Bayesian Investor Blog

Book review: The Plant Paradox, by Steven R. Gundry.

This book describes a good diet, which Gundry seems to have put together by combining ideas from other competent nutritionists, testing many variations on himself and his patients, and keeping the components that showed the best results.

Unfortunately, his rhetoric is designed to convince us that he’s got brilliant, revolutionary theories. That rhetoric bears little resemblance to careful reasoning. It seems more likely that he tried a bunch of arguments on his patients, and kept the ones that more effectively scared them into following his diet.

His puns are mightier than his scalpel (“No More Mr. Knife Guy”).

I first noticed Gundry via the ApoE4 website, at about the time I gave up hoping that I could eat coconut fat without raising my cholesterol. I saw that Gundry recommended no coconut fat for my genotype, but liked coconut fat for others. That led me to think: here’s a paleo-friendly nutritionist who knows more than average.

Then I switched for 5 or 6 weeks to a diet more in line with his Apoe4 advice [1], and was surprised at how much more my cholesterol levels dropped than I expected. I only have modest evidence suggesting that that dietary change was the main reason for the cholesterol drop, but it’s still a bit of evidence that Gundry knows something valuable that I would have otherwise have overlooked.

Here are several ways to understand his diet, starting from the more broad, theoretical views, and ending with the most empirical views.

Paleo

The obvious way to evaluate a food, in the absence of good experimental data, is looking at how much evolutionary pressure our ancestors had to adapt to it.

Of course, it’s not trivial to determine whether a food is like those our ancestors ate. Gundry uses criteria that are moderately different from other paleo gurus, paying more attention to families of plants (e.g. Nightshades) which are new to humans in the past few millenia.

Gundry used to be what he describes as “full-bore carbophobic” (like many paleo gurus), in spite of our ancestors’ long history of eating fruit. He now sends a confusing mix of messages, sometimes tolerating carbs, but then saying “fruit is candy, toxic candy. … fruit was good for one thing long ago and that was to fatten you for winter.” Some species eat fruit seasonally to gain weight, but as far as I can tell our ancestors ate fruit most of the year. Does Gundry have better evidence on this subject?

Gundry presents a new objection to typical modern fruit – it’s picked unripe, when it has biochemical defenses against being eaten prematurely, and then it doesn’t ripen properly. He wants us to get around this mainly by sticking to local fruit. I’m puzzled as to why frozen fruit isn’t a way to avoid this problem – there’s no obvious motive to pick it before it’s ripe.

I got more confused when he went on to recommend unripe bananas, mangoes, and papayas, due to their high resistant starch and low sugar. That suggests he doesn’t really believe his diatribes about unripe fruit, and is instead perpetuating a half-right heuristic for avoiding diabetes (a better heuristic would focus more on fiber or intermittent fasting).

His willingness to change his mind about carbs is somewhat encouraging, but I’d be more encouraged if he went further and recommended eating plenty of taro and sweet potatoes, rather than recommending them in moderation.

Lectins

Anyone with a decent understanding of evolution should see that plants engage in a variety of biochemical defenses against predators, and that predators evolve to handle those defenses, but can’t do so quickly.

We can make educated guesses about which plants have what chemical defenses to which we haven’t adapted well. So it’s almost common sense that we should be more suspicious of new types of plants than new types of animals.

Gundry identifies lectins as the main plant chemicals that we should worry about. This seems quite plausible, but there’s no obvious reason to think they’re the only ones to worry about. Gundry notes problems with specific lectins, but left me rather uncertain about the magnitude of the lectin problems – theory gives few hints about how harmful plants are. Gundry describes examples in which lectins seem to matter, but doesn’t say much about how many people are affected by lectins.

Gundry’s descriptions of lectins are fairly informative about why they exist and about why we should expect the effects that he describes. But he left me wondering what distinguishes them from other proteins. Why not include something like this simple Wikipedia introduction?

Lectins are carbohydrate-binding proteins, macromolecules that are highly specific for sugar moieties of other molecules.

Tradition

Henrich’s The Secret of Our Success points out that humans have often depended on non-obvious food preparation techniques, often needing to faithfully follow their inconvenient cultural traditions (e.g. cassava causes subtle cyanide-related diseases when people take short-cuts in processing it).

Some of Gundry’s advice returns us to this pattern (e.g. pressure cooking).

Of course, we’re not supposed to notice the tension between that and the “forget everything you thought you knew was true” rhetoric.

Gundry has also been influenced by Blue Zone diets such as the Mediterranean diet. I see some tension between that and has concerns about beans [2]. That tension might be resolved if blue zones prepare beans the way Gundry advises, but I haven’t found evidence of how they prepare beans.

Food Allergies

Gundry’s diet looks a lot like what I’d get if I were imitating the SCD diet but focused on protein allergies rather than sugar allergies.

Avoiding Junk Food

There’s a fair amount of agreement among nutritionists about how to identify junk foods, and Gundry’s diet outlaws almost all junk food. The arguably-junk foods that he accepts are few enough that he can enumerate most of them by brand, and often by flavor within a brand. The most questionable such foods are Laloo’s Goat Milk Ice Cream (why does Gundry accept the unlabeled eggs in it? Shouldn’t we assume they came from chickens who ate crap?) and several brands of protein bars (but he limits animal protein, so if I eat two Quest bars per day, he won’t let me eat any meat/seafood/eggs).

