Latest / Hardiness with Dr Paul Taylor / Wisdom Wednesdays: Why Lots Of Nutritional Research Just Cannot Be Trusted

Transcript

Wisdom Wednesdays: Why Lots Of Nutritional Research Just Cannot Be Trusted · Hardiness with Dr Paul Taylor

• 0:08
  Speaker 1: Hey, everybody, Welcome to another edition of Wisdom Wednesdays. Today,
• 0:13
  Speaker 1: I want to have a chat about epidemiological research, especially
• 0:18
  Speaker 1: in nutrition, and highlight the limitations of it by looking
• 0:23
  Speaker 1: at a study that was just released. But before we
• 0:25
  Speaker 1: get into that study, and there's a guy called Professor
• 0:28
  Speaker 1: John Eonidis who's at Stanford University. He's a professor there
• 0:33
  Speaker 1: and he is the most cited research scientist in the world.
• 0:37
  Speaker 1: So this guy is a brilliant research scientist. And he's
• 0:41
  Speaker 1: actually written a paper, and I think it's about a
• 0:44
  Speaker 1: book as well, called why most published research Findings are False?
• 0:50
  Speaker 1: And he highlights in that a whole heap of issues
• 0:54
  Speaker 1: such as what's called pe hacking, where you do a
• 0:56
  Speaker 1: whole heap of statistics and then you dive in and
• 0:59
  Speaker 1: try and find the one that are significant and then
• 1:01
  Speaker 1: you talk about them that that's how your study has
• 1:04
  Speaker 1: been designed. He also talks about conflicts of interest, and
• 1:07
  Speaker 1: I've talked about conflicts of interest in research and especially
• 1:11
  Speaker 1: in nutrition. You know, we've got conflicts of interest from
• 1:14
  Speaker 1: the dirty industry, from the meat industry sponsoring studies, and
• 1:18
  Speaker 1: now a whole heap on the other side from vegans
• 1:22
  Speaker 1: and sponsoring studies and real selective reporting, and all of
• 1:27
  Speaker 1: this can lead to incorrect, incorrect conclusions being drawn by
• 1:31
  Speaker 1: the data. Right now, he has also said, and I
• 1:35
  Speaker 1: love this quote, he said, we must recognize that nutritional
• 1:41
  Speaker 1: epidemiology is a dead horse and we need to bury
• 1:46
  Speaker 1: the carcass.
• 1:47
  Speaker 2: Right.
• 1:47
  Speaker 1: So what he means by that is observational studies in
• 1:53
  Speaker 1: every part of science apart from nutrition. Really we use
• 1:57
  Speaker 1: observational studies to generate highypothesis to then go and test hypothesis.
• 2:03
  Speaker 1: But in the world of nutritional science, we seem to
• 2:06
  Speaker 1: be using observational studies to actually inform public health. And
• 2:12
  Speaker 1: the worst one was in the nineteen seventies when there
• 2:14
  Speaker 1: was this whole association with cholesterol and heart disease, and
• 2:18
  Speaker 1: then this war on fat, which I think has then
• 2:21
  Speaker 1: spawned a whole heap of issues because we've switched to
• 2:25
  Speaker 1: high carbohydrate diets and a lot of it has been
• 2:28
  Speaker 1: ultra processed foods. Right now, that's an issue in and
• 2:31
  Speaker 1: of itself, But let's get back and examine those statements
• 2:35
  Speaker 1: through the lens of a very recent study that's just
• 2:38
  Speaker 1: been published. The title of this study is Meat Consumption
• 2:43
  Speaker 1: and Incident Type two diabetes and individual participant federated meta
• 2:48
  Speaker 1: analysis at one point ninety seven sorry million adults with
• 2:53
  Speaker 1: one hundred thousand incident cases from thirty one cohorts in
• 2:57
  Speaker 1: twenty countries, and it's published in Lancet Diabetes Indochronology. So
• 3:01
  Speaker 1: that's a very good publication and it's a big meta analysis.
• 3:07
  Speaker 1: So what they did, let's have a look at it.
• 3:11
  Speaker 1: And we've actually we've been here before, right, where these
• 3:13
  Speaker 1: observational studies are made, claims are made about the health risks,
• 3:18
  Speaker 1: and in this case of eating meat, the media gets
• 3:20
  Speaker 1: hold of it and then suddenly everybody's talking about it, right,
• 3:25
  Speaker 1: And so these findings were quickly picked up by the
• 3:28
  Speaker 1: New York Times, and now they are in lots of
• 3:31
  Speaker 1: different newspapers, and I've read them in a different one, right.
