Very low carb, ketogenic diets have some astonishing effects. I'll describe exactly what a ketogenic diet is and what some of those effects are below, but what's interesting is that simply by changing the amounts of carbohydrates, proteins, and fats you eat, it turns out you can cause profoundly different physical responses in the brain and the rest of the body. Given the range of powerful results that various studies have yielded, I find it inexplicable that more people aren't burning with curiosity and desire to understand ketogenic diets better. That some people actually oppose them is even more perplexing.
The oddest response of all, though, to me, is from those who can see the therapeutic value of ketogenic diets, but rather than let the implications of this new knowledge percolate through their prior beliefs and perhaps even update some of them, they instead try to reconcile the knowledge that a ketogenic diet can confer extraordinary health benefits with the idea that a healthy diet must necessarily be high in carbohydrates, low in fats, low in animal products, and high in plants to be healthy.
These ideas — despite lack of strong supporting evidence — are so entrenched, that some have gone to great lengths to construct dietary patterns that achieve ketosis while maximising plant intake, minimising meat and animal fat, and, yes, in some cases, even centering ketosis around a high carb diet . While these goals can be met, and the addition of ketosis to these paradigms is likely to add benefit, in my estimation their addition to the ketogenic paradigm serves to weaken the potential benefits, not strengthen them.
What is a ketogenic diet then? A ketogenic diet is any diet that causes you to generate a high level of ketone bodies , putting you in a state called ketosis . The word ketone comes from the same root as "acetone" which is one of the three ketone bodies . The other two are called acetoacetate (AcAc), and beta-hydroxybutyrate (BOHB). Generation of ketone bodies is called "ketogenesis" . I like to think of ketone bodies as a transport form of fat. In particular, unlike many fats, they can easily enter the brain. The brain can use ketone bodies for many things, but it's especially useful for energy. The brain needs a lot of energy, so the richer the blood is with ketone bodies, the more the brain will take up . Most circulating ketone bodies are made in the liver. The liver is a master metabolic organ that's often underappreciated. Based on signals from all over the body about what fuels are available, it regulates what kind and how much fuel to supply. When your food is providing mostly fat for energy, or when you are mostly getting energy from your own body fat, the liver provides a steady supply of both glucose, the form of sugar in the blood, and ketone bodies. These are both special fuels for a few tissues like the brain that use those fuels better than fat. Ketone bodies are made from fat. Inside most cells fat can be made into energy in a process called fatty acid oxidation (FAO). When FAO is at maximum capacity in liver cells, if there's still more fat available, it's made into ketone bodies. Maxing out FAO depends largely on two things: the supply of fat and the supply of glucose. The first point is kind of obvious. The higher the supply of fat, the more likely there will be more than can be oxidised at once even when FAO is going as fast as it can. The second point is a little less obvious. When glucose is low, it's the liver's job to make that, too. It turns out that the process of making glucose — gluconeogenesis (GNG), can sometimes limit FAO because it uses up some of the same chemical resources. Specifically, a substance called oxaloacetate (OAA) is required for both FAO and GNG. Oxaloacetate is commonly derived from glucose, although it can also be made from other sources. So when there is a lot of glucose around, OAA levels tend to be higher, which speeds up FAO, leaving less fat for making ketones, whereas when glucose is low, OAA is low and FAO is less likely to leave leftovers. Low glucose helps ketogenesis in another way, too. When glucose is lower, then insulin is normally lower, too. Lower insulin allows more flow of fat out of fat tissue, making more available to the liver. So the system works together harmoniously. In a normal, low glucose situation, insulin is low, allowing fat to flow freely through the blood to all the tissues that can use it efficiently for energy. This includes the liver. Meanwhile the liver can generate a slow, steady supply of glucose and ketone bodies for the few specialised tissues that don't use fat well. On the other hand, when glucose is more abundant, it's a good time to use it for energy. Even though the concept of preference doesn't really apply to body parts the way it does to minds, we say the body "prefers" to use the glucose up if it's there, rather than leave it accumulating. This "preference" happens through a coordination of many distributed effects. Some of the effects are triggered by insulin. Many cells can take in glucose faster when there is insulin around, so the body responds to a rise in glucose with a rise in insulin. Adipose tissue, or body fat, which was supplying the liver with fat through the bloodstream responds to higher insulin by reducing the release of fat. That means that there is less fat available to liver cells, and this slows down ketosis. In this way, a healthy body responds with agility to the fuel situation of the moment, by seamlessly changing modes according to what is available: "glucose" mode when there is a supply of that, and otherwise ketogenic mode. Much of the consternation over ketogenic diets comes from uneasiness about the less familiar ketogenic mode, especially when it is prolonged. However, when these misgivings have been articulated into specific concerns, they haven't been supported empirically. In some periods of our evolutionary past, ketosis was likely frequent, arguably even the more frequent state. Insofar as this is true, it seems that these fears would be unfounded. If modern civilisation is happily humming along in glucose mode, then why would anyone consider using the ketogenic mode that we know less about? The reason is that in many cases it appears to have advantages over glucose mode.
