Metabolic adaptation and weight loss plateaus

What adaptive thermogenesis actually is

The technical definition is narrow and worth stating precisely, because almost every popular account uses the term more loosely. Adaptive thermogenesis is the reduction in energy expenditure that remains after you have accounted for the change in body size and composition. If a person loses fifteen kilograms and their measured expenditure falls by more than the equations predict for someone of their new weight and lean mass, that residual is the adaptive component.

The review by Rosenbaum and Leibel sets out the physiology. Reduced energy expenditure after weight loss reflects both the smaller mass being maintained and an increase in the efficiency of skeletal muscle work, alongside changes in circulating leptin, thyroid hormones and sympathetic nervous system activity. The body behaves as though it is defending a weight, and it does so partly by spending less.

The classic measurement comes from Leibel, Rosenbaum and Hirsch, who fed subjects to maintain a body weight ten per cent or more below their usual weight and measured total energy expenditure directly. Maintaining that reduced weight was associated with a reduction in total energy expenditure of about 6 kilocalories per kilogram of fat-free mass per day in subjects who had never been obese, and about 8 in the subjects with obesity. For someone carrying 55 kilograms of fat-free mass, that is roughly 330 to 440 kilocalories a day. The study is the foundation of everything else in this field, and it is the reason the phenomenon is accepted rather than argued about.

What is argued about is the size, and the answer depends heavily on who is being measured and how much weight they lost. That is the subject of the next section, and the distinction between the extreme cases and the ordinary ones is where most of the public confusion lives.

How big is it really

The largest published figures come from the most extreme circumstances, and they are the ones that circulate. Fothergill and colleagues followed participants from a televised weight-loss competition and measured resting metabolic rate six years after it ended. Metabolic adaptation, meaning the shortfall against what body composition predicted, averaged about 499 kilocalories a day, and it persisted even though most participants had regained a large share of the weight. That is a real finding, and it describes people who lost an extraordinary amount of weight extraordinarily fast under conditions no ordinary dieter experiences.

For ordinary weight loss the measured residual is considerably smaller. The review by Muller and Bosy-Westphal, assessing adaptive thermogenesis across the available human studies, describes an effect that is detectable but modest, and notes how sensitive the measured size is to methodological choices: which prediction equation is used for the expected value, whether the measurement is taken during active weight loss or at a stabilised lower weight, and how body composition was assessed. Different reasonable choices produce meaningfully different estimates of the same phenomenon.

That methodological sensitivity is the honest headline. Adaptive thermogenesis is not a single number that applies to everyone, and any source quoting one is quoting an artefact of a particular study design. What can be said with confidence is the ordering: for a typical dieter losing a typical amount of weight at a sensible rate, the mechanical effect of a smaller body is the largest term, reduced daily movement is usually next, and the residual adaptive component is smaller than either.

The practical translation is unglamorous. Adaptation is a headwind that makes a deficit progressively less effective, and the response is to recalculate maintenance periodically rather than to assume the original figure still holds. It is not a reason to expect a deficit to stop working, and it is not a reason to believe a metabolism has been damaged.

Lost tissue versus genuine adaptation

Before invoking adaptation, the mechanical effect has to be subtracted, and the mechanical effect is easy to compute. In the Mifflin-St Jeor equation, body weight enters with a coefficient of ten, so each kilogram lost reduces predicted basal rate by ten kilocalories a day. Multiplied by a sedentary activity factor of 1.2, that is about twelve kilocalories a day of maintenance per kilogram of body weight lost.

Take a 100 kilogram man of 180 centimetres and 40 years, sedentary. His predicted basal rate is 1,930 kilocalories and his estimated maintenance is about 2,316. At 90 kilograms, the same arithmetic gives 1,830 and about 2,196, a fall of 120 kilocalories a day. At 80 kilograms it gives 1,730 and about 2,076, a fall of 240 from where he started. None of that is adaptation. It is simply that there is less of him to run, and it would have happened in a perfectly non-adaptive body.

This is why a deficit that worked in month one produces less in month four. A 500 kilocalorie deficit calculated against a 2,316 maintenance is a 500 kilocalorie deficit against 2,076 only if you recalculate, and most people do not. Left uncorrected, a fixed intake that started as a 500 kilocalorie deficit has quietly become a 260 kilocalorie one by the time twenty kilograms have gone, which halves the rate of loss without any physiology being involved at all.

The table below makes the size of the mechanical effect concrete, and it also makes the week-three point obvious. At three weeks into a sensible deficit, a 100 kilogram person might be 2.5 kilograms down, which is worth about thirty kilocalories a day. Thirty kilocalories is a third of an apple. It cannot stall anything.

