What is microbial diversity anyway?

We hear about the diversity of our gut microbiota quite a lot nowadays. But what that actually means is often a little vague.

Microbiotaologists calculate diversity by borrowing equations developed by those who have spent much longer studying the diversity of ecosystems, they would be ecologists.

While there are lots of different equations available, and often more than one is used in research, if you get your personal gut microbiota tested with a company like uBiome you will get one number back comparing yourself to everyone tested. In uBiome’s case, they currently seem to use a measure called the Inverse Simpson index.

Don’t worry about the equations (unless you enjoy that sort of thing (in which case you = strange) x 10). Put simply, his equation takes into account both how many species are there and also how many of each type of species are there.

Imagine a woodland of one hundred trees and ten different species of trees. If one species has ninety-one trees and there is only one tree each of the other nine species the woodland isn’t very diverse. But if there are ten trees of each species the woodland is much more diverse. The Simpson index tries to quantify this difference.

Using the Simpson index the diversity number gets smaller as diversity increases, which is counterintuitive, so it is usually divided by 1 to make the numbers go the right way around.

Hence we have the Inverse Simpson index.


Let’s make a very, very simplified gut microbiota with 10 species of bacteria with a couple of common species and 8 rarer species.

Bacteria Number of bacteria Percentage of each species
Species 1 20 40%
Species 2 18 36%
Species 3 2 4%
Species 4 2 4%
Species 5 2 4%
Species 6 2 4%
Species 7 1 2%
Species 8 1 2%
Species 9 1 2%
Species 10 1 2%
Total bacteria 50

It is important to note that when testing the gut microbiota current techniques can only show the percentages of each type of bacteria, but not how many of each type are there.

The diversity of this simple microbiota using the Inverse Simpson diversity is:

Diversity = 3.4

Now let us say that hypothetically we feed this gut microbiota with a diet containing diverse types of dietary fibre that feeds more of the rarer species of bacteria.

Bacteria Number of bacteria Percentage of each species
Species 1 25 25%
Species 2 22 22%
Species 3 7 7%
Species 4 7 7%
Species 5 7 7%
Species 6 7 7%
Species 7 7 7%
Species 8 6 6%
Species 9 6 6%
Species 10 6 6%
Total bacteria 100

The total number of bacteria has doubled while the percentages are spread more evenly between the different species of bacteria.

Diversity = 7.3

The diversity has doubled because the evenness has increased.

But now let us try a diet with much less dietary fibre reducing the food for any of the bacteria to very low levels.

Bacteria Number of bacteria Percentage of each species
Species 1 4 20%
Species 2 4 20%
Species 3 2 10%
Species 4 2 10%
Species 5 2 10%
Species 6 2 10%
Species 7 1 5%
Species 8 1 5%
Species 9 1 5%
Species 10 1 5%
Total bacteria 20

The rare bacteria remain rare but the previously common ones are much reduced.

Diversity = 8.3

The diversity number has now increased even more because we have lost the more common bacteria and so the percentages are again spread more evenly between the different species of bacteria.

Remember we can only see the percentages of each type of bacteria of the total, we can’t tell how many of each are there.

While this hypothetical example is incredibly simplified you can see that a single measure of diversity can be misleading when considering what is going on in the gut. It is difficult to interpret without taking into account what might be causing changes in diversity, what types of bacteria are making up that diversity, and what those bacteria are doing there. A low diversity gut microbiota dominated by a few beneficial bacteria could be better than a diverse one containing more pathogens.

Currently, the information available if you are having your gut bacteria tested as a consumer is limited due to the nature of the technology. In light of this, it is worth interpreting the microbial diversity with caution.


Posted in Articles, microbiota, Nutrition, Research, science, Uncategorized | Leave a comment

Has modern farming increased the retinol in the liver you eat?

Liver is about the most concentrated dietary source of pre-formed vitamin A (retinol) that you could hope to find, along with a number of other vitamins. Like in us humans, animals store retinol in their liver and the more they consume the more gets stored in their liver.

Historically, plant-eating livestock like cows and sheep would have sourced their vitamin A in the form of beta-carotene from the plants they ate, which they convert into retinol as required. Other more omnivorous animals like chickens and pigs may have gotten some pre-formed retinol in their diets, depending upon what they could eat.

The advances in animal nutrition in during the 20th Century have resulted in much more targeted livestock nutrition to optimise animal health and productivity. This includes the widespread use of nutritional supplements in modern livestock farming. When vitamin A is supplemented in the diet it is usually added in the form of retinol.

In 1992 M J Scotter and colleagues published an analysis of the liver retinol content of different farm animals in the UK in a survey of animal livers for vitamin a content.

In Table 1 (below) they compared their results to previous reports going back to 1960. You can see that, compared to 1960, the liver retinol content of all the animals increased if you look at the mean values. Looking at the range of individual results there were also more animals with really high levels of liver retinol. All groups showed a wide range of retinol content.

(There was no measure for chicken liver in 1960 to compare with)

Liver retinol

A later paper from 1998 reported that vitamin A levels hadn’t changed much during the 1990s. A survey of vitamin A concentrations in the liver of food-producing animals.

Liver retinol 2

They concluded:

“The concentrations of liver vitamin A were similar to those found in other surveys (Scotter et al. 1992). There was no evidence that vitamin A concentrations in liver had increased over the past decade, suggesting that the vitamin A supply in animal diets from all sources, including supplements, had remained at a similar level throughout.”

Another more recent study in 2006 found that the amount of retinol in the livers of farm animals in Austria was a lot lower than from other European countries with more industrial farming systems. Vitamin A content (retinol and retinyl esters) in livers of different animals

“In the small- and medium-sized farms (dominating in the agricultural production in Austria), animal breeding is being practised less extensively, without higher vitamin A supplementation, than in other European countries, such as Germany or the UK. Many of the farms tend to be quite self-sufficient and only a few of them buy modest amounts of commercial feeding supplies for the animals. Moreover, the slaughter of the animals was done on the farm or in the closed slaughter-house. Therefore the application of the vitamin A preparation to minimise the stress before slaughter could be excluded.”

As an example of the results:

“The mean amount of vitamin A observed in Austrian calf livers was one-fifth (1/5) of the Danish (10.3 mg RE/ 100 g), one-ninth (1/9) of the British (18.8 mg RE/100 g) and one-fifteenth (1/15) of the German (28.8 mg RE/ 100 g) average liver vitamin A contents.”

“The Austrian cattle showed a mean liver vitamin A content about 4.5–5.5-times lower than those of German (17.9 mg RE/100 g) and Danish cattle (22.3 mg RE/100 g)…”

“It was observed that the average amount of vitamin A measured in Austrian pig livers was…  just a third of that found in German pig livers (36.3 mg RE/100 g) (BgVV, 1995). The”

“The average content of vitamin A found in chicken livers was about 2, 4 and 6-times lower than the mean values reported for British (9.7 mg RE/100 g), for Slovak Republic (17–20 mg RE/100 g) and German chickens (33.5 mg RE/100 g), respectively.”

They concluded that:

“The variations of liver vitamin A concentrations among the species were a result of differences in race and age, but especially of different contents of vitamin A in the supplied feed.”


The liver you eat today may contain rather more vitamin A than the liver your grandparents ate when they were young. Whether this is of nutritional importance to you eating it is less certain. Liver is a valuable source of many nutrients and remains an important part of my diet.

One situation it may be of relevance is the question of liver and retinol consumption during early pregnancy, which is a topic beyond this blog post.

Perhaps the moral of this story is that what you food eats can be important and that modern farming can alter our food in ways that you might never think about.


