Educational

The appeal of time-restricted eating is that it asks nothing about what you eat. Just when.

The complication is that compressing food into eight hours doesn't reduce the carbohydrate — it concentrates it. What happens at the meal that breaks the fast turns out to matter a great deal.

What the trials found

The study most worth knowing is a 2018 trial by Sutton and colleagues, which did something unusual: it fed both groups identical food in identical amounts, so weight loss couldn't explain the results.

Men with prediabetes ate within a six-hour window ending in mid-afternoon, or across a standard twelve-hour day. The early-window group showed improved insulin sensitivity, lower blood pressure, and reduced oxidative stress — without losing weight.

That's a genuinely interesting result, because it means the timing itself did something. Later trials have been more mixed. Several comparing time-restricted eating with straightforward calorie restriction find similar outcomes for both, which suggests that in practice much of the benefit comes from eating less rather than from the schedule.

Early windows beat late ones

This is the most consistent finding in the area, and it runs against how most people actually do it.

Insulin sensitivity follows a daily rhythm and is better in the morning. The same meal eaten at 8 a.m. and at 8 p.m. produces different glucose responses in most people, with the evening version consistently worse.

Most people who compress their eating window skip breakfast and eat late, because that's the socially convenient version. That's the arrangement with the least support behind it. An eating window from roughly 8 a.m. to 4 p.m. matches the physiology; noon to 8 p.m. works against it.

The fasting-glucose surprise

People new to fasting often see morning glucose go up, and conclude something has gone wrong.

Usually nothing has. During a fast, the liver maintains blood sugar by releasing stored glucose and by making new glucose. Add the normal early-morning cortisol rise and you get a fasting number that can sit higher than it did when you were eating breakfast.

It's a physiological adaptation rather than a failure. What matters more is what happens across the day, and whether HbA1c and post-meal numbers improve over weeks.

Breaking the fast is the part that gets neglected

Here's where glycemic load comes back in, and it's the practical heart of this.

After 16 hours without food, the first meal lands on a system primed to absorb. If that meal is fast carbohydrate — pastry, cereal, a large bowl of white rice — the resulting spike can be larger than the same food would produce at any other time.

So the composition of the first meal matters more, not less, than usual. Protein and fat first, vegetables alongside, carbohydrate present but not leading. A meal built that way blunts the response substantially. A modest, well-built first meal, followed by the larger one an hour or two later, works better than one enormous plate at the moment the window opens.

The other common failure is arithmetic. Sixteen hours of fasting followed by three thousand calories in eight is not a deficit, and it puts a large carbohydrate load through a short window.

Who shouldn't do this without supervision

If you take insulin or a sulfonylurea, fasting introduces a genuine hypoglycemia risk, and doses that were correct for three meals a day will not be correct for one or two. That's a dose-adjustment conversation before starting rather than a discovery to make at hour fourteen.

Type 1 diabetes adds ketone risk to the picture, and prolonged fasting on an SGLT2 inhibitor raises the risk of ketoacidosis at normal glucose levels. Anyone with a history of disordered eating should be cautious for reasons that have nothing to do with glucose, since restrictive eating windows can reactivate patterns that were difficult to leave behind. Pregnancy is a clear no.

What it's genuinely good at

Two things, in my reading of the evidence.

It removes late-night eating, which is where a lot of unplanned carbohydrate lives and where insulin sensitivity is worst. For some people that single change accounts for most of the benefit, and it's available without adopting the whole framework.

And it's a simple rule. Not eating between 7 p.m. and 9 a.m. requires no counting, no tracking, and no decisions. Simple rules survive busy weeks in a way that detailed plans don't, and adherence beats optimisation.

What it does to the rest of the day

Two knock-on effects show up often enough to plan for.

The first is that appetite doesn't distribute evenly across a shortened window. Most people find the first meal is enormous and the last one is small, which loads carbohydrate into a single sitting. Deliberately splitting the window into two moderate meals rather than one large and one token works better on both glucose and satiety.

The second is caffeine. Black coffee during the fasting window is standard practice, and caffeine on an empty stomach raises glucose modestly in some people through the same adrenaline route that stress does. If your fasting numbers are higher than the fast alone would explain, the coffee is worth testing as a variable.

How I'd approach it

Start with a twelve-hour overnight fast, which is close to normal eating for most people and asks almost nothing. If that's comfortable, extend toward ten hours of eating, then eight, weighting the window earlier rather than later.

Keep the meals themselves built the way they should be built regardless — protein, vegetables, moderate carbohydrate. Fasting doesn't repeal any of that. It's a container, and what goes in it still decides most of the outcome.