Is Exceptional Control of Type 1 Diabetes Possible with a Low-Carbohydrate Diet?

, ,

Is Exceptional Control of Type 1 Diabetes Possible with a Low-Carbohydrate Diet?

Our study in the journal Pediatrics is available here. An accompanying commentary (worth reading) is available here. And a NY Times article about our findings, with a patients’ perspective, is available here

The biggest challenge facing people with any form of diabetes, and especially type 1 (“juvenile”), is controlling blood sugar around meals.

After eating a lot of carbohydrate, blood sugar rises rapidly for 1 or 2 hours. But the insulin needed to control that rise can cause low blood sugar later.

Despite the latest technologies for monitoring blood sugar and administering insulin, most people with type 1 diabetes mellitus (T1DM) face a kind of Charybdis or Scylla choice with every meal:

HYPERglycemia soon after eating versus HYPOglycemia after a few hours

For this reason, most people with T1DM have high HbA1c values (a measure of long-term blood sugar control), indicating substantial long-term risk for serious complications like kidney failure, blindness and limb amputation.

A conceptually appealing approach to improving diabetes control is a low-carbohydrate diet. With fewer carbohydrates (especially fast-digesting grain products, potatoes and sugary foods), blood sugar rises and falls more gently after eating, making for an easier target to hit accurately with insulin. And people on a low-carbohydrate diet require less total insulin, potentially providing metabolic benefits for weight control and heart disease prevention.

Interest in carbohydrate restriction and ketogenic diets (near total elimination of carbohydrate) for type 2 diabetes — the kind associated with obesity in adults — has recently grown, with publication of several promising pilot studies. However, low-carbohydrate diets are currently not recommended for T1DM, in which the body can’t make any insulin, for fear of complications like ketoacidosis, severe hypoglycemia and poor growth in children. This, despite the fact that an ultra-low-carbohydrate/high-fat diet was the most effective treatment for severe diabetes before insulin was discovered. Presently, the American Diabetes Association suggest a diet with 45% carbohydrate, including 45 to 60 grams per meal (although they also encourage individualization).

Making the Connection

I hadn’t given much thought to carbohydrate restriction for T1DM before meeting Dr. Carrie Diulus at a nutrition conference in 2016. Carrie is an orthopedic surgeon with T1DM who maintains NORMAL HbA1c levels on a very-low-carbohydrate diet.

As an endocrinologist, I’ve taken care of people with diabetes for more than 20 years. I’ve never seen someone with T1DM and normal HgA1c after the initial “honeymoon period.” And if I did, I would have expected that person to have an unrecognized hormone problem and frequent, severe hypoglycemia. Carrie was the picture of health. What’s more, she said her diabetes control wasn’t the exception, but the rule in a large social media community of children and adults following a very-low-carbohydrate diet and other recommendations of Dr. Richard Bernstein.

The Study

So, being a researcher at heart (and at work as my day job), I suggested we do a scientific survey of that community. A survey is what’s known as an “observational” study — not a clinical trial. But to document a phenomenon thought not to exist by most diabetes professionals, an observational study is the right first step.

After obtaining ethics approval from Boston Children’s Hospital, and as led by Dr. Belinda Lennerz, we and our collaborators studied about 300 members of the TypeOneGrit Facebook group, including adults and parents of children with T1DM. To conduct our study with as much rigor as possible, we also contacted diabetes care providers and obtained medical records: 1) to confirm that our participants actually had T1DM; and 2) to verify the laboratory tests and other clinical outcomes.

The Results

As hypothesized, the average HbA1c among our participants was 5.67% — in the normal range and well below the average values for people with T1DM of 8.2%. Quite a few of our participants had values in the 4% range (!!!).

Reassuringly, participants reported low rates of complications like diabetic ketoacidosis and severely low blood sugar. We found no evidence for an adverse effect on growth among children. And overall, cardiovascular risk factors were excellent, with a triglyceride to HDL-cholesterol ratio of 1.0, indicative of exceptionally low insulin resistance (though LDL cholesterol was elevated, probably due to high saturated fat consumption).

These findings suggest that a low-carbohydrate diet might help prevent the long-term complications of diabetes — a possibility that needs to be explored in high quality randomized-controlled trials.

How did study participants feel about this approach, which involves life-long restriction of many high carb stables of the modern Western diet?

Mostly, really good! They expressed high levels of satisfaction with their diabetes management and health. However, a substantial minority reported conflict with their diabetes care providers. Some were lectured about having such a low HbA1c and the “dangers” of restricting carbohydrate in diabetes. Some parents didn’t even tell their child’s doctor about the low-carbohydrate diet, out of fear that they’d be accused of child abuse. This distrust, more than any diet, could set the stage for a catastrophic event, if a patient doesn’t seek professional help at times of need, and makes diabetes management decisions beyond their competencies. We hope that, with greater awareness of the low-carbohydrate diet for diabetes, misunderstanding between patients and providers can be bridged — to the benefit of all involved.

What’s Next?

Our study has several important limitations. We don’t know the long-term safety and effectiveness of very-low-carbohydrate diets among the broader population with T1DM. Our participants were evidently highly motivated, and may not be representative of all people with T1DM. Other important questions await further study, such as whether less severely restricted diets (ie, allowing more whole fruits, beans and perhaps modest amounts of whole kernel grains) might provide qualitatively similar benefits; how best to dose insulin in conjunction with carbohydrate restriction and physical activity; and how to manage acute illness, when blood sugar and ketone levels rise. Therefore, the results of this study do not by themselves justify a change in diabetes management.


These limitations notwithstanding, our findings raise the possibility that much better diabetes control than currently thought possible might be achievable — with the hope of preventing feared long-term complications. After many decades focused almost entirely on new (and expensive) drugs and technology, it’s time to refocus on the medicinal power of diet.