Scotter MJ, Thorpe SA, Reynolds SL, Wilson LA, Lewis D J. Survey of animal livers for vitamin a content. Food Additives and Contaminants 9, 237–242 (1992).
Retail samples of livers from calf (23), ox (18), lamb (17), pig (15), chicken (16) and turkey (1) were analysed to determine levels of vitamin A (all-trans-retinol) and to aid assessment of the effects of using vitamin supplemented compound feedingstuffs for livestock. For comparison, 22 liver samples from lambs reared on diets not containing vitamin‐supplemented compound feedingstuffs and four samples of liver from ox which had received supplemented feed but not during the last four months prior to slaughter were also analysed. The chosen method of analysis utilized saponification, solvent extraction and normal‐phase high-performance liquid chromatography with fluorescence detection. For all species analysed, the levels of vitamin A ranged from 10 to 1100 mg/kg, with all but seven at or below 400 mg/kg. For lamb and ox livers, the mean levels were 310 mg/kg and 200 mg/kg respectively for retail samples. The mean levels were 220 mg/kg (lamb) and 120 mg/kg (ox) in liver samples from animals fed controlled diets. The results are of the same order as those reported over recent years.
Howells LC, Livesey CT. A survey of vitamin A concentrations in the liver of food-producing animals. Food Additives and Contaminants. 15, 1, 10-8 (1998)
“High oral doses of vitamin A have been shown to be teratogenic and accordingly the World Health Organization has recommended that the daily intake for pregnant women should not exceed 3.3 mg. Liver contains high concentrations of this vitamin and consequently it has become necessary to assess its level in the livers of food animals. A survey of fresh, frozen, imported and home-produced retail liver samples from calf (42), ox (121), lamb (228), pig (133) and chicken (125) was carried out between August 1992 and April 1993 using a modified version of the method of Ashoor and Knox (1987) to determine retinol and retinyl esters. Interlaboratory comparison showed no significant difference in results for this method and a method which employed hydrolysis of retinyl esters, and there were significant advantages in specificity, simplicity, cost and quality control. The mean +/- sd total vitamin A concentration calculated as retinol equivalents, for all species surveyed was 139 +/- 96 with a range of 3-1267 mg/kg liver. Mean values for individual species were: calf 188 +/- 125, ox 142 +/- 110, lamb 173 +/- 104, pig 174 +/- 118 and chicken 97 +/- 44 mg/kg liver. No change in the vitamin A concentrations in liver were observed compared with previous surveys in the UK, which suggests that the intake of vitamin A from dietary and supplementary sources by different species has not changed. The data from this survey confirm that liver cannot be recommended as a food for pregnant women.”
Majchrzak D. Fabian E. Elmadfa I. Vitamin A content (retinol and retinyl esters) in livers of different animals. Food Chemistry. 98, 4, 704-710 (2006)
In the present study, 90 animal livers of five different species (pig, cattle, calf, chicken, turkey) were examined for their vitamin A contents. The investigation of extracted vitamin A included all-trans retinol, retinyl palmitate, stearate, oleate and linoleate, expressed as retinol equivalents (RE). The separation of the various chemical forms was done using HPLC. The liver vitamin A contents ranged between 6.5 and 18.9 mg RE/100 g in pigs, from 1.1 to 6.7 mg RE/100 g in cattle and from 1.6 to 16.6 mg RE/100 g and 2.7 to 21.5 mg RE/100 g in chickens and turkeys, respectively. The livers of calves contained the smallest amount of vitamin A, with variation from 1.3 to 3.2 mg RE/100 g. Retinyl palmitate was the predominant form of vitamin A in the livers of investigated animals and contributed about 40% (avids) up to 75% (calf) of the total liver vitamin A contents. The results indicated that the lower levels of animal liver vitamin A, observed in our study, could be a result of small-structured agriculture in Austria. The variations of liver vitamin A concentrations among the species were a result of differences in race, age and the different feeding regimen.


Posted in Articles, Food, Nutrition, Research, science, Uncategorized | Tagged , , | 11 Comments

Does bacterial lipopolysaccharide (LPS) from the gut cause obesity?

Does lipopolysaccharide (LPS), also known as endotoxin, cause obesity? Reasons to think it might (in mice at least) come from a high-profile paper published in the journal Diabetes in 2007 and titled “Metabolic Endotoxemia Initiates Obesity and Insulin Resistance” (cited 3045 times according to Google Scholar, which is a lot).

LPS makes up a key part of the cell wall of many types of bacteria that live in the intestines and when these bacteria die the LPS is released into the gut. The bacteria in the gut contain a lot of LPS, which is okay as long as it stays in the gut. LPS triggers an inflammatory response if it gets into the body and is one of the main molecules that the body senses to detect the presence of bacteria.

Briefly, the authors of this study reported that feeding mice a high-fat diet* makes them fat and increases the amount of LPS leaking through into the blood from the gut. Mice with a gene knockout (CD14) that prevents the mice responding to LPS did not get as fat.

*”Mice were fed a control (A04, Villemoisson sur Orge, France) or a high-fat, carbohydrate-free diet for 2 or 4 weeks following protocols. The diet contained 72% fat (corn oil and lard), 28% protein, and <1% carbohydrate as energy content (17).”

The high-fat diet used in this study was a bit unusual but that is beyond this blog post.

This is quite a long and complex paper, but let’s focus on one of the key findings. Infusing LPS into mice over four weeks increased mouse body weight and body fat (highlighted in yellow). This seemed to confirm a direct role for LPS in making the mice fatter.

Cani 2007

In a follow-up paper, the authors hypothesised that LPS might play a key role in the development of metabolic disorders.


Is this replicable?

While exciting findings make for big papers it is important to know whether these results have been replicated. This was not easy to find and it probably would have altered my PhD research had I found the following rather earlier.

I discovered, much later, that the original authors had repeated this LPS infusion study in a later 2013 paper titled “Chronic endocannabinoid system stimulation induces muscle macrophage and lipid accumulation in type 2 diabetic mice independently of metabolic endotoxaemia“. In this study, they infused LPS over six weeks into mice fed both a control diet and a high-fat diet.

It appeared that they could not replicate the effects seen in the first study, although the wording of the text does not make this obvious.

Geurts 2013

If you look at Table 2 there were no significant effects of LPS infusion on body fat or weight gain compared to control mice (CT). Neither were there any significant effects of LPS in high-fat diet fed mice.

While the highlighted text says that LPS increased weight gain by 44%, if you look at the figures in Table 2 LPS treatment only increased the body weight gain by 24%, which was not significant (this could be a typo in the paper). While LPS did “increased the adipose tissue weight compared with the control mice” this was only by 0.03 grams (30 milligrams), not a meaningful difference. Effects of LPS on blood glucose or insulin were also not statistically significant.

Has anyone else repeated these results?

For a long time after, I struggled to find other studies that had tried to repeat these findings. Less exciting results tend to be less emphasised in papers or get published in less prominent journals. I eventually found them by searching for the brand of osmotic minipump that was used to infuse the LPS into mice or rats. So what papers did I find?

In Mice

Vila, 2014: Four weeks of LPS infusion…

“This treatment induced a 2-fold increase in LPS plasma level (Figure 7A), with no change in body weight (Figure S7A).”

Liang, 2014: Sixteen weeks of LPS infusion…

“Table 1 shows data of risk factors that are typically associated with NASH development in humans, for example, body weight, visceral fat mass, plasma lipids, fasting glucose and insulin, and ALAT. These risk factors were not affected by LPS.”

Iwashita, 2013: Four weeks of LPS infusion…

Body weights of all mice rose slightly for 4 weeks and the differences between groups were not significant (Fig. 1A). While fasting plasma glucose levels differed slightly among the groups after 4 weeks of LPS or PBS infusion, the differences did not reach statistical significance (Fig. 1Bi). There were no significant differences in GTT results among the groups (Fig. 1Bii).”

“Subsequently, the effect of a higher dose (1.0 ng/g/h) of LPS for 4 weeks was also investigated. None of the groups showed alterations in body weight or fasting glucose concentrations (data not shown).”

Nøhr, 2016: Twenty-eight days of LPS infusion…

“In the first few days following implantation of osmotic mini-pumps, LPS mice regardless of resveratrol dropped (≈ 10%) in body weight (Fig 1B). After 28 days of treatment, no differences in body weight were seen between the groups (Fig 1C).”

I found one paper that seemed at first to support the original findings in mice.

Ahn, 2015: Four weeks of LPS infusion…

“However, body weight and relative weight of epididymal fat were increased in the LPS group compared with the NC group (Fig. 1B and C). The area of fat tissue shown by MRI also increased (Fig. 1D). Interestingly, as shown by the results of H&E staining, the epididymal fat tissue in the LPS group showed hypertrophy (Fig. 1E), whereas there were no significant differences in other tissues (Supplementary Fig. 2).”

However, when looking at the results of this study the differences between groups is small.

Ahn weights

These differences were quite small and it is worth being cautious of data presented as percentage weight change without any details of the actual body weight.

In Rats

Hsieh, 2008: Four weeks of LPS infusion…

“As shown in Table 2, there were no significant differences in body weight, mean arterial blood pressure, fasting plasma glucose and insulin, AST, ALT, and albumin levels between the two groups during the study period.”

Wu, 2012: Fourteen days of LPS infusion…

“On day 14 after IP infusion of the endotoxin, there is no significant change in body weight (266±3 vs. 260±6 g, P>0.05, n=16), body temperature (36±1 vs. 37±2°C, P>0.05, n=16), daily food (22±5 vs. 23±5 g/day, P>0.05, n=16), or water (47±5 vs. 50±7 mL/day, P>0.05, n=16) intake between the saline- and LPS-treated animals. Chronic.”