Non-exercise activity falls quietly

Non-exercise activity thermogenesis is everything you burn that is not sleeping, eating or deliberate training: walking, standing, posture, housework, fidgeting, the stairs you take without thinking about them. It is the most variable component of daily expenditure between individuals, and the landmark overfeeding work by Levine, Eberhardt and Jensen showed how large that variability is, finding that the NEAT response to overfeeding explained most of the difference in how much fat people gained on an identical surplus.

The same component moves downward in a deficit, and it does so without any decision being made. Fewer steps accumulate over a day. Standing becomes sitting. The walk to the further shop becomes the walk to the nearer one. None of it is registered as a choice, and none of it appears in a training log, which is exactly why it is so effective at eroding a deficit invisibly.

The size is worth taking seriously. For most people the reduction in spontaneous movement during a sustained deficit is larger than the measured adaptive thermogenesis residual, which makes it the more useful thing to address. It is also far easier to address, because it responds to a deliberate target where hormonal adaptation does not.

The counter is boring and effective: a step target you actually meet, a walk at a fixed time each day, standing where standing is available. A step count is the only practical instrument most people have for watching this component, and it is worth recording alongside weight, because a step average that has fallen from 9,000 to 6,000 across a diet is a plausible two hundred kilocalorie a day explanation for a plateau that was about to be blamed on metabolism.

Water retention masking real fat loss

A stalled scale is not a stalled diet, and the most common reason for the difference is water. Body water moves on a scale of one to three kilograms in ordinary circumstances, which is enough to hide two or three weeks of genuine fat loss completely. The research measuring the composition of two-week weight changes in free-living adults found that the majority of short-term change was fat-free mass rather than fat, which is the quantitative version of the same point.

Several mechanisms contribute, and they often arrive together. Glycogen is stored with water at roughly three grams of water per gram of glycogen on the usual estimate, so a higher carbohydrate day or a restored training week can add a kilogram or more of scale weight that is entirely fuel and fluid. Sodium intake moves fluid balance within a day. The menstrual cycle produces a predictable fluctuation across the month that is large enough to swamp a weekly fat loss signal. And a new or harder training stimulus produces muscle damage and an inflammatory fluid response that adds weight for several days.

There is a further mechanism specific to dieting, which is sometimes described informally as a whoosh. Sustained dietary restriction is a stressor and raises cortisol, and cortisol promotes fluid retention. Fat cells that have lost their triglyceride can retain water in its place for a period, so the loss is real at the cellular level and invisible on the scale until the fluid is released, often abruptly and often after a day of relaxed eating or improved sleep. The observation is widely reported and the underlying cortisol and fluid physiology is sound, but the precise cellular mechanism is better treated as a plausible explanation than as an established finding.

The practical defence against all of this is measurement discipline rather than cleverness. Use a seven-day rolling average and never a single morning. Keep at least one measurement that water cannot move, which in practice means a waist measurement taken under fixed conditions, because a waist that is falling while the scale is flat is a diet that is working. And compare against three or four weeks ago rather than against last Tuesday.

Why a week-three plateau is almost never adaptation

Put the numbers together and the week-three case collapses. At three weeks, a 100 kilogram person on a sensible deficit is perhaps two and a half kilograms lighter, which is worth about thirty kilocalories a day of mechanical change. Adaptive thermogenesis at that point, if it is measurable at all, is a small fraction of a normal day's expenditure. The two together are smaller than the daily measurement noise on a bathroom scale.

Meanwhile, the alternative explanations are worth hundreds of kilocalories. A fortnight of slightly relaxed logging is easily two hundred a day. A step count that has drifted down by three thousand is comparable. A kilogram of water arriving with a return to normal carbohydrate intake after an initially low-carbohydrate fortnight hides a fortnight of progress outright. Each of these is an order of magnitude larger than the metabolic explanation, and each of them is checkable, which the metabolic explanation is not.

So the correct order of investigation at a plateau is by size of effect and by testability. Check the measurement first: are you reading a seven-day average or a single morning, and has the waist measurement moved even though the scale has not? Check the intake second, with a few days of genuinely weighed logging rather than an impression. Check movement third, using a step average. Only when all three come back clean does recalculating maintenance for the weight already lost become the next step, and only after months of sustained restriction does adaptation become a sensible thing to invoke.

There is one more explanation that belongs on the list and rarely appears: the deficit was never as large as it looked, because maintenance was overestimated at the start. A person who chose moderately active when they were lightly active began with a maintenance figure perhaps 250 kilocalories too high, and their twenty per cent deficit was really a ten per cent one. That produces slow, ragged progress from day one, which is easy to misread later as a plateau that arrived.

Diet breaks and refeeds: what the evidence supports

A diet break is a planned return to maintenance intake for a period of one to two weeks. A refeed is a much shorter increase, typically a day or two, usually achieved by raising carbohydrate. The two are frequently discussed together and the evidence behind them is not remotely equivalent.

