Perfect Health Diet Macronutrient Ratios ~ Part I: Breast Milk
If this post seems a bit out of the blue, it's because it was written a while ago and I'm currently going through the draft bin attempting to publish up some of that mostly done stuff.
First, it is important to note that there are actually three versions of The Perfect Health Diet book (perhaps four, I am unsure if the Kindle version was updated from the first print version, but I'll refer to the self-published Kindle version as v.2). I mention this because the first ebook was published in 2010 as was the self-published paperback and the latest version was completed late in 2012. While there are some minor changes, the core of the basis for the PHD has not changed much, if at all and if there were major changes I would expect Paul to highlight them in his various appearances. So, in v.1 and v.2 the rationale behind the macronutrient content was put forth in a very straightforward manner in a section entitled:
Four Reasons to Believe in a 20-65-15 Macronutrient Ratio (Carb-Fat-Protein)
- This was the composition of the Paleolithic Diet
- This is the composition of breast milk (adjusted for adult brain size)
- This is the composition of the human body - Eat what you are.
- Omnivorous animals prefer to eat this diet
Later in the text these ratios were relaxed a bit to 20% carb, 60-65% fat and 15-20% protein. In all versions, rationale #4 is the weakest. Which macronutrients lab mice prefer when given a choice is hardly convincing, and I've known field mice to get into both bars of soap and birdseed so I'm not going to conclude anything from that anytime soon! This will probably be the first and last time I refer to that rationale, but anything can happen. Also, point #1 has been discussed here quite a bit of late so I am not going to go into that again. Suffice it to say that he cites Cordain here, and there's a real problem with his interpretation -- or rather one might say super cherry picking of the sort to make Gary Taubes proud! This component of the argument is blurred somewhat in v.3 ... to fit the diet? Although not numbered in the same fashion, all of these points persisted in the most recent version of the book.
The basic structure of cells— a protein-rich intracellular compartment surrounded by fatty membranes— hasn’t changed in billions of years. The need for cells to be able to obtain energy by cannibalizing themselves hasn’t changed in that time either. So it should come as no surprise that all mammals have similar macronutrient needs. Mammals do need different foods— some are obligate herbivores, some obligate carnivores— but this is because they have different digestive tracts, not because their bodies have different nutrient needs. All mammalian diets point to the same basic macronutrient ratio: a macronutrient ratio that is majority fat, minority carbs and protein. The typical mammalian diet is about 10 percent carbs, 20 percent protein, 65 percent saturated and monounsaturated fat, and 5 percent polyunsaturated fat. Adjusting for our larger brain size, this suggests an optimal ratio for humans around 20 to 30 percent carbs, 15 percent protein, 50 to 60 percent saturated and monounsaturated fat, and 5 percent polyunsaturated fat. Which is strikingly close to the Paleolithic diet, the cannibal diet of fasting, and the composition of breast milk adjusted for brain size! (Kindle Locations 994-1004)
As you can see, there's a bit of hide-the-total-fat going on here with the breakout of the PUFA. Still there is a shift worth noting: 20-30% carb, 15% protein, and 55-65% fat. I would be remiss if I didn't point out that carb-fat issue aside, no, Paul, the Paleolithic diet is described in none of the common publications to this community as only 15% protein.
The remainder of this post will focus on Point 2, the breast milk issue,. I'll do a followup on what he refers to as some sort of universal mammalian diet or the "cannibal diet" which is related to Point 3, as time permits.
The Breast Milk Argument for PHD
Why now? Well I got side tracked yet again ;-) I was sorting through papers and stuff the other day and listening to podcasts in the background. Paul Jaminet did two new ones ... nothing new in the content really ... but a comment popped up in my reader challenging Paul Jaminet on this topic:
Tom wrote: You mention in the interview that the composition of milk across different species are fairly constant. This does not fit my reading, see for example: [link] . According to this website Antelope and Kengaroo would have 50% of energy from protein, horse milk 70% carbohydrates, reindeer 70% fats, some marine animals as much as 80-90% fats etc. ....
