Thirty years ago, when I started recommending the low-carb diets as a therapeutic tool for the treatment of obesity and related disorders, people routinely asked: where are the studies to back up your claims? Admittedly, back then, there were few well-controlled studies on the benefits (or lack thereof) of low carb dieting.
Today we have a large amount of data showing what happens when subjects follow low-carb diets and there have been at least 25 randomized controlled trials (RCTs) conducted comparing low-carb diets to low-fat, high-carb diets. These RCTs are quality prospective, randomized studies – not observations or epidemiological studies.
Despite this growing list of RCTs showing the benefits of low-carb dieting, people are still asking where the studies are.
Because as any social psychologist will tell you, when nutritional is involved, people won’t ever be satisfied by data. If study results don’t agree with their preconceived notions, people will say the studies are too short, too small, or fail to show enough evidence for effects on health long term.
As we shall discuss later in this talk, there has to be a change in narrative, the story that goes along with the evidence showing the superiority of the low-carb diet. And as of yet, there has not been a change in the narrative.
Given the circumstances under how diet studies are conducted—the fact that they simply can’t be as rigorously controlled as drug studies for various reasons– the 25 RCTs comparing low-carb diets with low-fat diets are good studies. (In a drug study, the manufacturer invests a ton of money in the study with the possible reward of a blockbuster drug at the end that could earn billions. In a nutritional study, the researcher spends as much money as he or she can cobble together from grants or other funding sources, with the hope that there will be good publishable data at the end. There isn’t a big payday for nutritional studies.) Despite that:
They show that:
- Low-carb dieters lose on average 65% more weight than on low-fat while consuming, on average, 45 kcal more;
- 17 out of 25 trials show statistically significant greater weight loss on low-carb diet; while 8 did not, but all showed low-carb lost more weight.
- Low-carb subjects had 3.5 times better health outcomes, such as weight loss, lipid values, blood pressure, blood sugar stabilization and other factors;
- Adherence levels on low-carb: 80.7.5% on low carb versus 77.5% on low fat, which was not statistically different. The trials that measured hunger levels showed those following low-carb diets experienced less hunger than those subjects on low-fat diets.
- [Comparing the diets in these RCTs:] Those on a low-carb diet on average ate: 8% CHO, 64% Fat, 28% Protein; Low-fat diet ate on average: 54% CHO, 29% Fat, 17% Protein
The above tabulation can be seen at: http://smashthefat.com/science/
Scientists have tried to explain the variability in these studies. How come most of the studies show a vast superiority of the low-carb diet while a handful of others show only a minor difference?
This question attracted the attention of Christopher Gardner at Stanford who was the lead author on the A to Z study in 2007, a study comparing four diets of varying carbohydrate content. Although the combined data from the study showed the diet most restricted in carbohydrate provided the best results overall, there were some subjects who did well on low-fat diets.
Two years later, Dr. Gardner revisited the data and divided the subjects into two groups: those who were insulin resistant at the start of the study and those who were insulin sensitive. When the weight loss data were evaluated as a function of whether or not the subjects were insulin sensitive or insulin resistant, it was apparent that those subjects who were insulin resistant at the start of the study lost vastly more weight on low-carb diets than on low-fat diets. Insulin sensitive patients, however, although still losing slightly more on low-carb diets did almost as well irrespective of whether they were on low-carb or low-fat diets.
This re-evaluation of the data showed that if individuals are insulin sensitive they can do well on either a low-carb or a low-fat diet. But if they are insulin resistant, dieters will do vastly better on a low-carb diet.
That explains some of the disparities in the RTCs. If the study recruited people who were insulin sensitive, you would expect weight loss to be the same, and if they picked people who were insulin resistant, the weight loss will be significantly less.
Other studies confirm Dr. Gardner’s findings and show that there is a difference between weight loss diets for people who are either insulin sensitive or insulin resistant. Examples are:
- A Low-Glycemic Load Diet Facilitates Greater Weight Loss in Overweight Adults With High Insulin Secretion but Not in Overweight Adults With Low Insulin Secretion in the CALERIE Trial – published in Diabetes Care in 2005
- Insulin sensitivity determines the effectiveness of dietary macronutrient composition on weight loss in obese women — published in Obesity Research in 2005.
The most generous interpretation of these two RCTs with pre-selected subjects, who did not have insulin resistance (IR), shows that low-fat diets did just as well, or better in some circumstances, for weight loss.