• 3:34
  Speaker 2: But let's have a look at the study.
• 3:36
  Speaker 1: So what they did was they set out to investigate
• 3:38
  Speaker 1: a potential link between meat consumption and diabetes on a
• 3:41
  Speaker 1: global scale. So they conducted a meta analysis. So that's
• 3:45
  Speaker 1: a study of studies. Here they pull all the data
• 3:48
  Speaker 1: and they use data from nearly two million adults across
• 3:51
  Speaker 1: thirty one independent groups or cohorts as part of a
• 3:55
  Speaker 1: big project called the Interconnect Project, which is an international
• 3:58
  Speaker 1: research too explore genetic and environmental risk factors for obesity
• 4:03
  Speaker 1: and diabetes. So this study spanned populations from America, Europe,
• 4:09
  Speaker 1: Asia and beyond to give a broad perspective on global health,
• 4:14
  Speaker 1: eating habits and health outcomes.
• 4:16
  Speaker 2: So the diets were self reported, which.
• 4:18
  Speaker 1: Is an issue in and of itself which I've talked
• 4:20
  Speaker 1: about previously. But then meat consumption was divided into three
• 4:23
  Speaker 1: categories unprocessed red meat, processed meat, and poultry. And then
• 4:28
  Speaker 1: they looked at the cases of type two diabetes. They
• 4:31
  Speaker 1: were tracked using medical records and diabetes medication use and
• 4:35
  Speaker 1: other self reports, and then validated with additional checks like
• 4:41
  Speaker 1: prescriptions and people's HbA one sea level.
• 4:44
  Speaker 2: So they didn't just ask them about it.
• 4:45
  Speaker 1: They went and looked at some of the data and
• 4:47
  Speaker 1: anybody who had diabetes at the start of the study
• 4:50
  Speaker 1: was excluded.
• 4:51
  Speaker 2: So they were.
• 4:51
  Speaker 1: Looking at the impact of meat consumption on driving type
• 4:56
  Speaker 1: two diabetes. And over a median period follow up of
• 5:00
  Speaker 1: ten years, there was more than one hundred and seven
• 5:04
  Speaker 1: thousand cases of type two diabetes were recorded. So these
• 5:07
  Speaker 1: are new cases studying people over time, and they found
• 5:10
  Speaker 1: in the study that increased intake of all three types
• 5:13
  Speaker 1: of meat, red meat, process meat, and poultry was associated
• 5:18
  Speaker 1: with a statistically significant increase in diabetes risk, even after
• 5:22
  Speaker 1: adjusting for factors like demographics, smoking, BMI, alcohol consumption, physical activity,
• 5:29
  Speaker 1: and other dietary habits. And I've talked about this before
• 5:31
  Speaker 1: where they try to use statistics to control for these
• 5:34
  Speaker 1: confounding factors, and processmate showed the strongest association, with a
• 5:40
  Speaker 1: fifteen percent increase risk for every fifty grams of processed
• 5:45
  Speaker 1: meat per day consumed, and the next was red meat,
• 5:48
  Speaker 1: and then the next was poultry. They had both smaller
• 5:51
  Speaker 1: but still significant increased risk of developing diabetes. Now, this
• 5:56
  Speaker 1: is where we need to hit pause and consider one
• 5:59
  Speaker 1: of the most important lessons in epidemiological research and especially
• 6:04
  Speaker 1: in nutrition. Correlation does not equal causation. So, for an
• 6:09
  Speaker 1: example on this ice cream sales are much higher on
• 6:13
  Speaker 1: hot days. So are robberies because people tend to be
• 6:16
  Speaker 1: out and they get robbed more. But that's a correlation,
• 6:21
  Speaker 1: is the increased incidents about ice cream sales and robberies.
• 6:26
  Speaker 1: That doesn't mean that ice cream sales cause robberies, right, So.
• 6:29
  Speaker 2: That's important distinction to know.
• 6:32
  Speaker 1: Now, these results in this study, they show an association
• 6:35
  Speaker 1: between red meat intake and the instant diabetes, but they
• 6:38
  Speaker 1: don't actually tell us if med meat consumption actually causes diabeties.
• 6:42
  Speaker 1: And I'm going to come back to this later because
• 6:44
  Speaker 1: it's really important. This distinction is critical. Just because two
• 6:48
  Speaker 1: things are linked doesn't mean that one causes the other.
