With the release of the long-awaited debate between myself and Tucker Goodrich (linked below) about seed oils and their impact on coronary heart disease (CHD) risk, I wanted to review the debate and discuss what went well and what could’ve gone better from my perspective, as well as highlight some of the apparent inconsistencies in Tucker’s position, which have become more visible to me after reviewing the debate. In addition to that, I think it is a good opportunity to share some background on why I decided to partake in this debate and how it came to fruition. Of course, I want to note that it was a cordial discussion, all communications prior to the debate were amicable, and both Tucker and I agree that Simon did a good job as a moderator, which you can hear us discuss at the end of the debate (3:39:45). As this point-by-point review is quite extensive, I have also written a “rapid review,” included at the beginning of this article, summarizing the main points for those who would prefer not to read the whole article (I don’t blame you). As a last note, you will notice me saying “yeah” while trying to interject at various points throughout the debate. This is a habit of mine in which I often say “yeah, but…” or “okay, but…” when interjecting and trying to respond. This is not me agreeing to the claim that preceded it. With that out of the way, let’s dive in.
Rapid Review/Highlights (Scroll past or click here to read the full, timestamped, point-by-point review)
Given that Tucker is well known in the anti-seed oil community, and I believe the best available evidence suggests that seed oils are health promoting, especially when replacing foods high in saturated fat, such as butter, I reached out via Twitter to ask if he was willing to publicly debate the topic. This allows those who may be influenced by his (or my) content to evaluate both sides of the argument when making decisions that can impact their health. We agreed to have Simon Hill as our moderator and upon the following debate parameters.
Debate Parameters
Parameter | Description |
Debate Proposition | It is more reasonable to believe that seed oils are beneficial, rather than harmful for CHD risk.
Matt’s position: Agree Tucker’s position: Disagree |
Debate Structure | Introductions: 5 minutes each
Body of the debate: Conversational Closing statements: 2 minutes each |
References | All references that we’d like to use during the debate must be shared in advance.
Primary references: March 8th Rebuttal references: March 18th |
Rule #1 | Questions are to be answered directly. No dodging questions or changing topics. |
Rule #2 | No interruptions except to clarify your position in cases where you believe your position may be misrepresented.
Note: In the case of either of us speaking for several minutes, the moderator may interject to allow for rebuttal. |
Rule #3 | No personal attacks/insults. |
Role of the Moderator | To introduce the debate, facilitate discussion, and ensure the aforementioned debate structure and rules are followed. |
For this rapid review, I will outline some key points of discussion and major inconsistencies that arose throughout the debate, but if you’d like to read a more detailed analysis with timestamps, you can read the full-length review below.
Nutritional Epidemiology – 23:55
After our introductory statements, Tucker targeted nutritional epidemiology and raised issues around some of the methodologies used. The problem with his position, as I pointed out in the debate, is that regardless of any issues he can raise, the fact of the matter is that the results from nutritional epidemiology agree with the results from randomized controlled trials in the majority of cases. I showcase multiple analyses where the concordance rate is around 65-67% (or higher) and when I simply asked if that is a majority, he refuses to answer, breaking the rules we agreed upon for the debate. It’s important to note that my arguments don’t hinge on nutritional epidemiology either. In fact, my strongest arguments were centred on randomized controlled trials, but I wanted to tackle his claim regardless.
Minnesota Coronary Experiment (MCE) – 36:08
MCE is one of the studies that Tucker holds in the highest regard, in part due to its large sample size, but there are several issues I raise with it. The first issue I raise is that the vast majority of participants dropped out leaving them with a much smaller sample size than they calculated they would need to detect an effect of the intervention. Tucker claimed that this doesn’t matter and cited non-statistically significant results that favour his position, while discounting non-statistically significant results that do not favour his position later in the debate. When I presented the data of only those individuals who stuck with the diet long-term, in which case those consuming the corn oil and corn oil margarine may have done even better than the control group, he claims that’s cherry picking. I’m still at a loss as to how looking at those who actually adhered to an intervention for more than a couple years is considered cherry picking if the goal is to detect an effect of the intervention on long-term health outcomes.
Throughout this section of the debate, he dodged questions and attempted to change topics multiple times, breaking the rules we set in advance. My final criticism for the study was the possibility of confounding due to trans fats in the margarine. He handwaved this and perhaps most astonishingly, later speculates that trans fats could be healthier than linoleic acid (omega-6 fat). This was truly absurd given the consistency and strength of the data on trans fats and health outcomes.
LA Veterans Administration Hospital Study – 57:01
Perhaps the strongest study I had on my side, the LA Veterans trial demonstrated what I would consider to be a large benefit (33%) for risk of death due to atherosclerotic cardiovascular disease. To combat this study, Tucker suggests it was a “blended intervention” because ALA (omega-3) intake also slightly increased. The issue with that argument is that this debate isn’t about isolated linoleic acid (LA). It’s about seed oils, and if the ALA is coming from seed oils, this study is fair game. This is an argument he continues to bring up throughout the debate even though it’s clearly irrelevant to the debate proposition. He also handwaves the reduction in risk and calls it a “slight” benefit, while simultaneously believing that the results from MCE are worrisome even though the magnitude of effect is far lower. This is a double standard.
Tucker attempted to formulate other arguments, but they all fell flat, and he broke more of the debate rules in the process. In addition, he did eventually concede that they showed a “slight benefit,” even though he doesn’t apply the same standards for what’s considered “slight” to his own studies. Regardless, this concession favours my position in this debate.
Lyon Diet Heart Study – 1:22:04
In addition to MCE, Tucker believes the Lyon Diet Heart Study is one of the strongest studies supporting his position, but there’s a major issue. There was a tiny reduction in LA intake relative to the increase in LA intake we saw in the LA Veterans trial, yet he thinks this reduction is largely responsible for the large reduction in risk of cardiac events. Not only was the change in LA intake small, but they modified numerous other dietary factors that would be expected to impact risk. One of Tucker’s arguments when I brought this to his attention was that they saw a larger reduction in risk than has been seen in any of the trials that increased LA intake or modified other dietary variables. The issue with this logic is that we shouldn’t expect a single dietary change to result in as large of an improvement as multiple simultaneous dietary changes. Not to mention that these various trials use different types of participants (age, CVD history, etc.).
Mechanisms – 1:33:22
We ended up having a rather lengthy discussion about the mechanisms behind atherosclerotic cardiovascular disease. I don’t put much stock in mechanisms at all given their low rate of translation into actual clinical outcomes, which is what matters; however, I indulged. I simply pointed out how Tucker’s own references actually don’t support his claims as it appears he was selectively reporting only the parts of those publications that speak to LDL oxidation, which he believes is the root cause. I agree that it is involved in the process, but it is not the initiating factor as exemplified by his own sources.
LA Veterans…Again – 2:06:54
Tucker suggested that one of the reasons we see such a striking benefit in the Lyon Diet Heart Study, which reduced LA intake, but don’t see a similar magnitude of harm in other trials was because there’s a threshold at which there is no further harm. He suggested that this threshold is around 5% of total calories, but I pointed out that the LA Veterans trial actually crossed that threshold, massively increasing LA intake, yet results in a reduction in risk. Tucker resorted to many of the same arguments he tried to raise earlier, and even brought up a truly bizzare argument, suggesting that corn oil isn’t a seed oil, even though its fatty acid content is very comparable to other seed oils, and even higher in LA content than many of them. I think he was just trying to grasp at straws, because if he truly held the position that corn oil wasn’t relevant to the debate, then he would need to toss out 2 out of 4 of his favourite trials: MCE and Rose Corn Oil.
As the discussion of LA Veterans continued, he broke more debate rules and even suggested that reducing CVD deaths by 33% in America (potentially 282,885 lives saved per year) wouldn’t be a significant benefit. Regardless, Tucker does end up acknowledging that LA Veterans did find an overall benefit .
Hooper Meta-Analysis – 2:16:05
Perhaps even more absurd than his last comments about the LA Veterans trial, when he transitioned the discussion to the 2020 Hooper meta-analysis I cited, he suggested that he doesn’t think a reduction in total cardiac events is important, and that it’s only important if it kills you. I found this to be quite insensitive given how damaging a non-fatal heart attack or stroke can be to one’s physical and mental health, as well as the burden it can place on those who care for that individual. I truly hope he doesn’t actually hold this position and was simply trying whatever he could, as ridiculous as it might be, to avoid admitting the interventions were beneficial.
At this point, Tucker ranted for an extended period of time and didn’t allow myself, or even Simon to interject, so I have addressed all the claims below in the full review. Ultimately, most of the claims he made here are tangential to what we had been discussing throughout the debate to this point anyway.
Ecological Data – 2:38:16
As Tucker had repeatedly cited ecological associations (ie. seed oil consumption and CVD have both increased over the past 100 years) to support his claims that the rise in CVD rates is due to the increase in seed oil consumption, I pointed out issues with this logic. For one, there are no adjustments made for confounders in that sort of analysis, and second, we could also use that logic to state that cigarette consumption is associated with longevity given that countries with higher per capita cigarette consumption also have higher average life expectancies. During this discussion, Tucker says that ecological data is the worst type of epidemiology (which he denounces), yet is “critical” at the same time. Quite the inconsistency.