Now since the link is from the Lactation Biology website for a class in the Animal Sciences Department at The University of Illinois, I'm going to take the information to be accurate to the best of the creator's knowledge. Thankfully it was a mere C&P into Excel, a few columns of formulas later and I had the percentages by calories for the macronutrients to play around with. I knew most of the diets off hand but did a quick search to fill in the gaps there as to which species are herbivores, carnivores or omnivores (please let me know if I made any error there.)
Regardless of what he claimed in the podcast, in PHD Paul states twice, perhaps more, that breast milk composition is similar in all mammals:
By calories, and excluding the contribution from milk oligosaccharides, human milk has a macronutrient profile of 54 percent fat, 39 percent carbs, and 7 percent protein.
Note the order:
• Fat provides the majority of milk’s calories.
• Carbohydrate is the second largest calorie source.
• Protein provides the fewest calories.
This order holds in the milk of EVERY mammalian species, not just humans. For instance, milk from cows is 52 percent fat, 29 percent carbohydrate, and 19 percent protein. (Kindle locations 636-643, emphasis mine)And:
But it is telling that most strains of wild-type mice chose to get a majority of calories from fat, a minority from carbs and protein— the same pattern we’ve seen repeatedly in other mammals, in humans, and in breast milk of all mammalian species. (Kindle Locations 895-897, emphasis mine)
This is quite a definitive statement and has persevered through the editions. I find that quite odd because I know that Paul had been challenged on this rationale in general, so I'm surprised he never double checked it. In his response to Tom's comment, it would appear his analysis was very limited to begin with:
Very interesting points. I’ve only looked at a restrictive set of species, and in those simple scaling laws, eg,- Protein scales with body mass- Fat+carb calories scale with energy expenditure (which depends on size [surface area: volume], ambient temperature, activity levels, and brain size)- Fatty acid composition scales with brain size- Fat:carb ratio is related to total energy expenditure : resting energy expenditure ratio
I have inquired if this scaling analysis is available and will link to it if Paul responds (update: no response) . Still, this is not what was stated in PHD. So here are a few bar graphs of the macro contents (by calorie %) of the various mammalian milks (human milk is shown in darker colors):
Milks sorted by FAT content, note humans near the middle of the pack and fat content ranging from 22 to 90%
Milk sorted according to CARB content. Note humans are near the top of the list here with carb content across species varying from essentially zero to 62%.
Milk sorted according to PROTEIN content. Note humans are at the bottom of the pack at 6% with the range up to 52%.
Lastly, I thought it would be interesting to see if there were any patterns that emerged for mammals based on the nature of the natural adult diet. The image below is sorted by carnivores at the bottom (up through seal), herbivores (antelope through elephant) and omnivores (monkey on up)
Milk by DIET type and secondarily on FAT consumption. There does not appear to be much of a pattern here either.
I think it is safe to say that the composition of mammalian breast milk varies widely between species, even within the classes of these species based on their adult diets.
Human Breast Milk vs. Adult Human Diet
I have to be honest here, when I first read PHD mostly in short skimming sessions in various waiting rooms, often interrupted sometimes by days, this argument really stuck out to me as being rather illogical. Infants of all species have different needs than the adults. No difference can be more stark than that of the herbivore, as all mammals by definition eat an all "animal based" diet until they are weaned. The explanation of conversion of nutrients -- the subject of the next post -- does not explain this difference fully, but I find that argument illogical as well. There is no reason to believe that the needs of an infant would match those of an adult. Further, the development of each animal and relative times of maturation vary quite widely. A foal gets up and walks within a couple of hours while it takes many months for a human baby to stand hanging onto something for support. Growth of brain vs. body differs widely, time to reproduction age, etc. I see no reason to even try to extrapolate infant food, macronutrient or otherwise, to adult diets.
Further, the main sugar in milk is lactose -- which isn't even paleo!! So if we're talking evolution, there's a long way for humans from infancy to reproductive age + time until offspring are independent (the important part for survival of the species) and they would not have had access to lactose.
