Can KETO fix Diabetes? (the VIRTA story)

We have a diabetes epidemic on our hands. Global  rates of diabetes have doubled over the last three decades and more than 800 million adults now live  with diabetes. One thing that's often put forward as a possible solution are low carb diets, and  that seems to make sense, if diabetes is a state where we have trouble processing, digesting  carbohydrates normally, then restricting the carbohydrates in the diet seems to make sense. 

But in science we always need to test things, hypotheses are just that. So a team of scientists  set out to investigate this exact idea. They enrolled a group of people living with type 2  diabetes and they advised them to start eating a low carb diet.

How low carb? Very low carb. If our  body doesn't get enough carbohydrate in the diet the absence of carbs.

This is why we call very low  carb diets ketogenic diets, because they generate Ketone bodies. So the team of scientists aimed for  this physiological range, which usually requires under 10% of our energy, of our calories, coming  from carbohydrate, and for reference the average Westerner and most diets out there get about 40,  50, 60% of calories from carbs, so a ketogenic diet is a pretty radical carb restriction. And  the main thing they were using as a metric were their diabetes parameters, specifically hemoglobin  A1c or glycated hemoglobin, and this reflects the average of our glucose levels over 3 months, and  the normal range of hemoglobin A1c is under 5.

7%, if it's higher than 5. 7 but lower than 6. 5 we  say the person is pre-diabetic and if it's 6.

5 or higher we say the person is diabetic.  So the average hemoglobin A1c of all the participants before the study was 7. 5, so that's  extremely high, clearly in the diabetic range, and that's on diabetes drugs, a lot of these  participants were taking medication to lower, to help lower their blood glucose level and still  it was this high.

Okay, so what happened on the diet? After one year of the program, the average  A1C of the participants was 6. 3%, so that's under the threshold for diabetes, the average of all  participants is.

And bear in mind this is with a lot less medication, most participants dialed  back and many even stopped taking diabetes meds so there was this substantial drop in A1C even  though they were taking a lot less meds. One out of every four participants was under that  diabetes threshold without any medication, so this is what we call remission, the  person is no longer diabetic naturally, without the help of medication. Out  of these about 25% of participants, about half were now in the pre-diabetic range,  so they're now pre-diabetic, and the other half, slightly more than half, were in the normal  range.

So this is what we call complete remission, having a normal hemoglobin A1c without medication.  So this is a very encouraging result, one out of four participants not diabetic anymore, one out of  eight participants completely in the normal range, and without taking any medication. Now, these  were the results after one year of the program, one of the best things about this study, this  is called the VIRTA study after the company that ran the program and one of the best things about  the study and something that I respect that the investigators did is they didn't stop there, they  didn't rest on their laurels, they kept following the participants up to a total of 5 years of  follow-up, and so at the 5 year mark 20%, one out of five participants are still in remission,  still under that diabetes threshold without medication.

Out of these 20%, about 13% were in  the pre-diabetes range and 6 to 7% had normal hemoglobin A1c, what we call complete remission.  Now, these are the numbers looking at the completers, only the participants that finished,  that went all the way to the 5 year time point. If we include, if we look at all the participants  who started the program, so including anybody who dropped out or gave up, then the numbers look  a little worse.

By the five year mark, 14% of all the participants who started the study were in  remission and about 5% were in complete remission. So this is a really interesting data set and I  think it can teach us a lot. First realization is that clearly this program worked for some people,  I've seen people on social media hate on keto, say that this is poison and nobody should ever be  on these diets, I don't really understand that, when you see somebody with type two diabetes which  is a very serious condition and by changing their diet in one year they can be in remission, they  can essentially not be diabetic anymore, kind of solve the problem, and some people maintain that  up to 5 years at least, that's as long as we know, why would you not have this on the table for some  individuals for whom it works?

By the same token, it's obvious from the numbers that  the program didn't work for everybody, the majority of participants were still diabetic  by the 5-year mark. The vast majority, over 90% it seems, were still either diabetic or pre-diabetic, and if we look at the average of these metrics at the 5  year time point for all the participants this is even more obvious. The average A1C at the 5-year  mark was 7.

2, so it improved significantly in the first year but then with time they gradually got  worse again, they rebounded. It's still lower than it was at baseline so it's not like they lost all  their gains but it's extremely high again. And if instead of A1C we look at fasting glucose, same  trend, the average before the study was around 158 milligrams per deciliter, that's extremely  high, normal glucose, fasting glucose is under 100.

After one year of the program the average  was a little over 125, so that's right under the cut off for diabetes which is 126 fasting  glucose, but then after 5 years back up to 151, almost back to baseline. And by the way, this type  of rebound is very common to see, this is almost routine in all kinds of dietary trials, it's not  specific to keto, we see this with all types of diets, and it's probably because people struggle  to adhere long term, to sustain the habits. They can do it for 6 months or a year or something like  that but then long term they struggle to maintain it.

