Eur J Nutr 39 : 67–70 (2000)

© Steinkopff Verlag 2000 SPECIAL ISSUE

Stanley Boyd Eaton Paleolithic vs. modern diets – selected

Stanley Boyd Eaton III
pathophysiological implications

Received: 24 February 2000

Summary The nutritional patterns rary and ancestral diets have many
Accepted: 20 March 2000 of Paleolithic humans influenced ge- pathophysiological implications. This
netic evolution during the time seg- review addresses phytochemicals and
ment within which defining charac- cancer; calcium, physical exertion,
teristics of contemporary humans bone mineral density and bone struc-
were selected. Our genome can have tural geometry; dietary protein,
changed little since the beginnings of potassium, renal acid secretion and
agriculture, so, genetically, humans urinary calcium loss; and finally sar-
remain Stone Agers – adapted for a copenia, adiposity, insulin receptors
Paleolithic dietary regimen. and insulin resistance.
Such diets were based chiefly on While not, yet, a basis for formal
wild game, fish and uncultivated recommendations, awareness of Pale-
plant foods. They provided abundant olithic nutritional patterns should
protein; a fat profile much different generate novel, testable hypotheses
from that of affluent Western nations; grounded in evolutionary theory and
high fibre; carbohydrate from fruits it should dispel complacency regard-
and vegetables (and some honey) but ing currently accepted nutritional
S. Boyd Eaton, MD () · S. Boyd Eaton III not from cereals, refined sugars and tenets.
Dept Anthropology dairy products; high levels of micro-
Emory University
2887 Howell Mill Road nutrients and probably of phyto- Key words Paleolithic diet –
Atlanta GA 30327, USA chemicals as well. insulin resistance – skeletal health –
e-mail: [email protected] Differences between contempo- phytochemicals – type 2 diabetes

Introduction While there has clearly been some Holocene genetic adap-
tation, primarily to infectious diseases (e. g. malaria), in-
The Paleolithic or Old Stone Age occupied nearly all the creasing population size and greater mobility have actually
last two million years of human evolutionary experience, reduced the likelihood of genetic innovation in recent mil-
an especially critical time segment during which the adap- lennia [2]. It is, therefore, not mere rhetorical hyperbole to
tations that define our genus were selected: body mass and assert that contemporary humans are, in the genetic sense,
shape, locomotive capability, masticatory apparatus, still Stone Agers and that, consequently, we remain adapted
growth/developmental schedule, relative (and absolute) for a preagricultural nutritional pattern.
brain size, resting metabolic rate, and daily foraging range
[1]. Experiential changes associated with agriculture and
industry have been dramatic and might readily have influ- Paleolithic diets
enced the human genome given sufficient time; however,
the very rapidity of these innovations has almost totally There was no universal nutritional regimen for Stone
overreached the capacity of genetic evolution to keep pace. Agers, but in terms of mathematical set theory, the super-
68 European Journal of Nutrition, Vol. 39, Number 2 (2000)
© Steinkopf Verlag 2000

Table 1 General paleolithic nutrition Table 2 Paleolithic nutrition

Nutrient Typical Variation with Latitude Paleolithic1 Current US2 Ratio
Protein (mg/d) (mg/d)
Animal Very High Positive MINERALS
Vegetable Moderate Negative Calcium 1622 920 1.8
Fat Copper 12.2 1.2 10.2
Total Moderate to High Positive Iron 87.4 10.5 8.3
(~ Mediterranean vs. Magnesium 1223 320 3.8
E. Asian) Manganese 13.3 3.0 4.4
C20 & Very High Positive Phosphorus 3223 1510 2.1
C22 LCPUFA1 Potassium 10500 2500 4.2
n–6:n–3 Ratio ~1 Positive Sodium 768 4000 0.2
Serum Cholesterol Zinc 43.4 12.5 3.5
Raising FA Low Positive VITAMINS
Cholesterol High (~ US levels) Positive Ascorbate 604 93 6.5
Carbohydrate Folate 0.36 0.18 2.0
Cereals None to Minimal Negative Riboflavin 6.49 1.71 3.8
Vegetables & Thiamin 3.91 1.42 2.8
Fruits Very High Negative Vitamin A 17.2 7.8 2.2
Dairy Foods None After Infancy – Vitamin E 32.8 8.5 3.9
Refined Sugars None (Honey) – 1 based on 3000 kcal/d, 35 % animal: 65 % plant subsistence
Fibre Very High Negative 2
average of US men and women; Food and Nutrition Board, 1989
Micronutrients Very High Negative
Phytochemicals (Probably High) (Probably Negative)
1long-chain
polyunsaturated fatty acids
Evolution, nutrition, and pathophysiology