Sydney Diet Heart Study – 2:42:03
While the Sydney Diet Heart Study design is pretty solid overall, the one big issue is that the margarine they used contained trans fats at the time. Given how harmful trans fats are, this made the margarine/seed oil intervention look worse than the control. Tucker argued that because I don’t know the exact composition of specific samples of the margarine used at that time, that I can’t claim they contained trans fats; however, we know through subsequent analyses that the methods used would produce trans fats, and the trans fat content of the margarine used at that time would be about 15%.
Olive Oil vs. Vegetable Oils – 2:47:22
There was one epidemiological study that both Tucker and I cited in our reference lists, and honestly I was pretty shocked to see him include it given that the results don’t suppor this position at all. They ultimately found that swapping olive oil for vegetable oils is a relatively neutral shift, and since we both agree olive oil is beneficial, surely that would mean he believes vegetable oils are similar beneficial, no? Well, he completely misinterpreted the study and didn’t appear to understand what or how adjustments for confounding variables are made. In fact, he seemed upset that they excluded polyunsaturated fat from their adjustment model, but that is absolutely the right thing to do if you’re trying to determine the total effect of a food that is high in polyunsaturated fats (ie. vegetable oils). Throughout this discussion, Tucker refused to answer questions and tried to change topics, but either Simon or I brought it back to the topic at hand in which case he had no good explanation for the results, even resorting to claiming that mayonnaise and isolated vegetable oils are essentially the same thing (but corn oil apparently isn’t according to his earlier comments) to avoid acknowledging the results.
Conclusion
Ultimately, I’m happy with how the debate went. Were there instances where I could’ve made a stronger argument or rebutted certain claims that Tucker made? Sure, but overall, I believe I provided sound arguments based on multiple converging lines of evidence (RCTs, epidemiology, mendelian randomization data, etc.) while providing valid criticisms for the majority of Tucker’s arguments, oftentimes by citing his own research. On the contrary, Tucker largely failed to criticize the strongest data I had in my corner, such as the LA Veterans trial, the meta-analyses, and the epidemiology, and most of his attempted criticisms revealed double standards that apply, sometimes to an even greater degree, to his own studies. In addition, he broke the various rules we had set dozens of times, most often to avoid answering questions that may require answers that are unfavourable to his position. I hope that those of you who watched the debate can see that, as my overall goal wasn’t necessarily to change Tucker’s mind, but to bring clarity to those of you who may be on the fence, or even more of the belief that seed oils are harmful. I too used to believe that, but upon evaluating the best available evidence through a more objective lens, I changed my view, and I hope you can too. As last note, I should mention that Simon has asked both Tucker and I to sit down for a part 2 of the debate, but we have yet to receive any availability from Tucker. Should Tucker finally accept Simon’s offer, I will gladly participate.
Full Review
Background
Seed oils (also often referred to as vegetable oils), such as canola, soybean, corn, and safflower oil, are heavily demonized by subsets of the online nutrition space, despite being recommended by most nutrition guidelines as heart healthy options. In fact, there is a very vocal crowd suggesting that seed oils are not only harmful for cardiovascular risk, but many of the chronic illnesses that are affecting us today. Tucker is one such seed oil critic, and since those conveying that message can influence others’ nutrition choices in ways that the best available evidence suggests could be harmful, such as swapping out seed oils for high saturated fat options like butter, I think it’s important to address.
A few months ago, a friend reached out to me and asked if I’d be willing to debate Tucker about the topic given his prominence in this space. Having had a few nutrition debates in the past, I thought it would be a good idea to have a conversation with him so that both his audience and mine could hear both sides of the story rather than living in the echo chambers that are so easy to create on social media. As such, I tweeted at Tucker and offered to have a debate, to which he agreed. It took a fair amount of time to organize because of scheduling conflicts between myself, Tucker, and the moderator he had originally proposed, so we asked Simon if he’d be willing and available to moderate on some proposed dates and he agreed. From that point on, we just had to sort out the finer details.
The agreed upon debate proposition was “it is more reasonable to believe that seed oils are beneficial, rather than harmful for CHD risk.” My position is that I agree with that statement while Tucker disagrees. The reason for honing in on a specific outcome is that we simply couldn’t cover everything in a single debate. This debate alone lasted nearly 4 hours, but I would be happy to have a subsequent debate with Tucker, should he want to cover other topics, although we should probably first have a follow-up discussion on this same topic (more on this later). We also agreed to share all of our primary references 2 weeks in advance (March 8th) and allowed for subsequent references for possible rebuttals to be shared up to 4 days before the debate (March 18). Tucker did originally suggest a shorter window; however, I wanted to ensure we had an adequate amount of time to review each reference in depth, so Tucker agreed to extend it to the dates mentioned above. However, he did end up sending several references less than 24 hours before the debate, including the morning of, in which case I surprisingly did manage to get through in time and responded with 1 additional reference of my own given that he appeared to be okay with that. We also decided that we would each have 5 minutes to make opening statements, 2 minutes to make closing statements, and that the bulk of the debate would be conversational. In addition, we both agreed that questions were to be answered directly without dodging or changing topics, and we made a rule against interruptions unless it were the case that one of us believed we were being misrepresented or one of us has been speaking for several minutes uninterrupted, in which case Simon as the moderator could step in to allow for rebuttal. All parties also agreed that the role of Simon as the moderator is to introduce the topic, step in to direct the conversation when we’re at a stand still or the above rules are broken, but otherwise to remain on the sideline and allow Tucker and I to converse.
Debate Parameters
Parameter | Description |
Debate Proposition | It is more reasonable to believe that seed oils are beneficial, rather than harmful for CHD risk.
Matt’s position: Agree Tucker’s position: Disagree |
Debate Structure | Introductions: 5 minutes each
Body of the debate: Conversational Closing statements: 2 minutes each |
References | All references that we’d like to use during the debate must be shared in advance.
Primary references: March 8th Rebuttal references: March 18th |
Rule #1 | Questions are to be answered directly. No dodging questions or changing topics. |
Rule #2 | No interruptions except to clarify your position in cases where you believe your position may be misrepresented.
Note: In the case of either of us speaking for several minutes, the moderator may interject to allow for rebuttal. |
Rule #3 | No personal attacks/insults. |
Role of the Moderator | To introduce the debate, facilitate discussion, and ensure the aforementioned debate structure and rules are followed. |
As a last note before reviewing the debate itself, I should mention that Simon reached out to both Tucker and I shortly after the recording of the debate to ask if we would be willing to sit down for a shorter follow-up discussion to hopefully resolve some specific points of discussion that Simon believes were left unsettled and may confuse viewers. We both said we’d be happy to, but after numerous attempts to schedule a time, Tucker has not provided his availability or even responded to many of our tweets, emails, or messages. At this point, he has suggested that he is willing to sit down for a follow up after the release of the debate, so hopefully we can make that happen soon, but I am unclear if it will come to fruition.
With that out of the way, we can dive into the debate itself. If discussing any specific claims, I will use timestamps so you can follow along as closely as possible. Most of the corresponding references will be directly on the screen while we’re discussing each point, but in the case that one isn’t, I will reference it in the text and provide screenshots for specific tables/figures when applicable.
The Debate
The first few minutes are spent discussing much of the background I outlined above, as well as short introductions by both myself and Tucker, so if you want to jump forward to the debate itself, it starts at 10:05.
Introductions – 10:05
10:05 – I think I laid out my introduction pretty well. My goal was to highlight the consistency in the research between randomized controlled trials (RCTs), prospective cohorts on both dietary intake and nutritional biomarkers, and mendelian randomization. Given that there was a lot of information to cram into 5 minutes, I didn’t get too bogged down into the details, but we got into the nitty gritty of some of the most important references during the body of the debate. In hindsight, I do wish I discussed some of the nutritional biomarker data more throughout the debate, but I think I did a fair job displaying the strength of my position regardless.
15:55 – Just prior to starting Tucker’s introduction, he said that he refers to CHD as CVD (cardiovascular disease), but those aren’t the same thing. CVD also encompasses other cardiovascular diseases, such as stroke. CHD is a type of CVD, and CHD events often make up the majority of CVD events, which is why I too did discuss some data looking at CVD as a whole, but I just want to clarify that referring to CHD as CVD is not accurate. During the introduction, Tucker brings up several sets of data, including ecological and mechanistic data, as well as some randomized controlled trial data, which are almost all discussed in depth later on in the debate, so I won’t comment further here.
22:08 – The first thing I wanted to ask Tucker was how he would define his epistemic framework. Essentially, I wanted to know how he evaluates evidence and comes to his conclusions. His first answer didn’t make much sense, and it wasn’t clear if he understood what I was asking, but I would say his second answer was very reasonable. Once we established that, my plan was to refer back to it if/when he strays from his own framework. Unfortunately, while I did have opportunities to point out inconsistencies later on, I did not think to refer back to it in the moment. That’s the reality of doing this in a live, verbal setting.