So clearly this program doesn't fix diabetes  for everybody, it's not a panacea, it's not going to single-handedly fix the diabetes epidemic, but  then again I don't think any single diet is going to do that. And bear in mind when we look at these  results that this study was very well designed in terms of giving the participants the best chances  of success, they were given a lot of resources, they had a dedicated app where they recorded all  their metrics and they got a lot of information and education from the app, they also had contact  with a coach and a medical professional, either a doctor or a nurse, through the app regularly. They  also received equipment from the research team to track their metrics, they got an electronic  scale, a glucose and a ketone meter.

Also, this program was designed and supervised by some  of the leading experts on keto diets, some of the academic researchers that have published  extensively on this topic and who believe this diet, are excited about this diet, so this setup  is completely different from me randomly saying I'm going to try a keto diet because I heard about  it and trying to do it by myself without this support system and all these experts and all these  resources. And another thing that's absolutely critical is this study, the VIRTA study was not  randomized, it's not a randomized trial, these participants chose to go on a keto diet, chose  to join this program, chose to do the keto diet, so there's an element of motivation there that's  massive in these type of interventions because people already come into it excited about or at  least open to the diet, to the specific diet, to the specific program, that's completely different  from a randomized trial where you randomly split participants and you tell them, hey, you go  on diet X and you go and diet Y, they didn't choose it themselves. Right?

So bottom line, in  a randomized trial or in real life, just a random person doing it by themselves, you expect much  worse results all else held equal than in this, let's call it, ideal setting. So the program,  the diet is not a fix all for everyone or for most people even, long term, but it does work for  some people, it works really well for some people, there's no contradiction, right? This doesn't  have to be cookie cutter.

We've covered a lot of studies, randomized trials using other diets  and putting diabetes in remission, I know the temptation is to compare these head-to-head and  come up with the winning diet, you can't really do that because the designs are so different,  comparing a randomized trial to a study like this that isn't randomized is not a very informative  comparison but what we consistently see for all of these studies is they all work for some people,  there's always a percentage of participants who achieve remission on all of these diets, none of  them work for everyone and we see that rebound pretty much always, so those are common themes.  I know what's popular and what's viral is to say that there's one diet to rule them all and if  you're not doing it you're killing yourself, that gets a lot of clicks, people love clicking  on that type of stuff but that's not supported by scientific evidence and I don't think that's going  to solve the real problem that we have in the real world. Another crucial lesson from this data set  is the problem with anecdotes because a lot of the discourse online is around anecdotes and personal  stories and these kind of flimsy narratives and what we see in general with these diets, with any  diet really, is those individuals who succeed, whatever percentage, 10%, 20% of people, whatever  the number is who do well on a diet, who do well on some program, they become the evangelists, they  start the blogs and the podcasts and they start calling themselves low carb Steve or low fat Mike  or keto doctor and stuff like that and they insist that this is life-changing, this diet is the right  diet, why isn't everybody doing this?

And they're not lying, they're telling you the truth. The  truth for them. And so if we get our scientific information from a Facebook group or a community  on the internet it looks like all there is is success because those are the people that are  in there very vocally sharing their excitement, we don't see the 80 or 90% or whatever the  number is of people who failed because they're usually not there, they're usually doing some  other diet in some other Facebook group.

Also, remember the rebound, right? Even among those  people that did succeed on the diet, that got really exciting changes over 6 months or a year,  a chunk of those people are going to rebound over time, right? So the anecdotes in the Facebook  group tend to not give us all that information, tend to kind of obscure the picture so this is why  we need data sets, this is why we need science, so we can look at things big picture, it's kind  of an unbiased approach and we don't have all this selection bias from the internet and these  communities.

So my takeaway from everything I've seen on this question is we offer people a  variety of options, all designed to be healthy, so a low carb diet, maybe even a keto diet,  designed to be healthy, I wouldn't do the bacon and lard internet version but you can do low  carb, you can even do keto with lots of greens, with lots of fiber, with healthy fats, fatty fish,  olive oil, nuts and seeds. Now, that's not going to fix the problem for everyone so maybe we have  a carb moderation diet that's not as drastic as keto, maybe we have a Mediterranean diet, maybe  we have a pescatarian diet, all designed to be healthy. So this way we maximize the number  of people that we can help.

We lose virality, the video doesn't become viral but that's  okay, we have real impact in the real world, hopefully. If you're interested in eating a low  carb diet for diabetes or some other purpose, we have a whole video covering how to do a  low carb diet in a healthy way, in a safe way, and I'll link that right here.