Monday, January 19, 2009

The Tokelau Island Migrant Study: Diabetes

This post will be short and sweet. Diabetes is a disease of civilization. As Tokelauans adopted Western industrial foods, their diabetes prevalence increased. At any given time point, age-standardized diabetes prevalence was higher in migrants to New Zealand than those who remained on Tokelau:

This is not a difference in diagnosis. Tokelauans were examined for diabetes by the same group of physicians, using the same criteria. It's also not a difference in average age, sice the numbers are age-standardized. On Tokelau, diabetes prevalence doubled in a decade. Migrants to New Zealand in 1981 had roughly three times the prevalence of diabetes that Tokelauans did in 1971. I can only imagine the prevalence is even higher in 2008.

We don't know what the prevalence was in Tokelauans when their diet was completely traditional, but I would expect it to be low like other traditional Pacific island societies. I'm looking at a table right now of age-standardized diabetes prevalence on 11 different Pacific islands. There is quite a bit of variation, but the pattern is clear: the more modernized, the higher the diabetes rate. In several cases, the table has placed two values side-by-side: one value for rural inhabitants of an island, and another for urban inhabitants of the same island. In every case, the prevalence of diabetes is higher in the urban group. In some cases, the difference is as large as four-fold.

The lowest value goes to the New Caledonians of Touho, who are also considered the least modernized on the table (although even their diet is not completely traditional). Men have an age-standardized diabetes prevalence of 1.8%, women 1.4%. At the other extreme are the Micronesians of Nauru, affluent due to phosphate resources, who have a prevalence of 33.4% for men and 32.1% for women. They subsist mostly on imported food and are extremely obese.

The same patterns can be seen in Africa, the Arctic and probably everywhere that has adopted processed Western foods. White rice alone (compared with the combination of wheat flour and sugar) does not seem to have this effect.

The data in this post are from the book Migration and Health in a Small Society: the Case of Tokelau.

The Tokelau Island Migrant Study: Background and Overview

The Tokelau Island Migrant Study: Dental Health
The Tokelau Island Migrant Study: Cholesterol and Cardiovascular Health
The Tokelau Island Migrant Study: Weight Gain


theoddbod said...

all these island studies are great. they're the perfect epidemiological experiments! Thanks again.

Drs. Cynthia and David said...

Thanks Stephan and Robert for your responses on the previous post. I'm just a skeptic- we all are at this point I think.

But what about the contradictions? Unless you want to claim it's all genetic and adaptation, then why is it that high carb is bad for some populations but not for all? Processed Western food (ground flour, refined vegetable oils and sugar) are different from sweet potatoes and other starchy root vegetables and rice? What about white potatoes- are populations subsisting on a large proportion of white potatoes exempt from diabetes and other Western diseases? Are the Kitavans showing signs of Western diseases with the introduction of Western foods?

Could it have to do with the bioavailability of the sugars and starches (fine particles allow more complete extraction/absorption of nutrients than larger particles)? or is it due to the influence of omega-6 and/or excess fructose? Or is it that we can have our starch or our fat but not both together?

Sorry for the barrage of questions...


Scott Miller said...

In your opinion, Stephen, why doesn't rice, nor starchy tubers, seem to have a negative effect?

It all seems to come down to grains. On the surface, gluten and lectins are the two things that standout as the culprits.

Fitnatic said...

Hi Stephan,

I enjoy reading your blog so much! By reading your blog and a few others, I (a layperson) am gaining new understanding about nutrition and bio-chemistry that affects our bodies so much! I am trying the no wheat route (though not completely there yet) and will soon start with reducing sugar in my coffee. Giving up coffee with milk and sugar and chocolates will be HARD. But I hope to do it by drawing inspiration online!

Anyway, just wanted to share that in India, ancient history talks about "Rishi's" ( a seer or saint) who lived all their lives in jungles and meditated 24/7 by surviving only on a few tubers. They lived long and very healthy lives.

I also suspect that oil might be a comparatively new phenomenon in Indian diet. All the ancient texts that talk about performing rituals for God always specify using Ghee. Pure cow ghee prepared a certain way is required. I could be wrong, but that makes me suspect that omega- 6 laden vegetable oils are a recent addition. Last few thousand years new.