If I were a doctor, dietitian or nutritionist setting out and practicing evidence-based medicine today, I would put people with insulin resistance on an LCHF diet. To me it’s a no-brainer. But many people still don’t see it that way.
Now that we know there are many recent RCTs showing the efficacy of the low-carb diet for treating obesity, high blood pressure, lipid problems, etc., let’s take a look back to when this wasn’t the case. To a time when the low-fat diet was ascendant.
Here are some important benchmarks in the timeline of to the low-fat mania that swept the nation:
- 1978: High fructose corn syrup enters the market
- 1980 Obesity in US has been stable between 12-14% since 1960
- 1980 USDA releases first ever low-fat dietary guidelines
- 1982 MRFIT study bombs
- 1984 NIH sponsored Lipid Research Clinic study bombs
- 1984 Anthony Gotto, president of AHA said: “We will conquer atherosclerosis by the year 2000.”
- 1985 - 1986 – NIH & AHA establish the NCEP (National Centers for Environmental Prediction)
- 1986 FDA says “no conclusive evidence” sugar causes chronic disease.
- 1987 Mevacor, the first statin, approved in record time
- 1988 Surgeon General’s report: “Highest priority given to reducing fat intake.”
In the midst of this burgeoning low-fat movement, I decided to start treating patients with low-carb diets in 1985. At that time, I began treating my patients for obesity and weight related health issues and giving them information that was the polar opposite of what they were getting everywhere else.
My own story began a few years earlier, in the early 1980s, when I had myself gained a lot of weight. I had been thin all my life and eaten whatever I liked, but suddenly, it seemed, I had found myself to be obese. I realized I had to do something for my own health
I heard about the Optifast program – a protein-based, low-fat, low-calorie, defined formula diet that was an offshoot of the Cambridge diet. It was a program franchised to doctors who offered it in a hospital setting. You could only get the meal-replacement supplements if you enrolled in a long-term treatment program; you had to go to the hospital weekly to see the doctor for physical exams, blood work, and check ups and to get the next week’s supply of Optifast. The program was effective, and people were losing substantial amounts of weight with it.
About that time, another doctor, who was interested in defined formula regimens, but was unable to secure an Optifast franchise, started a company of his own called Medifast that provided similar meal-replacement supplements and physician supervision, but in a doctor’s clinic setting instead of a hospital. His company developed the defined formula meal powders and would provide the products to physicians who completed his short course on the basics of treating patients with a protein-sparing modified fast.
I completed his course and started treating people with the Medifast program in my clinic. When I read all the Medifast and Optifast literature I realized that the engine behind the weight loss success was calorie and carb restriction and, to a degree, keeping patients in ketosis, but the real driver was that the formula diets, low in both calories and carbohydrates, were lowering insulin levels and improving insulin sensitivity.
In my own research to help myself, I had already stumbled onto the insulin connection to weight gain and metabolic imbalance and so I started thinking about insulin levels could be reduced with regular food instead of the formula diets.
And I noticed that when people got off the fasting stages of the Optifast and Medifast programs, they often put on all the weight they had lost. The maintenance protocol of both programs was to put them on a low-fat, high-carb diet—dry baked potatoes being among the first ‘real foods’ offered to them. Their success in losing had come by lowering carbs, which lowered insulin, and so it seemed to me kind of misguided to put make a 180 degree shift and put patients back on the same high-carb, low-fat diets that had gotten them fat in the first place.
I made modifications to the program to include a single low-carb food meal (chicken, green vegetables, and salad) each day along with the low-calorie, low-carb, low-fat shakes, and the patients did great and were able to stick to the plan much better. After a few thousand patients, I realized no one had any problems as long as the diet was modified. The vast majority of people could safely do this kind of modified fasting program on their own. The only problem was people couldn’t get the shake products outside a clinical or hospital setting.
There were a very few protein supplements available, and those that were tasted awful. So my wife and I set about trying to reverse engineer a palatable shake recipe made from these few egg white or whey proteins then on the market. My patients said I should write a book about it, which ultimately became my first book, Thin So Fast.
The physician who started the Medifast program heard that I was writing a book about protein sparing fasting, and asked if he could write the forward. When I sent him a draft of the book, he was infuriated (chiefly because I was giving the reader a recipe for the shake powders that they could make at home cheaply.) He railed that I would be responsible for the deaths of many people if I published this work. He brought up the spectre of The Last Chance Diet, which had been an immensely popular, ultra-low-calorie plan of 250 calories of incomplete protein from hydrolyzed collagen a day, and that had led to some deaths.
