FitProven
Evidence Breakdown
Based on 7 studies
THE BIGGER QUESTION
Whether this one statement holds is settled above. What to actually do about it is a wider question, weighed across every claim that bears on it.
- I'm in a calorie deficit but not losing weight — why? — this claim is one of 2 weighed there
Evidence map
For & against, at a glance
Non-Randomized Trial
Pro
Leibel RL et al. · 1995New England Journal of MedicineInpatient metabolic-ward study of 18 obese and 23 never-obese subjects whose body weight was experimentally driven up or down by 10% and then held stable while total energy expenditure was measured. Maintaining a 10%-reduced weight lowered total energy expenditure by 6 ± 3 kcal/kg of fat-free mass/day in never-obese and 8 ± 5 kcal/kg FFM/day in obese subjects — for a typical adult, roughly 300–400 kcal/day *beyond* what the smaller body size predicts. This is the foundational evidence that adaptive thermogenesis is a real, measurable phenomenon, though the design tests weight-reduced maintenance rather than a plateau occurring during an ongoing diet.
0.39
Inpatient metabolic-ward study of 18 obese and 23 never-obese subjects whose body weight was experimentally driven up or down by 10% and then held stable while total energy expenditure was measured. Maintaining a 10%-reduced weight lowered total energy expenditure by 6 ± 3 kcal/kg of fat-free mass/day in never-obese and 8 ± 5 kcal/kg FFM/day in obese subjects — for a typical adult, roughly 300–400 kcal/day *beyond* what the smaller body size predicts. This is the foundational evidence that adaptive thermogenesis is a real, measurable phenomenon, though the design tests weight-reduced maintenance rather than a plateau occurring during an ongoing diet.
Design Non-Randomized Trial (0.6) × quality 0.65 = impact 0.39
View sourceSystematic Review
Con
Nunes CL et al. · 2022British Journal of NutritionSystematic review of 33 studies (2,528 participants) measuring adaptive thermogenesis after weight loss. Adaptation was detected in 27 of 33 studies, confirming it is a real effect — but across the diet/exercise interventions "the values for AT ranged between 126–418 kJ/d" (roughly 30–100 kcal/day), and the authors note that the better-designed studies reported the *smaller* or non-significant values.
The honest exception: two studies in which participants lost enormous amounts of weight (−58.3 ± 24.9 kg and −57.6 ± 23.8 kg) showed a much larger adaptation of ~837–1255 kJ/day (~200–300 kcal/day), rising to ~2092 kJ/day during follow-up. So the magnitude scales with how extreme the weight loss is, and the 30–100 kcal/day figure is the range for ordinary dieting, not a universal ceiling.
Critically, adaptation "seems to be attenuated, or non-existent, after periods of weight stabilisation" — i.e. it fades in exactly the weight-stable state a plateau describes. For a typical dieter, a 30–100 kcal/day drag cannot erase a 500 kcal/day deficit, so adaptation of this magnitude cannot by itself stall weight loss.
0.68
Systematic review of 33 studies (2,528 participants) measuring adaptive thermogenesis after weight loss. Adaptation was detected in 27 of 33 studies, confirming it is a real effect — but across the diet/exercise interventions "the values for AT ranged between 126–418 kJ/d" (roughly 30–100 kcal/day), and the authors note that the better-designed studies reported the *smaller* or non-significant values. The honest exception: two studies in which participants lost enormous amounts of weight (−58.3 ± 24.9 kg and −57.6 ± 23.8 kg) showed a much larger adaptation of ~837–1255 kJ/day (~200–300 kcal/day), rising to ~2092 kJ/day during follow-up. So the magnitude scales with how extreme the weight loss is, and the 30–100 kcal/day figure is the range for ordinary dieting, not a universal ceiling. Critically, adaptation "seems to be attenuated, or non-existent, after periods of weight stabilisation" — i.e. it fades in exactly the weight-stable state a plateau describes. For a typical dieter, a 30–100 kcal/day drag cannot erase a 500 kcal/day deficit, so adaptation of this magnitude cannot by itself stall weight loss.
