- Top of page
- LITERATURE CITED
Paleolithic diets are increasingly acknowledged as templates for healthy diets, partly because very low age-adjusted rates of cardiovascular disease and other nutrition-related disorders have been observed among contemporary hunter-gatherers (Eaton and Konner,1985; Trowell and Burkitt,1981; Truswell and Hansen,1976) and traditional horticulturalists (Lindeberg and Lundh,1993; Sinnett,1977). Another reason is that the majority of Westerners are affected by atherosclerosis and associated abnormalities, and our understanding of the main underlying causes is very limited (Lindeberg,2005, 2010). If there is a healthy diet for humans in general, irrespective of ethnicity, it makes sense to focus on the time period up to the emergence of fully modern humans in Africa (Campbell and Tishkoff,2010). This brief review relates the recent evidence of Western diseases, as well as common concepts of healthy nutrition, to probable food patterns of our paleolithic ancestors before they left Africa.
The staple food items typically consumed by our bipedal ancestors in Africa is a matter of debate, but the principal food available included sweet and ripe fruits and berries, shoots, flowers, buds and young leaves, meat, bone marrow, organ meats, fish, shellfish, insects, larvae, eggs, roots, bulbs, nuts, and nongrass seeds (Gräslund,2005; Ungar,2007). In principle, this was the only type of food that was available during human evolution, but now only provides about one quarter of the caloric intake for the average European or American. In contrast, we currently get most of our energy from grains (grass seeds), dairy products, refined fat and sugar, and legumes. In addition, we have very little variation among plant foods today.
Fruits may have been consumed more or less regularly by our primate ancestors during 50 million years until they became bipedal around 6 million years ago (Bloch and Boyer,2002; Gräslund,2005; Ungar,2007). For chimpanzees, bonobos, and orangutans, fruit makes up more than 75% by weight of the diet. Fruit was also the most common plant food among 20th century hunter-gatherers in the Ethnographic Atlas (n = 229; Cordain et al.,2000a). Fruits differ from other edible plants in that they contain appreciable amounts of fructose, a monosaccharide, which typically constitutes 20–40% of available carbohydrates in wild fruits (Dzhangaliev et al.,2003; Ko et al.,1998; Milton,1999) and 10–30% in cultivated fruits (National Food Administration,1986). Honey, where 50% of the carbohydrate is fructose, may have been consumed in considerable amounts for a couple of months (Allsop and Miller,1996).
A high intake of fructose has been proposed to cause abdominal obesity and associated metabolic disturbances, including diabetes type 2, high blood pressure, blood lipid disorders (high triglycerides and low HDL cholesterol), and fatty liver (Johnson et al.,2009). However, a fructose intake below 60 g/day, corresponding to 4–5 kg/day of cultivated pineapples, is considered to be safe (Sanchez-Lozada et al.,2008), even for people with diabetes type 2 (Sievenpiper et al.,2009). Approximately two-third of dietary fructose in the US population is provided by non-natural foods and additives, mainly sucrose and high-fructose corn syrup (Park and Yetley,1993).
Starchy underground storage organs (roots, tubers, bulbs, and corms) may have become staple foods during periods of repeatedly dry and cool climates 1–2 million years ago (Wrangham and Conklin-Brittain,2003) or even before that (Laden and Wrangham,2005). Humans have a relatively high activity of salivary amylase (Perry et al.,2007; Samuelson et al.,1990), and our tooth morphology, including incisal orientation, seems to be well designed for chewing tubers (Lucas et al.,2006). To increase the caloric yield per workload, roots and tubers may often have been an adequate choice (O'Connell et al.,1983). The excellent health status among starch-eating ethnic groups (Lindeberg,2010; Sinnett,1977) contradicts the notion that starch per se is a cause of obesity and type 2 diabetes. Although a high-starch load undoubtably raises blood sugar after a meal, the main cause of an individual's inability to limit blood sugar rise after eating carbohydrates (i.e., glucose intolerance) remains obscure, and it is questionable if dietary starch plays a causative role (Brinkworth et al.,2009; Due et al.,2008; Kennedy et al.,2005; Noakes et al.,2006; Reaven,2005). Likewise, the proportion of carbohydrate in the diet has not been proven to significantly affect long-term body weight (Foster et al., 2010; Brinkworth et al., 2009).
