Category Archives: Pregnancy

Nutrition and Pregnancy, I: Nutritional Triage

Happy Mother’s Day!

2013 Mothers DayMother’s Day seems an auspicious time to begin a series on nutrition in pregnancy. It is an important topic, as I believe pregnant mothers are often alarmingly malnourished.

Triage Theory

“Triage theory,” put forward by Bruce Ames [1], is an obviously true but nevertheless important idea. It offers a helpful perspective for understanding the consequences of malnourishment during pregnancy

Triage theory holds that we’ll have evolved mechanisms for devoting nutrients to their most fitness-improving uses. When nutrients are scarce, as in times of famine, available nutrients will be devoted to the most urgent functions – fuctions that promote immediate survival. Less urgent functions – ones which affect end-of-life health, for instance – will be neglected.

Ames and his collaborator Joyce McCann state their theory with, to my mind, an unduly narrow focus: “The triage theory proposes that modest deficiency of any vitamin or mineral (V/M) could increase age-related diseases.” [2]

McCann and Ames tested triage theory in two empirical papers, one looking at selenium [2] and the other at vitamin K [3]. McCann & Ames used a clever method. They used knockout mice – mice in which specific proteins were deleted from the genome – to classify vitamin K-dependent and selenium-dependent proteins as “essential” (if the knockout mouse died) or “nonessential” (if the knockout mouse was merely sickly). They then showed experimentally that when mice were deprived of vitamin K or selenium, the nonessential proteins were depleted more deeply than the essential proteins. For example:

  • “On modest selenium (Se) deficiency, nonessential selenoprotein activities and concentrations are preferentially lost.” [2]
  • The essential vitamin K dependent proteins are found in the liver and the non-essential ones elsewhere, and there is “preferential distribution of dietary vitamin K1 to the liver … when vitamin K1 is limiting.” [3]

They also point out that mutations that impair the “non-essential” vitamin K dependent proteins lead to bone fragility, arterial calcification, and increased cancer rates [3] – all “age-related diseases.” So it’s plausible that triage of vitamin K to the liver during deficiency conditions would lead in old age to higher rates of osteoporosis, cardiovascular disease, and cancer.

Generalizing Triage Theory

As formulated by Ames and McCann, triage theory is too narrow because:

  1. There are many nutrients that are not vitamins and minerals. Macronutrients, and a host of other biological compounds not classed as vitamins, must be obtained from food if health is to be optimal.
  2. There are many functional impairments which triage theory might predict would arise from nutrient deficiencies, yet are not age-related diseases.

I want to apply triage theory to any disorder (including, in this series, pregnancy-related disorders) and to all nutrients, not just vitamins and minerals.

Macronutrient Triage

Triage theory has already been applied frequently on our blog and in our book, though not by name. It works for macronutrients as well as it does for micronutrients.

Protein, for instance, is preferentially lost during fasting from a few locations – the liver, kidneys, and intestine. The liver loses up to 40 percent of its proteins in a matter of days on a protein-deficient diet. [4] [5] This preserves protein in the heart and muscle, which are needed for the urgent task of acquiring new food.

Protein loss can significantly impair the function of these organs and increase the risk of disease. Chris Masterjohn has noted that in rats given a low dose of aflatoxin daily, after six months all rats on a 20 percent protein diet were still alive, but half the rats on a 5 percent protein diet had died. [6] On the low-protein diet, rats lacked sufficient liver function to cope with the toxin.

Similarly, carbohydrates are triaged. On very low-carb diets, blood glucose levels are maintained so that neurons, which need a sufficient concentration gradient if they are to import glucose, may receive normal amounts of glucose. This has misled many writers in the low-carb community into thinking that the body cannot face a glucose deficiency; but the point of our “Zero-Carb Dangers” series was that glucose is subject to triage and, while blood glucose levels and brain utilization may not be diminished at all on a zero-carb diet, other glucose-dependent functions are radically suppressed. This is why it is common for low-carb dieters to experience dry eyes and dry mouth, or low T3 thyroid hormone levels.

One “zero-carb danger” which I haven’t blogged about, but have long expected to eventually be proven to occur, is a heightened risk of connective tissue injury. Carbohydrate is an essential ingredient of extracellular matrix and constitutes approximately 5% to 10% of tendons and ligaments. One might expect that tendon and ligament maintenance would be among the functions put off when carbohydrates are unavailable, as it takes months for these tissues to degrade. If carbohydrates were unavailable for a month or two, there would be little risk of connective tissue injury. Since carbohydrate deprivation was probably a transient phenomenon in our evolutionary environment, except in extreme environments like the Arctic, it would have been evolutionarily safe to deprive tendons and ligaments of glucose in order to conserve glucose for the brain.