set believed to encompass Paleolithic dietary patterns (e. g. An appreciation of Paleolithic nutrient-nutrient interac-
high latitude, arid lands, equatorial, coastal) and the super- tions and of the relationships between diet and other as-
set including contemporary ones (e. g. traditional East pects of ancestral experience can enlighten topics of cur-
Asian, Mediterranean, typical American, vegetarians) are rent interest and/or controversy. Three (of many possible)
largely disjoint; they overlap surprisingly little. Preagricul- examples are phytochemicals, skeletal health and insulin
tural humans ate wild game, fish, uncultivated plant foods resistance.
and, when available, honey. Grains were for emergencies
and there were no dairy products, oils, salt, processed
foods, nor empty calories. Phytochemicals
In almost all cases, Stone Agers consumed more animal
protein than do current Westerners (Tab. 1). Total fat con- No free-living primates except humans consume cereal
sumption varied, chiefly with latitude, but intake of serum- grains, but from the emergence of agriculture, wheat, rice,
cholesterol-raising fat was nearly always less than for corn, and their like have provided from 40 to 90 % of our
Americans and Europeans [3] while there was more di- energy requirements. In so doing they have reduced the
etary long-chain (C20 and above) polyunsaturated fatty contribution of fruits and vegetables, the major energy
acid (LCPUFA) [4]. The preagricultural essential fatty acid source for Stone Agers and their simian predecessors over
ratio (ω6:ω3) approached equality; for average Americans the preceding fifty million year evolutionary span. Recent
this ratio is 10:1 or higher. Dietary cholesterol roughly epidemiological metaanalyses strongly suggest that fruits
equated U. S. levels. Carbohydrate consumption also var- and vegetables have far more cancer-preventive potential
ied with latitude, but, in all cases, came chiefly from fruits than do cereals [5], perhaps reflecting the phytochemical
and vegetables, not from cereals, refined sugars, and dairy content of non-cereal plant foods, phytochemicals to which
products. Compared with the Western European dietary current human biology became adapted through many mil-
pattern, Paleolithic foods provided much more fibre (both lion years of interrelationships. In contrast, the phyto-
soluble and insoluble) and from two to ten times more mi- chemicals of grains have interacted with the human
cronutrients (Table 2); there was probably a similar phyto- genome for only 10,000 years. While it required epidemi-
chemical discrepancy. ology to demonstrate that fruits and vegetables have cancer
preventive-potential (and more such potential than do
cereals), evolutionary understanding suggests why this
 S. Boyd Eaton et al. 69
Paleolithic vs. modern diets