Another food that could provide a high-energy yield is meat. It is consumed in considerable amounts by the chimpanzee (Stanford,1999). In one observational study, adult chimpanzees consumed an average of 65 g meat per day in the dry season (Stanford,1999). For humans, available archaeological evidence is consistent with, but does not prove, regular high-meat intake in the last 2 million years (Ungar,2007). Contemporary hunter-gatherers have generally been able to eat large amounts of meat or fish, although the figures are based on rather imprecise ethnographic data (Cordain,2000; Panter-Brick, 2001). Of the 229 hunter-gatherer populations studied during the 20th century, the majority (73%) were estimated to get more than half their caloric intake from meat, fish, and shellfish (Cordain et al.,2000a). Among those five African populations, for which more exact, quantitative data were available, meat and/or fish constituted on average 26, 33, 44, 48, and 68% of the food (Cordain,2006).
Hunter-gatherers have had exceptionally favorable levels of serum cholesterol, blood pressure, and other cardiovascular risk factors, even with very high-meat consumption (Lindeberg,2010). Truswell and Hansen (1976) found no evidence of sudden, spontaneous death when interviewing 96 adults among the San tribe, hunter-gatherers in the Kalahari desert, Botswana, South Africa. However, wild game meat has a lower fat content and a higher percentage of omega-3 fatty acids than domestic meat (Cordain et al.,2002; Mann,2000; Naughton et al.,1986). Animal experiments do not suggest that meat causes atherosclerosis, the main underlying cause of cardiovascular disease (Lindeberg,2010). The notion that “animal protein” causes atherosclerosis is based on studies with milk proteins, typically casein (Lindeberg,2010). In Western human populations, an association between coronary heart disease and reported consumption of meat/meat products has been found in a few case–control studies (Kontogianni et al.,2008), while other epidemiological studies have been less convincing (Hu et al.,2000; Keys,1980; Lea and Worsley,2002; Menotti et al.,1999; Mente et al.,2009; Moss and Freed,2003; Williamson et al.,2005).
Although Palaeolithic African humans may have lived at times by the shores of lakes and rivers where they could catch fish and shellfish, dependence on marine food seems to have emerged essentially after humans left Africa (Richards,2002). Omega-3 fatty acids from fish have been suggested to prevent coronary heart disease, although randomized controlled trials suggest that the effect is limited (Hooper et al.,2004, 2006). Observational studies have found an ∼ 20% lower cardiovascular risk among Westerners who eat fatty fish at least twice a week, than among those who do not (Hu et al.,2001; Schmidt et al.,2000). Possibly, this only applies to high-risk populations (Marckmann and Gronbaek,1999). The notion that fish is protective originally emerged from an observed very low incidence of myocardial infarction and sudden cardiac death among the Inuits of Greenland and Canada (Bang and Dyerberg,1972; Schaefer,1971). Of all the characteristics of Inuit dietary habits that could possibly explain these findings, all focus has been on omega-3 fatty acids, while the absence of other Western foods has rarely been considered.
Nuts can provide a high amount of energy for the amount of work involved and have probably made up an essential part of the diet during certain time periods. On the whole, nuts have a beneficial nutrient composition: they are typically rich in monounsaturated fat, protein, soluble fiber, and micronutrients, while low in saturated fat (Lindeberg,2010). In observational studies in the United States, a high intake of nuts has been associated with a lower risk of myocardial infarction (Kris-Etherton et al.,2001; Tunstall-Pedoe,1998). Nuts are low in water and have thus a high energy density, but their effect on body weight is not known.
The consumption of insects and larvae is thought to have been substantial in most African Paleolithic habitats, and they may have provided an important source of protein and fat (DeFoliart,1999; Ungar,2007). They are regularly consumed by nonhuman primates (Basabose,2002). Our knowledge of their health effects is extremely limited.