• 6:52
  Speaker 1: So in this world of diet and disease, there's countless
• 6:55
  Speaker 1: variables at play that could influence that both meat into
• 7:00
  Speaker 1: and diabetes risks. So things like socioeconomic status, access to healthcare,
• 7:05
  Speaker 1: other datary habits, lifestyle choices, they all play significant roles
• 7:09
  Speaker 1: whether or not you exercise or smoke. And we know
• 7:11
  Speaker 1: that people with higher process, especially meat intake, they are
• 7:16
  Speaker 1: often lower income because it's cheaper, so the elite that
• 7:18
  Speaker 1: stuff they often smoke more. They often had less access
• 7:23
  Speaker 1: to healthcare and less opportunity to engage in health promoting behaviors.
• 7:27
  Speaker 1: And we see this over and over again that lower
• 7:30
  Speaker 1: socioeconomic status people have got higher diseases across the board. Now,
• 7:36
  Speaker 1: this study did try to adjust for some of these,
• 7:38
  Speaker 1: but it is impossible to account for those things perfectly.
• 7:43
  Speaker 1: And if we dive in a little bit and look
• 7:46
  Speaker 1: at they adjusted for body mass index BMI in the study,
• 7:50
  Speaker 1: and when they included BMI in the statistical model, the
• 7:54
  Speaker 1: risk associated with meat consumption was significantly reduced by half
• 7:58
  Speaker 1: pretty much. So this on its own suggests that factors
• 8:02
  Speaker 1: that are related to general health rather than meat consumption
• 8:06
  Speaker 1: per se, might be driving the association with diabetes. I mean,
• 8:10
  Speaker 1: think about it makes sense if you've got a higher BMI,
• 8:12
  Speaker 1: that's a known risk factor for type two diabetes.
• 8:15
  Speaker 2: And if people who eat more meat also.
• 8:17
  Speaker 1: Tend to hire BMI, then it's not necessarily the meat
• 8:20
  Speaker 1: that's doing it, but it can be the BMI or
• 8:22
  Speaker 1: their overall health profile. So another point to note on
• 8:27
  Speaker 1: this is the inconsistency of findings across different populations. So
• 8:31
  Speaker 1: you dig into the paper. There was no significant association
• 8:35
  Speaker 1: between meat consumption and type two diabetes in the people
• 8:38
  Speaker 1: in the Eastern Mediterranean and people in Southeast Asia, and
• 8:42
  Speaker 1: the link between poultry intaken diabetes was only significant in
• 8:46
  Speaker 1: the European cohorts but not anywhere else. So why does
• 8:50
  Speaker 1: poultry cause increased diabetes in Europe but not anywhere else?
• 8:54
  Speaker 1: And why does red meat and particularly processed meat cause
• 8:58
  Speaker 1: increased risk of diabetes in some places, but not an
• 9:01
  Speaker 1: Eastern Mediterranea and not in Usia. So here we go again.
• 9:06
  Speaker 1: This is this observational study and a wave of media hype.
• 9:09
  Speaker 2: Now let's then follow this and think through it logically.
• 9:14
  Speaker 1: So what I did is I kind of tested the
• 9:20
  Speaker 1: kind of assumptions here, right, So if this observational study
• 9:24
  Speaker 1: is true and red meat does cause diabetes, then what
• 9:29
  Speaker 1: we should see, and this is the way that science
• 9:31
  Speaker 1: should work, is that is a hypothesis that we then
• 9:34
  Speaker 1: go and test. And the easy way to test that
• 9:37
  Speaker 1: hypothesis is to get a bunch of people and get
• 9:41
  Speaker 1: some of them to reduce their risk of red meat
• 9:44
  Speaker 1: consumption and process meat and then look and see if
• 9:48
  Speaker 1: they have less diabetes, then the similar amount of.
• 9:52
  Speaker 2: People who don't reduce theirs.
• 9:54
  Speaker 1: So what I did was I jumped onto illicit dot com.
• 9:57
  Speaker 1: So this is a research to use by researchers, and
• 10:00
  Speaker 1: you type in a research question, right, and I typed
• 10:04
  Speaker 1: into question what is the evidence that reducing red meat
• 10:07
  Speaker 1: consumption can manage diabetes?
• 10:09
  Speaker 2: Please find intervention studies. And what the illicit.
• 10:12
  Speaker 1: Does is then will find the top eight papers in
• 10:16
  Speaker 1: the field, and so I looked at these top eight
• 10:19
  Speaker 1: papers and I'll actually put links in them. Well, particularly
• 10:23
  Speaker 1: I'll put links to the ones that were in intervention
• 10:25
  Speaker 1: study to say if reducing it, because some of them
• 10:28
  Speaker 1: were observational studies. And let's talk about the first one,
• 10:31
  Speaker 1: potential effects of red meat. Reduced red meat compared with
• 10:34
  Speaker 1: increased fiber intake on glucose metabolism and liver fat content
• 10:38
  Speaker 1: a randomized and controlled atary intervention studies. So this is
• 10:41
  Speaker 1: the sort of study we're after. And it found caloric restriction,
• 10:45
  Speaker 1: not reduced red meat or increased fiber led to improved
• 10:49
  Speaker 1: glucose metabolism and reduce liver fat in subjects at risk
• 10:52
  Speaker 1: of type two diabetes. So reducing red meat did nothing.