Ketogenic diets suffer terrible misunderstanding for having the word "diet" in them. Diet is a bad word; it has all the wrong implications. For one thing, when people hear the word diet, they often assume it's a fat loss regime. To add to the confusion, ketogenic diets typically do result in fat loss! That is, they result in fat loss for those who are overfat to begin with, and for that purpose, it's as good an intervention as they come . Weight loss is so badly misunderstood in the current world that most people confuse cause and effect. We are taught that weight is the result of a delicate balance between voluntary energy intake and voluntary energy expenditure. As it turns out, eating less and moving more doesn't fix obesity, because obesity is the result of biochemical energy regulation signals telling the body to store more fat and not use it for energy, regardless of how much energy is coming in and how much the body could technically spare. If your diet does not affect these signals the right way, your fat loss efforts will either not work at all or will work only temporarily. Since typical fat loss diets do not address energy regulation signals, they are temporary by design, and they're inherently unsustainable. They can't be sustained because as long as the signals are insisting on fat storage, the dieter will be fighting against stronger and stronger urges to eat more and move less, in the form of ravenous hunger and exhaustion. The more diligently he applies his will, the more damaging it is to his body. From the body's perspective he is suffering malnutrition and starvation even if he is a hundred pounds overweight. So when someone well versed in the workings (or non-workings, such as it is) of typical "calorie control" diets, makes an educated assumption that ketogenic diets can't work any better than other diets, it's because they recognise that fat loss diets they know about are stressful and unsustainable and assume this applies to ketogenic diets, too. If you remember from the description of a ketogenic diet above, one of the signals it affects is insulin. Sufficiently low insulin is critical for the release of fat from adipose tissue. Ketogenic diets cause excess fat to be used up as a side effect. When your bloodstream is flowing with fatty acids feeding all your cells, you do not feel hungry and you do not feel tired, because unlike with energy restriction diets, you have access to energy. So it's perfectly sustainable. But it gets even better! The amount of fat that is released at a given insulin level is proportional to how much fat you have . That means that the leaner you get, the more slowly fat will come off your body, and the more hunger you will have for just enough food to offset what your fat stores can't supply. Think about the implications of this! A ketogenic diet will not cause you to lose fat indefinitely, but rather to lose fat until your fat stores are too low to supply you with more energy than you are taking in. And all of this is communicated through signals culminating in appropriate hunger. If there is nothing else interfering with these energy regulation signals then a ketogenic diet is a complete solution to obesity that requires no willpower or fight against your own body. For many that's all that was needed. For some, that's a big "if".
