Category Archives: HDL/LDL/cholesterol - Page 2

Serum Cholesterol Among African Hunter-Gatherers

We’re in the midst of a little diversion looking at the cholesterol levels of hunter-gatherers around the world. This investigation was precipitated by the surprising claim by some Paleo authorities that hunter-gatherers had much lower cholesterol levels than any modern human population (Did Hunter-Gatherers Have Low Serum Cholesterol?, Jun 28, 2011).

We started by looking at Eskimos and Inuit (Serum Cholesterol Among the Eskimos and Inuit, July 1, 2011). We found that healthy Eskimo and Inuit groups seem to consistently have mean serum cholesterol between 200 and 230 mg/dl, with lower mean serum cholesterol levels found only in short-lived populations suffering from tuberculosis and parasite infections.

Today we’ll look at the populations with the lowest reported serum cholesterol levels: African hunter-gatherers. Africans made up 3 of the 5 groups claimed to have low serum cholesterol in the original Eaton et al paper [1] and 4 of the 5 in the later Cordain et al paper [2].

Cholesterol levels in modern Africans

Before looking at African hunter-gatherers, I think it’s worthwhile to look at modern Africans and African-Americans.

African nations have the lowest serum cholesterol levels and highest mortality levels in the world:

  • The 39 sub-Saharan African countries in O Primitivo’s cholesterol database have an average serum total cholesterol (TC) of 165.7 mg/dl (4.285 mmol/l). For comparison, the other 122 nations have an average serum TC of 201.1 mg/dl (5.20 mmol/l).
  • When the 161 countries in the database are ranked by mortality, sub-Saharan African nations occupy all but three of places 123 through 161. The non-African nations with highest mortality are Laos (#130), Cambodia (#131), and Haiti (#137). Haiti’s population is 95% of African descent, by far the largest African population share in the Americas. The only sub-Saharan African nations ranking above #123 are the island nations of Cape Verde, Comoros, and Sao Tome and Principe.

Here are the nations in O Primitivo’s database. Sub-Saharan African countries are in blue, others in red:

Sub-Saharan African TCs cluster around 165 mg/dl, while in the rest of the world varies much more widely but generally ranges between 170 and 240 mg/dl. The only countries outside sub-Saharan Africa with a mean TC below 170 mg/dl are Bangladesh and Tunisia.

Now, African-Americans. Here are mean serum cholesterol levels from the various NHANES studies:

Mean serum cholesterol levels (mg/dl) in Americans age 20-74 from NHANES

Group 1960-1962 1971-1974 1976-1980 1988-1994 1999-2000
All 222 216 215 205 204
Whites 222 216.5 214.5 205 205
Blacks 215.5 216.5 213 202.5 195.5
Hispanics 209 205 202.5

This table makes visibly clear the triumph of modern medicine: mean serum cholesterol levels have been heroically brought down from 222 mg/dl in 1960 to 204 mg/dl today.

This data is useful, because African-Americans eat a diet similar to that of other Americans. They are also, by global standards, fairly healthy. All-cause mortality in 2005-7 in Contra Costa, California (a quick search failed to bring up numbers for the US as a whole) was 683.9 per 100,000 for whites and 1002.7 for blacks (Source).

If we look above at O Primitivo’s data, a mortality of 1002.7 is lower than that of any sub-Saharan African nation and would suggest a TC around 195 mg/dl – right where African-Americans are today. White mortality of 683.9 would suggest a mean TC between 200 and 220 – right where white Americans are.

To me, this data suggests:

  • If there is a genetic difference in Africans that lowers cholesterol, it’s not large – probably less than 10 mg/dl.
  • Africans fit spot-on the cholesterol vs mortality curve established by non-African nations. So if there are genetic differences in Africans, they haven’t altered the mortality risk of low cholesterol levels.
  • In any ethnicity or race, low serum cholesterol indicates impaired health – probably due to infections – and higher mortality.

In interpreting the serum cholesterol of African hunter-gatherers, we should keep in mind this pattern. Low TC is probably only minimally a matter of genetics or diet. It’s primarily about infectious burden.

Now let’s look at African hunter-gatherer populations. Unfortunately, there is a paucity of data in Pubmed.


A search in Pubmed for “Hadza cholesterol” produces one paper [4]. Unfortunately it dates from 1972 and electronic availability for that journal begins in 2001, so I was unable to read it.

Congo Pygmies

There are no papers in Pubmed searches for “pygmy cholesterol” or “pygmies cholesterol,” nor does “Congo cholesterol” return anything resembling a hunter-gatherer cholesterol study.

According to Wikipedia (“Pygmy peoples”), “The best known pygmies are the Aka, Efé and Mbuti of central Africa.” However, “Aka cholesterol,” “Efé cholesterol,” and “Mbuti cholesterol” return no results either.

!Kung San, Khoisan, Kalahari San, Bushmen

These people may be referred to by a variety of names, as Wikipedia (“Bushmen”) notes:

The indigenous people of southern Africa … are variously referred to as Bushmen, San, Sho, Barwa, Kung, or Khwe. These people were traditionally hunter-gatherers, part of the Khoisan group …

Searching on all these terms, I was able to find only a few reports on cholesterol levels, not all of whom were eating the hunter-gatherer diet. Here’s a summary.

