Healthy weight loss: the complete plan according to science

Healthy weight loss works when you press three buttons at once: a deficit you can keep, enough protein, and enough sleep. The numbers are sober. For every kilo you lose, your energy burn drops by 20 to 30 kcal a day. Your appetite, over that same kilo, rises by about 100 kcal a day (PMID 29156185).

So your hunger grows three to five times faster than your burn falls.

That is the heart of this plan. The metabolism that supposedly breaks is not your enemy. Your appetite is, and you can do something about that.

Below is the order that works. No diet name, no forbidden foods, no target weight. Just each step, and how big the effect honestly is.

When weight loss should not be your goal

Losing weight is not healthy for everyone, and not necessary for everyone. Is your waist under half your height, and are your blood values fine? Then weight loss buys you very little extra health. It can still be your goal. Just do not call it a health goal.

There is a reason to start here and not at the end. In young people, dieting is the strongest predictor of a new eating disorder (PMID 10082698).

That does not make weight loss taboo. It does make the way you approach it matter.

Do not follow this plan if you have or have had an eating disorder. Nor if you are pregnant or breastfeeding, under 18, underweight, using insulin, or have had bariatric surgery. In those cases, speak to your GP first.

Does the number on the scale set your mood for the rest of the day? Then that is a reason to stop weighing, not to diet harder.

Measure your starting point properly

Get a tape measure. Measure your waist and divide it by your height. Stay under 0.5, which simply means: keep your waist under half your height. That ratio says more about your health than your weight, and the UK NICE NG246 guideline uses it as the first screen (PMID 22106927).

The bands are simple. A ratio of 0.4 to 0.49 is healthy, 0.5 to 0.59 is raised. From 0.6 upward the risk is strongly raised.

And your BMI? It is less useful than you think. The 2025 Lancet Commission calls BMI a population screen, not a measure of one person's health (PMID 39824205).

BMI does not know whether you weigh 90 kilos because of muscle or because of belly fat. Your tape measure does.

Then weigh yourself sensibly. Take your weight every morning and work with a rolling seven-day average. Day-to-day swings are fluid, gut content and salt, not fat.

You can ignore the body fat percentage on your smart scale. That reading is not reliable enough to base decisions on.

Measure your waist once a month. That is your real progress meter.

Which diet? The answer is: the one you keep doing

There is no winning diet. In the DIETFITS trial, people on healthy low-carb lost 6.0 kilos and people on healthy low-fat lost 5.3 kilos over twelve months (PMID 29466592). The 0.7 kilo difference was not significant. A full year of serious testing, and it did not matter.

DIETFITS also tested the two most popular excuses. There was no interaction with genotype. There was no interaction with insulin secretion either.

So your genes do not decide which diet suits you. That idea sells well, but it did not survive the trial that studied it properly.

POUNDS LOST reached the same conclusion over two years, across every macro split (PMID 19246357).

And keto? On a metabolic ward, where every meal was controlled, people on carb restriction lost 53 grams of body fat a day. On fat restriction it was 89 grams a day (PMID 26278052). So the low-carb side lost less fat, not more.

A second keto study showed that energy expenditure rose by only 50 to 100 kcal a day, while fat loss actually slowed (PMID 27385608).

Time-restricted eating, the eight-hour window, adds close to nothing once calories are matched (PMID 32986097). A twelve-month randomised trial confirmed it (PMID 35443107).

There is one food choice that does move a lot.

When people ate ultra-processed food for two weeks, they took in about 500 kcal a day more than on unprocessed food, with the macros matched (PMID 31105044). Nobody asked them to eat more. They just did.

That is the reason to build your base on real, recognisable food. The Schijf van Vijf from the Dutch Voedingscentrum is a fine starting point, and it costs you nothing.

Then pick the format you can keep doing for a year. Adherence is the deciding variable, not the macro split.

Put your protein first

Protein is the one macronutriënt with its own job in a deficit. It protects your muscle mass and it satisfies you better than fat or carbohydrate. Aim for 1.6 to 2.0 grams per kilo of reference weight per day. Note that word reference weight, because that is where people go wrong.

What does protein buy you? A meta-analysis found that a higher protein intake in a deficit gave 0.87 kilos more fat loss and 0.43 kilos more fat-free mass retained (PMID 23097268).

In hard-training men it went further. At 2.4 grams per kilo, the protein group gained 1.2 kilos of lean mass and lost 4.8 kilos of fat. At 1.2 grams per kilo it was 0.1 kilos of lean mass and 3.5 kilos of fat (PMID 26817506).

The benefit flattens out around 1.6 grams per kilo per day (PMID 28698222). The ISSN uses 1.4 to 2.0 grams per kilo, with about 0.25 grams per kilo per meal (PMID 28642676).

Now that reference weight. At 120 kilos, 2 grams per kilo suddenly means 240 grams of protein a day. That is close to a thousand calories from protein alone, and it works against your deficit.

So use a reference weight that matches your height, or use your fat-free mass. For fat-free mass, 2.3 to 3.1 grams per kilo is recommended in a deficit (PMID 24092765).