should be so and also implies that the partial replacement Insulin resistance
of fruits and vegetables by cereals following the emergence
of agriculture probably increased our susceptibility to neo- Insulin resistance and hyperinsulinaemia are increasingly
plastic diseases. thought of as central factors in the pathophysiology, not
only of type 2 diabetes, but also of hypertension and coro-
nary heart disease. While it is clear that genetics, diet and
Skeletal health exercise are all involved in the development of insulin re-
sistance, the exact mechanism remains unclear. Since the
Despite their lack of dairy products, Paleolithic diets gen- field remains open, perhaps an evolution-based hypothesis
erally provided more calcium than do the diets of Ameri- merits consideration.
cans and Europeans. Wild plant foods often contain con- The relationship between excess adiposity and insulin
siderable calcium, the average for 119 such items being resistance is well recognised, but evolutionary considera-
132.6 mg/100 g (vs. whole milk ~120 mg/100 mg) [6]. tions suggest a parallel linkage involving relative skeletal
However, in far northern latitudes, plant foods sometimes muscle deficiency. Contemporary humans are distin-
made up only 5 % of annual energy intake, so that low bone guished from their predecessors not just by hyperadiposity,
mineral density (osteoporosis) was common among Stone but also by sarcopenia [10]. In addition, current physical
Agers living in such regions. Nevertheless, fractures were fitness levels are much below those typical in the past.
infrequent because a bone’s strength is determined not only These altered parameters distort the physiological milieu
by its mineral density, but also by its structural geometry, for insulin action from what it was when the genetic bases
especially its diameter and cross-sectional configuration for carbohydrate metabolic regulation was selected.
[7]. Structural geometry is affected by habitual physical Gramme for gramme, insulin-stimulated muscle can ex-
activity which tends to increase bone diameter and create tract far more glucose from the blood than can insulin-
cross-sections more oval than round. Such bones resist me- stimulated adipose tissue and exercise-conditioned muscle
chanical stresses effectively, even when osteoporotic. This extracts more than does non-conditioned muscle [11].
effect helps explain why Melanesians and black South Hence, a logical hypothesis is that functional insulin resis-
Africans, whose calcium intake is low, have far fewer age- tance, especially in its earliest stages, is directly propor-
related fractures than do black and white Americans whose tional to body fat mass and inversely proportional to the
calcium intake is greater. mass and metabolic activity of skeletal muscle. This rela-
Diets high in protein have been positively correlated tionship might reflect competition between the insulin re-
with hip fracture rates, presumably because protein in- ceptors of myocytes and those of adipocytes for available
creases urinary calcium loss. The mechanism involves en- insulin molecules. The initial effect would be repetitive
dogenous acid production which is increased by metabo- episodes of transient hyperglycaemia and hyperinsuli-
lism of protein; renal acid secretion and urinary calcium naemia. In genetically susceptible individuals, further
loss are highly correlated. Conversely, dietary potassium metabolic deterioration could result from secondary down-
produces a net alkalinising effect, opposite to that of pro- regulation of insulin receptors, glucose transporters, and
tein, and it appears that net endogenous acid production, intracellular enzymatic sequences, leading ultimately to
which largely determines net renal acid excretion, can be glucose intolerance and type 2 diabetes.
predicted by assessing the dietary protein (g/d) / potassium
(mg/d) ratio, with lower values being protective [8]. The
extremely high potassium content of Paleolithic diets (e. g. Conclusion
10 500 mg/d for a 3000 kcal, 35:65 animal:plant subsis-
tence pattern vs. ~2500 mg/d in the average American diet) At one level the insights arising from these examples – eat
thus may have exerted a critically beneficial influence: a more fruits and vegetables, reduce sodium intake, and in-
protein / potassium ratio of 0.84 for typical Stone Agers crease exercise – are banal. However, there are more pro-
compared with an average value of 1.24 for 141 American found implications. Basing the nutritional pyramid on
subjects consuming 20 different contemporary diets. grains is wholly out of line with primate and hominid evo-
Dietary sodium also increases urinary calcium loss [9]; lutionary experience. A Paleolithic model experimental
its effect is roughly similar in magnitude to that of protein. diet – high protein, high fibre, minimal serum-cholesterol
The typically low sodium intake of Stone Agers (~768 raising potential, balanced n–3:n–6 fatty acid ratios, high
mg/d vs. Americans ~4000 mg/d) would have exerted ad- LCPUFA content, etc. – deserves investigative support.
ditional protective influence on Paleolithic skeletal health. Micronutrient intake throughout evolutionary experience
exceeded current RDAs. The importance of bone structural
geometry and possible myocyte-adipocyte insulin receptor
competition both deserve study.
Paleolithic nutritional awareness is not, yet, a basis for
formal recommendations, but it can generate testable hy-
70 European Journal of Nutrition, Vol. 39, Number 2 (2000)
© Steinkopf Verlag 2000

potheses grounded in evolutionary theory. And it should

dispel complacency regarding currently accepted nutri-
tional tenets.

References

1. Wood B, Collard M (1999) The human Food, Nutrition and the Prevention of tein contents. Am J Clin Nutr 68:
genus. Science 284: 65–71 Cancer: a Global Perspective. American 576–583
2. Tattersall I (1998) Becoming Human. Institute for Cancer Research, Washing- 9. Devine A, Criddle RA, Dick IM, Kerr
Evolution and Human Uniqueness. Har- ton D.C, pp 506–507 DA, Prince RL (1995) A longitudinal
court Brace, New York; pp 239. 6. Eaton SB, Nelson DA (1997) Calcium in study of the effect of sodium and calcium
3. Eaton SB, Eaton SB III, Konner MJ evolutionary perspective. Am J Clin Nutr intakes on regional bone density in post-
(1997) Paleolithic nutrition revisited: a 54: 281S–2817S menopausal women. Am J Clin Nutr 62:
twelve-year retrospective on its nature 7. Ruff CB (1992) Biomedical analyses of 740–745
and implications. Eur J Clin Nutr 51: archaeological human material. In: Saun- 10. Rode A, Shephard RJ (1994) The physi-
207–16 ders SR, Katzenburg A (eds). The Skele- ological consequences of acculturation: a
4. Eaton SB, Eaton SV III, Sinclair AJ, Cor- tal Biology of Past Peoples. Alan R Liss, 20-year study in an Inuit community. Eur
dain L, Mann NJ (1998) Dietary intake of New York, pp 41–62 J Appl Physiol 69: 16–24
long-chain polyunsaturated fatty acids 8. Frassetto LA, Todd KM, Morris Jr RC, 11. DeFronzo R (1997) Pathogenesis of type
during the paleolithic. World Rev Nutr Sebastian A (1998) Estimation of net en- 2 diabetes: metabolic & molecular impli-
Diet 83: 12–23 dogenous noncarbonic acid production cations for identifying diabetes genes.
5. World Cancer Research Fund, American in humans from diet potassium and pro- Diabetes Rev 5: 177–269
Institute for Cancer Research (1997)