Nutritional Epidemiology – 23:55
23:55 – One of the first topics that is discussed is whether or not nutritional epidemiology is valid and reliable. It appears that Tucker has an issue with the dietary intake data, given that it’s possible that it includes improbable dietary intakes, at least for some individuals. The with that view is that we’re often talking about data collected from tens of thousands of individuals, where those sorts of outliers would be washed out and wouldn’t impact the larger data set. Even then, we don’t need the data to necessarily be accurate on an individual basis. What we need is for those who say they consume the most of a given food/nutrient to consume more than those who say they consume very little of that food/nutrient. I think it’s rather bizarre to hold the belief that the data is so confounded that those who say they consume something daily actually consume it to a similar degree or even less than those who say they almost never consume it. In fact, those without cognitive impairment fill out food frequency questionnaires (FFQs) with a greater degree of accuracy when measured against biomarkers than those with mild cognitive impairment.1 If it were the case that people couldn’t remember what they ate, and FFQ data was completely random, we would expect those without cognitive impairment to perform just as poorly as those with cognitive impairment. In addition, these food frequency questionnaires are often validated against either nutritional biomarkers or weighted food records, so we know they can predict intake with a fair degree of accuracy and I provided Tucker with multiple references showcasing how well intakes of linoleic acid (LA) are measured by FFQs. However, I could even grant that FFQs are not reliable (I don’t hold this position), and still make an argument for nutritional epidemiology being solid, which is what I did in the debate.
26:06 – To combat Tucker’s concerns, I pointed out that, regardless of any issues that may exist with nutritional epidemiology, the epidemiology is concordant with the RCT data the majority of the time (65-67% overall), as illustrated by multiple analyses on the topic (I said 67-69% during the debate, but I was actually citing the wrong statistic. That is my mistake, but doesn’t alter the point I was trying to make). In fact, if we make an apples-to-apples comparison of RCTs looking specifically at dietary intake of various foods/nutrients (rather than supplements) to prospective cohorts also looking at dietary intake the concordance rate is even higher (I unfortunately didn’t bring this specific statistic up during the debate). As you can see below, 21 out of 23 of the apples-to-apples comparisons (91%) were concordant between RCTs and epidemiology.2
Since Tucker holds the position that human RCTs are the highest quality data we have (22:45), I’m not sure how he can go on to state that epidemiology is wrong 100% of the time (27:40) when it’s actually concordant with what he claims is the highest quality data 91% of the time. He also follows that claim by stating that the FFQ data isn’t validated against measures of what people actually eat, which isn’t true, but I didn’t even bother combatting that point because it was tangential to what I was saying about how well the results from epidemiology match the results from RCTs. As you’ll see, I continuously bring it back to this point and am really just asking him to acknowledge that a 67% concordance rate is a majority, but he repeatedly refused to answer that simple question by stating that it’s irrelevant and I’m on a tangent when I’m directly combatting his claim that nutritional epidemiology is always wrong, breaking rule #1 multiple times. After several attempts, he finally answers “no” without providing any argument. When pressed on the matter, at 32:55 he states that he hasn’t had time to review the analysis, which is a cop out answer and I wish I drilled that point home further since it should go without saying that we shouldn’t make claims about publications we haven’t read. In fact, he himself cited several studies for me to review with less than 24 hours to go, breaking the rules we set around sharing data, yet I never complained and stayed up late to ensure I reviewed them. If he needed more time, we could’ve postponed it. Without commenting further on this issue, I just want to mention that my arguments do not hinge on nutritional epidemiology in the slightest. The RCT data also supports my position as you’ll see throughout this debate, but I was simply combatting Tucker’s claims regarding the epidemiology, as it was one of the first claims he made even though he himself cited a fair amount of epidemiology in the form of ecological data during his introduction (more on this later).
Minnesota Coronary Experiment (MCE) – 36:08
36:08 – Since Tucker states that MCE is the best RCT and was upset it was excluded from some of the meta-analyses, I suggested that it be the first one we discuss. He agreed when I said there was a high dropout rate in this study, but because they still had a large number of total participants, he doesn’t believe that is a valid reason to discount the results. While I have other issues with this study, I started by pointing out how the study lacked sufficient statistical power, essentially a study’s ability to ascertain an effect of the intervention. It’s important to understand that 2 of the major factors that impact a study’s statistical power are sample size and follow-up time. Sure, MCE had a rather large sample, but given the short follow-up time for most participants, there wasn’t enough statistical power to pick up on an effect. In fact, when I asked Tucker if they satisfied their own power calculations in the study, he claimed it was irrelevant and dodged the question, once again breaking rule #1. He went on to discount the power calculations as just a “statistical test,” while simultaneously holding the position that MCE’s large sample is reason to hold it in high regard, but since the reason a larger sample is beneficial is because it increases statistical power, I’m not clear on why he thinks the larger sample is meaningful. Regardless, he also claims that they showed harm, although it’s very possible the results were just due to chance alone, and in fact they are actually compatible with the intervention reducing risk, as demonstrated by the large confidence interval. When I raised this point, he tried to change topics to the LA Veterans Administration Hospital Study (38:55), which I did cite during my introduction, but I wanted to stay on topic and tackle these studies one at a time, which I think is reasonable. At 39:41 he actually suggested that we should move on, but we hadn’t resolved or even discussed half of the issues with MCE yet. I’m not sure why he thought it was time to move on if our goal is to have an in-depth scientific discussion of the data on this topic. Furthermore, during this exchange he actually walks back on calling MCE the highest quality study he has, and instead calls it “the biggest” (40:02).
40:29 – Following this, I brought up an earlier publication from this trial in which they separately analyzed those who actually stuck with the diet for 2 or more years, and in that case, if anything, those who were in the intervention group did better than those in the control group, on average. I did make a point to say that the results were not statistically significant, but as Tucker made clear earlier, that doesn’t matter to him. In his view, an average reduction in risk should be considered a reduction in risk. Instead, he discounts these results by stating that it’s a post hoc analysis (41:25) and later suggests that looking at post hoc analyses is “the most rudimentary fallacy of study analysis” (47:45), which I want you to remember, since he cites a post hoc analysis later on (2:45:03). At this point, he continues to refer to the total result, which included people who remained on the intervention for a year or less. It should go without saying that, if the goal is to identify an effect of an intervention on a disease outcome that can take decades to develop, excluding those who only remained on the diet for a matter of months is reasonable.
Side note: You’ll see throughout this debate that Tucker continuously refers to outcomes that are not relevant to the debate proposition. This debate is on coronary heart disease, but he often points to all-cause mortality, or even cancer when the CHD results are clearly not in his favour; however, even then, the results are explainable and don’t favour his position.
42:20 – Of course he tried to bring it back to the LA Veterans trial again and goes on to erroneously claim that more people died in the intervention group, which is a claim he admitted was wrong later on when I asked him to look up the results and provide evidence. It’s also interesting that he brings up other analyses and states that non-significant results show “no benefit” while simultaneously holding the position that the non-significant results from MCE show harm. This is another contradiction.
43:05 – Here I ask him point blank if he agrees that we see a lower death rate amongst those who had longer-term adherence in the intervention group with the data up on the screen in front of him. There is only one answer to that question if one holds the position that 6.1 is lower than 11.1 and 3.8 is lower than 14.2 (I should hope so), but instead he dodges the question breaking rule #1. Also, when asked if he thinks we’d be more likely to see an effect of the intervention in people who were on it longer, he dodges and claims irrelevance, once again breaking the same rule.
45:00 – I then asked if those people in this study were somehow immune the the effects of seed oil and he speculates that there are some people with genetic predispositions to being able to better handle omega-6 fatty acids; however, this doesn’t explain why they tended to do better, not just similar to the control group. Plus, he’d have to hold the position that the people in this subgroup, on average, had this genetic advantage, which is an unsupported claim. After further clarification, at 47:15 he acknowledges that he actually does not have an explanation for why they tended to do better.
48:56 – At this point, I asked him for clarity as to why the intervention group did better and he says “I think I’ve clarified it,” so I’ll leave you to judge whether or not that’s true. He directly follows this by switching topics yet again, this time to the Oslo-Diet Heart Study and raised issues around dietary confounding. In this case, I did bite on the topic change, and in hindsight I wish I brought it back to MCE to finish up with my last couple critiques, but after a short divergence, I was able to bring it back anyway. Since he takes issue with the dietary confounding of omega-3 fats in the Oslo-Diet Heart Study, it’s very odd that he holds Lyon Diet Heart Study in high regard, which is ripe with dietary confounding, as I will explain shortly.