Recently, Kobe Bryant suffered a ruptured Achilles tendon about six months after adopting a low-carb Paleo diet. It could be coincidence – or it could be that he wasn’t eating enough carbohydrate to meet his body’s needs, and carbohydrate triage inhibited tendon maintenance.

Triage Theory and Pregnancy-Related Disorders

I think triage theory may helpfully illuminate the effects of nutritional deficiencies during pregnancy. When a mother and her developing baby are subject to nutritional deficiencies, how does evolution partition scarce resources?

Nutritional deficiencies are extremely common during pregnancy. For example, anemia develops during 33.8% of all pregnancies in the United States, 28% of women are still anemic after birth [source].

It’s likely that widespread nutritional deficiencies impair health to some degree in most pregnant women.

Those who have read our book know that we think malnutrition is a frequent cause of obesity and diabetes. Basically, we eat to obtain every needed nutrient; if the diet is unbalanced, then we may need an excess of fatty acids and glucose before we have met our nutritional needs. This energy excess can, in the right circumstances, lead to obesity and diabetes.

But obesity and diabetes are common features of modern pregnancy. Statistics:

  • 5.7% of pregnant American women develop gestational diabetes. [source]
  • 48% of pregnant American women experience a weight gain during pregnancy of more than about 35 pounds. [source]

I take the high prevalence of these conditions as evidence that pregnant women are generally malnourished and the need for micronutrition stimulates appetite, causing women to gain weight and/or develop gestational diabetes.

Another common health problem of pregnancy is high blood pressure: 6.7% of pregnant American women develop high blood pressure [source]. This is another health condition which can be promoted by malnourishment.

It’s likely that nutritional deficiencies were also common during Paleolithic pregnancies. If so, there would have been strong selection for mechanisms to partition scarce nutrients to their most important uses in both developing baby and mother.

A Look Ahead


  1. Nutritional deficiencies are widespread during modern pregnancies.
  2. They probably lead to measurable health impairments and weight gain in many pregnant women.
  3. The specific health impairments that arise in pregnant women or their babies are probably determined by which nutrients are most deficient, and by evolutionary triage which directs nutrients toward their most important functions and systematically starves other functions.
  4. Due to variations in how triage is programmed, deficiency of a nutrient during pregnancy may present with somewhat different symptoms than deficiency during another period of life.

This series will try to understand the effects of some common nutritional deficiencies of pregnancy. Triage theory may prove to be a useful tool for understanding those effects. Based on the incidence of possibly nutrition-related disorders like excessive weight gain, gestational diabetes, and hypertension, it looks like there may be room for significant improvements to diets during pregnancy.

A Paleo Pregnancy Pitfall?

On Saturday’s Around the Web I linked to a study [1] that tied low-carb dieting early in pregnancy to obesity in the child at age 9. This made Ana concerned:

I’m somewhat worried about the pregnancy diet study. Actually I am trying to conceive, 3 months ago my husbund and I changed my diet to Paleo.

Now I see this study and even though I feel great, better than before, I’m not sure, how much credibility would you give it?

I presume that perhaps there could be too little fat, and with that too low calorie intake for a pregnant woman, perhaps that could be the case, opinions please!!!

We certainly don’t want Ana to be stressed out, and it’s hard to turn down three exclamation marks, so I thought I’d interrupt the cancer discussion to address her concerns.

The Study

The study [1] claimed two things:

  1. Women who ate less than a thousand carb calories per day during the early part of pregnancy were more likely to give birth to babies with an overly silenced gene for the Vitamin A receptor RXR-alpha.
  2. Babies born with an overly silenced gene for RXR-alpha were more likely to be overweight at age 9.

Let’s look at the second point, which is more solid, first.

RXR-alpha silencing is associated with obesity

Here is the data:

It looks like it’s normal to have about 50% RXR-alpha methylation in this promoter region and if you have 80% methylation, you’re likely to become a pudgy 9-year old.

How solid is the correlation? They replicated it in a second cohort. Their first study produced two epigenetic marks that were strongly associated with childhood obesity, RXR-alpha and eNOS. A replication study confirmed the RXR-alpha but not the eNOS association.