In practice that lands most people around 1.6 grams per kilo, and up to 2.0 grams per kilo if you train seriously.

High protein does not damage healthy kidneys. With existing kidney disease it is different, and that is a conversation with your doctor. We worked it through in a high protein diet and your kidneys.

Then set your deficit

The flat rule of 500 kcal less per day does not hold up. For someone weighing 55 kilos that is almost a quarter of their daily needs. For someone at 120 kilos it is not even an eighth. So work with a share of your maintenance, not a fixed number.

A deficit of 10 to 20 percent of maintenance is manageable for most people. Bigger is sometimes possible, but it costs you more muscle and more discipline.

Never go under 1200 kcal a day without medical supervision.

There is another error you see everywhere: assuming that a 7000 kcal deficit always equals one kilo of fat. As you get lighter, your maintenance falls with you. So your deficit shrinks on its own, without you changing anything.

That is not a broken metabolism. That is arithmetic.

How to work out your deficit, with a worked example and the pitfalls, is in calculating a calorie deficit.

Pick a pace that spares your muscle

A sensible pace is 0.5 to 1.0 percent of your bodyweight per week. At 90 kilos that is 450 to 900 grams a week. Faster is possible, but you pay for it in muscle mass and in staying power. In a steep deficit your body pulls a larger share of its fuel from muscle tissue.

In elite athletes this was compared properly. The slow group, around 0.7 percent a week, actually gained fat-free mass. The fast group, around 1.4 percent a week, did not (PMID 21558571).

One thing you should know here. Losing slowly does not prevent the yo-yo effect (PMID 26813524). People who lost slowly regained just as readily.

So the reason to go moderate is not that the weight stays off. The reason is your muscle mass and your head.

Strength training is not there to burn calories

Exercising without changing what you eat produces very little weight loss. Across trials, exercise alone came out at roughly 1.6 kilos over six to twelve months (PMID 21787904). That is not an argument against exercise. It is an argument against exercise as a weight-loss strategy.

Why is it so disappointing? Your body compensates. About 28 percent of the energy you put into activity is clawed back elsewhere in your day (PMID 34453886).

Strength training does something else, and something more important. It gives your body a reason to hold on to muscle while you sit in a deficit.

Two to three times a week, heavy compound lifts, and progression in weight or reps. You do not need much more.

The full programme and the logic behind it are in strength training and weight loss.

Sleep is the most underrated lever

People who extended their sleep by about 1.2 hours a night ate 270 kcal a day less. No diet, no instruction to eat less. Intake was measured with doubly labelled water, in people living their normal lives (PMID 35129580). That is bigger than most diet rules.

Sleep loss works the other way too. With calories matched, 5.5 hours versus 8.5 hours in bed, the share of weight lost as fat fell by 55 percent. Fat-free mass loss rose by 60 percent (PMID 20921542).

So you lose the wrong tissue.

The hormones behind it have been measured. After sleep restriction, leptin fell 18 percent, ghrelin rose 28 percent and hunger rose 24 percent (PMID 15583226).

And cortisol? The cortisol belly story is far weaker than the wellness industry claims. Outside genuine Cushing's disease, cortisol does not explain your weight.

Sleep is also the cheapest intervention in this whole article. It costs you nothing, only a decision about when you put your phone down.

What to measure, and how often

Self-monitoring is the best-evidenced behaviour in the entire weight-loss literature (PMID 21185970). People who track what they eat, weigh and do, lose more. The caveat matters: calorie counting is not suitable for anyone with a disordered relationship with food. If that is you, skip this section.

Keep it small. Weight daily, your average weekly. Waist monthly, and your protein intake on a few days a month.

Your training log is your fourth meter. If your lifts hold or climb while you lose weight, you are keeping muscle.

Below is what actually happens in the studies, with the honest effect sizes.

What you change	What it honestly delivers	Source
Exercise alone, without changing what you eat.	About 1.6 kilos over six to twelve months.	PMID 21787904
Eating low-carb instead of low-fat.	0.7 kilos difference after a year, and it was not significant.	PMID 29466592
Keeping protein high during your deficit.	0.87 kilos more fat lost and 0.43 kilos more muscle kept.	PMID 23097268
Sleeping about 1.2 hours longer per night.	270 kcal a day less eaten, with no diet at all.	PMID 35129580
Bringing less ultra-processed food into the house.	About 500 kcal a day less, with macros matched.	PMID 31105044
Squeezing your meals into an eight-hour window.	Close to nothing, once calories are matched.	PMID 32986097

Look at that table again. The biggest numbers sit with sleep and with the food in your house, not with the gym.

The plateau: it is your hunger, not your metabolism

This is the most important number in this article. For every kilo you lose, your energy burn falls by 20 to 30 kcal a day. But your appetite rises by roughly 100 kcal a day per kilo lost (PMID 29156185). Your hunger grows three to five times faster than your burn falls.

A plateau is almost never a stalled metabolism. It is a deficit that has quietly closed, because you eat a little more and move a little less without noticing.

