In my office, I’m seeing more and more people who aren’t trying to lose weight and they’re not trying to reverse diabetes – they’re worried about their cholesterol. I just want to remind everyone of Ketogenic: The Science of Therapeutic Carbohydrate Restriction in Human Health – the textbook I’m pretty proud of. We contributed two chapters, one on basic nutrition and the second one on the use of therapeutic carbohydrate restriction for obesity treatment and diabetes reversal.
If you do make it to my office, I like to pull out a little show and tell. The Banting Diet book from 1860, which is the beginning of this approach – not the keto diet for epilepsy (you’ve heard me talk about that) – was a diet book published in London at one shilling and is the basis for what I teach. Then, I pull out a book from 1923, The Osler Textbook of Medicine, and point out basically that the keto diet, 10 grams of carbs per day, was used to treat diabetes in 1923, which turns out to be exactly 100 years before the new textbook, Ketogenic: The Science of Therapeutic Carbohydrate Restriction in Human Health. It’s a clinical and research textbook. It took 100 years to rediscover and create the scientific background to support a keto diet, which had been used by humans for thousands of years.
Jeff Volek published his first paper in 2002, the same year we published our first paper on a low-carb ketogenic diet. He spoke at a recent conference, the Metabolic Health Summit, which is now the best keto science conference on Earth. I ask every year, Professor Volek, have you learned anything that would make me stop telling my patients to not eat carbohydrates? And he said no. The more he’s been studying keto – it’s been over 20 years – the better it looks. He’s even asserting that it’s normal to have a certain level of ketones in your blood.
I’ve only worked directly with Professor Volek once. We collaborated on a paper years ago when I learned that he had also approached Dr. Atkins to do research. I said, “Jeff, is there any data on low-carb diets?” And actually, there was. We put together a summary. Some small studies were showing a low-carb keto diet was beneficial. This is after I had learned that in a paper from 1980, 24 people from the NIH (National Institutes of Health in the US) intramural group set out to show a keto diet was bad – the low-carb diet, as it was called then. The 24 people all lost weight and it looked like everything got better except that the LDL cholesterol went up, so they said you couldn’t do this diet. It’s harmful; it’ll kill you. As I reflect on that paper all these years later, when we went to the literature, no studies were saying a keto diet was bad. There was no research saying a keto diet was bad. We would have found it. It’s never been proven that a low-carb keto diet is bad. It’s all been hearsay.
And so, through the years, we’ve been studying and publishing papers on randomized controlled trials with good follow-up and good control groups. There’s a body of knowledge to actually say a low-carb diet is better than a low-fat diet, which didn’t have much science behind it anyway. It actually can reverse type 2 diabetes and do all these things we’re just learning about now. It just struck me that there was never really any science to say it was bad. When you get to the topic of cholesterol, prove to me that LDL cholesterol is bad on a keto diet. Until you do that, I’m not worried about it. I say this with a little bit of passion because I’m getting people coming in worried and fear-mongered, even fired by their doctors because they don’t take the prescription medication that they’ve been taught to give people for an elevated cholesterol level. So just remember the context. There’s no evidence in the context of a keto low-carb diet that elevated cholesterol is bad.
People in power positions, physicians who have been taught that cholesterol is bad, might even treat cholesterol as a disease. It’s not a disease. There’s an old saying that in order to know something, you have to write it. The idea that cholesterol is bad started from rather weak science. You really didn’t need a study to show that meningitis could be fixed by penicillin because everyone died, and one person survived. They didn’t die from meningitis with penicillin or a similar antibiotic, so you didn’t need a randomized control trial. You just had to see it happen once and then again and again and again, and maybe it didn’t fix everyone at first, but you didn’t need to go through funding and the research process like an FDA approval trial is done today. The history of medicine teaches us that James Lind, who figured out that scurvy was caused by a lack of vitamin C, only had 12 people in his study to show that he could fix scurvy with citrus fruit. In Clinical Epidemiology, by Alvan Feinstein, one of my teachers, Feinstein says it was a good thing they didn’t have research methodologists to say that 12 wouldn’t be enough people to prove anything because it was enough.