Fischer, 2015: Four weeks of LPS infusion…

“Our results show that LPS infusion in male rats for 4 weeks neither increased body weight nor affected insulin sensitivity.”

“Furthermore, there was no difference in fat mass between groups…”

La Serre, 2015: Six weeks of LPS infusion…

“…there was no significant difference in final body weight between LPS-treated and LF-fed rats.”

“Chronic LPS treatment led to a redistribution of adiposity characterized by a decrease in epidydimal fat (8.8 ± 0.2 vs. 10.7 ± 0.8 g, p < 0.05) and an increase in mesenteric fat (8.3 ± 0.4 vs. 7.0 ± 0.2 g, p < 0.05) (Fig. 2C).”

In La Serre, 2015 total body fat remained about the same but there were small differences in the weight of different body fat stores. However, the reduction in epididymal fat is contrary to that found by Ahn, 2015 above.


Out of the five mouse studies and four rat studies that had both infused LPS, only one showed any increase in body weight and body fat, and that was small. The others studies reported no increase in body weight or fat after LPS infusion. This suggests that LPS is not a reliable way to make rodents fat.

I think this is a good example of negative results being rather difficult to find in the scientific literature, even while the original paper which showed a positive result retains a high profile and remains highly cited. Less interesting results tend to be relegated to a sentence in the results section that easily goes unnoticed, while the papers themselves focus on some other aspect of the study.

Posted in microbiota, Research, science, Uncategorized | Tagged , , , , | 1 Comment

Graphic faecal transplants and obesity

The gut microbiota has been a popular area of research over the past few years and few aspects of this research have gained quite as much attention as the role of the gut microbiota in obesity.

One of the studies that started this was published in Nature in 2006. This study has been rather influential as Google Scholar states that, as of today, this study has now been cited 5,842 times.

graphic 3


The reason for its importance is that it was the first study to show that transplanting the microbiota from obese mice into lean mice made the mice fatter than mice getting transplants from lean mice.

Some previous studies had reported differences in the composition of the gut microbiota between obese and lean individuals, but this was the first study to provide evidence that this might be causal.

The key graph in the paper for this is Figure 3c (yellow highlights are mine).


graphic 2


In germ-free mice transplanted with gut microbiota from obese mice, body fat increased by 47% compared to 27% getting a microbiota from lean mice.

While this looks like a big difference, it is worth noting that this is not body fat percentage. This is the percentage increase in fat in relation to the initial body fat. This is difficult to interpret without knowing how much fat the mice started with.

The study does not state the starting body fat of the mice, but it does tell us in the results how much fat the mice gained in grams.




As we know the percentage increase in body fat and the total weight of that increase, we can calculate the amount of fat each group of mice had before the faecal transplant.

Mice getting a faecal transplant from obese mice gained 1.3 grams of fat, an increase of 47% relative to their starting fat.

Mice getting a faecal transplant from lean mice gained 0.86 grams of fat, an increase of 27% relative to their starting fat.


If we calculate their starting fat:

Lean transplant recipients: 1.3/0.47 = 2.76

Mice getting obese mice transplants had 2.76 grams of fat at the start.

Obese transplant recipients: 0.86/0.27 = 3.19

Mice getting lean mice transplants had 3.19 grams of fat at the start.


If we add those up…

2.76 + 1.3 = 4.05

3.19 + 0.86 = 4.07

The final body fat of both groups was almost the same at about 4.1 grams


If we graph the total weight of the body fat in the two groups of mice:


(This graph probably wouldn’t get you a paper in Nature).

As you can see, while the mice getting a faecal transplant from the obese mice gained a little more fat, they ended up with the same body fat at the end of the study as the mice getting faecal transplants from lean mice. As mice at this age would typically weigh about 25 grams, at 4 grams of body fat both groups of mice would be considered lean and therefore neither group of mice were obese after the transplantation.

This suggests that the results may not be quite as exciting as they first seemed.

Percentage increases in fat do have valid uses in science but they can also be used to make minimal differences look larger and to make them statistically significant.

It is always worth reading graphs carefully.

Posted in Articles, microbiota, Research, science, Uncategorized | Tagged , , , , , , , | 10 Comments

Low-carb diets, folate, and neural tube defects.

A new study was published this week that attracted some attention.

Low carbohydrate diets may increase risk of neural tube defects

Zoe Harcombe wrote a blog post about it detailing criticisms of the study that did not appear particularly valid from my reading of the study. As these criticisms were being shared on social media today I thought I would write a quick response. Quoted sections are from the linked blog post.

“A study was published in the journal “Birth Defects Research” on 25th January 2018 (Ref 1). A press release accompanied it (Ref 2) claiming that “women with low carbohydrate intake are 30 percent more likely to have babies with neural tube defects, when compared with women who do not restrict their carbohydrate intake.” Newspapers worldwide reported the story (Ref 3), reiterating the scary headline that “Women on low carb diets may be at 30 percent greater risk of having a baby with a spinal and neurological birth defect, according to a new study.”

One of my lovely subscribers, Belinda Fettke, alerted me to the article and asked me to take a look at it as my weekly newsletter. So I did. It was a case control study. It was fundamentally flawed in a number of ways:

1) The characteristics table compared the control with the control, not the control with the cases.”

Exclusion of NTD cases is deliberate and is detailed in a footnote to Table 1. Table 1 compares the folate intakes and other characteristics of mothers eating lower-carb diet and those eating more carbs. This shows that estimated folate intakes were lower even when mothers with NTD cases were excluded in both groups. 

folate 6


folate 7

“2) The ‘correct’ characteristics table was available, as a Supplemental, but neither table included important data related to the study – not least carbohydrate and folate/folic acid intake.”

Table 1 does include estimated daily folate/folic acid intakes and does include the categories of carbohydrate intake in the table footnotes, although it does not detail the average intake of carbohydrate in each category.

folate 1


folate 1

This indicates that even when excluding cases of neural tube defects women in the restricted carbohydrate group (<95g) were estimated to be consuming less than half the folate of the nonrestricted group.

“3) Consequently the study did not adjust for material differences between the control and case groups.”

Incorrect, it was adjusted as is detailed in the methods section.

folate 2

“4) The study could not make the conclusion that it did.”

Yes, it could.

“I sent the email below to the authors on 30th January and sent a copy to the journal editor. I was minded to wait until I received a reply before posting this openly, but I have been asked to post this by a number of people who know that I found retraction-level errors because the headlines generated by this article need to be countered immediately. Pregnancy is worrying enough for women and men without fabrications like this trying to scare the life out of them.

Dear Dr Desrosiers,

I am currently reviewing your very interesting paper “Low carbohydrate diets may increase risk of neural tube defects.”

The article abstract reported “To assess the association between carbohydrate intake and NTDs [Neural tube defects], we analyzed data from the National Birth Defects Prevention Study from 1,740 mothers of infants, stillbirths, and terminations with anencephaly or spina bifida (cases), and 9,545 mothers of live born infants without a birth defect (controls) conceived between 1998 and 2011.” The article reported that these numbers became 1,559 (cases) and 9,543 (controls) with valid exclusions. That’s fine.”

Table 1 reported the characteristics of the control group only. Table 1 should show the differences between the cases and the controls, not least so that it is known what to adjust for. Supplemental Table 1 reported this data.”

Incorrect, this is a case-control study looking at differences between exposure to different categories of dietary carbohydrate intake, and by inference folate intake, so this is the correct table.

“Q1) Please can you explain why Table 1 is in the main paper and not Supplemental Table 1?”

See above.

“Q2) Please can you add the carbohydrate intake data to Supplemental Table 1 and please can you add the folate/folic acid intake data to Supplemental Table 1? These are standard inclusions in the characteristics table, so that readers can review, prima facie, the hypothesis being tested.”

Readers would be unlikely to be able to review the hypothesis being tested. The numbers of women with low-carbohydrate intakes were small in the total sample numbers, therefore any differences in average carbohydrate intake would be lost in the large numbers of subjects.

“Q3) Please can you add the calorie intake data to Supplemental Table 1, as this has been adjusted for, but not reported anywhere?”

Perhaps a valid point, although average calorie intakes could be added they would not add much to the interpretation of the study.

“Q4) Please can you provide the p values for Supplemental Table 1?””

These could be added but this would not add much to the study as these are not the main subject of this study and these comparisons have been made in many other research papers.

“Q5) Working on the assumption that case/control ratios outside 0.9-1.1 are likely to be statistically significant, Supplemental Table 1 suggests that adjustments should have been made for (I have used a (Y) and a (X) to indicate what was/wasn’t adjusted for): maternal race/ethnicity (Y); maternal birthplace (X); education (Y); household income (X); BMI (X); smoking (X); alcohol use (Y); folic acid antagonist medication use (X); and study centre (Y).