Cholesterol Levels of San Bushmen

Paper [ref] Mean TC Notes
Miller et al 1968 [23] 77 mg/dl Stale samples, poorly preserved samples, “difficult” technique not normally used for cholesterol measurement
Truswell & Hansen 1968 [24] 110 mg/dl Stale unfrozen samples, obsolete technique
Bersohn & Tobias 1966 123 mg/dl Unknown technique; reported by Truswell & Hansen [24]; farm bushmen were 145 mg/dl.
Tichelaar et al 1992 [26] 143 mg/dl Semi-settled
van der Westhuyzen J et al. [21] 158 mg/dl Semi-settled

Early reports showed the lowest TCs. The one from Klara Miller and colleagues in the Archives of Internal Medicine of 1968 [23], is not available to me as full text, but this is the abstract:

Samples of the plasma from nine male adult bushmen were analyzed for their lipid content. The mean results obtained (total cholesterol = 77 mg/100 ml, phospholipids = 107 mg/100 ml, and triglycerides= 49 mg/100 ml) were extremely low. The probable reasons for this finding are the very low dietary fat content and habitually high physical activity of Kalahari bushmen. [23]

These numbers are extraordinarily low for both TC and serum phospholipids.

My other early source was a 1968 letter to Lancet by Truswell and Hansen [24] commenting on the Miller et al report:

SIR,-We can add some more data to the values for serum lipids in 9 Kalahari bushmen which you reviewed in your annotation (Aug. 17, p. 395). Our findings confirm that these hunter-gatherers have among the lowest serum-cholesterol levels in the world.

In October, 1967, and again in April-May this year we joined the Harvard social anthropologists De Vore and Lee to investigate the medical and nutritional state of the !Kung bushmen they have been studying in the northern Kalahari, Botswana. We collected venous blood from some of these bushmen. Sera were allowed to separate by clot-retraction and decanted into tubes containing a little dried methiolate solution. The sealed serum-tubes were kept in a portable gas refrigerator while we were in the desert and were carried with us in a cool box, with ice-packs, when we travelled back to Cape Town by jeep and plane. Cholesterol was measured by the standard Abell method.

The results of 67 determinations in 54 bushmen are shown in the accompanying table. The mean serum-cholesterol levels were between 100 and 121 mg. per 100 ml. The extreme range in individual bushmen was from 57 to 178 mg. per 100 ml. A Bantu control serum collected in the desert on our first visit contained 166 mg. cholesterol per 100 ml. Our own serum-cholesterol levels, taken on the second visit, were 184 and 219 mg. per 100 ml….

Our values agree very well with the mean serum-cholesterol level of 121 mg. per 100 ml. reported by Bersohn and Tobias [Tobias, P. V. in The Biology of Human Adaptability (edited by P. T. Baker and J. S. Weiner); p. 190. Oxford, 1966.] in nomadic bushmen in the central Kalahari. These workers found that farm bushmen had rather high serum-cholesterol levels, averaging 145 mg. per 100 ml.

The lower concentrations found by Miller et al., averaging 77 mg. per 100 ml., might have resulted from the more difficult quantitative method they used, involving recovery from thin-layer chromatographic plates or, as they suggest, from harsher conditions in the southern Kalahari….

Some bushmen in the northern Kalahari obtain milk from Herero tribesmen who manage to keep cattle and have now settled between the bushmen in the same area. The 5 Dobe adults who said they got moderate amounts of milk had a mean cholesterol of 130 mg. per 100 ml., compared with 108 mg. per 100 ml. in 5 who drank little milk, and 114 mg. per 100 ml. in the 10 who were eating pure bush food. Most of the bushmen’s dietary fat probably comes from nuts, such as mongongo or mangetti (Ricinodendron rautanenii), which Lee estimates to be the largest source of calories in the Dobe region. They contain 38-58% fat.’

It looks like we have 3 reported measurements in San hunter-gatherers: 77 mg/dl by Miller et al, about 110 mg/dl from Truswell and Hansen, and 121 mg/dl from Tobias. All are remarkably low, far below the levels found in any contemporary population.

In all cases the samples were stored for weeks or months before being measured far from the place they were drawn – a thousand miles away in Cape Town in the case of Truswell & Hansen. Because the region lacked electricity, it was impossible to keep the samples frozen and difficult to keep them cool. In Miller’s case, the method is described as “difficult.”

We concluded from the study of Eskimos by Corcoran and Rabinowitch 1937 (Serum Cholesterol Among the Eskimos and Inuit, July 1) that stale samples preserved for a long journey and then measured by the method of Abell can produce false, low cholesterol readings. That may have happened also in the case of the three San investigators.

By the 1990s, sample and measurement techniques were greatly improved. Here are the methods used in a 1992 paper by Tichelaar et al [26]:

Blood samples were taken in heparinized containers, and separated plasma samples were deep-frozen until analysis could be carried out, which was always within two weeks. Total cholesterol (TC) and triacylglycerol (TAG) concentrations were determined enzymatically (Boehringer Mannheim GmbH Diagnostica, Mannheim, Germany; CHOD-PAP and GPO-PAP kits, respectively). [26]

Heparinized sample tubes, deep-frozen samples, measurement within two weeks, and enzymatic measurement in modern kits – none of these were features of the earlier lipid measurements on !Kung San hunter-gatherers.

Tichelaar et al studied young adult Bushmen, Hereros and Kavangos. The Bushmen were “in transition”:

The Bushmen studied were no longer purely nomadic; most of them live for variable times in association with Herero villages, but tend to migrate from one place to another depending upon work and food availability. They are accordingly very much “in transition.” [26]

Mean serum cholesterol was 3.71 mmol/l (143 mg/dl) in Bushmen, 4.21 mmol/l (163 mg/dl) in Hereros, and 3.61 mmol/l (140 mg/dl) in Kavangos. These numbers are low, but not as low as the numbers quoted by Eaton et al [1]. They are also not far removed from those of South African Xhosas and Vendas, who measured 4.08 and 4.23 mmol/l respectively (158 and 164 mg/dl). [26]

The Tichelaar sample of Bushmen had an average age of 29. Their plasma fatty acids were relatively good: omega-6 fatty acid levels were far below those of South African whites (9.3% vs 24.4%), and they had the highest EPA:AA ratio among the African groups. However, they drank 2 to 4 liters per day of home-brewed beer, and “vegetable and fruit consumption is little to none at all.” [26] No specific health statistics are given, but Tichelaar indicate that the infectious disease burden was high:

The change from their traditional lifestyle has resulted in a high incidence of malnutrition (9), especially reduced fat and protein stores (70% and 75%, respectively). The severe malnutrition is primarily responsible for the predominance of infectious disease in Bushmen, especially tuberculosis (9). One fact that was extremely clear was the high incidence of alcoholism.