Design Systematic Review (0.9) × quality 0.75 = impact 0.68
View sourceCross-Sectional
Con
Subar AF et al. · 2003American Journal of EpidemiologyThe OPEN study measured self-reported energy intake against doubly labelled water — the gold-standard objective measure — in 484 adults aged 40–69. Under-reporting was the norm, not the exception: on a food-frequency questionnaire men under-reported energy intake by 31–36% and women by 34–38%; even on 24-hour recalls, men under-reported by 12–14% and women by 16–20%. This shows the Lichtman finding generalises to a large, unselected sample — the average person tracking their diet is off by hundreds of calories per day, which is the same order of magnitude as the deficit they are trying to hold, and far larger than measured adaptive thermogenesis.
0.36
The OPEN study measured self-reported energy intake against doubly labelled water — the gold-standard objective measure — in 484 adults aged 40–69. Under-reporting was the norm, not the exception: on a food-frequency questionnaire men under-reported energy intake by 31–36% and women by 34–38%; even on 24-hour recalls, men under-reported by 12–14% and women by 16–20%. This shows the Lichtman finding generalises to a large, unselected sample — the average person tracking their diet is off by hundreds of calories per day, which is the same order of magnitude as the deficit they are trying to hold, and far larger than measured adaptive thermogenesis.
Design Cross-Sectional (0.4) × quality 0.90 = impact 0.36
View sourceProspective Cohort
Con
Martins C et al. · 2020American Journal of Clinical Nutrition171 premenopausal women with overweight followed for 2 years after a diet-induced weight loss of ~12 kg, with RMR measured by indirect calorimetry and body composition by a 4-compartment model — crucially, after 4-week weight-stabilisation periods, which removes the transient suppression caused by being in active energy deficit. Metabolic adaptation was only ~50–60 kcal/day (3–4% below predicted) immediately post-loss, and was absent at both the 1- and 2-year follow-ups. It did not correlate with weight regain. The authors conclude adaptation is "a mere reflection of the magnitude of weight loss" rather than a compensatory force driving plateaus or relapse.
0.30
171 premenopausal women with overweight followed for 2 years after a diet-induced weight loss of ~12 kg, with RMR measured by indirect calorimetry and body composition by a 4-compartment model — crucially, after 4-week weight-stabilisation periods, which removes the transient suppression caused by being in active energy deficit. Metabolic adaptation was only ~50–60 kcal/day (3–4% below predicted) immediately post-loss, and was absent at both the 1- and 2-year follow-ups. It did not correlate with weight regain. The authors conclude adaptation is "a mere reflection of the magnitude of weight loss" rather than a compensatory force driving plateaus or relapse.
Design Prospective Cohort (0.6) × quality 0.50 = impact 0.30
View sourceNon-Randomized Trial
Pro
Leibel RL et al. · 1995New England Journal of MedicineInpatient metabolic-ward study of 18 obese and 23 never-obese subjects whose body weight was experimentally driven up or down by 10% and then held stable while total energy expenditure was measured. Maintaining a 10%-reduced weight lowered total energy expenditure by 6 ± 3 kcal/kg of fat-free mass/day in never-obese and 8 ± 5 kcal/kg FFM/day in obese subjects — for a typical adult, roughly 300–400 kcal/day *beyond* what the smaller body size predicts. This is the foundational evidence that adaptive thermogenesis is a real, measurable phenomenon, though the design tests weight-reduced maintenance rather than a plateau occurring during an ongoing diet.
0.39
Inpatient metabolic-ward study of 18 obese and 23 never-obese subjects whose body weight was experimentally driven up or down by 10% and then held stable while total energy expenditure was measured. Maintaining a 10%-reduced weight lowered total energy expenditure by 6 ± 3 kcal/kg of fat-free mass/day in never-obese and 8 ± 5 kcal/kg FFM/day in obese subjects — for a typical adult, roughly 300–400 kcal/day *beyond* what the smaller body size predicts. This is the foundational evidence that adaptive thermogenesis is a real, measurable phenomenon, though the design tests weight-reduced maintenance rather than a plateau occurring during an ongoing diet.
