It's just totally wrong that the body "breaks down muscle as a last resort." Bodies break down muscle for energy almost as a first resort, because it reduces resting metabolism by a lot, and that's a long-term adaptation to starvation. If calories are a budget, then fat is your retirement account and muscles are your rent. If you take a pay cut at work, you cancel streaming services, drive fewer places, and eventually move into a smaller apartment before you dip into your retirement savings.
Remember every human being who now lives is the descendant of one of the 80,000 humans who survived a period of prehistoric famine so intense and so fatal, it irreversibly bottlenecked the entire population of Homo sapiens. Every human body is carrying a genetic load of adaptations to long-term famine, and that includes a lot of metabolic gun-jumping, like forward-looking reduction in muscle mass when it looks like you're not going to get enough to eat for a while.
Even in the Warsaw Ghetto Famine, fat people died of cardiac muscle depletion with substantial untouched fat deposits. Muscle depletion isn't a last resort, it's the first resort when glycogen is depleted.
In fact, it says: "Thus, after periods of starvation, the loss of body protein affects the function of important organs, and death results, even if there are still fat reserves left." That's because the body is cannibalizing muscle tissue right away, because the benefits are twofold compared to the onefold benefit of exhausting fat stores.
A lot of what we think we know about starvation comes from animal models, but the issue there is that relatively few animals are adapted to famine the way humans are, so their biochemistry isn't as tuned for absolutely last-stand, no-macronutrient survival as human bodies are, or as tuned for hair-trigger lipogenesis, either.
If famine wasn't such an evolutionary bottleneck probably more mammals would be like PEPCK+ supermice.
"And how do people lose fat without losing all their muscle first if this is the case?"
Well, they don't. People on diets typically lose more muscle than they do fat. Most nutritionists think that if you can keep the loss to 60% fat/40% muscle, you're doing really well but most people on caloric restriction go the other way and we've known that since the 70's. There's a cost function coefficient at work here, clearly; all of the Warsaw Ghetto victims did lose substantial fat reserves as they starved. But they lost muscle first, as nearly everyone does after they exhaust glycogen, and they continued to until they died from it. But the coefficient of loss, fat vs. muscle, will tend to be different for different people based on competing hormonal signals and your differing sensitivity to the signals. Muscle and fat cells compete to draw blood calories for their purposes, and they compete to donate calories, too.
It says that after this though: Starvation ensues when the fat reserves are completely exhausted and protein is the only fuel source available to the body. Thus, after periods of starvation, the loss of body protein affects the function of important organs, and death results, even if there are still fat reserves left.
So I'm very confused :) Apparently starvation is when all fat is exhausted, and protein is the only fuel source left. But also, you can die after starvation even though you still have fat. Wat?
People on diets typically lose more muscle than they lose fat: What? This also contradicts my own experience and everything I've ever heard. Most people think 25% lean loss is normal and that 40% from GLP-1 drugs is terribly bad.
It must be (I don't think normal weight loss or dieting is likely to harm anyone's heart) but the metabolic logic is more of a matter of a weighted cost function (it becomes more costly to the body to lose muscle as it loses muscle) than the hard logic of "WHILE starving, GOTO burn fat, IF fat == 0 GOTO burn muscle."
And I think one of the differences between fat people and thin people is that they have different coefficients for these functions - fat people might be people who respond to starvation by preferentially losing muscle over fat. In fact we know that if you've lost significant weight in the past, and are losing now, those cost functions in your body will have shifted so that you're even more likely to lose muscle before fat.
Yea I think it's more of a gradient, like reverse nutrient partitioning. Maybe the best case is 95% fat/5% lean loss, and the worst case the other way around. Thus it makes a huge difference how you are "starving." E.g. eating nothing at all is probably not that bad, but eating 100g carbs and otherwise nothing seems to be super bad, I remember reading a study about that. Basically, a certain threshold of carbs seems to prevent fat adaptation and then you starve real bad, whereas people who truly eat nothing, or just oil, can live off their fat somewhat comfortably.
