Evolution is bloody marvellous. It is, quite simply, the means by which we understand much of biology. The staggeringly talented British immunologist Peter Medawar, described by the evolutionary biologist and essayist Stephen Jay Gould as ‘The cleverest man I have ever known’, said that ‘For a biologist, the alternative to thinking in evolutionary terms is not to think at all.’ Gould himself was no stranger to lauding evolution, stating that it is ‘one of the half-dozen shattering ideas that science has developed to overturn past hopes and assumptions, and to enlighten our current thoughts’. So while others may have put it more eloquently than me, I think that, when faced with the mind-bending diversity of adaptation and life on Earth, ‘bloody marvellous’ does just as good a job.
For all its greatness, there is a problem with looking at life through the lens of evolution. We can sense this peril lurking in one of the most famous quotes of all about evolution, the title of an influential essay by Ukrainian-American geneticist and evolutionary biologist Theodosius Dobzhansky: ‘Nothing in Biology Makes Sense Except in the Light of Evolution’. It is such a wonderful title, bordering on an arrogance virtually impossible to ignore. But if you think about it, taken at face value it simply isn’t true.
The problem with the brilliantly sweeping statement of Dobzhansky is that not everything we see in biology, not every feature of animals, plants, fungi, single-celled protists or bacteria, is necessarily an evolved adaptation. In other words, not every feature has come about because there is a genetic component underlying that feature that confers an advantage to the bearer of those genes that translates into increased fitness and resultant increase in gene frequency. The problem is that the lens of evolution is so powerful and its explanatory power so great that viewing the world through it can lead to a tendency to see the hand of natural selection and evolution in absolutely everything. Adaptationism is the scientific position that ascribes an evolved adaptive origin for many of the features that we can document in organisms and it is a wonderful approach. But adaptationism can be taken too far. Take a simple physical feature of humans, our belly button. This universal structure of all mammals nurtured in their mother’s womb via an umbilical cord is wonderfully variant in form, and unparalleled in its ability to take a piercing and go septic, but it is not an evolved characteristic. There has been no selection for the belly button. Belly buttons are not a genetic characteristic that have conferred such an advantage to those that have them that the genes coding for them have swept through the population to fixation (when all individuals have the genes) in just a few generations. Belly buttons are merely the attachment point of the umbilical cord connecting you to your mother’s placenta. They are basically a scar. Thanks Theodosius, but belly buttons make total sense without the light of evolution. That said, the umbilical cord that produces the belly button is an evolved adaptation, so it can all get a little complex to disentangle.
‘Armchair adaptationism’ is sitting at a distance and concocting evolutionary Just So stories, named for the Rudyard Kipling short stories like ‘How the Leopard Got His Spots’. It is always a temptation to interpret every feature that we can see in organisms as an evolved characteristic, and that is never more of a problem than when we turn the lens of evolution on ourselves. Applying the beautifully simple hammer of selection logic is just so tempting and satisfying that we can end up seeing nails everywhere. So, why this cautionary preamble? Well, in this chapter, I’m going to explore the complex, emotive and biologically difficult topic of obesity. As we’ll see, the route to understanding why we are fat is littered with tempting nails dying to be hammered down with evolutionary arguments. We will need to remember that just because an evolutionary Just So story is seductive and ‘obvious’, it doesn’t mean it’s true.
Seductively thrifty genes
I’ll explore this statement more scientifically later but for now let’s just stick to a tabloid level of analysis and accept that humans are ‘getting fat’. The well-repeated evolutionary argument as to why we are so fat these days is that we are famine-adapted creatures that find ourselves living in times of feast. The story goes like this. Prior to the invention and spread of agriculture about 12,000 years ago in the so-called Neolithic Revolution (see Chapter 3), we were operating on a largely subsistence basis. This period, the Palaeolithic (the Stone Age), saw our ancestors and other hominins with whom we shared common ancestry living as hunter-gatherers. We’ll get into more details of their diet and environment shortly, but as a first approximation we can assume that these early humans would have been relatively non-selective omnivores. Plant products, including high-protein and calorie-dense nuts, vitamin- and carbohydrate-rich berries and starchy root tubers would have been particularly prized and, judging from archaeological evidence, meat also featured on our early menu. Meat would likely have been acquired opportunistically by grabbing whatever didn’t move fast enough. Small mammals, birds, eggs, insects and shellfish would have been consumed with gusto when the opportunity presented itself and the flesh of larger animals could have been scavenged from carcasses. We also developed increasingly technological hunting methods using tools, communication and social organisation to bring down large prey. In times of plenty, perhaps when a host of berries were in season, frogs were croaking in the lakes and game was plentiful, it is easy to imagine our ancestors full-bellied and satisfied. However, the hunting would not always have been good. Animals can be remarkably hard to find and kill (that’s evolution for you), while plant bonanzas are highly seasonal. Full bellies might be easy to imagine but hard times and hunger would never have been far away.
We are capable of ingesting far more calories in the food we eat than we need to survive. In times of plenty we could easily have had an excess of calories available to us. To get through hard times of famine it makes sense to be able to store the surplus energy available during times of feast. This is the precise reason why honey bees make honey; it is a food store laid down when nectar is highly available in the summer and can be converted to honey to last through the winter when no flowers are in bloom. While honey bees have evolved an external food store, we have evolved an internal food store, or fat as it is known. Our history of feast and famine, so the story goes, has led to Homo sapiens being highly adept at laying down fat stores in times of abundance to survive the leaner times ahead. In the modern world, and most especially in More Economically Developed Countries (MEDC), we no longer go through periods of feast and famine but instead are surrounded near-constantly with calorie-dense food. The upshot is that we live in times of absolute feast but are physiologically ever-prepared for a future famine that will almost certainly never come, but just in case it does, our bodies lay down fat and so we balloon. This evolutionary argument has become known as the thrifty gene hypothesis (TGH).
Since it was proposed back in the 1960s, the thrifty gene hypothesis has gained considerable ground in the public sphere and it has become the go-to evolutionary explanation in the media for why obesity is a crisis in the modern world. Indeed, such is the power of this hypothesis that obesity has become something of a figurehead for the notion that those of us in MEDCs have created a world for which we are not adapted. It also suggests a very simple remedy: eat a diet more like our ancestors and you shouldn’t get fat.
As simple as it sounds, the evolutionary cause of, and evolutionarily informed solution for, the obesity problem subdivides into a number of different and sometimes complex steps that need to be supported with evidence if we are to understand and solve the obesity problem with evolutionary thinking. We need to identify what fat is, what obesity actually means, how obesity is related to other conditions and whether this is really a feature of the modern world or whether obesity was around long before we started ‘going large’ on our burger and fries. Then, we need to dig down (in some cases literally) to examine what our ancestors ate and what their environment was like compared to ours, especially with respect to famine (a central feature of the TGH). If we are to blame an evolutionary mismatch for current trends in obesity, then we are going to have to explore the selection pressures our ancestors faced and compare them with those that are exerted by the environment in which we live today. Once that ground is laid, we need to bring out the big guns; we need to look for the genetic evidence for a ‘thrifty’ gene or genes because if evolution is to blame then it must have left its signature in our genome. Finally, the evolutionary-inspired claims that backdating our diets to those of our paleolithic forebears will solve the obesity problems need to hold up to scrutiny. Paleo diets and caveman nutritional plans need to be more than just marketing gimmicks. It’s a house of cards because if the evidence for any one of those steps is lacking, it starts to cast considerable doubt on the whole enterprise. So, let’s start with the simplest question: what is fat?
What is fat?
In Chapter 1 I touched on the fact that our bodies are composed of cells that can be organised collectively to form tissues that perform a particular function. One such cell-collective is adipose tissue and it consists of cells called adipocytes, also known as lipocytes (fats being biochemically part of the family of substances called lipids). More plainly, we can call these cells ‘fat cells’. Fat cells are wonderful at storing fat and they collect together to form the adipose tissue you can feel when you grab a chunk of soft yielding flesh.
