed for a median of 12.7 years.
The highest intake categories were ≥ 160 g (5.6 oz) per day for red meat and processed meat, and ≥ 80 g (2.8 oz) per day for poultry. A serving of meat in the U.S. is considered to be 3 oz which is about the size of a deck of cards. The comparison categories were 10 to 19.9 g for red and processed meat, and 5 to 9.9 g for poultry (in other words, not much).
They looked at the data in a number of different ways, adjusting results for age, gender, education, body weight and height, total energy intake, alcohol consumption, physical activity, smoking status and duration, and intake of the other meat intake categories.
After all was said and done, processed meat was associated fairly strongly with all-cause mortality. Processed meat intake was also associated with dying from cardiovascular disease and dying from cancer.
Red meat and poultry were neutral in the fully adjusted models, though red meat had some trends towards increased mortality and poultry had some trends towards decreased.
Adjusting for fruit and vegetable intake didn’t affect the results nor did removing the first two years of follow-up.
The body mass index (BMI) of the participants in the highest red meat category was higher than in the rest of the categories combined (27.0 vs. 24.4 for men, 24.8 vs. 22.9 for women). There was a similar difference in BMI for poultry categories. So it seems reasonable that if they had not adjusted for body weight and height there might have been more of an association for red meat.
But what would cause a higher BMI for those who eat more red meat (or poultry)? That would most likely be explained by a higher energy intake. However, they did perform a model where they removed the energy adjustment and it didn’t affect the results. That said, in the basic model adjusting only for age, gender, and study center, they found a strong association with red meat (1.37, 1.23-1.54), so I’m not fully convinced that removing the body weight/height adjustments would not have produced a strong association with increased mortality for red meat.
On the other hand, recent studies have not found such a strong association between a BMI less than 30 and increased mortality, so perhaps this isn’t a good line of reasoning. I have written the lead author to ask about this.
Why would processed meat be so much worse than red meat or poultry? According to the authors:
• Processed meat has a higher saturated fat and cholesterol content.
• Processed meat is treated by salting, curing, or smoking which leads to more carcinogens and precursors to carcinogens.
I will add that processed meat is also higher in sodium (which has been in the news quite a bit lately as being linked with increased mortality).
The authors declared no conflicts of interest.
Other than my question about the body weight adjustments, my remaining question is whether the meat intake categories in this study were too low to detect more of an effect for red meat or poultry. Certainly, there must be a point at which eating too much red meat has to be bad – one would think.
Or, on the other hand, maybe these meat intakes are entirely too high! Some will point out that, of course, people who eat a standard American diet (or European diet in this case), but a little less or a little more poultry or red meat, aren’t going to have much different mortality outcomes. But people who eat a whole-foods-only, low-fat vegan diet will definitely have better health outcomes. They just cannot be measured in a study like this.
Perhaps. Stay tuned…
1. Rohrmann S, Overvad K, Bueno-de-Mesquita HB, Jakobsen MU, Egeberg R, Tjønneland A, Nailler L, Boutron-Ruault MC, Clavel-Chapelon F, Krogh V, Palli D, Panico S, Tumino R, Ricceri F, Bergmann MM, Boeing H, Li K, Kaaks R, Khaw KT, Wareham NJ, Crowe FL, Key TJ, Naska A, Trichopoulou A, Trichopoulos D, Leenders M, Peeters PH, Engeset D, Parr CL, Skeie G, Jakszyn P, Sánchez MJ, Huerta JM, Redondo ML, Barricarte A, Amiano P, Drake I, Sonestedt E, Hallmans G, Johansson I, Fedirko V, Romieux I, Ferrari P, Norat T, Vergnaud AC, Riboli E, Linseisen AJ. Meat consumption and mortality – results from the European Prospective Investigation into Cancer and Nutrition. BMC Med. 2013 Mar 7;11:63. | link
You might have heard about the meta-analysis released January 2 in the Journal of the American Medical Association, Association of All-Cause Mortality with Overweight and Obesity (1), which found that “overweight” people, based on body mass index (BMI) had a lower risk of mortality than what is considered a healthy BMI.
Before I had a chance to read the study myself, I came across a USA Today article, Experts weigh in on the dangers of extra pounds, quoting the well-respected nutritionist, Dr. Walter Willett from Harvard School of Public Health, about the JAMA study:
“The most serious problem in the paper is that the normal-weight group included a mix of lean and active people, heavy smokers, patients with cancer (and) other conditions that cause weight loss, and frail elderly people who had lost weight due to rapidly declining health. Because the overweight and obese groups were compared to this mix of healthy and ill persons who have a very high risk of death, this led to the false conclusions that being overweight is beneficial and that grade 1 (moderate) obesity carries no extra risk. The new statistics are completely misleading for anyone interested in knowing about their optimal weight. … The paper is a pile of rubbish.”