A 1987 study by van der Westhuyzen et al from the Journal of Tropical Medicine and Hygiene [21] corroborates the picture provided by Tichelaar et al. They studied a settled population of !Kung San in Namibia. They confirm that alcohol abuse and malnutrition (including thiamin deficiency) were widespread. Mean serum cholesterol was 4.1 mmol/l (158 mg/dl) [21].

I got a little more insight into the health of settled !Kung San from another 1984 study [22]. It provides details about the timing and health effects of the transition from hunter-gatherer to settled lifestyle:

In 1969 we carried out a study of iron, folate, and vitamin B12 nutrition in a group of!Kung living a hunter-gatherer existence in the Kalahari desert in northwestern Botswana (1). We now report on a study of a comparable group of !Kung who have followed a settled way of life in Chum!kwe, Namibia, some 40 km west of the area ofthe original study (Fig 1).

This group, who are genetically and linguistically identical (2) to the subjects of the previous study (1) have resided at Chum!kwe for at least 15 yr. At Chum!kwe they underwent slow change during the 1960’s and now constitute a settled population which has ceased to hunt game or gather wild vegetation. The men undertake casual employment and food is purchased from local stores, or provided by Government agencies. The diet consists predominantly of maize, supplemented with small and irregular quantities of meat and vegetables, a diet which resembles that of the Black peoples inhabiting the rural areas ofSouthern Africa.

The subjects comprised a group of !Kung San (Bushmen) who, until relatively recently, were hunter-gatherers inhabiting the northwestern Kalahari (Fig 1). All the subjects now reside at Chum!kwe in Namibia 40 km to the west, where they have resided for at least 15 yr….

The diet of the San has changed considerably in the period since the previous study in 1969 (1). The staple food, commercial maize meal, is eaten as a porridge or fermented with cane sugar to make beer. Frequently, this porridge or beer is the only food consumed in a day. This basic diet is supplemented irregularly with canned fruit, vegetables, and meat. One family cultivated a few melons and squashes, but pastoralism is practised on such a small scale as to be negligible. There appears to be a high consumption of alcoholic beverages by the community, and beer, either homebrewed or commercial, is consumed by all members of the society, even very young children. The former is brewed in plastic containers.

Marshall and Ritchie (3) have reported on the economy of the Chum!kwe !Kung as they found it in 1981 and they have compared it with that existing in 1958 when the people still followed a hunting and gathering way of life. This latter way of life is described by Marshall (2). In 1981 the bulk of their dietary needs were met by purchases from the local store: made up of coffee, flour, sugar, maize meal, and canned foods; consisting of maize beer, and a small amount (<1%) of liquor. A smaller proportion of their diet came from Government agencies, and very little animal husbandary, gardening, and the gathering of wild fruits and vegetables. A negligible part of their diet came from hunting activities.

When one considers that in 1958, much of these people’s dietary intake came from the hunting of wild animals it will be appreciated that major changes have occurred….

Until 1980 !Kung inhabitants of Chum!kwe had lived in traditional settlements, well spaced from one another. In that year the government established three housing schemes or townships, consisting of 23, 18, and 14 single roomed houses built from cinder blocks with communal pit type lavatories and bathhouse and a single water tap situated in the centre of the township.

The people spend most of their time outside the houses and, in fact, do their cooking on, and sleep around, a fire which they light in front of the house, facing the central courtyard. The lavatories are never used because the inhabitants walk off into the surrounding bush to relieve themselves. Litter and refuse accumulate in these townships. We have observed that many of the traditional settlements in Chum!kwe moved twice, some moved three times, between August 1980 and December 1981. As a result, those that do move tend to be cleaner than the Government housing projects….

The present study has shown that the change in lifestyle undergone by the San between 1969 and 1981 has been accompanied by a deterioration in their previously excellent iron and folate nutrition and in an increase in the incidence of anemia, most of which is almost certainly nutritional in origin. Alcoholism, previously unknown, has become a major problem.

It is probably not feasible for the San to return to a hunter-gatherer existence because the available land could support only a fraction of the present populations in that mode; neither do the San wish to return to the past. [22]

So the transition to settled life occurred in the 1960s. The !Kung San now depend on government support. Their diet is corn porridge, corn beer, wheat and sugar. They are malnourished and show signs of liver damage. Their settlements are filthy, so much so that they don’t use the lavatories; when a settlement becomes intolerable they create a new one; some communities moved three times in 16 months.

Let me record here two other facts about the !Kung San that may be relevant:

  • They are small — comparable in height and weight to Central African pygmies. The average height of Bushmen males is 158 cm (5 foot 1 inch), and average weight is 46 kg (101 pounds). The average height of Bushmen females is 145 cm (4 foot 9 inch), and average weight is 38 kg (83 pounds). [27]
  • They are short-lived . The life expectancy of !Kung San in the period 1963-1974, which spans the period of transition from hunter-gatherer to settled life, may have been a bit over 50 years – or, conceivably, as low as 35 years:

How shall we interpret this?

