Does intermittent fasting really work for weight loss?

A little, and mostly because it makes you eat less. Pooled across 20 randomized trials and 1,242 people, a daily eating window took off 1.39 kg (Sun et al., International Journal of Behavioral Nutrition and Physical Activity 2025, PMID 41034862). When calories are held still the effect thins out: in the strictest test run so far, every meal was cooked and handed to 41 participants at a fixed calorie level, and 12 weeks of time-restricted eating changed weight by 0.3 kg (Maruthur et al., Annals of Internal Medicine 2024, PMID 38639542). Two honest qualifications: that trial was small enough that its result could have fallen anywhere between 1.2 kg for the window and 1.9 kg against it, and the same meta-analysis found that adding a window on top of prescribed calorie cutting took off a further 0.94 kg. The best reading is that a window works mainly by reducing intake, with a small extra effect that cannot be ruled out.

What is the best eating schedule for health?

No schedule has been settled on, and the one you can repeat beats the one that is theoretically best. The largest timing trial randomized 197 people to an early window, a late window, or one they chose, and measured deep belly fat by MRI: no window beat Mediterranean-diet counselling alone on that measure, though the early window did lower fasting and overnight blood sugar compared with the other groups (Dote-Montero et al., Nature Medicine 2025, PMID 39775037). The most-quoted human study behind eating windows had 19 people and no comparison group, and although nobody asked them to eat less, their own food logs show intake fell about 9 percent (Wilkinson et al., Cell Metabolism 2020, PMID 31813824). The same team later ran the randomized version, TIMET, in 108 people and found HbA1c improved by 0.10 percentage points (Manoogian et al., Annals of Internal Medicine 2024, PMID 39348690): statistically real, clinically small. Pick 8 to 10 hours you can hold.

Is a longer fast better than a daily eating window?

For fat loss the harder protocols have generally won. Pooled across 99 randomized trials and 6,582 people, alternate-day fasting was the only pattern that beat ordinary calorie cutting, by 1.29 kg, and time-restricted eating was not (Semnani-Azad et al., BMJ 2025). A 271-person trial found alternate-day fasting took off more body fat than a daily window over 12 weeks, although the two did not differ significantly on weight (Cai et al., BMC Gastroenterology 2019, PMID 31852444). A 2025 trial that scanned participants by MRI showed the mechanism: the alternate-day group ate 34 percent less, the window group 15 percent less (Derron et al., Clinical Nutrition 2025, PMID 40945487). In that trial alternate-day fasting also lowered resting metabolic rate, and harder protocols are the ones people abandon.

Does fasting trigger autophagy in humans?

Nobody has shown that it does. Autophagy is cells breaking down damaged parts to reuse them, and it is well established in animals. The best human study on the question opens by stating that the animal finding has never been shown in people (Bensalem et al., Journal of Physiology 2025, PMID 40345145). That trial did find a marker of autophagy move after six months, but only in an analysis run after the fact, and its participants were fasting 20 hours a day three days a week rather than keeping a daily window. The only other human study read a single gene in the blood of 11 people over four days (Jamshed et al., Nutrients 2019, PMID 31151228), and a gene reading is not a measurement of autophagy. No study has tested whether an ordinary 8-to-10-hour eating window affects it at all.

Is time-restricted eating bad for your heart?

There is one worrying study and it is much weaker than its headline. A 2024 conference summary linked an eating window under 8 hours to a 91 percent higher risk of cardiovascular death; that figure should not be used, because the finished paper reports something different. Published in 2025, it followed 19,831 people for about eight years and found a hazard ratio of 2.35 for eating inside under 8 hours, compared with a 12-to-14-hour window, with no link to death from all causes or from cancer (Chen et al., Diabetes and Metabolic Syndrome 2025, PMID 40849219). It is observational, and eating habits were estimated from two days of food diaries. Donald Lloyd-Jones, then president of the American Heart Association, argued the link may run backwards, since illness, shift work and disordered eating all shorten an eating window; 34 researchers published a formal reply making the same case, though they are the field's own scientists (PMID 41218440). Note that the signal was specific to windows under 8 hours, not to the 8-to-10 hours usually recommended.

Nutrition·Aug 10, 2026·7 min read

Does When You Eat Matter More Than What You Eat?

No, what you eat still wins. But the honest answer is more useful than either slogan: for weight the clock does almost nothing on its own, for blood sugar and blood pressure it may do something real, and there is one cost most articles skip.

Updated August 2026 · Reviewed by Roy Sañudo S.

Key Takeaways

No: what you eat and how much of it still wins. When you eat is mostly a lever on how much
When every meal was cooked and served at a fixed calorie level, the eating window changed weight by 0.3 kg
At fixed calories, timing did lower blood pressure and insulin resistance in one supervised trial, but it had 8 men and their triglycerides rose sharply
Lean mass fell by about half a kilo across 20 trials whether or not calories were cut. Keep protein up
The viral '91 percent higher heart death' figure is superseded, and the signal was for windows under 8 hours, not the 8 to 10 usually recommended

You have probably been told to stop eating after six, or to squeeze your meals into an eight-hour window. Maybe you tried it. Maybe the scale moved a little and you were never sure whether it was the clock or just the smaller portions.