Design Non-Randomized Trial (0.6) × quality 0.65 = impact 0.39
View sourceSystematic Review
Con
Nunes CL et al. · 2022British Journal of NutritionSystematic review of 33 studies (2,528 participants) measuring adaptive thermogenesis after weight loss. Adaptation was detected in 27 of 33 studies, confirming it is a real effect — but across the diet/exercise interventions "the values for AT ranged between 126–418 kJ/d" (roughly 30–100 kcal/day), and the authors note that the better-designed studies reported the *smaller* or non-significant values.
The honest exception: two studies in which participants lost enormous amounts of weight (−58.3 ± 24.9 kg and −57.6 ± 23.8 kg) showed a much larger adaptation of ~837–1255 kJ/day (~200–300 kcal/day), rising to ~2092 kJ/day during follow-up. So the magnitude scales with how extreme the weight loss is, and the 30–100 kcal/day figure is the range for ordinary dieting, not a universal ceiling.
Critically, adaptation "seems to be attenuated, or non-existent, after periods of weight stabilisation" — i.e. it fades in exactly the weight-stable state a plateau describes. For a typical dieter, a 30–100 kcal/day drag cannot erase a 500 kcal/day deficit, so adaptation of this magnitude cannot by itself stall weight loss.
0.68
Systematic review of 33 studies (2,528 participants) measuring adaptive thermogenesis after weight loss. Adaptation was detected in 27 of 33 studies, confirming it is a real effect — but across the diet/exercise interventions "the values for AT ranged between 126–418 kJ/d" (roughly 30–100 kcal/day), and the authors note that the better-designed studies reported the *smaller* or non-significant values. The honest exception: two studies in which participants lost enormous amounts of weight (−58.3 ± 24.9 kg and −57.6 ± 23.8 kg) showed a much larger adaptation of ~837–1255 kJ/day (~200–300 kcal/day), rising to ~2092 kJ/day during follow-up. So the magnitude scales with how extreme the weight loss is, and the 30–100 kcal/day figure is the range for ordinary dieting, not a universal ceiling. Critically, adaptation "seems to be attenuated, or non-existent, after periods of weight stabilisation" — i.e. it fades in exactly the weight-stable state a plateau describes. For a typical dieter, a 30–100 kcal/day drag cannot erase a 500 kcal/day deficit, so adaptation of this magnitude cannot by itself stall weight loss.
Design Systematic Review (0.9) × quality 0.75 = impact 0.68
View sourceCross-Sectional
Con
Subar AF et al. · 2003American Journal of EpidemiologyThe OPEN study measured self-reported energy intake against doubly labelled water — the gold-standard objective measure — in 484 adults aged 40–69. Under-reporting was the norm, not the exception: on a food-frequency questionnaire men under-reported energy intake by 31–36% and women by 34–38%; even on 24-hour recalls, men under-reported by 12–14% and women by 16–20%. This shows the Lichtman finding generalises to a large, unselected sample — the average person tracking their diet is off by hundreds of calories per day, which is the same order of magnitude as the deficit they are trying to hold, and far larger than measured adaptive thermogenesis.
0.36
The OPEN study measured self-reported energy intake against doubly labelled water — the gold-standard objective measure — in 484 adults aged 40–69. Under-reporting was the norm, not the exception: on a food-frequency questionnaire men under-reported energy intake by 31–36% and women by 34–38%; even on 24-hour recalls, men under-reported by 12–14% and women by 16–20%. This shows the Lichtman finding generalises to a large, unselected sample — the average person tracking their diet is off by hundreds of calories per day, which is the same order of magnitude as the deficit they are trying to hold, and far larger than measured adaptive thermogenesis.
Design Cross-Sectional (0.4) × quality 0.90 = impact 0.36
View sourceProspective Cohort
Con
Martins C et al. · 2020American Journal of Clinical Nutrition171 premenopausal women with overweight followed for 2 years after a diet-induced weight loss of ~12 kg, with RMR measured by indirect calorimetry and body composition by a 4-compartment model — crucially, after 4-week weight-stabilisation periods, which removes the transient suppression caused by being in active energy deficit. Metabolic adaptation was only ~50–60 kcal/day (3–4% below predicted) immediately post-loss, and was absent at both the 1- and 2-year follow-ups. It did not correlate with weight regain. The authors conclude adaptation is "a mere reflection of the magnitude of weight loss" rather than a compensatory force driving plateaus or relapse.