I came here from Eliezer Yudkowsky retweeting you, and don't usually read your blog, but let me report anecdotal evidence relevant to this post:
I've been taking 1.0mg/week of semaglutide for around 3 months now, and I've lost 30lbs (240lbs to 210lbs, at 5'10 height) during those 3 months, all of it in fat, which I know because my lifts in the gym have actually **improved** during this time, and I was not a beginner in the gym when I started (280lbs bench press 1RM, 365lbs squat, 495lbs deadlift). The subjective experience of the drug is not "I stop eating halfway through a meal" for me, it's much more of a "all food cravings are lowered, and I don't go to sleep hungry" effect. For me the drug has in fact been miraculous, and has allowed me to actually stick to my planned diet completely effortlessly.
I've been taking Rybelsus (pill form of semaglutide) since January, starting with 3mg and ramping up to 14mg. I've lost 7kg in this time. I'm lifting regularly at the gym and I have not had to lower the weights during this time. But sometimes I do get the "my wife can eat more than I can" effect (while normally I would have easily eaten 50% more than her).
Nice. What does your diet look like (maybe before/after the pill)? My theory is that semaglutide will produce weight loss, but if you eat a dorito-style SAD diet it'll be lots of muscle, whereas if you eat a healthy diet (whatever that means lol) it'll be mostly fat.
No big changes in ingredients before/after the pill. I still eat everything, including McDonald's 1-2 times a month. But sometimes I do get very strong "reactions" to food, which i didn't use to get. Sometimes I simply can't eat a food at a time, but a couple of hours later it might be fine.
I eat lots of home cooked food, so I guess it's healthier than average.
Nice, thanks for the input. I've heard from a handful of people now and all have had pretty different experiences, though it does work "miraculously" for many, like you. Others it did absolutely nothing. So some other factor is definitely at work.
You don't have diabetes, right? The SURMOUNT-2 trial was about obese diabetics. Weight loss is more difficult for them. If you look at this other trial, there is way more weight loss: https://www.nejm.org/doi/full/10.1056/NEJMoa2206038
At 72 weeks and 5mg weight went down 15%, for 10mg it went down 19.5%, for 15mg it went down 20.9%. Also, respectively 85%, 89%, and 91% had weight reduction of 5%, and 50% and 57% respectively in the 10-mg and 15-mg groups had weight reduction of 20% or more.
It still doesn't make a ton of sense to use yourself as a comparison group. Especially because this group had a lot of people with prediabetes and over 60% female, two things that would make it harder to lose weight than for you, a male with seemingly normal insulin sensitivity.
If you want to see a rough distribution of how many people lost how much weight, it's all there in Figure 1.
Another tidbit you may be interested in: "The ratio of total fat mass to total lean mass decreased more with tirzepatide (from 0.93 at baseline to 0.70 at week 72) than with placebo (from 0.95 to 0.88)." I don't think this means the placebo group is in stage III starvation.
Can't reply to your other comment (max comment depth reached?)
Yea, 30% having no significant effect is a somewhat large group, I think. Points to something else being at play, e.g. actual diet composition or maybe genetics.
Sure they might have otherwise gained weight, but why?
The results in this one definitely seem more promising. Still, the most successful group lost less weight than I did in 2x the time. Of course, that's an average - would be cool to see the result distributions mapped out. I tend to be a hyper responder in both directions with big swings. So you're right I'm probably not a great comparison group - except for myself ;)
The figure still doesn't break out what I'd like to see; it still only seems to have averages per dosage group. I'd want to see, for any given treatment group, how many people lost how much, broken out into finer detail. They have a 5, 10, 15, 20, 25% breakout, which I suppose is good. 15-30% (depending on dosage) had 25%, which is really good. But a large % didn't have any significant effect. So there's definitely still a factor at play.
You say you lost 21% in half the time (36 weeks). Look at chart B and you'll see that the 15mg group had an average weight loss of around that much at maybe 42 weeks and at 36 weeks they had average weight loss of maybe 18%. Considering the differences that we know exist between you and the control group, I'd say these results are on par with what you have experienced.
Also, where are you getting the idea that a large % didn't have any significant effect? In your blog post you define 15% as significant. Over 70% reached that threshold. I don't think I'd consider the 30% who didn't to be large, especially when most of them still lost some weight. These 30% are perhaps the ones that otherwise would have been gaining weight for all we know.
Interesting idea. For me, weight gain largely comes from eating at poor times, like before bed, as well as when I'm not hungry. I spent most of two decades overeating because of anxiety. Managing anxiety better, trading away late meals for late exercise, and reducing my overall calorie intake have been the keys in losing weight so far. I suspect that semaglutide would do nothing for most or possibly any of that.