If the potency of your intervention is strong, like meningitis and pneumonia being fixed by penicillin, you don’t need a large sample size or a lot of research behind it because it works where nothing worked before. The smaller the difference you’re trying to find, the larger number of people you need in a study, this is just basic research methodology. The weaker your intervention is, the more people you need to be able to show a difference there.
Back when I was in research training, clot-buster drugs were coming out and Streptokinase was used to open up the arteries in the heart and you would get an IV. Then, a tissue plasminogen activator (TPA) was discovered and they needed 40,000 people in a study in two groups, to show a difference between 8% and 7% outcomes, I think it was mortality, so life and death. That’s a 1% absolute risk reduction, and that’s why they needed 40,000 people in the clinical trial. Why bother to do a study that big to show such a small difference? It cost a lot of money, but it was funded by the company that had the drug that lowered that outcome rate by 1%, so the drug started being used in clinical practice. Who’s going to pay for a study to show a difference between 8% and 7% other than the company that’s going to be marketing that drug?
The Correlation with Cholesterol
What does this have to do with cholesterol? For primary prevention – meaning you don’t have heart disease yet – you don’t have clinical evidence of heart disease or a stroke or anything, it took 20,000 people in clinical trials to show the benefit of a statin drug. It doesn’t even matter how big that relative risk reduction is because it was small. You needed 20,000 people, which means the effect size and the potency of the intervention are going to be very, very small. But they needed to be able to show that difference, so they enrolled all of these people.
With secondary prevention (you do already have cardiovascular disease), statin medicines, really any cholesterol-reducing medicine for that matter, lower the absolute rate to a greater degree. And this is where I’m in the old paradigm and the new paradigm. If you have existing heart disease or stroke, the absolute risk reduction from a statin medicine might be worth doing. It only took 4,000 people in the clinical trials to show a significant difference, much different than the 180 people needed in a smoking cessation trial, looking at 20% and 40% relative risk reduction. You’re getting the idea that these are small effects.
Those studies came during an era when the fear about LDL cholesterol was high and the money available to do these studies from the companies and then by government sources was available. Gary Taubes’s book Good Calories, Bad Calories tells the story of how these clinical trials were not really ever very convincing about saturated fat and cholesterol in food, even the treatment of LDL cholesterol. These were very small effects if there were effects, but the effects were greater for those with secondary prevention, those who’ve already had a problem (like a heart attack or a stroke), and now you’re preventing the recurrence of it.
Cholesterol is not a disease. The disease we’re trying to prevent is atherosclerosis. Atherosclerosis is a disease of the arteries, interestingly not the veins, unless you put the veins around the heart as a bypass or you’re using a vein as an arterial bypass. The veins become more like arteries and you can get atherosclerosis in the veins, but only if they’re in the arterial side of the circulation. Maybe it’s the pressure difference, nobody really knows. But that’s why you can get blockages of bypass grafts even after coronary artery bypass surgery.
What you want to know is your status of atherosclerosis. I’ve been trying to make the language easier to understand. You want to assess the pipes. The arteries in the body are really just pipes. Yes, they’re living, they should be elastic like a hose when it’s fresh, not calcified like a garden hose that you forget to put in the house when it freezes, and you don’t want to have a buildup of atherosclerosis in the arterial walls. Atherosclerosis is the problem we’re trying to fix. The problem is, it’s hard to assess the pipes, so how could we do that?