The strongest controlled evidence for the longer version is the MATADOR trial by Byrne and colleagues, which compared continuous energy restriction with intermittent energy restriction delivered as alternating blocks of restriction and maintenance in men with obesity. The intermittent protocol produced greater weight and fat loss for the same total restriction, and the interpretation offered was that the maintenance blocks limited the compensatory fall in energy expenditure. That is a single trial in a specific population with a specific protocol, which is a real caveat, but it is proper controlled evidence and it is more than the alternatives have.

Short refeeds have a much weaker case. The physiological rationale usually offered involves leptin, which does rise with carbohydrate intake, but a day or two of raised intake producing a durable change in energy expenditure is not something the evidence establishes. What a refeed can reliably do is refill muscle glycogen, which improves training quality for a few days, and provide a psychological break, which improves adherence. Both are legitimate reasons. Neither is a metabolic reason, and presenting them as one is where most of the overselling happens.

There is an arithmetic point that is often missed. A diet break is not free: two weeks at maintenance is two weeks not losing weight, so the same total loss takes longer in calendar time. The MATADOR result suggests the trade is worth making because the restriction blocks themselves become more productive, but anyone planning a break should plan it as a deliberate extension of the timeline rather than a shortcut, and should return to the deficit on a fixed date rather than on a feeling.

The honest timescale for recovery

The question people most want answered is how long it takes for energy expenditure to return to normal after a period of restriction, and the honest answer has two halves that need separating. The portion caused by a smaller body does not recover at all while the body remains smaller, because it is not a deficit to be repaid. It is the correct expenditure for the current size, and the only thing that reverses it is regaining the weight, which is not the goal.

The adaptive portion is the part that can recover, and the evidence suggests it does so slowly and incompletely in some cases. The six-year follow-up of extreme weight loss found the adaptation persisting long after the competition, which is the most pessimistic data point available and also the most extreme population. For ordinary weight loss at ordinary rates the picture appears considerably more forgiving, though the measurement difficulties described earlier mean this is an area where confident numbers should be treated with suspicion.

What can be said usefully is procedural. Returning to maintenance for a sustained period, resistance training to retain or rebuild lean mass, and a deliberate effort to restore the daily movement that fell away are the three interventions that address the recoverable components. The first two have direct effects on expenditure, and the third addresses the component that fell furthest. None of them is fast, and all of them are worth doing regardless, which makes the uncertainty about the exact timescale less important than it sounds.

The more useful reframing is to stop treating expenditure as a thing to be restored and start treating it as a thing to be recalculated. Your maintenance figure after losing fifteen kilograms is a different number from the one you started with, and it will be lower for mechanical reasons whatever happens to the adaptive component. Measuring the new figure from a fortnight of weight data at the new weight is a better use of attention than waiting for the old one to come back.

Frequently asked questions

Put it into practice

Run your own numbers through the TDEE calculator, the calorie deficit calculator, the weight change % calculator and the BMR calculator. Related reading: TDEE explained, How many calories to eat and Track without DEXA.

Sources

1. Rosenbaum M, Leibel RL. Adaptive thermogenesis in humans. Int J Obes 2010;34:S47-55. doi.org/10.1038/ijo.2010.184
2. Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expenditure resulting from altered body weight. N Engl J Med 1995;332:621-8. doi.org/10.1056/NEJM199503093321001
3. Fothergill E, Guo J, Howard L, et al. Persistent metabolic adaptation 6 years after The Biggest Loser competition. Obesity 2016;24:1612-9. doi.org/10.1002/oby.21538
4. Muller MJ, Bosy-Westphal A. Adaptive thermogenesis with weight loss in humans. Obesity 2013;21:218-28. doi.org/10.1002/oby.20027
5. Levine JA, Eberhardt NL, Jensen MD. Role of nonexercise activity thermogenesis in resistance to fat gain in humans. Science 1999;283:212-4. doi.org/10.1126/science.283.5399.212
6. Byrne NM, Sainsbury A, King NA, Hills AP, Wood RE. Intermittent energy restriction improves weight loss efficiency in obese men: the MATADOR study. Int J Obes 2018;42:129-38. doi.org/10.1038/ijo.2017.206
7. Trexler ET, Smith-Ryan AE, Norton LE. Metabolic adaptation to weight loss: implications for the athlete. J Int Soc Sports Nutr 2014;11:7. doi.org/10.1186/1550-2783-11-7
8. Bhutani S, Kahn E, Tasali E, Schoeller DA. Composition of two-week change in body weight under unrestricted free-living conditions. Physiol Rep 2017;5:e13336. doi.org/10.14814/phy2.13336