Time-restricted eating: what changes, and what only looks like it changed

What the trials actually found

When randomised trials compare a compressed eating window against the same number of calories spread across the day, the difference in weight lost is small or absent. That result has now shown up often enough — including in year-long trials — that it is the reasonable default assumption. This does not mean time-restricted eating fails. It means the mechanism is mundane: a shorter window is a rule that makes most people eat less without counting anything, and eating less is what moves the scale. A method that works because it is easy to follow is still a method that works. It is just not doing anything metabolically exotic.

Where the window does matter: the hour, not the length

Your metabolism is not the same at nine in the morning as it is at eleven at night. Insulin sensitivity and glucose tolerance are measurably better earlier in the day and worse late in the evening, which is one of the more solid findings in circadian physiology. The consequence is that two eight-hour windows are not equivalent: eating from eight to four is not the same intervention as eating from noon to eight, even though both fit the same slogan. Where the studies do find advantages beyond calorie reduction, they tend to be in the earlier windows.

The part that shows up in your sleep

For most people the fastest and most reliable effect of shifting an eating window is not on the scale at all — it is on the night. A meal close to bedtime raises core temperature and keeps heart rate elevated into the first hours of sleep, and both of those are visible in overnight data long before any change in weight is. The trap in late-window schedules is that skipping breakfast is easy and skipping dinner is not, so the window drifts later, the last meal lands at nine, and you end up losing weight and sleeping worse at the same time.

How to tell which thing actually changed

Starting time-restricted eating changes three things at once: how long the window is, how many calories go through it, and what time the last meal lands. Those get credited to the diet as a single lump, which is why people end up believing things about it that are not true. Your own data can separate them if you let it — the weight trend answers the calorie question over weeks, the overnight resting heart rate and the first hours of sleep answer the timing question within days, and sleep latency answers whether the last meal moved too late. Tag the day you start, then compare three or four weeks either side rather than reading individual days. In Vitra that comparison happens on your own machine, against your own history, with nothing sent anywhere.

Who should be careful

A rule about when not to eat is a poor fit for some people. Anyone with a history of disordered eating should treat rigid eating windows as a genuine risk rather than a neutral tool. It is also a poor fit during pregnancy, for people taking glucose-lowering medication, and for anyone doing heavy training who needs the fuelling opportunities — a compressed window is a hard way to eat enough. If any of those apply, this is a conversation for a doctor before it is a conversation for a chart. This is not medical advice.

Frequently asked questions