51:56 – As a final note on MCE, I wanted to bring attention to the likely confounding by trans fats (TFA), which were common in soft margarines (MCE used corn oil margarine) at the time of the study. In response, Tucker claims that “nobody knows the composition” of the margarines at that time. That is not true. While they may not have had analyses of the margarines used in that trial at that time, we now know that the production methods used would create a fair amount of TFA, so we can infer based on subsequent analyses. However, to support his position, he cites changes in LDL-cholesterol, although they didn’t actually measure LDL-C, they measured total cholesterol, but in fairness, I also conflated the two at first. I explained that the issue is that cholesterol didn’t decrease as much as would be expected, not that it didn’t increase. It’s possible that the margarines led to a net reduction in cholesterol, but some of the cholesterol lowering properties of the polyunsaturated fats (PUFA) were blunted by TFA. That’s the point I tried to get across. He hand waves this as speculation, but if you recall, his rebuttals to my arguments about longer-term adherers was pure speculation as well, and in this case, I actually have some data to support my speculation. From here he argues that all cooked seed oils will have some level of TFA and in response to my question about if he thinks an oil with 1% TFA is the same as an oil with 15% TFA (55:26), he first dodges (rule #1) and when I re-asked the question, he actually suggests that an oil with 15% TFA is not necessarily going to be worse than an oil with 1% TFA (55:55). To go even further, he then suggests that LA may be even worse than TFA because TFA are saturated and not susceptible to oxidation. I think I was too stunned by hearing him suggest that TFA could be healthier than LA to pick up on this, but when he says that TFA are saturated (56:08), that is blatantly false. TFA are by definition unsaturated since they have a double bond and are not saturated with hydrogens.
Image source: BioNinja
On the topic of TFA and health outcomes, the data is very consistent in suggesting they are the most harmful of all the dietary fats for a variety of health outcomes, including total mortality.3,4
Isocaloric substitution of specific types of dietary fatty acids for saturated fatty acids in Wang et al3
Isocaloric substitution of specific types of dietary fatty acids for saturated fatty acids in Zhuang et al4
LA Veterans Administration Hospital Study – 57:01
57:01 – At this point we moved on to the LA Veterans trial and I asked Tucker to expand on his concerns around the supposed harm that it showed. Interestingly, he specifically quotes a part of the paper that suggests that the intervention didn’t cause harm. Regardless, at 58:31 he suggests that the intervention resulted in a higher risk of all-cause mortality (ACM). This is false, and when I asked him to show me where that was, he admitted he was incorrect and that the study actually did result in a benefit for ACM (note: whenever I refer to ACM in this trial, I was actually referring to deaths possibly related to the intervention/control. This is because there were more deaths due to unrelated causes, such as accidents, in the intervention group, which can skew the results. This is what I had presented on the screen while I was discussing this with Tucker, so I would assume he understood that, but in case anybody is wondering, I wanted to clarify). As a result, he shifts his focus to the idea of the trial being a “blended intervention” because the omega-3 fat, ALA, also increased slightly; however, that ALA came from seed oils. This debate is on seed oils, not isolated LA, so unless he holds the position that seed oils that also contain a modest amount of ALA (ie. soybean oil) provide a net benefit over saturated fat-rich sources, his criticism falls flat. However, at 1:00:56, he actually acknowledges that the studies do demonstrate that a seed oil intervention in which ALA also increases to a degree is “slightly beneficial.” When asked to quantify what he considers a “slight” benefit, given that they found a 35% reduction in risk of ACM and 33% reduction in risk of dying of atherosclerotic CVD, which I would certainly qualify as a fairly substantial benefit, he presented a figure from a review paper, in which the results contradict what is seen in the LA Veterans trial itself. When asked which statistic they were citing, he was not able to tell me, but I can confidently state that that result is not looking at ACM, which is what Tucker has been pointing to to suggest harm.
1:03:56 – Tucker subsequently suggests that the LA Veterans trial was also confounded by TFA because they used hydrogenated oils, but I explained how fully hydrogenated oils contain very little TFA, and he himself ended up acknowledging that the hydrogenated oils were unlikely to have impacted results after reading further into the study’s text (1:07:20). Of course, at this point he reverts back to stating that they showed a “small benefit” (1:08:32). To put this into perspective, he believes that the non-statistically significant increase in risk of ACM (7%) in MCE is worrisome, but the 35% reduction in risk in LA Veterans is “small.” For a more apples-to-apples comparison, if we were to consider the seed oil group in LA Veterans as the control and look at how much the saturated fat-rich foods increased risk, there was a 54% increase in risk… It appears he is holding a massive double standard here. Of course, I’d be happy to chat with him about this in part 2 of our discussion if he ever agrees to it.
1:08:45 – When asked for further clarification, he points to the non-statistically significant increase in cancer mortality and claims it’s a “large increase.” Once again, this debate isn’t about cancer, nor was the cancer rate high enough to lead to net harm, as is evident by the ACM results, but I decided to go along because even the cancer data is explainable. It turns out that those who didn’t adhere to the intervention were the ones who died of cancer. In other words, those who didn’t consume much of the seed oils, so it’s not accurate to claim that the oils were causing cancer. Before getting into the cancer data, I asked if he agrees there was a cardiovascular benefit overall, although he had already acknowledged there was a “slight benefit.” In this case, he dodged the question and said we can’t “slice and dice” the study like that (rule #1).
1:11:11 – Given that the criticisms didn’t pan out, he goes on to suggest that there was a slight benefit due to the increase in omega-3 fats, which I already discussed earlier, so I won’t bother reiterating. Simon stepped in and asked a great question about what amount of omega-3 would confound the results, and to what degree. Tucker didn’t have a specific answer, other than speculating about the effect, which is hypocritical given he criticized me for speculating earlier. On top of that, he seems to hold the view that ALA doesn’t have much of an impact, and that it’s EPA and DHA that do. He claims that high LA intakes block the conversion of ALA to EPA/DHA (1:12:48), so I’m confused as to why he even thinks ALA would be confounding an intervention where LA was increased to a large degree and ALA was only modestly increased. Based on his explanation, ALA should not even be considered a confounder.
1:14:56 – His next criticism was that the population in LA Veterans had a high background rate of disease, which I believe he is claiming as support for the idea that vegetable oils aren’t healthy, but are somehow replacing even worse foods? I’m not entirely sure of his point. His next argument hinges on ecological data, which is interesting since it is epidemiology and he heavily criticized epidemiology at the outset of this debate. Regardless, I pointed out that the majority of the populations he was referencing died at young ages, before we would expect them to have a heart attack (1:19:33). To suggest that they had low rates of heart disease is because they didn’t consume seed oils is quite frankly absurd. Regardless, to counter this point, I asked him if he would take the position that replacing saturated fat with vegetable oils would be beneficial in a population with a high baseline rate of CVD, if that’s what he’s claiming is the case in LA Veterans (1:18:09). To answer this, he refers to a different publication that claimed in the text that there was “little to no benefit” due to multiple non-statistically significant results, even though he cites non-statistically significant results that are in his favour. Yet another contradiction.
We do end up returning to the LA Veterans trial later on in the debate, but I think it was pretty clear during this initial exchange that Tucker’s criticisms did not pan out, he often tried to shift topics when faced with questions that he couldn’t favourably answer, and he did ultimately acknowledge that there was at least a “slight benefit” to consuming seed oils, which supports my position.
Lyon Diet Heart Study – 1:22:04
1:23:38 – Tucker acknowledges that the Lyon Diet Heart Study was multifactorial, meaning that the intervention group changed multiple aspects of their diet, but hones in on the fact that LA was lower and ALA was higher in the intervention group compared to the control group. He goes on to ask why this trial found the largest benefit of the CVD prevention trials while decreasing LA intake, if LA is supposed to be cardio-protective (1:23:48). I think I did a good job of responding to this in the debate by highlighting the reduction in SFA, increase in ALA, and how minute the difference in LA intakes was overall, in addition to the overall dietary shift. It is entirely possible that the multitude of changes in their diets resulted in a larger reduction in risk than trials that only really changed one thing by replacing saturated fat-rich foods with seed oils.
1:28:23 – At this point, Tucker claims that the author’s rationale for the study was to adjust the P/S ratio and explains why, but that tells us nothing about linoleic acid specifically, nor does it tell us what is responsible for the results, nor does it interact with the points I made.
1:29:40 – Thereafter, he acknowledges that the meta-analysis he cited found a benefit for heart attack, but “small benefits” for all-cause mortality. Now he’s at least acknowledging a benefit for heart attack risk, which is relevant to the debate proposition and supports my position, while handwaving the ACM finding without ever quantifying what is considered “small” in his view. I suspect he would also have to hand wave the results from most of the studies he cites if he were to quantify that.
1:30:00 – Tucker claims that the 2020 Hooper meta-analysis5 wrongfully excluded the Lyon Diet Heart Study because it apparently did not modify dietary fat. I would agree with Tucker in that Lyon did in fact modify dietary fat intake; however, another exclusion criteria was “multifactorial intervention,” so either way, it was rightfully excluded as both Tucker and I agree it was multifactorial. I am not clear on why they listed lack of fat modification as the reason for exclusion, but it could be a simple mistake.
Mechanisms – 1:33:22
1:33:22 – After what I think was a solid job by me explaining why we don’t see a similar magnitude of benefit in other trials compared to Lyon, Tucker says we have to talk about the mechanisms of heart disease. While I fundamentally disagree that we need to look at mechanisms when we have health outcome data (I explain why later), I went along with this because I knew which studies he was going to cite and that those very studies actually contradict his own position. He picked out specific aspects of the mechanisms behind the development of atherosclerosis while leaving out critical details like what actually initiates the process.