He wrote a letter to my book’s publisher, and co-opted a Harvard surgeon, who had done a lot of nutrition research, to write his own letter of dire warning to the publisher to support his view that the book was dangerous.
These warnings to the publisher threatened to derail the project. Then in November 1988, Oprah Winfrey finally revealed on her talk show what was behind her dramatic weight loss; she appeared in a pair of size 8 Calvin Klein jeans, pulling a little red wagon filled with 67 pounds of fat, showing how much weight she had lost on the Optifast diet. So here was the publisher with dire warnings of danger to readers on the one hand and a completed manuscript for a book on how to do at home what Oprah had just done on the other. The publisher had me write rebuttal letters to scientifically address all the concerns raised by the skeptics, which I did, and went ahead and published my book, getting it out by December 1988.
My book taught people to use available products to make a low-carb, low-calorie protein shake from easily available ingredients, and was pretty much identical to Medifast, with complete instructions on how to undertake my modified version safely on their own at home. It did well regionally—anywhere we were able to get media, but without an appearance on the big morning shows, not as well as they’d hoped nationally.
At the time my book came out, it was going head to head with another bestselling book, called The 8-Week Cholesterol Cure, with the take-home message was that people should eat a low-fat diet and muffins to lower their cholesterol. That kicked off the muffin craze that encouraged people to indulge in carb-laden baked goods, thinking they were taking steps to protect their health.
By this time, I had switched from a general family practice to solely focusing on treating obesity and weight related health issues with nutrition.
I joined the North American Association for the Study of Obesity (NAASO) and the American Society of Bariatric Physicians (ASBP), both heavily invested in the low fat, high carb conventional wisdom of the day. So I attended the meetings to glean the information I needed to further educate myself on the nutritional literature and kept my mouth shut. (I’d probably have been drummed out of the conference, had the organizers known my views.) About the only thing NAASO and the ASBP agreed on was that the ‘right’ diet was low-fat, high-carb; in those days there was not a lot of mixing between the two organizations.
Academics always seem to treat with some measure of disdain physicians in clinical practice. And back then, the members of NAASO looked upon the ASBP members as though they were more or less diet pill pushers. (The ASBP was the arm of obesity medicine that advocated using drugs for weight loss, because how else can you keep a person on a low fat high carb diet? They’re hungry all the time!)
Things have changed a lot since those days, and the ASBP has more generally come round to the low-carb side. These are doctors in practice and what they mostly want is for their patients to succeed and be healthier. Even academic organizations are now beginning to see the light. But back then, the low-fat diet was the treatment of choice for obesity.
Despite my own belief in the virtues of the low-carbohydrate diet, I actually had a lot of experience in treating patients with low-fat diets.
A major pharmaceutical company recruited our clinic to investigate the anti-obesity drug Orlistat (Xenecal, Alli) as a maintenance drug to prevent weight regain. We were part of a multi-centre study on whether the drug would help people maintain their weight loss after a low fat diet. Our clinic, it turned out, was the largest clinic in the world in terms of number of patients enrolled for the Orlistat maintenance trial.
Orlistat is a lipase inhibitor, which means that it prevents the absorption of some of the fat people eat—about 20% to 30%–and effectively puts them on a forced low fat diet. It was first tried as a cholesterol-lowering medication, but didn’t work very well, since dietary fat has very little if anything to do with cholesterol levels. So, since the medical establishment believed a low fat diet would make people thinner, it was tried as a weight-loss aid, but I can tell you it didn’t work very well for that either.
In the trial, subjects were put on a low-calorie, low-diet for 6 months, with lots of weekly hands-on monitoring and encouragement. To then make it into the randomized part of the study—where they’d receive the Orlistat to help the maintain weight–patients had to lose 4% of their body weight in those 6 months. For most participants, who weighed on average over 200 pounds, that meant losing just 8 pounds (or about 3.5 kg) in 6 months of low-calorie, low-fat dieting. I was amazed at the difficulty people had in losing just 8 pound of weight on a low-fat diet and how many were unable to lose even this modest amount. In stark comparison, the patients in our regular weight loss practice (who were on a low-carb, high-fat diet) were losing 2 to 4 pounds a week and this caused no small amount of consternation on the part of the study participants sharing stories in the same the waiting room.