I'd like to know what your conception of CICO is. I mean the CICO that you cite in phrases like "I guess that's your brain on CICO", not the concept of reducing calorie intake to being consistently several hundred kcals below calorie usage (or higher).
I think I've eaten out of boredom before, but fixing that didn't make me lose any (or much) weight. Anxiety is a whole other can of worms, I guess :) How have you gotten your anxiety under control, meditation? Therapy?
I'm still reading through while I prepare for my cardio workout. Two things:
An observation: The two layers of deficit, internal and external, makes sense to me based on what I've experienced. Hopefully it's right.
A question (an idle one): How do you, personally, pronounce "CICO"? I mean the acronym. I know it's goofy, just curious.
Edit: Another observation. The back-of-the-napkin math says the amount of calories daily that I have on hand (and belly and butt and...) to burn internally is the same, currently, as my birth year. Cool.
I manage anxiety through self-care and by paying attention to myself. I'm not always successful, but failure nowadays means I had a second helping or ate some biltong. Things are better than they were when I was younger. Caring about myself and choosing to be proactive makes a world of difference.
Nice. Yea, having a super harmless failure mode is great :) After a few years of keto, a "failure of willpower" for me meant that I bought a ketogenic energy drink or protein bar. Somehow, my brain had retrained and "cheat foods" were just different ones. I even went back and tried carbs last summer and confirmed I don't even miss them. Not that they were disgusting, just nothing I'd eat on purpose.
It's just totally wrong that the body "breaks down muscle as a last resort." Bodies break down muscle for energy almost as a first resort, because it reduces resting metabolism by a lot, and that's a long-term adaptation to starvation. If calories are a budget, then fat is your retirement account and muscles are your rent. If you take a pay cut at work, you cancel streaming services, drive fewer places, and eventually move into a smaller apartment before you dip into your retirement savings.
Remember every human being who now lives is the descendant of one of the 80,000 humans who survived a period of prehistoric famine so intense and so fatal, it irreversibly bottlenecked the entire population of Homo sapiens. Every human body is carrying a genetic load of adaptations to long-term famine, and that includes a lot of metabolic gun-jumping, like forward-looking reduction in muscle mass when it looks like you're not going to get enough to eat for a while.
Even in the Warsaw Ghetto Famine, fat people died of cardiac muscle depletion with substantial untouched fat deposits. Muscle depletion isn't a last resort, it's the first resort when glycogen is depleted.
That's not what the Wikipedia page on starvation says: https://en.wikipedia.org/wiki/Starvation
And how do people lose fat without losing all their muscle first if this is the case?
In fact, it says: "Thus, after periods of starvation, the loss of body protein affects the function of important organs, and death results, even if there are still fat reserves left." That's because the body is cannibalizing muscle tissue right away, because the benefits are twofold compared to the onefold benefit of exhausting fat stores.
A lot of what we think we know about starvation comes from animal models, but the issue there is that relatively few animals are adapted to famine the way humans are, so their biochemistry isn't as tuned for absolutely last-stand, no-macronutrient survival as human bodies are, or as tuned for hair-trigger lipogenesis, either.
If famine wasn't such an evolutionary bottleneck probably more mammals would be like PEPCK+ supermice.
"And how do people lose fat without losing all their muscle first if this is the case?"
Well, they don't. People on diets typically lose more muscle than they do fat. Most nutritionists think that if you can keep the loss to 60% fat/40% muscle, you're doing really well but most people on caloric restriction go the other way and we've known that since the 70's. There's a cost function coefficient at work here, clearly; all of the Warsaw Ghetto victims did lose substantial fat reserves as they starved. But they lost muscle first, as nearly everyone does after they exhaust glycogen, and they continued to until they died from it. But the coefficient of loss, fat vs. muscle, will tend to be different for different people based on competing hormonal signals and your differing sensitivity to the signals. Muscle and fat cells compete to draw blood calories for their purposes, and they compete to donate calories, too.
It says that after this though: Starvation ensues when the fat reserves are completely exhausted and protein is the only fuel source available to the body. Thus, after periods of starvation, the loss of body protein affects the function of important organs, and death results, even if there are still fat reserves left.
So I'm very confused :) Apparently starvation is when all fat is exhausted, and protein is the only fuel source left. But also, you can die after starvation even though you still have fat. Wat?
