Nutritional ketosis induced by carbohydrate restriction is often associated with major weight loss, which raises some important questions. Do ketones cause weight loss? Do ketones promote a metabolic condition whereby fat melts away to a greater extent than a non-ketogenic diet of equal energy content? Alternatively, can a person maintain or even gain weight while in nutritional ketosis? To explore the answers to these questions, we need to venture into the complex inter-relationships between keto-adaptation, appetite, energy balance and weight loss.
But before we do, we need to acknowledge that these questions have sparked an intense debate among serious scientists and the lay public over the last 15 years, and in the process sparking the expenditure of tens or millions of dollars on research in which both sides of the debate strove to prove the other side wrong. While we will discuss some of this research, much of it involving short-term studies done under artificially controlled conditions, we will also try to anchor this discussion with a perspective drawn from research utilizing well-formulated ketogenic diets for meaningful durations in the real world.
In almost every human study of overweight patients lasting 3 months or longer comparing a ketogenic diet to a low fat diet in an outpatient (aka ‘real world’) setting, the weight loss with the low carb diet is somewhat or significantly greater (Sachner-Bernstein 2015). And despite claims by skeptics that “most of that weight loss is water,” in fact most of the weight lost on a well-formulated ketogenic diet lasting a few weeks or longer comes from body fat. Add to this the common anecdotes of individuals who “went low carb” and lost a lot of weight seemingly effortlessly, and one could start believing that there is something about ketones or nutritional ketosis that mandates body fat loss.
Some scientists and journalists have concluded from this body of evidence that there exists a ‘metabolic advantage’ associated with ketogenic diets. In other words some believe that ketogenic diets cause a greater expenditure of energy (aka, calories) than non-ketogenic diets leading to the claim that ‘a calorie is not a calorie’ (Feinman 2003, Taubes 2007). However this flies in the face of a standard tenet of nutrition and dietetics that in order to lose weight, we always have to eat fewer calories than we burn – i.e., ‘a calorie IS a calorie.’
Unfortunately this reduces the debate to a very simplistic level. Why? Because we know that hunger, appetite, energy expenditure (i.e., metabolic rate), and even our propensity to be active are highly regulated by an increasing list of hormones and signaling molecules, not to mention our genetic inheritance (Bouchard 1994). Moreover these various factors interact with each other – for example: exercise stimulates hunger, calorie restriction increases hunger and decreases spontaneous activity (Keys 1950), calorie restriction reduces metabolic rate, and exercise plus calorie restriction markedly reduces metabolic rate (Phinney 1988).
And now add to this the ketosis piece that makes solving this puzzle even more complex. Is a calorie still a calorie if cutting carbohydrate to the point of inducing nutritional ketosis reduces appetite, allowing the same level of satiety with a reduced energy intake? (Boden 2005) Does the keto-adapted individual feel the same need to reduce spontaneous physical activity while losing weight? What might a barely measurable increase in energy expenditure in a short-term study translate to over a longer duration? Theoretically, just a one hundred Calorie per day deficit translates to 10 pounds of body fat lost over a year.
First, does being in nutritional ketosis necessarily cause weight loss? For individuals who have experienced fairly rapid weight loss with little effort, their answer is usually a resounding yes! But remember that this is typically based on one person’s experience (or one person and a few of his/her friends). This commonly happens in a person who is relatively insulin sensitive, so that when that individual gets to their new stable (‘maintenance’) weight, they probably did not need to remain in nutritional ketosis—i.e., they could eat a wider range of total daily carbs and still remain weight stable. So in that person’s experience, it looks like nutritional ketosis caused their weight loss and it stopped when they ate enough carbs to go out of nutritional ketosis. In scientific terms, we need to decide if this is a causal relationship, or just an association.
This leads us to the second question: Can a human remain in nutritional ketosis and not lose weight? For example, is it really true that some of us who have been in nutritional ketosis for a decade or longer could really maintain a stable weight and body composition? And if so, is this possible only by eating ridiculously large amounts of fat?
Let’s step back and think about this issue from a cultural/historical perspective. If Native American nomads (e.g., the Osage, Kiowa, Lakota, Blackfeet, Shoshone), the Inuit in the Arctic, or the Maasai in Africa ate moderate protein and virtually no carbs, they would remain in nutritional ketosis year around, and yet they manifestly did not keep losing weight until they wasted away.
