Time-restricted eating: TREAT, TIME, and the calorie-deficit confound

Key takeaways

• TREAT (Lowe 2020 JAMA Intern Med) randomized 116 overweight or obese adults to 16:8 time-restricted eating (eating window 12 noon to 8 pm) versus three structured meals per day, both without explicit calorie targets, for 12 weeks. Weight loss was 0.94 kg versus 0.68 kg (NS). Lean mass loss was greater in the TRE group .
• TIME (Liu 2022 NEJM) randomized 139 adults with obesity to 8-hour eating window (8 am to 4 pm) plus calorie restriction versus calorie restriction alone for 12 months. Weight change was negative 8.0 kg versus negative 6.3 kg (NS); body composition, lipids, and HOMA-IR were not different between groups .
• Cienfuegos 2020 randomized 58 obese adults to 4-hour, 6-hour, or control eating windows for 8 weeks. Weight loss was about 3 percent in both restricted-window arms versus 0 percent in control; the 4-hour and 6-hour arms did not differ from each other .
• The Varady 2022 narrative review of intermittent fasting (Nat Rev Endocrinol), which covers alternate-day fasting, the 5:2 diet, and time-restricted eating, concluded that TRE with eating windows of about 4 to 8 hours produces weight loss of about 3 to 4 percent over 8 to 12 weeks via an unplanned (spontaneous) reduction in daily energy intake of roughly 10 to 30 percent, with cardiometabolic effects that are inconsistent across trials and weight loss on a par with daily calorie restriction .
• Sutton 2018 (early time-restricted eating in 8 men with prediabetes) showed metabolic improvements (insulin sensitivity, blood pressure) at matched calorie intake, but the trial was small and the result has not been reproduced in larger isocaloric designs .
• The dominant signal in published TRE trials is that adherence-driven calorie reduction explains most observed weight loss; no consistent metabolic effect beyond calorie matching has emerged in well-powered trials.

What time-restricted eating actually is

Time-restricted eating (TRE) is a category of intermittent fasting that constrains the daily eating window without prescribing what or how much is eaten within the window. Common protocols include 16:8 (8-hour eating window), 14:10, 18:6, and the more aggressive 4-hour or one-meal-a-day patterns. Early TRE shifts the window to morning and early afternoon (for example 8 am to 4 pm); late TRE shifts it to afternoon and evening.

TRE is distinct from alternate-day fasting (ADF) and 5:2 dieting, which restrict on specific days rather than within each day. The mechanistic case for TRE rests on circadian alignment of metabolic rhythm, hepatic insulin sensitivity, and gut-microbiome stability, not just calorie reduction.

The TREAT trial: 16:8 versus three meals at free intake

Lowe et al. 2020 (JAMA Internal Medicine) randomized 116 adults with BMI 27 to 43 to consistent meal timing (CMT, three structured meals per day, 8 am to 8 pm) or 16:8 time-restricted eating (eating window 12 noon to 8 pm) for 12 weeks. Neither group received an explicit calorie target; both groups were free-feeding within their windows .

Weight loss was 0.94 kg in TRE versus 0.68 kg in CMT (NS, P 0.63). The TRE group lost more lean mass (negative 1.10 kg versus negative 0.65 kg in CMT, nominal P 0.07). Cardiometabolic markers (lipids, glucose, insulin, blood pressure) did not differ. The headline finding was that the TRE-specific weight-loss claim, when isolated from explicit calorie reduction, is small at best.

TREAT was widely criticized in the popular discussion for using a free-eating window. Defenders of TRE noted that the practical mechanism in real life is exactly the spontaneous calorie reduction that the free-eating-window design tested. The trial does not exclude that TRE works for some users via that path; it does suggest the effect is small and not larger than baseline three-meal patterns.

The TIME trial: TRE plus calorie restriction versus CR alone

Liu et al. 2022 (NEJM) randomized 139 adults with BMI 28 to 45 to (a) 8-hour eating window from 8 am to 4 pm with calorie restriction (1500 to 1800 kcal per day for men, 1200 to 1500 kcal per day for women) or (b) the same calorie restriction with a non-restricted eating window. Both groups were monitored for 12 months .

Weight change was negative 8.0 kg in the TRE plus CR group versus negative 6.3 kg in the CR-alone group; the 1.8 kg difference was not statistically significant (P 0.11). Body composition, abdominal visceral fat, subcutaneous fat, and liver fat all decreased similarly. Fasting glucose, postprandial glucose, HOMA-IR, lipid levels, blood pressure, and quality of life were not different between groups.

TIME is the longest, largest, well-controlled comparison of TRE plus calorie restriction versus calorie restriction alone. The result was substantially deflating for the TRE-as-metabolic-magic claim. Calorie restriction worked; adding the eating-window constraint produced no significant additional benefit.

The metabolic flexibility hypothesis

The mechanistic case for TRE relies on metabolic flexibility, the capacity to switch fuel sources between fed and fasted states. Sutton et al. 2018 conducted a 5-week crossover in 8 men with prediabetes, comparing early TRE (eating between 8 am and 2 pm) with a 12-hour control window at matched calorie intake. The TRE arm showed improved insulin sensitivity, beta-cell responsiveness, and blood pressure at the same calorie load .

Sutton 2018 was small (n equals 8) and has not been reproduced in larger isocaloric trials. Later isocaloric TRE designs in larger samples have generally shown either no metabolic advantage or small advantages that do not survive statistical scrutiny. The metabolic-flexibility hypothesis remains plausible but underpowered for a strong claim.

The adherence and cardiovascular questions

Adherence is the dominant practical lever in any weight-loss intervention. TRE has the unusual virtue of being a simple rule (do not eat outside this window) that some users find easier to follow than calorie counting. The Varady 2022 review concluded that the TRE-specific weight-loss effect, where it exists, is best understood as a behavioral simplification that drives spontaneous calorie reduction .

A 2024 conference abstract analysis of NHANES data linked self-reported 8-hour eating windows to higher cardiovascular mortality in observational follow-up; the abstract was widely discussed but has methodological limitations (self-reported eating windows, residual confounding, no clinical-trial replication, abstract-only at the time of writing rather than peer-reviewed publication). It belongs in the hypothesis-generating bucket rather than as evidence of TRE harm. The peer-reviewed RCT record summarized above remains the better guide .

What the evidence does not yet resolve

• Whether early TRE (eating window 8 am to 4 pm) produces different metabolic outcomes than late TRE (12 noon to 8 pm) in larger trials. Mechanistic data favors early TRE; trial-level data on the comparison is thin.
• Whether the small lean-mass-loss signal in TREAT generalizes. If it does, TRE could be a worse strategy for users on a GLP-1 already losing lean mass at a faster rate.
• Whether TRE produces meaningful long-term cardiovascular outcomes. No event-driven cardiovascular outcomes trial of TRE has been published.
• Whether the AHA 2024 NHANES abstract suggesting higher CV mortality with 8-hour windows survives methodological challenge or replicates in cleaner data.
• Whether TRE is a useful adjunct on top of pharmacologic weight loss (GLP-1, GLP-1 plus GIP) or whether the appetite reduction from those drugs makes the eating-window constraint redundant.

Editorial summary

Time-restricted eating works the way calorie restriction works, by reducing total intake. The metabolic-flexibility-bonus claim has not survived rigorous isocaloric testing in larger samples. For users who find TRE easier to adhere to than calorie counting, it is a reasonable behavioral simplification. For users expecting TRE to deliver a metabolic effect that calorie restriction cannot, the trial record does not support the expectation. The lean-mass signal in TREAT is worth watching, particularly for users on a GLP-1 already losing lean mass.