Adipose tissue actually comes in two types, brown and white. Brown adipose tissue (BAT) is found in almost all mammals and in humans it tends to accumulate around the kidneys and neck, between the shoulder blades and along the spinal cord. BAT is involved in producing heat by the process known as non-shivering thermogenesis. Shivering produces heat (thermogenesis) as a by-product of the involuntary contraction of our muscles. Non-shivering thermogenesis allows the chemical energy in fat to be converted to heat without the need for movement via a clever metabolic pathway present in BAT cells. BAT is medically intriguing, being associated with, among other things, bone density, bone health and increased longevity. As a consequence of all these functions BAT is metabolically active with a good blood supply, and the fat cells themselves have small, accessible droplets of fat stored within them. They also have plentiful mitochondria, the ‘powerhouses’ of cells that are required to fuel BAT cells in their complex (and still not fully understood) role.
As interesting as BAT is, it is white adipose tissue (WAT) that makes us ‘fat’. BAT is brown because of the plentiful mitochondria within the cells and rich blood capillary networks supplying them, whereas WAT is white because it doesn’t need to do anything especially interesting beyond storing energy. BAT cells have small droplets of fat stored within them but the fat cells of WAT have a lovely big space, or vacuole, right in the middle. This vacuole can fill up with energy-rich fat, and when full the vacuole pushes all the other cell contents (like the nucleus) tight up against the cell membrane.
WAT is present all over our bodies. Subcutaneous WAT fat forms just under the skin, whereas visceral WAT fat accumulates deep in the abdomen and surrounds our organs. Subcutaneous fat leads to beer bellies, love handles, amply cushioned buttocks, large thighs, a jowly face and flabby arms. It tends to be the fat we can see and the fat we worry about aesthetically, but in terms of our health the evidence suggests that subcutaneous fat is not causing the main problems. The really big negative health implications of fat storage come from the visceral fat we cannot see piling up around our internal organs. A high amount of visceral fat is a major player in the development of type 2 diabetes, which is one of the biggest medical problems associated with obesity. Type 2 diabetes is an increasingly common condition that causes the level of glucose (a type of sugar) in the blood to elevate because we become resistant to the effects of insulin, the hormone that regulates our blood sugar levels. Symptoms include excessive thirst, frequent urination, increased hunger, tiredness and, a symptom that isn’t talked about so much, an itching around the penis or vagina. These are relatively minor short-term symptoms compared to the very serious long-term implications of type 2 diabetes, which include heart disease, strokes and impaired blood flow that can lead to blindness, kidney failure and, in extreme cases, foot, hand and limb amputations. Other solid links, both direct and indirect, have been established between obesity and heart disease, strokes, certain types of cancer and psychological problems including depression and low self-esteem. To this we can add the pain caused from excessive joint wear, especially of the knee, and sleep disorders including sleep apnea (an interruption to normal breathing while asleep) and excessive snoring. As if these weren’t enough to persuade you that obesity is a bad thing, obese people are 25 times more likely to report problems with their sex lives, whether it be lower libido, erectile dysfunction or issues linked to body image and performance anxieties. It’s undeniable: being obese is bad.
Obesity and diabetes
As well as being one of the major health problems associated with obesity, type 2 diabetes plays an interesting part in our consideration of the importance of past evolution on our present state. It was in attempting to account for the rise of type 2 diabetes that the thrifty gene hypothesis was first postulated in 1962 by geneticist James Neel. Neel was an influential character in the development of an evolutionary understanding of human genetics and disease, although his work was not without controversy. In 2000, author Patrick Tierney in his book Darkness in El Dorado laid out a series of very serious accusations primarily against Neel and anthropologist Napoleon Chagnon, although other researchers also came under fire. The most serious of the accusations made by Tierney was that Neel had initiated a measles epidemic in 1968 among the Yanomami of the Amazon. The charge was that Neel had deliberately used a hazardous version of the measles vaccine in order to test theories on human evolution, the existence of ‘leadership genes’ and the evolution of infectious diseases in humans. According to Tierney, Neel had used the Edmonston B type of measles vaccine even though a safer and cheaper version was available. The motivation, so the claim went, was that the Edmonston B vaccine produced antibodies that would allow Neel to compare European and Yanomami immune responses. Tierney asserted that the use of this vaccine to fulfil a scientific purpose rather than using a safer alternative led to a measles epidemic. The Darkness in El Dorado incident as it related to Neel certainly illustrates how emotive human evolutionary biology can be, but in fact the measles epidemic claim, and indeed a great many other claims in the book, were later categorically refuted.1
What can’t be refuted quite so easily is the very clear link between obesity and type 2 diabetes. Paradoxically, given that both are disadvantageous, obesity and diabetes are both highly heritable. Heritability suggests a genetic basis and a genetic basis suggests evolution arising from selection, but how can selection operate to promote the increase in frequency of disadvantageous genes? It was in grappling with this conundrum that Neel hit upon the notion that insulin resistance (a precursor of type 2 diabetes) could promote the efficient storage of fat that would allow individuals to survive periods of famine. Thus, a disadvantage (a tendency to promote diabetes) is associated with an advantage during human history when food was scarce and the pros outweighed the cons. It was quickly realised that the link with diabetes was not actually required for selection of thrifty genes as long as they promoted survival against the current environmental background. Fat deposition was a good thing for survival against a background of regular famines and so was born the evolutionary explanation for the heritability (and so the genetic basis) of getting fat.
Feeling like you’re carrying too much timber, being able to grab a little more than you used to around the belly and having the feeling that your clothes are shrinking are all good everyday signs that you are likely to be putting on weight. Medically, obesity is defined a little more precisely. The globally accepted measure for whether you are obese or not is the Body Mass Index or BMI. Adopted by medical professionals and organisations around the world including the World Health Organisation (WHO), BMI is calculated by dividing weight in kilograms by height in metres multiplied by itself, or squared. So, I weigh 83kg (183lb) and I am 1.88m (6ft 2in) tall, which means my BMI is 83 divided by (1.88 x 1.88) = 23.5, give or take. Under current definitions, a BMI between 20 and 25 is ‘normal’ (phew), between 25 and 30 is ‘overweight’ and above 30 is ‘obese’. BMI then increases in jumps of 5, with every jump representing a different and progressively more severe level of obesity, passing through severely obese (35) to morbidly obese (45) and ending (currently) on hyper obese (60+).
BMI has a mathematical logic that is hard to ignore. Taller people are generally heavier and by dividing weight by height you can allow for that and compare more meaningfully between different people. In short, having a higher BMI than someone else means that you are carrying more weight for your height than they are. There are two main issues with BMI, the first of which the more mathematically astute among you may have noticed. When you scale up a three-dimensional object (like a human body) along one ‘axis’ or dimension (like height) you also have to scale up in the other two dimensions (width and depth). Imagine a cube with sides of 1cm (roughly ½in). If you double its height, you must also double its width and length to keep it as a cube. The volume (which equates to its weight) of the original cube is 1 x 1 x 1 = 1cm3, but the volume of the doubled cube is now 2 x 2 x 2 = 8cm3. A more logical formulation of BMI, therefore, would seem to be weight divided by height cubed rather than squared. This was pointed out by Professor Nick Trefethen in a beautifully concise letter to The Economist in 2013:
SIR – The body-mass index that you (and the National Health Service) count on to assess obesity is a bizarre measure. We live in a three-dimensional world, yet the BMI is defined as weight divided by height squared. It was invented in the 1840s [by a Belgian statistician and sociologist called Adolphe Quetelet], before calculators, when a formula had to be very simple to be usable. As a consequence of this ill-founded definition, millions of short people think they are thinner than they are, and millions of tall people think they are fatter.2
Trefethen explores this issue further and, acknowledging that people don’t scale in quite such a perfect way as cubes do (when we grow taller we don’t grow wider and deeper in the exact same proportions), suggests going midway, raising our height to the power of 2.5 rather than 2 (squared) or 3 (cubed). The wonderfully named Corpulence Index (CI) does actually use the cube of height but it never really gained any traction and, for now at least, BMI is here to stay.
