Time-Restricted Eating Cuts Weight But Leaves Thyroid Function Unchanged in Type 2 Diabetes
A randomized crossover trial finds TRE causes significant weight loss in obese adults with T2DM without disrupting thyroid hormone levels.
Summary
Time-restricted eating (TRE) is popular for weight loss, but some worry that losing weight or fasting for long stretches might slow the thyroid and metabolism. Researchers in Toronto tested this in 39 adults with obesity and early type 2 diabetes. Each person did two 6-week periods in random order: TRE with a 4-hour evening eating window (20 hours fasting), and a standard eating pattern. TRE produced about 3.5 kg more weight loss than standard eating. Even so, TSH, free T4, free T3 and reverse T3 did not change meaningfully, and resting metabolic rate stayed the same. The results suggest short-term TRE weight loss does not trigger the thyroid slowdown often feared during dieting. This was an exploratory analysis in people without known thyroid disease, so it does not cover longer-term use or people taking thyroid medication.
Detailed Summary
Weight loss often triggers compensatory changes that fight regain, and one proposed mechanism is an energy-sparing shift in thyroid hormones: lower active T3, higher reverse T3 and altered TSH. Time-restricted eating (TRE), which compresses food intake into a short daily window, is increasingly used for weight and glucose control. Its effect on the thyroid axis is poorly characterized, and prior studies of fasting and weight loss show mixed results. This analysis asked whether TRE-induced weight loss alters thyroid function in people with type 2 diabetes.
The researchers analyzed data from a randomized, open-label crossover trial at Mount Sinai Hospital in Toronto (NCT05717127). Thirty-nine adults with type 2 diabetes and no known thyroid disease were included. Mean age was 56.3 years, 64.1% were women, mean BMI was 32.4 kg/m², HbA1c was 6.6% and diabetes duration was about 2.9 years. Most were on diet alone or metformin. Participants were excluded if they took insulin, GLP-1 receptor agonists, SGLT2 inhibitors or sulfonylureas.
Each participant completed 6 weeks of TRE (a 4-hour eating window, 4-8 PM or 5-9 PM, with 20 hours of fasting) and 6 weeks of standard eating, separated by a roughly 4-week washout. Thyroid function was measured after an overnight fast at baseline, 6, 10, 16 and 20 weeks: TSH, free T4 and free T3 by immunoassay, and reverse T3 by mass spectrometry. Resting metabolic rate was measured by indirect calorimetry. The fT3:fT4 and fT3:rT3 ratios served as indirect markers of deiodinase activity.
TRE produced clear weight loss: -2.93 kg versus +0.59 kg during standard eating, a between-period difference of -3.50 kg (P<.001). BMI fell by 1.31 kg/m² more. Thyroid hormones did not differ between periods. TSH changed by -0.49 mIU/L (P=.32), free T4 by -0.55 pmol/L (P=.38), free T3 by -0.57 pmol/L (P=.19) and reverse T3 by -0.94 ng/dL (P=.35). Resting metabolic rate was also unchanged (P=.91), and neither ratio differed. Treatment-by-time interaction tests were all nonsignificant (P=.18 to .39). Sex-by-time interactions were not significant either (P=.16 to .93).
The authors also examined precision. The 95% confidence intervals for free T4 (-1.78 to 0.68 pmol/L) and free T3 (-1.45 to 0.31 pmol/L) stayed within their predefined minimum clinically important differences of 2 pmol/L and 1.5 pmol/L. This argues that the null result reflects a genuine absence of a meaningful effect rather than a lack of statistical power. The findings are consistent with other TRE trials showing stable TSH, and with the lack of change in resting metabolic rate.
Caveats are important. The trial was powered for pancreatic beta-cell function, so this analysis is exploratory. Six weeks is short, and the 4-week washout may not have fully stabilized TSH. Participants had treated early type 2 diabetes and no known thyroid disease, so results may not apply to people on levothyroxine, people with obesity alone, or those using earlier eating windows. Sleep changes were not assessed. Overall, short-term TRE appears to deliver weight loss without perturbing thyroid homeostasis in this population.
Key Findings
- TRE (20-hour daily fast, 4-hour evening eating window) produced -2.93 kg weight change vs +0.59 kg on standard eating, a -3.50 kg difference (P<.001)
- BMI dropped 1.08 kg/m² with TRE vs a 0.22 kg/m² rise on standard eating (between-period difference -1.31, P<.001)
- No significant between-period differences in TSH (-0.49 mIU/L, P=.32), free T4 (-0.55 pmol/L, P=.38), free T3 (-0.57 pmol/L, P=.19) or reverse T3 (-0.94 ng/dL, P=.35)
- Resting metabolic rate was unchanged despite weight loss (-6.9 kcal difference, P=.91), so no evidence of metabolic slowing
- 95% CIs for free T4 (-1.78 to 0.68) and free T3 (-1.45 to 0.31) stayed inside the predefined clinically important thresholds of 2 pmol/L and 1.5 pmol/L
- The fT3:fT4 and fT3:rT3 ratios, indirect markers of deiodinase activity, did not differ between TRE and standard eating
- No sex-by-time interactions for any thyroid hormone (P=.16 to .93) in a cohort that was 64.1% women (n=25)
Methodology
This was an exploratory analysis of a randomized, open-label, two-period crossover trial (NCT05717127) in 39 adults with type 2 diabetes and no known thyroid disease. Each person completed 6 weeks of TRE (4-hour eating window, 20-hour fast) and 6 weeks of standard eating, separated by a roughly 4-week washout, over 20 weeks total. TSH, fT4 and fT3 were measured by immunoassay, rT3 by LC-MS/MS, and resting metabolic rate by indirect calorimetry. Analyses used paired t-tests, a general linear model for treatment-by-time interaction and a confidence-interval approach against predefined minimum clinically important differences.
Study Limitations
The trial was powered for beta-cell function rather than thyroid outcomes, so the findings are exploratory. The 6-week intervention and short 4-week washout may be too brief to capture adaptive thyroid responses or to fully stabilize TSH. The sample was limited to adults with treated early type 2 diabetes and overweight or obesity, excluded those on levothyroxine, tested only an evening eating window, and did not assess sleep changes. No conflicts of interest or funding details were reported in the supplied text.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe:
