NuSI metabolic ward study on low-carb & the insulin hypothesis

So Gary Taubes’ organization, NuSI, helped fund a study by Dr. Yoni Feedhoff (which was also funded by the NIH).

They had 19 individuals on a carefully controlled diet, at a fixed calorie level, go through a baseline diet with normal carbs, and then a low-carb diet with 5% carbs for a month (in total, a 2-month study.) They had repeated 24-hour metabolic ward habitations during the study in order to watch their change in energy expenditure and their oxidation rates as they shifted from normal metabolism to ketosis. They measured fat levels with DXA.

Basically, this expensive study went with the gold standards for measurement accuracy on everything.

Most interesting (to me) findings:

  1. During the 15 days on the “baseline” normal diet with carbs, the subjects were already losing fat - even though they were supposedly eating the same “normal” number of calories they ate previously! This is just another example of how people are nearly always eating more than they think; and when they’re forced to eat only how much they thought they were eating, they often lose weight and lose fat.

  2. During the subsequent 30 days on the low-carb diet, the subjects lost only as much fat as they lost in the first 15 days. In other words, fat loss slower while eating low-carb, even though eating the exact same number of calories and getting the exact same amount of activity. (They did have a jolt of rapid non-fat weight loss at the beginning of the transition, as the body’s protein burning went up before the body could become adequately fat-adapted, but for the rest of the thirty days, fat loss was slower than while previously eating carbs.)

They conclude there’s no metabolic advantage to a low-carb diet, and that that the insulin hypothesis proves invalid as regards metabolism and fat storage / fat bruning. (But that there’s some interesting help on the appetite management side.)

Worth a listen.


Looks like a good study. I’d love to see them follow up with the appetite topic.


The problem being here that the diets are artificial and forced, not what happens naturally, because people eat more naturally. Insulin is not the only hormone that high/normal carbohydrate levels increase, they also increase leptin resistance, causing increased hunger and more painful hunger pangs when not satisfied. If you’ve ever tried to stick with a low calorie, low fat diet, you know this too well because you are miserable literally all the time. Even early on in this forum I saw repeated, many complaints about hunger and supplementing soylent use with snacking because they did not feel full and were still hungry after consuming. You can improve that with lots of extra fiber, but only so much.

Secondly, the time table for ketogenic adjustment of insulin sensitivity is on a much longer timescale, taking 6 months to a year to reduce that resistance down to healthy levels in the studies on the topic, so a one month study is specifically designed to avoid the parameters under which benefits are normally noted.

Third, there is a reason we use two separate groups for studies instead of putting them into fat loss prior to studying fat loss via their “baseline” diet, which was not baseline at all, but as you noted, less than they were eating as they were already losing weight, something that is not a normal condition. And we know that weight loss becomes more difficult over time because the body attempts to adjust your metabolic rate to maintain fat reserves and prevent starvation, and because fat shed from fat cells initially is much looser than when those cells are smaller and need to be further reduced or eliminated.

It looks to me like it was specifically designed to get these results given what we know of the science on the subject, not a “gold standard” study. Methods of data collection and monitoring may have been excellent, but if the design is fundamentally flaws I don’t know why I should trust the results.


No, we don’t know that. It’s an idea that floats around a lot but I’ve yet to see any evidence for it. It seems to be mostly an excuse people use to rationalize why they failed the diet failed.

The real reason weight loss becomes difficult overtime is the mental resolve weakening and people starting to long for all the foods they used to eat and love.

I’ve mentioned this else where and I think it’s worth repeating.

There’s no reason for the body to hold on to its fat storage. The reason it stored this fat is so that it can use it later on. When you have 20kg of extra fat, that’s way more than what the body should think it needs. Normal body fat percentage is around 10-15%


From the comments at the end of the video, it sounds like they cover the appetite topic in the paper’s discussion, but it’s just no presented here. The primary point of the study was to design a study that should show signs of the postulated mechanism underlying the insulin hypothesis.

I’m amused by the fact that the current study is, in fact, a longer study than his prior attempt. His previously study was criticized as being an otherwise well-designed study which would have shown clear benefit to the low-carb diet if only they had run it out to 30 days.

The reason NuSI supported this study with funding was so that they had the money to run the (costly effort) out to 30 days, being confident that this would produce a finding to support the insulin hypothesis.


