Few areas of nutrition have generated more confident advice on less evidence than meal timing. The advice has also reversed itself completely within living memory. In the 1990s and 2000s, the consensus among trainers and fitness magazines was that you should eat six small meals a day to keep your metabolism elevated. By the late 2010s the same publications were telling you to compress all your eating into an eight-hour window, or skip breakfast entirely, or eat once a day.
Both eras offered a plausible mechanism. Both were largely wrong. And the randomised trials that settled the question are more interesting than either claim.
The Metabolism Myth Was a Statistical Artefact
Start with the older claim, because its demolition is unusually clean.
The idea that frequent meals raise metabolic rate rested on a real observation: eating produces a measurable rise in energy expenditure, the thermic effect of food. The error was assuming that more eating occasions means more thermic effect. It does not. The thermic effect scales with the total calories and macronutrient composition consumed, not with how many sittings you divide them into. Six 400-calorie meals and three 800-calorie meals produce the same total thermic effect, because the total is the same.
The epidemiology that appeared to support frequent eating turned out to be contaminated. Bellisle, McDevitt and Prentice examined this directly in the British Journal of Nutrition in 1997 and concluded that the inverse relationship between meal frequency and body weight in observational data is almost certainly an artefact, driven by confounding and by the well-documented tendency of people with higher body weight to under-report both what and how often they eat.
Schoenfeld, Aragon and Krieger then ran the meta-analysis in Nutrition Reviews in 2015. Pooling the intervention studies, the apparent body composition advantage of higher meal frequency was driven entirely by a single outlier study. Remove it and the effect disappears.
Time-Restricted Eating Met Its Control Group
The modern claim received a harder test, and did not do much better.
The TREAT trial, published by Lowe and colleagues in JAMA Internal Medicine in 2020, randomised 116 adults with overweight or obesity to either 16:8 time-restricted eating or a consistent three-meal pattern for 12 weeks. This was the design the field needed: a real control group, not a before-and-after comparison.
The time-restricted group lost weight within their own group, 0.94 kg (95% CI 1.68 to 0.20; P=0.01). But the comparison that matters is against the control group, and there the difference was 0.26 kg (95% CI 1.30 to 0.78; P=0.63). Statistically indistinguishable from nothing.
The secondary finding was worse. Participants in the time-restricted arm lost significantly more appendicular lean mass than controls, a between-group difference of 0.16 kg/m2 (95% CI 0.27 to 0.05; P=0.005). Whatever weight they did lose contained a greater proportion of muscle.
A longer and larger trial reached a similar verdict. Liu and colleagues, publishing in the New England Journal of Medicine in 2022, followed participants for 12 months comparing calorie restriction plus time-restricted eating against calorie restriction alone. The combined group lost 8.0 kg (95% CI 9.6 to 6.4). The calorie restriction group lost 6.3 kg (95% CI 7.8 to 4.7). The difference of 1.8 kg was not statistically significant (95% CI 4.0 to 0.4; P=0.11).
Read those two trials together and the conclusion is not that time-restricted eating fails. It is that time-restricted eating works exactly as well as the calorie deficit it produces, and not one gram better. For many people an eating window is a genuinely effective way to eat less without counting anything, which is a real and useful benefit. It is just not a metabolic mechanism.
The Autophagy Claim Is Running Far Ahead of the Data
Autophagy is the most oversold concept in the fasting literature, and the gap between what is claimed and what has been shown in humans is wide enough to be worth spelling out.
The science is real and important. Yoshinori Ohsumi won the 2016 Nobel Prize in Physiology or Medicine for working out the mechanisms of autophagy. That work was done in yeast.
The human fasting evidence is thinner than almost anyone repeating it realises. It consists largely of surrogate markers measured in circulating blood cells. Pietrocola and colleagues, publishing in Autophagy in 2017, found enhanced autophagic flux in human white blood cells during fasting, but the effect was only detectable using ex vivo protease inhibition, a laboratory manipulation rather than a direct in vivo observation. A 2025 analysis in the Journal of Physiology found a between-group difference in peripheral blood mononuclear cell autophagy that emerged only at six months and only in a post-hoc exploratory analysis (P=0.04). A fasting-mimicking diet pilot trial published in GeroScience in 2025 found no significant difference in autophagic flux at all (P=0.42).
To be direct about what is missing: there is no study showing that fasting induces autophagy in human neurons, human liver, or any human solid organ in vivo. The claim that a 16-hour fast is cleaning out damaged proteins in your brain is an extrapolation from yeast and mice, not a finding in people.
Growth Hormone, and the Translation Error
A related overreach involves growth hormone, and here the underlying studies are solid but tiny and about something else entirely.
Ho and colleagues, in the Journal of Clinical Investigation in 1988, showed that a five-day fast raised 24-hour integrated growth hormone concentrations roughly threefold, from 2.82 plus or minus 0.50 to 8.75 plus or minus 0.82 micrograms per minute per millilitre (P=0.0002). Hartman and colleagues found roughly a fivefold increase in 24-hour growth hormone production across a two-day fast.
Now look at the details. The first study had six men. The second had nine. Both involved multi-day total fasts, not eating windows. Neither tells you anything reliable about what happens when you push breakfast back to noon. Presenting these as evidence for daily intermittent fasting is a translation error, and a common one.
Protein Distribution Is the Part That Survives
One aspect of meal timing does hold up, though it is narrower than usually presented.