0.30
171 premenopausal women with overweight followed for 2 years after a diet-induced weight loss of ~12 kg, with RMR measured by indirect calorimetry and body composition by a 4-compartment model — crucially, after 4-week weight-stabilisation periods, which removes the transient suppression caused by being in active energy deficit. Metabolic adaptation was only ~50–60 kcal/day (3–4% below predicted) immediately post-loss, and was absent at both the 1- and 2-year follow-ups. It did not correlate with weight regain. The authors conclude adaptation is "a mere reflection of the magnitude of weight loss" rather than a compensatory force driving plateaus or relapse.
Design Prospective Cohort (0.6) × quality 0.50 = impact 0.30
View sourceShowing the 4 strongest of 7 studies. Tap any node to expand its detail.
Evidence
PRO (2)
PRO Non-Randomized Trialn=41INDUSTRY0.65 Leibel RL, Rosenbaum M et al. (1995)
Maintaining a 10%-reduced body weight lowered total energy expenditure by 6 +/- 3 kcal/kg fat-free mass/day in never-obese and 8 +/- 5 in obese subjects -- roughly 300-400 kcal/day beyond what the smaller body size predicts; a 10% weight GAIN raised it by 9 +/- 7 and 8 +/- 4 respectively
Inpatient metabolic-ward study of 18 obese and 23 never-obese subjects whose body weight was experimentally driven up or down by 10% and then held stable while total energy expenditure was measured. Maintaining a 10%-reduced weight lowered total energy expenditure by 6 ± 3 kcal/kg of fat-free mass/day in never-obese and 8 ± 5 kcal/kg FFM/day in obese subjects — for a typical adult, roughly 300–400 kcal/day beyond what the smaller body size predicts. This is the foundational evidence that adaptive thermogenesis is a real, measurable phenomenon, though the design tests weight-reduced maintenance rather than a plateau occurring during an ongoing diet.
Weighted 0.65 — 41 subjects (18 obese, 23 never-obese) studied on an inpatient metabolic ward with doubly labelled water and indirect calorimetry -- an exceptionally rigorous, tightly controlled design, which is why it remains the foundational demonstration of adaptive thermogenesis. Not randomised, modest n, and indirect for the usual claim: it measures maintenance at an experimentally ALTERED weight, not a plateau arising during an ongoing diet. Partly funded by the Weight Watchers Foundation, the foundation of a commercial weight-loss company, alongside NIH.
Funding: US National Institutes of Health (DK30583, GCRCRR00102) and the Weight Watchers Foundation
New England Journal of Medicine
PRO Prospective Cohortn=140.40 Fothergill E, Guo J et al. (2016)
Metabolic adaptation -275 kcal/day at week 30, widening to -499 kcal/day at 6 years, despite regaining 41.0 of the 58.3 kg lost
Six-year follow-up of 14 of the 16 original "The Biggest Loser" contestants, with RMR by indirect calorimetry and body composition by DXA. Metabolic adaptation — the residual RMR after adjusting for lost body mass and age — was −275 ± 207 kcal/day at the end of the 30-week competition and had widened to −499 ± 207 kcal/day six years later, despite participants regaining 41.0 ± 31.3 kg of the 58.3 ± 24.9 kg they lost. This is the single strongest demonstration that adaptation can be large and persistent. Two caveats limit how far it generalises: the sample is 14 people who underwent an extreme, televised diet-and-exercise intervention, and within this cohort the participants who kept the most weight off had the greater metabolic adaptation — so adaptation did not determine who regained.
Weighted 0.40 — Only 14 people, no control group, and an extreme televised diet-and-exercise intervention that generalises poorly. Measurement quality is high (DXA, indirect calorimetry) and the 6-year follow-up is unmatched, but within the cohort the greatest adaptation occurred in those who kept the most weight off - so it cannot support a causal story about regain.