First, let’s consider the recently acquired numbers of 143 and 158 mg/dl, which used modern measurement techniques on well-preserved samples and should be trustworthy.

These numbers are precisely what we would expect from an infection-ridden African population with short (~50 year) lifespan. Most sub-Saharan African countries with high infectious disease burdens and shortened lifespan have mean population serum cholesterol close to this.

For instance, Mauritania has mean serum cholesterol of 157 mg/dl. Life expectancy in Mauritania is 61 years. Nigeria has mean serum cholesterol of 136.9 mg/dl and a life expectancy of 47.6 years. Cameroon has mean serum cholesterol of 120.9 mg/dl – the lowest in the world – and a life expectancy of 54 years. Infectious disease risk is rated high in Mauritania and very high in Nigeria and Cameroon.

If TCs among the settled !Kung San meet our expectations, where does this leave us in regard to  to the low serum cholesterol reported for the hunter-gatherer San in the 1960s? My expectations would be:

  • If San hunter-gatherers were healthier than the settled !Kung San, then their serum TC should have been higher than the settled !Kung San TC of 143 to 158 mg/dl. In this case the 1960s measurements would have to be in error.
  • On the other hand, if San hunter-gatherers were sicker than the settled San, then a TC of 120 – similar to that of modern Cameroon – or even lower becomes plausible.

I don’t think either possibility can be ruled out.

The hunter-gatherers may have had poor health. Although anthropologists consider malnutrition more prevalent in the settled !Kung San, their data is not sufficient to show that life expectancy or infectious disease burden became worse when the !Kung San gave up their hunter-gatherer life. And recall also the statement by Fernandes-Costa et al: “neither do the San wish to return to the past.” [22] The hunter-gatherer lifestyle seems to have been no idyll.

The data may have been wrong. If the early measurements performed on stale, poorly refrigerated samples with pre-modern techniques underestimated !Kung San hunter-gatherer TC by 33%, a plausible estimate for the amount by which Corcoran & Rabinowitch 1937 may have underestimated Eskimo TC, then it would bring the measurements of Truswell & Hansen [24] and Bersohn & Tobias right in line with modern-day African TCs.


The !Kung San provide some support to the claims in Eaton et al [1] that hunter-gatherers had low serum cholesterol.

However, the data offers no support to the thesis, argued vigorously in O’Keefe et al [2], that low serum cholesterol levels are healthy. Rather, the evidence from Africa is that low serum cholesterol levels reflect a high burden of infectious disease, especially parasitic disease, and are invariably associated with shortened lifespan.

African hunter-gatherer serum cholesterol levels were probably below 160 mg/dl, maybe well below, and their low cholesterol levels marked a lifespan that was little more than 50 years.

Related Posts

The posts in this series are:


[1] Eaton SB, Konner M, Shostak M. Stone agers in the fast lane: chronic degenerative diseases in evolutionary perspective. Am J Med. 1988 Apr;84(4):739-49. Full text:

[2] O’Keefe JH Jr, Cordain L, Harris WH, Moe RM, Vogel R. Optimal low-density lipoprotein is 50 to 70 mg/dl: lower is better and physiologically normal. J Am Coll Cardiol. 2004 Jun 2;43(11):2142-6.

[3] Konner M, Eaton SB. Paleolithic nutrition: twenty-five years later. Nutr Clin Pract. 2010 Dec;25(6):594-602. Full text:

[4] Barnicot NA et al. Blood pressure and serum cholesterol in the Hadza of Tanzania. Hum Biol. 1972 Feb;44(1):87-116.

[21] van der Westhuyzen J et al. Thiamin status and biochemical indices of malnutrition and alcoholism in settled communities of !Kung San. J Trop Med Hyg. 1987 Dec;90(6):283-9.

[22] Fernandes-Costa FJ et al. Transition from a hunter-gatherer to a settled lifestyle in the !Kung San: effect on iron, folate, and vitamin B12 nutrition. Am J Clin Nutr. 1984 Dec;40(6):1295-303.

[23] Miller K et al. Lipid values in Kalahari bushmen. Arch Intern Med. 1968 May;121(5):414-7.

[24] Truswell AS, Hansen JD. Serum-lipids in bushmen. Lancet. 1968 Sep 21;2(7569):684.

[26] Tichelaar HY et al. Plasma lipids and fatty acids in urbanized Bushmen, Hereros and Kavangos of southern Africa (Namibia). Lipids. 1992 Sep;27(9):729-32.

[27] Joffe BI et al. Metabolic responses to oral glucose in the Kalahari Bushmen. Br Med J. 1971 Oct 23;4(5781):206-8.

Serum Cholesterol Among the Eskimos and Inuit

We’re investigating the surprising claim, set forth by S. Boyd Eaton, Melvin Konner, and Marjorie Shostak 1988 [1] and reiterated by Loren Cordain and collaborators in 2004 [2] and Boyd and Konner 2010 [3], that hunter-gatherers had total serum cholesterol below 135-140 mg/dl (3.49-3.62 mmol/l), and that this was healthy. A summary of their claims can be found in Tuesday’s post (Did Hunter-Gatherers Have Low Serum Cholesterol?).

The original Eaton et al paper [1] listed Hadza, Kalahari San (Bushmen), Congo Pygmies, Canadian Eskimos, and Australian Aborigines as hunter-gatherer populations with low cholesterol. So I’m going to survey cholesterol levels in those groups, and maybe a few others.

I had intended to cover all groups in one post, but I found that doing the topic justice requires more elaborate discussion. Therefore, I’ll give each group its own blog post, starting with the Eskimos and Inuit.

A word of caution: Please do not jump to conclusions until the series is complete. I will withhold most of my analysis until the end of the series. For now I am just presenting data, assessing its quality, and including relevant facts about the health of the study population.