Furthermore, when we studied 9 lean healthy young men and fed them first a ‘balanced diet’ for a week locked in a metabolic ward and then the exact same calories as a ketogenic diet for 4 weeks, they did not lose any body fat (only about 2 pounds of water) despite being consistently and strongly in nutritional ketosis. (Phinney 1983) And finally, the 10 low carb runners from our FASTER Study had been following a low carb or ketogenic lifestyle for at least 6 months (average blood ketones of 0.6 mmol/L); but while thin, none of them were continuing to lose significant weight (Volek 2016).
These conflicting perspectives challenge us to think more deeply about the relationships among ketones, appetite, and energy metabolism. With apologies for getting into scientific details, here are a few relevant facts about ketones.
Coming back to this question of ketones and weight loss, when someone with some extra body fat begins a ketogenic lifestyle, perhaps it is the increased ability to burn these stores, coupled with the reduction in appetite and cravings, that facilitates initial weight loss. In this scenario, keto-adaptation facilitates weight loss, but only as long as the reduced hunger and cravings allow one to comfortably eat fewer calories per day than one burns.
Over time, most people who sustain a ketogenic lifestyle stop losing weight and find a new stable weight (Hallberg 2018). This is achieved when their natural instincts of hunger and satiety lead to an increase in dietary fat intake to balance out one’s daily expenditure. But as long as dietary protein is kept moderate and carbs low, this dietary fat is used in place of body fat to produce ketones, so clearly nutritional ketosis can be maintained without any further weight loss. We have named this a ‘eucaloric ketogenic diet’ (Phinney 1983).
But here’s a problem that many people experience. They have been told that increasing blood ketones will speed their weight loss. However, rather than cutting back on carbs and avoiding extra protein to boost ketone levels, they are led to believe that they can get the same effects by adding extra MCT oil, coconut oil, or exogenous BOHB to push up blood ketone levels. As noted above, this does not enhance their body’s ability to burn fat. It just gives them a type of fat that has to be burned (some of it as ketones) in place of body fat. No wonder they are usually disappointed when their weight loss stalls well above the goal they want to reach.
So finally, let’s discuss if there really is a ‘metabolic advantage’ to nutritional ketosis, meaning that the body burns more energy per day at any set level of physical activity compared to when a non-ketogenic diet is consumed. The simple answer is that we still don’t know the answer for sure. But we do know enough to get a rough idea how much of a metabolic inefficiency might be associated with nutritional ketosis.
As noted above, our bike racer study (Phinney 1983) involved 9 lean men locked up while eating a precisely controlled ketogenic diet for 4 weeks. In addition to daily weights, three methods were used to determine changes in body composition. As a group, these subjects lost 1 kg of body weight in the first week of the ketogenic diet, all of which was attributable to reduced muscle glycogen stores (which were directly measured). After that, their weights were stable for the next 3 weeks. Unfortunately in this study we did not have the opportunity to measure metabolic rate, but based on our body composition data, anything over a 3% increase in energy expenditure associated with the ketogenic diet would have shown up as non-water associated weight loss (by implication, a loss of body fat). Clearly these 9 men did not demonstrate an obvious increase in body fat loss in the first 4 weeks of keto-adaptation.
And then there is the controversial NuSi study* (Hall, 2016). This involved 17 individuals confined for 2 months in a metabolic ward and fed two different diets containing identical energy contents but differing in carbohydrate contents – one ‘balanced’ and one ketogenic. After 4 weeks of adaptation to each diet, the subjects had their metabolic rates monitored by two different methods: one using isotope analysis over the last 10 days and the other with continuous indirect calorimetry in a chamber for 24-hrs. The average chamber energy expenditure over 24-hrs was 75 kcal/day greater during the ketogenic diet. Given that the average subject in this study was consuming about 3000 kcal/day, that translates to about a 2% difference. Using the different isotopic method to assess average metabolic rate over the last 10 days of each 4-week study period, the calculated increase in daily energy expenditure on the ketogenic diet was closer to 150 kcal (i.e., 4%).
