How Fasting & Caloric Restriction Impact Health | Dr. Satchin Panda & Dr. Andrew Huberman
00:00:00.000 | So there is a famous experiment that was published last year by Joe Takahashi's lab, and it came
00:00:08.660 | out in Science, and that relates to caloric restriction. And we kind of started with this
00:00:16.120 | idea, we started discussing that the rat experiments were done with caloric restriction, and researchers
00:00:22.360 | gave reduced calorie consumption by 20 percent or 30 percent and gave that food, the rats,
00:00:29.340 | and then subsequently mice, and they all lived longer. What is interesting is, in all those
00:00:38.200 | experiments, the researchers came and gave this bolus of food at one time, whereas the
00:00:44.300 | ad libitum fed mice or rats, they had access to food all the time. So they were eating
00:00:48.700 | all the time, and then these rats were given 20 percent less. And what happens is, these
00:00:56.000 | mice or rats, they're not going to take that less food, which is rationed now, and just
00:01:01.800 | eat a little bit of lunch and then snack after three hours or snack after three hours. They
00:01:06.300 | gobble up all that food. Within two to three hours, maximum four hours, food is gone.
00:01:12.140 | So they're sort of on the OMAD diet, the one meal a day.
00:01:15.320 | Yeah. They're almost like in one meal a day, three to four hours, food is gone. Or you
00:01:21.120 | can say they're on four hours eating or feeding and 20 hours fasting. So then the question
00:01:28.940 | became, well, the benefit of caloric restriction, as we know, is it due to reduced calorie or
00:01:38.060 | time-restricted feeding? There is a timing component to it, that they're eating all of
00:01:42.220 | that within three to four hours, and then there is a long fasting. And this is a difficult
00:01:47.340 | question to answer, because now you have to ask these poor grad students or technicians
00:01:52.820 | to come and split that food into eight or 10 or 15 different small portions and then
00:01:59.460 | give them to mice in every two hours. So Joe Takahashi, who actually published the first
00:02:06.340 | paper in 2017 showing that most caloric restrictions, I mean, he used the protocol that was used
00:02:13.220 | by caloric restriction field, it actually creates a condition of time restriction. So
00:02:20.220 | he showed that, and then he went back and worked with engineers to come up with a smart
00:02:25.460 | case where he could actually tell, he could program how much food is given to mice at
00:02:32.660 | what time of the day or night, completely programmed. So then he took this, for example,
00:02:39.700 | let's say the ad libitum fed mice eats five grams of chow in a day. And if you want to
00:02:46.500 | reduce calories by 20%, then the CR mouse should get four grams of food. And he divided
00:02:54.620 | this into nine or 10 meals, and then gave them in every 90 minutes. So in this case,
00:03:03.060 | they are eating small meals throughout day and night, so there is no fasting. So you
00:03:08.900 | can say that, well, this mouse actually is not getting into fasting because in every
00:03:13.620 | few hours is getting some food. And then he measured how long the mouse is going to live.
00:03:21.620 | And he used account, means this is a very standard protocol, people count how many mice
00:03:27.660 | are dying on which day and then examine them to see whether they died because there was
00:03:32.700 | an accident or they actually, there was a natural cause. And then they calculate at
00:03:38.380 | the end, what is the half life? So 50% survival, because that's on an average, that's a good
00:03:47.220 | indicator because if there is an outlier that will live for a long time, then that can skew.
00:03:52.500 | So what was interesting was the ad libitum fed mice, of course, they live certain number
00:03:57.220 | of days, and then this caloric restricted mice that never got into super fasting, but
00:04:03.940 | kind of eating, snacking throughout day and night, that also lived 10% extra, 10% longer.
00:04:10.700 | So that means caloric restriction extended lifespan by 10%.
