Since his days playing with Gary Taubes and the Nutrition Science Initiative (NuSI) back in 2016, it seems that NIH’s Kevin Hall has been on a mission to create better studies to shoot an arrow into the heart of the Ketonians’ Sacred Cow, the Carbohydrate Insulin Hypothesis. His latest study may just be the kill shot!
With a Bachelor’s Degree in Physics and a PhD in biophysics, Kevin Hall, PhD now runs the Integrative Physiology Section of the Laboratory of Biological Modelling at the National Institute of Health. I’m impressed at how he continues to access some substantial funding to do some solid research to settle some of the hottest arguments in the Twitterverse.
Previous major studies
The calculus of calories
I first heard about Hall after he did some fascinating work to mathematically model human metabolism in 2015 which he detailed in the paper Validation of an inexpensive and accurate mathematical method to measure long-term changes in free-living energy intake and describes in this video.
Warning: Only watch this presentation if you are a maths nerd and want your mind blown at how complex trying to model human metabolism can be!
Fat vs carbs
In short, due to the slightly higher thermic effect of carbohydrates vs fat, it seems that people lose a little more weight on a low-fat diet compared to a low-carb diet when matched calorie for calorie.
While interesting, this is only really relevant if you are able to perfectly track and control your calories (which you can’t). In the real world, most of us eat to satiety and stop when we’re no longer hungry, so it’s more useful to think about what makes us eat (and absorb) more.
Effect of ultra-processed food
In his 2019 study, Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake demonstrated that we tend to consume more calories AND absorb more energy from highly processed foods.
This result should be no surprise if you’ve tried to eat just one doughnut or a potato chip. But nevertheless, it is useful to quantify the diabolical impact that highly processed food has on our ability to override our best intentions and self-control.
Beyond macronutrients and micronutrients in our food, the thermic effect of our food (aka dietary-induced thermogenesis) is a highly underrated parameter. Protein takes more energy to convert to energy than carbs, and carbs more than fat. When we convert food sources with a higher thermic effect to ATP for use in our body more is lost to heat or used up in the process of digestion rather than being available for use by our body or storage as fat.
But within macronutrients, the degree of pre-processing makes a big difference to how our bodies respond to food. That’s why protein powder is different from a raw steak, raw spinach and broccoli are different from refined flour and high fructose syrup, and perhaps why nutrients that come in real food are better than fortification and supplements.
I will come back to this again later, but before we go on, I think it’s worth noting the DIT of the different macros (shown in the table below).
|macro||cal/g||DIT assumed||DIT range||energy yield (cal/g)|
|protein||4||20%||20 to 35%||3.20|
|carbs||4||5%||5 to 15%||3.80|
|fat||9||3%||3 to 15%||8.73|
You will note that the DIT values have wide ranges. For example, a low DIT protein could be protein powder vs raw game meat and a low DIT carb could be sugar or HFCS vs raw potato or broccoli.
The bottom line when it comes to satiety (beyond macros, low carb vs low fat and animal vs plant-based food) is that we lose more energy when we digest and use the energy from minimally processed food.
Not only is more of the energy available for use, but our body LOVES these easy energy sources and tends to eat a lot more of them whenever they are available.
Hall’s latest study: Keto vs plant-based low fat
Hall’s latest study A plant-based, low-fat diet decreases ad libitum energy intake compared to an animal-based, ketogenic diet: An inpatient randomized controlled trial is pretty darn cool.
One of the critical challenges with typical nutrition studies in a controlled setting is that calorie intakes are matched, so they don’t test the satiety response.
In experiments with free-living participants, people don’t track what they eat very well (and they often go off plan and “cheat”), so it’s hard to rely on the data.
This yet to be published study compared a 75% fat keto diet with a low fat 75% carb plant-based diet.
They effectively pitted two of the most popular but diametrically opposite diets against each other in a controlled setting and tested which one led to:
- greater satiety (when they ate as much as they wanted to),
- weight loss,
- lower insulin, and
- more stable blood sugars.
While participants were not diabetic, the 20 participants were not particularly lean or metabolically healthy either (average BMI of 27 and an average body fat percentage of 32% for men and 41% for women). They are pretty similar to the general population who is looking to lose some weight and improve their diet.