Please can you explain why the factors marked with an (X) weren’t adjusted for?”

The answer to this question is detailed in the methods section of the paper.

folate 2

“Q6) If I interpret Table 2 correctly, it means that of the 1,559 cases, 93 restricted carbohydrate (6%) and 1,466 (94%) didn’t and it means that 479 controls (5%) restricted carbohydrate and 9,064 (95%) didn’t.

The study “hypothesised that some women who restrict carbohydrates may have suboptimal folate status and subsequently may be at higher risk of having an NTD-affected pregnancy.”

Notwithstanding that you set out to prove a hypothesis (and not to disprove the null), please can you confirm that you failed to prove this hypothesis? Table 2 could conclude that “94% of NTD-affected pregnancies occurred in women not restricting carbohydrate.” Table 2 could also conclude that “of the women who had an NTD-affected pregnancy, fractionally more (1 in 100) restricted carbohydrate.” Table 2 cannot conclude the other way round – that those who restrict carbohydrates may be at higher risk of having an NTD-affected pregnancy.”

This question seems to result from a misunderstanding of odds ratios, which are tricky things to understand. Many explanations can be found online including this one.

Also this one: Explaining Odds Ratios.

Odds ratios are calculated like this:

Microsoft PowerPoint - Lesson 5 revised slides.ppt

These are some of the results of the study:

folate 3

An odds ratio of 1.20, in this case, indicates that for every 5 children with any neural tube defect in the nonrestricted group there were 6 children with any neural tube defect in the restricted group.

folate 4

After adjusting for the other factors listed in the footnotes the odds ratio is 1.30

Although this difference is small it could add up to many more cases of neural tube defects if the association held true across the population as a whole.

“The study “hypothesised that some women who restrict carbohydrates may have suboptimal folate status and subsequently may be at higher risk of having an NTD-affected pregnancy.”

Notwithstanding that you set out to prove a hypothesis (and not to disprove the null)…”

This appears to be a misunderstanding of the difference between a research hypothesis, that poses a specific hypothesis to be tested, and statistical hypothesis testing.

One of Zoe Harcome’s own papers appears to contain just such a research hypothesis that she seems to have set out to prove “Evidence from prospective cohort studies did not support the introduction of dietary fat guidelines in 1977 and 1983: a systematic review“.

folate 5

“Q6) Having added carbohydrate intake, folate/folic acid intake, calorie intake and p values to Supplemental Table 1, please can you adjust for all the differences between cases and controls and then re-calculate the odds ratios accordingly?”

Adjusting for all possible variables when there is no rationale for doing so is not good science (it is stated in the methods why some of these were not adjusted for).

folate 2

“Please can you then revise and reverse the directionality of the press release and correct the newspaper articles world-wide, which reported that: “women with low carbohydrate intake are 30 percent more likely to have babies with neural tube defects, when compared with women who do not restrict their carbohydrate intake.”

Many thanks

Kind regards – Zoë”


Almost every study has minor errors or additional things that could potentially be added. They type of study this is means that the authors have to rely on food frequency questionnaire and other questionnaires. While these are imperfect they are about the only way carry out a preliminary study into questions like this. If they find associations studies like this can then form the basis for further research.

However, from my reading of the study, the criticisms detailed by Zeo Harcombe do not appear to be valid and I do not think these are “retraction-level errors”.

While the results of this study are only epidemiological associations, together with the reduced estimated folate/folic acid intake reported, it makes a plausible link that should be investigated further.

Low-carb diets are not intrinsically high-or-low in folate. But as many people now rely upon fortified flour for their dietary folate, if this is removed from the diet and not replaced with other sources then folate intakes will decline.

Rather than trying to discredit the study, this could be used as a point to discuss supplementation or the inclusion of more low-carb foods that are naturally high in folate for women eating low-carb diets who may become pregnant.

Posted in Uncategorized | 1 Comment

What should we eat? A cacophany of advice

Many people want to know what and how they should eat to lose weight, improve their health, avoid illness, live forever, or for any number of other reasons. One popular source of information are books relating to food, diets, and nutrition. However, there are many different such books published. While many people do not read the books themselves, these books tend to generate newspaper and magazine articles, blog posts, online discussion, and television coverage exposing more people indirectly to their ideas. But what kind of such books are published?

Some time ago I began saving the images of the covers of these books that I came across from a popular book selling website together with the year they were first published in the UK. Here are a number of non-systematically collected covers published over the last four years.*

Some of these books are actually rather good, some are really quite bad, and I suspect many others are perhaps somewhat mediocre. What stands out though is the sheer amount of information published each year on what we should eat and why. What drives this appetite for such books fascinates me. What people are looking for and how do they navigate this vast amount of often conflicting information?


Book Covers 2017


Book Covers 2016


Book Covers 2015


Book Covers 2014

*These books were selected if they related to a combination of food, diet, nutrition, and health in some manner. For example, a book of “vegan recipes” would not be included, while a book of “vegan recipes to improve your health” would be.  Hopefully, there are not too many major errors or books missing although there are likely to be a few as they were not collected systematically. Self-published books were generally excluded due to the large number of them (I had to draw a line somewhere).
Posted in Articles, Food, Nutrition, Uncategorized | Tagged , , | Leave a comment

Hippocratic misquotations: Let thy quotations not be by Hippocrates


“Let food be thy medicine and medicine be thy food”

– Hippocrates.

This is perhaps one of the best-known quotes on diet and health attributed to Hippocrates, the physician in Ancient Greece considered to be the father of medicine. Only it seems that he never said it.

Dr Diana Cardenas, in her 2013 paper “Let not thy food be confused with thy medicine: The Hippocratic misquotation”, investigated the origin of this supposedly Hippocratic quotation. With the assistance of Professor Jacques Jouanna, an expert in ancient Greek and an author of books on Hippocrates, she confirmed that the quotation does not exist in any of the writings of Hippocrates.

hippocrates 1

In contrast, Dr Cardenas argues that, although diet was a major part of his practice of medicine, Hippocrates would not have made this statement as it fundamentally contradicts his principles of Hippocratic practice. Namely that food and medicine were considered to be different, even if both very important.

hippocrates 2.png

Diana Cardenas concludes this misquotation has been accepted due to the iconic reputation of Hippocrates.

hippocrates 3

This leads to the obvious question posted by Diana herself. If it did not originate from Hippocrates, then where did this quote originate from? This led me to some interesting, if frustrating, searching in the realm of Google Books Ngram Viewer, a searchable archive of millions of printed publications over the past three centuries showing trends in the use of words.

Using “food be your medicine“, which seems to be the earliest used version of the phrase,  there is nothing before the 1920s, when this first appears.

hippocrates 4

Similarly, the alternate archaic wording “food be thy medicine” first appears in the around 1959, but then its use really takes off during the early 1970’s.

hippocrates 5

The use of the phrase has increased steadily since its use caught on in the 1970s. It is now so ubiquitous that it is hard to imagine now that its origin is so recent. Although the lack of any results before the 1920s does not mean it was not used earlier, the lack of any written documents in the Google Books archive suggests that it was not widely used before that time.

Your food must be your medicine

The first time that I can find a similar phrase used, “Your food must make and keep you well; your food must be your medicine”, is in a periodical published in London in 1921 called Public Opinion (Volume 120), although there is no indication that this is linked to Hippocrates.


content (1)

Let your food be your medicine

The first mention of the phrase “Let your food be your medicine” appears to be in a book by Otto Carqué called The Key to Rational Dietetics published in 1926. Otto was an early Los Angeles based diet and health diet book guru. The very first page of this book begins with this quote but this does not appear to be linked to Hippocrates, who is not mentioned in the book.



An edition of the magazine “National Magazine of Health” repeats the phrase in an apparent review of Otto’s book.

A year later in 1927, in page 296 of the Wisconsin Beekeeping (Volumes 4-6), we find “Let food be your medicine” used by one Mrs E Nedvidek of the Riese Naturopathic Sanitarium in relation to honey, but apparently not linked to Hippocrates.


Mrs. E. Nedvidek, Manager of the Riese Naturopathic Sanitarium, La- Crosse, in telling how she uses honey in her sanitarium, said, “We use all natural foods, and our motto is ‘Let food be your medicine‘. We use honey as a sweet in every instance. Honey is the best blood builder you can get. We start the little babies on honey. It is not only a food, it is also a medicine.