• 10:56
  Speaker 1: And then the next one, modern consumption of red meat
• 10:58
  Speaker 1: compared to soy or non soi leg gloom has no
• 11:01
  Speaker 1: adverse effect on cardio metabolic factors in patients with TAPE
• 11:05
  Speaker 1: two diabetes. They found moderate consumption of red meat has
• 11:08
  Speaker 1: no adverse effect on cardio metabolic factors compared to soy
• 11:11
  Speaker 1: or non soi legooms in patients with type two diabetes.
• 11:14
  Speaker 1: So they were comparing an intervention. Then they looked here
• 11:18
  Speaker 1: and another one faculty opinions recommended of effect of lower
• 11:23
  Speaker 1: versus higher red meat intake on cardio metabolic and cancer outcomes.
• 11:27
  Speaker 1: A systematic review of randomized trials. Remember cardio metabolic includes diabetes.
• 11:31
  Speaker 1: They found load of very low certainty evidence suggests that
• 11:34
  Speaker 1: reducing red meat intake may have little or no effect
• 11:37
  Speaker 1: on cardio metabolic and cancer outcomes. And then there's another one,
• 11:42
  Speaker 1: red meat, dirty and insulin sensitivity or randomized crossover intervention study.
• 11:46
  Speaker 1: This is the really good study of randomized one. We
• 11:49
  Speaker 1: then cross people over to the different arms. They find
• 11:53
  Speaker 1: a diet high in lean red meat improved insulin sensitivity
• 11:57
  Speaker 1: compared to high dirty diet in overweight and obese subjects.
• 12:02
  Speaker 1: And then another one substituting lin beef for carbohydrate in
• 12:05
  Speaker 1: a healthy dietary pattern does not adversely affect the cardio
• 12:09
  Speaker 1: metabolic risk factor of profile of men and women at
• 12:13
  Speaker 1: risk for tapto biobetes. They found that actually taking out
• 12:17
  Speaker 1: carbohydrate and putting in beef actually has no adverse effect
• 12:21
  Speaker 1: on cardio metabolic risk factors and the risk of developing
• 12:25
  Speaker 1: type two diabetes. And then the last one, effects of
• 12:28
  Speaker 1: total red meat intake on glycemic control and inflammatory biomarkers,
• 12:31
  Speaker 1: and meta analysis of randomized control trials. Total red meat
• 12:36
  Speaker 1: consumption for up to sixteen weeks does not affect biomarkers
• 12:40
  Speaker 1: of glycemic control or inflammation in adults at risk of
• 12:44
  Speaker 1: cardio metabolic disease. So all of this taken together, suggests
• 12:51
  Speaker 1: that John Ionides is absolutely right, and we need to
• 12:55
  Speaker 1: realize that observational nutritional studies or a dead horse.
• 13:01
  Speaker 2: And we need to bury the corpse. Now.
• 13:02
  Speaker 1: I do just want to call it one little caveat
• 13:05
  Speaker 1: on that, because I have reported on observational studies, particularly
• 13:08
  Speaker 1: stuff from the UK Biobank. I've done a number of
• 13:11
  Speaker 1: things looking at risks of devention and stuff like that,
• 13:14
  Speaker 1: and exercise and strength. When we see observational studies that
• 13:19
  Speaker 1: are consistent that consistently so the same thing in different populations,
• 13:25
  Speaker 1: and there is a plausible biological mechanism, then you can
• 13:30
  Speaker 1: start to say, okay, well this is actually given us
• 13:33
  Speaker 1: some reasonable evidence that we then need to go and
• 13:35
  Speaker 1: test further with the randomized control trial.
• 13:38
  Speaker 2: And it's not.
• 13:39
  Speaker 1: Always possible to do a randomized control trial, but we
• 13:43
  Speaker 1: need to realize that you can't just take one study
• 13:48
  Speaker 1: and then have all of these headlines around it, and
• 13:50
  Speaker 1: especially have people influence or change their diets based on
• 13:56
  Speaker 1: one or a couple of pretty poor observational stuf allies.
• 14:00
  Speaker 2: That's it for this week, folks, Catch you next time.