Subjects were randomised to one diet for two weeks and then switched to the other for the following two weeks. You could say it would have been nice for the trial to go on for longer, but there is a limit to how long you can lock people up for when it comes to cost and sanity.
All the food was provided in the NIH research facility. They were given twice as much food (in the form of meals and snacks) as they would need to maintain their weight and told to eat as much of it as they wanted (i.e. to satiety). Then, by measuring the food leftover at the end of each day, Hall and his team were able to see how much participants actually ate.
So, what did they find?
Both approaches were equally enjoyable and familiar
Participants reported similar enjoyment on both dietary approaches and found both approaches familiar and enjoyable. You can check out the photos of the meals in the full-text preprint of the article here.
Hunger, fullness and satisfaction were similar on both diets
Participants reported similar levels of hunger, satisfaction and fullness.
Participants ate fewer calories on the plant-based diet
The headline of the study was that, on average, participants ate significantly fewer calories on the low-fat plant-based diet compared to the keto diet!
In fact, all participants ate less on the plant-based diet.
When the calories came mainly from fat (blue in the chart below), the overall calorie intake was clearly higher. They actually ate more energy from fat on the keto diet than they did total energy on the plant-based diet.
Blood glucose levels were more stable and a little lower on the keto diet. HbA1c, average glucose and insulin levels reduced on both diets (but slightly more on the keto approach).
There was definitely a flatter and lower blood sugar response on the low carb diet compared to the low-fat diet.
While not exactly optimal, you couldn’t call the low-fat blood sugars particularly unhealthy or alarming. On average, participants saw a rise of up to 30 mg/dL or 1.7 mmol/L after eating, which is not great, but not horrific either. Conversely, free fatty acids in the bloodstream on the low-fat diet were lower and decreased a lot more after the low-fat meals.
It shouldn’t be a surprise that the fat in our bloodstream is different on a low fat vs low carb diet. We can run our bodies on either fat or glucose. When we eat a high carb meal, the pancreas raises insulin to shut off the release of stored body fat into our bloodstream to allow the glucose to be burned up.
The problem comes when we try to run on a similar mixture of carbs and fat together. This formula is rare in nature but common in our modern processed food.
The keto diet was definitely ketogenic!
There is no argument that the people on the keto diet were in ketosis. Blood ketones levels reach 2 mmol/L towards the end of the two weeks.
In fact, the keto approach was designed to mimic “Well Formulated Ketogenic Diet“ as defined by Steve Phinney and Virta.
We burn less energy when we eat fewer calories
The table below shows the energy expenditure on the two diets, with:
- lower energy expenditure on the plant-based diet (less energy eaten typically means we adapt by burning less energy), and
- a higher respiratory quotient (RQ) on the plant-based diet (meaning that they were burning more carbs while the low carb group were burning more fat).
It’s worth noting here that the resting energy expenditure and sleeping energy expenditure were lower on the plant-based approach with less calorie intake. This lower energy expenditure means that you are going to feel cold (and perhaps a little sluggish and depressed) on your diet, which may affect long term compliance.
A diet with a higher thermic effect of food (i.e. more energy lost to heat) will leave you feeling warmer and likely more energetic. Bodybuilders who intentionally try to overeat protein often report the meat sweats as their body heats up and they feel energised with heaps of excess energy to burn. A dietary approach that leaves you feeling energised and satiated is much more likely to lead to long term compliance.
People lost more weight on the keto diet
Presumably, because there was an improvement from their baseline diet, people lost weight on both the low fat and low carb diets.
You can see initially people on the low carb lost more weight as their glucose lowered and they lost “water weight”. Conversely, in the first few days, we see people gain a little weight on the plant-based diet due to an increase in fibre in their gut and possibly more water weight.
But technically, due to the overlapping error bars, the difference in weight loss was not statistically significant (make of that what you will). You would really need to see how the diets performed head to head for at least six weeks to verify any statistically significant difference in weight loss.
For reference, the chart below from our 6 Week Nutrient Optimiser Weight Loss Challenge (which used high satiety nutrient-dense foods and meals) saw a more rapid weight loss in the first week or so, with a slower rate thereafter for the following five weeks.
But they lost more fat on the plant-based diet
However, interestingly, at the end of the two weeks, it was the people on the low-fat diet that lost more fat from their bodies (which is what most people are wanting). The weight loss on the keto approach was more water weight.
