Is Keto the Cure for Type II Diabetes?

A drug-free approach might be the best treatment we have for America's most ubiquitous lifestyle disease.


A keto-compliant salad featuring collard greens and bacon crumbles. Photo courtesy of Brian Ambrozy/Flickr

Type II diabetes is one of America's most ubiquitous—and expensive—chronic diseases. Patients often require a suite of pharmaceutical products to manage high blood glucose levels, and the complications that arise over the long term, ranging from loss of vision and limbs to kidney failure and coronary artery disease, strain the resources of patients, their families, and the health care system.

The financial strain on insurance companies, employers, and Medicaid and Medicare is even more enormous. A 2013 study in the American Journal of Preventive Medicine put the lifetime direct medical costs for type II diabetes treatment at $124,000 for patients diagnosed in middle age. With nearly 30 million Americans affected by the disease, the American Diabetes Association estimates the national cost of direct diabetes care to be roughly $176 billion per year.

But unlike type I diabetes, which is an autoimmune disorder that destroys insulin-producing cells in the pancreas, type II diabetes is a lifestyle disease, and thus reversible. Over time, people with type-II diabetes can be made more receptive to their own insulin, which in turn allows their bodies to effectively clear glucose from the blood without insulin medication. The trick for the vast majority of type II patients is as simple as losing weight. ("The relationship between obesity and diabetes is of such interdependence that the term 'diabesity' has been coined," two diabetes researchers wrote in 2005.)

But that "trick" is actually pretty hard. Permanent weight loss without bariatric surgery is practically impossible at the population level. A 2014 study by Kaiser Permanente that looked at incidents of non-surgical diabetes remission in 122,781 patients found that it basically doesn't happen. The most commonly cited number among obesity researchers, meanwhile, is five percent—only five percent of people who lose weight without surgery will succeed in keeping it off over the long term.

Now Virta Health, a Silicon Valley startup, has developed a lifestyle modification system that can reverse the markers of type II diabetes. In a study published this month, Virta researchers found that over the course of a year, they were able to achieve remission of symptoms and a cessation of several pharmaceutical products in nearly two hundred patients using a "novel metabolic and continuous remote care model."

Virta uses a combination of the ketogenic diet—which involves moderate fat and protein intake combined with very low carbohydrate intake—and frequent remote contact with a physician and a health coach to help patients change their lifestyle and lower their body weight, their blood glucose, and their HbA1c (a biomarker for diabetes). In its February study, 94 percent of patients in the intervention arm of the trial were able either to cease using insulin or to radically lower their insulin dose, and all of the patients in the intervention arm were able to stop using a class of antidiabetic drugs called sulfonylureas, which increase the amount of insulin released by the pancreas. The control group meanwhile, increased its insulin use over the course of that same year.

At $370 a month, Virta's model isn't cheap, but it's cheaper over the course of a year than the suite of drug therapies many type II patients require. And if Virta's model is scalable, the long-term savings on dialysis, hospital stays, and management of diabetic foot ulcers could be massive. Already, the company has partnered with Purdue University and Nielsen to offer the Virta system as a covered health benefit to employees with type II diabetes. The company's stated goal is to reverse diabetes in 100 million people by 2025.

I recently spoke to James McCarter, Virta's head of research, about the company's treatment model and the broader landscape of type II diabetes care. McCarter received his A.B. in biology from Princeton, and he got his M.D. and a PhD in genetics from Washington University in St. Louis, where he's an adjunct professor at the medical school. Our conversation has been edited for length and clarity.

Reason: Every doctor recommends lifestyle modification as the first course of treatment for patients with type-II diabetes and pre-diabetes. Yet most patients end up on metformin and eventually insulin and some other pharmaceutical products. Physicians seem jaded about lifestyle modification as a viable treatment. Do they have good reasons to be jaded?

