Bayesian Investor Blog

Book review: The Paleo Cure, by Chris Kresser.

I wish I had read this when I went paleo 7 years ago. It’s more balanced than the sources I used. Alas, it was published shortly after I finished a big spurt of learning on the subject.

It still has a modest number of ideas that seem new to me, and many ideas that I’d have liked to have known when the book was first published, but which I found through less organized sources.

The diet

Kresser advocates starting with a very strict 30-day reset diet, which allows no chocolate, no sweeteners, no butter (ghee is allowed), no alcohol, and of course no grains or legumes. He outlines how to then relax the restrictions to end up with something a bit more convenient than the stereotypical paleo diet.

Kresser strikes a good balance between the conspicuous pro-fat gurus, and the more cautious paleo advocates who are neutral on fat, and who are widely overlooked because they’re insufficiently controversial. Kresser recommends a low carb diet for some, with a simple and crude test for whether it’s likely to be valuable for you. He’s apparently focused mainly on readers who don’t use blood tests. I suspect that an HgA1c level above 5.6 would be a more sophisticated rule to use.

There’s plenty of nuance in Kresser’s specific advice (e.g. raw broccoli is good for most people, but causes thyroid problems in some), but when talking about motivations for avoiding categories of food, he abandons most nuance and rigor, and sounds more like a salesman (although less so than some promoters of paleo-like diets, such as Gundry or Dave Asprey).

He also has less strict advice to follow something resembling the UNpacked Diet (i.e. avoid most packaged foods). Kresser doesn’t explicitly mention the UNpacked Diet, since it was invented after The Paleo Cure was first published, but Kresser says things that sound awfully similar (part of a section title: “if it comes in a bag or a box, don’t eat it!”). The UNpacked Diet is intended as a joke, and I don’t expect anyone to implement it, but it seems significantly more sensible than the standard American diet (SAD).

Measuring nutritional quality

Kresser emphasizes nutrient dense foods. He complains that standard measures of nutrient density are biased toward some non-paleo foods (i.e. grains), and points to an alternate measure create by Mat Lalonde (which, alas, seems to be explained only in that long video).

Organ meats are at the top, grains and oils are at the bottom. Where are insects? They should be a close second to organ meats on any paleo-friendly version of this list (and I prefer their taste over liver), but they’re missing from this book.

I approve of attempts to improve on measures of food nutrient value, and Lalonde’s measure has some advantages over the alternatives.

Lalonde tries to build his measure objectively, by only considering nutrients that are classified as essential. That causes him to treat fiber as neutral – it’s not “essential” because its absence won’t reliably kill people, but I’m guessing that fiber deficiencies cause more deaths in the U.S. than any other single nutrient deficiency.

Sodium is an essential nutrient, but Lalonde excludes it because he doesn’t like the implications of encouraging sodium consumption. He doesn’t address subtler versions of that kind of problem, such as men getting too much iron.

I’m torn between deciding that Lalonde’s measure is better than the available alternatives, versus wanting to invent a better measure myself.

Breadth

Kresser has a better explanation of fructose than I’ve seen before: it’s mainly the fructose to glucose ratio that matters. The ratio that’s found in most fruit is fairly safe, but getting more fructose than that can be hard for our bodies to handle.

The Paleo Cure is not just a diet, it’s a paleo lifestyle. Some examples of his lifestyle advice:

• Walk/run barefoot, or with minimalist footwear such as Vibram FiveFingers
• Sit on a balance disk
• Give up pointless arguments
• Take a partner yoga class
• Get a dog

That portion of the book seems less memorable than the diet portions, mostly because the diet requires more thought and attention to get right.

Reference quality

The book has some references. They’re not in the book, but are available on a part of his website that’s restricted to readers of the book. That would be a good approach if it enabled him to find room for more scientific references than he was willing to fit into a book. Alas, it only provides references for a small fraction of the book’s claim, and often the reference is to one of Kresser’s blog posts. His blog posts typically contain a few links to technical articles that are somewhat relevant, and generally provide some inconclusive support for the book’s claim.

Kresser also provides 12 supplemental chapters online. Those have references that look close to what I’d expect in a peer-reviewed journal. That’s a weird difference from the main book.

I made some quick attempts at fact-checking some claims that I guessed were not well supported by evidence:

Sitting increases the risk of death

I was under the impression that the evidence for this was rather weak, so I checked a bit, and it looks like the experts are more confident than they were 6 years ago that the causality likely works as implied.

Coincidentally, as I was writing this review, I read a theory in the book Move Your DNA which slightly improves my model of why we ought to expect sitting to be harmful.

The evidence on sitting still seems a bit shaky, but it looks a bit stronger than it looked a few years ago. I’ll estimate a 80% chance that this claim is correct.

Freeze them for at least two weeks to kill any potential pathogens in the raw meat.

I was surprised to find that this claim seems to be about half-right. A paper from PubMed says:

The few outbreaks of food-borne illness associated with frozen foods indicate that some, but not all human pathogens are killed by commercial freezing processes.