For every gram of glucose stored as glycogen, we also store 4 grams of water, so a reduction in glycogen would lead to a loss of water weight, so you would look less inflamed. Participants on a plant-based diet would also likely have a much higher weight of food in their gut given the extremely high level of fibre on this diet.
Energy density matters!
Finally, this is perhaps the most interesting chart, showing that:
- participants ate faster in terms of calories per minute on the animal-based low carb diet (ABLC), but
- ate more quickly in terms of grams per food per minute on the plant-based low-fat diet (PBLF).
When we can eat more slowly we feel fuller with fewer calories, particularly when our food contains fewer calories per gram.
What about insulin?
When it comes to insulin, Hall’s latest study shows that insulin levels reduce in an energy deficient and/or as we lose weight (regardless of diet). Although people on the plant-based diet lost more fat from their body, insulin was lower on the keto diet.
While it would certainly be great to data over a longer time period to understand if there is any significant difference, data from this two-week study don’t support the idea that lower insulin levels equate to more fat loss (i.e. the Carbohydrate Insulin Hypothesis).
As shown in the chart below, insulin secretion across the day (not just the bolus insulin secreted after meals) is largely proportional to body fat levels. Insulin works to hold our stored energy in storage, so the more fat we have to store, the higher our insulin levels need to be.
The Carbohydrate – Insulin Hypothesis simplistically assumes:
carbs -> insulin -> fat storage
But in reality it is a little bit more complex. A rise in insulin is a response to increased body fat (not the cause):
low satiety nutrient-poor foods -> increased cravings and appetite -> increased energy intake -> fat storage -> increased insulin
Hence, the solution to managing your diabetes, blood sugar, insulin levels and avoiding the myriad of complications of metabolic syndrome is:
How does this align with our previous analysis?
I’m thrilled that this study looked at how much we eat ad libitum (i.e. as much as you want) with different dietary approaches. So I thought it would be useful to compare Hall’s data to our previous satiety analysis.
We eat less when we move away from the carb+fat danger zone
All participants lost weight when they switched to a low carb or low-fat diet. Presumably, this was because their diet “improved” on either dietary approach as they moved away from the typical carb+fat danger zone.
The chart below from our previous analysis of half a million days of MyFitnessPal data shows that we naturally tend to gravitate towards a diet that is a mixture of carbs and fat when it is available. We love these energy-dense processed foods that allow us to get more energy more quickly and thus get a bigger dopamine hit.
This ‘moderate’ carbohydrate approach (i.e. 40 – 50% carbs) tends to align with all the processed hyperpalatable junk foods that don’t occur in nature (e.g. doughnuts and croissants).
Don’t eat “fat to satiety” (unless you want to get fat)
The results of Hall’s latest study also align quite nicely with our analysis of half a million days of MyFitnessPal data. As the energy density increases due to more dietary fat, we find these foods easier to overeat. The common keto adage of “just eat fat to satiety” doesn’t seem to work in real life if your goal is to lose fat from your body.
Our analysis of sixty thousand days of food diary data from Optimisers also showed a similar trend. That is, a higher percentage of energy from fat leads to greater energy intake.
Ketones don’t equate to fat loss.
We have also noted previously that people with the best metabolic health tend to have lower glucose AND blood ketone levels and less energy floating around in their bloodstream (see Optimal ketone and blood sugar levels for ketosis).
Once you exceed your Personal Fat Threshold and your body fat can no longer hold the excess energy from the food you eat, it backs up into your bloodstream and you see elevated levels of ketones, free fatty acids and glucose.
While blood sugars can be a useful gauge of your energy status, measuring blood ketones is of little use to guide your weight loss on a keto diet. There is no way to tell if those ketones are from the fat being released from your bum and belly or from the three bulletproof coffees you just drank to try to boost your ketones.
The chart below shows our pooled analysis of four thousand blood ketone results from a range of different people whose highest ketone reading (blue bars at the bottom of the chart) also have the highest total energy (i.e. glucose plus ketones) in their bloodstream.
If you’re on a low carb or keto diet, then you might see ketone levels of 0.3 – 1.5 mmol/L. But chasing super high ketones with lots of extra dietary fat is not the way to lose body fat.