1:34:57 – I found this part to be truly bizarre. Tucker shares his screen of a slide where he wrote “oxidized LDL is THE explanation for CVD” while also sharing a figure highlighting 6 other mechanisms on the very same slide.
1:35:45 – Tucker isn’t sure who referenced the olive oil paper. To clarify, we both included it in each of our references. I was actually surprised he included it since it doesn’t seem to support his position, but we return to discuss this paper more in depth near the end of the debate.
1:38:58 – I discuss this during my rebuttal, but I just want to make abundantly clear for anyone reading this that oxidation of lipids is not step 1 of atherogenesis as Tucker claimed. In fact, the Borén paper6 he cited makes it very clear that an elevated concentration of LDL-C precedes oxidation, but he skipped right to Figure 3 of the paper, ignoring what came before it (Figure 1 outlines the entire process). In fact, him claiming that there is “no other explanation” is pretty unbelievable given that every single one of the paper’s he himself cited here clearly suggests otherwise.
Figure 1
Figure 3
1:39:55 – In response to me clearly showing that his own reference counters his claims, he retorted with “no, that was a hypothesis,” but that’s irrelevant. He claimed that his explanation was “the only explanation,” while citing this reference that shows that is clearly false. Plus, his hypothesis is just that, a hypothesis as well.
1:45:35 – From this point on, you’ll notice that Tucker cut me off and spoke over me multiple times (rule #2). This was the toughest part of the debate to handle. Not because I didn’t have the information to counter his claims, but because as soon as I started to provide an explanation I would be interrupted.
1:47:55 – I asked Tucker if he has data to support the position that reducing LA intake eliminates the possibility of LDL oxidation, but he dodged the question and instead cited data that found a reduction in oxidizability (rule #1). When I said that wasn’t what I was asking, he claimed that it was what I was asking… but it clearly wasn’t. He then went on to talk about different levels of oxidation susceptibility, which again, isn’t what I was asking.
1:53:46 – After I summarized my position, Tucker followed up by suggesting that I may be misunderstanding the process of oxidation because it is a requirement for the development of atherosclerosis. This doesn’t interact with anything I said, and in fact, I agreed that it is involved in the process, and clearly (or at least I think it was clear) explained why it is not the initiating factor. He also claimed that oxidation cannot happen without omega-6 fats, but was never able to provide me with any data suggesting we can abolish LDL oxidizability by limiting LA in our diets, and instead dodged the question when I asked previously.
1:55:02 – At this point, he brought up some issues with the 4E6 antibody assay that was used in a study I didn’t even reference during the debate. As such, I didn’t even bother addressing it during the debate because it’s tangential to the claims I’ve made, but if you’d like to read more about why Tucker’s interpretation is incorrect, you can do so here.
1:55:42 – I asked Tucker why the harmful effects of LA are not seen in the LA Veterans trial, and in doing so I stated that adipose tissue LA values begin around 10% of fatty acids, but I also state that dietary intake was around there too. I want to clarify that I’m referring to dietary intake as a percentage of fatty acids, not as a percentage of total calories, which is referred to when we discuss the 5% threshold later. I just want to make sure that nobody gets confused by the shift in percentages later on.
1:56:08 – In response to my question regarding where the threshold of effect is, Tucker first went on a tangent (rule #1), then when I asked again where that threshold is, he went on another tangent (rule #1), and finally suggested the threshold is around 5% of calories after my 3rd attempt (1:58:57). This is important since Tucker believes LA is only harmful if increasing intake from very low levels and that’s a reason why he believes we don’t see harm in trials where intakes are higher across the board.
1:59:18 – Since Tucker agreed that the proposed threshold is around 5% of calories, I brought up the results from the 2018 Hooper analysis7 that he himself cited, in which they found benefit when increasing LA intake from below 5% of calories to above. This is damning evidence against his position given that the only evidence he has been using to support the idea that risk increases across that threshold was in the form of a multifactorial intervention (Lyon Diet Heart Study), but a study that does a better job isolating LA as the primary exposure actually found the opposite. When I brought this up he accused me of cherry picking. I’m not sure how looking at the specific results that fit his own criteria for the type of study that should demonstrate harm could be considered cherry picking, but I’ll let you be the judge of that. It’s also interesting that Tucker was not aware of the subgroup analyses in his own reference (I didn’t provide this reference, he did).
2:03:09 – Tucker even agreed with me here when I said there can be multiple mechanisms at play, but that contradicts his earlier claims about oxidation being the only mechanism.
2:04:39 – When I clarified that we see the opposite effect to what Tucker proposed when crossing the 5% threshold, he says he was simply speculating. That’s fair, as it was speculation, but if he now takes the position that the threshold doesn’t explain why we see benefit in certain trials, then that argument goes out the window. He followed this by claiming that I came up with “some table that we can’t seem to find,” even though I clearly have it and he’s the one having difficulty accessing it. I believe I later figured out why he was having difficulty. He was referring to an older version of the paper, while I was referring to the updated version.
2:05:27 – Tucker suggested that we move along because he won’t have time to review the paper (again, it was his reference, but he appears to have had an outdated version).
2:06:48 – Tucker asked which study was being referenced in the Hooper analysis, and it was LA Veterans, where LA intake increased from ~4% of calories, so we dove back into a discussion on that trial. Of course, he resorted to the old argument that it was confounded by ALA, which is irrelevant given that this is a debate on seed oil consumption, an issue I’ve already discussed above in this article.
LA Veterans…Again – 2:06:54
2:07:35 – In response to me listing the seed oils that were used in the LA Veterans trial (soybean oil, cottonseed oil, safflower oil, and corn oil), Tucker responds by saying that “corn oil has virtually no seed oil.” That’s an odd statement because it’s not a mix of various types of oils. Either corn oil is or is not a seed oil. Unfortunately he didn’t allow me to respond immediately when I tried to, but based on his follow-up comments, it sounds like he meant to say it has virtually no LA. That is false (see image below). It has the 7th highest LA content of all the listed oils in the figure below based on data from the USDA Food database. It has a very similar fatty acid content to soybean oil, and actually has less omega-3 and slightly more omega-6 (LA), while containing more LA than other seed oils like cottonseed oil. If he actually meant to say that corn oil is not a seed oil, then it’s a discussion of semantics because the fatty acid content fits with what Tucker claimed was the important factor of the oils he wanted to focus on (8:57).
Source: USDA Food Database
I suspect this was just Tucker grasping at straws because his other criticisms of the trial had fallen flat thus far, and if he truly held the position that corn oil is off the table, then he would have destroyed many of his own arguments. 2 out of the 4 trials he cited in favour of his position specifically used corn oil. These trials are MCE (corn oil and corn oil margarine) and the Rose Corn Oil trial… It’s right in the name. What’s funny about this exchange was that he followed up his claim about corn oil by citing the 3 trials he favours, which exclusively used corn oil. In reality, if we were to throw out any trial that used corn oil, he would only be left with the Lyon Diet Heart Study, a multifactorial trial. Perhaps my biggest folly of this entire debate was not catching this hypocrisy in that moment.
2:09:47 – I claimed that the LA Veterans trial primarily used soybean oil. I was mistaken as I rechecked the publications and it doesn’t explicitly state that, as far as I can tell. I have to call out my own mistakes as well.
2:10:11 – I asked tucker if he considers a 35% reduction in risk of ACM and a 33% reduction in risk of fatal ASCVD to be a “small benefit,” as was seen in the LA Veterans trial, but he dodged (rule #1) and referred to the written text in a separate meta-analysis. At 2:10:36 I repeated the question, and yet again he dodged (rule #1) and referred to the 3 trials (2 of which used corn oil) that he favours.
2:11:01 – He strawmanned me by suggesting I’m ignoring meta-analyses in favour of a single trial to “prove an effect,” but that’s not the case at all. This is the only trial that fit his criteria for crossing that 5% threshold discussed earlier, and I wanted him to either explain why we see the results we do, or to concede that seed oil consumption is beneficial when replacing saturated fat, if LA intake crosses that threshold.
2:11:38 – I repeated my question about the “small effect” for a 3rd time in the span of a minute and a half. In response Tucker claimed it’s a “small reduction” because it’s a large relative risk, but a “small” absolute risk reduction. Let’s see how the absolute risk stacks up against some of his favourite trials, shall we?
Difference in absolute risk of all-cause mortality in the intervention group compared to the control group of the included trials.
Trial | Absolute Risk Difference
(all-cause mortality) |
Minnesota Coronary Experiment | +0.43% |
Lyon Diet Heart Study | -3.95% |
Sydney Diet Heart Study
|
+5.85% |
LA Veterans Administration Hospital Study
|
-7.18% |
Note: There are issues with focusing on ARs in this case, such as differences in trial length, which I’ve discussed here, but I performed these calculations just to point out the inconsistency in Tucker’s position.
If Tucker holds the position that the LA Veterans trial resulted in a “small” absolute risk reduction, then surely he holds the position that the other trials he favours also resulted in small, or even very small increases in risk, no? In addition, if we focus on CVD deaths, which is more relevant to the topic of the debate, the results are fairly similar. I went with ACM here because Tucker himself seems to favour it, even though it’s tangential to the debate proposition.