Critics

Critics tend to focus mostly on Gundry’s style, complain about “good” foods that he tells us not to eat, and pay minimal attention to what Gundry eats.

For example, take this HuffPost article by David L. Katz.

Katz disputes Gundry’s attacks on lectins, whole grains, etc., and complains about some confusion over topics, like raw food and salt, which seem unrelated to anything Gundry says.

But does Katz think Gundry’s diet is unhealthy? This quote looks like it’s intended to attack Gundry, while leaving some ambiguitty as to what diet Katz is attacking:

the new contention that we should avoid all of the most nutritious plant foods … Following this advice will decimate the quality of your diet, and for anyone who actually sticks with such silliness over time (an unlikely eventuality with any diet)- your health.

The link in the first part of that quote looked like it was the only available hint about how Gundry’s diet might cause harm, so I followed the it to this more responsible paper by Katz and Meller: https://www.annualreviews.org/doi/abs/10.1146/annurev-publhealth-032013-182351. That paper sure looks like it supports Gundry’s diet. See this summary of consensus dietary advice:

The aggregation of evidence in support of (a) diets comprising preferentially minimally processed foods direct from nature and food made up of such ingredients, (b) diets comprising mostly plants, and (c) diets in which animal foods are themselves the products, directly or ultimately, of pure plant foods – the composition of animal flesh and milk is as much influenced by diet as we are (31) – is noteworthy for its breadth, depth, diversity of methods, and consistency of findings.

Gundry stridently supports that consensus [3], even though he’s not about to describe himself as supporting a consensus.

Katz and Meller also complain:

In pursuit of marketing advantage, notoriety, or some other bias, the defenders of competing diets tend inevitably to emphasize their mutual exclusivities.

That’s a good description of Katz and Gundry when they’re trying to sell their writings to a large market. Like Katz, Gundry sounds more responsible when talking to a sophisticated audience (I recommend this Gundry AHS talk as better than his book).

Nutritional Characteristics

So how plausible is it that Gundry’s diet would decimate my health? I entered some of his recipes into a nutritional web site.

My first attempt used one serving each of the first recipe in each of these Phase 2 categories: breakfast, snacks, main and side dishes, and desserts. That was clearly not enough, so I added the second recipe from main and side dishes. That yielded a bunch of nutrient deficiencies. Its calorie count clearly qualified as significant calorie restriction, and didn’t contain anything that looked like a main course, so that isn’t quite how most people would follow the diet.

Then I noticed that he instructed me to have desserts on a “few occasions”. So I changed the dessert portion to 1/4 serving (to reflect eating it once every 4 days). I also switched to using the first, middle, and last recipes from the main and side dishes category, in place of the first two, to get a soup, a main course, and a salad.

That yielded a 2032 calorie day [4]. It’s a bit low in some nutrients [5]. It has a controversial 48.5 grams of saturated fat. The omega-6/omega-3 ratio is poor, but his supplement advice probably makes up for that. 44 grams of fiber – not quite as good as my current diet or a typical hunter-gatherer diet, but nearly 3 times the US average.

Then I remembered his meal plans, which, confusingly, aren’t near his recipes, and which are harder to interpret because he doesn’t specify quantities. I made some mildly optimistic guesses about how people would implement day 1 of his Phase 2 plan. That produced 2317 calories, 40 grams of saturated fat, and 41 grams of fiber. Still a bit low in B12 and D (he recommends supplements for those). B1 and potassium are 97% of the RDA, and magnesium is 99% of the RDA (close enough, given the uncertainties involved). All other vitamin and mineral levels are comfortably above the RDA.

The diet might decimate the weight of an average American, but any nutrient deficiencies that it produces will be mild compared to what most Americans experience.

The diet’s saturated fat will be bad for some people. Gundry says

I am the world’s expert on the dietary approach for people with ApoE4.

That might be true, but don’t look in this book for any such expert knowledge; see the ApoE4 web site for reports on how Gundry’s advice for ApoE4s conflicts in fairly large ways with his book’s advice (less saturated fat, no cheese, beef, poultry, or coconut oil). How the @#&% can an expert on a gene that 30% of us carry write a book without noting that nearly half the recipes in the book are inappropriate for those of us with that gene? Accommodating ApoE4s would only be a little harder than the accommodation he provides for vegans!

How does it really work?

Gundry tells many anecdotes about his patients doing well after adopting his diet. What are the most likely explanations for this?