There is a real metabolic adaptation, and it is larger than your change in body composition alone explains (PMID 7632212). In The Biggest Loser participants, resting metabolic rate was still about 500 kcal a day below prediction six years later (PMID 27136388).

That sounds like a broken metabolism, but it is not. That study involved an extreme deficit and extreme training volumes, far beyond anything you are doing.

For most people the plateau sits in the fork and the diary, not in the thyroid.

Why your plateau is not a slow metabolism is worked through in the weight loss plateau. And whether a slow metabolism exists at all is covered in slow metabolism.

Plan your maintenance before you start

Maintenance is not a rest stop. Holding on to lost weight costs you a permanent 300 to 500 kcal a day of effort (PMID 29156185). People who do not plan for it regain. Not out of weakness, but because the body actively pushes back. Count on it from day one.

We know quite precisely how stubborn that push-back is. A full year after losing 13.5 kilos, appetite hormones were still set to hunger (PMID 22029981). Not a few weeks. A year.

On average, more than 80 percent of lost weight comes back within five years (PMID 29156185). A meta-analysis landed at roughly 3 kilos, or 3 percent, still gone after five years (PMID 11684524).

Still, the picture is not hopeless. In Look AHEAD, 27 percent of participants held at least a 10 percent loss after eight years, against 17 percent in the control group (PMID 24307184).

And that famous line that 95 percent of diets fail? It is not true. It traces back to a case series from 1959 and was never a reliable figure.

In the American National Weight Control Registry, more than 87 percent of participants held at least a 10 percent loss after ten years (PMID 24355667). Read that number carefully. It is a volunteer registry of people who were already succeeding.

So it shows what maintainers do, not how often it works.

And what do they do? They weigh themselves regularly, they eat fairly consistently, and they move a lot. Boring, and that is exactly the point.

On GLP-1 medication we should be honest here. Semaglutide 2.4 mg produced 14.9 percent weight loss versus 2.4 percent on placebo at 68 weeks (PMID 33567185). Tirzepatide came in at 15.0 to 20.9 percent versus 3.1 percent (PMID 35658024).

But stopping means regaining. After semaglutide was withdrawn, two-thirds of the loss came back (PMID 35441470). In SURMOUNT-4 the group that continued lost a further 5.5 percent, while the placebo group regained 14.0 percent (PMID 38078870).

These are chronic medicines, not a course of treatment. Which blood values to monitor alongside them is covered in what GLP-1 medication is.

The biology behind the yo-yo effect, and what does help against it, is in the yo-yo effect.

Is it your hormones?

Almost never. A hormonal cause of obesity is rare, and running a routine hormone panel because weight loss is not working is not indicated. Cushing's disease occurs in roughly 1 to 2 people per million per year. That is close to nobody.

An underactive thyroid gives you a few kilos, and those kilos are mostly fluid, not fat. Subclinical hypothyroidism should not be treated in order to lose weight.

In fact, a raised TSH is often a consequence of excess weight and normalises again after weight loss. There is more on that in the thyroid and your weight.

Insulin resistance is mostly a consequence of fat mass, not its cause (PMID 28074888). Weight loss improves your insulin sensitivity, not the other way round. In the DiRECT trial a substantial share of participants even put type 2 diabetes into remission after losing weight (PMID 29221645, PMID 30852132).

PCOS is the exception that does deserve attention. It affects roughly 10 to 13 percent of women, and it genuinely makes losing weight harder.

And then the cause almost nobody mentions: your medicine cabinet.

Medication is the most under-recognised real cause of weight gain. Olanzapine added an average of 2.4 kilos, pioglitazone 2.6 kilos (PMID 25590213). Mirtazapine, paroxetine, lithium, valproate, gabapentin, pregabalin, insulin, sulfonylureas and corticosteroids are on that list too.

Never stop a medicine on your own. Put the question to the doctor who prescribed it, because there is often an alternative.

So what is a blood test good for? A blood test does not tell you why the scale has stopped moving. It shows what your weight is doing to your health, and lets you watch that improve while you lose.

That is a different goal, and a more honest one. Your blood sugar, your HbA1c, your cholesterol and your liver values move with weight loss, often before the scale shows anything.

A complete metabolic panel puts those values side by side, before and after. What that gives you in practice is described in what weight loss does to your blood values. And how to spot insulin resistance is in recognising insulin resistance.

What such a test does not do is tell you why you are stuck. That answer lives in your week, not in your blood.

When this article is not for you

Weight loss is not a healthy goal for everyone. Do not follow this article if you have or have had an eating disorder, are pregnant or breastfeeding, are under 18, are underweight, use insulin, or have had bariatric surgery. Speak to your GP first.

If thoughts about food or your body start to take over, talk to your GP or contact an eating disorder helpline. That is not weakness. That is sense.

Your first step this week

Get a tape measure tonight. Measure your waist at the level of your navel and divide that number by your height. Write it down with the date, and repeat it in four weeks.

Then do one thing, not five. Count your protein intake for three days, without changing anything else.

Most people are startled by how low that number comes out. That is exactly where your plan begins.