I Ate 6000 Calories Per Day. Here's What Happened to My Cholesterol

I spent the last week eating 6,000 calories per day. I know, the first question everybody asks is, "What the heck were you eating?" We definitely have a gross fascination with watching people eat absurd amounts of food. But I'll get to that, I promise. First, I want to describe why I was doing this in the first place so you don't think I'm crazy, or at least if you think I'm crazy, I'm crazy with a purpose. This was part of an overfeeding experiment in which I wanted to show I could lower my LDL cholesterol levels by going hypercaloric, eating enough that I gained body fat in a ketogenic context. Now, for some background, this is something that's been done by other people as part of what's called the Feldman Protocol. A little bit more background on that linked below. The bottom line is, by increasing your calories massively, if you're someone who's low carb, typically kind of lean with high LDL, you can drop your LDL cholesterol. This has been replicated by other people, so I wanted to do it myself. But remarkably, the results were completely the opposite of what I expected. And when I say unexpected, the opposite of what I expected, I'm not feigning surprise. I genuinely predicted this to go the other way, and what I found was results that contradicted my initial hypothesis. Now, I'm sharing this with you because I don't want to be defensive about the fact that I was wrong. I was wrong. I was really wrong. And I think that's incredibly exciting. My dad always used to say, "Nick, the best part about science, the most exciting part about science, is not when you prove your hypothesis correct, but when you're completely and utterly wrong, because that's when you learn something." So let's learn something together and unpack what went wrong with my experiment, why I'm a little bit different, what we can learn from that, and have a little bit of fun along the way. Oh, and I promise some of the numbers that I'll share will blow your mind. They definitely scared scared me. Welcome to my channel. Stay curious. First, for a bit of background, this video is actually a follow-up to a prior video I did, the Butter Drop Experiment, where initially I was setting out to do just this: eat a ton of calories over a short time frame to force rapid expansion—rate matters—of fat cells and thereby lower my LDL cholesterol. I'll explain why I think that would work in a minute. But that prior experiment failed for two reasons. One, the lab actually forgot to order my final lipid panel, so I didn't have any data to use at the end. And two, I actually didn't end up gaining weight. I lost 0.8 pounds despite eating over 4,300 calories per day. Yeah, 4,300 calories, including over 2,000 calories per day in butter. That week I ate over 4.5 pounds of butter. But yeah, I lost weight. Now, weight gain, fat gain in particular, is necessary for this hypothesis to bear out, and here's why. The hypothesis goes that in lean people who are low-carb, that see the massive increases in LDL, HDL, and the decreases in triglycerides, the so-called lean mass hyperresponders, dynamics at the fat cell really control lipid levels, and one dynamic is their expansion and their shrinkage. So if you force the fat cell to expand, which presumably you can do by just force-feeding calories, if you force it to expand rapidly, then not only are you shoveling triglycerides into their core, the stored form of fat, but you have to expand their surface, you have to expand the cell membrane, and the cell membrane includes cholesterol. So there's a structural demand for more cholesterol as the fat cell's rapidly expanding, which can theoretically be drawn from the serum pool. So the idea is if you take someone like me, a lean mass hyperresponder—lean people who are insulin sensitive on low-carb diets with super high LDL, high HDL, and low triglycerides—then if you cause their fat cells to rapidly expand by overfeeding, even in a ketogenic context, then what can happen is that the cholesterol can be taken out of the blood to feed the cell membrane as it grows, and your cholesterol goes down. Now, importantly, there is precedent for this. A lot of people, dozens by now, have done the Feldman Protocol, which is just this: Take people with high cholesterol on low-carb diets, force-feed them tons of calories in a ketogenic context, and their LDL goes down. It's been replicated by now, again, dozens of people. One example is Siobhan Huggins, who did this experiment primarily with heavy cream. She was drinking, I think it was 1.6 liters of heavy cream per day, and consuming 4,000 to 6,000 calories per day for 5 days, and her LDL went from 252 to 105. So based on the sensible hypothesis, be it speculative, and prior data, I thought this the same thing would happen in me, that I could overfeed calories and my LDL would go down. But I was wrong. I was