Thy food shall be thy medicine – Hippocrates

In the same year of 1927, we find the quote “Thy food shall be thy medicine” first attributed to Hippocrates in “Principles of Diet: A Simplified Index to the Most Important Facts of Food Science and an Appendix of Normalizing Menus” by Dr Fred Reinhold edited and revised by Emma Regina Way.


hippocrates 11


This archaic phrasing does not seem to have caught on and “Let food be thy medicine” does not appear again until 1962 in a book called “Barefoot in Eden: The Macfadden Plan for Health, Charm and Long-lasting Youth” by Johnnie Lee Macfadden, with no mention of Hippocrates.

Let your food be your medicine – Hippocrates

To find the next use of this wording of the phrase “Let your food be your medicine” and an allusion to Hippocrates we have to move on to the 1940’s and another beekeeping journal in a 1941 print of Modern Beekeeping (Volume 25 – Page 259), that while not naming Hippocrates alludes to him as a man “Four hundred years before Christ…”.


content (1)


The next direct attribution to Hippocrates that I can find appears to be in “The Useful Soybean: A Plus Factor in Modern Living” by Mildred Lager published in 1945, although I cannot access the full text of this. According to Wikipedia Mildred was “an early American pioneer of natural foods and health food. On October 15, 1933 (in the depths of the Great Depression), she founded a health food store named The House of Better Living at 1207 West Sixth St., Los Angeles, California.”

Let your food be your medicine and medicine be your food

The first use I have found of the full phrase “Let your food be your medicine and medicine be your food is in a book called “Bloodless Surgery: With Technique and Treatments” published in 1945, although this does not attribute it to Hippocrates.



The first use I can find of the full quote attributed directly to Hippocrates himself  is in a 1959 book “Why?: Use Suncooked Juice Foods Daily!” by Jesse Mercer Gehman who was active in the emerging fields of alternative medicine, health reform, and naturopathy.


hippocrates 10


After 1959 this use of the quote gradually increased before taking off a few years later.

This may have been helped by the use of the full quote “Let thy food be thy medicine and medicine be thy food” in 1979 by Michael Lesser, MD in his testimony before Senator McGovern to the Select Committee on Nutrition and Human Needs of the United States Senate in its hearings on Diet related to killer diseases. Michael Lesser was one of the early proponents of Orthomolecular Medicine.


hippocrates 122


By the early 1980’s various form of the quote started to become commonly used in many situations and the origin of the quote and its attribution to Hippocrates does not seem to have been questioned until Dr Diana Cardenas’s paper in 2013.

The originator of the quote may be lost to history now, as it was likely used for the first time before it was written down. But it is fascinating to see how a quote like this spreads and how we often do not question things that sound like they are correct.

(I may update this post if I come across earlier examples, which is not unlikely).

Posted in Articles, Food, Nutrition, Uncategorized | Tagged , , , , , , , , , , | 11 Comments

Evolving salt: Did humans evolve on a high salt diet?


Recently I was reading some of a new book called “The Salt Fix: Why the Experts Got it All Wrong and How Eating More Might Save Your Life”.

This contains the interesting claim that humans evolved on a diet high in salt. This is rather surprising as salt and sodium are generally considered to have been a rather scarce in human diets before the development of salt production technologies in the past few thousand years.

Reading this I thought I would delve into the evidence for this in a little more depth, which ended up with me fact-checking Chapter 2 of this book, which contains this claim. I thought that this would be interesting as firstly, evolutionary justifications are often used now in nutrition, and secondly, I had already seen statements from this part of the book quoted on social media.

As I don’t have the time or interest to review a whole book a subsection of Chapter 2 reviewed here is quoted in full from the subheading “Prehuman Primates” to the end of the subsection “The Case Is Clear” in Chapter 2. My comments are found between each quoted excerpt.

“Prehuman Primates

Even today, most people believe that prehuman primates (such as orangutans, monkeys, baboons, and macaques) subsisted mainly on fruit and terrestrial vegetation.”

This is not a great start as orangutans, monkeys, baboons, and macaques are not pre-human primates.

Most people believe that orangutans, monkeys, baboons, and macaques subsist mainly on fruit and terrestrial vegetation because they do.

“Thus, one group of scientists has insisted that our prehuman bodies evolved on a low-salt diet. But that is clearly not the case.”

This suggests that other groups of scientists have insisted we evolved on a high-salt diet, which does not seem to be the case.

“Millions of years ago, climate changes that featured intense dry seasons were thought to have forced nonhuman primates to seek out wetlands. 19 Their diet would have consisted of aquatic vegetation, with a sodium content five hundred times that of terrestrial plants. 20”

The only source I could find for the sodium content of aquatic plants was a paper from 1973 called “Sodium Dynamics in a Northern Ecosystem“, which examined the role of water plants in the diet of moose living on Isle Royale, Lake Superior. While floating aquatic pondweed can be a good source of sodium it is only really relevant if you are a herbivore like a moose that can eat large quantities of pond weed, in which case this can be a decent source of sodium. I could find no details for the African environments in which our early ancestors evolved.

While it may be true that aquatic plants can have a higher sodium content than terrestrial plants, this is still not very high because terrestrial plants have very low levels of sodium.

“This may also be when nonhuman primates started eating meat, which they would have first encountered when fish and aquatic invertebrates were trapped in aquatic vegetation – providing primates with the original seafood salad. 21”

I think catching fish is usually a little more difficult than this…

“Once these foods were “inadvertently” eaten, nonhuman primates probably got a taste for them and started seeking them out deliberately. Their first fish were thought to have been easier prey, such as catfish that were injured, washed ashore, or trapped in shallow ponds. (Catfish were plentiful where ancestral primates and early humans roamed, making this a plausible notion.) This dietary switch— toward consuming more fat and omega-3s— certainly makes sense for its potential to foster the development of a larger (more human-sized) brain. Dozens of nonhuman primates have been reported to eat fish and other aquatic fauna that would have supplied their diet with ample amounts of salt. 22”

While eating fish has occasionally been reported in nonhuman primates, this would not supply “ample amounts of salt” as fish do not contain high levels of salt. African catfish contain 406 mg of sodium per kilogram of raw catfish, which is not a high level.

A better explanation of the wetlands theory (not referenced here) is a book chapter called “The Case for Exploitation of Wetlands Environments and Foods by Pre-Sapiens Hominins“. This explains a new theory that wetlands were important to our earliest hominin ancestors and that fish may have been an early item on the menu.

But this does not imply that they were eating a high sodium diet.

“They would have encountered such things as shark eggs, shrimp, crabs, mussels, razor clams, snails, octopus, oysters and other shelled invertebrates, tree frogs, invertebrates in the river mud, snapping turtle eggs, water beetles, limpets, tadpoles, sand-hoppers, seal-lice, and earthworms. 23 These abounded at seashores and in swamps, freshwater and marine water, and other tropical and temperate locations. Based on this list, it’s obvious that the diet of prehuman primates (and thus early humans) would not have been low in salt; in fact, it could have been extremely high in salt.”

This rather diverse list contains a mixture of very different creatures and is copied out of context from reference [23].

  • I suspect that our early hominin ancestors did not spend a lot of time eating water beetles, tadpoles, sand hoppers, or seal-lice.
  • The shark eggs, shrimp, crabs, razor clams, octopus, and oysters are sea creatures that would not be found in the inland freshwater wetlands described in the theory described in the previous paragraph.
  • The mussels, snails, tree frogs, and snapping turtle eggs are at least potential food sources for our early ancestors, but none of these is high in sodium.

No evidence is referenced that the creatures in this list available to early human ancestors are high in sodium (they don’t appear to be). The reference cited does not mention sodium. Therefore, it is not obvious that early human diets could have been high in sodium.

“The taste for fish and other aquatic creatures may have led these prehuman primates to begin deliberately trying to catch fish by hand and eventually using tools such as sticks, sand, and food to catch fish— which represented a huge leap forward in cognitive development. Think of that twist of fate: eating fish by happenstance may have enabled early primate brains to develop the intellect to actively catch fish through the use of tools. Exactly how they were able to obtain these salty creatures is more of a mystery, but it is thought that they used rocks to crack shells open and tapped on bamboo to find frogs living inside it. At least five other species, beyond orangutans, have been found to use tools to obtain fish and other salty aquatic prey. Thereafter, hominins – both modern and extinct humans – would have used primate fish-catching practices. 24″”

“…begin deliberately trying to catch fish by hand and eventually using tools such as sticks…”

“…and tapped on bamboo to find frogs living inside it.”

I’m not quite sure what to make of this. Have you ever tried catching fish with a stick or tapping on bamboo to find frogs? Additionally, neither fish nor frogs are high in sodium.