But leaving this aside, Paul seeks to reconcile the various means by which he arrived at the macros in PHD -- it's story telling with quite a bit of stretching to go along with. The macros don't really converge between what the paleo diet is (forget fat/carb, the protein content claimed is far lower than reported in the sources he cites) , and while the jump from the 54% he cites for human milk to 65% sounds nominal (smaller still when you hedge to 60%), this is rather difficult to attain with a "normal" diet that pretty much any humans have eaten, even the Inuit. He tries to go about this task of bringing milk in line with the paleo macros by mathematically adapting the infant macros to the adult brain. Since the infant brain is growing, it needs more carbs for total body weight. The very low protein content makes sense because skeletal muscle and all that is not required as much for the infant as is brain development. However, Paul focuses only on the carbs and surmises that the adult needs about half what the infant needs, and partitions the other 20% between protein and fat. This appears arbitrary, if not biased to preserving the preconceived optimal fat level, and largely to bring protein within range of what human adults generally consume.
The prevalence of ketosis in human babies is a phenomenon ripe for the exploitation of diet gurus. We are born fat burners Mark Sisson has exclaimed! Nora Gedgaudas likes to evoke the ketotic infant quite a bit too. But ketosis is prevalent in certain species early in life for specific reasons. We are used to thinking of ketones as being glucose sparing. In other words, they reduce the need for glucose by providing an alternate energy source for the obligate cells like those in the brain. But here's an interesting paper: Survival of the fattest: fat babies were the key to evolution of the large human brain (a related work by Cunnane is cited in PHD).
... an brain was a product of having first evolved fat babies. Hence, the fattest (infants) became, mentally, the fittest adults. Human babies have brains and body fat each contributing to 11–14% of body weight, a situation which appears to be unique amongst terrestrial animals. Body fat in human babies provides three forms of insurance for brain development that are not available to other land-based species: (1) a large fuel store in the form of fatty acids in triglycerides; (2) the fatty acid precursors to ketone bodies which are key substrates for brain lipid synthesis; and (3) a store of long chain polyunsaturated fatty acids, particularly docosahexaenoic acid, needed for normal brain development. ...
Emphasis is mine. Taken in conjunction with the "high" fat content of human milk (including those MCTs), ketones are of clear importance early in life. I don't want to turn this into a discussion of the role of ketones in adults, so I won't, but as per the bolded statement, ketones are probably more important as building blocks for lipids than they are for energy (though no doubt they play a role there as well). So ... who is to say what fraction of the fats in breast milk are there specifically for this purpose of brain growth ... and when said growth tapers off early in life, it would seem logical that fat requirements are reduced substantially. Indeed, the statement in another paper by the same author that "ketones are important brain lipid precursors and their role in this regard is supported by glucose which generates the necessary NADPH for de novo lipogenesis" would argue strongly that in this context, glucose is still primary fuel, and thus humans with the bigger brains and needs for brain growth should have relatively high carbohydrate content in breast milk. Yep ... right near the top while fats are modestly in the middle. One might argue that adults are merely suppressing this preferred metabolism, but in my opinion (and that of many others) the fact that true ketosis is so difficult to attain and maintain argues against this.
Halving carbohydrate intake on a percentage basis is based on the following: The weight of the newborn brain is 11% of body weight yet uses 75% of the baby's enery, and the adult brain is only 2% of body weight yet uses 23% of energy (these values are referenced to this paper in PHD, but are also in the Cunnane paper). Yeah, and? On a %E/%W basis, the infant brain ratio is 5.5 while the adult ratio is 11.5. Here is where he gets the idea to halve the glucose needs from the diet. OK, I see the math ... sort of ... But this is based on various assumptions such as (1) All of the energy used by the infant brain is from carbohydrate, and (2) the adult brain is the only organ that uses energy from carbohydrate.We know this not to be true.
When published, this section will contain live links to:
Part II: Nutrient conversion & Eat what you are
Part III: Paleo macros