Time-restricted eating versus dietetic guidance on glycaemic outcomes in adults at risk of type 2 diabetes: a non-inferiority randomised clinical trial.

AIMS/HYPOTHESIS: Time-restricted eating (TRE) consolidates daily energy intake to a consistent 6-10 h window. The aim was to assess whether TRE is non-inferior to individualised dietetic guidance (IDG) in changing HbA1c at 4 months in adults at risk of type 2 diabetes.

METHODS

In a two-arm, parallel-group, multi-centre randomised, non-inferiority clinical trial at two Australian clinical research institutes, adults at risk of type 2 diabetes (i.e. with overweight or obesity and a score of ≥15 on the Australian type 2 diabetes risk assessment tool) were recruited. Participants were randomly assigned using an electronic system (1:1) to TRE (9 h, self-selected eating window, last eating occasion by 19:00 hours) or IDG. Staff conducting assessments and analyses were blinded to allocation; participants and dietitians were unblinded. Both groups received five personalised telehealth consultations between 0 and 3 months (3 h total), specific to their allocated group. The primary outcome was the between-group difference for change in HbA1c at 4 months assessed via covariate linear regression using multiple imputation with a non-inferiority margin set at 1.1 mmol/mol (0.10%). Secondary outcomes were change in HbA1c at 12 months and changes in fasting glucose, insulin, HOMA-IR and nocturnal glucose AUC. Exploratory outcomes included changes in cardiometabolic outcomes and body mass and composition. Adverse events were tracked for 12 months.

RESULTS

One hundred and twenty four participants were randomised to TRE and 123 to IDG (mean ± SD HbA1c: 40 ± 3 mmol/mol [5.8 ± 0.3%]). At 4 months, TRE was non-inferior but not superior to IDG for HbA1c (mean [95% CI]: -0.22 [-0.72, 0.30] mmol/mol; -0.02 [-0.07, 0.03]%; p=0.40). At 12 months, the upper bound of the difference between groups was larger than the non-inferiority margin (0.10%) meaning that non-inferiority could no longer be concluded (mean [95% CI]: 0.47 [-0.19, 1.23] mmol/mol; 0.05 [-0.02, 0.11]%; p=0.14). However, the absolute changes in HbA1c were small and not clinically meaningful in either group at either time point. Adverse events were minor and not different between interventions.

CONCLUSIONS/INTERPRETATION: TRE 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.

TRIAL REGISTRATION: ClinicalTrials.gov; NCT04762251 FUNDING: This research was funded by the Australian Government Medical Research Future Fund (MRFF 1200555).