Moore and colleagues, in the American Journal of Clinical Nutrition in 2009, established the dose response: muscle protein synthesis was maximally stimulated at roughly 20 g of high-quality protein after resistance exercise, and 40 g produced no further synthetic response, only increased leucine oxidation. The study had six participants, which is typical for this invasive tracer methodology and worth remembering.
Areta and colleagues, in the Journal of Physiology in 2013, then tested distribution directly over 12 hours. Four 20 g feedings every three hours produced greater myofibrillar protein synthesis than either eight 10 g feedings or two 40 g boluses. Moderate doses at moderate intervals beat both grazing and stacking.
This is where the practical case against one meal a day is strongest. It is difficult to consume adequate protein in a single sitting, and the distribution evidence suggests that even if you managed the total, concentrating it into one feeding is not the optimal stimulus for muscle protein synthesis.
The narrow post-workout anabolic window, by contrast, does not survive. Schoenfeld, Aragon and Krieger examined it in the Journal of the International Society of Sports Nutrition in 2013 and found that the apparent effect of protein timing vanished once total daily protein intake was controlled for. The studies that seemed to show a window were mostly showing that the timing groups ate more protein.
So What Should You Actually Do
The honest answer is that meal frequency is a preference variable, not a lever, and that is genuinely liberating information.
Total energy intake determines weight change. Total protein intake, distributed across roughly three to four feedings of about 20 to 40 g, supports muscle. Beyond that, the number of meals should be chosen on the basis of what you can sustain, what fits your day, and what keeps you from being so hungry that you undo the whole thing at nine in the evening.
If an eating window helps you eat less without tracking, use it. If three meals suit you, use those. If you are training hard, the case for spreading protein and against a single daily meal is reasonable. None of these choices is metabolically privileged, and the confidence with which each has been sold should make you more sceptical of the next timing protocol, not less.
Where the Evidence Runs Out
Almost every trial in this area faces the same problem: adherence is self-reported, and people are unreliable narrators of their own eating. Trials also run for weeks to months, while the claims made for fasting concern outcomes over decades.
The muscle protein synthesis literature rests on very small samples, six to twelve participants being typical, because the tracer methods are invasive and expensive. Acute synthesis is also a surrogate for actual muscle growth, and the two do not always agree.
There may be populations for whom meal timing matters more than these averages suggest, particularly people with type 2 diabetes or significant circadian disruption such as shift workers, where the evidence is developing and genuinely more promising. And a null result in a 12-week trial is not proof of no effect, only proof that any effect is smaller than the trial could detect.
Anyone with diabetes, a history of disordered eating, or who is pregnant should treat extended fasting as a medical question rather than an optimisation one.
References
Lowe DA, Wu N, Rohdin-Bibby L, et al. Effects of time-restricted eating on weight loss and other metabolic parameters in women and men with overweight and obesity: the TREAT randomized clinical trial. JAMA Internal Medicine. 2020;180(11):1491-1499. doi:10.1001/jamainternmed.2020.4153. PMID 32986097.
Liu D, Huang Y, Huang C, et al. Calorie restriction with or without time-restricted eating in weight loss. New England Journal of Medicine. 2022;386(16):1495-1504. doi:10.1056/NEJMoa2114833.
Schoenfeld BJ, Aragon AA, Krieger JW. Effects of meal frequency on weight loss and body composition: a meta-analysis. Nutrition Reviews. 2015;73(2):69-82. doi:10.1093/nutrit/nuu017. PMID 26024494.
Bellisle F, McDevitt R, Prentice AM. Meal frequency and energy balance. British Journal of Nutrition. 1997;77(Suppl 1):S57-S70. doi:10.1079/BJN19970104. PMID 9155494.
Moore DR, Robinson MJ, Fry JL, et al. Ingested protein dose response of muscle and albumin protein synthesis after resistance exercise in young men. American Journal of Clinical Nutrition. 2009;89(1):161-168. doi:10.3945/ajcn.2008.26401. PMID 19056590.
Areta JL, Burke LM, Ross ML, et al. Timing and distribution of protein ingestion during prolonged recovery from resistance exercise alters myofibrillar protein synthesis. Journal of Physiology. 2013;591(9):2319-2331. doi:10.1113/jphysiol.2012.244897. PMID 23459753.
Schoenfeld BJ, Aragon AA, Krieger JW. The effect of protein timing on muscle strength and hypertrophy: a meta-analysis. Journal of the International Society of Sports Nutrition. 2013;10(1):53. doi:10.1186/1550-2783-10-53.
Ho KY, Veldhuis JD, Johnson ML, et al. Fasting enhances growth hormone secretion and amplifies the complex rhythms of growth hormone secretion in man. Journal of Clinical Investigation. 1988;81(4):968-975. doi:10.1172/JCI113450. PMID 3127426.
Hartman ML, Veldhuis JD, Johnson ML, et al. Augmented growth hormone secretory burst frequency and amplitude mediate enhanced growth hormone secretion during a two-day fast in normal men. Journal of Clinical Endocrinology and Metabolism. 1992;74(4):757-765. PMID 1548337.
Pietrocola F, Demont Y, Castoldi F, et al. Metabolic effects of fasting on human and mouse blood in vivo. Autophagy. 2017;13(3):567-578. doi:10.1080/15548627.2016.1271513.
Bensalem J, et al. Intermittent time-restricted eating may increase autophagic flux in humans: an exploratory analysis. Journal of Physiology. 2025. doi:10.1113/JP287938. PMID 40345145.