Another issue with BMI, and one that many people seem to cling to without always taking much account of the evidence in front of them, is that more muscular people tend to be heavier. The argument goes that muscle is denser than fat and so if you are carrying a lot of muscle mass then you may well be heavier for your height than you ‘should’ be according to BMI calculations. Consequently, you could have a BMI that suggests you are overweight when in fact it is your well-developed musculature that tips you over the edge. If you’re thinking of using this argument then please do take some time to look in your trophy cabinet of Olympic medals for rowing and weightlifting, or at your bodybuilding schedule; if either of them are full then maybe you have a point. For most people though, the ‘it’s not fat, it’s muscle’ argument is not likely to be a winner. BMI also does not tell us anything about the relative proportion of subcutaneous or visceral fat, so it is possible to be carrying potentially dangerous levels of visceral fat but not fall into the 25+ overweight category. Likewise of course, it is possible that your BMI could be 25+ but that your fat is all carried around the buttocks and thighs with fewer health implications.
I guess what I am saying is that BMI may have a few issues when it comes to specific individuals, but arguing over the fine details doesn’t detract from its usefulness when we apply it at a population level. BMI is an effective and accepted measure that allows us to compare between different populations and between and among populations over time, and when we do that very clear patterns emerge. We are fat and getting fatter.
Global obesity patterns
Worldwide obesity has nearly tripled since 1975, and in 2016 39 per cent of adults (aged 18 or over) were overweight (BMI 25–30) and 13 per cent were obese (BMI 30+). What that means is that about one in seven adults on planet Earth right now is obese and two out of every five adults are medically overweight. These numbers are only going up, at least for now. Largely related to this rise in obesity, we have also seen a global increase in diabetes (from 108 million in 1980 to 422 million in 2014).3 But if we are to blame the rise in obesity and the subsequent decline in our health on a mismatch between our evolutionary past and the developed world in which many of us now live, we might expect there to be a relationship between obesity and some measure of ‘development’. To work that out requires more exploration of the data.
Examining the global distribution of those with a BMI of 30 (‘obese’) or more is an interesting first step in understanding the rise of obesity, not least because it yields some rather surprising findings. For example, the top eight obese nations are all Pacific Island nations, headed by Nauru with a scale-busting 61 per cent of people being obese. The first non-Pacific Island nation is Kuwait at number 9 (38 per cent obese) and then comes the USA, flying the flag for fatness at number 10 with 36 per cent of its population being obese. Arab nations feature highly in the back half of the top 20, representing seven of the next eight nations, from Saudi Arabia to UAE (the eighth nation being Turkey at number 15). In the bottom nine of the table, with obesity levels less than 5 per cent, there are poorer nations such as Ethiopia, Eritrea and Nepal but also Japan, a wealthy and highly developed country. Simply ‘eyeballing’ the numbers isn’t enough to determine whether being part of the developed world leads to an evolutionary mismatch and therefore obesity. To do that, we need to take a more considered approach.
It is difficult to find a single measure of nation ‘development’ that everyone feels happy with and so rather like BMI we have to accept that there is a degree of give and take in our deliberations. The United Nations Human Development Index (HDI) for example is a statistical measure that takes into account factors like life expectancy and education, whereas Gross National Income (GNI) or Gross Domestic Product (GDP) are blunter economic measures of productivity. A more nuanced but less snappily named economic measure, ‘GDP (at purchasing power parity) per capita’, accounts for the cost of living and inflation rates of different countries and provides an arguably better measure of ‘lifestyle’ and development. A number of analyses have used a variety of different measures and in fact found much the same thing. Despite ‘noise’ in the data, with some spectacularly obese poorer nations (see below) and some very trim richer nations (e.g. Japan), there are relationships between most measures of economic development and obesity.4 We should be wary of the old adage that correlation does not imply causation, and we should also be very mindful that global obesity cannot be simply waved away with a trite explanation that as we get rich we get access to food aplenty and get fat. But equally clearly, something is going on.
The Pacific Obesity Cluster
The standout anomaly when considering the relationship between obesity and development is the prevalence of extreme obesity in the relatively poor Pacific Island nations, occupying the top eight positions of the global obesity league table. Surely this ‘Pacific Obesity Cluster’ presents a problem for the overall idea that development relates to obesity despite the overall trend? As we’ll see, in fact the clustering of these nations supports the underlying hypothesis and, rather neatly, provides us with an excellent example of the power of recent human evolution to drive mismatches with the modern environment we have created.
We are using development as a proxy measure for access to calorie-dense foods and arguably also a more sedentary lifestyle, which is another driver of obesity. The Pacific Island nations actually support the idea that access to calorie-dense food is the primary driver of obesity, but the major difference between those nations and the wealthier nations that follow them in the league table is that the supply of calorie-dense food is driven not by wealth but by relative poverty. In Tonga (number 5) for example, the diet changed in the mid-twentieth century from fish, root vegetables and coconuts to imported offcuts of meat. Fatty ‘turkey tails’ from the US and ‘mutton flaps’ from New Zealand were cheap and rapidly became popular. Mutton flaps are the low-quality end of sheep ribs and are a staggering 40 per cent fat. Such calorie-laden delicacies are frequently the only meat available and many lay the blame for present-day obesity in these nations firmly at their door. There are likely other cultural factors at play in the South Pacific, such as the historical association of large size with wealth, beauty and power, but the main factor causing present-day obesity in these nations is ready access to surplus calories. Genetic research though reveals that there are relatively recent evolutionary factors at play in the South Pacific that have led to greatly elevated tendencies among these islanders to lay down fat stores. When coupled with a recent environmental change resulting in ready access to calorie-dense foods like mutton flaps, this genetic background conspires to produce the most obese nations in the world.5,6
The story of Samoa
The selection pressure that caused the evolution of ‘super-obesity’ goes a long way to support the thrifty gene hypothesis, at least in the South Pacific. To find out more we must visit the sixth most obese nation on Earth, the beautiful Pacific Island nation of Samoa.
Samoa, like the other Pacific Island nations that lead the global fat league, was settled by people that most likely travelled from Taiwan through the Philippines and eastern Indonesia. This consensus ‘out-of-Taiwan’ model of expansion and settlement of the Pacific is supported by evidence from linguistics (the study of language similarities and differences), genetics and archaeology, and the expansion likely happened around 3,000 years ago. The settlement of Samoa and other island nations would clearly have required considerable ingenuity in boat building, including the development of double-hulled voyaging canoes, and skill in open ocean navigation. It would also have required physical resilience, toughness and the ability to cope with serious food shortages. Boats have limited capacity to carry stores, fresh food would not last long and voyages were long in terms of both distance and time. There would be a clear advantage to those individuals that were able to lay down fat stores more effectively when food was available, perhaps before the voyage or during stop-offs along the way, and to make good use of those fat stores during long periods of on-board famine. Indeed, the long voyages necessary to settle the islands of the South Pacific would seem to provide an ideal selection regime for genetically ‘thrifty’ individuals. But with the caution I advised earlier about the perils of armchair adaptation and evolutionary seduction, we need to examine the genetic evidence to see if our Just So story holds up.