• The diet works quite well, for reasons that loosely resemble the reasons given in the book.
• The diet works wonders, not because it avoids lectins, but due to a combination of only allowing foods that most experts agree are good, and to Gundry’s skill at persuading people to stick to the diet.
• The diet works well. Gundry mostly found it by observing what works, but he’s fairly confused about why it works, and the actual reasons will surprise us all when they’re discovered.
• Selective reporting makes the diet look good. The book provides little evidence against this hypothesis, but I found moderate evidence against it here and here (“95 of 102 patients achieved complete resolution of autoimmune markers and inflammatory markers within 9 months”).
• The patients didn’t really improve, and careful measurement would show that. Once again, the book says little to rule this out. He says he does lots of testing (critics complain about overtesting!), and reports a few good numbers for valuable tests, but mostly leaves me frustrated at how little he says about which tests he runs and what improvements he sees.
• Gundry has unusual skill at recognizing people who will benefit from his diet, and only accepting them as patients. E.g. maybe he’s unusually good at recognizing symptoms of food allergies.
• Gundry has unusual skill at recognizing people who will recover with just placebos, and only accepting them as patients.

Most of these explanations seem somewhat plausible.

Sloppy Writing / Research

• His list of foods to avoid in Phase 1 includes all roots and tubers, but his list of YES foods in that phase includes carrots, beets, radishes, and onions. Now onions are technically bulbs, but the others pretty clearly contradict his ban on roots. So what does he mean by root? Maybe “roots prohibited by the SCD diet”? That interpretation works well, provided I ignore the differences in reasoning behind the two diets.
• He says humans invented fire “about one hundred thousand years ago” – Wikipedia says 200 thousand to 1.7 million years ago.
• He repeats the myth that frogs won’t jump out of water that’s slowly brought to a boil. Humans who suffer as a result of a bad diet often switch diets (but switching to a random diet often fails), and peer pressure often contributes to dietary problems in ways not hinted at by the frog myth.
• His arguments against using BHT include: “BHT is used in embalming fluid” – see the Dihydrogen Monoxide FAQ for similar arguments against water.
• “two-leaf plants (bicotyledons)” – apparently an attempt to dumb down dicotyledons, which have two seed leaves.
• “I have yet to see a case of vitamin D toxicity. I doubt that it exists.” – I’m not very reassured by Gundry not having seen a case, given the authoritative-looking claims that he’s disputing. Gundry’s testing advice looks adequate to catch any toxicity that vitamin D might cause, so why does he think this controversy is relevant?
• He claims “we are collectively far less healthy than our parents were at a comparable age”, then cites a study, about early detection of health declines, which has no apparent relevance to his claim. It’s true that he’s pointed to evidence of new health problems, but he doesn’t have much evidence they’re worse than were the lasting effects of iodine deficiency, lead exposure, smoking, polio, etc.

Conclusion

Gundry seems to be a pretty good doctor who has learned a good deal from trial and error. But there isn’t much of a market for evidence from the small-scale trials that he’s done, so he wrote this book more as a pep talk than as an appeal to our intellect.

How upset should I be about the non-intellectual approach?

Intellectual approaches don’t have a good track record at convincing people to eat healthy diets. Some parts of Gundry’s approach probably work better. I suspect his case reports are more effective than are statistics from RCTs at changing peoples’ behavior, for somewhat the same reasons why charities typically appeal to the need to save one child rather than many children.

I find it’s often easier for me to avoid eating something if I classify it as “not food” than it is if I take a more intellectual approach. Gundry’s argument style resembles those which convinced me to classify grains as “not food”.

But many of the problematic parts of the book just hurt Gundry’s credibility without providing any benefit. I’m unsure whether he cares about that. Provoking criticism probably helps book sales compared to more careful authors.

I’m seriously considering trying the diet, but I want to find a 6.5 week block of time when I can completely avoid restaurants, and that will take a while. And after those 6.5 weeks, I’ll probably return to eating at least a few potatoes and peanuts. I’ve done enough testing to be pretty sure that peanuts don’t affect me much. I expect that my diet a year from now will be more Gundry-like, but not strictly adhering to his diet.

Footnotes

[1] – Replacing about half my pasture-raised chicken consumption with shrimp. That change was an accidental byproduct of something unrelated to my health goals. I also reduced my whey protein consumption to near zero, partly due to Gundry’s advice. I’ll have more details in a later post.

[2] – I’m actually increasing my bean consumption in response to Gundry’s report that pressure-cooked beans are fairly safe. My opinion about beans had been fluctuating due to concerns stemming from paleo theory.

[3] – I’m assuming that Katz and Meller didn’t mean that animals we eat need to be vegan. They seem to be making a clumsy attempt to describe animals that eat a natural diet. It’s implausible that they’re trying to say that wild-caught salmon (which eats smaller fish) is unhealthy.

[4] – I.e. a bit more than a typical Kenyan, and less than a typical North Korean. That’s rather far from the U.S. average of 3750 calories. [Added 2018-09-16: Oops, these estimates include food that gets wasted before being eaten; what people actually eat is likely in the 1800 – 2800 calorie range. ]

Official recommendations for moderately active adults range from 1800 to 2800 calories, depending on age and sex.

So this interpretation of Gundry’s advice produces results that are weird, in a way that’s very consistent with the weirdness of mainstream nutritional advice.

[5] –

• 2.6 micrograms of B12 – 51% of the RDA
• 604 grams of Calcium – 60% of the RDA
• 12.4 mg of B3 – 62% of the RDA
• 39 IU of D – 10% of the RDA, but there’s not much reason to expect that from food