Attack of the Vegan Twins (aka Netflix Plant Based Party Part 10)

[TLDR: The new Netflix documentary is a rehash of debunked talking points and an attempt to pretend that a well done but mostly meaningless study is saying something that it doesn't. For more content like this subscribe to my newsletter.]

What’s that you say? Netflix has a new documentary out? And it’s showing that vegans are healthier than omnivores? What year are we in again? 2011? 2018? Nope we’re in 2024, baby. It’s remix time.

This particular movie - “You Are What You Eat” - is based on a recently completed study of identical twins done by Dr. Christopher Gardner at Stanford. That study took sets of twins and put one on a vegan diet and one on what was described as a “healthy omnivorous” diet. The twins were given food for the first month and for the second month they were asked to continue eating to the plan (either vegan or healthy omnivore) but they were not given food.

My first thought when I heard about this study was, "Of course the vegans will do better. Vegan diets tend to be low fat, so a low fat vegan diet will do better than a medium fat omni diet because the combination of fats and carbs is really the issue." But looking at the supplementary materials for the study, I see I was wrong. The researchers were aware of this particular criticism and therefore did their best to keep the macros similar, especially in terms of fat content. The main difference in macros was that the vegan participants were getting more carbs, less protein and slightly less dietary fat than the omnivorous participants. 

The other main difference between the two arms was really about diet satisfaction. As you can see from this chart, the vegan twins reported a significant decrease in overall diet satisfaction as compared with the omnivorous twins (who remained more or less the same as baseline).

So total diet satisfaction went from 3.5 (out of 5) at baseline to 3.0 after 2 months on the vegan diet. As you might expect, the vegan twins simply ate less than their omnivorous counterparts given that they were enjoying their food less:(Average total daily energy intake was 1,688 for the vegan arm as opposed to 1,862 for the Omni arm.)

The difference of almost 200 calories a day is not insignificant and in itself would be enough to explain all the “benefits” of the vegan diet. If any group eats about 10% less than the other I can pretty much guarantee that you'll find differences between the two groups.That’s probably the case on any diet - vegan diet, carnivore diet, Oreos diet, grapefruit diet, whatever. But it's going to be particularly pronounced on a bad diet. If you're eating steak, eggs and vegetables I'm not sure you'll see benefits should you eat less of that. You might see some. But if you're eating a diet of mostly Krispy Kreme doughnuts everyday, I can guarantee benefits from cutting down the number of Krispy Kremes you eat. And they'll be dramatic benefits.

So before I had even heard of the Netflix piece based on the same study, I knew what to expect. Limiting "caloric intake" leads to improvements? Ya, OK. And in related news, I hear that dihydrogen monoxide (take a second if you need it) is wet. Yawn. 

But, for reasons beyond my comprehension, many people were paying attention. People who I know for a fact knew nothing about health and nutrition science were popping up on my social media feeds extolling the virtues of a vegan diet because of this amazing twins documentary. “Alright,” I said to myself, “I'm always up for torturing myself for the greater good.” In retrospect not one of my better decisions, but here we are. 

“You Are What You Eat” has almost no new information in it aside from the data related to twins study. The other stuff can be pretty easily debunked by looking at previous analyses of vegan propaganda. I’m particularly partial to this lecture by Zoë Harcombe on The Game Changers, but there are plenty of sensible people out there who have taken this apart piece by piece. The following list is therefore an incomplete but (I hope) important addition to the growing body of literature on the vegan tendency to put ideology before veracity. 

1) If ever you see Dr. Michael Greger in a documentary, you should fact check everything he’s saying. If he says that the sky is blue, you should probably get a second opinion. He’s the ultimate cherry picker and cites evidence that supports his beliefs while ignoring evidence (sometimes in the very paper he’s citing) that refutes his beliefs. Early in the video he claims that US meat consumption reached unprecedented levels after World War 2. As journalist Nina Teicholz has pointed out, per capita meat consumption is actually lower today than it was in the beginning of the 19th century when people were eating 1.5 to 2 times as much meat as today (150-200 lbs per person as opposed to 100 lbs today). And there are societies - long-lived Hong Kong among them - that consume more meat than the US does with much much better health outcomes. 