The tip of the iceberg Not only do most diets have abysmal success rates for fat loss, they're also mostly entirely useless for anything else. In the face of a severe medical condition, typical dietary therapies have little to no power. If you have a serious medical condition and someone tells you they know of a diet that will have a real impact on it, you should be highly skeptical. Most dietary therapies, when tested in clinical trials, turn out to have very little effectiveness. Ketogenic diets stand out in stark contrast to this, because they have a long established clinical history in one particular area: epilepsy. Ketogenic diets are at least as effective as the best anti-epileptic drugs. Around 15% of epilepsy patients who are put on a ketogenic diet by their neurologist become completely seizure free. About a third have a 90% reduction in seizures, and a third have better than a 50% reduction in seizures . That's state-of-the-art treatment for that disease. This may sound quite niche. Epilepsy is not rare, exactly, but nor is it very common. The World Health Organization estimates that somewhere between 0.4 and 1% of people in the world have active epilepsy . In the grand scheme of medical knowledge, the fact that a ketogenic diet can put epilepsy into remission, may not sound very interesting unless you or someone you love is affected, because it doesn't sound generally applicable. But this is short-sighted. There are neurological benefits of a ketogenic diet that are just beginning to be appreciated, in part because we don't know exactly why it treats epilepsy so effectively. In the process of trying to determine which of the many effects a ketogenic diet has on the brain and the rest of the body are responsible for its therapeutic value, we have come to learn of multiple simultaneous mechanisms, several of which have been proposed to be the contender . But each discovered mechanism suggests wider application. As more and more conditions appear to be potentially positively affected by one or more of the mechanisms of a ketogenic diet, the more it appears that epilepsy is the tip of a gargantuan neurological therapy iceberg. It would be beyond the scope of this book to give a detailed review of the neurological applications now being studied for ketogenic diets, but some of the proposed areas of study include Alzheimer's disease, Amyotrophic lateral sclerosis (that is, ALS, or "Lou Gehrig's disease"), Multiple sclerosis, Parkinson's disease, and traumatic brain injury. In general, the state of ketosis can be said to be "neuroprotective" . We are taught not to think of it this way, but psychiatric disorders are also neurological, and there is preliminary evidence that ketogenic diets can improve those as well . Given that growing brains, both in the uterus and in breastfed babies make extensive use of ketone bodies , it really oughtn't be surprising that a brain in need of repair does particularly well in a ketogenic state. We will come back to the role of ketosis in normal, already healthy brains in a later chapter, but for now, what is important to know is that not only is ketosis not harmful to the brain, it's apparently uniquely beneficial.
The other set of conditions that appear to respond well to ketogenic diets are those associated with what many call "insulin resistance". I don't like that name, because insulin resistance itself isn't always bad. In fact, insulin resistance can be a normal, healthy response. Muscle cells need a constant source of energy, which they get primarily from glucose or fat. Glucose and fat from the bloodstream are escorted into the cells at entry points called transporters. Some of these transporters depend on insulin. Insulin is a hormone and works through receptors, which can be thought of a bit like locked copies of programs that run when they are activated with the right key. When a receptor is "unlocked", it activates chemical pathways. Insulin receptors on muscle cells activate pathways that facilitate glucose transport into the cell, and upregulate other glucose metabolism processes. But this activation isn't just an on and off switch, it's graded. How much effect activating those receptors has depends in turn on how well the chemical reactions in the activation sequence run. So even if a receptor is activated, its effect could be weaker or stronger depending on context. For example, if the activated pathways need enzymes to work, then even if they are activated, they won't work very well if the enzyme levels are low. Modifying the surrounding context is one way a cell could become more or less sensitive to activation by insulin. The effectiveness of a hormone like insulin can also be changed by how many receptors the cell makes available. The more copies of the program available to be keyed by insulin, the more sensitive the cell is to a given concentration of insulin, just because the keys are more likely to find the locks. So that's another sensitivity control mechanism. When a cell needs energy, it becomes more insulin sensitive, but when it has plenty, it starts to limit the effectiveness of the receptors and transporters. In other words, the cells resist insulin's promotion of uptake of available fuel. This resistance is an indication of cellular satiety, and being able to signal that and