One method of working out whether there is a genetic link to a particular condition, trait or characteristic is to perform a Genome-Wide Association Study, or GWAS. In principle, GWAS are straightforward. You identify a group of people that have the trait you are interested in, in this case people who are obese, and a group of people who don’t. You then genotype all the individuals in both groups and compare the resulting long list of genetic ‘letters’ (the sequence of DNA bases we met in Chapter 1), looking for consistent differences between the groups. Genetic characteristics shared by those in the group with the condition and lacking in the group without the condition are the smoking gun required to ascribe a genetic association to the condition. Of course, in practice GWAS can be very difficult. To be robust, studies require a great deal of genetic know-how and some heavy-duty statistical analysis using tools developed in the incredibly important field of bioinformatics. You also need to control for factors like sex and age in your study population as well as the underlying geographical and ethnic background of participants. These factors can all produce associations that are caused not by a genetic association with the trait of interest but by genetic associations between the participants in each group. Nonetheless, these issues can be resolved and GWAS have been applied medically with great success since the start of this century. Developing through a series of landmark studies of myocardial infarctions in 2002 and age-related macular degeneration (a cause of vision loss in older people) in 2005, the most notable early success came with a 2007 study that revealed, through the study of 17,000 participants, a genetic underpinning for seven diseases including type 1 and 2 diabetes, coronary heart disease and rheumatoid arthritis.
Even allowing for a more generous BMI cut-off of 26 to allow for differences in skeletal build and muscle mass between Europeans and Samoans, 80 per cent of men and 91 per cent of women in Samoa were overweight in 2010. As I discussed in Chapter 1, there could be a solely environmental reason for this (mutton flaps streaming in because of a relatively poorer economy), a solely genetic explanation (actually impossible because you cannot get fat without food, which is always a product of your environment), or an interaction between environment and genes. In this case, the interaction would be between mutton flaps and some putative ‘thrifty genes’. In fact, even before the onset of the modern-day obesity epidemic, the prevalence of obesity in Samoans was greater than in other populations. This strongly suggests that modern-day environmental changes in diet may not be the only factor in the rise of obesity and lends some support to there being an inherent, genetic component to obesity in these islands. Further indication that there might be a genetic influence on BMI in Samoa, and strong evidence at that, is the fact that there is an estimated heritability (the component of a trait determined by genes and capable of being passed on to children) for BMI of 41 per cent. Something is clearly going on in the population of Samoa and a GWAS is the best way to find out exactly what.
The year 2016 saw the publication of a GWAS focused on high BMIs in Samoa. Studying more than 3,000 Samoans, the research team were able to identify a specific variant of a gene called CREBRF that was very strongly associated with BMI. CREBRF is a gene identified widely across vertebrates and expressed (in other words, read and used by cells to make a protein) in many different tissues, including adipose tissue. Studies suggest CREBRF has a fundamental role to play in the function of our cells; these studies have also implicated the proteins coded for by CREBRF and related genes in the metabolism of fat cells. Overall, the background evidence suggests that this gene could well have a role to play in fat storage and obesity. Those Samoans with a higher BMI were more likely to have the variant of the CREBRF gene, which was also associated with body fat percentage, abdominal circumference and hip circumference. CREBRF seems like a good candidate for a thrifty gene.
Nearly one in two Samoans carry at least one copy of the variant CREBRF gene associated with high BMI in Samoa. In contrast, the variant is unobserved or extremely rare in other populations (one in 18,300 Europeans and one in 2,100 East Asians). That the Samoa variant is rare in other populations is not perhaps surprising. Modelling using a computer tool called Polyphen-2 predicts the impact of genetic mutations on the structure and function of proteins, and this approach suggests that changes to the specific part of the gene impacted in the Samoa study are highly likely to be damaging and are therefore unlikely to spread. To unravel this conundrum, the researchers studied the effects of the genetic variant using what is known as a ‘cell model’. Cell models use established ‘lines’ of specific cells that can be grown in the lab and used to investigate how cells function. The cell model the researchers used was 3T3-L1, a cell line of adipose cells established in 1962 from the embryo tissue of a Swiss albino mouse and widely used to study the cellular basis of fat metabolism. Using this cell model they found that the variant gene decreased energy use and increased fat storage in these cells, which is surely the very definition of thrifty from a fat-cell perspective. Taken together with the history of the Samoan people and the population genetics in Samoa and beyond, the study presents a very convincing evidence-base for the CREBRF variant found in Samoa being a thrifty gene.
In an interesting twist to the original Neel formulation of the thrifty gene hypothesis, as well as promoting fat storage the Samoan variant also seems to protect against type 2 diabetes. Carrying the variant of the gene which causes a 1.3-fold increase in the risk of obesity actually reduces the risk of type 2 diabetes 1.6-fold, which is certainly a counter-intuitive finding given the well-established link between obesity and type 2 diabetes. That the gene provides protection against one of the key diseases associated with the condition it promotes is likely linked to the effect that it has on metabolism at a cellular level and could certainly explain why the gene is still around and at such a high frequency in the Samoan population.
Further evidence for South Pacific genetic thriftiness has come from additional studies of obesity in people from this region. A study of Māori and Pacific people living in New Zealand replicated the findings of the Samoa study, revealing the same genetic variant having the same association with higher BMI and reduced risk of type 2 diabetes.7 What is more, the magnitude of the risk associated with the variant was also similar. So, related people with a similar recent evolutionary history share a rare genetic variant that promotes fat storage. Couple that with a shared heritage of long sea voyages and onboard famines that likely killed off the thin and it feels like the TGH is well and truly over the line. But things are not all they seem.
Problems with the thrifty gene hypothesis
The Samoan study does provide excellent support for the thrifty gene hypothesis, but when it came along in 2016 the TGH was losing ground despite being widely parroted in the media and supporting a number of dietary fads. The problem for the TGH was not in establishing the truth behind the current environment because we quite clearly have better access to calorie-dense food now than we did in the past. I can nip to one of the many supermarkets within a five-minute walk of where I work, buy a kilogram of sugar and wash it down with a wine glass full of molten lard for less than £1. Though not the nicest meal (and eating that much sugar is certainly not recommended), I could, in no time at all and with very little financial outlay, consume around 6,200 calories, or enough to keep me going for a good two and a half days. The problems for the TGH actually came in establishing the feast-and-famine selection environment in human history that would be required to cause the evolution of thrifty genes, and then finding definitive evidence of those genes in our genome. These are not insignificant problems by any means.
A particularly major problem with the TGH is that if there have been a great many famines over the course of human history, and a greater risk of death to leaner individuals lacking the fat stores to make it through, then the force of selection on genes promoting fat storage would have been massive. Massive selection pressures promoting certain genes should push those genes to fixation: the point at which everyone (or virtually everyone) carries the beneficial gene. Such evolutionary change can happen surprisingly rapidly, even with quite modest selection pressures. John Speakman, a major critic of the TGH, has done the mathematics and the resulting numbers are not good news for the hypothesis. The calculation goes like this. First, Speakman assumes that a mutation in some gene linked to fat storage arises and that this mutated version of the gene promotes obesity in some way. Carrying this gene gives an individual only a 0.5 per cent selective advantage over the rest of the population, in other words they are just 0.5 per cent more likely to survive a famine than someone carrying the normal version of the gene (the ‘wild type’ as geneticists would call it). Speakman then proposes a schedule of famine events, such that there is one famine every 150 years. With these conditions in place, none of which seem especially challenging or controversial, a mutation that was present in a single individual would spread to fixation and be found in the entire population after 6,000 selection events (famines) or, in this case, 900,000 years. Granted that extends beyond the period of anatomically modern humans, but it is well within the period we might consider to be our recent evolutionary history. We should, says Speakman, all be obese against such a selection background, but we are not. The variation in obesity and tendency to deposit fat among humans works against the TGH, because a beneficial thrifty gene should have spread to fixation.
Another issue with the TGH arises from the pattern of mortality that we typically find in famines; in other words, who tends to die when famine strikes. The TGH relies on body composition being the prime factor driving mortality, with lean individuals more likely to die than those that are fatter. Providing that the fatter individuals are fatter as a consequence of some genetic mechanism promoting body fat, then a thrifty gene can spread. The problem is that overall there seems to be very little evidence supporting the idea that fatter people are more like to survive famines than thinner people. As Speakman cautions though, the lack of evidence for an effect does not necessarily mean that the effect is lacking and we have no studies comparing BMI and the likelihood of death in a population before and after famine.