How plausible is it that RXR-alpha silencing would contribute to obesity? Very plausible, because RXR-alpha is the hub of a network of genes regulating most aspects of metabolism and cell activity.

Vitamin A binds to two types of nuclear receptor, RAR and RXR. When it binds to RXR, a vitamin A – RXR complex is imported into the nucleus. This then looks around for a partner. Partners of RXR-alpha include:

  • Vitamin A – RAR complexes.
  • Vitamin D – VDR (vitamin D receptor) complexes.
  • T3 thyroid hormone – TR (thyroid hormone receptor) complexes.
  • LXR (liver X receptor).
  • CLOCK, the circadian rhythm regulation gene.
  • PPAR-gamma (peroxisome proliferator-activated receptor), a regulator of lipids whose deficiency leads to high cholesterol and hyperglycemia.
  • MyoD, a factor that triggers muscle creation.
  • Many others; a list can be found at Wikipedia.

The vitamin A – RXR-alpha complex “dimerizes” with these other nuclear receptors, forming a new complex that acts as a transcription factor to turn on gene expression. Most of those other partners cannot act to turn on DNA transcription unless they dimerize with RXR.

This means that the absence of RXR-alpha would be functionally equivalent to being low in vitamin A, vitamin D, T3 thyroid hormone, CLOCK, and all those other partners. It is like being born a sun-starved hypothyroid with messed up circadian rhythms who can’t form muscle and is hyperglycemic and dyslipidemic.

All of those things are associated with obesity.

RXR-alpha silencing might be a universal component of the metabolic damage in obesity:

[A]n association between increased RXRA methylation and adiposity is consistent with the observation of strongly diminished RXRA expression in visceral white adipose tissue from obese mice (35).

Personally, I think it’s very likely that silencing of RXR-alpha promotes obesity. This is the most solid part of the paper. They have data, and the mechanism makes sense.

Conclusion: some babies are getting off to a bum start in life due to epigenetic silencing of an important gene.

Does maternal diet affect RXR-alpha silencing?

This is the really weak part of the paper. Here was their data:

When I say this was their data, this was all of it. No scatter plots, no information about how other characteristics of the diet correlate with RXRA methylation, no information about health or lifestyle characteristics of the various carbohydrate intake cohorts so that we can evaluate the possibility of confounding factors.

It is unlikely that low carbohydrate intake was causing the problem. Aside from the fact that dietary carbohydrate intake is only weakly correlated to any factors seen by the baby in the womb (eg blood glucose, insulin, etc), 261 g/day is a substantial amount of carbohydrate – well above physiological needs. So the low-carb quartile included women in glucose deficiency, glucose moderation, and glucose excess; the other quartiles only women in glucose excess. If a glucose deficiency caused RXRA hypermethylation and glucose excess caused RXRA, there would have been a much larger scatter in RXRA methylation levels among the low-carb quartile compared to the 3 higher-carb quartiles. But we can see from the graph that the standard deviations are the same in every quartile.

So there is likely to be some other factor besides carbohydrate intake that was responsible for the RXRA hypermethylation. What are the possibilities?

One possibility alluded to in the paper is that the women had low carbohydrate intake because they were starving. The paper notes that “famine during pregnancy is associated with obesity in the adult offspring (5).” However, I am unaware of recent famines in Southampton UK.

Another possibility is an excess of some other macronutrient. Those mothers who ate fewer carbs were eating more fat and possibly more protein. Given the ubiquity of vegetable oils in modern fats, the increased fat was probably largely omega-6. This raises two possibilities:

  • High maternal omega-6 intake causes RXRA methylation.
  • High maternal protein intake causes RXRA methylation.

Both possibilities have support from studies in rodents: maternal high protein intake and maternal omega-6 fat intake are both associated with obesity in offspring. For more on the risks of high protein, see The Danger of Protein During Pregnancy, Jul 12, 2010.

Another possibility is that the low-carb high-fat diet produced a vitamin A excess. As we discuss in the book, this is a common problem, especially among people taking a multivitamin; probably due to widespread vitamin D and vitamin K2 deficiencies, large numbers of people exhibit evidence of impaired health with vitamin A intake above 10,000 IU/day. As a fat-soluble vitamin, vitamin A intake is more or less proportional to fat intake.

