Calorie targets

Eating too little to lose weight? What the research actually found.

Cutting harder does stop working. But not for the reason everyone says. Your metabolism is not the thing that gives out first. Your appetite is.

This page goes through what has actually been measured: how much metabolic slowdown really happens, why the scale stalls, and what the trials on fast versus slow dieting found. Every figure is attributed. Some of it contradicts advice you have been given, including advice on this subject that we think is well-meant and wrong.

Published 26 August 2026. This is a summary of published research, not medical or dietary advice.

Why has my weight loss stopped?

Almost always because the deficit closed, not because the body sealed itself shut. Three things drift at once, and none of them announces itself: appetite climbs, portions grow, and logging gets looser. The number on the app stays the same. The number going in does not.

Franz’s 2007 review of 80 trials with at least a year of follow-up found weight loss tends to plateau around six months, across diets that had almost nothing in common with one another. When the same curve shows up regardless of what people were told to eat, the cause is more likely to be how long anyone can hold a behaviour than anything about macronutrients.

Is “starvation mode” real?

The mechanism is real. The story built on top of it is not.

Adaptive thermogenesis. The technical name. Was established in 1995, when Leibel, Rosenbaum and Hirsch held people 10% below their usual weight on a metabolic ward and found energy expenditure dropped by 6 to 8 kcal per kg of lean mass per day beyond what the smaller body explained. Later work found the effect is largely driven by leptin: restore it, and expenditure comes back up.

So how big is it? Müller’s 2015 study put 32 men through three weeks at half their energy needs, with whole-body MRI and isotope dilution to measure body composition properly. Resting expenditure fell 266 kcal a day. Of that, 108 kcal was adaptation rather than lost tissue, and roughly 72 kcal once shrinkage in the organs themselves was accounted for.

Seventy to a hundred calories a day. Real, measurable, worth knowing about, and nowhere near enough to stop a deficit that is genuinely being held.

The Biggest Loser study, and what it actually said

This is the study everyone cites. Fothergill and colleagues followed 14 contestants for six years and found resting metabolic rate still 704 kcal a day below baseline, with most of the weight regained. It gets quoted as proof that dieting permanently breaks your metabolism.

Read the paper and that reading falls apart. Metabolic slowing at the end of the competition did not predict who regained weight. The correlation was −0.1, with a p-value of 0.75. The association that was significant ran the other way: the contestants who kept the most weight off had the most metabolic slowing. The slowdown was a consequence of staying lighter, not the cause of failure.

A companion paper on the same people found what did separate them. The maintainers had increased their physical activity by 160%; the regainers by 34%. Regain tracked activity closely (r = −0.82) and not intake at all (r = −0.15). Fourteen people, no control group, after a televised contest involving four to six hours of daily exercise, so generalise carefully. But if you are going to cite the study, cite what it found.

“I’m eating 1,200 calories and not losing weight”

It is usually not 1,200 calories. This is the least popular sentence on this page and the best evidenced.

In 1992, researchers took ten people who reported eating under 1,200 kcal a day and believed themselves resistant to dieting, and measured them properly. Doubly labelled water for expenditure, indirect calorimetry for metabolic rate. Their metabolic rates were normal, within 5% of predicted. Their actual intake was 47% higher than they had reported. They were also over-reporting their exercise by about half.

That was ten hand-picked people, so treat the 47% as an illustration rather than a population figure. The population figure exists too: Freedman’s 2014 pooling of five large validation studies, checked against recovery biomarkers, found average under-reporting of 15% to 28% depending on the method, with higher body mass predicting bigger gaps.

Nobody in those studies was lying. Oil in the pan, the last third of the children’s dinner, a handful of something while cooking, a splash more milk than you pictured. Measurement error runs one direction, and it runs there for everybody.

Does losing weight fast mean gaining it back fast?

Almost everyone believes this. The trials do not support it.

Purcell’s 2014 study randomised 200 people to lose 15% either rapidly over 12 weeks or gradually over 36, then followed them for three years. The proportion of lost weight regained was identical. 76.3% in both arms by intention to treat. More of the rapid group actually reached the target in the first place: 81% against 50%. The authors wrote plainly that their findings were “not consistent with present dietary guidelines which recommend gradual over rapid weight loss”.

Nackers’ earlier analysis pointed the same way: the fastest early losers had lost the most at 18 months and were regaining no faster than anyone else.

One caveat that matters more than the finding. The rapid arm was a supervised, nutritionally complete meal-replacement programme run through a hospital. That is not the same activity as deciding by yourself to eat 800 calories a day, and the trial says nothing about the second one. What it does retire is the folk claim that speed itself causes rebound.

So what does going too low actually cost?

Not the thing people fear. Several other things instead, and they are better documented than the metabolic story.

Hunger that does not settle down. Sumithran’s 2011 study in the New England Journal of Medicine followed 50 people for a year after a 13.5 kg loss. Leptin, ghrelin, peptide YY and the rest were still shifted from baseline at 62 weeks, and subjective hunger was still significantly elevated. A year later. That is a physiological signal, not a willpower problem, and it is the single best- evidenced reason a deficit you cannot sit with will not last.

Lean mass, if protein and training are wrong. Longland’s 2016 trial ran 40 men at a 40% deficit with resistance training. At 2.4 g of protein per kg of bodyweight they gained 1.2 kg of lean mass; at 1.2 g they gained 0.1 kg, and lost less fat. The deficit was identical. What differed was what was eaten inside it.