Stephan Guyenet said...


I'm not aware of any contradictions to the hypothesis that starch coming from tubers and certain grains, carefully prepared, are healthy on a cultural level (from the standpoint of modern, non-communicable diseases). If you have examples of this please pass them on.

There are obviously some people in modern society who do not tolerate high-carb. I believe this is because they have already sustained permanent damage, for example type I diabetics. But I'm not aware of any traditional society that suffers from the diseases of civilization on a diet of starchy tubers (without eating any modern foods).

I don't have data from any population that eats potatoes but not modern foods. It may be out there, but I haven't found it yet. That used to be true of certain S American groups, I don't know if it still is. Weston Price found them to be in excellent health, but he didn't look for diabetes. I suspect potatoes can be part of a healthy diet.

Lindeberg's studies were all conducted before modern foods arrived on Kitava. I don't have any data about their health now, assuming they're eating more modern foods.

A mashed starchy tuber has a very high reactive surface and a very high glucose availability, so I don't think that's a factor. In fact, when you look at starch-eating cultures, they seem to go to great lengths to make their starch as quickly and easily assimilated as possible. Grains are soaked, pounded, fermented and cooked into porridges and mashes. Bran is removed if possible.

Fructose could definitely have been a factor on Tokelau. In my opinion, it's a big piece of the puzzle. n-6 doesn't seem to have played a role because their total PUFA intake didn't change much and remained low.


I can't say for sure, but my guess is the gluten. We've been eating tubers for around 1.5 million years, so we're probably better adapted to their toxins. They also tend to be lower in toxins than seeds in general.


Good luck with your transition. Ghee was the main fat in many parts of India, and modern seed oils are mostly (but not entirely) a recent introduction. I think ghee and butter are the ultimate fats.

Monica said...

"I don't have data from any population that eats potatoes but not modern foods. It may be out there, but I haven't found it yet. That used to be true of certain S American groups, I don't know if it still is. Weston Price found them to be in excellent health, but he didn't look for diabetes. I suspect potatoes can be part of a healthy diet."

I'm pretty sure that until the potato famine, the Irish were almost completely dependent upon potatoes. It would be fascinating if there was some study of health at that time period. My guess is that they would have done poorly if they did not have any source of animal protein or fat, but I wonder if there is any evidence of diabetes at that time. My guess is that there is not (but if there is no evidence either way I suppose it doesn't tell us anything).

Robert Andrew Brown said...


Re Omega 6

The NZ population I would guess had rising Omega 6 intake.

Can you let me know where I could find the dietary data. Is it only in the book? I have not so far managed to find much re linoleic acid and Tokelau searching NCBI

Did processed foods increase as many include hidden fats?

How about fish intake in islanders, did that fall? Omega 3 and 6 is a balance. Severe imbalance at any intake level seems to have negative effects.

NZ from memory has quite low fish intake.

Cynthia and David

Japanese in Japan v Hawaii

Omega 3 genetics heart disease. Long chain Omega 3 gives Japanese men by higher HDL.

Stephan Guyenet said...


I believe the NZ population probably ate more n-6 than the Tokelauans, but their total PUFA was still at 4%. They ate fish fairly regularly even in NZ so I have to assume their ratio wasn't too bad.

Tokelauans probably saw their ratio skew toward n-6 as well, but I'd be surprised if it went above 2:1 n-6:n-3, given the fact that fish remained a staple and their total PUFA didn't exceed 3%.

Monica said...

Stephan, I decided to look into this potato thing a little more seriously.

I went back to one of my favorite books to confirm my assumption about potatoes in the Irish diet prior to 1850. It appears I am correct. Not only did the average family of six consume 250 lbs. potatoes weekly, the population doubled from 1800 to 1845. Because the climate was so remarkable for growing potatoes, people were able to be fairly leisurely and abandon other food production -- and married earlier, had larger families, and were able to nurse more newborns.