If a balance between vitamin D and vitamin A is needed because the vitamin D-VDR complexes and vitamin A-RXR complexes have to be in proper proportion, then the body may respond to an excess of vitamin A and a deficit of vitamin D by upregulating VDR expression and downregulating RXR expression. Such downregulation may be achieved by RXR-alpha methylation.

Another possibility is some confounding factor that happens to be correlated with carbohydrate intake. In the US Nurses Health Study, nurses with the lowest carbohydrate intake were “rebels” who rejected not only the health advice to eat vegetable-rich and whole-grain rich diets, but also every other standard bit of health advice. The low-carb nurses smoked more, exercised less, and drank more alcohol and more coffee.

So it could be maternal smoking, lack of exercise, or drinking too much alcohol or caffeine that causes RXRA hypermethylation and childhood obesity.

Another possibility, raised in the comments by Amber, is that mothers of the obese children were obese themselves, ate low-carb diets for weight control reasons, and passed on their obesity to their children. It is indeed the case that obese mothers tend to have children who are obesity-prone, and it is suspected that epigenetics may be responsible for this “inherited” obesity. If low-carb diets have indeed become popular among the obese mothers of Southampton UK, then this is a possibility that must be considered.

Low-Carb Paleo Pitfalls?

Should Ana modify her diet because of this study?

I think it’s important to avoid a glucose deficiency. But I don’t think it’s necessary to eat 1,000 calories per day of carbs to achieve that.

I think it’s important to eat a moderate amount of protein, neither too much nor too little; and to limit the amount of omega-6 fats eaten.

I think it’s a good idea to avoid alcohol or excessive consumption of bioactive beverages like coffee during pregnancy. Also to avoid smoking, and to get some exercise and sun exposure.

If you’re doing all these things, I don’t think you need to be concerned. Ana says, “I feel great, better than before”; that’s good evidence that she’s well prepared for a healthy pregnancy.


The paper presents solid evidence that hypermethylation of RXR-alpha in the womb predisposes children to become obese at age 9.

The paper gives us essentially no evidence at all as to what causes hypermethylation of RXR-alpha in the womb, except that it correlates with low carbohydrate consumption in the women of Southampton UK.

I hate it when journals do this. If you’re going to link carb intake to RXRA methylation, give some real data and analysis. Probably the authors are saving their dietary analysis for a future paper. The carb graph was included as a “teaser” to make the work seem more interesting.

There are many known health dangers which are known risk factors for obesity and which correlate with low carbohydrate consumption in the general population. So until more evidence emerges, I think there’s little here for low-carb Paleo dieters to be concerned about.


[1] Godfrey KM et al. Epigenetic gene promoter methylation at birth is associated with child’s later adiposity. Diabetes. 2011 May;60(5):1528-34.

Low Serum Cholesterol in Newborn Babies

Don Matesz, who has embraced low-fat and low-cholesterol dieting, recently stated that “I now consider anything over ~160 mg/dl [to be] excess serum cholesterol” and cited in his support the Cordain-Eaton claims that healthy hunter-gatherers had low serum cholesterol. Of course, we looked at that and found that healthy hunter-gatherers generally had serum cholesterol over 200 mg/dl and that hunter-gatherers with low serum cholesterol generally had high infectious burdens and short lifespans. See:

When Erik referenced our series and asked, “What do you think of the argument that low cholesterol in hunter gatherer populations stems from infections and parasites?”, Don replied:

Mean total blood cholesterol of healthy human neonates is about 72 mg/dl.

Is this due to infections and parasites?

In case this question was not merely rhetorical, let me answer: No.

But it’s an interesting biology question. Why do neonates have low serum cholesterol?

Neonates and Infants

The study that Don cited [1] looked at cord blood from neonates. Cord blood is blood that circulates on the fetal side of the placenta in utero. As soon as the baby is delivered, the cord is cut and blood ceases to circulate.

So the cord blood serum cholesterol of 70.3 mg/dl is really sampling fetal cholesterol – the blood of babies who have never eaten and never breathed.

The not eating part is relevant, because HDL is generated from the metabolism of chylomicrons created in the intestine when fat is eaten, and LDL is generated from VLDL particles that carry excess calories as triglycerides from the liver. So eating generates LDL and HDL. We might expect that LDL and HDL, and thus TC, levels will rise as soon as the neonate starts feeding.