entirely wrong. I needed to eat 3 meals because my normal 2 just wouldn't do it for the amount of food I needed to get in. So I started eating around 8:00 AM with a Keto Brick, which is 1,000 calories, and about 6 tablespoons of macadamia butter. So at 8:00 AM, I had 1,642 calories. And then around noon, I'd have lunch. A representative lunch would be 10 large eggs cooked in 4 tablespoons of ghee, with some macadamia butter mixed in and drizzled on top, about half a cup or 8 tablespoons of macadamia butter, and then another half Keto Brick. Then have dinner around 6:00 PM, some various fatty meat like a ribeye, maybe 12 ounces, with 4 tablespoons of olive oil, the rest of that half Keto Brick that I started at lunch, and then the rest of the jar of macadamia butter. They come in 16 tablespoons, so I was eating about 16 tablespoons of macadamia butter a day. Or about a cup. So a total for this day would have been, or was, 6,384 calories, 585 grams of fat, 40 grams of net carbs with 34 grams of fiber, and 232 grams of protein. So 82% of my calories were coming from fat. You can see the statistics here. And that was a pretty typical day. So what happened? What were the results? My weight did go up. It went up by about 0.7 pounds over the course of the week. And the weight gain was pretty linear of about 0.1 pounds per day. I don't think it was water weight also because I started in a ketogenic context and I continued in a ketogenic context. So I wasn't eating a ton of carbohydrates to gain a lot of glycogen. But bottom line, there was 0.7 pounds of weight gain, presumably some was fat. And to document that, I did do body caliper measurements, or had a friend do body caliper measurements, on the right side at 4 points with triplicate measures at each site at the beginning and the end of the week. And there was an increase in the sum of 4 from 5.5 millimeters to 6.25 millimeters. That's something like 7.5% body fat to 8% body fat, although don't take those numbers as absolute accurate values because again, this isn't a DEXA. It was just an adjunctive tool to measure some expansion or suggest some expansion of fat cells. So I wanted that. The most important thing is there was some weight gain and the caliper suggested that it was some degree of fat gain. Now, what happened to my LDL. My LDL did go up. It went up from actually a very high baseline of 494 to an even higher level of 541 milligrams per deciliter, which is crazy high. It's actually the second highest I've ever documented. The highest I ever documented was 545, which occurred in the context of actually a very, very low cholesterol and low saturated fat diet where I was eating 1 gram of saturated fat for every 5.67 grams of unsaturated fat, which is consistent with the idea that this isn't a saturated fat-driven phenomenon. So I was still short of my personal high, but nevertheless, my LDL cholesterol did increase from 494 to 541 milligrams per deciliter. So it went up, and remember, I had predicted it to go down. So what else changed? Well, two other things I was measuring during this experiment were resting heart rate and core body temperature, because I wanted some estimate or some suggestion of changes in energy expenditure via compensatory mechanisms that my body would be engaging in in order to attenuate weight gain, since I'm historically a hard gainer. And it was really remarkable, actually— I've never done this before— but how consistent and striking the changes in these parameters were. My resting heart rate started in the low 40s and jumped to the high 60s almost immediately the day after I started overfeeding, and stayed there until I ceased overfeeding. And a similar thing happened to my core body temperature. People are going to ask how I measured this. Sean Baker already has. It was rectally. Yeah, I went there. But I was measuring my core body temperature and resting heart rate every morning at around 5 AM, and my resting heart rate went from the low 40s to the high 60s. My core body temperature went up by about 1.5 degrees °F. So there were increases suggesting an increase in basal metabolic rate and, and maybe some degree of NEAT, non-exercise activity thermogenesis. This was more subjectively assessed via just how much I was bouncing around per day. The pedometer on my phone did measure an increase in step count by about 5,000 per day. So basically all this pointing to compensatory mechanisms whereby my body was increasing energy expenditure via increase in heart rate, body temperature, and non-exercise activity thermogenesis in order to attenuate weight gain. So bottom line, what you can end up doing here, maybe to explain the results and why the results ended up being not what I expected, is there are variables that would be tending to increase LDL and those tending to decrease LDL. The decrease LDL variables are primarily the fattening Again, that was the primary, the core