2) Later on, Dr. Greger says dairy consumption increases the “risk” of Parkinson’s disease and cancer. He’s reaching this conclusion based on very weak epidemiological data. A google search on Parkinson’s brings us to articles like this one that use FFQ’s (Food Frequency Questionnaires) from other studies to make conclusions about Parkinson’s prevalence. As a reminder, these studies usually ask completely unreasonable questions of participants in the first year (how many salads did you eat in 2023?) and then use that “data” to make conclusions about why an individual may have died or had a heart attack years later. This methodology - which I think we can all agree is flawed to say the least - is one reason why Prof. John Ionnidis, perhaps the most cited scientist on the planet, calls for an end to this kind of nutrition science. “We’ve wasted enough resources and caused enough confusion,” he says in this interview.

The interesting thing about this particular epidemiological data is that it’s even more improbable than usual. It finds an association between low-fat dairy consumption and Parkinson’s but not high-fat dairy consumption. Why would that be the case? And given that it’s the case, if you believe these data shouldn’t you be recommending people drink full fat milk as opposed to skim milk? That seems to be the logical conclusion unless one is religiously committed to avoiding dairy.

And the data for an association between dairy and cancer is even weaker. Limiting ourselves to looking at meta-analyses (these are studies that bring together the results of a number of earlier studies) we find that dairy is associated with a slight increase in the prevalence of prostate cancer (RR=1.07) but also with a (slight but bigger) decrease in the prevalence of breast and colorectal cancers (RR=0.85 and RR=0.83 respectively). None of this data is convincing enough for me to tell you that dairy is healthy, but it certainly doesn’t seem to be unhealthy, especially if these are the best data the vegan zealots can cite.

(As a reminder, RRs inform on relative risk. If something increases the prevalence of a disease from 1% to 2%, that’s an RR of 2.0, i.e. it’s doubled the prevalence. As a general rule, in epidemiology if the RR is less than 2.0, or greater than 0.7 for something that’s supposed to decrease the disease prevalence, it’s probably not worth even testing the hypothesis. So the data that Greger is citing is not even worthy of hypothesis generation. Even if it were, epidemiological data like this can never prove causation, it can only show association. So there would be further steps to take even if he was right, but he's so wrong that it's not worth even checking to see if he might be right.)

3) Dean Sherzai has what I can only describe as the audacity to compare San Bernadino with Loma Linda and to claim that the fact that residents of the latter are more likely to be vegetarian explains why they're healthier. This one made my jaw drop. According to the US Census Bureau, San Bernadino residents are almost twice as likely to live in poverty (12.8% vs 20.2%) as Loma Linda residents. They’ve also earned on average only 60.4% the income that Loma Linda residents have earned in the last year. These differences are more than enough to account for the longer life expectancy. In recent years, San Bernadino has been listed as one of the poorest counties in the USA.

Sherzai’s implication is, frankly, insulting. If your neighbourhood is better known for its meth than for its brunch, you have problems. Not being vegan ain't one of them. Poverty is the primary determinant of health and longevity in the United States, and Sherzai's ignorance doesn't excuse his bigotry. 

4) At about minute 19 of the opening episode we learn that saturated fat raises cholesterol. This is one of those things that you hear and nod your head. Saturated fat raises cholesterol? Why yes of course it does. But why does it raise cholesterol? Because we’ve been told it does.

So where is saturated fat found? Yes there is some saturated fat in meat and even more in dairy. But would it surprise you to learn that the food with the most saturated fat is coconut oil? At 87% saturated fat, if saturated fat increases cholesterol we would expect the effect to be most pronounced with coconut oil. And yet many human trials - including this one - show that coconut oil does not significantly increase cholesterol. In that trial, coconut oil didn’t differ much from olive oil, which we’re told is the healthiest fat. So no, saturated fat does not increase cholesterol. Furthermore in 2020 the Journal of American College of Cardiology (which is about as mainstream a journal as you can find) put out this statement: "Whole-fat dairy, unprocessed meat, and dark chocolate are SFA-rich foods with a complex matrix that are not associated with increased risk of CVD."

