select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
Research, published in JAMA Internal Medicine, reports that higher intake of processed meat, unprocessed red meat, or poultry, but not fish, was significantly associated with a small increased risk of incident cardiovascular disease (CVD).
Dr Duane Mellor, Registered Dietitian and Senior Teaching Fellow, Aston Medical School, Aston University, said:
“This study adds more data to the debate about whether meat is good, bad or neutral with respect to its effect on health, in this case to looks at risk of heart disease. Unlike a previous series of publications in Annals of Internal Medicine published in September 2019 which suggested meat might not be as bad as we thought, this paper in JAMA Internal Medicine reports eating too much meat showed the opposite.
“This paper showed that higher red meat, both processed meat like sausages and unprocessed like steak are linked to a small increased risk of heart related and all cause death in a combined analysis of 6 studies. To do this the researchers took data about the food people reported that they ate varied and could have been affected by people needing to remember what they ate. Also, as food intake was only measured once, over the up to 30 years people were followed up for any changes in eating habits would have been missed. One of the stronger points of this study was they looked at absolute risk difference of less than 1% suggesting that eating some meat may only have a very tiny impact on an individual person’s health, with differences only being seen in big populations such as those used in this paper.
“It was also interesting that what this study defined as portion of red meat was at 4oz (or 120g) more than currently recommended by most governments including the UK which is around 70g per day. If the research is looked at this way, the findings could have been quite similar to the research from last year which suggested that eating the recommended amount of meat is probably safe.
“It can be difficult for people to make sense of what can seem to be conflicting messages on food. In this case eating moderate amounts of meat including red meat (less than 3oz per day) is likely to be safe, and in some groups, it can supply valuable amounts of iron and vitamin B12. However, in the interest of sustainability as well as health reducing meat intake in line with this study to the recommended less than 70g per day would be sensible. But if someone chooses to reduce meat further to do this, alternative sources of nutrients such as iron e.g. seeds, lentils, beans or dark green vegetables with a source of vitamin C such as green vegetables or citrus fruit to help its absorption is key. Although, it seems research like this seems to contradict what was written before, if these studies are looked at together, the overall message is fairly consistent we should eat meat in moderation as part of a varied and balanced diet.”
Prof Gunter Kuhnle, Professor of Nutrition and Food Science, University of Reading, said:
“This is a very well and thoroughly conducted study. It is based on 6 observational cohort studies which ensures that there is a fairly diverse study population and thus results are more representative of the general public, even though the study population is comparatively small (30,000 participants). An important limitation of the study is that diet has been assessed at one time-point only – and while this is often a good estimate of the typical diet of participants, it is likely that this has changed over time.
“The outcomes are very interesting, although not very surprising: they show an increased risk for cardio-vascular diseases (and all cause mortality) for meat (processed meat, red meat and poultry), but no association for fish intake. It is interesting that there was no beneficial association for fish intake, but that might be – as the authors explain – due to the fact that they could not distinguish between different types of fish preparation. In general, a beneficial effect could have been expected.
“While the observed effect is statistically very significant – so there is very little doubt that these findings are due to chance – it is important to emphasise that the absolute risk increase is very small – less than 2% over 30 years for two additional servings of meat per week (one serving of processed meat: 2 slices of bacon, 2 small sausages or one hot dog; one serving of meat/poultry: 110 g).
“What does this mean? It is important to keep in mind the red and processed meat are not only associated with an increased risk of CVD, but also of cancer. However, the increase in absolute risk is so small that it is unlikely to be relevant for the individual – there is no need to stop eating meat. However, on a population level, this is more important. With about 1 million people being diagnosed with heart disease every year, even a small reduction in absolute risk can have a considerable effect and reduce the number of people suffering.”
Prof Kevin McConway, Emeritus Professor of Applied, The Open University, said:
“This new study makes careful use of the right kind of statistical methods, and does an excellent job, given the inevitable limitations of this type of research. It’s very honest about its own limitations (in the section just before the final conclusions in the report). Most importantly, this study on its own can’t prove that eating greater amounts of the meats that it considers are what causes the increases in risk that it observed. It’s an observational study. The people who eat different amounts of the meats will also differ in many other ways, and perhaps these other differences are what really cause the differences in the risks of cardiovascular disease (heart attacks, strokes, etc.), and of deaths that are earlier than would otherwise have occurred. The researchers did make a good job of adjusting their findings to account for many possible differences. But, as they themselves say, this process cannot be perfect. And there are issues about how close the match is between what people say they eat and what they actually eat, particularly as these people were asked only once when they entered the study, and their diets may have changed during the average 19 years of follow-up after that.
“The researchers take care to point out that the increases in risk that they found were small. Several of the increases were statistically significant, so that it’s not easy to explain them away in terms of random differences between individuals. But statistical significance isn’t the same thing as real-world importance. Just how small were the effects?