How We Detect Atherosclerosis
Just about every woman who’s been pregnant in the last few decades has had an ultrasound. And ultrasounds are done to assess the uterus, the growing baby, the fetus. You can get ultrasounds of the neck artery, the aortic artery, and the abdominal artery. You can’t do an ultrasound of the coronary artery; those are too small and the resolution isn’t good enough, but you can assess the squeeze of the heart. Ask the doctor, “Do I have atherosclerosis? Can we assess it?” Typically, doctors aren’t into that, the prevention thing. They wait for the manifestation of a problem and then give medication for the result of the problem. It’s rare in my area that traditionally trained doctors ever ask about the food their patients consume. That’s sad. Food is my secret weapon. If the doctor says you can’t assess the problem, go and get an ultrasound screening. The company I’m aware of in my area, please let me know if there are other ones, is called Lifeline Screening. They come around and set up shop with ultrasound machines. You go in and ask for these pipes to be assessed. They’ll send you a report and if it says you have no atherosclerosis, that’s good information to know. If it says you have mild, moderate, or severe, that’s good information to know. I start with the ultrasound technique of checking the arteries. If someone is really concerned about their coronary artery system, these ultrasounds don’t assess that. While there might be a strong correlation between the two, there are exceptions where the heart is affected while the other arteries are not.
Some people now are going to get a test called a coronary artery calcium score or CAC score. You have to know the background of this kind of test. It’s been around a while, but its popularization was because a wealthy man had a calcium score and the focus of where the calcium was right at the beginning of the two arteries that come off, it’s called the left main. The main artery goes into these two left branches. He didn’t have any symptoms at all. He got the score and they said, “You have calcium at your left main.” It’s not just the score, it’s where the calcium is that is important. He went for further evaluation and had a blockage there. It’s called “the widowmaker” because if you have a total blockage there, the heart stops because all the blood to the big side – the left ventricle – stops, and no oxygen can get there. This man had a procedure done and he felt it saved his life, which it seemingly did. He said, “I want to popularize the idea of people getting this calcium score,” so he wrote a movie called The Widowmaker and asked some people to popularize the coronary calcium score.
Meanwhile, Dr. Agatston in Miami is doing a lot of these scores. In fact, the name on the calcium score is called the Agatston score for Dr. Agatston. He’s publishing papers and the mainstream cardiologists don’t know what to do with this guy. They’re putting in a new test that unfortunately tells you a lot and tells you nothing; it depends on what your score is. In other words, it’s a very sensitive test, meaning if it’s zero, it means you don’t have a problem. But if it’s not zero, it’s not very specific; it doesn’t tell you if you have the disease. If the CAC score is zero, it really means you have no calcified plaque in the coronary arteries. In large studies looking over a period of years, your risk for a heart attack is very low. It’s not zero; it’s not a perfect test, but it gives you a risk stratification that your coronary artery calcium score is really low and your risk for a heart attack is really low.
Important Research Around Cholesterol and Keto
What does that mean? I sleuthed out this paper, “The Role of Coronary Artery Calcium Score in Primary Prevention of Cardiovascular Disease.” Now remember, primary prevention is that you have not had a heart attack or stroke yet. Most people are in this category. If your doctor, for primary prevention, wants you to be on medication to treat cholesterol (which is not a disease but a predictor, remember), the number of people that had to be in studies to show a change for primary prevention was a lot of people, meaning the effect was low. They argue in this paper that if your coronary score is zero, it doesn’t matter what your cholesterol is, you don’t need treatment for cholesterol with statins or other kinds of cholesterol medicine. This is a chink in the armor of the idea that everyone needs statins. If the doctor says, “Everyone needs it,” well, actually, there are other well-respected doctors who don’t feel that way. Do you have the disease atherosclerosis or just high cholesterol? That’s a big difference. I’ll save you the trip to my office. If you’ve never had a clinical event and your cholesterol is high, get a coronary calcium score. If your calcium score is not zero, we’re in a bit of a pickle because you might have narrowing of the arteries. The coronary calcium score is a bit of a Pandora’s box, meaning you open this and you let out all of this non-specific information.