The extent to which alcohol could have been a regular part of human's original environment is unknown (Kiple and Ornelas,2000). If storage vessels were made of leather or plants in earlier prehistoric times, they have long since disappeared without a trace. The attraction of intoxication is evident in most ethnic groups in the world and seems to be independent of the effect of culture (Smith,1999). Even without deliberate production of alcoholic beverages, a low-level dietary exposure to ethanol via ingestion of fermenting fruit may have characterized our lineage of humans and human-like ancestors for about 40 million years (Dudley,2002). The expected consequences of ethanol in the relevant amounts would seem slightly beneficial (Ronksley et al.,2011).
It is often impossible to determine, for any particular habitat, the percentage of food that came from each of the available foodstuffs. In the context of public health, it may be crucial to consider some of the foods that were not staple diets during human evolution. On average, roughly three-fourth of the calories in Western countries is today provided by foods that were practically unavailable during human evolution: wheat and other cereal grains, dairy foods, refined fats, and sugar. In addition, sodium intake is now high.
Wild seeds were available from various plants, but not from the grass family (Poaceae), which includes today's wheat, rice, and maize, and rarely or never from one plant species every day. Seeds from legumes apparently became staple foods during the emergence of agriculture, as evidenced by gradual changes in their form and quality in consequence of domestication (Berger et al.,2003; Zohary and Hopf,1973). Contemporary hunter-gatherers, in particular those living in arid, hot, marginal environments (Australian Aborigines, Kalahari Bushmen), often include large, fatty seeds in their diet, but these provide a relatively small amount of energy, and each of them not by far as much as is now provided by wheat, rice, or maize (Cordain et al.,2005; Lee,1968; O'Connell et al.,1983).
When seeds from any one particular plant species are consumed in large amount on a regular basis, an interesting situation arises. The plant kingdom contains thousands of bioactive substances and other natural chemicals, phytochemicals, many of which are thought to be part of the defence system against herbivores (Herrera and Pellmyr,2002; Wynne-Edwards,2001). The highest concentrations are generally found in the most vital parts (sprout, seeds, and beans). Such phytochemicals can often make up 5–10% of the plant's dry weight (Perantoni,1998). Prehistoric foragers were supposedly able to limit the negative health effects by having access to a large number of various plant species, by consciously avoiding the most poisonous ones (Ulijaszek and Strickland,1993), and by the use of cooking (Wrangham and Conklin-Brittain,2003). One example is plant lectins (Chrispeels and Raikhel,1991; Van Damme et al.,1998). Unrefined grain products, on the whole, have a higher lectin content than refined seed products. Lectins in wheat, rye, rice, and potatoes bind to receptors in the “host organism” (Lindeberg,2010). Compared to other dietary proteins, plant lectins are unusually resistant to enzymatic breakdown in the intestines, and can penetrate the intestinal mucous membrane, finally being deposited in the internal organs. The long-term, potentially negative effects of plant lectins alone may include atherosclerosis (Freed,1999; Kritchevsky et al.,1998; Lindeberg,2010), diabetes (Freed,1999; Jonsson et al.,2005; Lindeberg,2010; Wang et al.,2001), and autoimmune diseases (Cordain et al.,2000b; Freed,1999; Lindeberg,2010).
The impact of dairy foods on human health has been debated for decades, especially with regard to cardiovascular disease and bone strength. A human autopsy study in 1960 found myocardial infarction to be more common among patients who had undergone the milk-based Sippy diet than among those who had been treated at hospitals where the Sippy diet had not been prescribed (Briggs et al.,1960). Internationally, cardiovascular mortality has been positively associated with the intake of dairy products (Segall,1994), and recent changes in milk consumption have correlated with changes in mortality rates of coronary heart disease in Europe (Moss and Freed,2003). Comparisons within, rather than between, countries suggest that other lifestyle factors than dairy foods are more important to explain the variation of coronary heart disease in such populations (Elwood et al.,2005). Animal experiments have found that casein, the dominant milk protein, promotes atherosclerosis (Wilson et al.,2000) and other forms of intracellular fat overload (lipotoxicity; Lindeberg,2010).