2:12:30 – He goes on to say that it’s a “small reduction” because most of these people will go on to die of CVD anyway, but he has no way of proving that and that could be true of every trial we’ve discussed, including the ones he cited.
2:12:45 – This claim was truly astonishing to me. Tucker suggested that reducing the risk of death due to cardiovascular disease in the USA by 33% and potentially bringing it down to the 3rd highest cause of death isn’t a significant improvement since it’s still a “huge death rate.” According to the CDC,8 857,226 people died in 2020 due to heart disease and stroke combined, which would account for most CVD deaths. Tucker is essentially handwaving the possibility of preventing about 282,885 CVD deaths in the USA alone per year by simply having people swap out saturated fat-rich foods for seed oils.
2:13:02 – As I’m explaining why these results are significant, Tucker cut me off (rule #2) and changed subject to a completely different publication (sort’ve breaking rule #1, although it wasn’t in response to a specific question).
2:13:42 – Here he says that “I do a thing” where I cherry pick data, which could arguably be considered an attack on my character (rule #3), while misrepresenting the reason we’re focusing on LA Veterans. Once again, it’s the only trial that fits his criteria for crossing the 5% threshold, which is why we returned to discussing it, and we agreed to systematically discuss each individual trial as per Simon’s suggestion. I think it’s pretty clear he doesn’t have an argument for the results of the LA Veterans trial, which in and of itself would suggest it’s the highest quality trial of the bunch.
2:14:21 – Tucker claims there is “no evidence” that seed oil can reduce CVD by ~30%. We can debate the quality of that evidence, sure, but his claim is patently false as I just showed evidence.
2:14:57 – Simon asks Tucker what his major issues are with the LA Veterans trial and Tucker vaguely says “there are some problems with it” without expanding on what they are, but acknowledges they aren’t significant and the study did ultimately show a “small benefit.” After this he says that the results are not consistent across other studies, but remember that there’s a reason we’re focusing on this one. He also suggests I’m ignoring contradictory evidence, but I’m not. I have commented on the trials he has brought up and continued to do so throughout the rest of the debate.
Hooper Meta-Analysis – 2:16:05
2:16:05 – Once again, Tucker wanted to focus on the authors’ summary for the 2020 Hooper meta-analysis5 to suggest “little to no benefit,” but those comments are irrelevant to the debate proposition. He’s referring to the commentary on saturated fat reduction, not the results specific to polyunsaturated fats/vegetable oils (I wish I noticed this earlier, but didn’t catch it very quickly). In fact, the only commentary on polyunsaturated fats that was included in the summary is highlighted below.
They explicitly state that swapping out saturated fat for polyunsaturated fat is beneficial for cardiovascular event risk, and even when he read the authors’ notes suggesting that reduction of saturated fat alone probably reduces cardiovascular events, he appeared to be less excited than he was when he read “little to no benefit” for mortality outcomes.
2:17:42 – As I was explaining why they didn’t find a statistically significant reduction in risk of mortality outcomes, Tucker claimed that “power is a statistical game,” but I’ve already commented on this above (36:08). Also, I should note that if he doesn’t care about power or statistical significance (he’s made this pretty clear thus far), then he should actually accept that the very results he claimed show “little to no benefit” actually do suggest benefit, since there were non-statistically significant reductions in risk.
2:18:31 – Tucker went on to suggest that, if you need a high amount of power to detect a difference in risk, then that risk is small. Well this is interesting given that MCE required a much larger sample (based on their own power calculations) than LA Veterans to detect an effect, so he would have to accept that LA Veterans found a larger effect. In fact, every single criticism he raised actually applies to MCE. Yet another contradiction.
2:19:29 – Here I asked Tucker if he agrees that reducing cardiac events (heart attacks, strokes, etc.) is meaningful. He replies by saying it depends on what the trade off is, so I clarify that all else is held equal when asking this question. He dodged the question at first (rule #1), but I clarify it’s a hypothetical and I’d like to hear his answer. He continued to dodge. He then went on to say that he would rather have 2 “minor” heart attacks than 1 fatal heart attack. I intuitively think many people would agree with that, but again, this doesn’t answer the question given that I made clear all else is held equal in the hypothetical.
2:20:19 – At Simon’s request, I re-stated my question and made abundantly clear that I would just like to know if reducing the number of cardiac events is a good thing, provided everything else is held equal (ie. no downside that is causing harm in another way). To my astonishment, he says “not necessarily.” By saying that, Tucker essentially suggests that reducing one’s risk of having a heart attack or stroke is not necessarily a good thing if we knew that there were no additional risks to the given intervention. To support this position, he again suggested that he would prefer to have multiple non-fatal heart attacks rather than a single fatal heart attack, and after a follow-up question from me he said the only reason we care about having a heart attack is because it can kill you. If there is any point during this debate that my emotions rose, this was it. I found his comments to be incredibly insensitive, as I (and probably most you) have had people close to me suffer a heart attack and survive, only to have it significantly impact their quality of life. To go further, since the study findings were talking about cardiac events as a whole, that includes stroke. Tucker may also be suggesting that having a non-fatal stroke and suffering permanent physical or cognitive impairment is not a big deal. I sincerely hope that nobody watching this was triggered by his comments, and if so, I apologize on his behalf. I truly hope (and believe) that Tucker doesn’t actually hold the position he’s presenting, but was simply claiming to in order to avoid admitting that the interventions were beneficial.
2:22:31 – Tucker asked if I care about silent heart attacks, so I follow that up by asking if he is saying the bulk of the cardiac events in the trials we’re discussing were “silent,” in which case he said we’re on a tangent, which is a dodge (rule #1). This is not at all tangential, and in fact, it directly interacts with the current topic of discussion. One thing I should note is that I did not mean to agree with Tucker when he said silent heart attacks don’t matter. This goes back to what I mentioned in the introduction of this article. Saying “sure” or “okay” is just a habit of mine when trying to interject and respond.
2:22:56 – After a bit of a runaround, he again cited the data for saturated fat reduction, which is not the topic of the debate, nor is it relevant to the specific stats I cited from that paper, while also running through arguments I’ve already discussed above.
2:24:41 – When Tucker shared his screen I finally realized he was referring to the saturated fat statistics and not the polyunsaturated fat substitution statistics. When I pointed this out, he says he’s looking at my paper… That’s such an odd response. This paper is long, as he acknowledged, this debate is about seed oils, and when I shared this reference with him before the debate, I wrote a line specifying which results I wanted to discuss. Him essentially suggesting that he’s countering my claims because he’s citing a couple of irrelevant lines of text from my reference suggests to me that he doesn’t understand my arguments and is grasping at straws. In fact, he goes on to say that the authors made the same “little to no benefit” claim regarding the polyunsaturated fat substitution, but that appears to be false. There is nothing I can see in the text, other than the commentary I referenced above, that pertains to the substitution.
Tucker’s Rant – 2:25:33
2:25:33 – Tucker made a very bold claim when he states that the biggest change in this population, in which CVD is the primary cause of death, over the last 100 years is an increase in consumption of omega-6 polyunsaturated fats. I’m not even sure how to quantify that and pit it against non-dietary changes like medical advancements, smoking habits, etc. If he was specifically referring to dietary changes, then his argument still falls flat. If we look at average food intake changes by percentage of calories between 1909 and 1999 in one of Tucker’s own references,9 the largest change was actually a 14.32% reduction in grain intake. By Tucker’s logic, the reduction in grain intake should be responsible for the increase in CVD, no? Also, if we instead focus on changes in consumption of various foods, measured by mass, then the increase in poultry intake is the largest increase (35.4kg per person per year). The changes in oil intake pale in comparison. Now he may point to the relative changes in intakes, but given that he previously claimed that large relative risks are not important if they translate into small absolute risks, I don’t see why he wouldn’t be more concerned with the absolute changes in intake.
2:26:24 – Tucker claimed that CVD is so low in the Tsimane (hunter-gatherer) population that it is “undetectable.” This is false. Based on an analysis comparing CAC scores between the Tsimane and US populations, we see that a higher proportion of the Tsimane have coronary artery calcium (CAC) scores of 0 and a lower proportion have a CAC score over 100.10 So based on that alone, his statement is incorrect. However, I will add that they essentially guessed the Tsimane participants’ ages. They didn’t use actual birth records, and based on subsequent data that evaluated their ages using a more objective measure, DNA methylation, it’s likely that their ages were overestimated by a significant degree, which would place their CVD rates, based on their CAC scores, in a similar ballpark to the US population in the previously mentioned study.11 What’s interesting is that I actually sent him these references prior to the debate, so he should already know this if he reviewed the references.
2:26:30 – He also mentioned the Masai population, another hunter-gather population, but they too have CVD, and in fact, based on a reference I sent him,12 they have extensive atherosclerosis, which is comparable to what we see in older American men. They just have other compensatory changes in their vasculature as well, such as the widening of blood vessels, which is suggested to be due to their high level of physical activity, not the fact that they don’t consume seed oils. Again, I sent him this reference prior to the debate, so I’m not sure why he’s making a blatantly false claim.