control fuel intake is necessary for the health of the cell! Given what we've already discussed, it shouldn't be surprising that cells adapt which fuels they take up readily based on context. Glucose transporters that use insulin can be dialed up or down in sensitivity. Fatty acid transporters can, too . These tend to happen reciprocally (ibid.) Because of the myopic insistence on viewing glucose metabolism as the normal default, and fat metabolism as pathological or an emergency alternative, the state of those transporters is called normal when there is high sensitivity to insulin-dependent glucose uptake, and low activation of fatty acid transporters. In a different world, rather than calling it normal, we might have called this state "fat resistance", but we don't. When your cells are insulin resistant simply because they're primarily using fat for fuel or they just don't need any more fuel at the moment, that's adaptive behaviour. They can quickly resume using glucose if fatty acid supply recedes and glucose and insulin levels rise. This kind of insulin resistance is called "physiological" or "benign" insulin resistance. The key is that "reversibility" aspect. Physiological insulin resistance is a useful signal, indicating satiety at the cellular level. The problem comes when cells are insulin resistant, but there are still high levels of glucose in the blood. This could happen, for example, if fat stores were not taking it in as fast as it is appearing in the bloodstream. High levels of glucose get noticed by the pancreas, which secretes more insulin in response. Extra insulin can partly make up for the fact that there is resistance, by activating a given number of receptors more frequently. As you can probably imagine, these processes can ratchet each other up in a feedback loop. You can tell this is happening because insulin levels rise above normal. This is called hyperinsulinemia. High insulin levels are unlikely to be harmful if they're transient, but when high insulin is chronic, it is associated with all the symptoms of metabolic syndrome. Metabolic syndrome, like all syndromes, is a name for a cluster of symptoms that often occur together. These symptoms are: central obesity (that is, extra body fat resulting in a large waist, as opposed to fat accumulating in the hips or thighs), high blood pressure, high blood sugar, high triglycerides, and low HDL cholesterol. Having any three of these qualifies as having metabolic syndrome. Each of them alone is also a risk factor for a large number of diseases, including two of our top killers, type 2 diabetes and heart disease. Many believe the relationship between hyperinsulinemia and metabolic syndrome is causal, perhaps by having a common cause . Here are just some of the more common conditions for which metabolic syndrome is a risk factor. Preliminary studies using carbohydrate restriction to treat several of these conditions have been encouraging , but insofar as they are real effects, the reasons are unclear and may differ by condition. Nonetheless, it should not be completely surprising if low carb diets help conditions that metabolic syndrome is a risk factor for, simply because low carb diets reduce every metabolic syndrome criterion . Any therapy that reduces risk factors for leading causes of death and disability warrants our attention.
Sidebar: But what causes hyperinsulinemia? When we see low carb diets reverse symptoms of metabolic syndrome, and the dreaded diseases that accompany it, it is tempting to think that carbohydrates themselves are the cause. Of course, a high carbohydrate diet alone can't be the cause, because, for one thing, we know of societies that eat high levels of carbohydrates where metabolic syndrome isn't prevalent . and we probably all know people who eat high levels of carbohydrate and don't have it. This means that high carbohydrate intake in and of itself can't be the problem. It doesn't mean that high carbohydrate diets can't be a factor, though. The best theory in defense of a causal role of carbohydrate is the "carbohydrate-insulin hypothesis". The idea behind that hypothesis is that if you keep eating high levels of carbohydrate, perhaps especially sugar, or other highly refined carbohydrates, it keeps that insulin resistance feedback loop going higher and higher, resulting in hyperinsulinemia, metabolic syndrome, and everything that goes with it. However, this explanation isn't wholly satisfactory by itself. For one thing, if the cells are insulin resistant, that means they are signalling cellular satiety. It seems reasonable to think that cellular satiety and whole body satiety are connected. The brain integrates a variety of signals about fuel availability to decide if we sould be hungry or not. Part of this equation probably comes from how much energy is in the bloodstream. If the muscle cells start saying no to glucose, then glucose will stay high until fat cells take it up. As long as glucose is high, it would seem odd to feel hungry. So then why would anyone with systemic insulin resistance be hungry enough to keep eating more glucose? There are many different proposed answers to this question. The carbohydrate-insulin hypothesis answer has to do with another function of insulin. Just as insulin unlocks glucose entry programs, it can also