In fact the primary factor affecting mortality in famines is not body composition but age. Famines tend to kill the young (10) and the old (rather ungenerously defined as 40+). Selective death affecting the old will have little or no effect on evolution. Deaths that occur after individuals have produced offspring are invisible to natural selection, unless perhaps we are dealing with a social species where older non-breeding individuals have some genetically underpinned and significant part to play in the survival of closely related kin. An example would be grandmothers providing care for grandchildren and promoting their survival and subsequent ability to have offspring (see Chapter 3). At the other end of the age-scale, the deaths of children are also unlikely to play much of a part in the evolution of thrifty genes. Since childhood obesity is a relatively very recent phenomenon, Speakman’s argument is that famine-related mortality cannot have been biased towards leaner children in the past. In essence, all children were lean during early human famines and were equally likely to live or die regardless of whether they carried a thrifty gene. Perhaps most damning of all, where data do exist on the overall mortality imposed by famines, the numbers are not especially high. A study of 190 identified famine events occurring in the UK across 2,000 years indicated that famines occurred every 10 years. However, examination of the likely mortality effect of those famines showed that only one famine every 100 years produced a significant increase in mortality above the baseline mortality for the previous 10 years. The overall conclusion is that famines are not especially common and when they do occur they do not result in particularly high increases in mortality. In summary, famines, the primary selection factor in driving the evolution of putative thrifty genes, do not happen very often, do not have a large effect when they do and the individuals killed by famine in a population are not those that are consistent with the hypothesis.
Speakman explores a number of different selection environments and pressures and through his calculations concludes that ‘If the thrifty gene idea is correct, we should all have inherited advantageous mutations in thrifty genes, and if these mutations cause obesity, as the hypothesis suggests, we should all be obese.’ Of course, you cannot get fat on genes alone; it is still essential to have an environment of calorific plenty in order for that genetic background to be expressed. It could be argued that the fact we are not all obese is not necessarily clinching evidence against the TGH, because we can, at least notionally, control what we eat even if we are surrounded by cakes, crisps, cookies and chocolate. Furthermore, if we look at the UK then we are certainly in a period when the majority of people are overweight, with 62 per cent of adults having a BMI greater than 25 in 2014. It is predicted that as many as one-third of adults in the UK could be obese by 2020 and that proportion has already been comfortably breached in the USA. Nonetheless, Speakman makes a strong case by stating that if selection for thrifty genes were a long-standing feature of our evolution then thrifty genes should be widespread and, crucially, that fatter people should fare better in times of famine.8
It is nearly always difficult to disentangle the different effects of genes and environment on a particular characteristic, and this is also the case if we are looking for thrifty genes that were important in our early evolution. Global obesity statistics combine a great number of intermeshed environmental effects, including recent history, culture, economic development and food availability, that will interact with genes that might be common to all humans, or might be, as we have seen in the South Pacific nations, more or less contained within specific populations. If thrifty genes were essential in our early survival then we might be able to see their effect in populations of modern humans that live in environments and have lifestyles more in keeping with our evolutionary past, or at the very least have lifestyles that are not ‘modern’, characterised by readily available high-sugar, fatty, processed foods. In fact, the effect of a thrifty gene should be easy to spot. The crucial advantage of having such a gene is that between periods of famine, individuals carrying the gene can get fat, building up stores that allow them to survive the next food scarcity. So, the logic goes, we should see people getting fat between famines. The reality is that we don’t.
The year 1816 became known as ‘the year without a summer’. In April 1815 Mount Tambora, a volcano in Indonesia, erupted or, to put it more accurately according to contemporary reports, exploded. It was the largest observed eruption in recorded history and the sound of the mountain tearing itself asunder was heard a staggering 2,600km (over 1,600 miles) away. Ash filled the skies and resulted in a global climatic phenomenon known as a volcanic winter. Particles of ash and clouds of sulphur dioxide gas spewed into the atmosphere. Reacting with water in the atmosphere, the sulphur dioxide formed sulphate aerosols that together with the ash particles obscured the sun, reduced the heat reaching the surface of the Earth and caused temperatures to decrease across the world by 0.4–0.7°C. The drop, although a good deal less than a degree, was sufficient to provoke severe effects on the climate in many places. In North America, for example, the atmospheric pollutants from the eruption caused a persistent fog to form, and the lowered temperatures resulted in unseasonal frosts in May and June of 1816. Further frosts were reported in July and August. The effects on agricultural production that year were devastating and the consequent crop failures have been called ‘the last great subsistence crisis in the western world’, promoting near-famine conditions.9
If we treat 1816 as a famine event then, if the argument for thrifty genes holds up, we would expect the population to lay down fat from that point on in thrifty preparation. However, historical data from the US reveal that in the late 1890s, a full 80 years later, the level of obesity in the USA was still only 3 per cent,10 well below the level we see now and certainly not consistent with thrifty genes squirrelling away fat as an adaptive response to famine. It seems much more consistent with a small number of people having access to plentiful food while the majority of people ate a simpler, more calorie-controlled diet without the fatty processed food and abundant sugar that we find in more recent times.
As well as looking back through time, if we want to remove the influence of the modern world on BMI we can look to contemporary populations that are relatively unaffected by the modern world. Hunter-gatherer societies are now rather unusual and tend to be found in the less accessible and more remote parts of the world such as the San people of the Kalahari, the Hadza of Tanzania and the Jarawas of the Andaman Islands in the Indian Ocean. Collecting food predominantly by foraging, such people are effectively living a similar lifestyle to our pre-agricultural ancestors 12,000 years or more ago. It is important to say here that these people are not Stone Age humans. They are modern humans with their own recent history of evolution. My point here is merely that their lifestyle is not what we would call modern. As well as hunter-gatherers, there are many more communities around the world that practise subsistence agriculture, relying purely on what they themselves produce. Hunter-gatherers and more particularly people relying on subsistence agriculture might expect to exhibit strong evidence for thrifty genes, since food shortage and interruptions in availability cannot be smoothed out by reliance on wider society or alternative means; if you don’t find it or grow it, you don’t eat it. However, studies on such communities during non-famine conditions (when thrifty genes should be promoting fat deposition) consistently find BMI from 17.5 to 21, which is at the very lean end of normal.
The thrifty late hypothesis
That existing hunter-gatherers are lean, well fed and do not seem to show any sign of thrifty genes, at least from the studies that have been carried out, is interesting because it points us towards an elaboration of the TGH that could help us to get around some of the issues that Speakman raises. Around 12,000 years ago the development of agriculture saw a major shift in our feeding behaviour. No longer were we solely tied to the vagaries of the environment and natural food availability; but equally, the development of more centralised, larger agricultural societies made us potentially more prone to famines. The development of agriculture saw more mouths to feed and with all our eggs in fewer baskets, both literally and metaphorically, the impact of droughts, crop failures and animal diseases might be felt ever more keenly by a population that was now too large and lacking in knowledge to revert suddenly to a foraging existence. To account for this seismic shift in the lifestyle of most modern humans, the ‘thrifty late hypothesis’ suggests that intense pressure for thrifty, fat-storing genes would only have occurred in the last 12,000 years or so. ‘Thrifty late’ explains why hunter-gatherer populations do not become fat between famine events, because they have not been under the same agriculture-induced selection pressures as other populations. By having a shorter period of more intense selection, the thrifty late hypothesis also seems to go a long way in explaining why we see such a variance in fat deposition between people in response to modern diets (some people gain weight more readily than others) and why there are still a large number of unfixed variants of genes related to obesity. Another supporting element was inserted into the adaptive thrifty gene hypothesis by British nutritional scientist Andrew Prentice in 2008. Prentice agrees that famines are likely only to have been a significant factor in relatively recent, agricultural human history but argues that it is reduced fertility rather than increased mortality that is their primary effect.11
Obesity: are there benefits?