I don't understand the economics of it all, but because of trade obligations (likely in place through English laws?), the Irish were producing some meat, vegetables, and dairy but most of it was exported. When corn was imported from America to aid in alleviating the famine, the Irish rejected it, even though they were starving. Apparently their digestive systems were very conditioned to potatoes.

In light of this, it would be fascinating to discover more about Irish health at that time... particularly this insulin issue and any evidence of diabetes.

The book I am referencing is "Magical Mushrooms, Mischievous Molds" by George Hudler. (It's discussing the potato famine caued by the infamous Phytophthora infestans.) He cites his sources for this historical information in the back of the book. Maybe you can get a copy of it and further investigate the potato dietary issue. If you want to exchange further ideas about this, feel free to email me at monicabeth10ATgmailDOTcom

Stephan Guyenet said...


By the way, the data are all available on PubMed if you search for the Tokelau island migrant study.

Drs. Cynthia and David said...

Thanks Stephan. Monica's mention of high potato diets of the Irish was what I had in mind. Haven't bothered to search it as yet.

One contradiction I refer to is that all of the studies I know using healthy people (particularly young athletes) put onto high carb diets show deleterious responses in terms of lipid profiles - high TGs, low HDL, high LDL/apoB levels, poor total cholesterol/HDL ratios, etc. I don't think we consider this "healthy," though that depends on how much you believe these markers mean anything. My impression is that high carb is somewhat safer in the presence of high physical activity (elevated lipase activity sucks the TGs out of the VLDL particles and lowers blood TG) or some other factor that keeps the total caloric intake quite low. Traditional cultures would not have had the variety and abundance that we do and perhaps this may have restrained the enthusiasm for overeating that we have. Excess carbs result in excess fat storage, particularly problematic in the form of visceral fat/ fatty liver, as you've discussed previously. This does indeed "damage" the metabolism somehow, and hopefully is reversible.

I still don't understand how traditional cultures eating high carb diets whether corn or rice or tubers avoid the pitfalls that we've seen in the West. The Sumo wrestlers have no problem fattening up on their high rice diet after all. But clearly all Japanese don't suffer so.

However, India is known for its high carb intake (lentils, beans, rice, wheat, little or no meat) and prevalence of CVD and diabetes as well. You might be interested in this paper
discussing ethnic Indians in America vs India and Pakistan. The authors of course recommend more exercise and less fat in the diet of their Indian subjects.

Is the traditional diet in India healthy? This may be a contradiction to the other traditional cultures you mention - again using lots of grains and legumes in both unprocessed and processed forms. It's complicated too because there are so many regional differences, but the overall macronutrient balance may be similar at least in some regions. It should be worth a look at some point.

Robert Andrew Brown said...

Thanks for the trial David and Cynthia.

I noted the following comment in the summary.

"In addition, Indians and Pakistanis usually consume snacks fried in vegetable oil."

This is also interesting

Human Requirement for N-3 Polyunsaturated Fatty Acids

See graph straight line relationship diabetes to Omega 3 :6 ratio page 963

Clearly such conditions are multi factorial and higher vegetable oil consumption will also mean likely lower Vit D K and maybe minerals. Urban living likely means less sun exposure so lower internal Vit d production too.

Robert Brown

Author Omega Six The Devils Fat.


gunther gatherer said...

Maybe we can replace the kind-of-carb discussion with a kind-of-cooking discussion when asking why some hunter gatherers are eating high carbs and still not getting diseases of civilization.

The white man's food is baked, broiled, barbecued, glazed, pasteurized, processed, seared and fried. This creates huge amounts of exogenous AGEs that western cultures consume in huge quantities.

I really think glycotoxins and advanced glycation endproducts are part of the picture, as HGs are eating mostly raw foods or foods cooked in water.

Scott W said...

Doctors Cynthia and David:

In the athlete studies you reference, you must be certain that they address carb source in order to use them in this discussion since that is the point that Stephan is making.

Usually, macronutrient studies lump all carbs, including improperly processed grains and gluten grains, into the "carb" category.


Before I found Stephan's blog, I was very committed to low carb. I felt that any information on a healthy primitive culture that ate high carb must be flawed. I remember thinking this as I read about the ultra-running Tarahumara who reportedly subsist largely on corn porridge.

