Keto Diet

Making sense of conflicting science

A fundamental shift in thinking is underway in dietary science, and the process is messy and fascinating.

At issue is a diet in which people replace carbohydrates with fat as their primary source of energy, sometimes called the “keto diet.” While medical science has long considered fat the archnemesis of good health, some researchers now argue fat is a solution to some of the world’s greatest health problems. The result: controversy.

The debate arouses emotions unusual for scientific papers. In a 2015 review, Richard Feinman et al. express frustration that many doctors refuse to prescribe fat-based diets for diabetes patients: "It is not known who decides what constitutes evidence-based medicine but we feel that [current evidence is] sufficiently strong that the burden of proof rests on critics.” In a paper from the same year, Pamela Dyson criticizes some scientists, apparently in response, for using overheated rhetoric: “it is worth reminding ourselves that ‘passion in science is an infallible marker of lack of evidence.’”  She then goes on to accuse low-carb proponents of, “cultural elitism.”

Multiple meta-analyses have produced contradictory results in the last three years on whether low-carb, high-fat diets are beneficial or healthy. For proponents, the benefits include treating epilepsy, cancer, weight loss, metabolic syndrome, heart disease, depression, and more. Skeptics warn the diet may exacerbate depression, kidney issues, hypoglycemia, and heart disease. Low-carb diets often lack critical nutrients. Short-term symptoms of starting the diet can include bad breath, hair loss, constipation, and dry mouth.

The public cannot stop eating while scientists work this out. How do we navigate this debate?


What are low-carb diets and ketosis?

There exist many trendy low-carb diets — including Atkins, Protein Power, and Paleo — and proponents make a wide variety of claims for their benefits, some based on valid evidence, some less so. Most recent variations in these diets aim to replace glucose with fat as the body’s primary source of energy. When a diet has little carbohydrates and only moderate protein — two things the body can use to create glucose — mitochondria throughout the body switch to relying more on fat. There are roughly one-thousand-trillion mitochondria in the human body, and so changing their fuel source may have far-reaching consequences.

Most fats cannot pass through the blood-brain barrier, and so when insufficient glucose is available, the body begins to create a special kind of fat the brain can use called “ketones.” Recent variants of the low-carb diet, called the “keto diet,” seek specifically to switch the food the brain uses in this way, a state called “ketosis.”


Are the studies talking about the same thing?

“Low-carb diet” is a label that can apply to a wide range of diets, and so not all studies on the topic are measuring the same thing. A 2015 meta-analysis by H. J. van Wyk et al. found they were ineffective at helping diabetics lose weight or reduce long-term levels of blood glucose. This study defines “low-carb,” however, as diets in which 45% of food intake is made up of carbohydrates or less. Ketosis requires something closer to roughly 10% or less. The study measures something of value: the impact of partially reducing carb intake. The results may not, however, correspond with what many proponents now consider to constitute “low-carb.” Such fuzzy and evolving definitions can confuse scientific debate.


Is the current science reliable?

One reason there may be wide variation in positions on the keto diet is the wide variation in the quality and reliability of nutritional science. John Ioannidis recently criticized researchers in the field for making strong causal claims about the health impacts of food based on studies that can only establish correlation. A common technique is to compare what thousands of people say they eat and their health outcomes, an approach that will necessarily result in some random or spurious correlations.

The primary studies I cite in this article draw on randomized control trials, which can better establish causality. Even in these cases, however, it can be difficult to isolate the precise impacts of diets, since there is so much variation in what people eat, even when on the same diet, and small differences may matter. “Individuals consume thousands of chemicals in millions of possible daily combinations,” Ioannidis points out.

To overcome the ambiguity in how diets are defined, and what exactly people are eating, it would be valuable to study the health impacts of ketosis, rather than the keto diet. Ketosis is a physiological state that people are either in or not, distinct from whatever diets they may use to achieve it, and it is verifiable with a blood test. More randomized control trials are needed that distinguish groups based on whether they are in ketosis, rather than by what they are trying to eat.

In any case, Ioannidis’ criticism is a valuable reminder to read all nutritional health results with healthy skepticism, and for scientists in the field to express them with humility.


Effectiveness of the diet vs The challenge of compliance

Pamela Dyson finds low-carb diets are ineffective at treating diabetes because patients found it hard to maintain the diets over time. After a period of about 6 months, people in the low-carb experimental group had begun to eat nearly the same level of carbohydrates as the high-carb group. Unsurprisingly, there was little difference in outcomes. She concludes the keto diet, “cannot be recommended as the default treatment for people with type 2 diabetes”.

Dyson’s study suggests staying in ketosis is a challenge, but it leaves open the question whether ketosis itself would be an effective treatment, if there is a way to help people follow the diet. Recent work has attempted to tackle that issue. In a 2017 study, Laura Saslow et al. helped patients follow a high-fat, low-carb diet using “behavioral adherence strategies based on positive affect and mindful eating.” The approach succeeded in maintaining adherence over a year. HbA1c levels (a measure of long-term glucose levels in the blood plasma) were on average lower than the control group, and with less variation.