I don’t know why you’re looking for a conspiracy. NuSI helped fund this study because they want to show that carbs are bad for you. If anything, we should be wary of bias in the design favoring the ketogenic diet.

I’m not at all commenting on your specific complaints about the study, which may very well be valid. I haven’t done enough research to comment there. But it’s a much more severe claim to accuse the researchers of deliberately designing the study to produce dishonest results. Is that honestly what you believe?


What resolve? I’m on a very low carb diet and I have virtually no desire to eat more or for foods I no longer eat. It just isn’t a thing that people have to deal with when they are not eating a high carb diet. And why diets fail is just as important and critical as the quantity of weight gained or lost. If leptin sensitivity is too adversely affected by a normal carb diet, then even monitoring your calories will make you more and more miserable if you try to control only the number of calories, which is both the findings of the research on diet failure rates and obesity generally.

And we still have to contend with weight gain to begin with, which we find happens in most of the population in any area impoverished that relies heavily on high carbohydrate diets out of necessity, in literally thousands of examples going back well over a hundred years.


The rise in obesity in obesity in the US tracks extremely well with the rise in overall calories consumed:


Some people will always blame the carbs, but for me, all the evidence points to the total calories.


Which leads to the appetite question. Did the diets shift to more carbs, and are people overeating because carbs are less satiating?


Can we at least agree that fats are delicious? Mmm, fats.


Yeah, that low-fat craze was a bomb.


Much of the carb vs fat research seems to be done with regard to weight loss. Given the obesity epidemic this is understandable. While calorie deficits still seem to be the only real solution, high fat diets may help with appetite.

But for the Soylent/soylent market, where macro nutrient types and quantities can be fine tuned, perhaps the bigger question is which diet is better for over all health?

Granted for niche markets such as diabetics seeking low glycemic loads or endurance athletes seeking long term energy… There may be advantages to a high fat diet.

But what about the non diabetic, average weight, moderately athletic (or non-athletic) population? There seem to be advantages with respect to blood cholesterol. Diets high in unsaturated fats may help lower LDL and triglycerides. Anything else? Or put conversely, are there any health advantages to a high carb diet?

Going extremely high carb or high fat might not be practical for most people. So for a product like Soylent, what should be the ratio of carbs to (unsaturated) fat for optimum health? I think that’s the key question here…


High fat diets also lower inflammation and increase insulin sensitivity.

Not that I am aware of.


I think the weight of evidences suggest that the main factor is that protein is most satiating.

True, carbs - especially simpler carbs - are less satiating than fat, but protein seems to be the big hitter on satiation. This is likely the dominant reason why many earlier low-fat vs. low-carb studies appeared to show a metabolic benefit to low-carb; people are terrible at controlling their food intake, throwing the metabolic calculations off. Low-carb people, who are eating high-fat, end up consuming more protein… which is most satiating… resulting in lower calorie intakes. Low-fat diets tend to lead to high-carb, but not so often high protein… so people eat more.

Another nice things about the study in the OP: it carefully maintained constant protein intakes, as well as total calorie intake, so it’s not a confounding factor.

Such a good study.

That is, in fact, a very tough question to answer.

In reality, I think we’re finding that a lot of the advantages of moving a low-carb diet - such as increased insulin sensitivity - are really mostly an advantage when switching to low-carb after having spent years overeating in a normal or high-carb diet.

In other words, if your insulin sensitivity is reduced because you’ve been getting no exercise, and have been overeating, and have been sugar-bombing yourself between meals with sugary snacks and drinks… well, you’ll get a nice benefit from switching to low-carb.

You’d also increase your insulin sensitivity by getting some exercise.

You’d also increase your insulin sensitivity by just moving to a calorie deficit for a while.

If you’ve burned out your insulin sensitivity, these are all good ideas!

But, what if you’ve been eating a healthy (or even slightly low) number of calories? If you’re already lean and healthy, as many athletes are who actually have use for all those calories they eat?

Well, for them, there may well be no health benefit from low-carb, whatsoever. Maybe if there’s no excess-carb “damage” to heal, then going low-carb doesn’t help at all.

(And you totally can gain fat eating low-carb, if you keep overeating.)