That is the actual question, and the trials have now answered it.

What did the famous mouse experiment show?

Two groups of mice ate the same high-fat food, one around the clock and one only inside an 8-hour window. After 18 weeks the around-the-clock group weighed 47.4 g against the other group's 34.2 g, with liver trouble and blunted insulin, the hormone that manages blood sugar (Hatori et al., Cell Metabolism 2012, PMID 22608008).

Two things before you get excited. The paper says both groups ate the same amount, but that is a finding of *no detectable difference* across 24 cages, not a measured match, and no calorie figure appears anywhere in it. And mice are nocturnal, on a laboratory light cycle, already eating most of their food at night. Their day is not your day.

Does an eating window help you lose weight?

A little, and not because of the clock. Pooled across 20 randomized trials and 1,242 people (Sun et al. 2025, PMID 41034862), a daily window takes off 1.39 kg.

Then researchers started holding calories still, and it mostly vanished. With 116 adults eating freely (Lowe et al., JAMA Internal Medicine 2020), a 16:8 window gave no advantage. With 139 people on the same calorie target for a year (Liu et al., NEJM 2022, PMID 35443107), the window group lost 1.8 kg more, which was not significant. And in the strictest test run so far (Maruthur et al., Annals of Internal Medicine 2024, PMID 38639542), where every meal was cooked and handed to 41 people at a fixed calorie level, twelve weeks of time-restricted eating changed weight by 0.3 kg. Noise.

That evidence cuts both ways, so be fair to it: with 41 people the result could have landed anywhere from 1.2 kg for the window to 1.9 kg against, so it does not prove the clock is useless. And the same meta-analysis found that adding a window on top of prescribed calorie cutting took off a further 0.94 kg. Most likely reading: the window works by making you eat less.

What changes when the calories stay the same?

Here the clock does earn something, and here the evidence is thinnest.

The one trial built to test it fed every meal under supervision and held weight constant: five weeks of an early 6-hour window (Sutton et al., Cell Metabolism 2018, PMID 29754952) improved insulin resistance and dropped blood pressure about 11 points, with no weight loss at all. But it was eight men with prediabetes finishing their last meal before 3pm, and their triglycerides rose sharply, which most summaries leave out.

Larger work has not confirmed it. A crossover trial in 31 women (Peters et al., Science Translational Medicine 2025, PMID 41160666) shifted their body clocks and changed nothing metabolic. And in the largest timing trial, 197 people scanned by MRI (Dote-Montero et al., Nature Medicine 2025, PMID 39775037), no window beat plain Mediterranean-diet advice on deep belly fat, though the early window did lower fasting and overnight blood sugar. Timing is not irrelevant. It is just not a weight-loss lever.

Is a longer fast better?

For fat loss, yes. Across 99 randomized trials and 6,582 people (Semnani-Azad et al., BMJ 2025), alternate-day fasting was the only pattern that beat ordinary calorie cutting, by 1.29 kg. A daily window was not. A 2025 trial that scanned its participants (Derron et al., Clinical Nutrition, PMID 40945487) shows why: the alternate-day group ate 34% less, the window group 15% less. It also lowered their resting metabolic rate, and it is the protocol most people quit.

And autophagy, the cellular clean-out fasting is famous for? It has never been demonstrated in people. The detail is in the FAQ below.

Is a short eating window bad for your heart?

You may have seen the 2024 headline: eating inside 8 hours, a 91% higher risk of dying of heart disease. Do not use that number. It came from a conference summary. The finished 2025 paper (Chen et al., PMID 40849219) followed 19,831 people for about eight years and reported a hazard ratio of 2.35 against a 12-to-14-hour window, with no link to death from all causes or from cancer.

They watched people rather than testing anything, and worked out eating habits from two days of food diaries. Donald Lloyd-Jones, then president of the American Heart Association, argued the link may run backwards, since illness, shift work and disordered eating all shorten an eating window. Another 34 researchers replied in print (PMID 41218440) that two days of recall cannot support a mortality claim, though they are the field's own scientists.

And the signal was for windows under 8 hours. The 8 to 10 hours recommended below sits outside it.

The part most articles skip: muscle

In the JAMA trial, the time-restricted group lost significantly more lean mass: roughly 65% of what they lost was muscle, against a normal 20 to 30%. The authors flagged it as a caution. The 20-trial pooled analysis found the same, about half a kilo of lean mass gone with or without calorie cutting, and no improvement in body-fat percentage. The year-long NEJM trial did not see it. None of these trials prescribed extra protein.

Who should not try this

Talk to a doctor first if any of these apply: a history of an eating disorder (clinicians treat this as a hard stop, because a rule about when you are allowed to eat is an easy place for an old pattern to hide); insulin, or any medication that lowers blood sugar, where long gaps carry real risk; pregnancy or breastfeeding; being underweight, or being older and already losing strength, because of the muscle finding above.

