Keto diets may cause ‘significant risks’ for pregnant women, physicians say

David S. Seres, MD, ScM, PNS, FASPEN

This article raises some important and valid concerns about the keto diet but should not be considered definitive. Rather, it is more reflective of the authors’ opinions.

Scientific credibility has come under serious attack, and we in science should do our best to explain how we do what we do so that the public is better equipped to understand what they hear from the media.

Most of the authors are affiliated with and/or employed by the Physicians Committee for Responsible Medicine (PCRM), which funded the paper. PCRM is a loud proponent for plant-based diets. This is a popular and probably correct philosophy — that less meat is better. Keto diets depend heavily on meat to create a diet high in protein and fat and nearly devoid of carbohydrates. If you are anti-meat, then being anti-keto goes hand in hand.

The authors properly disclose their affiliation, but the reader should always be concerned about the authors’ objectivity and the strength of their conclusions when such a conflict of interest exists.

The Crosby paper is what is called a narrative review. It is not direct research, and there are no new findings. To create a narrative review, authors often find studies that support what they believe to be true.

This is the least rigorous kind of scientific publication because there is no systematic extraction of available research. For every study included in a narrative review that reports a negative effect, there is a good chance that there are other studies that aren’t included that report a positive or null effect.

While the authors of the Crosby paper are accurately depicting what is in the various articles that they are reviewing, the method with which their own article was written allows for a high risk of biased reporting.

Keto diets restrict carbohydrates and require strict adherence to create a metabolic disarray in which more fat is burned. Instead of using glucose, its usual fuel, the body uses ketone bodies, which are produced by the breakdown of fat. This kind of diet takes a lot of work. There are many unanswered questions about whether a keto diet can even be sustained in the long-term.

There are theoretical concerns with the keto diet. I stress that they are theoretical. But theoretical concerns should be weighed heavily when the value of a keto diet has not yet been proven in rigorous long-term randomized trials, except perhaps for control of drug-resistant epileptic seizures, and is itself theoretical.

The initial rapid weight loss in a keto diet is thought to be due mostly from water losses and may not be indicative of true weight loss. People also often feel badly for a long time, with fuzzy headedness, fatigue, aches and so forth. High protein intake is associated with a higher risk of kidney stones and kidney dysfunction, and it is thought that long-term high-protein intake may predispose people to osteoporosis.

So, I agree with the authors that the keto diet is not ready for prime time until high-quality long-term randomized studies quantifying these effects have been done. But I do not share the strength of their convictions, given the quality of evidence.

Randomized studies of the effect of diet are difficult to do. You would need a lot of people to agree to be randomly assigned to stay on keto diets long enough to see, for instance, how many heart attacks or how much cancer they have. Such a study may take hundreds if not thousands of subjects to adhere to the diet, likely for decades.

More importantly, before any of these dietary studies can be carried out, we must figure out how to motivate people to change their dietary behavior. In this era of personalized medicine and nutrition, the very first thing that should be personalized when trying to change the way people eat is behavior and motivation.

Find out what makes each individual want to change their behavior. For some, it’s reward. For others, it’s avoiding something punitive. When we learn how to best motivate behavior change, then we may be able to study these things in a proper manner.

There needs to be a better balance between the public’s desire to know what to do and science’s transparency about the strength of our knowledge. Even though the overall conclusions in the paper agree with my own biases, the strength of the data does not merit the strength of the conclusions made by Crosby and colleagues.