2:26:32 – He then referenced the Navajo Indians, and did acknowledge they have some level of CVD, albeit less than we see in other populations. While that may be true, I don’t see how an argument could be made for seed oils (or in particular linoleic acid) being the difference maker given that tissue LA levels are actually very similar between the Navajo Indians,13,14 Tsimane15 and baseline values in the LA Veterans trial,16 which then increased by a very large degree after consumption of the oils. Funny enough, the baseline values in LA Veterans was actually lower than what has been seen in Inuit populations.17 In hindsight, I wish I brought all of this up during the debate, but there were too many claims to tackle all at once, so I had to pick my battles.
Note: LAVAT = LA Veterans Administration Hospital Study
2:26:41 – Here Tucker raised a question about when we expect the increase in seed oil consumption in the USA to “work,” as far as decreasing risk of CVD/CHD. In response I point out that the primary source of seed oil in the American diet is ultra-processed food (UPF). These foods are unhealthy for a multitude of reasons, so we can’t pin it on seed oils. In response, Tucker suggested that I was conceding that seed oils are harmful when consumed with UPFs, but that’s a strawman and not what I said at all. These foods can be unhealthy despite containing seed oils, just like soda can be unhealthy despite containing water. In fact, I would take the position that if you replaced the seed oils in UPFs with butter, they’d be even worse.
2:28:42 – After clarifying my position on UPFs, Tucker simply said “okay fine” and then once again claimed that the single biggest change in the American diet is the increase in vegetable oil consumption, completely ignoring everything we just discussed about how they’re consumed. Of course, I also tackled whether or not it was the “single biggest change” above, but unfortunately didn’t address that on the spot.
2:29:55 – To his credit, Tucker did concede that smoking is a confounder for the associations he’s been making claims about.
2:32:29 – Tucker again tries to bash epidemiology. Of course, we discussed this at the outset of the debate, in which case Tucker refused to answer questions that ultimately showcased nutritional epidemiology’s reliability. In this case, he refers to work by an epidemiologist, John Ioannidis, and claims that he wrote a paper suggesting that “every food causes cancer.” That’s not true. Ioannidis’s paper18 claimed that 72% of the evaluated studies found that the tested food was associated with a higher or lower risk of cancer. On top of that, he didn’t actually share this reference in advance of the debate, but he did share a different review19 by Ioannidis, which makes similar claims, but instead of cancer, focuses on overall mortality and lifespan. He may have simply mixed up the two, which is understandable, but even then, the latter has a published response that points out very basic errors, including calculation errors, in the methodologies used by Ioannidis and colleagues.20
2:32:41 – Tucker claimed that you can find a study to suggest that basically any food causes cancer, which appears to me to be an attempt to discredit epidemiology due to the fact that there can be contradictory results. If that’s what he meant, then that same criticism applies to RCTs, which he values more highly. In fact, we spent a large portion of the debate discussing why there are contradictory findings between various RCTs.
2:33:22 – Tucker claimed that the papers I shared looking at the validation of FFQs found weak correlations with actual intakes of LA. He goes further and claims that it was one of the worst data points they found. When I said it’s not true, he just retorts with “it is absolutely true” and doesn’t allow me to respond, so I’ll clarify here. One analysis that evaluated correlation between FFQs, 24-hour diet recalls, and biomarker measurements found pretty solid validation coefficients for total fatty acids, linolenic acid, and linoleic acid (see the slide I prepared for the debate below).21
In addition, a systematic review and meta-analysis of the validity of FFQs, which I also shared with Tucker, found that LA actually had one of the higher correlation coefficients with 24 hour recalls or food records (highlighted in the figures below).22
2:33:55 – He brought up the NHANES dataset, which I don’t believe I ever even referred to, and claimed that 50-80% of the data they collected had “implausible” levels of caloric intake. I have no clue where he’s getting that figure from. At the end of the day, all of his criticisms of nutritional epidemiology fell flat. If I were to grant that there are issues with the validity of FFQs, or any other problem he may have, those issues clearly can’t be that problematic since nutritional epidemiology is concordant with the RCT data about 91% of the time as I discussed earlier. Of course, he didn’t allow me to respond to any of these claims any time I tried to speak up and instead continued on his rant.
2:35:14 – He brought up a study I sent to him, which has absolutely nothing to do with FFQs or anything else we were discussing for that matter. As he had made so many claims without allowing me to respond, I didn’t bother tackling this one due to its irrelevance. If you’d like to see how Tucker is grossly misrepresenting the study, you can read more here (section 2a, reference 16).
2:37:02 – Simon attempted to interject to finally allow me to respond to some of these points, given that Tucker has had the floor for close to 15 minutes almost uninterrupted. Tucker didn’t even allow Simon, as the moderator, to interject, and instead continued his rant. This is very poor debate etiquette.
Ecological Data – 2:38:16
2:38:16 – Simon interjected again and Tucker finally handed it over to me to respond. As you can see, there was far too much for me to respond to over that segment, so I went all the way back to the point I was trying to make 10 minutes prior regarding Tucker’s reliance on ecological associations (ie. seed oil consumption and CVD rates both increased during a similar period, therefore one causes the other). Tucker actually interjected at 2:38:42 and stated that ecological data is the worst type of epidemiology, while stating that it’s “critical” just seconds later. To make absolutely clear for you, Tucker holds the position that the whole of nutritional epidemiology is useless, but has heavily relied on what he himself considers to be the worst type of epidemiology during the debate, because it’s apparently “critical.”
2:39:16 – To exemplify why his standards are absurd, I simply showed the correlation between population level cigarette consumption and lifespan and asked if we should consider it “critical data” in the discussion of smoking and lifespan, since he’s essentially looking at the same sorts of correlations with population level consumption of a single food/ingredient. In response he laughs and says “fine, sort’ve makes my point.” I really should have asked for further clarification, because watching back I’m not sure if he’s saying that he accepts that it is critical data or what kind of point it makes.
Sydney Diet Heart Study – 2:42:03
2:42:03 – When I asked if he wanted to address the Sydney Diet Heart Study and Rose Corn Oil trial, he first acknowledges that there are issues with them without mentioning specifics and goes on to say we can’t just “toss” studies for invalid reasons. Keep in mind that he hadn’t been able to actually find a valid criticism for the LA Veterans trial to this point.
2:42:36 – I pointed out that the Sydney Diet Heart Study can be “tossed” for a valid reason. While the overall trial design was actually pretty solid, they used a type of margarine that is known to contain trans fats (TFA). Tucker’s response was the same response he gave me when I brought up the probable TFA confounding in MCE. Unfortunately, I didn’t have a good response when he asked me exactly how they know the composition of the margarine, but again, I wrote about that concern earlier in this article. I addressed most of his subsequent comments about this study in the debate itself, either in that moment or when we discussed MCE earlier. However, I will point out that Tucker cited a post hoc analysis regarding serum cholesterol levels and mortality in MCE to support his claim, but when I cited a post hoc analysis during our initial discussion of MCE he called it cherry picking and claimed it was “the most rudimentary fallacy of study analysis” (47:45). I’d say that is rather hypocritical.
2:47:15 – When I made my final point in my argument as to why the corn oil margarine could result in a reduction in cholesterol levels, despite containing a fair amount of trans fats, Tucker did not address it at all and instead pivoted to discussing a different study altogether. Note that he referred to this as my reference. While that’s true, he also cited this paper in his references before I ever shared my references, so he clearly believed it supported his views in some way. You’ll notice throughout the subsequent discussion that he tried to distance himself from it by continuously stating that it was my reference, particularly when it became clear he was misinterpreting it. I won’t add as much much additional commentary here, other than highlighting a few specific comments that were made, because I believe the misinterpretation and/or misrepresentation on Tucker’s part was made abundantly clear in the debate itself.
Olive Oil vs. Vegetable Oils – 2:47:22
2:48:11 – Not only did Tucker appear to misunderstand what adjustments were made here (this will become clear shortly), but he was referring to the adjustments made in a sensitivity analysis, not even the main analysis or substitution analysis, which were the results we actually ended up discussing throughout this section of the debate. In addition, the reason they removed PUFA but not other fatty acids from the healthy eating index score was because the score didn’t include other fats. So really, saturated and monounsaturated fats were also “removed” from the model, but it isn’t explicitly stated unless you were to look at the components of the diet score. Tucker must not have looked into it further and didn’t seem to understand what he was talking about here. I unfortunately didn’t catch these mistakes during the debate.
2:49:11 – Tucker agreed with their results regarding olive oil consumption and a reduction in mortality risk, as well as the results for margarine and mayonnaise increasing risk, but this is nutritional epidemiology, so based on his standards, he either shouldn’t care what the results were or should actually oppose the results because nutritional epidemiology is wrong 100% of the time according to him (27:40).
2:55:36 – Instead of acknowledging his misinterpretation of the study, Tucker called the subject a “tangent” and tried to change the topic to discuss the mechanisms again.