We can check this out by looking at cholesterol levels in infants. The following data is from Japan [2], but any healthy population would give similar results:

Serum total cholesterol in infants, mg/dl, by feeding method

Infant Age Formula-fed Partially breastfed Breastfed
One month 117 142 163
Six months 140 162 194

Source: Tables 2 and 3, Isomura et al 2011.

The key data is in the rightmost column, the breastfed babies. By one month postpartum, TC is 163 mg/dl (“excess serum cholesterol” on Don’s view). By six months, it is 194 mg/dl.

Formula fed babies had a much smaller rise in TC.

To understand the pattern of this data, let’s look at three issues:

  • Why do formula-fed babies have lower TC than breastfed babies?
  • Why do neonates have low TC?
  • Why do breastfed babies end up with TC near 200 mg/dl?

Formula is a lipid-deficient food

Why do formula fed babies have lower serum cholesterol? One contributing factor may be a dietary lipid deficiency.

Human breast milk is rich in cholesterol. One study found that the cholesterol content of human breast milk follows a diurnal rhythm with a low of 140 mg/L during sleeping hours and early morning, and a high of 220 mg/L in the afternoon and evening. Other studies agree that human breast milk always has more than 100 mg/L cholesterol. Babies typically drink 750 mL/day, so a breastfed baby’s daily cholesterol intake is 100 to 200 mg.

Scaled by body weight, this would be the equivalent of 1.5 to 3 grams cholesterol per day for adults – approximately ten times the typical cholesterol intake of American adults.

Clearly, evolution thinks babies should get plenty of cholesterol.

But cholesterol levels in formula are much lower:

Since … infant formulas contain very little cholesterol (10 to 30 mg/L) (Huisman et al., 1996; Wong et al., 1993), it is not surprising that plasma cholesterol concentrations are higher in infants fed human milk than in formula-fed infants.

I guess the formula makers don’t consider cholesterol to be a desirable nutrient. This may be an extremely consequential mistake.

Low TC in Neonates May Have Evolved to Suppress Immunity

So why do neonates have a very low TC?

In addition to fat and cholesterol transport, LDL and HDL both have immune functions. Low serum cholesterol signifies a loss of these immune functions. Normal immune function is associated with TC around 200 mg/dl or higher.

But infants are well known to have suppressed immunity. This is important: if the fetus had an ability to generate antibodies and mount an immune response, it might generate immune attacks against the mother leading to miscarriage.

After birth, a baby’s immune system gradually matures:

A baby’s immune system is not fully developed until he/she is about six months-old. In the meantime, pregnant mothers pass immunoglobulin antibodies from their bloodstream, through the placenta, and to the fetus. These antibodies are an essential part of the fetus’s immune system. They identify and bind to harmful substances, such as bacteria, viruses, and fungi that enter the body. This triggers other immune cells to destroy the foreign substance….

Immediately after birth, the newborn has high levels of the mother’s antibodies in the bloodstream. Babies who are breastfed continue to receive antibodies via breast milk…. This is called passive immunity because the mother is “passing” her antibodies to her child. This helps prevent the baby from developing diseases and infections.

During the next several months, the antibodies passed from the mother to the infant steadily decrease. When healthy babies are about two to three months old, the immune system will start producing its own antibodies. During this time, the baby will experience the body’s natural low point of antibodies in the bloodstream. This is because the maternal antibodies have decreased, and young children, who are making antibodies for the first time, produce them at a much slower rate than adults.

Once healthy babies reach six months of age, their antibodies are produced at a normal rate.

LDL particles, by presenting pathogen toxins to macrophages which can then present them on MHC molecules, play an important role in the generation of antibodies. (See Blood Lipids and Infectious Disease, Part II, July 12, 2011.) Low LDL signifies a reduced ability to generate antibodies.

Low LDL is therefore highly desirable as long as the baby remains in the womb, and in fact LDL levels are very low in utero.

But persistent low LDL after birth is dangerous: it makes the infant vulnerable to infections. Likewise, HDL has important immune functions (see HDL and Immunity, April 12, 2011). So LDL and HDL gradually rise to normal physiological levels, finally reaching a TC of 200 mg/dl after 6 months in breastfed babies – precisely when the babies attain normal immune function.

If TC of 190 mg/dl or higher signifies normal immune function, then formula fed babies are still immune suppressed at 6 months. Extrapolating the rise in TC, partially breast fed babies might achieve normal immune function at 12 months and formula fed babies might not achieve normal immunity until age 24 months!