Allen and Cheer (1996) found a striking inverse association between diabetes and lactase persistence rates with the latter explaining almost half of the worldwide variation in diabetes prevalence, indicating cow's milk as a potential cause of diabetes (type 2), although junk food is likely worse (Elwood et al.,2010). Animal experiments suggest that the possible culprit should be sought in the nonfat portion, such as milk proteins or lactose (Hugi et al.,1998; Lavigne et al.,2001; Lindeberg,2010; Nilsson et al.,2004; Tovar et al.,2005). Furthermore, several observational studies suggest that cow's milk may contribute to prostate cancer: among US physicians, high-milk consumption has been associated with an increased risk of prostate cancer (Chan et al.,2001), and other prospective cohort studies have generally found similar results (Food, nutrition and the prevention of cancer: a global perspective,2007; Grant,1999; Park et al.,2007a, b). Milk has has also been discussed as a cause of breast cancer (2007; Hakkak et al.,2001), the incidence of which covaries by almost 20% with prostate cancer internationally (Parkin et al.,2005). Notably, a high intake of cow's milk in childhood may lead to greater attained height, which is one of the most consistent risk factors for breast cancer (Food, nutrition and the prevention of cancer: a global perspective,2007).
The role of milk for bone strength is uncertain. Prospective observational studies do not suggest that a high-milk intake appreciably lowers the risk of fracture in the modern world (Benetou et al.,2011; Bischoff-Ferrari et al.,2011; Warensjo et al.,2011). Furthermore, calcium supplementation has only had a marginal beneficial effect in randomized controlled trials (Bolland et al.,2010). Hence, recent evidence suggests that the influence of milk and calcium intake on bone remodeling has been overestimated (Jackson et al.,2006; Lanou et al.,2005; Nordin et al.,1995; Winzenberg et al.,2006), while absorption and excretion of calcium may have been underrated (Lindeberg,2010). Absorption is inhibited by phytate from grains and beans (Reinhold et al.,1973), and losses are possibly increased by acid-producing foods, such as dietary salt and cereal grains (Alexy et al.,2007; Buclin et al.,2001; Frassetto et al.,2007; Sebastian et al.,2002).
There are no studies of the occurrence of osteoporosis and bone fractures among contemporary hunter-gatherers, but existing data suggest that hip fractures have been uncommon among traditional aboriginal populations (Adebajo et al.,1991; Barss,1985; Bloom and Pogrund,1982; Prentice et al.,1990; Scrimgeour,1992; Solomon,1979). Despite a large number of documented fractures in fossilized bones from Palaeolithic hunter-gatherers, osteoporotic fractures are also considered to be relatively rare in prehistoric skeletons south of the Arctic circle (Agarwal and Grynpas,1996; Roberts and Manchester,1997; Webb,1995). Eskimos constitute an apparent exception, with low-bone density in the 20th century (Harper et al.,1984; Mazess and Mather,1974, 1975; Pawson,1974), and vertebral compression fractures commonly encountered in prehistoric skeletons (Merbs,1989; Yesner,1981). The question is whether vitamin D deficiency is unavoidable north of the polar circle. In addition, the traditional Eskimo diet, rich in meat and fish, and low in vegetables, tubers, and fruit, would result in a net acid load (Sebastian et al.,2002).
In addition to being more or less devoid of grains and milk, refined fatsand sugar were obviously not included in Paleolithic diets. Hence, the intake of fiber and most micronutrients was generally higher (Cordain et al.,2005; Lindeberg,2010). There are no known nutrients in grains, milk, or refined fats, required by humans, that are not provided by a mixture of meat, fish, shellfish, fruit, vegetables, nuts, and eggs (Lindeberg,2010). Calcium intake may often have reached current recommendations, especially when the intake of green leafy vegetables was high (Cordain et al.,2005; Lindeberg,2010). Because iodized salt and dairy products were not available, only those ancestors with high-regular access to fish or shellfish would be expected to have reached the currently recommended intake of iodine (unless they were regularly consuming animal thyroids; Lindeberg,2009). Paleolithic diets were often high in protein, typically 15–35% of energy (E%), with quite variable proportions of fat and carbohydrate (Cordain et al.,2000a; Kuipers et al.,2010). The significance of the proportion of these macronutrients for public health is a matter of debate (Lindeberg,2010). In the treatment of obesity, macronutrient proportions appear much less important than calorie balance (Brinkworth et al.,2009; Clifton,2008; Foster et al.,2010; Sacks et al.,2009; Shai et al.,2008).