Someone came to me recently and said, “I really want to know the anatomy of my coronary arteries.” There is a second test you can do called a coronary CT angiogram. The first coronary artery score does use a CT scan, but there’s no dye injected and it can’t see the inside of the arteries. But with a CT coronary angiogram (CCTA), you go in that same CT scanner, but they inject dye into your vein. It lights up the arteries, and you can actually get a scan that shows you the inside of your coronary arteries, but it’s non-invasive. It’s pretty amazing. If you have chest pain or tightness up the arm or the jaw, typically the left side – even atypical heart pain could be all the way through to the back; the hallmark is that it’s during exertion, we take that very seriously as physicians. If you came to the ER complaining of that, if it looked like you might be having a heart attack, the gold standard to look at the coronary arteries is to take you to the catheterization lab or cath lab, put a tube in the artery through the leg or the arm, and inject dye directly into that left main coronary artery, then take X-ray pictures. It’s invasive, meaning you have a tube put in there, but it’s done all the time. It’s so simple but it’s not risk-free and much more expensive because you have to mobilize other humans to be there in the lab. A heart catheterization is the gold standard to see if you have narrowing inside your coronary arteries.
It’s pretty unusual for someone to have that level of testing without symptoms, but there are some circumstances where you might get a functional test called a treadmill test or a stress test. This is where you get on a treadmill, they make it go faster and faster, raise the incline, and they have a heart monitor, or they might even use some injectable dye into the heart to see if your heart gets enough blood flow during exercise. That’s called a functional test, which will tell you if you have sufficient flow to a good degree. It’s not perfect. An anatomic test like the coronary CT angiogram tells you the anatomy, but it doesn’t tell you the function.
When I was in training, I did some time in the coronary care unit. There’s always a trade-off between knowing the anatomy – whether there’s a blockage – and knowing the function and whether the heart is getting enough blood even if there’s a blockage. You want to know that because there may be a blockage but it might not be limiting the flow. You don’t want to put in a stent or some other invasive thing like a bypass graft if it’s not limiting the flow. If you wait till you have a heart attack to figure out that there’s flow being limited, you might already have lost part of your heart function. This is the reason why it’s not so clear-cut in thinking what to do.
Recently I had someone who had a really high CAC score. There is arterial damage, there is some calcification, which is a sign of healing, actually. The calcified lesion is a sign of healing, it’s not a sign that necessarily there’s something bad. That’s hard for people to grasp. The calcium score might go up, but it doesn’t mean there’s more narrowing happening. Again, that’s confusing. He really pressed the issue and found an area where he could, out of pocket, pay for the CT coronary angiogram, the more invasive one. The cardiologist in my area didn’t really want to do it. It was going to be a couple thousand dollars out of pocket. He found a place where it was cheaper than that. The scan was done, and most of the arteries were actually clear, even though the calcium score was really high. There was one little place with about a fifty percent blockage, which is not typically flow-limiting. We were left with, well, there is a little blockage there, what does it mean? I sent them back to the cardiologist. I don’t know the result yet. If it were me and I was really concerned, I’d get on a treadmill with monitoring and I’d check the heart to see how far I could go. Or what I do is ask the patient, how much exercise are you doing? How much exertion are you doing? Are you able to do what you want to do without any problem? This is personalizing it to that person. You may have blockages, but if you’re not running marathons, you’re not climbing up ten flights of stairs every day, you’re not at the gym doing all this stuff, you may not need perfect blood flow going to all of your heart all the time. If your adaptive response allows you to live normally, you don’t want to just go in and assess your arteries willy-nilly because the interventions can cause problems. You can have a stent put in to keep the artery open, but these stents block over time. You can go in and revascularize, put arteries around veins that become arteries, but then those block over time.
In another discussion I had, someone came and asked my judgment whether he should go from the coronary score to the CT angiogram or further testing. After discussion and learning about someone the best I could in an hour, I decided that he didn’t really need the assessment. His metaphor was, it’s like I’m driving my car, it runs just fine under the conditions I’m running it. I don’t need to go in and make sure that the pipes are fine in my car. Why would I want to do that for my heart? You may have blockages, but you’re going to live your life fine and perhaps never have a problem. Sometimes getting that information and intervening might not be the right thing to do. It’s rare or uncommon, but sometimes with a heart catheterization or intervention, you might have a stroke or bleed where they do the test. These aren’t totally benign tests.