Immunity Matters for Infant Health

I don’t want to delve too deeply into this, but infants are vulnerable to infections – this is why infant mortality has always been high. It still is today, and 6 months of age is still the canonical age when the danger lessens:

Globally, approximately 4,000,000 children less than 6 months of age die each year at a rate of 450 deaths per hour. In addition, high hospitalization costs for infected infants are incurred in the United States with an annual estimated cost of $690,000,000.

Formula feeding definitely escalates the risk:

In the United States, more than 40% of all infant hospitalizations are attributable to infectious disease … Diarrhoeal diseases and digestive tract infections are the most common infectious diseases in infants….

Breast feeding has been shown to have a number of beneficial effects in infants, including protection against infectious and allergic diseases. [3]

In this study, 41% of formula-fed infants developed infections between ages 5 and 8 months. [3]

A study from Brazil [4] shows that breastfeeding makes a huge difference in infant mortality:

In a population-based case-control study of infant mortality in two urban areas of southern Brazil, the type of milk in an infant’s diet was found to be an important risk factor for deaths from diarrhoeal and respiratory infections. Compared with infants who were breast-fed with no milk supplements, and after adjusting for confounding variables, those completely weaned had 14.2 and 3.6 times the risk of death from diarrhoea and respiratory infections, respectively. Part-weaning was associated with corresponding relative risks (RR) of 4.2 and 1.6. [4]

Now, deficient serum cholesterol is not the sole factor accounting for higher mortality in formula fed babies. But it is a contributing factor.


If serum cholesterol is healthiest below 160 mg/dl, then formula fed babies have excellent blood lipids despite a high disease and mortality rate, but breastfed babies are already in trouble at age one month and are suffering a shocking dyslipidemia at age six months, despite excellent health.

I think that’s absurd. A more logical interpretation of the evidence is this.

Healthy babies achieve serum cholesterol levels around the adult norm of 200 mg/dl by age six months.

Serum cholesterol levels below 190 mg/dl or so indicate immune suppression and increased risk of infectious disease – whatever the age of the human in question. Formula fed babies are immune suppressed for an extended period – well beyond the six month period of a healthy breastfed baby.

There are multiple causes of low serum cholesterol. A high infectious burden is one; never having eaten is another; a lipid-deficient diet is a third. But there is no evidence I am aware of suggesting that low serum cholesterol is a desirable condition.


[1] Mishkel MA. Neonatal plasma lipids as measured in cord blood. Can Med Assoc J. 1974 Oct 19; 111(8):775-80.

[2] Isomura H et al. Type of milk feeding affects hematological parameters and serum lipid profile in Japanese infants. Pediatr Int. 2011 Mar 21.

[3] Picaud JC et al. Incidence of infectious diseases in infants fed follow-on formula containing synbiotics: an observational study. Acta Paediatr. 2010 Nov;99(11):1695-700.

[4] Victora CG et al. Evidence for protection by breast-feeding against infant deaths from infectious diseases in Brazil. Lancet. 1987 Aug 8;2(8554):319-22.

The Danger of Protein During Pregnancy

At we’re advocates of protein restriction. We recommend:

  • Avoiding all protein-containing plants, as plant proteins tend to be toxic;
  • Striving to eat fatty, not lean, meats and fish, in order to keep protein intake down and fat intake up.

Protein restriction helps protect against viral and bacterial infections by promoting autophagy, the process of intracellular protein scavenging, digestion, and recycling.  During autophagy, bacteria and viruses, as well as junk human proteins and damaged organelles, are digested.  Autophagy has been strongly linked to longevity [1] and is protective against many diseases.

Our advocacy of low protein intake separates us from many other Paleo bloggers.  Loren Cordain, the dean of the Paleo movement, has long advocated consumption of lean meats.  Although he has moderated his stance somewhat, the front page of his site still places lean meats first among his favored foods:

Learn how a diet based on lean meats

The Paleo Diet is a way of eating in the modern age that best mimics diets of our hunter-gatherer ancestors – combinations of lean meats

(The other major difference we have with Dr. Cordain is his exclusion of starchy foods from a “Paleo” diet, even though starchy tubers have been part of the ancestral human diet for 4 million years. But that is a story for another day.)