Then I read a book about the bushmen, who were great hunters but could only intermittently bring home meat. While the men were spending days tracking prey, the women were consistently bringing home large tubers for the day-in and day-out subsistence of the tribe. Kills were shared so widely that everyone got just a taste of meat.

The "carb/insulin is always bad" hypothesis (Taubes, et al) must be modified until it can accomodate populations that healthily tolerate carbs.

What is really needed are double-blind studies where all macros are held steady and carb source, then oil source, is changed. I suspect there will be not only a difference between races in how this is tolerated, but for all populations based on Price's observations around the world.

Scott W

Fitnatic said...

About diets in India, each region has its own traditional diet. The northeast and east region consumes a lot of fish or meat with rice. The northwest region consumes a lot of wheat with meat and yogurt. The south consumes a lot of fermented rice and lentil with little or no wheat.

I am originally from western part of India near Mumbai. People of my state consumed different kinds of millet with some wheat and lot of green leafy vegetables and small amounts of potato. The people in the coastal areas of my state consume a lot of rice with coconut. Interestingly the neighboring state consumes a lot of fried food, ghee and wheat.

Unfortunately, everyone in India has started consuming a lot of wheat and other stuff fried in vegetable oil- cherry picking the BAD aspect of each others diets.

Activity levels have literally plummEted ( but still much higher than in USA one does not get ahywhere without a car) and stress levels have skyrocketted - all for earning more money to follow the western lifestyle.

On top of it the younger generation is wholeheartedly giving in to the insidious advertising of western diet (McDonald's anyone?).

Result- alarming trend of CVD at a very young age - late 20s or early 30s. Incidence of diabetes in atleast one family member. And this is though just eating diet not suitable to your particular state and activity level. One can only wonder what will happen when western diet is fully adapted.

On top of that, doctors say that its in the genes (?1?) - so nothing one can control though diet and lifestyle change apparently. So people who are really concerned about their health just try and eat more vegetables and wheat chapati, exercise and do yoga. Most do not even try this and give in to their "destiny".
Low carb will never find much acceptance in India where eating meat is prohibited to a large extent and maybe event he hot weather does not support eating a lot of meat.

Okay, getting off the soapbox now.

Fitnatic said...

Sheesh , excuse all the typos! :)

Monica said...

Scott, you echo my thinking on this as well. After reading Taubes and Eades I was firmly committed to low carb. In fact, I'm still somewhat committed as far as my personal weight loss efforts -- I think the combo of low carb and intermittent fasting has worked really well for me. But, based on the information I see here, I've started to question that assumption.

I think reading Taubes and eating low carb has great benefits -- but it's important to remember, I think, that at the end of his book Taubes calls for controlled trials to evaluate the carb hypothesis. It's not a sure thing yet -- even Taubes admits that. Ideally, trials would compare high carb diets of different types over a long time period.

Drs. Cynthia and David said...

Well, it's good we all have open minds here, or are trying to have them anyway. I don't exclude all carbs, even now that I'm trying harder to lose weight.

The studies I've seen on athletes don't give details on diet, at least not in the abstracts. I haven't gone to the library or paid for full text articles to be sure, but mostly the investigators were concerned about the amount of fat and perhaps saturated fat in the diet, and not much else. It's worth a second look at some point to see if any of the papers discuss how much starch and from what source, sucrose, fruit, and fructose (and lactose for that matter) the athletes were getting. There must be some studies on low vs high glycemic diets that would provide more info as well that I haven't seen.

It's also not a given that athletes are necessarily "healthy" either, though they are generally healthier than the sedentary population. I've heard of lean marathoner types coming down with type II diabetes- imagine their shock!

As P says, the dietary diversity of India is amazing, so it would be important to look at regions separately. There is probably a vast literature there too.

I'm still wondering what the optimal balance is too, and not sure that we have accurate information yet. We need to ask the right questions too.

Robert Andrew Brown said...

Hunter gatherer

AGEs and oxidised food products generally are a contributory factor.

I agree.

Half Navajo said...