Optimising Nutrition advisor Ted Naiman developed this chart from Hall’s study data showing that the low-fat plant-based diet had a much lower energy density than the typical American diet.
Our satiety analysis also shows that people who consume food with a greater weight of food tend to eat less. We just find it harder to put away as many calories when you need to chew through tons of food.
The problem with using energy density though is that you need to get greater than around 4.0 kg/2000 calories to get a significant impact and few people seem to be able to achieve this on a regular basis living in the real world,
The chart below from the analysis of our series of 22 recipe books shows that a diet with a greater weight per calorie also tends to be more nutritious.
This next chart (also prepared by Ted from the Hall study data) shows that the low-fat plant-based diet had a lot more fibre than both the keto diet and the typical American diet.
Again, our analysis shows that more fibre aligns with greater satiety. The average intake of Optimisers is about 5% and we have set a stretch target of 7% as an achievable target that aligns with greater satiety.
The remarkable thing about the Hall plant-based diet is that they are consuming 12% fibre (which is largely unavailable to the body to be used for energy). The fibre distribution chart below shows the distribution of fibre intake from sixty thousand days of data from Optimisers. Very few people are getting that level of fibre in the real world, whether or not they are following a plant-based or keto diet.
Do carbs or fat have the biggest impact on satiety?
Our analysis suggests that lowering carbohydrates will reduce your ad libitum energy intake by about 17%.
But lowering your dietary fat will reduce your ad libitum calorie intake by about the same amount.
Combining these hacks together (i.e. reducing both fat and carbs), will decrease ad libitum energy intake by around 20%.
In line with this, the chart below shows the change in energy availability from carbs, fat and protein in our food system over the past century. Over the past fifty years, the growth in obesity rates (orange line) has tracked fairly closely with the increase in calories (purple line) from both carbs and fat.
So what changed? To help answer this, the next chart shows that the increase in energy in our food system has come from added vegetable fats and oils along with flours and cereals (i.e. the plant-based products of large-scale agriculture). Meanwhile, meat, eggs, nuts, dairy, vegetables and even sugar is largely unchanged.
The combination of refined carbs together with fats (both from “plant-based” sources) seems to be the elephant in the room of the diet debate that continually gets stuck in a seemingly infinite loop of fruitless debate over plans vs animals and carbs vs fat.
After all our analysis, we found that one of the most powerful satiety factors is the percentage of protein in your diet.
People who consume a higher percentage of protein end up consuming around 60% fewer calories. While it’s not necessary to go to the full extreme of a super high Protein:Energy ratio or a Protein Sparing Modified Fast, tweaking your diet a little in that direction is likely to help
That’s why, in our 6 Week Nutritional Optimisation Masterclass, we ensure participants are getting adequate protein, and then guide them, based on their biometrics (i.e. blood sugars, weight and body fat) to dial back their intake of processed carbs and fat until they see the results they are chasing.
It’s worth noting here that both the low fat plant-based and the keto diet in Hall’s latest study had around 14% of calories from protein which aligns with the poorest satiety outcome.
However, because we crave easy energy from refined carbs and fat, we tend to gravitate to lower protein intake. However, while it’s not easy to maintain a very high percentage protein intake, it seems to be much more achievable to get some positive satiety benefit by marginally increasing protein percentage (by reducing energy from refined carbs and fat).
Once the nutrition world is ready to think in terms of optimal (rather than just debating carbs vs fat and plant vs animal to win arguments on Twitter) It would be fascinating to see another arm of the same study done with a nutrient-dense approach with high protein.
Thermic effect of food/net energy
As noted earlier, the thermic effect of food is a major determinant of satiety. We eat less calories when the food is less processed and our bodies need to do more of the processing.
- The blue line in the chart below from our previous satiety analysis shows the % energy from protein vs how much people eat. You can see as people eat a higher percentage of protein they seem to eat fewer calories.
- Meanwhile, the red line shows the net energy available from the food after we account for losses from dietary-induced thermogenesis.
While protein has a big effect on our satiety, the net energy available after losses seems to have an even greater effect on the number of calories we consume. While taking your percentage of protein from one extreme to the other can have a 65% impact on your calorie intake, our analysis of data from Optimisers indicates that a lower net energy availability in your food (i.e. higher dietary-induced thermogenesis) can have a massive 75% impact on your calorie intake!