Yikes! That’s quite a rip on the JAMA study’s authors. Not only that, but why would a research group put so much effort into a meta-analysis but fail to adjust for some of these obvious confounders, and why would JAMA publish it?! But Willett’s criticisms would explain how they could have reached such a counter-intuitive finding and all would still be good in the nutrition universe.
I read the study fully intending to find the criticisms by Dr. Willett to be valid. Alas, I did not. From my reading of the paper, they did adjust for smoking and age, and analyzed the study for any bias due to “frail elderly people.” They did some testing to make sure previous heart disease and cancer were not affecting the results.
The control groups for the studies in the meta-analyses were typically people with a BMI of from 18.5 to < 25 or from 20 to < 25. The study found that people with a BMI of 25 to < 30 had a statistically significant 8% reduced risk of mortality (.92, .88-.96). People with a BMI of 30 to < 35 had essentially the same risk of mortality. It wasn’t until you got to the group of people with a BMI of 35 or greater that risk of mortality significantly increased.
So, what could be going on here? Well, a BMI of 18.5 to 20 has often been considered to be unhealthfully thin, but from what I gleaned from the paper’s discussion, excluding such people did not appreciably affect the results.
BMI doesn’t account for muscle mass, as Willett alludes to above, but it is hard to believe that there were enough muscle-bound people in the meta-analysis to confound the results to any significant degree.
Willett went on to say in the interview:
“In the last several years, two other major analyses, involving the collaborative efforts of more than 150 scientists, have been conducted on the relation of body weight to mortality…these studies showed clearly that both overweight and all grades of obesity are associated with increased mortality.”
In contrast, the JAMA authors mention that their results are consistent with two previous meta-analyses. And they give some reasons why being overweight might be associated with lower risk of mortality:
“Possible explanations have included earlier presentation of heavier patients, greater likelihood of receiving optimal medical treatment, cardioprotective metabolic effects of increased body fat, and benefits of higher metabolic reserves.”
Occam’s razor would dictate that the best explanation is simply that it’s healthier to be “overweight,” and I don’t like doing backwards somersaults to wish away findings that, had they gone the intuitive way, would be unquestioned as solid. That said, given the prevalence of type 2 diabetes and the fact that vegans have been shown to have a much lower rate of type 2 diabetes and much lower average BMIs, I’m not going to try to gain weight. But it’s a little demoralizing to have one of the most basic ideas in nutrition ̵ that being overweight is not better than being a normal weight ̵ significantly questioned.
In conclusion: Rubbish? Not from what I can tell. Food for thought? Yes. As is often the case, this isn’t the last word.
Flegal KM, Kit BK, Orpana H, Graubard BI. Association of all-cause mortality
with overweight and obesity using standard body mass index categories: a
systematic review and meta-analysis. JAMA. 2013 Jan 2;309(1):71-82. | link
A new meta-analysis on vegetarian mortality and cancer has been released (1). It found a statistically reduced rate for vegetarians in terms of ischemic heart disease mortality and cancer incidence. It did not find a difference for all cause mortality or stroke.
“Although the 2012 meta-analysis by Huang et al. (1) is more recent, it may not be as reliable as the 1999 meta-analysis [by Key et al.] because it includes a 1984 study on Zen priests (2) who were mostly semi-vegetarian and which used a standardized mortality ratio (comparing all the Zen priests to the greater population rather than comparing the “vegetarians” to non-vegetarians within the same group). The Heidelberg Study results were also included and its control group was semi-vegetarians, which means there were semi-vegetarians in both the “vegetarian” and “non-vegetarian” group in the 2012 meta-analysis; while this is not ideal, it should have biased the results against finding a beneficial effect of a vegetarian diet. In its favor, the 2012 meta-analysis includes data from EPIC-Oxford that was not available for the 1999 meta-analysis.”
There is a table with the confidence intervals at the link above.
1. Huang T, Yang B, Zheng J, Li G, Wahlqvist ML, Li D. Cardiovascular Disease Mortality and Cancer Incidence in Vegetarians: A Meta-Analysis and Systematic Review. Ann Nutr Metab. 2012 Jun 1;60(4):233-240. (Link)
2. Ogata M, Ikeda M, Kuratsune M. Mortality among Japanese Zen priests. J Epidemiol Community Health. 1984 Jun;38(2):161-6. (Link)
A new study has been making the rounds showing that increased red meat intake is associated with an increased risk of early death. Ginny updated her article Bad news for red meat is bad news for chickens to reflect this new study (she addresses it in the final paragraph).