Dr. Peter Attia, Responds to IF Study Published in JAMA

“A really interesting study came out this week looking at time-restricted feeding and comparing it to a continuous meal structure. It was published in a journal called JAMA, the Journal of the American Medical Association, and it has garnered a lot of attention, including a piece written in the New York times by Anahad O’Connor, who’s a great science reporter. I thought I would take an opportunity to explain the implications of this study. What does it tell us about fasting? What does it not tell us about fasting? That sort of thing. And I’d like to do that here in this very informal video. My team and I are working on a much more comprehensive evaluation of this study through the lens of the nuances of design and such.

So if you’re really interested in this topic, and in particular what this paper says about it, I would encourage you highly to read it, but I want to talk about a couple of macro principles.

So first of all, what did this study do? This study set out to compare two groups of overweight people, and one of those groups, which was the control group, was basically told to eat three square meals a day during a set of predefined time intervals that roughly corresponded to breakfast, lunch, and dinner—they were permitted to snack as well. If they chose to, the other group was put onto a time-restricted eating window that was fixed between noon and 8:00 PM. So this is what most of you would think of as a 16:8 feeding window. So for 16 hours, from 8:00 PM to noon the next day, there would be no eating. And then from noon to eight, you would eat. Okay. So the question—because remember, every study begins with a question—the question being posed by the investigators was would the group on the time-restricted eating side of the equation lose more weight than the control group? It’s a fair question to ask, right? Many people approach fasting through the lens of, “Hey, one of the benefits of this might be weight loss.” So what happened? The study was run for a reasonable period of time, almost three months, about 12 weeks. And during that period of time, there was really no statistically significant difference in weight loss between the groups, both of the groups actually lost a trivial amount of weight. I would consider it a clinically irrelevant amount of weight, something between the order of about one and two pounds. So both of them had a statistically significant reduction in weight relative to their baselines, but between the groups, there was no difference in weight loss.

Now it would be tempting to just stop there and say, well, the implication of this is that time restricted eating offers no weight loss advantage at least to overweight people over a period of 12 weeks. And that might be the case, but I think there are a couple of things that warrant some consideration before we jump to that conclusion immediately.

“Basically there are three levers that you can always pull with respect to your nutrition. What you eat, how much you eat, and when you eat.”

The first is taking a step back—let’s remember the framework that I’ve presented previously. If you haven’t seen it, I’ll go through it again briefly. Basically there are three levers that you can always pull with respect to your nutrition. I think of them as levers, you can think of them as three variables you can manipulate. The first is dietary restriction. That is what you are eating. So if you choose to restrict sugar or carbohydrates or animal products—anytime you choose to restrict something in the diet—that is called dietary restriction. And by the way, that can come without any other form of restriction, right? A ketogenic diet or a vegan diet are two extreme examples of dietary restriction that don’t often come with an implicit amount of caloric or time restriction, which are our other two variables. Of course, I’ve just explained what those other two variables are. The next one is caloric restriction. This says, I am actually going to restrict how much I eat of something, though that does not imply you’re necessarily eating less of something or eating it during a certain period of time. And of course, then finally, the third lever is time restriction, which is when you eat. Now, I want you to think about these three levers as three things you get to pull every day, you get to decide how hard you want to pull them. Okay? You can even pull none of them.

You can have no dietary restriction, which means “I’m going to eat anything, any type of food I see,” you can have no caloric restriction, which means “I’m going to eat as much of it as I want,” and you can have no time restriction, which means “I’m going to eat whenever I want.” And that’s called the standard American diet. We’ve been doing a really good natural experiment for the last 50 years that’s demonstrated that the standard American diet will kill almost everybody. Very few people are immune to the standard American diet. So I think we can all agree that not pulling any of the leavers is probably not a great idea. You’re going to have to pull some of them.

“So what did this study set out to test? It didn’t provide any input on dietary restriction, it didn’t provide any caloric restriction, and it provided some time restriction.”

So what did this study set out to test? Well, it didn’t provide any input on dietary restriction. The participants were not instructed to refrain from eating anything in particular. It did not provide any caloric restriction. They were not told to refrain from any quantity of food. What they were testing was a group that had no restriction on when they ate versus a group that had some restriction on when they ate; about an extra four hours to five hours of restriction. So in other words, I would say this group, the control group, didn’t pull on any of the levers, the treatment group didn’t pull on the first one, didn’t pull on the second one and pulled a little bit on the third one. And at least in that setting, it didn’t appear to make that much of a difference.

“If one wants to not pull on the first two levers, and they want to achieve weight loss, they have to pull really hard on the time restriction lever.”