00:04:15.460 | - I've wondered about this because recently, you know, there's been, there were a bunch
00:04:20.440 | of news headlines about intermittent fasting. And frankly, I was frustrated if you looked
00:04:26.020 | at one major news outlet, they would say time restricted feeding affords no additional benefit
00:04:32.940 | beyond caloric restriction for weight loss. Then another popular press venue, let's call
00:04:40.300 | it that, same study described as time restricted feeding doesn't work. Right. And then another
00:04:48.340 | one, maybe someplace even more extreme, you know, time restricted feeding only beneficial
00:04:55.940 | because of caloric restriction or something like that. So what you've essentially got
00:04:59.620 | are three different interpretations of the same data, all of which are, well, two of
00:05:04.660 | which are true, one of which is false in my opinion. But what I think people take away
00:05:08.940 | from that is, oh, time restricted feeding isn't valuable, which is not the case. I think
00:05:13.940 | for many people, it's a convenient way to eat because at least for people like me, it's
00:05:18.060 | simpler to designate between portions of my day when I'm eating and portions of my day
00:05:23.140 | when I'm not eating, as opposed to portion control. For other people, portion control
00:05:26.660 | can work. But all of that is related to either maintenance or loss of weight. None of it
00:05:34.060 | deals with the potential health benefits independent of weight loss. And so I think that if we
00:05:41.420 | can segment those out, obviously in humans, it's hard to know if a given treatment or
00:05:48.500 | experiment is extending life because you don't really know how long people would live anyway.
00:05:51.940 | Right. Whereas with mice, you have some sense of when the mortality was likely to occur.
00:05:57.060 | So what can we say about time restricted feeding and longevity in terms of biomarkers or in
00:06:03.340 | terms of any other indication that people who start and stop their feeding window at
00:06:07.980 | a consistent time, somewhere between eight and 12 hours per 24-hour cycle, are tilting
00:06:14.340 | the scales towards living longer as opposed to living shorter?
00:06:18.580 | This example of this news article that you mentioned is really interesting because that
00:06:22.540 | relates to Joe Takahashi's study, because I described that if you split calories and
00:06:28.420 | eat throughout the day, throughout day and night, then the mice live 10% extra. But if
00:06:33.900 | you now give mouse the same calorie restricted diet and feed them during daytime, whether
00:06:41.540 | within 12 hours or two hours, then the mice live 10% extra.
00:06:46.380 | Beyond that?
00:06:47.380 | Yeah.
00:06:48.380 | So 20% longer.
00:06:49.380 | 20%.
00:06:50.380 | So, okay. So let me make sure I understand so that I make sure I understand. If you take
00:06:55.460 | a certain number of calories and you distribute them throughout the 24-hour cycle, it's caloric
00:07:00.660 | restriction, the mice will live 10% longer. If you, however, restrict that to the active
00:07:06.300 | cycle, so for humans, the daytime, then they live 20% longer.
00:07:14.420 | 20% longer.
00:07:15.420 | So it's not just total caloric intake, meaning it's not just important to be sub-maintenance
00:07:20.620 | in calories for sake of longevity. It also is important as to when in the 24-hour cycle
00:07:26.700 | you eat those calories. Do I have that right?
00:07:28.940 | So now, that's still, the story is not over because these mice are fed during daytime
00:07:34.980 | when they're not supposed to eat.
00:07:36.420 | That's right. So for us, it would be the equivalent of being on the night shift and only eating
00:07:39.380 | at night, but a sub-caloric, sub-maintenance calorie diet, I guess is the right way to
00:07:43.940 | say it.
00:07:44.940 | But when you fed mice during nighttime when they're supposed to eat and they're getting
00:07:50.300 | the same number of calories within 12 hours or two hours, then the mice live 35% longer
00:07:56.620 | than they control.
00:07:57.620 | 35% longer. So scale to human lifespan, which we don't know, but 35% longer would mean that,
00:08:05.940 | and again, no one knows, but humans now, what is the average mortality in the United States,
00:08:11.940 | somewhere around 80?
00:08:12.940 | Yeah, so it's around 80. It used to be 80, now it's reduced a little bit because of COVID,
00:08:18.220 | but let's take 80.
00:08:19.220 | Okay, so people are then now living somewhere between 25 and 35 years longer. I'm putting
00:08:24.500 | some error bars on that.
00:08:25.500 | Yeah, yeah, yeah.
00:08:26.500 | Amazing.