hypothesis, that fat cell expansion would create a cholesterol demand, a structural cholesterol demand, which can be pulled from the LDL pool. So I thought that would be the dominating factor, but there's also factors like an increase in insulin. So if you're overfeeding, what can happen is your insulin can go up, and that can increase the removal of LDL from the serum, so that would also decrease LDL. And because of the diet I was eating, I was eating a little bit more fiber, about 34 grams of fiber per day. So the fattening being the primary, but also an increase in insulin presumably— I didn't measure it, but I'm presuming at 6,000 calories per day my insulin would go up a little bit— and an increase in fiber, those would all tend to drive down LDL. Now, what would be the factors that would increase LDL? Well, this comes to the lipid energy model, whereby we explain the lean mass hyper-responder phenotype, that triad of the high LDL, the high HDL, and the low triglycerides, as a function of changing from carb fuel to fat fuel in lean people who go low carb. And so what you have to have is this massive increase in trafficking of fat around the body, which uses the cholesterol system. The VLDL gets shipped out from the liver containing triglycerides. Those triglycerides get sucked out, put into fat cells and muscle tissue, and the leftover— the VLDL shrinks into an LDL with the surface components of the VLDL being shed and given to HDL. So you end up with a high HDL the high LDL and the low triglyceride phenotype. And one of the predictions of the lipid energy model, which we've demonstrated previously, I've demonstrated previously, is if you try to control other variables and you increase your energy demands, what happens? Well, if you're keto and you're a lean mass hyper-responder, the flywheel of lipid energy model needs to spin faster and faster, right? And so what should happen is that everything just gets, the phenotype gets amplified in magnitude. Your LDL goes up. So what you see here is kind of a balance of factors whereby, you know, the fattening could drive down the LDL, the insulin could drive down the LDL, the fiber could drive down the LDL, but I'm having an increase in energy expenditure, which is suggested again by my increase in body temperature, my increase in resting metabolic rate, and my increase in non-exercise activity thermogenesis. The bouncing around, compensatory mechanisms to prevent fat gain, to prevent weight gain, that would also drive up my energy expenditure. And we can very easily presume this was the case because despite eating over 6,000 calories per day and being in a calculated calorie surplus of over 26,000 calories, I didn't gain over 7 pounds. I gained a meager 0.7 pounds. So it's very clear that there was a massive increase in my energy expenditure, and what I think happened is that that dominated over the meager fat cell expansion.. And what I love about this, what I love about this explanation is it explains why I, someone who's a typically hard gainer, I think it's clearly an outlier response to only gain 0.7 pounds if you're eating 6,000 calories per day, had a response that was opposite of what everybody else who has done the Feldman Protocol before has experienced, the drop in LDL. Because presumably in those people, their fat cells were expanding much more quickly, and they had a much smaller compensatory response. So you'd expect the balance of factors to weigh for them in favor of decreasing LDL as their fat cells rapidly expand, but they don't increase their energy expenditure much. Versus in me, my fat cells expanded but not that rapidly. They didn't expand that much because I only gained 0.7 pounds, whereas I had a massive upregulation in energy expenditure. You can see how beautiful this is because it takes a hypothesis That is a general model of physiology, the lipid energy model, which if you don't have background on that, please click the links below. But it takes that general model and it explains inter-individual variability, why I'm an outlier and had this increase in LDL response and everybody else had a decrease in LDL response. A general model that explains inter-individual variability. Isn't that freaking cool? So I think it's really cool. But with that, I want to go over some limitations also strengths, and then some speculations. One, uh, limitation was it was a 1-week study and an n 1. Those come with the obvious caveats: acute doesn't equal chronic, don't generalize from an n 1. Enough said. I also couldn't keep fiber and protein constant just because of how much I needed to eat. I needed to liberate my diet a little bit, which included Keto Bricks and, uh, more macadamia butter than even I'm used to. So I had a little bit of an increase in fiber and also an increase in protein. Um, although those variables shouldn't predict an increase in LDL, so I think they don't really have major impact