Funding: US National Institutes of Health (NIDDK Intramural Research Program)
Obesity
AGAINST (5)
AGAINST Cross-Sectionaln=4840.90 Subar AF, Kipnis V et al. (2003)
Against doubly labelled water, men under-reported energy by 31-36% and women by 34-38% on a food-frequency questionnaire; even on 24-hour recalls men under-reported by 12-14% and women by 16-20%
The OPEN study measured self-reported energy intake against doubly labelled water — the gold-standard objective measure — in 484 adults aged 40–69. Under-reporting was the norm, not the exception: on a food-frequency questionnaire men under-reported energy intake by 31–36% and women by 34–38%; even on 24-hour recalls, men under-reported by 12–14% and women by 16–20%. This shows the Lichtman finding generalises to a large, unselected sample — the average person tracking their diet is off by hundreds of calories per day, which is the same order of magnitude as the deficit they are trying to hold, and far larger than measured adaptive thermogenesis.
Weighted 0.90 — n=484 adults aged 40-69, validated against doubly labelled water -- the objective gold standard, not another self-report instrument. That biomarker anchor plus a large, unselected sample makes this about as strong as a cross-sectional design gets, and the effect is enormous and consistent across both instruments and both sexes. Funding statement not retrievable (paywalled); NCI/NIH support is likely given the OPEN study's provenance but is not asserted here.
American Journal of Epidemiology
AGAINST Systematic Reviewn=25280.75 Nunes CL, Casanova N et al. (2022)
Adaptive thermogenesis detected in 27 of 33 studies, but in lifestyle interventions it ran only ~126-418 kJ/day (~30-100 kcal/day); the exceptions were two studies with very large weight loss (~58 kg), where it reached ~837-1255 kJ/day. Better-designed studies reported the smaller or non-significant values, and adaptation was attenuated or non-existent after weight stabilisation
Systematic review of 33 studies (2,528 participants) measuring adaptive thermogenesis after weight loss. Adaptation was detected in 27 of 33 studies, confirming it is a real effect — but across the diet/exercise interventions "the values for AT ranged between 126–418 kJ/d" (roughly 30–100 kcal/day), and the authors note that the better-designed studies reported the smaller or non-significant values.
The honest exception: two studies in which participants lost enormous amounts of weight (−58.3 ± 24.9 kg and −57.6 ± 23.8 kg) showed a much larger adaptation of ~837–1255 kJ/day (~200–300 kcal/day), rising to ~2092 kJ/day during follow-up. So the magnitude scales with how extreme the weight loss is, and the 30–100 kcal/day figure is the range for ordinary dieting, not a universal ceiling.
Critically, adaptation "seems to be attenuated, or non-existent, after periods of weight stabilisation" — i.e. it fades in exactly the weight-stable state a plateau describes. For a typical dieter, a 30–100 kcal/day drag cannot erase a 500 kcal/day deficit, so adaptation of this magnitude cannot by itself stall weight loss.
Weighted 0.75 — 33 studies and 2,528 participants — broad, careful coverage of the literature. It is a narrative systematic review rather than a pooled meta-analysis, methods for quantifying adaptation differ widely between studies, and the authors themselves note the better-designed studies reported the smaller or non-significant values. Funding could not be verified (paywalled).
British Journal of Nutrition
AGAINST Mathematical Model0.65 Thomas DM, Martin CK et al. (2014)
Metabolic adaptation could not reproduce the observed plateau: imposing a 10% suppression of energy expenditure raised the predicted stabilization weight by 11% but did not move the plateau, which the model still placed at 365-730 days. An intermittent-non-adherence model did reproduce the real ~6-month plateau and its weight oscillations (calculated CALERIE adherence in women: 80% in month 1, 60% in month 2, 50% in month 3, 40% thereafter)
A mathematical modelling study (classified here as a secondary/non-primary-data review, as the evidence schema has no modelling category) that directly pits the two explanations against each other, validated against four datasets including CALERIE, the Minnesota Starvation Experiment and the Bouchard twin overfeeding study. Metabolic adaptation could not produce the plateau people actually experience: even imposing a 10% suppression of energy expenditure raised the predicted stabilization weight by 11% but did not shift the timing of the plateau, which the model still placed at 365–730 days. An intermittent-non-adherence model did reproduce the observed ~6-month plateau and its characteristic weight oscillations — in CALERIE, calculated adherence in women was 80% in month 1, 60% in month 2, 50% in month 3 and 40% thereafter, and the free-living CALERIE weight curves oscillate while the in-residence Minnesota Starvation Experiment curves decrease monotonically. Notably, a plateau appeared even at what would be considered high adherence, meaning a plateau is not evidence of a lack of willpower either.