Amy McKenzie et al. achieved similar results using an at-home digital coaching platform. The guidance helped a large proportion of patients adhere to the diet, and they saw a major reduction in HbA1c and diabetes medication.


But what about all that fat?

Health guidelines for Canada, Uk, Australia, and the US recommend eating very little saturated fat to avoid risk of heart disease. Many low-carb, high-fat diets, in contrast, recommend eating so much saturated fat that it is the body’s primary source of energy.

None of the above meta-analyses found that low-carb, high-fat diets were worse than traditional low-fat diets for weight loss, heart disease, or reducing diabetes symptoms. As Feinman points out, 11 major studies have failed to find any connection between saturated fats and heart disease. Whatever the other outcomes of this debate, it appears the reputation of saturated fats may be undergoing a rehabilitation.


Is ketosis a pseudoscientific fad?

The keto diet is popular among diabetics as a strategy for avoiding high blood glucose. (Glucose can damage the nervous system at high quantities, one of the primary risks of diabetes). More than a quarter of those who follow the diet, however, reported hiding this choice from healthcare specialists, according to a recent study. I was surprised to find that online forums for these patients are full of stories of conflict with doctors and nutritionists over the diet. The disconnect between patients and specialists is in part caused by concerns over saturated fat, but it is also likely amplified by the stigma associated with the keto-diet as a pseudoscientific fad.

The keto diet has many of the hallmarks of a fad, including inflated claims for its capacity to cure everything. A documentary, The Magic Pill, advocates that everyone should follow the keto diet, drawing on a number of non-scientific sources such as naturopaths. Near the end of the piece, the narrator tells the audience, “We can be healthier if we just listen to what nature intends,” which assumes nature is capable of intending.

The work of careful scientists who study the potential benefits of ketosis, however, should not have their work impugned because less cautious thinkers speak on the same subject. While it would be irresponsible for health specialists to become over-excited about a fad, so too would it be irresponsible to reject a potential health intervention because it is a fad.


Is Ketosis a solution for everyone and everything?

In The Magic Pill, a nutritionist claims the diet is what humans evolved to eat. Feinman et al. note the keto diet, “may well approximate the diet used by much of humanity for tens of thousands of years before the rise of agriculture.” Could high-fat diets shift from being the enemy of health guidelines to being their official recommendation for everyone?

The claim that humans evolved to depend primarily on the keto diet is dubious. Carbohydrate-rich roots constituted a major part of many prehistoric hunter gatherer diets, and our modern appetite for sugar suggests our evolutionary forbearers did not hesitate to eat fruits when they found them. It can be a challenge to eat sufficient nutrients in a ketogenic diet, such as potassium and magnesium, drawing into question whether these diets are somehow more “natural” or optimal for everyone. Also, the modern life expectancy is above 80 years old while the majority of people eat high-carb diets.

That humans can undergo periods of ketosis suggests, however, that we are adapted to do so when necessary, likely for periods when carbohydrates were unavailable or when there was no food at all. (Ketosis is also initiated during fasting, when the body draws on its long-term fat storage, rather than food, for energy.) Some experts now argue that our bodies have evolved to depend on periods of hunger, and so to maintain optimal health, it is important to fast intermittently. From this perspective, if the keto-diet is not the best for all people at all times, it may be a tool that could be helpful when used at the right time in the right ways.

One useful outcome of ketosis may be to train the body to use fat efficiently. Some runners adopt the keto-diet for short periods to prepare their muscles to burn fat during endurance races, because the body does not carry enough glucose to rely on it exclusively during long competitions like marathons. They are not the only ones who might benefit. One of the risk factors of obesity and the high-sugar diets is called “metabolic inflexibility,” the inability to sufficiently metabolize (i.e., burn) the fat in food. As a result, excessive levels of it can build up in the blood, raising the risk of heart disease and stroke. Helping people become more metabolically flexible may be one example of where the keto diet will be a valuable tool, rather than a panacea.


Thinking clearly about the Keto Diet

The keto diet triggers a profound transition for the body whose risks and benefits remain only partially understood. This piece could only survey a small fraction of these potential impacts, and it did not seek to decide whether, or for whom, this transition is healthy and desirable. The current state of research appears to suggest, however, that the diet is at the least a legitimate area of research, rather than a mere fad, and one that may come to occupy a role in health recommendations of the future.

To advance our understanding of that potential, scientists should clearly define the diet in terms of whether or not it initiates ketosis. We should also be careful to distinguish the challenge of staying on the diet from whether following it successfully is itself beneficial. While saturated fats appear to be less dangerous than once thought, the keto diet is also likely not the risk-free cure-all that some suggest it is. To avoid unnecessary confusion, careful thinking is in order. While shifting our thinking on the role of saturated fats and ketosis for health, we must not trade one set of errors for another.

  • Tristan Cleveland

Photo credit: stevepb.