This long study, lasting over two years, gave me the opportunity to see first hand the difference between the outcome of the low-carb diet I used with my own patients for weight loss and the results of the low-fat diet the protocol of the Orlistat study imposed.
At that time, the conventional medical wisdom taught that a calorie was a calorie. (In many quarters that view persists to this day.) By this way of thinking, there was no real difference between the calories in meat and an equivalent number of calories as pure sugar. Anyone with common sense knows that while those calories might contain equal amounts of energy, they are not equal in their effects on the body and in particular on blood sugar and insulin levels.
One illustration of the error in a calorie is a calorie can be found in a small but very well controlled study in the American Journal of Physiology in 1992. Healthy insulin-sensitive young males were used as their own controls. After a 12 hour lead-in fast, half the subjects were totally fasted for another 72 hours while the others had 105% of their caloric need given intravenously as pure fat (Intralipid). The results are quite astounding: the two groups showed almost the same weight loss whether they were completely fasting or eating more calories than they needed as pure fat. The other blood values – ketone levels, blood glucose, insulin levels, etc. – affected by the total fast were duplicated in those subjects getting the intravenous fat.
From these results, the authors concluded that carb intake was important for normal fuel homeostasis. I came to a very different conclusion: that carb restriction, not the presence of a negative energy balance, was responsible for initiating the positive metabolic response to fasting. After all, early man fasted a lot and ate a lot of fat, and from all we know, didn’t have a weight problem. I started coming round to the idea that carbs were problematic for most people.
This is a study that most people don’t know about, but I was stunned by it. It made me more convinced than ever that I was on the right track. And it helped me to better understand the interplay between insulin and glucagon and how these two powerful hormones function in the gain and loss of stored fat.
Another pivotal point or epiphany in my medical and nutritional journey came when I read the book Napoleon’s Glands by Arno Karlen, now out of print and difficult to find. Before I read this book, I had never particularly thought about our evolutionary heritage as it pertained to diet. But the book contained a fascinating chapter on paleopathology, discussing how we can learn much about disease from examining the remains of ancient people.
From the science of paleopathology, for instance, we know that the ancient Egyptians suffered many of the medical ills we do—the so-called diseases of civilization–with obesity, high blood pressure, and rates of heart disease likely as high as do people alive today.
And they did it on a diet most modern nutritionists would call ‘nirvana’: they didn’t have sugar, though they had a little honey; they ate fish, fowl, very little red meat, the made beer, and they ate a lot of grains. In fact, the Roman word for Egyptian soldier translates to ‘the bread eaters’ because each soldier was issued about five pounds of bread each day as his ration.
And the Egyptian story got me thinking about hunter-gatherers and what they ate and about the evolutionary basis for a LCHF diet.
And there were other books I stumbled across that helped complete the puzzle for me, including one called The Stone Age Diet, by Walter L Voegtlin, a gastroenterologist and comparative anatomist. He went into this detailed breakdown of digestion between carnivores and herbivores and showed that humans are more like dogs than sheep, yet we encourage them to eat all this plant food.
I had already hit upon the pivotal role of insulin and insulin resistance in obesity and the diseases associated with it and by 1992 I was working on my next book, tentatively being called ‘The Insulin Connection’. The publisher liked the idea of my wife and I writing the book together (she had written five books on health on her own by then) and so we teamed up on that book, which became Protein Power and that went on to become a best seller. It was at its core a high fat, low carb diet, adequate in protein, but the publishers didn’t want to publicize the high-fat nature of the diet at the time, as the mainstream were still adamantly of the opinion that fat was bad.
At the time, I was still in the mindset that cholesterol was meaningful and the publicity team, for sure, was concerned about our recommending eating too much saturated fat, so we advised people to eat lean meats, trying to find something that would resonate in the milieu of fat terror.
But in every interview or television appearance, we’d be queried about a laundry list of myths about low-carb diets. I call them Vampire Myths, because they just won’t die:
- Low carb diets clog your arteries;
- Low-carb diets will destroy your kidneys;
- Low-carb diets cause osteoporosis
- Low-carb diets decrease endurance
- Low-carb diets make your thinking fuzzy
- Low-carb diets cause cancer
- Low-carb diets cause dangerous ketoacidosis
- Low-carb diets defy the laws of thermodynamics
Good science has laid them all to rest, but they’ve now morphed into other a pair of other Vampire Myths:
- Low-carb diets are simply calorie-restricted diets; and
- Low-carb diets don’t work any better than low-fat diets.