Cycle disruption, with a dose-response. Williams’ 2015 controlled-feeding study found menstrual disturbances tracked the size of the energy deficit, appearing at deficits of roughly 470 to 810 kcal a day. Not an abstract risk. A measured relationship between how deep you go and how likely it is.

And the thing that actually decides the outcome. Dansinger’s 2005 trial put 160 people on four different diets. After a year, weight loss correlated with how well people stuck to the diet they were given (r = 0.60) and not at all with which diet that was (r = 0.07). Adherence was the whole story. A deficit you can hold for a year beats a deeper one you abandon in March, and it is not close.

How low is too low?

Current UK guidance sets the line higher than most people expect. NICE guideline NG246, published in 2025, says diets of 800 to 1,200 kcal a day belong only inside a specialist weight-management service; diets under 800 only where there is a clinically assessed need for rapid loss; and neither for more than 12 weeks or without clinical supervision.

Which is worth sitting with, because 1,200 is the number a great many people set for themselves on a Sunday night without telling anybody.

And the familiar 1,200 and 1,500 figures are not a discovery about human physiology. They began as prescribing ranges in a 1998 US clinical guideline. The bottom of what a supervised low-calorie diet would typically be written for, explicitly adjusted for body size. No trial established them as thresholds below which harm begins. They are conventions, and they have been repeated for so long that they read like findings.

Where this leaves the number you set

Pulling the threads together: the metabolic penalty for dieting is real but small. The speed penalty appears not to exist. What does consistently decide outcomes is whether the deficit is one you can still be doing in six months, and the two things that end it early are hunger you cannot sit with, and drift you cannot see.

That is a more useful conclusion than “don’t crash diet”, because it tells you what to watch. Not the scale week to week. Whether you are hungry all the time, and whether what you logged resembles what you ate.

What Fuel8 does about it

Fuel8 is a food diary for iPhone, and it takes a position on this rather than leaving it to you. It will not generate a target below 1,200 kcal for women or 1,500 for men, whatever the arithmetic produces, and it caps the deficit at 20% below your estimated expenditure regardless of how aggressive a goal you pick. Those three numbers are pinned by tests written against the literal values, so nobody can quietly lower them later.

Every target is still editable, and the formula that produced it is shown in Settings. The floor is a floor on what the app will recommend, not a lock on what you can do.

On the drift problem, the honest answer is that no app can measure what you ate. It can only make recording it fast enough that you keep doing it. Fuel8 searches a database that lives on the phone, scans barcodes, and takes a whole meal spoken in one sentence. It also reads your steps from Apple Health and shows them without adding them to what you can eat, for reasons covered here. And there are no streaks to break, no red failure states and no scores. A day logged badly is worth more than a day not logged at all.

Common questions

Sources

1. Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expenditure resulting from altered body weight. New England Journal of Medicine, 1995;332:621–8. n=41, inpatient metabolic ward.
2. Müller MJ et al. Metabolic adaptation to caloric restriction and subsequent refeeding. American Journal of Clinical Nutrition, 2015;102:807–19. n=32 men.
3. Fothergill E et al. Persistent metabolic adaptation 6 years after “The Biggest Loser” competition. Obesity, 2016;24(8):1612–19. n=14, no control group.
4. Kerns JC et al. Increased physical activity associated with less weight regain six years after “The Biggest Loser”. Obesity, 2017;25:1838–43.
5. Purcell K et al. The effect of rate of weight loss on long-term weight management: a randomised controlled trial. Lancet Diabetes & Endocrinology, 2014;2:954–62. n=200, 3-year follow-up.
6. Nackers LM, Ross KM, Perri MG. The association between rate of initial weight loss and long-term success. International Journal of Behavioral Medicine, 2010;17:161–7. Secondary analysis, n=262.
7. Sumithran P et al. Long-term persistence of hormonal adaptations to weight loss. New England Journal of Medicine, 2011;365:1597–604. n=50, 62-week follow-up.
8. Freedman LS et al. Pooled results from 5 validation studies of dietary self-report instruments using recovery biomarkers. American Journal of Epidemiology, 2014;180:172–88.
9. Lichtman SW et al. Discrepancy between self-reported and actual caloric intake in obese subjects. New England Journal of Medicine, 1992;327:1893–8. Intensive protocol n=10.
10. Longland TM et al. Higher compared with lower dietary protein during an energy deficit. American Journal of Clinical Nutrition, 2016;103:738–46. n=40, randomised.
11. Williams NI et al. Magnitude of daily energy deficit predicts frequency of menstrual disturbances. American Journal of Physiology. Endocrinology and Metabolism, 2015;308:E29–39. n=34.
12. Dansinger ML et al. Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets. JAMA, 2005;293:43–53. n=160, randomised.
13. Franz MJ et al. Weight-loss outcomes: a systematic review of trials with a minimum 1-year follow-up. Journal of the American Dietetic Association, 2007;107:1755–67. 80 trials.
14. NICE guideline NG246, Overweight and obesity management. Published January 2025.

This page summarises published research and is not medical or dietary advice. Fuel8 is a food-logging tool, not a medical device. Individual needs vary and the studies described here report group averages. If you have a medical condition, are pregnant, or have any history of disordered eating, speak to a qualified professional before changing how you eat. The full health disclaimer is below.