5) The decision to feature veg-for-life (and vegan for a while) bodybuilder Nimai Delgado as one of the trainers is… gutsy, perhaps? Although he denies it, most people who know anything about body building believe that he’s on some kind of performance enhancing drug regimen. That includes this vegan YouTuber who goes over some of the evidence. I don't really care, but I do think it's pretty funny that the guy who's telling people they can gain muscle eating plants (and protein powders, I imagine) has a back that looks like this:

6) To its credit, the film talks a little about regenerative grazing as a way to rebuild soil and sequester carbon. But unfortunately it dismisses regenerative agriculture as unscalable without even exploring the question. Is it unscalable? Maybe. George Monbiot mentions a study that says it is. I’m not sure which study he’s citing but the studies he’s cited on this previously were pretty shaky. As Diana Rodgers points out, he’s relying on largely speculative pieces from the 1990s to make these conclusions. The research has gone a lot further in recent years (though I think it’s still an open question). In any case the film misses a trick here. If it had talked more specifically about the problems with the current agricultural model based on mono-crops and industrial inputs and the generation of huge amounts of waste, it might have actually been doing a good deed and informing people. Unfortunately it's just parroting the vegan line on this without delving into the real issues.  

And in any case, the film is ostensibly about healthy diets. Why are we discussing sustainability issues in the context of a movie that’s not going to spend enough time to do these issues justice?

That's an abbreviated list, but hopefully a useful one, of the easily debunked "data" in the movie that isn't pertinent to the main event - the twins study.

The primary outcome that study was designed to look at was lipids, specifically total cholesterol and LDL (low density lipoproteins). For decades, LDL and cholesterol (they’re separate molecules and LDL is not “bad cholesterol”) have been vilified as causing atherosclerotic cardiovascular disease (ASCVD). That’s despite evidence that people with higher LDL tend to live longer (sometimes called the LDL paradox) and that most people with ASCVD have low or normal LDL levels.

The timing of a study that puts so much emphasis on cholesterol is interesting. There is new data from another study - the LMHR study which has been matched with the Miami Heart study - that basically shows absolutely no correlation between LDL levels and the progression of ASCVD as measured by a CT angiogram. The data - this is just the baseline data from a study that will take another year to publish - is being touted as showing that for a specific population of people with very high LDL (so-called Lean Mass Hyper Responders or LMHR), there is no difference in disease progression when matched with a similar population with “normal” LDL. But that’s not the most important thing. Look carefully at this slide presented by Dr. Budoff: 

Forget about the Keto (LMHR) side of the slide for a moment. Just focus on the MiHeart side of the slide, which is a “normal” population without elevated cholesterol. The person with the highest plaque score, meaning that he actually has measurable heart disease, has an LDL somewhere around 120. That's in the "normal" range in many countries. There is an individual with an LDL over 200 who has no plaque, i.e. they don’t have any evidence of disease at all. If these individuals were being given medicine for lowering their risk of heart disease based on LDL, the wrong person would be getting the medication. 

Perhaps even more importantly, look at the shape of that weighted average line. It doesn’t move much but to the extent it does move, it’s moving down. That means that higher LDL was (very slightly) protective against heart disease in the general population (the MiHeart population).

This is just one data point and shouldn’t be used to draw too many conclusions. But it’s a data point that is consistent with a wide body of peer-reviewed data showing that high LDL is not the highest risk factor for heart disease. In some populations it might even be protective.

So why was LDL the primary outcome for this study? I have my own theories, but for now let’s park that question.

There were secondary outcomes to the study, most of which are more important than LDL-C. Let’s take a look at the relevant table:

OK, so what do we see here? LDL has gone down but so has HDL (so-called good cholesterol). That’s not a good thing. And triglycerides have gone up. The ratio of triglycerides to HDL (we want HDL high and trigs low) is known to be a good predictor of risk for heart disease and even all-cause mortality. According to this data, some of the vegan twins may have increased their risk of dying by any cause despite the fact that they lost weight. There are lots of ways to lose weight, but to lose weight in such a way that you actually increase your chance of dying? That's impressive.

