Expenditure below prediction — and "below what" decides everything#
Adaptive thermogenesis is the gap between the calories you burn and the calories a body of your size and composition ought to burn. Lose weight and your expenditure falls for the boring reason first: there is less of you. Adaptive thermogenesis is whatever is left over after you subtract that — expenditure running below the prediction, not merely below where it started. In a one-year lifestyle intervention where 94 adults averaged a 270 kcal/day deficit and lost 4.8 percent of their weight, researchers computed that leftover thirteen different ways and got answers spanning roughly −65 to −230 kcal/day3. Every one of those numbers is real. They differ because "prediction" is a modelling choice.
That is the shape of the whole topic, and it is why the dimmer-switch framing beats the popular one. This is not a furnace that shuts off; it is a light being turned down by a knob whose markings nobody agrees on. What follows is the part that is settled — that it happens, where in the body it happens, and that it can be switched back on — separated from the part that genuinely isn't, which is how big it is. The related question of whether it can stop a diet is answered elsewhere, and the answer is no: see is 'starvation mode' real for the magnitude debate and the persistence argument. This article is about the mechanism.
The direction nobody mentions#
Almost every account frames adaptive thermogenesis as a starvation defense — a body fighting to keep the fat it is losing. The study that defined the field says something that framing cannot accommodate: it runs the other way too.
Rudolph Leibel, Michael Rosenbaum and Jules Hirsch put 18 subjects with obesity and 23 who had never been obese through experimental weight perturbation in both directions, measuring 24-hour total, resting and non-resting expenditure and the thermic effect of feeding at each plateau. Holding a weight 10 percent or more below the starting point cut total expenditure by 6 ± 3 kcal per kilogram of fat-free mass per day in the never-obese and 8 ± 5 in the group with obesity. Holding a weight 10 percent above the usual one raised total expenditure by 9 ± 7 and 8 ± 4 kcal/kg FFM/day respectively — all four results at P < 0.001, and none of them related to how fat or what sex the subject was1.
Overfeed a person and their body starts wasting energy just as reliably as underfeeding makes it conserve energy. Whatever this system is defending, it is not fat — it is the previous weight.
That symmetry rules out a lot of popular explanation. A mechanism evolved purely to survive famine has no reason to burn off a surplus, and "metabolic damage" cannot explain a metabolism that speeds up on command. The data describe a controller with a setpoint, pushing back against displacement whichever way it comes.
Where the dimmer actually sits: your muscles#
Now read Leibel's components against his totals, a contrast the paper reports but does not spell out. On the loss side, resting and non-resting expenditure each fell 3 to 4 kcal/kg FFM/day — roughly half the 6-to-8 total apiece. On the gain side, non-resting expenditure rose 8 to 9 kcal/kg FFM/day against a total rise of 8 to 9: essentially the entire increase came from the moving half of the day, with the thermic effect of feeding contributing 1 to 2. Both readings point away from the organ you keep being told about. Adaptive thermogenesis is not mostly a slowed resting rate. It is mostly a change in what movement costs.
A decade later the same group found the mechanism. Ten inpatient subjects — five men, five women, three never obese — were studied at usual weight, at 10 percent reduced weight, and at 10 percent reduced weight with leptin. Maintaining the reduced weight came with a fall in total expenditure of roughly 300 to 500 kcal/day beyond what mass and composition predicted, a 20 to 30 percent decline in energy spent on physical activity, and this: cycling to generate 10 watts of power, gross mechanical efficiency of skeletal muscle was about 23 percent higher than at the starting weight2.
Twenty-three percent better mileage on the same fuel. That is what the dimmer physically is: not a banked furnace, but muscle re-tuned to do the same external work for less — which is also why it is so hard to feel. Nothing about you is obviously slower. You are simply cheaper to run.
The dimmer has a knob, and it is leptin#
The same experiment then did the thing that turns a description into a mechanism. Subjects held at 10 percent reduced weight were given low-dose leptin, enough to restore circulating leptin to pre-weight-loss levels. Total energy expenditure, non-resting energy expenditure, skeletal muscle work efficiency, sympathetic nervous system tone and circulating thyroxine and triiodothyronine all returned to pre-weight-loss values2. The 23 percent efficiency gain was fully reversed.
A hormone put the expenditure back. Whatever adaptive thermogenesis is, it is a signal being obeyed — not a component that wore out.
The authors' reading is that the weight-reduced state is one of relative leptin insufficiency: fat cells make leptin, you have fewer of them, the brain reads the drop as a deficit and turns down the parts of expenditure it can reach. Two caveats — this is ten people in a metabolic ward, and leptin injections are not a treatment to extrapolate to, since leptin does little in people whose levels are already high (leptin resistance explained). As a demonstration of what kind of thing this is, though, ten people in a clean crossover beat a thousand with a questionnaire.
Why nobody agrees on the size#
Here is where the literature genuinely splits, and the split is not about physiology. It is about the denominator.