block the exit programs for fat from fat cells. With high circulating insulin, fat escapes fat cells more slowly, so there is less available to feed muscle cells. Thus it is reasoned that once circulating glucose starts running low, without adequate access to fat, hunger should ensue. But this still doesn't fully make sense. First of all, in a healthy person, once glucose goes down, insulin should follow suit, opening the access to stored fat. If insulin isn't going down when blood sugar does, there must be more going on. It's not obvious how this feedback loop would get bootstrapped in the first place. For the carbohydrate-insulin mechanism to come into play, the insulin response must be exaggerated out of proportion to glucose levels. There is evidence that certain forms of carbohydrates have this effect . If this were the only reason for elevated insulin, then metabolic syndrome could be reversed simply by avoiding refined carbohydrates. Indeed, there are people who can achieve good health simply by eating a so-called "whole foods" diet . Others however cannot do this restore their health without strict carbohydrate restriction. This raises the question of why their insulin levels don't normalise. There are other reasons for insulin to become elevated besides high, fast levels of glucose. For example, elevated insulin is part of the immune response . So damage or infection could cause it. If you have elevated insulin for a reason other than responding to elevated glucose, that insulin will still have the effect of clearing what glucose there is into cells. You will now potentially have not enough blood glucose, and simultaneously limited access to fat. In this situation, hunger seems more plausible. So now what happens if you eat more carbohydrates in response to that hunger? The glucose will stimulate yet more insulin. The cells already don't want any of it, and so are becoming more and more insulin resistant. This is one reason why, no matter what the initiating cause of hyperinsulinemia, low carbohydrate diets can help. Just reducing the glucose load goes a long way toward stopping the vicious cycle. Moreover, once you switch from glucose mode to ketogenic mode, you can continue along comfortably by eating enough fat to keep that going, even if your elevated insulin doesn't resolve enough to let you access your fat stores as much as you'd like. At least you won't be getting worse, and the symptoms associated with hyperinsulinemia in glucose mode will go away. Metabolic syndrome goes into remission because you have bypassed it.
These healing properties of the ketogenic diet are not, like many other diets, due to adding specific nutrients, for example to address deficiencies. And they aren't due to special chemicals in the food purported to have a drug-like effect, as has been proposed for plant "phytochemicals". Nor are they due to removing specific components that someone might have an immune response to. Ketogenic diets are almost completely agnostic about nutritional qualities in that sense. You can get into ketosis eating so-called "whole" foods, or highly processed foods. You can get in ketosis eating only plants or only animals or any combination of those. You can get in ketosis by greatly restricting calories or exercising a lot. Technically, you can even get into ketosis while eating carbohydrates if you eat enough of the right kind of fat (e.g. medium chain triglycerides) or take ketones as a supplement. Different approaches have different relative advantages. Getting your own body to generate enough ketones to be in ketosis is really all about the energy dynamics; you just have to manipulate the signals that dictate whether you are primarily storing energy or using it. This is a great strength of ketogenic diets, but it is also a great weakness. If something you are eating is detrimental to your health, a ketogenic diet might actually mask that damage just because it is in other ways an improvement from your baseline. If your condition is affected by something that isn't high in carbs, you may get only limited benefit from a ketogenic diet until that aspect is addressed. One way to recognise whether this is the case for you is if fasting gives noticeable benefit over a ketogenic diet. Fasting is extremely ketogenic. Most of what we know about ketosis originated in studies on fasting, because in our modern world fasting is the only time most people abstain from eating carbohydrates! Nonetheless, almost all of the benefit attributed to fasting actually comes from ketosis . But ketosis doesn't require fasting at all in humans. This is a great advantage of our species. If you feel significantly better when fasting than you do on a deeply ketogenic diet with full calories, then clearly it is not just the ketosis of fasting providing you benefit, but rather the removal of something (or a class of things) you normally eat. One of the two pillars of the Carnivore Diet is plant removal, because, empirically, many people seem to get benefit from it. The other pillar, of course, is meat eating, because of the unique nutritional contributions of animal sourced foods. The high contribution of plant foods and the low contribution of animal foods is sometimes why a low carb diet by itself is simply not enough to address all health conditions.