People on diets typically lose more muscle than they lose fat: What? This also contradicts my own experience and everything I've ever heard. Most people think 25% lean loss is normal and that 40% from GLP-1 drugs is terribly bad.
Is it just skeletal muscle that is lost first? Wouldn't nonskeletal muscle like the heart be preserved more carefully?
It must be (I don't think normal weight loss or dieting is likely to harm anyone's heart) but the metabolic logic is more of a matter of a weighted cost function (it becomes more costly to the body to lose muscle as it loses muscle) than the hard logic of "WHILE starving, GOTO burn fat, IF fat == 0 GOTO burn muscle."
And I think one of the differences between fat people and thin people is that they have different coefficients for these functions - fat people might be people who respond to starvation by preferentially losing muscle over fat. In fact we know that if you've lost significant weight in the past, and are losing now, those cost functions in your body will have shifted so that you're even more likely to lose muscle before fat.
Yea I think it's more of a gradient, like reverse nutrient partitioning. Maybe the best case is 95% fat/5% lean loss, and the worst case the other way around. Thus it makes a huge difference how you are "starving." E.g. eating nothing at all is probably not that bad, but eating 100g carbs and otherwise nothing seems to be super bad, I remember reading a study about that. Basically, a certain threshold of carbs seems to prevent fat adaptation and then you starve real bad, whereas people who truly eat nothing, or just oil, can live off their fat somewhat comfortably.
It's also not like a switch where the body depletes one source and then moves on to the next.
Yea, like many things, it seems to be more of a grey scale than a switch.
I came here from Eliezer Yudkowsky retweeting you, and don't usually read your blog, but let me report anecdotal evidence relevant to this post:
I've been taking 1.0mg/week of semaglutide for around 3 months now, and I've lost 30lbs (240lbs to 210lbs, at 5'10 height) during those 3 months, all of it in fat, which I know because my lifts in the gym have actually **improved** during this time, and I was not a beginner in the gym when I started (280lbs bench press 1RM, 365lbs squat, 495lbs deadlift). The subjective experience of the drug is not "I stop eating halfway through a meal" for me, it's much more of a "all food cravings are lowered, and I don't go to sleep hungry" effect. For me the drug has in fact been miraculous, and has allowed me to actually stick to my planned diet completely effortlessly.
I've been taking Rybelsus (pill form of semaglutide) since January, starting with 3mg and ramping up to 14mg. I've lost 7kg in this time. I'm lifting regularly at the gym and I have not had to lower the weights during this time. But sometimes I do get the "my wife can eat more than I can" effect (while normally I would have easily eaten 50% more than her).
Nice. What does your diet look like (maybe before/after the pill)? My theory is that semaglutide will produce weight loss, but if you eat a dorito-style SAD diet it'll be lots of muscle, whereas if you eat a healthy diet (whatever that means lol) it'll be mostly fat.
No big changes in ingredients before/after the pill. I still eat everything, including McDonald's 1-2 times a month. But sometimes I do get very strong "reactions" to food, which i didn't use to get. Sometimes I simply can't eat a food at a time, but a couple of hours later it might be fine.
I eat lots of home cooked food, so I guess it's healthier than average.
Nice, thanks for the input. I've heard from a handful of people now and all have had pretty different experiences, though it does work "miraculously" for many, like you. Others it did absolutely nothing. So some other factor is definitely at work.
I squat heavy once or twice a week, so I'd likely notice losing muscle mass pretty quickly. A non-lifter, though, yeah.
"the average overweight or obese person would probably gladly take a 15%, or even a 5%, weight reduction."
Depends on how the TTRPG rules read. Many bonuses don't stack, alas.
"So does my Ozempic bonus multiply with the keto reduction, or do I add up the percentages..?"
You don't have diabetes, right? The SURMOUNT-2 trial was about obese diabetics. Weight loss is more difficult for them. If you look at this other trial, there is way more weight loss: https://www.nejm.org/doi/full/10.1056/NEJMoa2206038
At 72 weeks and 5mg weight went down 15%, for 10mg it went down 19.5%, for 15mg it went down 20.9%. Also, respectively 85%, 89%, and 91% had weight reduction of 5%, and 50% and 57% respectively in the 10-mg and 15-mg groups had weight reduction of 20% or more.