If the TGH, or amendments like ‘thrifty late’, are correct then we really should be able to detect evidence that at least some of the genes associated with obesity are providing an adaptive advantage. After all, it is the counter-intuitive fact that obesity is clearly bad, and yet equally clearly obesity has a genetic component, that is really at the heart of the TGH. We can argue about selection pressures, the magnitude of famines and the impact of agriculture all we like, but the TGH really does demand that there are some genes that cause obesity that also benefit those who carry them. We can investigate that by looking for signals of selection in the human genome at those places (termed ‘genetic loci’) that are associated with obesity.
The 1,000 Genomes Project is the ideal tool for probing the genome (our DNA) to find out more about its selection history. The project was launched in 2008, and by 2012 1,092 human genomes had been sequenced (their long sequences of DNA bases determined) and made freely available for scientists to study. Using these genome sequences, Guanlin Wang worked with Speakman to look at the 115 genetic loci that were known in 2016 to be associated with obesity. By examining the genome data in detail, it was possible for them to determine that only nine of these loci showed evidence of the positive selection that would be a signature of the thrifty hypothesis. What is more, of those nine loci, five (so more than half) actually involved positive selection for leanness rather than obesity.12
So, by the time the Samoan study came along, the TGH was not faring well; as a human-wide global explanation for the recent rise in obesity it arguably still isn’t. The prevalence of obesity in South Pacific nations and the identification of a thrifty gene associated with obesity and protective against diabetes in those populations is excellent evidence that the TGH has some merit, but looking across the wider human population the evidence is, like our hunter-gatherer ancestors, rather lean. We cannot, it seems, firmly pin the obesity epidemic on a mismatch between our thrifty past and our calorie-rich present. A much simpler explanation is that like other mammals we have the ability to convert excess fat into adipose tissues, and these days it is simply far too easy for many of us to eat far too much.
If not thrifty then drifty?
Despite some clear issues with the thrifty gene hypothesis, we are still left with an intriguing genetic association with obesity and a relatively high heritability of obesity, both of which suggest that something interesting is going on. The prevalent counter-hypothesis has become known as the ‘drifty gene hypothesis’ (DGH). Like the TGH, DGH is an evolutionary explanation, but unlike the TGH it is not an adaptive hypothesis. To understand how the DGH might explain obesity, we first need to understand a process called genetic drift.
Evolution, remember, is the change in gene frequency over time and in general we are concerned with the evolution of adaptive traits through natural selection. We may also be interested in other adaptive selection mechanisms, including sexual selection (the peacock’s tail, for example) or kin selection (honey bee workers committing suicide for the good of the colony). However, selection for and against favourable and unfavourable traits is not the only way that genes change in frequency over time.
Consider a bag full of ten marshmallows, five of them white and five of them pink. Now imagine that these marshmallows can reproduce, their offspring are exactly the same colour as them and there is no real difference between the two colours in terms of the likelihood of surviving and having offspring. To start filling a second bag with our ‘next generation’ we simply pick out, without looking, a marshmallow. If it is pink then we put a new pink marshmallow into the ‘next generation’ bag and we put the ‘parent’ marshmallow back in the original bag. Then we pick again, without looking, and put our second ‘next generation’ marshmallow into the second bag, returning the parent to the first bag as before. This carries on until we have ten marshmallows in each bag.
Assuming we are unbiased in our selection, it feels like we should end up with five white and five pink marshmallows in the next generation because we are picking from an evenly proportioned population. In reality though, we would only get five white and five pink in our next generation an average of one in four times. By sheer chance we could actually end up with a new population that was entirely white (one out of every 1,024 times we did it) or entirely pink (the same chance as entirely white). Odds of one in just over 1,000 indicate events that could happen but are pretty unlikely, so let’s say we get some intermediate value, like seven white and three pink (120 chances out of 1,024). White would have increased in frequency (evolution) but without any adaptive advantage. The next time we sample (to create a third generation) we could, again just by chance, end up with an even more biased population. Indeed, at some point the proportion of white marshmallows might have changed towards the point where it has gone to fixation and pink no longer exists. This non-adaptive change in gene frequency occurring because of the random sampling of individuals is known as genetic drift.
The possible role of genetic drift in obesity was presented by John Speakman at the 2007 Obesity Society meeting in New Orleans as part of the presidential debate, with Andrew Prentice. Prentice, you’ll recall, is the nutritionist we met earlier, who proposed that reduced fertility through famine was more important than increased mortality through famine in the evolution of thrifty genes. Speakman’s argument, which he has developed considerably since that first outing, combines models of how we control our weight and fat storage with an ecological perspective of human evolution. Central to the argument is the threat of predation.
There are a number of theoretical approaches, or models, that have been proposed to understand how our bodies control ‘fatness’. The three basic models are the set point, the settling point and the dual intervention point, and there is plenty of debate and experimental work developing and investigating each of them.
1. The set point model proposes that there is some, you guessed it, ‘set point’ of fatness and that the body is trying to regulate fat stores up or down to reach this ideal. It works well as a model to explain weight gain and weight loss in rodents, and some evidence supports it in humans, but the obesity epidemic suggests that perhaps a set point model isn’t the full picture. If we have a set point for fatness then why are so many of us so far away from it, and why is the set point seemingly so variable across even reasonably closely related populations?
2. The settling point model proposes that the set point can move depending on nutritional conditions and the environment. So, in an environment with plenty of food and relatively little energy expenditure through physical exercise our body weight ‘settles’ to a new higher point. The settling point model seems more consistent, initially, with what we see in human populations, but observations of people on diets and taking part in controlled feeding experiments are actually more consistent with the set point model. With aspects of both models supported or refuted by studies on mice and humans, neither seems a satisfactory explanation for fatness in humans.
3. The dual intervention point model tries to account for the seemingly conflicting observations that there is evidence supporting a set point but that environmental factors also seem to be able to alter body fatness. The dual intervention model proposes that as well as an upper limit controlling our maximum weight, there is also a lower limit preventing us from becoming dangerously thin. The lower limit, or intervention point, is the weight at which our body intervenes to prevent starvation and activates fat storage. In a sense, this lower limit is a kind of ‘thrifty’ limit, a saving regime imposed when times are tough and the belt has, literally, been tightened too far. The upper limit is far more interesting. The upper limit is proposed to be linked to our ability to avoid predators.
The logic behind the predation-imposed upper limit, or intervention point, is simple. If you are fat you are less able to escape the claws and teeth of animals pursuing you for food. It is frankly very hard to argue against the logic here. Given just how bad being eaten is in terms of survival, future reproduction and lifetime fitness, it certainly makes sense that avoiding it by remaining a lean, mean, running-away machine could have been a vital component of evolutionary history. I have a reasonable amount of personal experience in examining the insides of various African antelopes and a very striking thing about them is the almost complete absence of fat. These prey animals are the very definition of lean and while that does put them at risk during periods of food scarcity (such as droughts), that is a long-term, ‘cross that bridge when they get to it’ concern (see Chapter 10 for more on this). In the short term, that leanness translates into a remarkable ability to run away that is pretty useful when you live next to lions and leopards.