Yet, as we now know from the 25 RCTs showing the low-carb diet’s superiority, that’s not the case.
So, why do people keep saying this?
A couple of reasons.
First, there is the confirmation bias (a tendency to search for or interpret information in a way that confirms a person’s preconceptions, leading to statistical errors), and cognitive dissonance (the discomfort that arises when beliefs don’t match up to reality, and that more often than not resolves in further confirming the beliefs, not overturning them).
The second reason is in way data is presented in virtually every study done today: the intention to treat (ITT) analysis. ITT was developed basically to analyze drug studies, but it’s use has become so ubiquitous that it’s almost impossible to get a study approved unless this method of analysis is used to portray the data.
ITT ignores non-compliance, protocol deviations, withdrawal, and anything that happens after randomization. It ignores and distorts data, and more than often than not, it gives erroneous results.
By its very nature, ITT always hurts the better therapy, in this case, the low carb diet. ITT analyses will fail to show that low-carb diets outperform low-fat diets. These erroneous results are then repeated as gospel.
Since the results, especially for weight loss, are typically so much greater with low-carb diets, any dropouts from the low-carb arm of a study have a much greater impact on the average outcome than do the lesser results for dropouts on the low-fat arm. ITT ends up making the study results trend toward being the same rather than showing the larger differences in outcomes for those who actually followed the two therapies to the end.
In 1973, Horst Rittel and Melvin Webber, a couple of urban planners at UC Berkeley, wrote a seminal paper on why it was so difficult to reach agreement on some problems while it was so easy on others.
Problems that carried an emotional or political component were much more difficult to reach a consensus on as compared to problems that didn’t carry this emotionally laden baggage.
Something simple, such as the best antibiotic for treating strep infections, has no real emotional impact on people, so it’s pretty easy to look at the data and come to a consensus. Rittel and Webber termed these kinds of problems as Tame Problems.
Other issues, say, gun control, taxation, urban planning, immigration, etc., all have an emotional component and are much more difficult to solve as a consequence, because more than just data are involved. The authors called these problems Wicked Problems.
Wicked problems are complicated. They’re nebulous things based on perception and feeling, not fact. And people perceive things differently, with many posited explanations for why a problem exists. Depending on which explanation you choose, there’s a different solution. In the minds of many, data is irrelevant when emotion is involved.
The only way to solve Wicked Problems is by changing the narrative or the story around the problem.
In my view, the issue of what a human should correctly eat for health ought be a Tame Problem. And looking at the data, the solution should be the low carb diet. Plenty of well-controlled trials have shown it to be superior to the low-fat diet, but people have an emotional investment in diet. Perhaps they are anti-eating meat, so ethical issues get into the mix. Suggesting that we should be raising cattle and putting everybody on a low-carb diet—which some perceive is harmful to the environment or too resource intensive or couldn’t be achieved in third world settings where hunger is rife–creates a situation in which data is ignored because of the emotional undercurrent. It is difficult to persuade with the data alone, which converts what should be a Tame problem into a Wicked Problem.
And since we can’t solve Wicked Problems simply by throwing data at them, we need to change the narrative about the value of low-carb dieting.
Look at the current landscape during the period when these 25 RCTs were published. The Google trends line for the search term ‘low-carb diets’ over the past 10 years is flat. The problem with the LCHF, Protein Power, or any other low carb diet is the narrative associated with them. For years Atkins, which became synonymous with LCHF, was vilified. People were brainwashed by the media about its dangers and in knee-jerk fashion became scared of it, even though they have the wrong information about it. Ten years of good, solid science supporting the LCHF diet as better hasn’t overwhelmingly changed public opinion, because, you’ll remember that data doesn’t change minds, particularly in Wicked Problems.
Look instead at a Google trends line for the search term ‘Paleo diet’. It came out of nowhere and towers over low-carb in search frequency. Why? Because it has a narrative – eat like our Paleo ancestors ate, eat the diet we cut our genomic teeth on over 3.5 million years of natural selection, eat the diet in sync with our Paleolithic past. It’s a story that people can embrace and understand. It’s not about macronutrients (protein, carb, or fat) but rather about our history.
To persuade, the LCHF narrative needs to change.
On a more personal note Drs Michael and Mary Dan Eades are truly two of the most amazing people I have had the privilege of meeting. Follow their blog here: http://www.proteinpower.com