Summary of the Data

Searches on “Eskimo cholesterol” and “Inuit cholesterol” turned up a lot of papers. I didn’t look at all of them but I looked at most of the pre-1988 papers (which Eaton et al could have cited) and a sampling of recent ones.

Here is a table summarizing what I found. Papers are ordered from oldest to newest:

Cholesterol Levels of Eskimo & Inuit Populations 

Paper [ref] Mean TC Notes
Corcoran & Rabinowitch 1935 [9] 141 mg/dl Stale samples, obsolete technique, small sample size; tuberculosis common
Wilber & Levine 1950 [11] 218 mg/dl
Rodahl 1955 [13] 215 mg/dl
Pett & Lupien 1958 [10] 204 mg/dl
Scott et al 1958 [12] 214 mg/dl
Davies & Hanson 1965 [8] 182 mg/dl Diseased (tuberculosis), life expectancy 32 years
Ho et al 1972 [7] 221 mg/dl
Dyerberg et al 1975 [6] 216 mg/dl
Young et al 1993 [15] 205 mg/dl Average of 4 age and gender cohorts
Howard et al 2010 [14] 211 mg/dl TC calculated from LDL 125, HDL 62, triglycerides 118
Makhoul et al 2010 [5] 223 mg/dl

Results are remarkably consistent. Nine of the eleven papers reported mean total cholesterol (TC) between 204 mg/dl and 223 mg/dl. Let’s look more closely.

Studies that Reported Low Serum Cholesterol

The TC of 141 mg/dl reported by Corcoran & Rabinowitch 1937 [9] among Hudson Bay and Baffin and Devon Island Eskimos is precisely the number quoted for Canadian Eskimos by Eaton, Konner and Shostak [1]. Presumably this was their source for that number.

How solid is the number? Corcoran & Rabinowitch acquired samples from only 27 non-fasting Eskimo men. The measurement method was archaic and the samples were not fresh:

None of the tests was completed during the voyage. The work then was confined to collection of the blood samples and their necessary treatment to preserve the different constituents to be examined. All analyses were made on oxalated plasma. [9]

Corcoran & Rabinowitch note the unreliability of lipid measurements from the 1930s:

A survey of the literature shows wide variations of the different lipoid constituents of blood, both in fasting experiments and following ingestion of food, in animals and in man, and whether the analyses were made upon whole blood, red blood cells, plasma or serum. Correlation of these data is difficult because of the variety of technical methods with which they were obtained. [9]

Possible evidence for deterioration of the samples is the fact that, although twenty of the twenty-seven Eskimos were eating zero-carb diets, no ketones were found in any sample:

Also suggestive of an unusual mechanism for the utilization of fat is the absence of ketosis in these natives, whereas the urines of both [Stefansson and Andersen during the Bellevue All-Meat Trial] contained acetone. The explanation of this absence of ketosis is not entirely clear. [9]

My guess is the ketones had evaporated before measurement, or were otherwise degraded.

I might add that in 1935, when the Corcoran & Rabinowitch samples were collected, the Eskimo had already begun to deviate from the hunter-gatherer lifestyle. Rabinowitch 1936 [16] reports:

These Eskimos are employed by the Police, and live in huts for a great part of the year. Their food and clothing are also to some extent the products of civilization … The food supplied by the Police must be supplemented by the natural foods of their environment- seal, etc. These Eskimos still spend much time hunting. [16]

Rabinowitch observed flour in about half the Eskimo tents he visited, and was told at Lake Harbour, which may have had the lowest flour consumption of the sites he visited, that the average annual flour consumption was 130 pounds for a family of three Eskimos. [16]

Infectious disease, notably tuberculosis, was common:

Tuberculosis was common in the Straits and Bay. At Chesterfield Inlet, of 62 persons examined 22 had some respiratory disturbance; and of these 12 had coughs with no detectable adventitious sounds in the lungs; 2 had what appeared to be bronchitis only; and 8 had active pulmonary tuberculosis. In addition to these 8 cases, active glandular tuberculosis (cervical) was found in 4 children. At Port Burwell, of 31 natives examined 8 had coughs with no detectable adventitious sounds; 2 had what appeared to be bronchitis; and 5 adults had active pulmonary tuberculosis. Two children had masses of confluent glands in the neck. There was one case of tuberculosis of bone (phalanx); there was no reason to suspect lues in this case. At Coral Harbour there was a child with tuberculosis of a knee joint. Two cases of active pulmonary tuberculosis were found at Lake Harbour. At Wolstenholme one child was found with a mass of confluent glands in the neck. [16]

They also had parasites and worms:

From reports of the Institute of Parasitology of McGill University by Drs. TWM Cameron and IW Parnell it is obvious that the Eskimo is exposed to a variety of parasitic infections. These authors have found that at least three-quarters of all the animals examined, birds, duck, geese, etc., harboured parasites. The polar bear, walrus, and weasel were found free, but most of the seals were infected with Ascarides and intestinal flukes. The Eskimo lives in intimate contact with his dogs, and carcasses and feces of these animals are heavily parasitized with hookworm, Ascarides, flukes, and tape worms. Ascarides, taenia and hookworm were found as far north as Craig Harbour, and hares from Ellesmere Island were heavily infected with worms. Nail scrapings of Eskimos were found high in content of Oxyuris vermicularis…. Our pathologist, Dr LJ Rhea, in his search for parasites found 6 cases of eosinophilia amongst 34 blood smears. [16]

The other paper showing a mean population TC below 200 mg/dl was Davies & Hanson 1965 [8], who found a mean TC of 182 mg/dl. I liked this paper because it gave a lot of textual background concerning the health and diet of the 727 Canadian Northwest Passage Eskimos studied. Some quotes from this study:

[Seventy to ninety percent] live at sealing or fishing camps and visit the trading posts twice yearly or more often, depending on distance.