“The researchers don’t fall into the trap of just telling us that eating a certain amount more meat will increase risk by a certain percentage, without telling us what that increase is a percentage of. They do give that kind of figure (the hazard ratios, or HRs). But they also give measures of how much the risk of cardiovascular disease, or the mortality risk, would increase in absolute terms if everyone ate two extra weekly servings of the meats they consider. (These are the figures for absolute risk difference over 30 years, or ARD.) Those absolute figures do look small. But they are perhaps difficult to make sense of, because the research report seems not to tell us what the baseline risk is – that is, what percentage of people would actually have an episode of cardiovascular disease, or die, in 30 years after they entered the study. I don’t have the data to work that out for cardiovascular disease at all. For mortality, I can give some relevant rough estimates. The average age of the participants, when they first entered the study, was about 54. Based on UK data, out of a group of 100 average people aged 54, we’d expect very close to half (about 50) to die in the next 30 years, by the age of 84. The new research tells us that, if they all ate two more servings of processed meat a week, then roughly one more of the 100 would die in that 30-year period. If they ate two more weekly servings of unprocessed meat, the observed change is risk would be slightly smaller, but it still rounds to one extra death in 100 people. Some caution is necessary here – these calculations are rough and ready, not all the people in the study were aged exactly 54, and they were in the US, not the UK. The researchers’ estimates are for the change in risk over 30 years but they come from a statistical model rather than direct observation, because adjustments were made and because very few, if any, of the participants were followed up for as much as 30 years. The researchers warn that their findings may not apply in different countries anyway. And the findings are subject to statistical uncertainty. But the changes in mortality are by any measure very small, I’d say. I haven’t given figures for the additional risk from eating more servings of poultry or fish, because the estimates were even smaller, and because (for mortality) they were not statistically significant, so may quite plausibly have been zero. (The increase in cardiovascular disease risk, associated with eating more poultry, was statistically significant, however.)
“Another way of getting a handle on the increased mortality risk is to look at what difference eating more meat would make to life expectancy – if in fact the increases in risk are caused by the extra meat (which we don’t know for sure). Roughly, if a large group of people like those in the study ate two more servings a week of processed meat for their entire adult life, then based on the figures from this study, that would decrease their life expectancy by something between 2 and 6 months. (Bear in mind that life expectancy is an average – they would not all live between 2 and 6 months less.) Eating two more weekly servings of unprocessed meat would decrease life expectancy by between 1 and 6 months. Again these are pretty small changes, given that life expectancy at birth in the UK is 79 years for men and 83 years for women.
“Although the increases in risk are small, the new research concludes that the findings “have critical public health implications”, on the grounds that diets can change, and that most people (in the US at any rate) consume the food types that were investigated on a daily or weekly basis. What they seem to be saying is that it would be possible to make a substantial difference to the number of heart attacks and strokes, and of relatively early deaths, in the US population, if the associations with meat eating really are causal and if a lot of people could be persuaded to reduce their meat consumption. This looks very different from the recommendations published last autumn in Annals of Internal Medicine (*), which suggested that adults should continue with their current levels of consumption of processed meat and of unprocessed red meat. (Those recommendations did not cover consumption of poultry or fish.) How can the conclusions from two large-scale studies be so different? Well, it isn’t because the statistical findings were different. They are slightly awkward to compare, because the new research reports the changes in risk for two extra weekly portions of meat, while the 2019 research reports changes in risk for three fewer weekly portions of meat. But, translating the findings to the same basis, they don’t differ much, with the differences generally being within the bounds of accuracy of the data. The new research report does agree that its findings “are comparable to those reported in the literature”, referencing the very comprehensive systematic review (**) on which the 2019 Annals recommendations were based.
“So the differences in conclusions must relate to differences of opinion between the researchers on how important the small, statistically significant, differences in risk are in the real world. The researchers on the new study clearly think that they definitely are very important. The researchers on the 2019 study clearly did not. They did point out that their conclusions are based on low-certainty evidence and they say they are ‘weak recommendations’, but there are problems with the reliability of the evidence for the new study too, since it uses broadly the same methods as the research that was the basis for the 2019 advice.
“I think that a major reason for the differing conclusions is that the 2019 Annals researchers specifically stated that “The target audience for our guidance statement was individuals who consume unprocessed red meat or processed meat as part of their diet. The panel took the perspective of individual decision making rather than a public health perspective.” In other words, they took the line that the risk differences were small, and that many individuals who eat meat would consider that such small differences in disease risk are unimportant when looked at beside the enjoyment they get from what they eat or the awkwardness of changing eating habits. The researchers on the new study are taking a public health perspective; they note that people can choose to eat less meat, and if they do so and the relationship between meat eating and disease risk is indeed causal, then fewer people would have heart attacks and strokes, and on average people would live a bit longer. Individual trade-offs between small risk increases and personal preferences don’t really come into that view – but it is individuals who decide how much meat they are going to eat, if any. (And of course, none of this takes into account animal welfare, or environmental effects of livestock raising.)”
* Johnston et al. (2019) ‘Unprocessed Red Meat and Processed Meat Consumption: Dietary Guideline Recommendations From the Nutritional Recommendations (NutriRECS) Consortium’. Ann Intern Med. 2019;171(10):756-764.
** Zeraatkar et al. (2019) ‘Red and Processed Meat Consumption and Risk for All-Cause Mortality and Cardiometabolic Outcomes: A Systematic Review and Meta-analysis of Cohort Studies’. Ann Intern Med. 2019;171(10):703-710.
‘Associations of Processed Meat, Unprocessed Red Meat, Poultry, or Fish Intake With Incident Cardiovascular Disease and All-Cause Mortality’ by Victor W Zhong et al. was published in JAMA Internal Medicine at 16:00 UK time on Monday 3rd February.
DOI: doi:10.1001/jamainternmed.2019.6969
Declared interests
Dr Duane Mellor: No financial declarations of interest. I have worked on an unpaid basis with the Vegan Society
Prof Gunter Kuhnle: No interest to declare.
Prof Kevin McConway: Prof McConway is a member of the SMC Advisory Committee, but his quote above is in his capacity as a professional statistician.