James McCarter: I think endocrinologists and primary care doctors have reason to be jaded, in that they've seen lifestyle modification fail so many times. The conventional advice of "eat less and exercise more" has been shown to not work. People can do caloric restriction for a while, but you know what happens when you calorically restrict without any overall strategy other than just eating less? You get hungry. Exercise, and you get hungry. You can battle that hunger and craving for a while, but eventually it's going to come back. What's generally seen with most lifestyle interventions is that people will lose five pounds and gain it back over the course of a year.

Exercise is great for overall fitness and something that I believe in strongly in terms of maintenance of overall health. But it's not a good strategy for weight loss.

Reason: So instead, type II diabetes patients and their doctors end up treating symptoms instead of trying to reverse the disease itself.

McCarter: I think that's right, and I think to your point about physicians being jaded, they've seen that lifestyle modification only works in a minority of people. They're not surprised when a patient comes back three months, six months, or a year later and the disease has progressed.

Reason: If one of the problems with lifestyle modification is that only a tiny fraction of patients can self-motivate or self-direct an effective change, what does the Virta model do differently? How do you help the patient who can't pull off lifestyle modification?

McCarter: If you look at a large study that was done by Kaiser a number of years ago, they saw that their remission rate of type II diabetes was well under 1 percent. We're seeing well over 50 percent. So what's causing that 50-times improvement in our results? It really comes down to two things. Let's talk first about the physiology and then we'll talk about the behavior change.

First of all, using nutritional ketosis as an underlying part of the physiology approach has a tremendous impact on people's ability to succeed in getting glycemic control, which is reduction in medications, improvement in metabolic health, and reduction in weight.

The reason for that is that unlike a willpower approach, where you're just trying to force yourself to eat less, using nutritional ketosis allows you to tap into body fat for fuel. That means you have incoming energy from your body fat storage and from dietary fat. As a result, people will naturally tend to eat fewer calories because they are satiated.

We ask people everyday, "Tell us on a four-point scale how you feel about your hunger, cravings, mood, and energy. What we see is that as people achieve nutritional ketosis, energy and mood go up, hunger and cravings go down.

In effect, the physiology of ketosis is providing you with a tail wind. It's making the whole process much easier.

Reason: Various kinds of diets can work for almost everyone for at least a little while. How do you make those new eating habits stick?

McCarter: The other part of our model is the coaching. The number of people who can just read a diet book and implement its recommendations without any monitoring or coaching is small. What we're doing with Virta is we're providing five things: We're providing a physician with telemedicine for medication management; a health coach with an ongoing, one-on-one daily relationship who consults on nutrition and behavior change; education that's provided online; biometric feedback; and an online community.

Essentially, we're providing a whole support environment that allows people to understand what they're doing and why.

Reason: Is there a calorie deficit? It seems like there would have to be for weight loss, but I'm also guessing that it would be a small one because you seem to be very skeptical of crash diets or excessive calorie restriction. Or does keto somehow defy the claim that calories out need to exceed calories in?

McCarter: Rather than measuring calories, what we're doing is having people monitor their approximate macronutrients. Roughly how many carbohydrates am I having on a daily basis? How much protein am I having on a daily basis? It's a low-carbohydrate, moderate-protein diet.

What we have people do instead of an elaborate food diary is measure daily blood beta hydroxybutyrate. That's one of the ketone bodies, and by seeing an elevation in beta hydroxybutyrate, we're able to say, "Oh, you're actually in nutritional ketosis which means you're burning fat for fuel which is what we want to achieve."

It doesn't mean you have a caloric deficit necessarily, but at least you're getting your diet and other aspects of your lifestyle correct in a way that supports nutritional ketosis. And we're looking at glucose as well. That way we can say on a daily basis, are you on track or off track?

Now, calories still do matter. What generally happens, even though we are not asking people to count calories, is that because they are feeling satiety in their meals at an earlier point, they're creating a deficit. Rather than having a second or third helping, they're saying after one helping, "Gosh, I feel you know, adequately full." They are generally eating less, especially in the first six months.