2:57:13 – I said the substitution analysis was looking at replacements of 100g of olive oil with other fat sources, but I misspoke. It was 10g.
2:57:57 – After I highlighted the results of the study pertaining to vegetable oils, which is the result that is actually relevant to the debate, Tucker responded by saying “it’s an odd study.” This was a cop out answer. Remember that he himself cited this very study. He should have read it, and clearly did based on the highlighting on his screen share, but did not appear to fully understand it.
2:58:00 – When asked if he has any actual criticisms of the study, he said “yes,” but then decided to make conspiratorial comments about a completely different study (rule #1), which was never directly referenced or shared with me in advance, so I cannot fact check.
2:58:58 – Tucker commented on yet again another study without any reference so I cannot fact check.
3:00:32 – When I brought it back to the results regarding vegetable oils and asked if he had any issue with the results, he dodged the question (rule #1) and said the claimed mechanism (singular) supports his argument. Well, they actually highlight seven different mechanisms, not just one mechanism related to oxidative susceptibility.
3:02:45 – I asked why Tucker’s proposed mechanisms aren’t born out in the outcome data from this study and he dodged again (rule #1) to reiterate his claims about the mechanisms.
3:03:04 – When asked again for him to provide an argument, he hesitated, then said “I don’t know that this paper gives an answer to that.” I think this suggests that Tucker is ideologically attached to his position. He doesn’t have a valid criticism for the results of the paper he himself cited, yet refuses to accept the results.
Why I Don’t Highly Value Mechanisms – 3:04:33
3:05:20 – Since Tucker continued to resort to his proposed mechanisms (which aren’t even supported by his own references, as I previously explained) as being reason to discount the results of this study, I asked Tucker if it would be okay for me to explain why I don’t put much stock in mechanisms. He said “go ahead,” but seconds later took issue with me wanting to discuss the translation rates between animal models (which is often used in this mechanistic research, and at the end of this debate he actually proposes a follow-up study that he claims would need to be done in animal models at 3:30:28) and human outcomes by claiming that his main arguments had been focused on human data. Well, the papers he cited earlier from Steinberg23 and Witztum24 discuss plenty of animal research conducted on mice, rabbits, and other animals, and the Witztum paper actually predominantly discusses animal research. On top of that, most of the human research was using isolated human cells, which doesn’t translate very well either. Regardless, after Simon gave me the go-ahead to finish my commentary, I decided to skip over the translation rate of cardiovascular research between animals and humans (it’s quite poor) since I thought Tucker would just claim it was tangential afterwards anyway.
3:06:30 – To further highlight the poor translation between mechanistic research and clinical outcomes, I mentioned data on the translation of “highly promising” cancer research. In response, Tucker once again claimed this was tangential (3:10:41) because it’s “totally outside the scope of our discussion.” I’m confused as to how since I said I wanted to discuss why I don’t put much stock in mechanisms alone, Tucker agreed to let me do so, then as I was doing so he claimed I was on a tangent by discussing animal data even though this reference wasn’t exclusive to animal data, it was referring to mechanistic research as a whole.
3:07:35 – Tucker claimed that a paper he cited by Reaver24 used human data, and while he did share this reference with me prior to the debate, as far as I can tell, he didn’t once discuss it during the debate unless it was some off hand remark I don’t recall. Even then, this was in response to me explaining how oxidative susceptibility is typically measured in vitro (in a petri dish) and this study did exactly that, while also using mouse macrophages (immune cells), not that I think that’s super important in this discussion.
3:08:12 – I finally brought up the very poor translation rate of phase 1 clinical trials (in humans) through phase 3 and approval. Only 6.9% of the mechanisms are later accepted as effective treatments. To accept this but discount nutritional epidemiology, which has a fairly high translation rate to clinical outcomes, is a double standard. Tucker said this is “totally a tangent,” but I am at a loss as to how this is at all tangential.
To recap: I asked him if I can explain why I don’t place much stock in mechanistic research. He said okay. When I started to discuss translation of animal models, he said his arguments hinge on human data (although that statement contradicts the very references he cited during the debate), so I instead pointed to the translation rate of strictly human data, and he claimed I was on a tangent. How that is a tangent is beyond me. I’m honestly not sure if my argument could’ve been any more on point.
Back to Olive Oil vs. Vegetable Oils – 3:11:01
3:11:01 – Once again Tucker suggested that the olive oil study we were discussing was my reference when in reality it was both of our references. I think he could see that his interpretation was flawed at this point.
3:11:46 – To combat Tucker’s claim that we can’t look at vegetable oil consumption in isolation (outside of mayonnaise, margarine, etc.) because vegetable oils are everywhere I attempted to make a smoking-related analogy, but I lost my train of thought in that moment and I wasn’t able to convey what I wanted to. I meant to say that it’s like suggesting we can’t isolate the impact of second hand smoking from directly smoking in a study because they both involve inhaling some level of smoke.
3:13:06 – Tucker claimed that his big problem with the paper is that they separated vegetable oil containing foods like mayonnaise from isolated vegetable oils. Earlier his big problem was the lack of PUFA adjustment, but at this point I think he accepted that he misinterpreted what that meant.
3:14:17 – Tucker claimed that eggs have been shown repeatedly to not be associated with risk, which is a claim I suspect is almost solely based on epidemiology, and is a topic I’d love to debate him on in the future. However, as I wanted to stay on point, I didn’t dive down that rabbit hole. If you want to see perhaps the most in-depth analysis on eggs and CVD that I’ve ever read, you can do so here.
3:14:32 – I asked Tucker if we could agree that margarine and mayonnaise are not isolated vegetable oils and replied: “no.” This was truly astounding. Simply looking at the ingredients list on any body of vegetable oil and comparing them to any mayonnaise or margarine should make it pretty clear that that is not the case. At this point in the debate, it appeared as though Tucker held the position that corn oil is not a seed oil, but mayonnaise is. Let that sink in.
3:15:05 – To further clarify, I asked Tucker if a bottle of canola oil and a bottle of mayonnaise were exactly the same thing and he acknowledged that they aren’t. That should be case closed, but then he goes on to say “they are the same at the end of the day,” whatever that means.
3:15:38 – Simon interjected to bring it back to to the substitution analysis from the study we were originally discussing and Tucker continued to hold onto his original position yet had no argument for why the results contradict that other than saying “it’s epidemiology,” even though he himself cited this paper and the translation rates of epidemiology are greater than the translation of the mechanistic research he values so highly.
3:17:00 – He repeated the false claim that this study found harm from vegetable oil consumption by conflating vegetable oils with mayonnaise, despite the egg yolk, salt, and other contents of mayonnaise that distinguish it from isolated vegetable oils. From here on he repeated his misrepresentations of the mechanistic research, which I’ve already commented on and finished by suggesting that we summarize our positions and bring the debate to a close. Keep in mind, he never once actually provided a rebuttal for the results.
To finish up, Simon asked each of us a few wrap up questions prior to getting into our closing statements, but I don’t think they require additional commentary.
Conclusion
Ultimately, I’m happy with how the debate went. Were there instances where I could’ve made a stronger argument or rebutted certain claims that Tucker made? Sure, but overall, I believe I provided sound arguments based on multiple converging lines of evidence (RCTs, epidemiology, mendelian randomization data, etc.) while providing valid criticisms for the majority of Tucker’s arguments, oftentimes by citing his own research. On the contrary, Tucker largely failed to criticize the strongest data I had in my corner, such as the LA Veterans trial, the meta-analyses, and the epidemiology, and most of his attempted criticisms revealed double standards that apply, sometimes to an even greater degree, to his own studies. In addition, he broke the various rules we had set dozens of times, most often to avoid answering questions that may require answers that are unfavourable to his position. I hope that those of you who watched the debate can see that, as my overall goal wasn’t necessarily to change Tucker’s mind, but to bring clarity to those of you who may be on the fence, or even more of the belief that seed oils are harmful. I too used to believe that, but upon evaluating the best available evidence through a more objective lens, I changed my view, and I hope you can too. As last note, should Tucker finally accept Simon’s offer and provide his availability for a part 2 of this debate, I will gladly participate.
Special Thanks
I want to conclude by thanking a few people. First off, thanks to Simon for being so generous with his time in moderating this marathon of a debate, as well as providing us with a well-known and respected platform. Of course, thanks to Tucker for agreeing to the debate and keeping it respectful. Thanks to Matt Madore, a nutritional science major and friend of mine, who helped with my literature review prior to the debate. A big thank you to Nick Hiebert, who’s content regarding seed oils was a phenomenal resource I used during my preparation for this debate. Lastly, thanks to all of you who sat through nearly 4 hours of 2 people bickering about seed oils.
Dr. Matthew Nagra is a Naturopathic Doctor in Vancouver who is passionate about evidence-based nutrition and medicine. In addition to his online and public speaking work, he employs a wide range lifestyle interventions, including plant-based nutrition, as well as physical therapies in his practice to help his patients prevent or recover from chronic illnesses and physical injuries. Holding a Plant-Based Nutrition Certification from Cornell University and the T. Colin Campbell Center for Nutrition Studies, he has penned several articles on the topic, underlining his specialized training in nutrition.