I’ve gone from if you want to talk about blood cholesterol, we’re going to talk about the old way – total cholesterol and LDL, and the new way – triglyceride and HDL.There are several chapters in the ketogenic textbook that talk about the shift in looking at the blood cholesterol markers. It’s called metabolic syndrome. High triglyceride and low HDL in the blood are what we want to target. We don’t worry about the total and LDL cholesterol. If you want to talk about your risk over time, you can plug in your numbers to the Mayo Clinic Statin Choice Decision Aid tool or the cardiovascular risk assessment tools online, get a risk at baseline, and then see what your risk would be if you are treated with a drug. I did that recently for someone in the office, looked at it, and the change was from two percent to one percent that there would be a problem over ten years. The person looked at that and said, “That’s nothing. I’m not worried about that.” You can personalize that. Of course, the assumption is that data from that database is reflective of what would happen if you’re not a carb eater. These studies have been done, and the prediction models are done on people who eat carbohydrates. My hunch is that we’ll do even better, but that’s a hunch based on all these other things. It’s not based on prospective clinical trials. I’d love to do an outcome study comparing different diets, including a keto diet. It hasn’t been done yet. I don’t even think it’s been envisioned yet.
If you want to assess your pipes, you might get an ultrasound, coronary calcium score, or CT angiograms, which are more expensive and invasive, but not as invasive as a heart catheterization looking at the coronary arteries. If this is all too much, you’re welcome to come to Durham, North Carolina or to any keto-friendly physician to discuss it. The company I’m watching and very supportive of is Dr. Tro, which is doing telemedicine. Dr. Tro is hiring doctors to be able to do this work all over the country.
Virta Health has a big corporate structure behind it and a much higher and more complicated intervention. They send you a kit with a Bluetooth scale, a ketone blood monitor, and then you talk to the doctor to get off the medicine through your app. The best study on diabetes reversal with a keto diet is done by the Virta Health company. I’ve had a few people actually be in the Virta model. My technique is very different, much less intensive and simpler. Perhaps one might say easier, but they both use keto methods, the keto diet. That’s why it’s so effective at reversing type 2 diabetes.
Another paper I want to highlight was a substudy or reanalysis of a Stanford study. Lucia Aronica did a sub-study looking at the people who were in the most extremes in something called the Dietfits study. They took people who did what would be like an Atkins high-fat, low-carb diet versus an ultra low-fat or Ornish type diet within the same study. Some people went to that extreme and stayed with it, and she compared the changes in insulin resistance and blood cholesterol. The upshot is that they both worked and they both lowered insulin resistance. If there was any stronger change, it was with the people who were doing the low-carb diet versus the low-fat diet, but the brilliant message is they both work, and the commonality may be that there was no junk food, no highly processed sugary foods or some other mechanism yet unknown.
To me, the type of diet you do is less important than the metabolic outcome. While some people peg me as just a keto guy, you’ll be surprised perhaps that End Your Carb Confusion – my book that came out over three years ago now – has three levels of carbs. It’s not always a keto-level of carb restriction. If you want to lose body fat or reverse type 2 diabetes, I will always say to use the strictest version. Remember, it’s total carbs, not net, and people get tripped up on that. New “keto” products come out, and people are duped by the marketing of net carbs. I always look at total carbs. It doesn’t take much extra time to look at the nutrition facts. It’s a really important study saying that maybe everyone’s been right all along – that an extremely low-carb diet and extremely low-fat diet, if you do it and you do it properly, actually reduces insulin resistance.
I wanted to also highlight some other research. One was a research presentation. Dave Feldman and the Cholesterol Code and a group of people who have super high LDL cholesterol on a low-carb diet. Many of them are lean and have high HDLs at the same time and low triglycerides. They basically have the metabolic syndrome fixed, but the LDL is really high. The initial study is enrolling 100 people who say they’ve been on a low-carb keto diet, doing a scan, that coronary CT angiogram. These people are being flown to one of the best centers in the world for that particular test. They presented the baseline data from 80 of the 100 people. The remarkable thing is that of the average 4.5 years of self-reported keto diet, 40 of the 80 people had no evidence of any blockage in this CCT. We’re not talking about calcium scores here. It’s the more invasive one.