This next chart shows how the plant-based and keto approaches stack up with our 22 recipe books in terms of dietary-induced thermogenesis assuming the following losses for each macronutrient:
- protein – 20%
- non-fibre carbs – 5%
- fat – 3%
- fibre – 50%
You can see at the bottom of the chart, the keto approach used in the Hall study has the lowest dietary-induced thermogenesis of all the approaches. It was destined to fail.
In the middle, we have the Hall plant-based approach with a massive amount of fibre (yellow) compared to all the other approaches (WARNING: THIS IS NOT YOUR TYPICAL PLANT-BASED DIET).
At the top, we can see that the high protein:energy ratio and the fat loss approaches will provide a much greater dietary-induced thermogenesis and hence greater satiety. Meanwhile, food choices with a lower dietary-induced thermogenesis will leave you hungry, cold and potentially fat.
Is a plant-based diet better than keto?
While the plant-based camp will be claiming Hall’s latest study as a victory over the evil keto and carnivore camps, I don’t think we should dumb it down to thinking simplistically in terms of plant-based versus animal-based yet again. This study did an excellent job of comparing two diametrically opposing extremes, but they are not optimal in terms of either nutrient density or satiety. Optimal is rarely found at the extremes.
I wanted to do an analysis of the nutrient profile of the diets, but there wasn’t enough detail of the ingredients in the recipes, so I thought it would be interesting to compare the exemplar versions of the plant-based diet and keto diet from our series of 22 nutrient-dense recipe books tailored for different goals and preferences.
The nutrient fingerprint charts below show our exemplar version of a plant-based diet from our series of recipe books. While you can create a reasonably nutritious plant-based diet, it still lacks a number of essential nutrients towards the top of this chart.
This next nutrient fingerprint chart is from our recipe book designed for therapeutic keto (which has similar macros to the recent Hall study). Again, there are a number of vitamins and minerals that are harder to get when you prioritise high fat as the dominant macronutrient.
Note: We designed the book or keto recipes for people who require elevated ketones for therapeutic purposes (e.g. Alzheimer’s, dementia, Parkinson’s, epilepsy etc), not fat loss.
Blood sugar & fat loss
If your goal is to stabilise blood sugars and lose fat from your body (via a diet with greater satiety and nutrient density), we can create a diet that has a much better nutrient profile and allows the energy from fat to come from our body rather than our diet. The nutrient fingerprint is for our book of recipes designed for blood sugar and fat loss.
Then, if we are really serious about fat loss, we can focus even more on nutrient density and satiety which results in less fat, fewer carbs, more fibre and more protein. The micronutrient fingerprint chart below shows the micronutrient profile that is possible in our exemplar version of recipes designed for fat loss.
Finally, we have the high protein:energy ratio book (hat tip to Ted Naiman) which isn’t quite as nutritious but will have even higher satiety due to the very high thermic effect of food.
What is the best way to judge optimal nutrition?
As demonstrated in the Hall study, our bodies adapt our energy expenditure based on how much energy we consume. Both sides of the calories in vs calories out equation are incredibly hard to calculate accurately!
Unless you are able to weigh and measure everything you eat for the rest of your life accurately, the best approach is to aim for a dietary approach that:
- provides you with greater satiety with a higher dietary-induced thermogenesis (so you eat less intuitively with less hunger),
- maintain your lean muscle mass (so you maintain a high metabolic rate and continue to burn energy), and
- ensure your diet is not overly processed (to ensure you assimilate less of the calories).
While not considered in the Hall study, we thinking focusing on getting the essential nutrients without excess energy is a pretty good idea too.
There are many different moving parts to nutrition.
At the simplest level, we believe nutrition should primarily be about nutrients. While focusing on foods and meals that contain more protein and fibre (along with less refined carbs AND fat) tends to lead to greater satiety, we have found the best approach is to focus on getting more nutrients per calorie.
Rather than trying to restrict how much you eat, once you focus on giving your body what it needs, everything else falls into place.
If you want to lose body fat without having to track everything you eat and fight against your instincts and hunger, focusing on getting more nutrients per calorie tends to lead us to eat fewer calories.
Rather than wasting any more energy debating plants vs animals or carbs vs fat, hopefully, in time more people will get on board seeking optimal nutrition.