This comment shows some graphs saying that bacteria continue dying after many weeks in a freezer. Apparently they get refrozen many times, and each time kills a modest fraction?

From a supplemental chapter, on treating people without pre-existing heart disease who are at risk for heart disease:

Yet in spite of this marked reduction in LDL cholesterol in the statin group, there was no difference in lifespan between the two groups.

Kresser is summarizing Statins and All-Cause Mortality in High-Risk Primary Prevention A Meta-analysis of 11 Randomized Controlled Trials Involving 65 229 Participants” (free download here).

The paper’s stated conclusion backs up Kresser’s claim:

Conclusion: This literature-based meta-analysis did not find evidence for the benefit of statin therapy on all-cause mortality in a high-risk primary prevention set-up.

But that’s not at all how I’d summarize the paper’s evidence:

with 1447 deaths occurring among 32606 participants assigned to placebo arm and 1346 deaths among 32623 participants assigned to statin-treated arm, reflecting about 100 fewer deaths in the statin-treated group. In a random-effects model meta-analysis of these 11 trials, the risk ratio for all-cause mortality associated with the use of statins was 0.91 (95% CI, 0.83-1.01). The corresponding risk ratio using a fixed-effect model was 0.93 (0.86-1.00).

I summarize that as saying that my best guess should be that statins reduce mortality by something like 7 to 9%, and that the sample size (65229 patients!) was not quite big enough to reach a definitive conclusion. It clearly supports Kresser’s main claim, that statins are not at all wonder drugs, but 100+ fewer deaths does not seem like “no difference”.

The main problem seems to be the paper’s use of the term “evidence” to mean something like “admissible in court”, whereas I want to use the data for a more Bayesian influence on whether I should take the drug.

I was easily able to find newer meta-analyses which did show statistically significant mortality reductions.

I still can’t tell whether a statin would be good for me. The main uncertainties involve side effects, and how much I can reduce my risk via other means. Kresser provides a number of ideas about better means, and the near-zero rates of heart disease in many traditional cultures suggest that better means exist.

Fans of nominative determinism will be happy to note that Kresser cites, as evidence that some authorities promote statins as wonder drugs, a proposal by Dr. John Reckless to put statins in the water supply.

The contemporary hunter–gatherers were superior in every measure of health and physical fitness.

That is almost an accurate summary of The western diet and lifestyle and diseases of civilization, and it’s important to focus attention on the many ways in which traditional cultures are healthier. But let’s not go overboard. Inflammation, as measured by hs-CRP is much higher in the Tsimané than in the U.S., reflecting more infectious disease in the Tsimané.

Other complaints

Kresser frequently mentions correlations without explaining why the causality behind the correlation supports his advice. Maybe it isn’t possible to do much better than look at correlations on those topics, but I wish he’d show more skepticism about them.

Kresser’s defense of red meat looks pretty weak, because his main discussion of the topic looks only at correlations within the U.S., some of which suggest that red meat is ok. A careful reader might notice hints of the much more dramatic pro-red meat evidence we get by looking at traditional Maasai society: a good deal of red meat, and near zero heart disease. But to notice that, the reader would need to combine two pieces of information about the Maasai that show up hundreds of pages apart. Shouldn’t a book with paleo in the title emphasize this kind of evidence more than differences between several types of questionable industrial diets?

Kresser carelessly alternates between advocating free range eggs in some places, and pasture-raised eggs in other places. “Free range” in this context is misleading, and the “pasture-raised” label is typically the only version that are raised on a healthy diet.

How paleo affected me

I’ve been following a fairly paleo lifestyle for nearly 8 years, with maybe an 80% overlap between what I practice and what Kresser advises. The effects that I’ve experienced haven’t seemed much like a cure for anything. But I didn’t have much in the way of problems that I hoped it would solve. My main motivation was to prevent Western disease.

I have noticed that my digestive tract feels better on a paleo diet, and it’s easier to control my weight. My grass pollen allergy bothers me less, but that could be mainly caused by me learning how to avoid high-pollen areas.

My LDL and homocysteine skyrocketed after my first version of a paleo diet (Bulletproof). Kresser admits that a few people have this problem, and should minimize saturated fat, while still recommending that most people treat saturated fat as healthy.

My HDL and my neutrophil/lymphocyte ratio improved on a paleo diet, and I had a slight improvement in blood glucose. I don’t have many other pre-paleo tests, so I can only guess at whether other important biomarkers improved.

Conclusion

The book’s most important claims look quite reasonable. My criticisms of the book have been about moderate exaggerations and peripheral issues.

Enough time has passed since the book was written that we have better evidence on some of the topics that he discusses, and statins are the only subject on which the evidence against one of his claims has increased a bit. The evidence that sitting is harmful seems a bit stronger now. Mostly the evidence seems unchanged.

It’s a shame that paleo’s reputation has been molded more by gurus who are less careful and too focused on hot-button controversies.