00:08:27.500 | So that was really profound, but now you pointed out biomarker and other stuff. So now, if
00:08:35.820 | you look at any given time within that experiment, and actually Joe went back and had a separate
00:08:42.540 | cohort of mice, very similar, and so that he could take tissue samples. And of course,
00:08:47.100 | in this case, you have to sacrifice the mouse. And he looked for, he did a lot of molecular
00:08:53.620 | analysis with known markers, for example, hemoglobin A1C equivalent, or glucose control,
00:09:00.660 | cholesterol, all this stuff. He could not find anything that predicted the benefit of
00:09:07.940 | caloric restriction. So that means in this experiment, whatever we know so far, the predictor
00:09:14.380 | of longevity, none of them could predict whether this CR only mouse, which ate throughout day
00:09:23.140 | and night, that mouse is going to live less than the night fed mouse that was going to
00:09:29.180 | live 25% extra.
00:09:31.700 | Does that mean that there are biomarkers related to longevity that we just haven't discovered
00:09:35.580 | yet?
00:09:36.580 | Yeah, so that's exactly. So that means whatever we know so far about biomarkers, he could
00:09:44.020 | not use to predict. Maybe there was a lot of noise, maybe he had to use more number
00:09:49.820 | of mice to get that, because biomarkers are not going to predict in every instance that
00:09:55.540 | there is some error. What is also very interesting is, if you look at the body weight and body
00:10:01.740 | composition of all these mice, there is no difference in body weight and body composition.
00:10:07.400 | Across all these different groups?
00:10:08.400 | Across all these groups.
00:10:09.400 | So it doesn't matter when they ate, provided they were sub-maintenance calorie intake,
00:10:15.900 | fewer calories than is required to maintain their weight, didn't matter what pattern of
00:10:20.620 | eating, they were the same weight. So that in many ways seems to mimic the human studies
00:10:24.900 | where they say, "Look, it doesn't really matter whether or not you use caloric restriction,
00:10:28.980 | or you start your feeding window in the morning, or start your feeding window in the evening,
00:10:33.940 | or you portion control for sake of weight loss."
00:10:36.900 | Because they're taking a snapshot of that. And then another thing with the human study
00:10:41.300 | that we are referring to here, in that human study, people are actually already eating
00:10:49.380 | within 10 hours window, habitually, when they selected these people to have them enroll
00:10:56.740 | in the study. So they were already eating for 10 hours and fasting for 14 hours. All
00:11:03.260 | participants had to reduce their caloric intake, and they reduced by almost 25%. The CR group
00:11:10.740 | continued with 10 hour sitting window, and the CR plus time-restricted group had to eat
00:11:16.500 | the same number of calories within eight hours.
00:11:18.660 | - So it's just a two-hour difference.
00:11:19.980 | - It's just a two hours difference.
00:11:21.460 | - Okay, so that people, I just want to make sure people can understand. So in this human
00:11:25.340 | study, which is the one that I felt that the popular press venues, all except one venue,
00:11:30.860 | got either semi-wrong or badly wrong in terms of their conclusion. That was my interpretation
00:11:35.620 | anyway, was that either people came into the study eating basically in a 10 hour feeding
00:11:41.300 | window, which goes back to my first question, which is that most people are not eating in
00:11:44.180 | the middle of the night, or if they're on shift work and they are, then they're sleeping
00:11:47.660 | during the day anyway. So they're eating in a 10 to 12 hour feeding window anyway. So
00:11:50.940 | you're saying they either did caloric restriction, portion control within the 10 hour window,
00:11:55.700 | or another group within the study ate sub-maintenance calories. So caloric restriction, CR, as we're
00:12:02.980 | calling it, the acronym CR, but restricted that to an eight hour feeding window. And
00:12:07.260 | they didn't see any difference in terms of weight loss. But it's not all that surprising,
00:12:11.420 | right? I mean, if it's just a two hour difference.
00:12:13.460 | - Yeah, exactly. So we have done that experiment in mice and we don't see difference in not
00:12:18.940 | only weight loss, many other markers. And I was telling you about this Joe Takahashi's
00:12:24.460 | paper where I told you that he allowed these mice to eat within two hours or 12 hours,
00:12:31.380 | sub-caloric diet.
00:12:33.540 | - Two or 12?
00:12:34.540 | - Two or 12.
00:12:35.540 | - Yeah, that's dramatic.
00:12:36.540 | - But still, he did not see change in longevity even within those two. So that means when
00:12:43.540 | you do caloric restriction and then at least for mouse, and you are within 12 hours window,
00:12:51.860 | that is giving the mice the best benefit, the optimum benefit. And two, three, or five
00:12:58.540 | or 12 for a mouse doesn't matter, at least for longevity.