It still doesn't make a ton of sense to use yourself as a comparison group. Especially because this group had a lot of people with prediabetes and over 60% female, two things that would make it harder to lose weight than for you, a male with seemingly normal insulin sensitivity.
If you want to see a rough distribution of how many people lost how much weight, it's all there in Figure 1.
Another tidbit you may be interested in: "The ratio of total fat mass to total lean mass decreased more with tirzepatide (from 0.93 at baseline to 0.70 at week 72) than with placebo (from 0.95 to 0.88)." I don't think this means the placebo group is in stage III starvation.
Can't reply to your other comment (max comment depth reached?)
Yea, 30% having no significant effect is a somewhat large group, I think. Points to something else being at play, e.g. actual diet composition or maybe genetics.
Sure they might have otherwise gained weight, but why?
Right, I don't have diabetes.
The results in this one definitely seem more promising. Still, the most successful group lost less weight than I did in 2x the time. Of course, that's an average - would be cool to see the result distributions mapped out. I tend to be a hyper responder in both directions with big swings. So you're right I'm probably not a great comparison group - except for myself ;)
The figure still doesn't break out what I'd like to see; it still only seems to have averages per dosage group. I'd want to see, for any given treatment group, how many people lost how much, broken out into finer detail. They have a 5, 10, 15, 20, 25% breakout, which I suppose is good. 15-30% (depending on dosage) had 25%, which is really good. But a large % didn't have any significant effect. So there's definitely still a factor at play.
You say you lost 21% in half the time (36 weeks). Look at chart B and you'll see that the 15mg group had an average weight loss of around that much at maybe 42 weeks and at 36 weeks they had average weight loss of maybe 18%. Considering the differences that we know exist between you and the control group, I'd say these results are on par with what you have experienced.
Also, where are you getting the idea that a large % didn't have any significant effect? In your blog post you define 15% as significant. Over 70% reached that threshold. I don't think I'd consider the 30% who didn't to be large, especially when most of them still lost some weight. These 30% are perhaps the ones that otherwise would have been gaining weight for all we know.
Interesting idea. For me, weight gain largely comes from eating at poor times, like before bed, as well as when I'm not hungry. I spent most of two decades overeating because of anxiety. Managing anxiety better, trading away late meals for late exercise, and reducing my overall calorie intake have been the keys in losing weight so far. I suspect that semaglutide would do nothing for most or possibly any of that.
I'd like to know what your conception of CICO is. I mean the CICO that you cite in phrases like "I guess that's your brain on CICO", not the concept of reducing calorie intake to being consistently several hundred kcals below calorie usage (or higher).
I think I've eaten out of boredom before, but fixing that didn't make me lose any (or much) weight. Anxiety is a whole other can of worms, I guess :) How have you gotten your anxiety under control, meditation? Therapy?
I wrote about CICO before: https://exfatloss.substack.com/p/a-tale-of-two-caloric-deficits. I should've linked it, I'll do that now :) Basically I think it's a useless tautology that prevents people from thinking straight.
I'm still reading through while I prepare for my cardio workout. Two things:
An observation: The two layers of deficit, internal and external, makes sense to me based on what I've experienced. Hopefully it's right.
A question (an idle one): How do you, personally, pronounce "CICO"? I mean the acronym. I know it's goofy, just curious.
Edit: Another observation. The back-of-the-napkin math says the amount of calories daily that I have on hand (and belly and butt and...) to burn internally is the same, currently, as my birth year. Cool.
Haha that means you're not over 1000 years old and not over 5000 then? ;)
I pronounce it "kee-koh" but no clue how other people say it or if that's correct, lol.
Confusingly, this sounds almost exactly like "keto" and so people get confused when I rant about keko and proudly annouce I'm in year 7 of keto ;)
I'll read over that. Thanks.
I manage anxiety through self-care and by paying attention to myself. I'm not always successful, but failure nowadays means I had a second helping or ate some biltong. Things are better than they were when I was younger. Caring about myself and choosing to be proactive makes a world of difference.
Nice. Yea, having a super harmless failure mode is great :) After a few years of keto, a "failure of willpower" for me meant that I bought a ketogenic energy drink or protein bar. Somehow, my brain had retrained and "cheat foods" were just different ones. I even went back and tried carbs last summer and confirmed I don't even miss them. Not that they were disgusting, just nothing I'd eat on purpose.