Ten to 25% of their food is obtained from trading posts …

[L]ife expectancy of the Eskimo is about 32 years … [8]

Health was poor, as you’d expect from the short life expectancy. Infectious disease was a serious problem. Many had had tuberculosis or brucellosis; chronic coughs were common. Many had abnormal blood cell counts, such as eosinophilia, neutrophilia, and lymphocytosis. Some had diabetes, despite low-carb diets; I suspect the combination of alcohol and omega-3 fats (also low vitamin D) to have been the culprit. However, iodine status was excellent and hypothyroidism was extremely rare.

My guess is that these Eskimos were bringing a lot of tobacco, alcohol, and infectious disease back from those trading posts.

Considering that 75% to 90% of their food was acquired in the traditional way, a life expectancy of 32 years is not exactly a ringing endorsement of the healthfulness of the Eskimo/Inuit diet. The poor health of this group of Eskimos may have contributed to their relatively low TC of 182 mg/dl.

Studies that Reported High Serum Cholesterol

The other nine studies reported mean serum TC between 204 and 223 mg/dl. These levels fall in the minimum mortality region of O Primitivo’s database and are suggestive of good health.

I particularly like one recent paper, Makhoul et al 2010 [5], because it sampled Yup’ik Eskimo eating the traditional diet, and included scatter plots showing each individual’s cholesterol numbers. They write:

Because of their traditional diet, which is based largely on fish and other marine foods (20), Yup’ik Eskimos have a mean intake of EPA and DHA that is >20 times the current mean intake of the general US population (3.7 compared with 0.14 g/d in men and 2.4 compared with 0.09 g/d in women) (21). Studies of Yup’ik Eskimos offer a unique opportunity to examine how a broad range of EPA and DHA intakes (22) affect chronic disease biomarkers. [5]

Some Eskimos in their sample got as much as 15% of calories from EPA+DHA. Their cholesterol levels:

TC is mostly between 200 and 240 mg/dl, LDL between 100 and 160, and HDL between 50 and 70. Cholesterol increased as fish oil intake increased – evidence that cholesterol gets higher as the diet becomes more traditional.

Other studies also found that the more traditional the Eskimo diet, the higher were total cholesterol levels. Here is the discussion in Ho et al 1977 [7], who found mean total cholesterol of 221 mg/dl in Arctic Eskimos:

This value was in general agreement with that obtained from other mass samplings of Arctic Eskimos (8-11) but was slightly higher than those values obtained from the Eskimos living on the Pacific Coast of Alaska, as reported by Scott et al. (9). [PJ: Scott et al, my reference [12], found mean TC of 214 mg/dl.] A generalization was made by Scott and co-workers from their study on 842 Eskimos that northern Alaskan Eskimos have higher serum cholesterol levels than southern Eskimos. The reason for this difference might well be related to the differences in their diets, as the main staple of northern Eskimos is marine mammals, whereas that of the southern Eskimos includes some vegetables and fish (12). [7]

In general, the studies reporting mean TC over 200 mg/dl all reported that their study population were eating a diet resembling the traditional hunter-gatherer diet. In Arctic populations, this diet featured high intake of marine mammals and low intake of carbohydrates.


The vast majority of studies show that Eskimo and Inuit populations have mean serum cholesterol over 200 mg/dl. The only studies showing mean serum cholesterol below 200 mg/dl sampled tuberculosis-ridden populations with short life expectancy. The study showing the lowest mean serum cholesterol used obsolete sample preparation and measurement techniques on stale samples.

The most parsimonious explanation of the data is that TC of 200-230 mg/dl is normal for Eskimos and Inuit, that lower TCs indicate the presence of infectious diseases such as tuberculosis, and that the very low TC of Corcoran & Rabinowitch 1937 may have suffered from sample degradation during the two-and-a-half-month voyage (July 13-Sept 29) before samples could be measured in Montreal.

It would be difficult to attribute the low TC in Corcoran & Rabinowitch 1937 to diet, as the subjects ate flour and other government-provided foods and did not obviously eat a more traditional diet than the Eskimo of later studies. The most salient difference between the Corcoran & Rabinowitch subjects and those of later studies was the high incidence of tuberculosis in 1935. Perhaps the low TC in Corcoran & Rabinowitch was due to tuberculosis, but Davies and Hanson found a mean TC of 182 mg/dl in another tuberculosis-ridden Eskimo population.

The Corcoran & Rabinowitch 1937 paper will be useful to us because it gives us an indication what may happen to measured TC levels when an older measurement method, that of Abell, is applied to stale samples. It appears that such measurements may under-report TC by as much as 1/3 (210 mg/dl to 141 mg/dl).

Related Posts

The posts in this series are:


[1] Eaton SB, Konner M, Shostak M. Stone agers in the fast lane: chronic degenerative diseases in evolutionary perspective. Am J Med. 1988 Apr;84(4):739-49. Full text:

[2] O’Keefe JH Jr, Cordain L, Harris WH, Moe RM, Vogel R. Optimal low-density lipoprotein is 50 to 70 mg/dl: lower is better and physiologically normal. J Am Coll Cardiol. 2004 Jun 2;43(11):2142-6.

[3] Konner M, Eaton SB. Paleolithic nutrition: twenty-five years later. Nutr Clin Pract. 2010 Dec;25(6):594-602. Full text:

[5] Makhoul Z et al. Associations of very high intakes of eicosapentaenoic and docosahexaenoic acids with biomarkers of chronic disease risk among Yup’ik Eskimos. Am J Clin Nutr. 2010 Mar;91(3):777-85.

