News · Nutrition & Diet
Time-restricted eating matched a dietitian, then drifted
A Diabetologia trial randomized 247 Australians at risk of type 2 diabetes to a 9-hour eating window or personalized dietitian advice, and found the two matched at 4 months but not at 12.
Based on a peer-reviewed randomized trial in Diabetologia
- Researchers at Australian Catholic University and the University of Adelaide randomized 247 adults at risk of type 2 diabetes, publishing in Diabetologia.
- One group ate within a self-selected 9-hour window, finishing by 7pm. The other got individualized dietitian guidance. Both received five telehealth consultations totalling 3 hours.
- At 4 months, time-restricted eating was non-inferior to dietitian guidance for HbA1c, but not superior (difference -0.02%; 95% CI, -0.07 to 0.03).
- At 12 months non-inferiority could no longer be concluded, meaning the eating window had drifted behind the dietitian arm.
- The most important caveat comes from the authors: absolute HbA1c changes were small and not clinically meaningful in either group, at either time point.
- Adverse events were minor and did not differ between the two approaches.
- This tested blood sugar in people at risk, not people with diagnosed diabetes, and everyone got substantial professional support.
Time-restricted eating has spent a decade being tested against doing nothing, which is a low bar. A trial in the journal Diabetologia finally ran it against the thing it is supposed to replace: a dietitian.
Researchers in Australia randomized 247 adults at risk of type 2 diabetes to either a nine-hour daily eating window or individualized dietetic guidance. Over four months the two were indistinguishable. Over twelve, they were not.
The comparison that had been missing
Most fasting research asks whether restricting hours beats eating freely. Until now that question was the one being asked, and it was never very useful, because the realistic alternative to fasting is not chaos, it is dietary advice.
This trial asked the harder question directly: whether time-restricted eating is non-inferior to individualised dietetic guidance in changing HbA1c at 4 months in adults at risk of type 2 diabetes. HbA1c is a measure of average blood sugar over roughly three months.
How it was run
The design was a two-arm, parallel-group, multi-centre randomized, non-inferiority clinical trial at two Australian clinical research institutes, enrolling adults with overweight or obesity who scored 15 or more on the Australian type 2 diabetes risk tool.
One hundred and twenty four participants were randomized to time-restricted eating and 123 to dietetic guidance. The fasting group ate within a 9 h, self-selected eating window, with the last eating occasion by 19:00 hours.
Crucially, both arms got real support: both groups received five personalised telehealth consultations between 0 and 3 months, three hours in total. Nobody was left alone with a rule.
What happened, in two stages
At four months the answer was a tie. Time-restricted eating was non-inferior but not superior to dietetic guidance for HbA1c, with the difference sitting close to zero.
By twelve months the tie had gone. The upper bound of the difference between groups was larger than the non-inferiority margin, meaning that non-inferiority could no longer be concluded. The fasting group had drifted, though the gap was small enough that chance could account for it.
The sentence that reframes everything
Then the authors add the line that any honest write-up has to lead with rather than bury: the absolute changes in HbA1c were small and not clinically meaningful in either group at either time point.
Neither approach did much. The trial is not a contest between a winner and a loser; it is two interventions producing changes too small to matter, one of which held its position slightly better.
That is worth sitting with, because it applies to the dietitian arm too. Five telehealth consultations with a professional also failed to shift blood sugar meaningfully in this group over a year.
What it is fair to take away
The participants were at risk of type 2 diabetes rather than diagnosed with it, and their starting HbA1c averaged 5.8%, which is close to normal. There was not much room to improve, which partly explains the small changes and limits how far this transfers to people with established diabetes.
Safety was unremarkable: adverse events were minor and not different between interventions.
The authors’ conclusion is modest and, given their own numbers, appropriately so. Time-restricted eating may offer a pragmatic and practical short-term alternative when access to dietetic support is limited, or if the approach is preferred by an individual.
Which is a real finding, just not the one the internet wants. Not that eating in a window fixes blood sugar, but that if you cannot get a dietitian, a nine-hour window is not obviously worse than one for a few months.
People also ask
Does this mean time-restricted eating works?
It means it did about as well as seeing a dietitian for four months, on one blood-sugar measure, and that neither approach moved that measure much. The authors are explicit that absolute changes in HbA1c were small and not clinically meaningful in either group at either time point. So the fairest reading is that both did roughly the same modest amount, not that fasting is powerful.
What is a non-inferiority trial, and why does it matter here?
It asks whether a new approach is no worse than an established one by more than a pre-agreed margin, rather than whether it is better. Here the margin was 0.10% of HbA1c. At 4 months time-restricted eating stayed inside that margin, so it counted as non-inferior. At 12 months the upper bound crossed the margin, so that conclusion no longer held. Non-inferior never means superior.
Why did the two approaches diverge by 12 months?
The trial does not explain it, and the difference was not statistically significant, so it should not be over-read. One plausible account is that structured dietary advice builds habits that persist while a time window is easier to let slip once support ends. Both groups had their consultations in the first three months only. Adherence beyond that was not sustained by the study.
What was the eating window exactly?
Nine hours, self-selected, with the last eating occasion by 7pm. That is more permissive than the 16:8 pattern often discussed online, which allows an 8-hour window, and the early cut-off matters: eating late is thought to sit worse with circadian rhythms. Participants chose when their window started.
Should I try this instead of seeing a dietitian?
This is general information, not advice for your situation. The practical framing the authors offer is that time-restricted eating may be a pragmatic and practical short-term alternative when access to dietetic support is limited, or if the approach is preferred by an individual. That is a statement about access and preference, not about superiority. Anyone with diabetes, or on glucose-lowering medication, should not change eating patterns without medical advice.
References
- Parr EB, Charrouf R, Hutchison AT, et al. Time-restricted eating versus dietetic guidance on glycaemic outcomes in adults at risk of type 2 diabetes: a non-inferiority randomised clinical trial. Diabetologia (2026).
- National Institute of Diabetes and Digestive and Kidney Diseases. Preventing Type 2 Diabetes.