Science

A small trial tested fasting windows in type 1 diabetes — carefully

Medically reviewed by Longevity Benchmark Editorial Last reviewed 2026-08-21

A 32-person UIC trial found time-restricted eating lowered HbA1c versus calorie counting in type 1 diabetes, without added hypoglycemia risk.

August 21, 2026 · Science · Nutrition

What the study found

A randomized controlled trial led by Krista Varady, professor of kinesiology and nutrition at the University of Illinois Chicago, tested time-restricted eating in adults with type 1 diabetes and reports that it lowered blood sugar without raising the risk of the complications that make fasting a hazard in this population. The work was published in Diabetes Care (DOI 10.2337/dc26-1093) and announced by UIC on August 20, 2026. Thirty-two participants from around Chicago, all adults with type 1 diabetes classified as obese by CDC body mass index standards, were randomly assigned for six months to one of three arms: time-restricted eating, defined as eating only between noon and 8 p.m. with no calorie counting; a 25% calorie reduction; or a control group making no change in behavior.

After six months, the time-restricted group had lower blood sugar than the calorie-counting group, with the HbA1c test showing what UIC describes as an average decrease of about 0.5%. The safety finding is the one the researchers emphasize: time-restricted eating did not increase the risk of extreme high or low blood sugar levels, or of diabetic ketoacidosis. UIC describes this as the first study to examine intermittent fasting in people with type 1 diabetes, and Varady's stated next step is larger, multisite work.

The short version

Fasting windows are promoted to a general longevity audience as though metabolic starting points were interchangeable, and the group with the most to lose from getting it wrong — people dosing insulin — has had essentially no trial evidence to consult. This does not resolve that. Thirty-two people across three arms, six months, one site, with the headline comparison drawn against calorie restriction rather than against no change. What it does is move the question from unstudied to studied at pilot scale, and the finding that matters most is the absence of a safety signal, not the size of the HbA1c difference.

What it means for you

Read the comparison carefully, because it is narrower than the headline suggests. The reported blood sugar advantage is time-restricted eating versus a 25% calorie reduction — two active interventions — rather than versus the control group that changed nothing. That is a legitimate and arguably more useful comparison, since the practical question for someone already inclined to change their eating is which approach to pick, not whether to do anything at all. It also means the result should not be read as "fasting lowers HbA1c by 0.5% compared to your current routine." With roughly ten or eleven people per arm, the more honest summary is that the two approaches were in the same range and time-restricted eating came out somewhat ahead on this measure in this trial.

The safety result deserves more weight than the efficacy result, and it comes with an important qualifier. In type 1 diabetes the concern with any fasting protocol is not whether it helps with weight or average glucose — it is hypoglycemia and diabetic ketoacidosis, which are acute and dangerous. This trial found no increase in either. It found that under research conditions, over six months, with a research team involved. Our own fasting protocol guide lists type 1 diabetes as a contraindication without specialist supervision, and nothing in this study changes that line. If anything it illustrates it: the way this protocol was shown to be tolerable was inside a supervised study, which is a description of specialist supervision, not an alternative to it. Insulin dosing has to move when an eating window moves, and that is a clinical decision.

For the broader question this site exists to answer — whether a given test or program is worth its price — the read-across is modest and mostly cautionary. A continuous glucose monitor can show you what an eating window does to your own curve, and that is genuinely informative, but a device that reports glucose does not make an insulin adjustment safe, and none of the consumer CGM programs we review are built for type 1 management. Likewise the distinction we draw between diabetes reversal and diabetes management holds here without amendment: type 1 diabetes is an autoimmune condition requiring exogenous insulin, and a dietary protocol that improves a marker is management, not reversal. A trial that improves HbA1c by a modest margin in a small sample does not alter that framing, and anyone selling it as though it does is ahead of the evidence.

What this does not say

It does not say you should try time-restricted eating, that fasting is now approved for type 1 diabetes, or that any reader qualifies for this protocol. It does not establish a safe fasting window for anyone outside the trial, and it does not tell you how to adjust insulin — that is a decision for the clinician managing your diabetes, and changing an eating schedule while on insulin without one carries real risk of hypoglycemia and ketoacidosis. The trial enrolled adults with type 1 diabetes who met CDC criteria for obesity; results in that group do not automatically extend to people with type 1 diabetes at other weights, to type 2 diabetes, or to people not using insulin. With 32 participants across three arms, this is a pilot, and its own authors are calling for larger studies.

How this fits what we already know

Our nutrition guides have treated time-restricted eating as a protocol with reasonable evidence for modest metabolic effects in general adult populations and thin evidence at the edges — older adults, people on glucose-lowering medication, anyone with a history of disordered eating. Type 1 diabetes has been one of the thinnest edges, and the guidance has accordingly been conservative. This trial is the first data point at the pilot level rather than a reason to loosen that. We are not revising the fasting protocol page on the strength of 32 participants; the contraindication language there already anticipates supervised use and remains accurate. If the multisite work Varady describes reports out and holds, that will be the point at which the guidance is worth reopening.

Sources

University of Illinois Chicago, "Intermittent fasting may help Type 1 diabetics control their blood sugar levels," today.uic.edu, August 20, 2026 (accessed August 21, 2026).
MedicalXpress, "Intermittent fasting may help type 1 diabetics control their blood sugar levels," medicalxpress.com, August 20, 2026.
Varady K. et al., Diabetes Care, 2026, DOI 10.2337/dc26-1093 (paywalled; not retrievable at time of writing).

Frequently Asked Questions

How large was this trial?

Thirty-two participants from around Chicago, randomized across three arms for six months. That works out to roughly ten or eleven people per arm. It is a pilot-scale randomized controlled trial, and the lead researcher has said the next step is larger, multisite work. A trial this size can establish that a protocol is worth studying further; it cannot establish how a protocol performs across a population.

What were the three arms?

Time-restricted eating, defined as eating only between noon and 8 p.m. with no calorie counting; a 25% calorie reduction; and a control group making no change in behavior. The reported blood sugar result is a comparison between the time-restricted group and the calorie-counting group, which is a narrower comparison than time-restricted eating versus doing nothing.

Does this mean intermittent fasting is safe for people with type 1 diabetes?

It means that in this trial, under research conditions, time-restricted eating did not increase the risk of extreme high or low blood sugar or of diabetic ketoacidosis. That is a meaningful result and it is not the same as a general safety clearance. Participants were enrolled in a supervised six-month study with a research team. Anyone using insulin who is considering a change to their eating window needs that conversation with the clinician managing their diabetes, because insulin dosing has to change alongside it.

How big was the blood sugar effect?

The University of Illinois Chicago release reports an average decrease of about 0.5% on the HbA1c test in the time-restricted group relative to the calorie-counting group. MedicalXpress renders the same figure as about 0.5 percentage points. The paper itself is behind a paywall and we were unable to open it to confirm which reading is correct or to retrieve confidence intervals, so we are reporting both wordings as published rather than picking one.

Does this change anything about buying a CGM or a metabolic program?

Not directly. This trial studied an eating schedule in people who already have type 1 diabetes and already manage insulin, not a consumer product. The indirect relevance is that fasting windows are marketed widely as a longevity intervention to people whose metabolic situations vary enormously, and the evidence in insulin-treated populations has been close to absent until now. One pilot trial narrows that gap slightly; it does not close it.