In my clinical experience, that’s not very surprising. In fact, somebody asked me about the study the day it came out on Twitter and I made the comment that doesn’t really surprise me that much because both personally and with my patients, if one wants to not pull on the first two levers, meaning they don’t really want to pull on dietary restriction or restrict or change the quality of foods, and they don’t want to pull the second lever, which is how much they eat, and they want to get some metabolic benefit, which would include weight loss, but not only weight loss, but also an improvement in metabolic parameters, they have to pull really hard on the time restriction lever. In my opinion, they generally have to restrict calories from somewhere between 20 and 22 hours a day, which effectively reduces you to one meal a day. And by the way, you can still manage to screw that up if you overcompensate and do silly things on some of the other variables, I’ll give you an example.

“If you’re willing to be gluttonous enough and eat horrible enough food, you can undermine any of the benefits.”

I had a patient who was really committed to doing one meal per day, but he somehow had it in his mind that during that meal, he had to eat an entire pizza, two liters of cream soda and a box of donuts. And guess what? He got worse. So even if you’re restricting calories for 22 or 23 hours a day, if you’re willing to be gluttonous enough and eat horrible enough food in that hour, you can undermine any of the benefits.

Now, what do we know about this study? Well, unfortunately, the lead author, Ethan Weiss, who’s been very gracious and communicated with our team about what they tried to do and not do, and his team were unable to get any total food recall. That was a big part of this study. One of the hypotheses that was being tested here was if there was going to be a benefit from time-restricted feeding. Was it going to come down to simply eating less because you were given less time to eat? Or was it going to be independent of how much you ate? You might actually have a metabolic benefit that comes from the fasting period. Unfortunately, despite their best efforts, they were not able to do that using what are called recall food frequency questionnaires. So there was no data to tell us what either group of these patients ate. There are several other limitations of the study that I will not go into here, but I do want to close with one very important point.

Many of you know about this idea of time-restricted feeding because of work that was done in mice several years ago. That’s what brought this to people’s attention probably about six years ago is when some of the seminal work was done in this space. In fact, there was a very interesting paper that was published in The Cell—I believe it was about 2014, and there were several experiments, but the gist of it was if you took a group of mice and divided them into two groups, and one group was allowed to consume really junky mouse food, which is basically very high-fat, very high-sugar mouse food. And they were allowed to consume that as much as they wanted. And then another group was allowed to consume that only in an eight hour feeding window, but not during a 16 hour feeding window, how would they differ? And the group that was allowed access to mouse junk food, 24/7 became morbidly obese and metabolically destroyed the group that was allowed access to as much mouse junk food as they wanted during only eight hours—and by the way, they ate a ton of it during that eight hours—they remain lean and healthy. There was a sister experiment to that which took obese mice and did the same thing. And lo and behold, the obese mice who had unlimited access to mouse junk food stayed sick and unhealthy, whereas the obese mice who were only allowed to eat mouse junk food for an eight hour window, actually got lean and healthy.

“A mouse will be dead of starvation in about 48 hours. Their metabolism and our metabolism have pretty much nothing in common. So for a mouse to go 16 hours without eating is a very different proposition than for a human to go 16 hours without eating.”

Here’s the catch: a mouse will typically die of starvation in about 48 hours. Let me repeat that. A mouse will be dead of starvation in about 48 hours. Their metabolism and our metabolism have pretty much nothing in common. So for a mouse to go 16 hours without eating is a very different proposition than for a human to go 16 hours without eating. I spent many hours discussing this exact question with scientists and there is no clear agreement on what the equivalent is. In other words, 16 hours of fasting in mice equals how many days in humans. I don’t think we know the answer, but it’s clearly on the order of days, whether it’s five days or seven days, we simply don’t have the biomarkers or biology to tell us, but I just want you to keep in the back of your mind that, a 16 hour fast in humans really isn’t what we would consider a fast.

“And that goes back to my original point, which is, if you’re not making any attempt to change the quality of what you eat or the quantity of what you eat and you want a significant benefit, you’re probably going to have to restrict your feeding window by more than 16 hours.”

And that goes back to my original point, which is if you’re not making any attempt to change the quality of what you eat or the quantity of what you eat and you want a significant benefit, you’re probably going to have to restrict your feeding window more than 16 hours, conversely, for 16 hours to work in my experience, patients have to make some changes in these other two variables and clinically, I think most patients find it easier to do that on the dietary restriction side. So in other words, eat a couple of meals during that eight hours, but make sure that the quality of those meals is at its finest. Again, we’re going to go into much more detail in a written format that goes through all of the nuances of the paper, talking about intention to treat analysis, all sorts of other things that some of you may or may not find interesting. But hopefully this short overview gives you a sense of how to put this study in the context of the broader infrastructure of fasting.”