Teenagers should not do this on their own. A growing body needs steady food, and restrictive eating rules are a known on-ramp to disordered eating.

What to do tomorrow

Tomorrow, change nothing. For three days, just write down the time of your first bite and your last. That is your real window. Then pull it in by 30 minutes a week until you reach 8 to 10 hours.

Protect dinner rather than breakfast. In one adherence study, people managed to push breakfast back 70% of the time and managed to skip the evening meal only 48% of the time. It was 16 people, so treat it as a hint. Dinner is the meal with other people in it. Worth knowing the trade: the only trial that improved blood pressure and blood sugar used a window ending before 3pm, which almost nobody sustains. Protecting dinner is the advice people actually follow, and it gets you the eat-less effect rather than that metabolic extra.

Keep protein up, because muscle is what you can lose here. Water and plain tea do not count against the window, and no human trial shows black coffee undoes anything. Miss a day and nothing is undone.

Then judge it in a month by how you feel, and by your blood pressure and blood sugar rather than the scale.

What you eat still matters more. When you eat is a lever on how much you eat, plus a possible small bonus. If you want the bigger levers, we have looked at the tests that claim to measure your biological age and at what the evidence says about keeping your thinking sharp. Here is how we check claims like these, and here is how I live with it.

Where the science stands

Our read after cross-checking every study and expert view.

Key findings

Strong evidenceFor weight, the clock does almost nothing on its own. In the strictest test run so far, every meal was cooked and handed to 41 participants at a fixed calorie level, and 12 weeks of time-restricted eating changed weight by 0.3 kg (Maruthur et al., Annals of Internal Medicine 2024). Pooled across 20 trials and 1,242 people, a window takes off 1.39 kg, which is about what eating slightly less produces on its own.
Early signalFor blood sugar and blood pressure at fixed calories, the picture is genuinely open. One supervised trial that held weight constant found blood pressure dropped about 11 points and insulin resistance improved, but it ran on 8 men with prediabetes finishing dinner before 3pm, and their triglycerides rose sharply. A newer crossover in 31 women shifted their body clocks and changed nothing metabolic.
Moderate evidenceLean mass is the consistent cost. Across 20 randomized trials it fell by about half a kilo whether or not calories were cut, and the proportion of body fat did not improve. In one trial roughly 65 percent of the weight lost was muscle, against a normal 20 to 30 percent, and its authors flagged it as a caution.
Strong evidenceThe harder protocols beat the daily window on fat loss, for the least mysterious reason available: people eat less on them. Pooled across 99 trials and 6,582 people, alternate-day fasting was the only pattern that beat ordinary calorie cutting; time-restricted eating was not.

Where the research agrees

That a shorter eating window makes most people eat less, and that this is where most of the benefit comes from. Everyone lands here, including the trials that found no benefit at all, and including the researchers who defend the practice most strongly.

Where it's still debated

Whether the clock does anything once intake is held equal. The positive trials are consistently tiny and short; the null trials are the larger and longer ones. That pattern can mean the effect is small and real, or that it is not there. Settling it needs a controlled-feeding trial large enough to measure blood sugar and blood pressure as primary endpoints, and nobody has run one.

Scientific References

This article synthesizes research from the following institutions and studies. All content is derived from peer-reviewed scientific literature and leading research centers.

• Salk Institute for Biological Studies, Time-Restricted Eating Research (Dr. Satchin Panda)

• Hatori et al., Cell Metabolism: "Time-Restricted Feeding without Reducing Caloric Intake Prevents Metabolic Diseases in Mice Fed a High-Fat Diet" (2012). PMID 22608008. DOI 10.1016/j.cmet.2012.04.019. Mouse study — the widely cited "same food, opposite result" experiment.

• TIMET Clinical Trial (Wilkinson, Panda et al.): "Ten-Hour Time-Restricted Eating Reduces Weight, Blood Pressure, and Atherogenic Lipids in Patients with Metabolic Syndrome" (Cell Metabolism, 2020). PMID 31813824. n=19, single-arm, no control group.

• Lowe D.A. et al., JAMA Internal Medicine: "Effects of Time-Restricted Eating on Weight Loss and Other Metabolic Parameters in Women and Men with Overweight and Obesity" (TREAT, 2020). Randomized, n=116; time-restricted eating showed no weight advantage over three structured meals a day.

• Liu D. et al., New England Journal of Medicine: "Calorie Restriction with or without Time-Restricted Eating in Weight Loss" (2022). DOI 10.1056/NEJMoa2114833. Randomized; with calories matched, the eating window added nothing.

Curated by Roy Sañudo S.

Research Curator, Anima Cosmi

Anima Cosmi translates peer-reviewed research from Harvard, Oxford, Stanford, the Salk Institute, and other leading labs into plain language. Every article names its researchers, institutions, and publication years, so you can check the sources yourself.

Not medical advice. Anima Cosmi is a research curator, not a medical practice: this is education, not a prescription. Talk to a qualified healthcare provider before starting any new supplement, diet, or routine, especially if you take medication or have a health condition.