Reason: So instead of having the goal be "I will eat 500 fewer calories today than my body needs," patients are focused on getting the macro ratio roughly right and checking their efforts against the ketone blood test?

McCarter: Right, and it's going to be a different journey for everyone. Some people just get it right out of the gate. Other people will take quite a few weeks or even months before they really figure out exactly how to do this. One of the key things we've found is a need for individualization. We want this to work not just for the quantified self-optimizer, but for somebody who has had diabetes for 10 or 20 years, who is on insulin or other potent diabetes drugs they want to stop taking. It has to work for somebody who is a stay-at-home parent, for a business traveler, for somebody working the night shift. It has to work for different ethnic cuisines; it has to work for different dietary restrictions.

That's what both our software and our health coaches aim to do. Make changes that work for specific individuals.

Reason: Where do you think the rest of the medical community is on the utility of a low carbohydrate diet for weight loss? In the realm of nutrition science, the debate over dietary fat still seems pretty contentious.

McCarter: Conventional wisdom has shifted somewhat. Many physicians would describe it as effective for weight loss, but most would say that it is a short-term measure that is not sustainable. Many do worry about dietary fat. There is a growing movement that counters the status-quo. While the number of physicians that recommend a sustained low carb approach for weight loss and metabolic health is still limited, it is growing rapidly. For instance, the international petition [Dr. Sarah Hallberg of Virta and Indiana University Health] started for Dr. Tim Noakes last week has garnered nearly 35,000 signatures, including many physicians and medical professionals. [Editor: Tim Noakes is a South African physician who pioneered early research into the low-carb, high-fat diet as a treatment for type II diabetes. The Health Professions Council of South Africa is attempting to revoke Noakes' medical license because he told a woman on Twitter that she could feed her baby such a diet when the child finished breastfeeding.]

We find that when we start taking care of a patient, their primary care doctor, who often begins as a skeptic, quickly converts to a supporter based on the results we obtain and the supporting scientific literature we provide.

Reason: It seems like the ability to do a lot of this coaching and guidance remotely is what's going to make this model scalable for Virta and anybody else that wants to help large numbers of people make lifestyle modifications. Because if everybody needed to check in with someone who lived where they lived, this seems like a thing that could maybe only go so far.

Jim: You're exactly right. It doesn't work without technology and it doesn't work without the ability to provide what we call continuous remote care. We actually tested that in our clinical trial. Of the 262 people with type II diabetes that were in the intervention arm, half received everything remotely—the physician, the health coach, the education, all of it.

The other half received all the remote stuff but also came to an in-person classroom setting with 10 to 20 other people and a health coach at our clinic. Initially, it was once a week, and then less frequently over time. The outcomes between the in-person group and the remote-only group were statistically indistinguishable.

Reason: Most of my own weight loss was self-directed, but I recently signed up for a remote coaching service with daily lessons as a way to get better about my eating habits, and I noticed that the combination of check-ins with a real person and daily lessons on a website is strangely compelling for reasons that aren't entirely related to the content. I feel watched, but not in a bad way.

Jim: That's part of the reason we structured Virta the way we did, with an actual one-on-one relationship with a coach. There are prior clinical studies that have shown that when you have this coaching relationship, as opposed to entirely automated or entirely self-directed program, people will do better.

Reason: What do you make of the fact that some patients in your trial couldn't come off of metformin? Does that mean some aspects of type II diabetes are not actually reversible? That it could take longer to reverse the symptom that metformin treats?

Jim: Let me talk you through the medications a little bit and address Metformin as part of that. The first thing I should say is that medication reduction is symptomatic of an overall improvement in metabolic health.

Of all the medications for type II diabetes, there are things that can be done right away and there are things that take more time. First off, we want to avoid hypoglycemic events, which is low blood sugar. If you're on potent hypoglycemic medications, which are medications like sulfonylureas and insulin that lower your blood sugar, and then you go on a carbohydrate-restricted diet, that's going to drive you toward low blood sugar levels. So what we try to do early on is very aggressively remove sulfonylurea. Fully everyone in our trial were off that within the first three months.