“Early Humans

Intriguingly, the emergence of tool-assisted fish catching in early Homo dates to around 2.4 million years ago. Primate fish-eating habits suggest that hominins would have also started eating aquatic plants first, then accidentally sampled the aquatic animals clinging to their nightly feeding, and, having acquired a taste for a newfound meat, eventually transitioned to catching fish and other aquatic prey. 25 Some researchers assert that an early human, Paranthropus boisei, and early Homo dug into wetlands to add vertebrates and invertebrates to what had previously been their predominantly plant-based diet. These aquatic animal foods yield plenty of salt and novel, high-quality nutrients, such as docosahexaenoic acid (DHA). Similar to how these essential fatty acids may have led to brain growth in prehuman primates, DHA allowed for the brain to increase in size in early humans. 26 The fact that DHA is important for the growth of the human brain creates the unavoidable suggestion that aquatic foods— and the hunger for salt that drew our ancestors to them— were an important player in how the human brain evolved into what it is today. 27 Terrestrial plants are low in DHA, which suggests that this transition to aquatic vegetation and prey was essential to increasing our brain size. 28 Imagine: our hunger for salt may have played a role in early humans’ great leap forward.”

Aquatic animal foods yield more sodium than land plants, but no reference is given here for how much this is. I have not been able to find any values for the sodium in aquatic animals in African wetlands where our early ancestors may have lived.

This is one theory of early human evolution that is presented as fact. The author attempts to link his theory of a high-salt diet without any evidence.

“Even early humans who lived far from the ocean’s brackish waters had this hunger for salt.”

I do not know how the author knows this.

“Data suggests that early humans roaming East Africa’s noncoastal regions between 1.4 and 2.4 million years ago may have consumed a diet extremely high in salt. An ancient ancestor to humans known as “Nutcracker Man” was said to have lived on large amounts of tiger nuts. 29 The fossils of this early human, discovered in 1959 in Tanzania, feature strong jaw muscles as well as wear and tear on molars, indicative of a diet high in tiger nuts.”

This is incorrect.

The “Nutcracker Man”, (Paranthropus boisei), was not a human ancestor.

A diet of tiger nuts is one theory to explain the highly specialized skull adapted for heavy chewing and the C4 carbon isotope ratios found in this species. However, reference [29] is actually a Daily Mail newspaper article… which I won’t link to.

While this branch of the hominin family is interesting it does not directly relate to the evolution of modern humans and our diet.

“Tiger nuts are extremely high in salt (up to 3,383 milligrams of sodium per 100 grams, the average amount of sodium we modern humans eat in an entire day). 30 Just a handful (3 ounces) of these nutlike tubers would have provided an entire day’s worth of sodium in today’s world.”

The single reference [30] cited gives a very high value (3,383 mg per 100 grams) for sodium in tiger nuts. Several other studies (not referenced in the book) show much lower levels of sodium in tiger nuts suggesting that reference [30] may not be representative of tiger nuts in general.

These suggest that tiger nuts are not necessarily a rich source of sodium and that the study cited in the book may be an outlier.

“Nutcracker Man did not live by nuts alone. He also survived on a diet largely composed of grasshoppers. A close relative of the grasshopper, the cricket contains a very good amount of sodium (about 152 milligrams of sodium per five crickets). 31”

This is incorrect.

The crickets tested, Acheta domesticus, weigh an average of around 0.4 grams and contain 152 mg of sodium in 100 grams of crickets. That means these crickets contain about 152 mg of sodium per 250 crickets, not per five crickets.

“Most likely, certain insects are so high in sodium because it allows them to move and fly faster and thus avoid being eaten by their brethren. 32 Scientists have observed that sodium deficiency can lead to cannibalism in insects (and probably other animals, too). 33 The theory goes that the animals instinctively know that salt is contained within blood, interstitial fluid, skin, muscle, and other parts of their bodies.”

Insects are not so high in sodium. As Reference [32] and [33] suggest, actually both are the same reference, that sodium deficiency is bad for insects and they will eat each other, but this is not really relevant to the discussion here as insects are rather different to humans.

“Not surprisingly, experts believe humans have been getting protein and micronutrients from wild insects for several millennia— and continue to do so to this day, particularly in parts of Africa, Asia, and Mexico. 34”

This reference [34] actually shows that insects are not high in sodium. Even for crickets, which had the highest sodium level of the insects tested, you would have to catch and eat more than 2 kg (4.4 pounds) of crickets each day (more than 5,000 crickets) to get close to the average sodium intake of Americans.

“”The Case Is Clear

From an evolutionary standpoint, evidence does not suggest that we evolved on a low-salt diet. Instead, much of our evolutionary theory seems to support the fact that we evolved on a high-salt diet. So where does this persistent misconception about our original diet come from?”

Is the case clear? No evidence has been presented so far to suggest that we evolved on a high-salt diet.

“The idea that our human ancestors consumed very little salt, generally less than 1,500 milligrams of sodium per day, is both old and current. 35 Some of the debate about evolutionary diet seems to stem from one influential paper on the topic, which was published in 1985 in the New England Journal of Medicine, one of the world’s most prestigious medical journals. The authors of this paper estimated that during the Paleolithic era (from about 2.6 million years ago until about 10,000 years ago), our intake of sodium was just 700 milligrams per day. 36 But this figure was based on the sodium content of select land animals (and only the sodium content of the meat) as well as land plants available to hunter-gatherers. This estimate does not include the sodium that would have been obtained from tiger nuts, insects, or aquatic vegetation or prey, nor does it include the other large stores of sodium found in animals besides the meat, such as that found in the skin, interstitial fluid, blood, and bone marrow (which we know hunter-gatherers did eat). We can’t forget that, aside from their meat, animals themselves (muscle, organs, viscera, skin, blood) are extremely good sources of salt.”

None of the evidence presented in this book questions the estimates made in 1985.

Muscle, organs, viscera, skin, blood) are not “extremely good” sources of salt. A diet based mainly on eating animals nose to tail can supply around 1,500 milligrams of sodium per day. But this is significantly lower than modern intakes of salt.

“For example, muscle contains approximately 1,150 milligrams of sodium per kilogram. Australian Aborigines would eat 2 to 3 kilograms of meat per sitting during a kill. 37 This is equal to 3,450 milligrams of sodium per day, the exact amount of sodium that current-day Americans consume (when they’re not straining to achieve the low-salt guidelines, that is!).”

This is incorrect. Meat does not contain that much sodium, but no reference is provided.

According to the USDA food database, 1 kilogram of “Beef, top sirloin, steak” contains 520 milligrams of sodium.

As another example, kangaroo meat contains between 40 and 60 milligrams of sodium per 100 grams. Eating 2-3 kilograms of kangaroo meat could contain 1000-1500 milligrams of sodium.

“Organs of animals are even higher in salt than meat: just 10 ounces of bison ribs (about one-quarter of a kilogram) provides 1,500 milligrams of sodium, the same amount in just 13.5 ounces of bison kidney or 2 pounds of bison liver. And remember, this doesn’t even include the salt that is found in the skin, interstitial fluid, blood, and bone marrow.”

This is incorrect. However, no reference is provided.

  • 10 ounces of bison ribs contain 300 milligrams of sodium, not 1,500. Beef ribs seem to contain a bit less.
  • 2 pounds of bison liver contains 882 milligrams of sodium, not 1,500 milligrams. This is equal to 276 milligrams in 10 ounces, similar to that found in meat.
  • I could not find any measurement of bison kidney. But 100 grams of beef kidney contains 182 milligrams of sodium, more than in muscle meat but still not a lot.

This indicates that organ meats are not going to provide you with a lot of salt.

I could not find any measurement of sodium in the skin, bone marrow, or in “interstitial fluid” in animals, or any idea how you would go about eating interstitial fluid. If the author of the book has these figures he is not revealing them.

“Early humans probably got salt in other ways as well. Some would have also eaten soil, as is still done by Kikuyu women of Africa, who are known to make dishes from sodium-rich soil. 38”

If you read reference  [38]  it states that Kikuyu women of Africa did eat sodium rich soils, but that their diets were still very low in sodium despite this. To quote the reference:

“A nutritional study carried out a generation ago in Africa (Orr and Gilks, 1931) reports that Kikuyu women, especially during pregnancy and lactation, prepared a special dish making use of sodium-rich soils; men did not eat this. It is of interest to note that in spite of their regular daily utilization of edible earths, Kikuyu women obtained only between 0.88-1.56 grams of sodium per day, whereas males derived only 0.48 grams. Obviously, without the emphasis on sodium-rich soils, the Kikuyu diet would have been virtually devoid of sodium.”

“Our ancestors also likely had salt licks and drank rainwater, providing clear evidence that previous estimates of sodium intake during our evolution are most likely drastic underestimations.”

Our ancestors may have had salt licks, although as shown above, eating sodium-rich soil does not supply a lot of sodium.