Speakman uses similar evolutionary thinking combined with ideas of genetic drift to explain why the upper intervention point has ‘drifted’ upwards over time in modern humans. Around 2–4 million years ago our hominin ancestors, little more than small primates living in open savannah habitats, were fair game for a suite of predators no longer around today. Those early hominins were the antelopes of their day, prey animals that hid, cowered and ran away. Sabre-toothed cats belonging to the wonderfully named genus Dinofelis (‘terrible cat’) and the fearsome Megantereon cultridens roamed these plains, and evidence from gnawing marks on bones suggests early hominins were most certainly on their menu. Against such an environmental background of teeth and claws there would be a very strong selection pressure on early hominins to remain lean and to have a tightly constrained upper intervention point.13 Being fat would in all likelihood mean being dead. But then, along came Homo erectus…
Larger in body size, Homo erectus made use of fire and tools and was a more sophisticated hominin in every way. We also know that H. erectus formed larger social groups than previous species. It is thought by some that this social behaviour, which would have involved evolutionary changes to the brain, may well have been an adaptation against predation, although sociality could also increase hunting success and allow for bigger prey to be taken. A bit later, around 1–2 million years ago, we know from the fossil record that a number of large-bodied carnivores went extinct in east Africa. Suddenly, predation wasn’t such a big concern for members of the genus Homo. Suddenly, you didn’t need to be quite so thin, quite so lean and quite so ready to run. Suddenly, if you had a tendency to put on a bit of timber, it might not matter and you might make it to an age when you could have some little Homo-lings, who might inherit your tendencies towards a more ample waistline.
With a relaxation of selection pressure on that upper intervention point, mutation and genetic drift could have become very important factors in the evolution of obesity. Adding to their contribution is the fact that early populations were small. Small populations tend to exacerbate the sampling effects we saw earlier with marshmallows, which can be ironed out to some extent when you have more individuals to sample. Overall, Speakman contends that, unconstrained by predation, the upper intervention point has drifted around and has resulted in the large diversity of upper intervention points (i.e. the large diversity in obesity) that we find in modern-day humans.
Neither the thrifty gene nor the drifty gene hypotheses have yet gained universal approval and a number of scientists continue to develop different arguments and counter-arguments, amendments and counter-amendments, tweaks and refinements to both hypotheses.11,14 Human populations are highly variable in all kinds of ways and the discovery of thrifty genes in some populations but not others might simply be a reflection of that diversity. It may also be that drift has been more important in some populations than in others. Overall though, the simple idea that we are fat because we are storing up for a forthcoming famine, as seductive as it undoubtedly is, doesn’t currently stand up to full scrutiny across all populations. Even if the drifty gene hypothesis is correct, at least for some populations, then we still get to blame evolution for getting fat, albeit in a far less satisfying way. Rather than relatively recent ancestors being metabolically good boy scouts, we have far more distant ancestors drifting towards obesity because they didn’t need to run as much. Ancestral laziness, not thriftiness, may be the reason our clothes are shrinking.
An evolution-inspired solution?
If we stand back from the fine print and accept that evolution has had a role in determining how we store fat (and that is really beyond dispute), then the big question is, can we counter the current global obesity crisis simply by adopting the diet of our ancestors? The Palaeolithic diet, also known as the paleo diet, the Stone Age diet or the caveman diet, is the idea that shunning the foods of the modern world and eating only those foods that were available to our ancestors will align our diet with our evolved bodies and prevent us from getting fat.15 Whether this is a good plan or not rests on several things: knowing what our ‘caveperson’ diet actually was; whether adopting that diet will help us to lose weight, the goal of most people choosing a diet plan; whether the diet is healthy; and whether the diet is sustainable in the long-term so that people can stick to it and keep the weight off.
Determining what our ancestors ate requires us to examine a wealth of different types of archaeological evidence. We know from bones showing tool marks and scorching from cooking for example that our ancestors ate meat. Hunting of many species is also depicted in cave art. We also know that they enjoyed seafood when they could, with great piles of shells stacked up on beaches where our ancestors feasted. Likewise, fruits, berries and nuts were eaten, eggs were consumed where they could be collected and even honey would have appeared in the diet.16 Given their availability and the fact that they are eaten widely across the world today, it also seems likely that insects would have made at least the odd starter here and there.
We can delve deeper into ancient nutrition by analysing teeth and bones using a technique called stable isotope analysis.17 This technique relies on the fact that nitrogen and carbon both exist in different forms, called isotopes. The nucleus of an atom of carbon always has six positively charged protons and this ‘atomic number’ (6) is what defines the atom as carbon. Atomic nuclei also have neutrons, sub-atomic particles that carry no charge and have the same mass as a proton. The number of neutrons in the nucleus of carbon can vary and this variation results in heavier and lighter forms, or isotopes, of carbon. The same applies to other elements like nitrogen, which has an atomic number of 7 (seven protons) but can have six or more neutrons, with the most common number being seven.
Carbon actually has 15 isotopes but only two of them, carbon 12 (six protons and six neutrons) and carbon 13 (six protons and seven neutrons) are stable for any meaningful length of time. Carbon 14, used in radiocarbon dating, decays slowly into nitrogen with half of it decaying in 5,700 years (its half-life). It is this decay that allows us to use it as a kind of atomic clock to date organic material. Nitrogen has 16 isotopes but again only two are stable, nitrogen 14 and the less common nitrogen 15. The ratio of the different stable isotopes of carbon and nitrogen present in a sample of bone produce an isotopic signature that enable us to make inferences about the diet that helped to form that bone. Different plants, for example, produce different ratios of carbon isotopes, with diets rich in subtropical grasses producing a ratio more skewed to carbon 13 than diets richer in products derived from trees and temperate-zone plants. Nitrogen isotope ratios can reveal the extent of meat-eating as well as indicating whether a diet had a large component of seafood.
Archaeological and isotopic evidence allow us to piece together a reasonable picture of the sort of foods that were consumed in different places and at different times. Notable components of a modern diet missing from the Palaeolithic diet include grains in any great number and dairy foods, both of which are products of the later development of agriculture (see Chapter 3). Of course, the paleo diet also lacked the processed foods so familiar to us today, so no cornflakes, bread, sausage rolls, ready meals, cakes, biscuits, sweets or fizzy drinks.
To go from the list of food types and some general overview of proportions of different broad types of food (plants versus meat, seafood versus ‘land-food’) that we can derive from archaeological evidence to an actual day-to-day schedule that we can adopt in the modern world is something of a leap. Cave paintings might tell us that our ancestors ate meat and collected honey, but they are pretty light on the sort of detail we need for a diet plan. To help us make that leap we can look to existing hunter-gatherer societies that are largely isolated from the influence of a modern diet for more clues. This approach, combining archaeological evidence with contemporary observations, underpins arguably the most influential book in the area of caveperson nutrition, The Paleo Diet, published in 2002 and written by nutritionist and exercise physiologist Loren Cordain.
Cordain didn’t come up with the idea of the paleo diet, but he was perhaps the most important force in propelling the concept into the mainstream. In fact, the nutritional principles underlying the book mostly derived from a paper published in 1985 by Stanley Boyd Eaton and Melvin Konner that laid out the foundations for the paleo diet. Their paper, entitled ‘Paleolithic Nutrition’, appeared in the prestigious medical journal New England Journal of Medicine and in it Eaton and Konner explore archaeological dietary evidence, lay out their findings of a review of contemporary hunter-gatherer nutrition and identify the mismatch between the diet of the past and modern human nutrition with its links to obesity. Their final line sets the stage for what would become a very profitable dietary trend. As they put it, ‘The diet of our remote ancestors may be a reference standard for modern human nutrition and a model for defense against certain “diseases of affluence”.’18
Eaton and Konner developed this idea further, publishing the book The Paleolithic Prescription in 1988 with Marjorie Shostak, an American anthropologist who conducted fieldwork with the !Kung San bush people, a group of hunter-gatherers living in the Kalahari desert. This book greatly promoted the evolutionary ‘discordance hypothesis’ (as they called the mismatch between our evolved heritage and the modern world), and brought the concept of evolutionary mismatch into the mainstream. It was against this background that Cordain’s book was published and the paleo diet has been firmly in the public eye ever since. Indeed, a quick search on Amazon reveals more than 10,000 books on the topic and a Google search yields more than 94 million hits. Despite competition from the carnivore diet, the keto diet, the Dubrow diet, the noom diet and even the biblically inspired shepherd diet, the paleo diet is still very much alive and thriving.