on the interpretation of the data we just discussed. I also couldn't easily add a single pure fat source. I tried to do this before with the butter, but it was too satiating as butter alone to get to the levels I needed to get to force fat gain in a ketogenic context. So I ate more fat sources, the keto bricks, the macadamia butter, ghee, olive oil, etc., etc., in order to get to 6,000 calories. Again, I don't think it has bearing on the outcomes, but it is a limitation. Um, strengths include that I think that there was a very clear signal in terms of, um, evidence for changes in energy expenditure. The graphs for my changes in body temperature and heart rate were so clean, way more clean than I would have expected in terms of a response. So, I think it's pretty fair to make the assertion that, yeah, my energy expenditure did go up given the tiny amount of weight I gained versus what will be predicted and concurrent increases in body temperature and heart rate. Also, the results from this experiment are in clear and striking contrast to other experiments done in the subject—me. For example, the Oreo versus statin study. I was right with that hypothesis that eating Oreos would drop my LDL. Via the lipid energy model, and we had a very different result in this experiment, and we can contrast those to gain insight into physiology. Also contrasting my odd response compared to the other people who have tried the Feldman Protocol, again, contrasting the outlier, me, from other people allows us to gain some insight into physiology, at least hypothesis generation. So I find that a great strength, figuring out what's unique in me, What's unique about my response, be it being a lean mass hyper-responder who eats Oreo cookies and drops his LDL, whereas that wouldn't be the predicted response in your average Joe, or me as a lean mass hyper-responder among lean mass hyper-responders seeing an increase in LDL despite others seeing a decrease. Again, contrasting the outlier to the group allows you some insight into physiology. It's really useful for hypothesis generation. Now, in terms of predictions, speculation, I think I could have made the Feldman Protocol work, i.e., drop my LDL cholesterol, if I did "sico" better, i.e., was more hypercaloric so my fat cells actually rapidly expanded. The issue in me is that I have this absurd upcompensation with energy expenditure, so there's a much stronger balance of factors. So I hypothesize, speculate, that if I were able to drive my calories even higher or was somehow able to attenuate my compensatory response in increases in energy metabolism, then I would actually see a decrease in my LDL cholesterol. But as it happened, we didn't see that, and we've gone over why I think that was the case. And now, finally, for a very necessary disclaimer, and I'm going to read something I wrote in order to be precise in my wording. I'm not putting a value statement or making any medical recommendations in this video. In particular, I'm not putting a value statement on LDL cholesterol or medications used to treat LDL cholesterol. I've long been public about my medical background, my personal health history, and have made the conscious decision to share my personal data for the benefit of public discussion and academic discourse, not to endorse any particular lifestyle choice. And while data on lean mass hyperresponders are outstanding, I'm personally monitoring my cardiovascular risk, my actual plaque accumulation, via coronary CT angiography, which you can choose to do if you so wish. But I really encourage everybody to have an open-minded discussion with their doctor about what's the best choice for them. We are currently in a place where there is very little, basically no data on coronary risk in lean mass hyperresponders as a specific population. And with that in mind, it is appropriate to take conservative clinical stances because you don't want to gamble with your health. But every patient case is complex. There is context wrapped around every patient. So I'm not gonna say lower is better, because I think that is a gross oversimplification, and it overlooks the complexities of individual patient management, especially for those who are using ketogenic diets therapeutically. So have a discussion with your doctor, and don't get medical advice off social media. I share what I share for academic discussion, because I think it can be productive. But it becomes counterproductive when people try to manipulate my messaging to use as confirmation bias for something that I'm not saying. Again, value judgments with respect to LDL cholesterol or medications used to treat them. I think that's very important, and I'd really encourage everybody listening to stand with me in this disclaimer, because I think not doing so hinders the academic debate. It distracts from it. So that's my call to action from you. I hope you found this video Interesting.