After that, we're aggressively titrating the insulin downwards, so that about half of the insulin was gone by three months and another nearly half of patients had it reduced. Ninety-four percent of people in the intervention group were able to either reduce or eliminate their insulin use.

But the one that we generally will leave alone and not aggressively reduce is metformin. The reasons for that are that it is generally well-tolerated, it's generally inexpensive, and there's a growing body of evidence that it's effective in prediabetes. The American Diabetes Association now recommends metformin for people with prediabetes to prevent progression to diabetes, and there is also emerging evidence that it may have some other benefits, including longevity benefits.

That's the rationale to leave metformin in place if it's well-tolerated. Our definition for having reversed diabetes is that patients have glycemic control, which means they've lowered their hemoglobin A1C lower than 6.5, which is the diabetes threshold, without diabetes medications other than metformin.

Reason: Chronic diseases require decades of expensive treatment, which means effective lifestyle modification could save payers—be they insurance companies, patients, or employers—thousands of dollars a year per person. Can you talk a little about Virta's disruption potential and what kind of blowback that might attract?

McCarter: There's plenty of work to be done, so I'm not worried about what's going to happen to many of the incumbent players. For instance, there are not enough endocrinologists to take care of all the people with type II diabetes in the United States. If Virta is successful over time, maybe this allows endocrinologists, who are incredibly well-trained, to concentrate on more challenging diseases, like type I diabetes and other extremely challenging endocrine disorders.

As for the pharmaceutical industry, there's plenty of disease out there to be handled. There are also people for whom the types of behavior change we're advocating are not a fit for them. It's not as if we're going to get 100 percent adoption. The industry is going to be disrupted over time, but it will adjust.

Reason: So it's overly dramatic to say that one thing will work for everyone, or that an effective new treatment option will crater incumbents overnight?

McCarter: To get a sense of how industries adapt, it helps to look back at the late 1970s, which is when the dietary recommendations for low-fat foods came out. You can see that within about five years, the food industry rolled out thousands of products where they just removed fat and put in sugar and starch. That probably didn't do consumers any favors, but it showed the speed with which industry can respond.

Reason: The food industry seems to be changing again right now. I've noticed with delight that a ton of products now advertise their protein content on the package, the same way they used to advertise their low-fat content.

McCarter: Yeah, people are focusing on protein. In the coming five years, I think you're going to see a return to the idea that fats can be beneficial. As opposed to saying something is low-fat, I think you're going to see things that advertise as being high-fat.

Reason: That still feels far away to me, but maybe not that far. The number of products and recipes that incorporate chia and coconut and almonds has increased quite a bit. Those are all very fatty, delicious, and thus satiating things. It seems like the next logical step for manufacturers is being more explicit about why they think these products are good for us.

McCarter: There was actually a report from Credit Suisse a couple of years ago, maybe two years ago now, where they basically predicted all of this. It was kind of an industry direction report suggesting that fat was a marketing tool.

Reason: A health care tool potentially being scalable and scaling a health care tool are two different issues. How does Virta scale?

McCarter: On the commercial side of things, our goal is to make this available and affordable to everyone over time. To begin to break through, we've been concentrating on employers. Self-insured employers are on the hook for the costs. What we can do is we can go to that employer and say, "Hey, work with us to have your folks with type II diabetes join the Virta Clinic and turn that around." We put a fair amount of the fee structure at risk. It's outcomes-based, so if we don't succeed in reversing type II diabetes, we don't get paid.

Reason: Does Virta currently have a way to follow study participants past the 12-month mark? The Kaiser study you mentioned covers a pretty long time horizon, and I know obesity researchers like to point out that weight regain gets likelier with each passing year.

McCarter: Absolutely. The Virta-IUH trial (see listing) was originally designed for two years and has recently been extended to five years. You can read the description on our blog. We will be publishing two-year and five-year outcomes. We are also following our commercial patients long-term, with more to come on that in a few months. The longterm view is super important.