Hosting Thanksgiving? Get the free keto game plan.

Men with diabetes / What men should know / Fact 4

What Happened in the Keto Trials

Ketogenic diets have been tested in type 2 diabetes, and the results are good and qualified. In the best-known study, average HbA1c fell from 7.6% to 6.3% in a year. In a pooled analysis of 23 randomized trials, low-carb diets were clearly ahead at six months and mostly level again by twelve.

Check this first. If you take insulin, a sulfonylurea or an SGLT2 inhibitor, cutting carbs without your prescriber can be dangerous. Talk to them before you change what you eat. Read the medication page.

The one-year study

Researchers followed 262 adults with type 2 diabetes who chose a ketogenic diet delivered with continuous remote medical care, and 87 who chose usual care. After a year in the keto group, HbA1c had fallen from 7.6% to 6.3%, weight was down 13.8 kg, insulin had been reduced or stopped in 94% of the people using it, and sulfonylureas had been stopped entirely. The usual-care group did not change.

Now the limits. People picked their own group, so the two were not comparable to begin with. The company that sells the program funded the study. LDL cholesterol rose 10%. And the key ingredient was not in the food: physicians adjusted every participant's medication as their blood sugar fell. The paper's own first line says carbohydrate restriction necessitates prompt medication changes.

The pooled analysis

A BMJ analysis combined 23 randomized trials with 1,357 participants. At six months, 57% of people on low-carb diets had an HbA1c below 6.5%, against 31% on comparison diets. Weight, triglycerides and insulin sensitivity all improved by clinically important amounts.

By 12 months most of that advantage had gone, and there were signs of worse LDL cholesterol. The benefit was also smaller when remission was defined as being off medication, and much weaker in studies that included people on insulin.

Why it fades

The analysis points at adherence. Very-low-carb diets worked well among the people who stuck to them and poorly among those who did not. A diet you stop following stops working, which is not a deep finding, but it is the one that decides the result.

It is the reason these pages keep coming back to food you would choose to eat anyway.

This week

Free · take it to your doctor

The Doctor’s Brief

A printable page for your next appointment: your medications, the questions to ask before you cut carbs, and a 14-day log.

No spam, ever. Unsubscribe in one click. By subscribing you agree to our privacy policy.

The evidence

What this page is based on

Each source with what it found and where it falls short. Nearly all of this research is in type 2 diabetes.

Hallberg et al., Diabetes Therapy, 2018

349 adults with type 2 diabetes: 262 chose a ketogenic diet with continuous remote medical care, and 87 chose usual care.

What it found. After a year in the keto group, average HbA1c fell from 7.6% to 6.3%, weight fell by 13.8 kg, insulin was reduced or stopped in 94% of the people using it, and sulfonylureas were stopped entirely.

The limits. People chose their own group, so it was not randomized. Doctors adjusted medication throughout, which a book cannot do. The company that sells the program funded the study, 17% left the keto group within the year, and LDL cholesterol rose 10%.

Goldenberg et al., BMJ, 2021

A pooled analysis of 23 randomized trials (1,357 people with type 2 diabetes) of low- and very-low-carbohydrate diets lasting at least 12 weeks.

What it found. At six months, 57% of people on low-carb diets had an HbA1c below 6.5%, against 31% on comparison diets, with large improvements in weight, triglycerides and insulin sensitivity.

The limits. The advantage was smaller when remission meant being off medication, was much weaker in studies that included people on insulin, and had largely faded by 12 months, when there were also signs of worse LDL cholesterol. Very-low-carb diets did well only among people who stuck to them.

Quick answers

Will I get the results from the one-year study?

Nobody can promise that. Those participants had daily medical support and their medication adjusted for them, and they chose the program. Your results depend on your health, your medication and how long you keep going.

Does keto raise cholesterol?

LDL cholesterol rose 10% in the one-year study, and the pooled analysis saw signs of it at 12 months. It does not happen to everyone. Ask for a lipid panel before and after so you and your doctor can see what it does in you.

General information, not medical advice. It does not diagnose, treat or cure diabetes or any other condition, and results vary. Never start, change or stop a medication without your prescriber. See our health disclaimer.

The Long Game · the playbook

Two weeks, planned for you

A 14-day keto menu with total carbs and net carbs on every dish, a one-page brief for your doctor, and a log for your meter readings. Built to be used with your doctor, not instead of one.

  • 14 days of menus with total carbs, net carbs and protein for every meal
  • 36 recipes and two shopping lists
  • The medication checklist and the Doctor’s Brief
  • A 14-day log for readings, meals and walks

Get the playbook — $19See everything inside

Instant PDF download. 14-day satisfaction guarantee. A meal plan, not a medical treatment. Never change a medication without your prescriber.