The LDL story started with a little girl who, by the age of five, had a heart attack. Familial hyperlipidemia with super high LDLs in the 700-800 mg/dL level (18-20.7 mmol/L), caused problems over just a matter of a few years. If LDL is the problem then everyone should have CAD disease, if it’s a problem in everyone. At least there’s a subgroup of people who on average four to five years following a keto diet had no blockage at all. The half that had some blockage didn’t have much, and it was comparable to those in a comparative group in Miami who had no heart disease at all, a healthy group. Compared to typical Americans, which, of course, is not a great comparison, there wasn’t any more blockage than a group that had low LDLs and didn’t have any heart disease in that study.
The study is not done. They’ve been following people now for a year with documentation of the labs that they’re in ketosis and so the full study is taking people for a year and repeating that coronary CT angiogram to see if there’s any change. It may show a subset of people who don’t have to worry about LDL. At most, it starts to whittle away at the idea that LDL matters at all. If this is repeated other people might wake up to the idea that it took 20,000 people in a clinical trial to show a difference in a statin treatment. Don’t freak out if your LDL goes up to 150 or 200 mg/dL (3.8 or 5.1 mmol/L)if all the other factors are there, your HDL is high, over 60 or 80 mg/dL (1.5 or 2.0 mmol/L). I had someone who recently told me their HDL was 120 mg/dL (3.1 mmol/L). You might actually have a doctor start saying your HDL is too high. No, it can’t be too high. They just haven’t seen it that high in this context. It can’t be too high.
I was trained in the LDL paradigm that it was bad, and I’m coming out of that paradigm, at least saying that we’re finding a different way to go about it. Don’t let anyone fear monger you. Don’t treat cholesterol as a disease. Atherosclerosis is the disease.
Finally, I want to talk about Nick Norwitz, a medical student at Harvard. He had severe ulcerative colitis as a young man and nobody knew what was going on. He had to figure out on his own that a low-carb keto diet could fix his ulcerative colitis. He was really ill. He did a recent study – he calls it the Oreo study – where he was eating a roll of Oreos every day on top of his diet and comparing the statin LDL reduction to the LDL reduction he got from eating Oreos. It was meticulously done – two weeks of adding Oreos to his diet. His LDL plummeted. The LDL went down by adding 100 grams of carbs in the form of cookies. It goes down twice as much as a statin drug without the Oreos, all other things being equal. It was a short-term study. He openly says that, “I just did this to get attention. That the mere fact that you can lower an LDL cholesterol doesn’t mean the food that you’re eating is good.” The basic idea is that just looking at an LDL without the context, what else are you doing, is really irrelevant. Showing his result, he knew that was going to happen.
Some people straddle the old paradigm and the new paradigm, even doctors, and say, “If your LDL is up on a keto diet, we’ll add back some carbs and lower the LDL.” I’m not sure that’s necessary. I want proof that LDL is harmful. The data coming from Jeff Volek and other scientists on keto diets say it’s not harmful. Everyone goes into ketosis, every human, and mammal for that matter, if you don’t eat for two days. I want to end on that note. What kind of system would be designed to go into self-destruct mode or a toxic mode if there was no food around? Ketosis is, if anything, a protective thing and it’s anti-inflammatory. You’re getting the anti-inflammatory effects of not eating the inflammatory foods as well. It’s normal to have certain levels of ketones in the blood if you’re doing a keto diet. It’s a normal thing, we store fat on our bodies and you start burning the fat on your body if you don’t eat food for a couple of days. If you don’t eat carbs for a couple of days, your body looks around and says, “I have plenty of energy here, it’s just that energy,” so you start to burn it. I’m putting my eggs in the basket of not eating that terrible, bad food and that this is going to be an approach that is going to be healthy. The LDL goes up because it’s good for you, not because it’s bad for you and you need to worry about it. It’s a different context. Doctors aren’t used to it. It’s not what they’ve seen, but please take all of these things into account. The burden of proof should be on the people pushing the drugs to say that someone eating this way needs that.
Watch the full video here.