So, if you’re at healthy weight, and your metabolism is healthy, and you’re not over-consuming, your body may run totally fine on many wildly different mixes of macronutrients. That seems to be what the science suggests: get your protein, get your essential fatty acids, and then you can largely fill up the rest of your calorie count with almost any mix you like. Until you start over-consuming, the body seems to do a really good job of taking care of things.


Well then, it sounds like in terms of studying satiety then we need to study all 3 macros and their impacts…


That might depend on the athlete… For short/medium distance athletes who need peak performance for less than 90-120 minutes perhaps a high carb diet is fine (so long as all those calories are burned on physical activity). For long distance athletes (2+ hours) there may be advantages to moving ones metabolism towards lipids. What I’m having a hard time figuring out however is how much I wish my metabolism skewed towards lipids and what kind of macro profile I need in my diet to get there.


That’s true, for ultra-endurance athletes, there may well be a performance benefit… but I meant a health benefit. In general, athletes use up their carbs, so there are no excess carbs, so they don’t suffer the ills of excess carbs.

The lastest AARR has a good article covering a range of studies on fat adaptation for atheletes that basically concludes that there may be a benefit above the range you discuss, but that sprint performance is definitely impacted (negatively). Top-class marathoners are still using a kick during their runs, and the top marathoners are typically big carb eaters… but as you get further beyond the 130 or 140 minute run time of a marathon, it may be better to sacrifice the ability to put on a sprint in favor of being able to maintain a high pace for a longer period of time.

Utlramarathoners have 6, 12, and 24 hour run-time events… but even those have stations every 12 to 22 miles apart where runners can eat.

It’s worth noting that athletic performance when fat-adapted varies very widely by individual - it can be worth experimenting on yourself to see exactly how much you’re impacted for endurance and for sprinting performance when fully fat-adapated… just make sure you’re making objective measurements, because how you “feel” running may vary, especially if you took a long time getting fat adapted. You may feel like it’s not helping when it is, or you may feel like you’re coping better while you actually have a worse time!


I have been experimenting… I’m about a 3:30-3:45 marathoner and the most I’ve ever tried is a 50% fat, 30% carb ratio. While I can’t say for sure about long distance endurance, I know for a fact that my sprinting capability is impacted. My sprinting (or even a 5k distance) certainly does better with a good dose of simple carbs.

This is wandering OT a bit, but I have experimented in the marathon… rather than carbo loading with the usual pasta feed the night before I’ve kept to Soylent (usually straight 1.5 so 40% fat) for the days prior and won’t take anything but water for the first 13-16 miles after which I’ll start taking in the sugar (gatorade, gu etc)… What’s interesting in my last two marathons I did fade between miles 20 and 23 but then surged in the final 3 miles so perhaps I’m either not very fat adapted, or maybe I need to kick in the sugar a little earlier… Still playing with it. I know I can’t handle sugar for 26 miles… makes me nauseous. I’m 165 pounds and can burn upwards of 4000kcal in a marathon…


You probably know this, but while the body can switch on carb burning very rapidly (immediately, even), becoming fully fat-adapted may take weeks.

You might experiment with taking the fat to an even higher level for a longer period before the race or long run.

Good luck!

Afterthought: Most people can switch to running on “all fats” in just a couple of days, but the maximum fat-using capacity may rise more slowly over time and take weeks to peak - this is relevant for your experiments. The good news is that this maximum fat-using capacity doesn’t got down immediately when you eat carbs for a few days; eating a carb meal will immediately halt the fat usage while the carbs are available, but the maximal capacity to use fats is not affected by just a day or two of consuming carbs.

That suggests an interesting progression to experiment with:

  1. Train for several weeks on a very-low-carb diet, doing endurance work at the limits of your oxidative capacity. Purpose: give your cells the impetus to ramp up maximum fat-using capacity.
  2. Carb up traditionally, as before a race.
  3. Do a long training run - something well past where you normally hit the wall.

Theory: when you do the long run, you’ll have the performance benefits of being carb-loaded during the main course of the race. Eventually, you’ll hit the wall as the carbs run out… after you push through that, your cells will be working on nearly all-fat, but will be doing so at a higher capacity because of the recent no-carb training. Best of both worlds?


1 Like