While true that rainwater can contain sodium this is misleading as the sodium content of rainwater is very low. If you have ever tasted rainwater it does not taste salty.

“But alas, the mantra has always been that the strict vegetarian diet of our early ancestors only provides around 230 milligrams of sodium per day, and that even a carnivorous diet only provides around 1,400 milligrams of sodium.”

These values appear to be reasonable estimates.

“These low estimates led most experts to believe that our current salt intake is two to twenty times what our ancestors would have consumed. And if we didn’t eat that much salt during our evolution, then our current intake can’t be good for us! (Or so the mantra goes.)”

This is how theorising based on evolutionary principles is often carried out.

“No one truly knows how much salt our Paleolithic ancestors ate or how much salt our human brain evolved on— but it’s probably much more than what most experts think. Some experts believe that 45 to 60 percent of our Paleolithic ancestors’ calories came from animal foods 39 that are naturally high in salt.”

As stated above, animals foods are not “naturally high in salt”, at least compared to modern foods with added salt.


Salt may not be as bad for people as has often been claimed. But the poorly constructed evolutionary rationale in this book used to justify a diet high in salt is badly referenced, contains little to no actual evidence, and includes a number of factual inaccuracies. The references that are used appear to have been chosen to fit into the narrative of the book and references that did not fit the narrative have been inaccurately reported. Factual statements that were not referenced were often inaccurate.

I did not find any evidence in this chapter that supported the suggestion that humans evolved on a diet high in sodium and the poor quality of this chapter does not give a great deal of confidence in the rest of the book. If you want to enjoy reading popular books related to nutrition I recommend that you do not attempt to check their factual accuracy.

Posted in Articles, evolution, Research, science, Uncategorized | Tagged , , , , , , , | 41 Comments

NOD mice: Gluten-free diets and diabetes

Type 1 diabetes mellitus is an autoimmune disease where both genetic and environmental factors, such as diet, combine to trigger the disease. Type 1 diabetes has been found to occur more frequently in individuals who also suffer from coeliac disease. This may be because both diseases share the same genetic risk factors and exposure to gluten could potentially be a dietary risk factor for both diseases.

This is where we turn to the non-obese diabetic or NOD mouse. First developed in 1980 in Japan, this mouse spontaneously develops Type 1 diabetes as its immune system attacks the cells of its own pancreas,  with 60-80% of females and 20-30% in male mice becoming diabetic. This mouse has been used as one model in research to try and understand Type 1 diabetes in human.

Back in 1999, evidence was suggesting that plant proteins in the diet of NOD mice was the environmental factor that was triggering their diabetes, particularly wheat and soy proteins. David Funda and his colleagues from Copenhagen and Prague decided to test this by feeding NOD mice either regular mouse food made from cereals and legumes, or a gluten-free equivalent of the the same diet with the wheat removed.

They found that by feeding the diet without wheat, the NOD mice develop far less diabetes and occurred later in life, compared to those fed the regular food.


Source: Funda, 1999.

This dramatic result suggested that gluten was a key factor in triggering diabetes in these mice. It should be noted that they fed the sames diets to the mothers of these mice so that the mice used in the study also had/didn’t have exposure to gluten before they were born. The same authors have recently shown that it is exposure during pregnancy that is most important in triggering diabetes in these mice (Antvorskov 2016).

The details of these diets and the ingredients used are shown in the table below. If you take the wheat out you have to replace it with sometime to keep the levels of proteins the same. In this case, they replaced it with meat protein while trying to keep other nutrients at similar levels.


Source: Funda, 1999.

However, it is good science to try to find evidence that the effect you saw was actually caused by the thing you think caused it. Was it actually gluten triggering the diabetes in these mice?

To test this, the same authors carried out another study using the same diets and the same type of mice as before. Only, this time, they included a third group of mice that were fed the wheat-free/gluten-free diet with pure gluten added back in (Funda, 2008).


Source: (Funda, 2008).

Perhaps surprisingly, as you can see from the chart above, the mice fed the gluten-free diet with added gluten were protected from diabetes just as much as those fed the gluten-free diet. This suggested that the story wasn’t so simple as first thought.

In fact, the diet with added gluten contained rather more gluten than the normal diet, as the scientists were expecting it to have a bad effect, but they wanted to make sure it was bad.


Source: (Funda, 2008).

This suggested it wasn’t gluten itself that was triggering diabetes in NOD mice.

However, taking the wheat out of the diet did prevent the diabetes, which leaves us with the question as to what was having the effect. Wheat is more than just gluten and taking the wheat out of the diet removes more than just gluten.

A recent study from Israel added an interesting twist to this question by testing different types of wheat (Gorelick, 2017). The modern bread wheat that we eat today has changed a lot in recent decades through selective breeding and hybridisation. This study compared the normal mouse diet that caused diabetes in NOD mice with versions of the same diet in which the type of the wheat was changed.

This included modern wheat and four old varieties of wheat called landraces.

  • Diet 1: A non-wheat diet replacing wheat with maize.
  • Diet 2: Standard modern bread wheat (Triticum aestivum).
  • Diet 3: An old bread wheat (Triticum aestivum), a landrace wheat from Israel.
  • Diet 4: Wild emmer wheat (Triticum turgidum ssp. dicoccoides).
  • Diet 5: Emmer wheat (Triticum turgidum spp. dicoccum), also known as farro.

As you can see in the charts below, only the modern wheat caused high levels of diabetes in NOD mice. Mice fed the diets made with old wheat varieties showed almost no diabetes during the time of the study, similar to the previous gluten-free diets.*


Source: (Gorelick, 2017).

This suggests that modern wheat has a key effect but it is unlikely to be due to gluten, both because isolated gluten from modern wheat did not cause diabetes, and old varieties of wheat did not, even though they also contain gluten.

This suggests something else in modern wheat is having the effect on these mice. Wheat flour contains a complex mixture of several types of storage proteins such as gliadins, glutenins, globulins, and triticins. The α-gliadins possess the ones with the most links to celiac. However, other wheat proteins may be involved in triggering Type 1 diabetes in these NOD mice. For example, a wheat globulin protein, Glb1, has been identified as a trigger of diabetes in another animal model, the BioBreeding rat (MacFarlane, 2003). The proteins from the older varieties of wheat have yet to be analysed to unravel why they are less likely to trigger diabetes in NOD mice.

As usual, it is worth bearing in mind that these NOD mice are a model used to study diabetes, but they are not small furry humans. Caution should be used when trying to transfer results from the mice directly to human diabetes. However, it is interesting to see such contrasting effects of different types of wheat. While the grains may look similar on the outside, the proteins they contain can be quite different.

As always, more research will be needed to understand exactly what is going on.


*Curiously, in this study, the maize diet also triggered diabetes in the NOD mice. Plant proteins other than wheat are known to be a diabetic tigger in these mice.