"Ok it's good, but it's not "natural"" There are many doctors, researchers, and others who understand the therapeutic value and exciting potential of ketogenic diets, but still feel that they are not particulary healthy. They view the ketogenic diet as a medical biohack, like a drug intervention that has value in extreme cases of need, but also involves significant trade-offs. If this were true, then going on a ketogenic diet wouldn't be worth the cost unless you were dealing with something severe enough to justify it. As discussed above, there are many ways for ketosis to happen; they aren't all equivalent! Because a ketogenic diet is so flexible, it is tempting, and relatively easy to try to cut and paste a state of ketosis into the current nutritional paradigm. And it makes intuitive sense to do that. If our modern ideas about what makes a diet healthy represent cutting edge science and are mostly correct, then of course we would want to integrate this therapy, when appropriate, into a diet we already think offers the best health outcomes. I believe this is a huge mistake, exactly because many of our modern conceptions about what makes a diet healthy are not, in fact, correct. We have built our nutritional knowledge on top of assumptions that haven't been adequately tested. In many cases, they haven't been tested at all. In some cases, they have been tested, found wanting, and yet they are still believed. The reality is that all of recorded history happened after the agricultural revolution. That means that almost everything we have learned about health and diet pertains to the context of a diet high in carbohydrates from grains. Grain agriculture is recent on the scene, from the perspective of the age of the human species. That means that any evidence we gain from the most ancient of texts on how people lived, is still a reflection of an extremely modern way of living that differs dramatically from most of our hominid existence. This can't be overemphasised. Evidence does not support the idea that we ate a high carbohydrate diet before grain agriculture, at least not with significant regularity. Even our prehuman ancestors did not likely get most of their calories from carbohydrates, nor do our closest primate relatives now. Part II of this book: "Human: A Whole Different Animal" will delve into these facts in much more detail. This century has seen major innovations affecting our available food supply. We have fast, efficient transportation to bring fresh produce from across the world. We have precise environmental controls for cultivating plants where they otherwise wouldn't thrive because they require set temperatures, additions of nutrients, or a vast supply of water. We have plant oils and plant proteins that need to be extracted and so require recent technology to acquire in large quantity. As we will see, even the kinds of plants we have available to eat has increased a lot in recent centuries. When we look at the early history of low carb diets we do not see diets based on avocados, olive oil, and loads of high fibre vegetables. We do not see pea protein shakes and almond flour cakes. The reason is simply that those things were not readily available common foods. Even olive oil, which has been collected since historic times, was probably not mostly used for consumption until recently . To illustrate, take this description from William Banting describing his low carb diet in 1864, not much more than a century ago . "For breakfast, I take four or five ounces of beef, mutton, kidneys, broiled fish, bacon, or cold meat of any kind except pork; a large cup of tea (without milk or sugar), a little biscuit, or one ounce of dry toast. "For dinner, five or six ounces of any fish except salmon, any meat except pork, any vegetable except potato, one ounce of dry toast, fruit out of a pudding, any kind of poultry or game, and two or three ounces of good claret, sherry, or Maderia [sic] - champagne, port and beer forbidden. "For tea, two or three ounces of fruit, a rusk or two, and a cup of tea without milk or sugar. "For supper, three or four ounces of meat or fish, similar to dinner, and a glass or two of claret. "For nightcap, if required, a tumbler of grog - (gin, whisky, or brandy, without sugar) - or a glass or two of claret or sherry." Notice how the meat chioces are described in variety and detail, but the only mention of vegetables is to distinguish potato from non-potato. Notice also the implication that there aren't any other high carbohydrate vegetables in quantity to speak of, and that there is no need to go into any detail about the varieties of vegetable to choose from. Likewise with fruit. In other words, putting a human on a low carb plant-based diet is something that never would have been plausible before now. Whereas, a low carb diet based on meat is something I submit we are already adapted to. Is it silly to suggest that there is more to ketosis than ketosis? That is, to suggest that the benefits of ketosis may be improved if we induce ketosis in a way that more resembles how it would have occurred for us in the past? Maybe. We can't know exactly where to draw the line, and certainly don't want to be caught up in a Paleo re-enactment scenario. But that doesn't mean that there aren't more things going on in tandem when we eat a meat-based ketogenic diet than when we achieve ketosis other ways. When you try to recreate a biochemical pathway in isolation, then you aren't normally creating a concordant state. That is, you may be creating signals in the body that actually oppose one another, because they've never adapted to occurring together. It's a reasonable null hypothesis that achieving ketosis with a diet that we would have been in ketosis on often in our evolving past, might have better effects than imposing ketosis under conditions it has never occurred in before. While some are trying to bring plant-based ideology to the ketogenic world, other believe that the middling results of ketogenic therapies in cancer are due to this discordance . As I will present in depth in a subsequent chapter, humans have a very special relationship to fat and ketosis. It is entirely reasonable to suppose that ketosis happened regularly in humans and prehumans in past millenia. But not in the context of tofu and canola oil.