Our own contributions to the field include one survey in the Trobriand Islands, Papua New Guinea, and two intervention trials among Swedish patients with Western disease. In Kitava, Trobriand Islands, we noted an apparent absence of cardiovascular disease and associated risk factors among Kitava's 2,300 inhabitants (6% of which were 60–95 years old) as well as among the remaining 23,000 people in the Trobriand Islands (Lindeberg,1994, 2010; Lindeberg and Lundh,1993; Lindeberg et al.,1997). Yam, sweet potato, taro, and fruit were staple foods while grains, dairy, refined fats, and sugar were absent. The study adds to the notion that some of our most common diseases are preventable and that a high-carbohydrate intake is not a problem in itself (Lindeberg,2010).
In our clinical trials, we randomly assigned people with diabetes type 2 or glucose intolerance to one of two prudent diets, with or without grains and dairy products (Jönsson et al.,2009; Lindeberg et al.,2007). In these studies, we noted beneficial effects of the “Paleolithic” diet with regard to waist circumference, glucose tolerance, blood sugar, blood pressure, and blood lipids (Jönsson et al.,2009). We also found some evidence that the “Paleolithic” diet was more satiating, such that the meals gave a feeling of fullness at a lower level of energy intake (Jönsson et al.,2010). In an uncontrolled study among nonobese sedentary healthy volunteers, Frassetto et al. (2009) found improved levels of plasma insulin in the fasting state (by 68%) as well as after a glucose load (by 39%), blood lipids, and blood pressure. Similarly, a sharp improvement in body weight, blood sugar, blood lipids, and blood pressure was noted in urbanized aborigines with type 2 diabetes when they returned to a hunter-gatherer lifestyle for 7 weeks (O'Dea,1984). Physical activity was also increased in that study.
There are very few obvious risks with a paleolithic diet. The effect of a high-protein intake on kidney function is debatable and is expected to be outweighed by the beneficial effects on abdominal obesity and other health-related variables (Franz et al.,2002; Gannon and Nuttall,2006; Lindeberg,2010). People with genetic hemochromatosis, a hereditary disease that results in enhanced iron absorption, need to limit their intake of meat and fish but should otherwise do well on a Paleolithic diet. Heterozygous carriers are advised to check their iron status regularly after middle age. People who are treated with ACE-inhibitors, AII-blockers, or diuretics should only switch slowly to a salt-free palaeolithic diet in order to avoid a sharp drop in blood pressure (Milan et al.,2002). Patients with type 2 diabetes who are on sulphonylurea preparations (glipizid, glibenklamid, and glimepirid) are at risk for low-blood sugar when making the radical switch to a palaeolithic diet. Switching to a palaeolithic diet during on-going warfarin treatment should be done in consultation with a physician or nurse.
In summary, a paleolithic diet may serve as a model for healthy foods, in particular, in clinical trials in comparison with other prudent diets. Hypothetically, food choice is more important than counting calories or macronutrients in order to avoid common health problems in the Western world. Lean meat, fish, shellfish, vegetables, tubers, fruit, berries, nuts, and eggs can relatively safely be tried in the prevention and treatment of disease. Hypothetically, dairy products, margarine, oils, refined sugar, and cereal grains, which provide 70% or more of the dietary intake for modern humans, are not an optimal food choice for long-term health. However, much more research is needed. Although this review has focused on what most humans would have in common, studies on possible ethnic differences in susceptibility to diet-induced disease are also needed.