Funda DPKaas ABock TTlaskalová-Hogenová HBuschard K.  (1999) Gluten-free diet prevents diabetes in NOD mice. Diabetes Metab Res Rev. 15(5):323-7.
“BACKGROUND: Epidemiological as well as animal studies have shown that environmental factors such as nutrition contribute to the development of diabetes. In this study we investigated whether the early introduction of a gluten-free diet can influence the onset and/or incidence of diabetes, as well as insulitis and the number of gut mucosal lymphocytes, in non-obese diabetic (NOD) mice. METHODS: Gluten-free and standard Altromin diets (with the same milk protein and vitamin content) were given to breeding pairs of NOD mice as well as to the first generation of NOD female mice, which were then observed for 320 days. RESULTS: A substantially lower diabetes incidence (chi(2)=15.8, p=0.00007) was observed in NOD mice on the gluten-free diet (15%, n=27) compared to mice on the standard diet (64%, n=28). In addition, mice on the gluten-free diet developed diabetes significantly later (244+/-24 days SEM) compared to those on the standard diet (197+/-8 days, p=0.03). No differences in the number of CD3(+), TCR-gammadelta(+), IgA(+), and IgM(+) cells in the small intestine were observed. CONCLUSION: We showed that gluten-free diet both delayed and to a large extent prevented diabetes in NOD mice that have never been exposed to gluten.”
Funda DPKaas ATlaskalová-Hogenová HBuschard K. (2008) Gluten-free but also gluten-enriched (gluten+) diet prevent diabetes in NOD mice; the gluten enigma in type 1 diabetesDiabetes Metab Res Rev. 24(1):59-63.
BACKGROUND: Environmental factors such as nutrition or exposure to infections play a substantial role in the pathogenesis of type 1 diabetes (T1D). We have previously shown that gluten-free, non-purified diet largely prevented diabetes in non-obese diabetic (NOD) mice. In this study we tested hypothesis that early introduction of gluten-enriched (gluten+) diet may increase diabetes incidence in NOD mice.METHODS: Standard, gluten-free, gluten+ modified Altromin diets and hydrolysed-casein-based Pregestimil diet were fed to NOD females and diabetes incidence was followed for 310 days. Insulitis score and numbers of gut mucosal lymphocytes were determined in non-diabetic animals.RESULTS: A significantly lower diabetes incidence (p < 0.0001) was observed in NOD mice fed gluten-free diet (5.9%, n = 34) and Pregestimil diet (10%, n = 30) compared to mice on the standard Altromin diet (60.6%, n = 33). Surprisingly, gluten+ diet also prevented diabetes incidence, even at the level found with the gluten-free diet (p < 0.0001, 5.9%, n = 34). The minority of mice, which developed diabetes on all the three diabetes-protective (gluten+, gluten-free, Pregestimil) diets, did that slightly later compared to those on the standard diet. Lower insulitis score compared to control mice was found in non-diabetic NOD mice on the gluten-free, and to a lesser extent also gluten+ and Pregestimil diets. No substantial differences in the number of CD3(+), TCR-gammadelta(+), and IgA(+) cells in the small intestine were documented.CONCLUSIONS: Gluten+ diet prevents diabetes in NOD mice at the level found with the non-purified gluten-free diet. Possible mechanisms of the enigmatic, dual effect of dietary gluten on the development of T1D are discussed.
Gorelick J, Yarmolinsky L, Budovsky A, Khalfin B, Klein JD, Pinchasov Y, Bushuev MA, Rudchenko T, Ben-Shabat S. (2017 ) The Impact of Diet Wheat Source on the Onset of Type 1 Diabetes Mellitus-Lessons Learned from the Non-Obese Diabetic (NOD) Mouse ModelNutrients. 10;9(5).
Nutrition, especially wheat consumption, is a major factor involved in the onset of type 1 diabetes (T1D) and other autoimmune diseases such as celiac. While modern wheat cultivars possess similar gliadin proteins associated with the onset of celiac disease and T1D, alternative dietary wheat sources from Israeli landraces and native ancestral species may be lacking the epitopes linked with T1D, potentially reducing the incidence of T1D. The Non-Obese Diabetic (NOD) mouse model was used to monitor the effects of dietary wheat sources on the onset and development of T1D. The effects of modern wheat flour were compared with those from either T. aestivumT. turgidum spp. dicoccoides, or T. turgidum spp. dicoccum landraces or a non-wheat diet. Animals which received wheat from local landraces or ancestral species such as emmer displayed a lower incidence of T1D and related complications compared to animals fed a modern wheat variety. This study is the first report of the diabetogenic properties of various dietary wheat sources and suggests that alternative dietary wheat sources may lack T1D linked epitopes, thus reducing the incidence of T1D.
Posted in Articles, Gluten, Nutrition, Research, Uncategorized, Wheat | 2 Comments

Diet and Nutrition of the Lion

Do lions eat the organs of their prey, or can a lion survive only on muscle meat? I’ve seen this question asked online but clear answers are hard to find, at least not answers with sources backing them up. In an attempt to answer that question online I found that lion nutrition is an interesting topic, and this is the result of my intrepid Googling.

TL:DR Captive lions suffer serious nutritional deficiencies if fed only muscle meat. Lions preferentially eat the intestines and organs of prey animal, continuing to eat virtually everything else including some bones. Those organs and bones are essential to provide the vitamins and minerals required for good lion health.

The book The Serengeti Lion: A Study of Predator-Prey Relations by George B. Schaller gives a description of the feeding practices of the lions of the Serengeti. Very little of the animal is left uneaten by the lions.

lion a

The book goes on to say that the lions often gorge themselves first on the gut and internal organs first, suggesting a preference for the internal organs of the body, perhaps due the importance of these to their nutritional requirements in providing both fat and vitamins found in and around the gut.

lion b


Husbandry Guidelines for the African Lion by Annemarie Hillermann also provides some interesting details as to the nutritional requirements for lions. Perhaps unsurprisingly, feeding lions whole animal prey to eat is recommended as the simplest way to provide a balanced nutrition and diet for lions.

lion diet

These guidelines go on to describe how obesity is a potential problem in captive lions. Given their sedentary lifestyle in captivity their food must be restricted to meet their needs, if given too much they will overeat. Additionally, starve days are recommended to mimic their natural feeding frequency.

lion diet 2

ZUTRITION, another great source of information, has some comprehensive advice regarding the nutrition of lions.

“Cats are obligate carnivores. They derive most of their energy requirements from protein. The natural diet of cats is rich in proteins and therefore cats had no evolutionary need to synthesize as many amino acids as omnivores did. They have an absolute requirement for protein, and cannot synthesize the amino acids taurine, arginine, methionine and cystine. Meat diets will provide these amino acids, however diets that contain more carbohydrates may be deficient.

All-meat diets pose potential problems, however. A calcium:phosphorus ratio imbalance may lead to growth problems or metabolic bone disease. The Ca:P ratio in the body is 2:1. The Ca:P ratio to aim for in the diet is between 1:1 and 2:1. All-meat diets are high in phosphorous and have little-to-no calcium. They also may be lacking in vitamins A, E and D, which are found in adipose or organ tissues.

Vitamins B and K are provided by gut contents of whole prey and would be lacking in an all muscle-meat diet. Organ meats such as liver, kidney and heart tend to have the worst ratios of Ca: P and may run as high as 1:44.”

ZUTRITION has the following advice captive lion nutrition and the potential for nutritional deficiencies.

Lions seem to have a high requirement for preformed retinol (Vitamin A) and need a regular source of liver to supply this, particularly in young lions.

Vitamin A deficiency has been implemented in causing abnormalities in the cranial bones, especially atlanto-occipital malformation with ankylosis, hypertrophic osteopathy in the occipital bone and parietal bone, and osseous tentorium cerebelli, leading to progressive ataxia in young lions. Most affected lions are between 6–15 months of age. Many of the lion cubs exhibiting this syndrome were fed a diet of chicken parts with little other supplementation.”

Thiamin deficiency can also occur, as also noted in this study, as lions have a high requirement for B vitamins.

Vitamin B1 (thiamin) deficiency has caused anorexia, ataxia (more pronounced in the rear limbs), hypermetria (more pronounced in the front limbs), progressing to generalized weakness and recumbency lasting several minutes to hours, or longer. Young lions presenting with this problem were often fed a beef muscle meat diet supplemented with calcium.”

Feeding lions a diet of chicken meat can also result in severe copper deficiency, also noted in this study in the United Arab Emirates where lions and other big cats are often kept as pets.

Copper Deficiency can result from feeding a diet almost exclusively comprised of poultry.  This deficiency results in ataxia.

An imbalanced ratio of phosphorus to calcium, from feeding only meat and organ meats without any bones, is now understood to cause metabolic bone disease in lions. If an all muscle-meat or a muscle and organ meat diet is fed, the diet must be supplemented with some form of calcium.

Metabolic bone disease (MBD) may result from feeding a diet with an inappropriate calcium: phosphorus ratio. The calcium should be approximately 1.7 times the phosphorus in the diet. Organ and muscle meat are very high in phosphorus and calcium must be supplemented in appropriate amounts to prevent this problem. Thin bone cortices as well as hairline fractures and intermittent lameness are all signs of metabolic bone disease.”

It appears that big cats require a substantial amount of calcium from bone consumption to balance out the phosphorus in the meat and organ heavy diet.

“Private owners often feed poor diets such as ground turkey meat or chicken breast to young growing cubs. They fail to supplement the cubs with sufficient calcium (see Calcium Supplementation chart above), and the cubs they hand-rear suffer terrible consequences. It takes about 12 GRAMS of calcium carbonate to balance one kilogram of muscle meat fed.”

The effects of this imbalanced calcium to phosphorus ratio in lions was noted early on in 1960 in a rather sad news article in the New Scientist magazine as the cause of a long-standing problem in captive lions.

lion diet new scientist

Source: From New Scientist magazine, 12th May 1960.

A paper titled Nutritional and metabolic bone disease in a zoological population was published in 1976 detailing the health of two lion cubs suffering from multiple bone fractures and thin under-mineralised bones. This was attributed to them being fed a diet of muscle meat and offal, lacking in bones. As meat is high in phosphorus compared to calcium a high ratio between these minerals was suspected as causing these bone problems, although vitamin A poisoning could not be ruled out.

lions phosphorus 1lions phosphorus 2


Posted in lions, Nutrition, Research, Uncategorized | Tagged , , , , , , , | Leave a comment