[6] Dyerberg J et al. Fatty acid composition of the plasma lipids in Greenland Eskimos. Am J Clin Nutr. 1975 Sep;28(9):958-66.

[7] Ho KJ et al. Alaskan Arctic Eskimo: responses to a customary high fat diet. Am J Clin Nutr. 1972 Aug;25(8):737-45.


[9] Corcoran AC, Rabinowitch IM. A study of the blood lipoids and blood protein in Canadian Eastern Arctic Eskimos. Biochem J. 1937 Mar;31(3):343-8.

[10] Pett LB, Lupien PJ. Cholesterol levels of Canadian Eskimos. Federation Proc. 17(1958): 488, 1958.

[11] Wilber CG, Levine VE. Fat metabolism in Alaskan Eskimos. Exp Med Surg. 1950 May-Nov;8(2-4):422-5.

[12] Scott EM et al. Serum cholesterol levels and blood pressure of Alaskan Eskimo men. Lancet. 1958 Sep 27;2(7048):667-8.

[13] Rodahl K. Diet and cardiovascular disease in the Eskimos. Trans Am Coll Cardiol. 1955 Apr;4:192-7.

[14] Howard BV et al. Cardiovascular disease prevalence and its relation to risk factors in Alaska Eskimos. Nutr Metab Cardiovasc Dis. 2010 Jun;20(5):350-8.

[15] Young TK et al. Cardiovascular diseases in a Canadian Arctic population. Am J Public Health. 1993 Jun;83(6):881-7.

[16] Rabinowitch IM. Clinical and Other Observations on Canadian Eskimos in the Eastern Arctic. Can Med Assoc J. 1936 May;34(5):487-501. Full text:

Did Hunter-Gatherers Have Low Serum Cholesterol?

Emily raised a great question in response to last Tuesday’s post:

[W]hat of the reports of hunter gatherers having low cholesterol. Is it the product of fringe environments, or low infectious burden, or what?

Let’s look into this. Do hunter-gatherers in fact have low cholesterol? If so, why?

The Claim

As far as I know, this idea originated with and was promoted by the fathers of Paleo dieting, S. Boyd Eaton and Loren Cordain, and their collaborators.

Its first appearance, to my knowledge, was in a 1988 paper by S. Boyd Eaton, Melvin Konner, and Marjorie Shostak called “Stone agers in the fast lane: chronic degenerative diseases in evolutionary perspective” [1]. Here’s their data:

The footnote to Table IV reads as follows:

The published paper has 101 references and takes 11 pages in the journal, yet no supporting references for the cholesterol data were included.

Here is a graph from a 2004 paper by Loren Cordain, William Harris, and some pro-statin medical doctors [2] (thanks Stabby!):

The caption states that total cholesterol (TC) ranges between 70 and 140 mg/dl in hunter-gatherers, and LDL cholesterol (LDL-C) between 35 and 70 mg/dl. However, this claim is unsourced. The paper provides references for assertions that LDL tends to be around half TC, and that modern Americans have TC around 208 and LDL-C around 130, but there are no references for hunter-gatherer cholesterol levels.

The data in this graph seem to be drawn from the Eaton, Konner, and Shostak paper [1]. The Hadza number is the same as the 109.5 mg/dl (2.83 mmol/l) average of the male and female Hadza in [1]; Inuit at 141 mg/dl (3.65 mmol/l) is the same as “Canadian Eskimos” in [1]; !Kung and San (probably the same people) are both listed very close to the 119.5 mg/dl (3.09 mmol/l) average of “Kalahari San (Bushmen)” in [1]; Pygmy looks identical to the 106 mg/dl average of male and female “Congo Pygmies” in [1]. It looks like they just copied from Eaton et al but deleted the Australian Aborigines who in [1] had a male-female average TC of 139 mg/dl (3.59 mmol/l).

Eaton and Konner were sticking to the low hunter-gatherer cholesterol claim in 2010 [3]; they cited only their original 1988 paper [1] when they wrote:

Our review of various health measures in HG and other nonindustrial populations showed that average HG serum total cholesterol was always below 135 mg/dL … [3]

So over 23 years, to judge by these papers, Eaton et al and Cordain et al have yet to cite a peer-reviewed article in support of the proposition that hunter-gatherers had low cholesterol. Where did this idea come from? And is it true?

The Evidence Is Worth Looking Into

The claim that healthy hunter-gatherers had serum cholesterol below 140 mg/dl is quite surprising, given that contemporary populations are healthiest when their serum cholesterol is over 200 mg/dl, and mortality rises and life expectancy falls sharply as serum cholesterol falls below 180 mg/dl. (See Blood Lipids and Infectious Disease, Part I, Jun 21, 2011.)

Are hunter-gatherers – either their diets or their genetics – so different from modern populations? Or is the claim that healthy hunter-gatherers have low serum cholesterol a mistake?

I think this is an interesting question, with implications both for the design of Paleo diets and for our interpretation of serum lipid results. When we discussed HDL, I argued that some dietary methods to raise HDL might benefit us by enhancing immunity (see HDL and Immunity, April 12, 2011; HDL: Higher is Good, But is Highest Best?, April 14, 2011; How to Raise HDL, April 20, 2011). Might a similar strategy for dietary manipulation of LDL be desirable too?

Next Steps

I’ll examine the issue in 3 parts:

  • On Thursday I’ll survey the literature on hunter-gatherer cholesterol. What are their numbers really?
  • Next week I’ll continue the Blood Lipids and Infectious Disease series by looking at the immune functions of LDL cholesterol. What happens to LDL when people get infections? Is there an optimal LDL level?
  • In conclusion of the series I’ll return to the issue of human populations – whether hunter-gatherer, horticultural, pastoral, or modern – and what their cholesterol levels tell us about their health. Why do some populations have low serum cholesterol and other populations have much higher cholesterol?

