A “new” study once again claims butter is harmful. But does the data actually support this conclusion? And does the data apply to you?
Transcript:
Another study claims butter is unhealthy
Here we go again. Another study claims, seed oils are healthy and butter is unhealthy. The study is only a few days old and the press is running with the headlines already. And many are saying that this data reinforces what we’ve known to be true all along that butter is unhealthy. But does it? Unfortunately not. I mean this is just another example of extremely low certainty science being touted as high level evidence that meaningfully answers the question of butter versus seed oils. But that’s unfortunately far from the truth. So let’s get into the details.
The problem with nutrition epidemiology
Now if you followed our channel before, you’ve likely heard us say “The strength of the recommendation should match the strength of the evidence” or you know something similar to that effect. Well, when it comes to nutrition epidemiology studies like this one in question, the strength of the recommendation is often very strong, despite the strength of the study being very weak. So let’s talk about that. But in my opinion, the worst thing that can happen from a study like this is that, someone who eats butter or saturated fat as part of a therapeutic low-carb or keto diet that’s improving their mental and metabolic health, they’ll see a headline like this and start to worry like “Oh wait a second, like maybe the diet I’m eating isn’t healthy and maybe I need to make a change, and by making a change, maybe they lessen the therapeutic potential of their diet, right? So instead they need to realize that this study doesn’t apply to them at all and has little to no meaning. So are you curious to learn more about that? Well, I’m curious I’m excited to talk about it because it’s so important for us to understand these concepts.
Breaking down Harvard’s “new” study
So the study is called “Butter and Plant-Based Oils Intake and Mortality” and it’s from the TH Chan Harvard School of Public Health which, I don’t know to be honest seems to turn out one of the new version of this type of study every month or so but here’s the crucial point, none of these papers that they produce are new studies, they’re simply new analyses of these troves of data previously collected from the Nurses Health Study cohorts 1 and 2, and the health professionals follow up cohort. So they simply enrolled individuals in these cohorts decades ago, had them answer a dietary food frequency questionnaire which they repeat every decade or so or few years, and then they follow them to see who gets cancer or heart disease who lives and who dies. So it’s important to realize, there’s no intervention, there’s no treatment and there’s no substitution they’re just simply following them and the substitution part is important because the conclusions often say like “Oh if you just substitute seed oils for butter, you reduce your risk by this much” or “If you substitute plant protein for animal protein” but that’s all like computer modeling based on low quality evidence, it’s not based in reality and not something they did, so that’s important but let’s jump to the conclusions of this one.
Study results & conclusions
The conclusions state that eating butter is associated with an increased risk for total mortality, and eating more seed oils or what they call plant-based oils reduce or associated with the reduced mortality but here’s the key point when it comes to eating butter and the increased mortality it’s important to note when they attempt to adjust for the confounding variables, there’s no statistically increased risk except, for the absolute highest quartile of butter eaters compared to the absolute lowest and not in between. So meaning as people ate more, the risk didn’t seem to increase except at this highest level, so why is that important?
Healthy user bias
Well, for a number of reasons but one is because when you look at the characteristics of these groups the baseline characteristics, we see that those who ate the highest level of butter, also ate far more calories than the lowest butter group, 400 more calories per day, they exercise less, they smoked more, they drank more alcohol and they were less likely to take a multivitamin and interestingly these were consistent across all three cohorts and it’s what we see in every single one of these studies, produced by this group. So this is called the healthy user bias, or in this case you could call it the unhealthy user bias and we have a whole video specifically exploring this concept but here’s the short version. if you’re told eating butter is unhealthy, as we’ve been told for decades, then those more likely to eat butter are those who probably aren’t as concerned about their health. Therefore, they’re less likely to take a vitamin they’re less likely to exercise they’re more likely to smoke etc. Of course this isn’t true for everyone, but in a study like this, it’s true enough or i should say it’s true in enough people, to sway the results. So you know for the results to overcome this bias they have to be really robust, meaning a big difference between the two groups, but that’s not the case, in fact it was a minuscule result, minuscule as far as statistics go. The hazard ratio was 1.15, meaning there was a 15% increased risk of dying young or developing some disease, in this case, dying young, for those eating the most butter, compared to those eating the least. And for reference you know we can hear examples of smoking and the risk of cancer. Well, there the hazard ratio was around 20 or a 2000% increased risk. Not 15 but 2000, so those types of findings can overcome healthy user bias but not a hazard ratio of 1.5. No way!
Do these results apply to you?
So really important to wonder if these results apply to you as an individual, but remember, the level of recommendations should match the level of the evidence and no matter how you shake it in this study, this is very low quality evidence. And as we frequently point out here at Metabolic Mind, these studies, they don’t control for overall dietary quality they don’t control for the amount of carbs or fine carbs, whole foods versus ultra processed foods, snacks etc. So in my opinion, if someone’s eating a predominantly whole foods, low carb diet that contains a fair amount of butter, this study has exactly zero relevance to that individual, and i think the worst thing that can happen from these low quality studies is that someone questions their otherwise healthy low carb diet that has helped treat their bipolar disorder, their schizophrenia, their type 2 diabetes polycystic ovarian syndrome or other metabolic disorder they think well if butter is unhealthy and i’m eating butter as part of my diet, maybe i should reconsider. No, not at all! That’s where this kind of science can do far more harm than good.
What we can takeaway from this type of research
So please always ask what is the quality of the science, what’s the quality of the evidence, and how does this apply to me as an individual or for clinicians how does it apply to this one individual patient i’m seeing at this moment. What you may find is that the answer is, frequently, it doesn’t apply at all, despite the bold headlines. So I hope this was helpful and gives you another perspective on nutrition science and there are plenty of other problems with this study but I tried to highlight the most important ones as I see it so, if it was helpful, please like and subscribe so you won’t miss any of our content. Thank you for watching I’m Dr Bret Scher, and we will see you here next time at Metabolic Mind, a non-profit initiative of Baszucki Group. Please remember our channels for informational purposes only. We’re not providing individual or group medical or healthcare advice or establishing a provider patient relationship. Changing your diet, changing your medications can be dangerous if done without proper clinical supervision. So always consult your healthcare provider first.