In terms of other markers, glucose has actually increased despite the fact that the vegan twins were eating roughly 10% less. That is odd. Insulin has decreased (as we would expect with a diet that’s restricting calories) and that might be important. We’d probably need to look at other markers - inflammation markers such as CRP and homocysteine, maybe other hormones such as glucagon, and incretins such as GIP and GLP1 - in order to fully understand what’s going on here. Such measurements either weren't taken or weren't published. (If you're not following along with this paragraph, don't worry. It just means that you're a normal person and not a nutrition nerd like me.)

In any case the results do not paint an unambiguously positive picture for the vegan twins.

The most concerning finding to me when looking at the vegan arm of the study was the Dexa scan data. As explained (very well) in the movie, a Dexa scan can measure the composition of our body in the way a scale cannot. In particular we use Dexa scans to determine how much fat, muscle and bone an individual has. In general we would want to see increases in muscle mass and decreases in fat mass on any kind of a weight loss or fitness program.

So how did the vegan twins do? They only share the data for the four sets of twins in the video. Actually, they don’t even share that. For the last set of twins, the young men who were trying to gain muscle, they only share part of the data, i.e. how much muscle they gained. And in that partially shared case study, the non-vegan gained more than 7 pounds of muscle while the vegan gained 2.3. So the vegan diet inhibited muscle gain by something like a factor of 3x.

And that pattern holds with the other twins for whom we have more complete data, Pam (vegan) lost 1.7x more muscle than Wendy (6.6 lbs and 3.8 lbs); Carolyn (vegan) lost an astonishing 30x more muscle than Rosalyn (3 lbs and 0.2 lbs); Charlie (vegan) lost muscle (0.4 lbs) while his omnivorous twin gained an impressive 3.6 lbs of muscle.

If we consider muscle to be the organ of longevity as Dr. Gabrielle Lyon and others argue we should, the vegans in this experiment have done themselves a disservice. Gaining muscle mass is not easy, especially for the older twins. Losing it, on the other hand, is all too easy. The data are pretty clear that even a very short amount of bed rest can cause loss of muscle (atrophy). When you consider that muscle mass is the best predictor of when a person will die, that's scary. Again, it might be the case that the vegan twin has done themselves a real disservice here. 

Whether or not muscle mass is the most important metric to be looking at, it’s certainly more important than LDL-C. So why was the latter the primary outcome of this study? When I first looked at the study, I wasn't sure. "Maybe they're not aware of all this data," I thought. But it's worse than that.

Let's look at an earlier study undertaken by the same Dr. Christopher Gardner, the “mostly” vegan researcher behind this project. Back in 2007, Dr. Gardner was first author on another study. This one pitted the Atkins diet (low-carb) against the Ornish diet (vegan) and a couple of other ones. The data he himself (this very same Dr. Christopher Gardner) published here concluded the following: "In this study, premenopausal overweight and obese women assigned to follow the Atkins diet, which had the lowest carbohydrate intake, lost more weight and experienced more favorable overall metabolic effects at 12 months than women assigned to follow the Zone, Ornish, or LEARN diets.”

In other words Dr. Gardner learned as far back as 2007 that the best diet for human health was likely to be a low-carb diet that included relatively more animal foods than even a standard diet, let alone a vegan diet. This was against a whole list of metrics: weight, BMI, triglycerides, HDL, insulin and glucose. The Atkins (low-carb) diet did better overall and on nearly every metric.

In that study, the only metric by which the vegan diet was superior was lowering LDL-C. Is it a coincidence that all these years later Dr. Gardner is designing trials testing vegan diets against the metric that he knows they tend to do best in? Or is he putting his finger on the scale because of an ideology that he holds for other reasons? Might "veganism is good for you!" be the noble lie that Gardner is telling his audience, when what he really believes is something like "as humans we need to reduce our ecological footprint"?

Only one person can answer those questions. Based on the data presented here, there really is no reason to eat more plants if one is concerned about health and longevity. And I strongly suspect Dr. Gardner knows it.