Those who have read the pre-publication draft of our book know that we place high store on human breast milk as an indicator of the optimal composition of the human diet.  Human breast milk provides only 7% of calories in the form of protein. (Carbs are about 38% and fats about 55%.) One can debate whether 7% is the right level of protein for adults; but, if the principle of natural selection is sound, it must be that infants need a low-protein diet.

Science bears this out.  As our book notes, diets containing 20% of calories as protein are highly toxic to infants. Pre-term infants fed 20% protein diets had more fever, lethargy, and poor feeding than infants fed 10% protein diets, and lower IQs at ages 3 and 6 years. [2] Even a slight increase in the protein content of formula, from 7% to 9%, significantly increased the likelihood that babies would be overweight by age 2. [3]

Given our skepticism toward high-protein diets, especially for babies, we were pleased to see Dr. Cordain in his most recent newsletter [The Paleo Diet Update v6, #20 – Protein Intake for Pregnant Women] acknowledge the dangers of high protein intake by pregnant mothers. Dr. Cordain advises a pregnant mother:

[Y]ou probably should increase your fat and carbohydrate consumption, and limit protein to about 20-25% of energy, as higher protein intakes than this may prove to be deleterious to mother and fetus for a variety of physiological reasons….

“Protein intakes above this [25% of total calories] threshold may affect pregnancy outcome through decreased mass at birth and increased perinatal morbidity and mortality.” [4]

The physiological basis for this aversion stems from a reduced rate of urea synthesis during pregnancy that is evident in early gestation [5] as well as increases in the stress hormone cortisol [6]. Hence, pregnant women should include more carbohydrate and fat (i.e. fattier meats) in their diets and limit dietary protein to no more than 20-25% of their total caloric intake.

What are the long-term effects of a high-protein diet during pregnancy on the offspring?  In long-term follow-up studies of the adult children of mothers who ate high protein diets while pregnant between 1948 and 1954, it was found that by age 40 offspring commonly had high levels of the stress hormone cortisol [6] and high blood pressure [7,8].  The effects of faulty maternal diets can be long-lasting.

At, we think 20% is still likely to be a bit more protein than is desirable. We would advise pregnant mothers to restrict protein to about 15% of calories and to strive to obtain 30% of calories as carbohydrates.  As long as adequate carbs are obtained, there is only a modest need for protein and as little as 10% of calories as protein may be sufficient.

Note that this advice is very close to the ratios of 30% carb, 15% protein, and 55% fat that we recommend to adults and children generally.  Pregnant women may benefit from slightly more starch and slightly less protein than others; but on the Perfect Health Diet, pregnancy should not require a significant change in eating habits.

[1] Jia K, Levine B. Autophagy is required for dietary restriction-mediated life span extension in C. elegans. Autophagy. 2007 Nov-Dec;3(6):597-9.

[2] Goldman HI et al. Clinical effects of two different levels of protein intake on low-birth-weight infants. J Pediatr. 1969 Jun;74(6):881-9. Goldman HI et al. Effects of early dietary protein intake on low-birth-weight infants: evaluation at 3 years of age. J Pediatr. 1971 Jan;78(1):126-9. Goldman HI et al. Late effects of early dietary protein intake on low-birth-weight infants. J Pediatr. 1974 Dec;85(6):764-9.

[3] Koletzko B et al; European Childhood Obesity Trial Study Group. Lower protein in infant formula is associated with lower weight up to age 2 y: a randomized clinical trial. Am J Clin Nutr. 2009 Jun;89(6):1836-45.

[4] Speth JD. Protein selection and avoidance strategies of contemporary and ancestral foragers: unresolved issues. Philos Trans R Soc Lond B Biol Sci. 1991 Nov 29;334(1270):265-9; discussion 269-70.

[5] Kalhan SC. Protein metabolism in pregnancy. Am J Clin Nutr. 2000 May;71(5 Suppl):1249S-55S.

[6] Herrick K et al. Maternal consumption of a high-meat, low-carbohydrate diet in late pregnancy: relation to adult cortisol concentrations in the offspring. J Clin Endocrinol Metab. 2003 Aug;88(8):3554-60.

[7] Campbell DM et al. Diet in pregnancy and the offspring’s blood pressure 40 years later. Br J Obstet Gynaecol. 1996 Mar;103(3):273-80.

[8] Shiell AW et al. High-meat, low-carbohydrate diet in pregnancy: relation to adult blood pressure in the offspring. Hypertension. 2001 Dec 1;38(6):1282-8.