Weighted 0.65 — No sample size of its own: this is a MATHEMATICAL MODELLING study, filed as narrative_review only because the evidence schema has no modelling category -- the type label understates what it is. It carries no participants, but is validated against four independent datasets (CALERIE n=23, the Minnesota Starvation Experiment n=13, the Bouchard twin overfeeding study, plus NHANES). Independently NIH-funded (one author declares advisory roles with Weight Watchers, Pathway Genomics and Nike SPARQ). Rated well above a typical narrative review for rigour and external validation, but capped because the conclusion is model-derived, not experimentally tested.
Funding: NIH (NIDDK R15 DK090739-01, U01 DK094418, DK089051, P30 DK072476; NIA U01 AG20478; NHLBI HL102166; NICHD R00HD060762)
American Journal of Clinical Nutrition
AGAINST Cross-Sectionaln=100.55 Lichtman SW, Pisarska K et al. (1992)
Measured energy expenditure and RMR fell within 5% of predicted values, ruling out abnormal thermogenesis; the subjects under-reported food intake by 47 +/- 16% and over-reported physical activity by 51 +/- 75%
The canonical test of the "my metabolism is broken" hypothesis. Ten subjects with a history of "diet resistance" — who reported failing to lose weight on under 1,200 kcal/day — had actual intake and total energy expenditure measured over 14 days by doubly labelled water and indirect calorimetry. Their measured energy expenditure and RMR were within 5% of predicted values, ruling out abnormal thermogenesis. Instead they under-reported actual food intake by 47 ± 16% and over-reported physical activity by 51 ± 75%. The plateau was an intake problem, not a metabolism problem. Sample size is small (n=10), so this establishes that under-reporting can fully account for apparent diet resistance, not the population prevalence.
Weighted 0.55 — Only 10 self-described diet-resistant subjects, but measured with the gold standard -- doubly labelled water plus indirect calorimetry over 14 days -- which is why it remains decisive despite its size, and it is publicly funded. It establishes that misreporting CAN fully account for apparent diet resistance in such people; with n=10 and a self-selected sample drawn from 224 screened, it cannot establish how COMMON that is, and it is not a randomised design.
Funding: US National Institutes of Health (P01-DK42618, DK-26687, RR 00047)
New England Journal of Medicine
AGAINST Prospective Cohortn=790.50 Martins C, Gower BA et al. (2020)
metabolic adaptation was only ~50-60 kcal/day (3-4% below predicted) immediately after weight loss and absent at both 1 and 2 years; it did not correlate with weight regain
171 premenopausal women with overweight followed for 2 years after a diet-induced weight loss of ~12 kg, with RMR measured by indirect calorimetry and body composition by a 4-compartment model — crucially, after 4-week weight-stabilisation periods, which removes the transient suppression caused by being in active energy deficit. Metabolic adaptation was only ~50–60 kcal/day (3–4% below predicted) immediately post-loss, and was absent at both the 1- and 2-year follow-ups. It did not correlate with weight regain. The authors conclude adaptation is "a mere reflection of the magnitude of weight loss" rather than a compensatory force driving plateaus or relapse.
Weighted 0.50 — 171 women entered, but the analysed sample shrinks sharply with follow-up — 156 immediately post-weight-loss, 131 at 1 year, 79 at 2 years, and only 46 with complete data at every timepoint — so the headline 2-year null rests on 79 people, and dropouts skewed toward higher baseline BMI and less weight loss. Measuring RMR after 4-week weight-stabilisation periods is a real methodological strength that removes the transient deficit effect. Publicly funded (NIH); premenopausal women aged 20-41 only.
Funding: US National Institutes of Health
American Journal of Clinical Nutrition