I think fasting and not over eating are to big factors in staying lean, also not eating wheat, vegetable oils, and to many processed foods. Traditional cultures didn't seem to pig out as much as western civilizations do, and were in a fasting period more often, and stayed in better shape whether it was a hi carb or high fat diet. I still choose eating high fat low carb though, more nutrients for me, but i am not afraid of carbs.

Robert M. said...


After looking at that study on advanced glycation endproducts I think this is a lot of smoke and no fire. Look at figure 4, the relationship between ingestion of AGEs and the oxidatize stress markers. There's nothing there.
R^2 of 0.37 is miserable. For a study with 172 people, the data are amazingly noisy. The authors appear to have not done any multivariate analysis to see if these oxidative stress markers are correlated more tightly to any other input parameters.

The section on discussion of results is largely speculation.

This is essentially the same result as cholesterol consumption studies.

Stephan Guyenet said...


Wow, that makes potatoes look pretty good! They are certainly a very unusual crop. They're extremely productive in temperate climates and they're actually pretty nutritious. They contain 10-20 percent protein by dry weight and it has an amino acid profile similar to milk ("complete" protein). Add a little bit of fatty meat and I think you could get by just fine on a potato-heavy diet. It would indeed be interesting to get health data from then. Although I'm sure there's nothing to be found on diabetes.


Yes the problem with these carb-feeding studies is they act like all carb is interchangeable (as Scott mentioned). The carbohydrate typically comes largely from grains, especially wheat.

Then there's the issue of blood lipid subfractions. Elevated trigs and low HDL predict heart attacks in Americans but not in Kitavans. Carbohydrate seems to cause those changes in both populations. Is it possible high trigs and low HDL are simply markers of wheat/sugar consumption in the West? I'm not sure it's quite that simple, but that may be part of the answer.

The diets in parts of India where they have high rates of CHD often have an exceptionally poor ratio of n-6 to n-3. Other populations such as the African Bantu seem to do pretty well on grain and legume diets that are mostly vegetarian but don't include linoleic acid-rich vegetable oil. But they treat their grains and legumes with the utmost care, soaking and fermenting them where appropriate.

I believe the reason non-Western cultures escape the diseases Westerners get is because the carbohydrate itself is not the problem. The pattern is pretty clear globally: replace traditional diets with wheat and sugar, get fat and diseased. It seems to be essentially independent of macronutrient changes from what I can tell.

I'm still on the fence about whether carb restriction is the key to low-carb diets, or whether it's simply restricting wheat, sugar and processed foods. I'm open to the possibility that it's partly the carbs. But there are hints that paleo diets are great for weight loss, independent of carb intake.


Roasting meat and tubers on the fire must be the oldest cooking method. Hunter-gatherers roasted, smoked and baked food regularly in many groups.


Thanks for the description. I see your point about Indians not embracing low-carb. I think that could probably be said about a lot of Asia. Fortunately, I don't think it's necessary to eat low-carb to be healthy. I think the traditional South Indian diet of fermented beans and rice could be quite healthy if they replaced vegetable oil with ghee and ate meat or fish regularly. I love Indian food, I have a batch of idlis fermenting right now.

Robert M. said...

In looking at rat chow used for scientific studies:

Fat content is typically trans-saturated lard.

Protein content is typically cow milk casein, a well known allergen. Sometimes soy isolate is used.

Carbohydrate content is typically sucrose, corn flour/syrup, or wheat flour.

The chow is then typically enriched with vitamins and minerals.

gunther gatherer said...

Robert M and Stephen,

If you don't like that study, search Dr. Helen Vlassara on PubMed for more information on AGEs, and you will find tons of her rat studies. She has done two decades of research on exogenous glycotoxins and their effect on the body and diseases of aging. This first link has the rats on low AGE diet weighing 13% less than high AGE rats.$=relatedarticles&logdbfrom=pubmed

Monica said...

"Add a little bit of fatty meat and I think you could get by just fine on a potato-heavy diet."

It definitely seems to work from some people. Maybe there are some data on the McDougall dieters not eating grains?

Of course, they don't eat meat...

It seems to have solved this guy's inflammation problems, though:

Stephan Guyenet said...


Thanks for the information. I'll keep it in mind.