Because adaptive thermogenesis is defined as a residual — measured expenditure minus predicted expenditure — its size depends entirely on how you build the prediction. Change the equation and you change the answer, sometimes changing its sign. A 2025 secondary analysis of 131 adults through a 16-week weight-loss program made this unusually concrete: participants lost 5.6 kg (95% CI −6.3 to −5.0), and resting rate adjusted by the Katch-McArdle equation rose by 0.19 kcal/kg (95% CI 0.14 to 0.23; p < 0.0001), implying no adaptation at all, while the same participants' bioimpedance-adjusted rate trended downward, reaching −0.21 kcal/kg at week 125. One dataset, two defensible methods, opposite conclusions. (The source trial was funded by the Herbalife Nutrition Product Innovation Center, which is worth stating; the finding is a methodological one and cuts against tidy claims in either direction.)
That is the thirteen-method spread from the top of this article, and it propagates into the reviews. Pooling 33 studies and 2,528 participants, adaptive thermogenesis was detected in 27 — but the verdict is careful: values "may be small or non-statistically significant when higher-quality methodological designs are used," and adaptation "seems to be attenuated, or non-existent, after periods of weight stabilisation"4.
| Question | Status |
|---|---|
| Does expenditure fall below prediction during weight loss? | Yes — detected in 27 of 33 studies |
| Does it also rise above prediction during weight gain? | Yes — +8 to +9 kcal/kg FFM/day at +10% weight |
| Where does it mostly sit? | Non-resting expenditure; ~23% higher muscle work efficiency |
| Can it be reversed? | Yes — restoring leptin returned every measure to baseline |
| How big is it? | Contested: ~65 to ~230 kcal/day depending on the method used |
So Rosenbaum's 300-to-500-calorie figure and Nunes's 65-to-230 are not a contradiction to be split down the middle. They measure different things: total daily expenditure in people held at exactly 10 percent below their maximum on a metabolic ward, versus resting expenditure after a moderate 4.8 percent loss in free-living adults. Bigger loss, tighter control, broader measure, bigger number. Say which one you mean and the disagreement mostly evaporates.
Planning around a headwind#
What the dimmer model changes about practice is modest, which is itself the point.
- Budget for it, don't fear it. Roughly 65 to 230 calories a day — plausibly more after a large loss — is enough to slow a diet and make a lower weight harder to hold, nowhere near enough to stop one. If your loss has truly stopped, the arithmetic sits elsewhere (weight-loss plateaus).
- Protect the moving half. The adaptation concentrates in non-resting expenditure and muscle efficiency, so the term it attacks is the one you can consciously defend: hold your daily movement steady as you get lighter.
- Expect maintenance to be the harder job. A controller defending a previous weight makes regain the default, which is why the skills in keeping weight off are different from the skills that lost it.
- Recalculate on your own numbers. The prediction equations are exactly what the size dispute is about, so a formula is a poor instrument for detecting the thing it defines — the target you steer by should come from your own weight trend, recalibrated as you shrink (why your calorie needs drop as you lose weight).
Adaptive thermogenesis has spent a decade being cited as evidence that dieting breaks you. The experiments say something less dramatic and more interesting: a system that resists displacement in either direction, sitting mainly in how efficiently your muscles move you, and switchable — at least under laboratory conditions — by a single hormone. That is a dimmer. Dimmers turn back up.
FAQ#
What exactly is adaptive thermogenesis?#
It is the portion of a change in energy expenditure that mass and body composition don't explain. Lose weight and your burn drops mostly because there is less tissue to run; adaptive thermogenesis is the extra drop below what your remaining tissue predicts. Because it is defined as a leftover, its measured size depends on which prediction equation you use — thirteen approaches applied to one study gave answers from about −65 to −230 kcal/day.
Does adaptive thermogenesis also happen when you gain weight?#
Yes, and that is one of the strongest findings in the field. In the study that established the effect, holding a weight 10 percent above usual raised total energy expenditure by 8 to 9 kcal per kilogram of fat-free mass per day — a mirror image of the fall seen at 10 percent below. The system defends a previous body weight in both directions, which is why "starvation defense" is an incomplete description of it.
Why do studies report such different sizes for metabolic adaptation?#
Because they are subtracting different predictions from different measurements. Adjusting resting rate with the Katch-McArdle equation and with bioimpedance produced opposite conclusions in the same 131 people. On top of that, studies vary in how much weight was lost, whether participants were weight-stable when measured, and whether resting or total expenditure was assessed — reviewers pooling 33 studies found the effect shrinks or disappears under higher-quality designs and after weight stabilization.
Sources#
- Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expenditure resulting from altered body weight. N Engl J Med. 1995;332(10):621-628.
- Rosenbaum M, Goldsmith R, Bloomfield D, et al. Low-dose leptin reverses skeletal muscle, autonomic, and neuroendocrine adaptations to maintenance of reduced weight. J Clin Invest. 2005;115(12):3579-3586.
- Nunes CL, Jesus F, Francisco R, et al. Adaptive thermogenesis after moderate weight loss: magnitude and methodological issues. Eur J Nutr. 2022;61(3):1405-1416.
- Nunes CL, Casanova N, Francisco R, et al. Does adaptive thermogenesis occur after weight loss in adults? A systematic review. Br J Nutr. 2022;127(3):451-469.
- Tang M, Wang J, Xiang Y, Xu R. Metabolic adaptation fluctuates with different prediction equations: a secondary analysis based on a weight-loss clinical trial. Front Nutr. 2025;12:1543263.