There is no official paleo diet. It is more of a dietary philosophy, with different authors proposing different variations in how to interpret the basic tenet and providing different daily suggestions and recipes. Whether this is to provide some clear water in a very crowded marketplace is not for me to say, but taken overall, some clear paleo-diet trends emerge. Paleo-diet advocates reject dairy, grains, sugar, legumes, processed oils and salt, since such foods would not have been available in Palaeolithic times, while embracing fruit, vegetables, nuts, meat and fish. Cordain recommended that 55 per cent of daily calories come from an equal mix of lean meat and fish and the remainder from an equal mix of fruit, vegetables, nuts and seeds. Other authors tweak these figures but most remain in that ball park. Basically, it seems to come down to looking at the food in front of you and asking, if it was 20,000 years ago and you were standing naked on the savannah with a spear, or on the shoreline with a pointed stick, would this food have been possible? A pepperoni pizza washed down with a glass of milk scores a big no, a piece of meat with some vegetables scores a yes. Of course, in answering this question we do need to overlook the fact that most of the vegetables we identify today have evolved from anthropogenic artificial selection in recent times and bear little resemblance to ‘paleo veg’.
The really big question though is not whether your fish and chips counts as ‘paleo’ (as long as you don’t fry it in oil or coat your fish in flour) but whether the diet really works. In 2017 Jonathan Obert and colleagues published a review of the scientific and medical literature associated with four weight strategies, one of which was the paleo diet. If you are looking to lose weight rapidly then their results are encouraging. At least nine trials have shown short-term benefits of the diet that include weight loss and reduction in waist circumference. This is positive news for paleo advocates, but Obert et al. are critical, stating that trials were short in duration and ‘underpowered’, which is essentially saying that there were too few people being tested.
A ‘gold standard’ procedure for investigating the effects of a particular intervention or treatment is to use a randomised control trial (RCT). RCTs allocate subjects randomly to either a treatment group or a control group, with comparisons between the outcomes of the two groups providing evidence on the efficacy or otherwise of the treatment or intervention. An RCT of the paleo diet was undertaken in Sweden, the results of which were published in 2014. Seventy post-menopausal women with BMIs in excess of 27 were allocated to either a paleo diet (35 women, average BMI 32.7) or to a Nordic Nutrition Recommendations (NNR) diet (35 women, average BMI 32.6). The paleo diet was such that 30 per cent of calories consumed came from protein, 40 per cent from fat and 30 per cent from carbohydrates, and it was based on the usual suspects: lean meat, fish, eggs, vegetables, fruits, berries and nuts. Additionally, subjects were allowed avocado (a fat source), and rapeseed and olive oil to be used in food preparation and dressing. Neither of those oils, by the way, pass the ‘standing naked on the savannah’ test. As would be expected, dairy products, cereals, added salt, refined fats and sugar were all excluded. The NNR diet aimed for a daily calorie breakdown of 15 per cent from protein, 25–30 per cent from fat and 55–60 per cent from carbohydrates, with an emphasis on low-fat dairy products and high-fibre fruit and vegetables. Overall, both groups had a calorie intake that averaged 2,000kcal a day on the diet compared to around 2,300kcal before, so by no means a crash diet regime. In fact, the trial lasted for a full 24 months and inevitably involved some pretty big lifestyle shifts, especially for those on the paleo diet. To support them, and to ensure that they stuck to the diets and provided useful data, the subjects were given ample back-up including recipes, cooking classes and group meetings.19 It is always worth remembering that these sorts of trials rarely take place in a clinical setting and so a degree of trust is required; if a paleo dieter sneaked a quick burger and milkshake, the researchers would never know.
After six months, both groups had lost body weight, and BMI and waist circumference were significantly lower than when they started. The paleo dieters lost an average of 6.5kg (14.3lb) of body fat in six months, which is more than double the 2.6kg (5.7lb) lost by the NNR dieters. That is an impressive result: a stone of weight in old money. Further losses occurred over the next six months but interestingly both groups of dieters hit a plateau in terms of body weight and the other measures after twelve months. That six-month lead for the paleo dieters didn’t last and by 24 months both groups were level pegging.
Despite losing its early advantage, the paleo diet clearly ‘works’ in terms of losing weight, and works well. As well as overall weight loss, rapid early losses are beneficial because when people can see and measure the difference it tends to keep them motivated. Lest we get too starry-eyed and atavistic here though, we should consider the fact that in the long-term (two years) the paleo diet is no better for weight loss than a ‘normal’ calorie-restricted diet. So, in terms of losing weight on a long-term plan, the components of the diet (twenty-first century versus Stone Age) are less important than the simple measure of energy surplus. What the study confirms is the boring and predictable fact that if the ‘energy in’ is less than the ‘energy expended’ you will lose weight. The conclusion is that, from a weight-loss perspective, the paleo diet works simply because by following any of the many variants of it you are restricting your calorie intake. There is nothing ‘magical’ about the components of the diet, and if you were to eat double paleo-portions every day you would gain weight just as surely as if you were ‘going large’ down at the burger joint.
Worryingly for those thinking of going Stone Age, a number of potentially serious health implications of the diet have been identified, at least in some studies. Participants in a study of the effectiveness of the paleo diet in Australia reported significantly more episodes of diarrhoea on the diet than those following a conventional calorie-restricted regime, and 69 per cent of participants reported an increase in their grocery bills.20 Interestingly, and certainly not to be discounted, 43 per cent of participants reported a belief that the diet was not healthy (perhaps greatly swayed by their regular and doubtlessly unpleasant toilet visits). One health risk that is commonly associated with the paleo diet is reduced bone density leading to osteoporosis, a condition especially associated with the elderly where weakened bones commonly lead to fractures. Dairy is a major source of calcium in our diet and by excluding dairy the paleo diet tends to be low in this vital bone-forming mineral (more on this in Chapter 3). Long-term studies of the effect on bone density of following the paleo diet are lacking, but paleo-diet advocates are quick to point out that calcium is present in foods other than dairy and that magnesium and vitamin D are also required for bone formation. All this is true, but when sources of magnesium are flax seeds and artichoke hearts, and calcium source examples include bone broth, oysters and figs, I can’t help but question the day-to-day sustainability of such a diet for most people.
Mismatches and our evolutionary heritage
The mismatch between our twenty-first-century diet and the simpler diets of our ancestors is the best known of the evolutionary mismatches being blamed for our current woes, but as we’ve seen, the famine-adaptation thrifty gene hypothesis is not widely supported by the evidence, at least across the human population overall. The drifty gene hypothesis is gaining ground and is an evolutionary explanation, but it is a more complex and less satisfying formulation. Drifty genes is a non-adaptive hypothesis that takes us far deeper into our ancestry and involves more wide-ranging ecological factors than we might have expected. Whichever hypothesis wins out, and in reality we might be looking at a complex mix of both and others we have yet to devise, it is clear that the heritage of our evolutionary past has laid the ground for our present and future, and that the calorie-dense, predator- and famine-free world in which we live conspires to mean that many of us will struggle constantly to keep to a healthy weight. The idea that we can delve in to our pre-agricultural roots and devise a diet more in keeping with those simpler times in which we evolved is appealing and has generated a considerable industry to support it. As we have seen, the approach can work for weight loss, but it is a fiddly and expensive diet that likely provides no advantage over conventional calorie-controlled diets. It may also have some negative effect on our skeletons, but of course this could be reduced by taking supplements or seeking out specific nutrient-rich foods.
The crux of the paleo diet is that it is predicated on a single and simple fact: that we modern humans are identical to pre-agricultural humans. As an intervention and a philosophy, the paleo diet assumes that we have not evolved in the last 12,000 years, at least in terms of our ability to handle different foods. Given just how big an effect the development of agriculture has had on our environment, this seems on the face of it to be a pretty bold position to take. As we’ll see in the next chapter, the development of agriculture and the availability of grains and dairy in fact led to considerable evolution and, later on, to yet more important mismatches.