As explained below, the folks who did the NuSI metabolic ward study committed 2 errors. The first was a design flaw: the low fat diet was administered first to all 17 study subjects, and then they were given the ketogenic diet for the second month. In a well-designed trial, the diet sequence would have been randomized. The second flaw in the study was that these scientists underestimated their subjects daily energy needs by about 300 kcal/day – a 10% error. The combination of these two flaws would predictably tilt the playing field against seeing a significant increase in energy expenditure during the ketogenic phase of this study.
Thus these observed 2% and 4% increases in energy expenditure on the ketogenic diet underestimate the true metabolic effects of a well-formulated ketogenic diet. From the short-term perspective of a month or two, whether this is 75 Calories per day or 300 Calories per day, the added effects on weight loss would not be spectacular. But from the longer term view, a sustained daily increase in energy expenditure of 200 Calories would translate to 20 lbs of adipose either lost or not gained.
In summary, being in nutritional ketosis will accelerate the rate at which the body burns fat, and this is a fundamental key to the short- and long-term benefits of a ketogenic diet. If the extra fat that is burned is compensated by an increase in dietary fat, then no body fat loss will occur (but there still will be other benefits). However, most people carrying excess fat tissue who achieve nutritional ketosis by eating natural low-carbohydrate foods initially feel more satiated, allowing them to eat less fat than they burn, which results in net fat loss. But eventually, even when one is in sustained nutritional ketosis, our natural instincts prompt us to increase fat intake to meet our daily energy needs resulting in a stable weight and body composition.
Bottom line: While there might be a ‘metabolic advantage’ to a ketogenic diet over one containing a fair amount of carbohydrate, on average the difference is small. This does not explain why some people seem to lose weight relatively easily when carbs are restricted, which may be due to inter-individual variation. However the common observation of significant and sustained weight loss over months and years (Hallberg et al, 2018) is more likely a result of the benefits of nutritional ketosis on fuel flow, appetite, and cravings; as well as the reduced inflammation that is triggered by modest levels of beta-hydroxybutyrate. Clearly humans following a long-term ketogenic diet can eventually remain weight stable by adjusting fat intake to balance daily fat use for fuel. For those wishing to lose weight additional rather than remain weight stable, one’s goal should be to reduce dietary fat intake down to the margin of satiety (just enough, but not too much) and avoid or limit non-satiating energy sources such as alcohol.
Addendum – NuSI Study Details
*NuSI stands for ‘Nutrition Science Initiative.’ It was a not-for-profit organization that sponsored research attempting to objectively answer some basic questions about low carbohydrate and ketogenic diets. In this particular study, they paid for a very meticulous study at 4 different top-notch academic sites comparing energy expenditure, first on a high carb and then a ketogenic diet in a total of 17 individuals. In order to have the best chance of changing the negative mainstream bias against nutritional ketosis, NuSI chose some of the most steadfast ‘ketone skeptics’ in academia to run this study.
Unfortunately, despite the expertise of these highly regarded scientists, they systematically mis-calculated the energy needs of the study subjects. As a result, instead of everyone getting an accurately-dosed energy balancing diet, they were under-fed by about 300 Calories per day. Because all of the subjects got the high carb diet for the first month and the ketogenic diet for the second month, the natural effect of restricting energy intake by 10% would have been a small but significant reduction in resting energy expenditure that would have occurred progressively over the 2-month study duration. In other words, by significantly underfeeding the study subjects and always administering the ketogenic diet in the second month, the deck was stacked against the ketogenic diet. Taking this into account, the 2-4% rise in daily energy expenditure in the second month would probably have been double this amount had the study been properly designed. Whatever the true effect, these numbers are fairly small as noted above; but this is a study in which the odds were tipped in favor of the status quo through clever experimental design, and in which the conclusions were not justified by the data.
Have more questions about nutritional ketosis? Check out our FAQ by Dr. Steve Phinney and the Virta team.
The information we provide at virtahealth.com is not medical advice, nor is it intended to replace a consultation with a medical professional. Please inform your physician of any changes you make to your diet or lifestyle and discuss these changes with them. If you have questions or concerns about any medical conditions you may have, please contact your physician.