This might lead us into issues such as: Has there been recent human evolution toward higher cholesterol levels? Are there biological differences in optimal cholesterol levels among different human populations – for instance, Africans and Eurasians, or aboriginal populations and descendants of Neolithic farmers?

Should be fun!

Related Posts

The posts in this series are:


[1] Eaton SB, Konner M, Shostak M. Stone agers in the fast lane: chronic degenerative diseases in evolutionary perspective. Am J Med. 1988 Apr;84(4):739-49. Full text:

[2] O’Keefe JH Jr, Cordain L, Harris WH, Moe RM, Vogel R. Optimal low-density lipoprotein is 50 to 70 mg/dl: lower is better and physiologically normal. J Am Coll Cardiol. 2004 Jun 2;43(11):2142-6.

[3] Konner M, Eaton SB. Paleolithic nutrition: twenty-five years later. Nutr Clin Pract. 2010 Dec;25(6):594-602. Full text:

Blood Lipids and Infectious Disease, Part I

If there will be a general theme to our second year of blogging, it will be chronic infections – how they interact with the body to promote disease, and how we can use diet-related techniques to successfully combat them.

Blood lipids, such as LDL and HDL cholesterol, provide a fascinating window into health. We’ve already discussed both LDL and HDL (Low Carb Paleo, and LDL is Soaring – Help!, March 2, 2011; Answer Day: What Causes High LDL on Low-Carb Paleo?, March 3, 2011; HDL and Immunity, April 12, 2011; HDL: Higher is Good, But is Highest Best?, April 14, 2011; How to Raise HDL, April 20, 2011), but there’s quite a bit more to be said.

The extraordinary Portuguese blogger Ricardo Carvalho, better known as O Primitivo, of Canibais e Reis did some great work a few years back assembling World Health Organization statistics into an Excel database. One of the fruits of this labor was that he was able to correlate disease rates against serum cholesterol levels for all the countries in the database. His post about that is here and he created a great graphical representation of the results which I’ve reproduced here (click to enlarge):

There’s a lot of interesting information in this graph.

On the upper right are some correlation coefficients between serum cholesterol and disease incidence. Most diseases are either uncorrelated with cholesterol, or negatively correlated, meaning that mortality goes up as cholesterol goes down.

Minimum mortality is found in countries with average cholesterol between 200 and 240 mg/dl. Mortality rises sharply as cholesterol levels fall below 200 mg/dl.

No ecological fallacy

When aggregating over populations, it’s possible to get misleading results – an outcome called the “ecological fallacy”. However, I think this relationship is pretty solid.

It’s been confirmed in individuals in clinical trials. For instance, in the Japan Lipid Intervention Trial, here was the table of mortality as a function of serum lipids:

You’ll note that the “relative risk” of dying was lowest for TC between 180 and 260, LDL-C between 80 and 160, and HDL-C between 60 and 70.

The minimum mortality region of O Primitivo’s national samples falls right in the middle of the JLIT minimum mortality region.

Infectious disease mortality depends strongly on cholesterol levels

By far the strongest dependence of mortality on cholesterol levels is found for infectious diseases. Infectious disease mortality is shown by the dashed green curve.

Infectious disease mortality approaches zero where TC averages between 215 and 245 mg/dl. It rises very sharply as TC falls. Over a wide range of TC, for every 10 mg/dl drop in average TC, mortality rises by 200 infectious deaths per year per 100,000 population.

Other diseases are partly infectious in origin

As regular readers know, I believe that many other “noncommunicable” diseases are actually caused in part by chronic infections. Cardiovascular disease is one – atherosclerotic lesions are universally found to be infected, macrophages usually need to be infected in order to become “foam cells” which contribute to atherosclerotic plaque formation, and the infectious burden in lesions drives the risk of fragmentation of plaque into the clots which cause heart attacks and strokes.

If cardiovascular disease is partly due to infections, and the protective effect of cholesterol against infections is present here, then we should expect cardiovascular disease mortality (shown in the red dashed line) to rise as TC falls.

It does. Mortality from cardiovascular disease starts rising as TC falls below 205 mg/dl and rises by about 200 infectious deaths per year per 100,000 population for every 25 mg/dl drop in average TC.

Causality probably runs in both directions

We can’t directly infer causality from these statistics. Consider two possible directions of causality:

  • Serum cholesterol – LDL and HDL – help defend against infections. As long as you have these infection-fighting lipids in your blood, infections can’t kill you.
  • Infections destroy LDL and HDL, and the more severe the infection, the lower blood cholesterol goes. By the time the infection is severe enough to kill you, TC is very low. So countries with high infectious burden have both low TC and high infectious mortality.

The first line of causality is certainly true. I’ve already discussed the important infection-fighting properties of HDL – in fact I’ve argued that infection fighting, not cholesterol transport, is the primary function of HDL (See HDL and Immunity, April 12, 2011). As I’ll discuss, LDL also has immune functions.

The second line of causality is also quite likely. Pathogens evolve ways to suppress immunity; that is what makes them effective pathogens. If LDL and HDL are crucial for immune defense, then pathogens will have found ways to destroy or damage them.

In this series I would like to explore these interactions between blood lipids and infections a little more deeply. They may lead us to ways we can tweak our diet to improve our defenses against disease — or help doctors recognize under what circumstances taking statins will raise, not lower, cardiovascular disease risk.


[1] Matsuzaki M et al. Large scale cohort study of the relationship between serum cholesterol concentration and coronary events with low-dose simvastatin therapy in Japanese patients with hypercholesterolemia. Circ J. 2002 Dec;66(12):1087-95.