BACKGROUND

The median eating duration in the U.S. is 14.75 h, spread throughout the period of wakefulness and ending before sleep. Food intake at an inappropriate circadian time may lead to adverse metabolic outcomes. Emerging literature suggests that time restricted eating (TRE) may improve glucose tolerance and insulin sensitivity. The aim was to compare 24-h glucose profiles and insulin sensitivity in participants after completing 12 weeks of a behavioral weight loss intervention based on early TRE plus daily caloric restriction (E-TRE+DCR) or DCR alone.

METHODS

Eighty-one adults with overweight or obesity (age 18-50 years, BMI 25-45 kg/m) were randomized to either E-TRE+DCR or DCR alone. Each participant wore a continuous glucose monitor (CGM) for 7 days and insulin sensitivity was estimated using the homeostatic model assessment of insulin resistance (HOMA-IR) at Baseline and Week 12. Changes in CGM-derived measures and HOMA-IR from Baseline to Week 12 were assessed within and between groups using random intercept mixed models.

RESULTS

Forty-four participants had valid CGM data at both time points, while 38 had valid glucose, insulin, HOMA-IR, and hemoglobin A1c (A1c) data at both timepoints. There were no significant differences in sex, age, BMI, or the percentage of participants with prediabetes between the groups (28% female, age 39.2 ± 6.9 years, BMI 33.8 ± 5.7 kg/m, 16% with prediabetes). After adjusting for weight, there were no between-group differences in changes in overall average sensor glucose, standard deviation of glucose levels, the coefficient of variation of glucose levels, daytime or nighttime average sensor glucose, fasting glucose, insulin, HOMA-IR, or A1c. However, mean amplitude of glycemic excursions changed differently over time between the two groups, with a greater reduction found in the DCR as compared to E-TRE+DCR ( = 0.03).

CONCLUSION

There were no major differences between E-TRE+DCR and DCR groups in continuous glucose profiles or insulin sensitivity 12 weeks after the intervention. Because the study sample included participants with normal baseline mean glucose profiles and insulin sensitivity, the ability to detect changes in these outcomes may have been limited.