Weight loss results in obligatory reductions in energy expenditure (EE) due to loss of metabolically active fat-free mass (FFM). This is accompanied by adaptive reductions (i.e. adaptive thermogenesis) designed to restore energy balance while in an energy crisis. While the ‘3500-kcal rule’ is used to advise weight loss in clinical practice, the assumption that EE remains constant during energy restriction results in a large overestimation of weight loss. Thus, this work proposes a novel method of weight-loss prediction to more accurately account for the dynamic trajectory of EE. A mathematical model of weight loss was developed using ordinary differential equations relying on simple self-reported inputs of weight and energy intake to predict weight loss over a specified time. The model subdivides total daily EE into resting EE, physical activity EE, and diet-induced thermogenesis, modelling obligatory and adaptive changes in each compartment independently. The proposed model was tested and refined using commercial weight-loss data from participants enrolled on a very low-energy total-diet replacement programme (LighterLife UK, Essex). Mathematical modelling predicted post-intervention weight loss within 0.75% (1.07 kg) of that observed in females with overweight or obesity. Short-term weight loss was consistently underestimated, likely due to considerable FFM reductions reported on the onset of weight loss. The best model agreement was observed from 6 to 9 weeks where the predicted end-weight was within 0.35 kg of that observed. The proposed mathematical model simulated rapid weight loss with reasonable accuracy. Incorporated terms for energy partitioning and adaptive thermogenesis allow us to easily account for dynamic changes in EE, supporting the potential use of such a model in clinical practice.

The observation that 64% of English adults are overweight or obese despite a rising prevalence in weight-loss attempts suggests our understanding of energy balance is fundamentally flawed. Weight-loss is induced through a negative energy balance; however, we typically view weight change as a static function, in that energy intake and energy expenditure are independent variables, resulting in a fixed rate of weight-loss assuming a constant energy deficit. Such static modelling provides the basis for the clinical assumption that a 14644 kJ (3500 kcal) deficit translates to a 1 lb weight-loss. However, this ‘3500 kcal (14644 kJ) rule’ is consistently shown to significantly overestimate weight-loss. Static modelling disregards obligatory changes in energy expenditure associated with the loss of metabolically active tissue, i.e. skeletal muscle. Additionally, it disregards the presence of adaptive thermogenesis, the underfeeding-associated fall in resting energy expenditure beyond that caused by loss of fat-free mass. This metabolic manipulation of energy expenditure is observed from the onset of energy restriction to maintain weight at a genetically pre-determined set point. As a result, the observed magnitude of weight-loss is disproportionally less, followed by earlier weight plateau, despite strict compliance to a dietary intervention. By simulating dynamic changes in energy expenditure associated with underfeeding, mathematical modelling may provide a more accurate method of weight-loss prediction. However, accuracy at an individual level is limited due to difficulty estimating energy requirements, physical activity and dietary intake in free-living individuals. In the present paper, we aim to outline the contribution of dynamic changes in energy expenditure to weight-loss resistance and weight plateau.

Adaptive thermogenesis (AT) has been proposed to be a compensatory response that may resist weight loss (WL) and promote weight regain. This systematic review examined the existence of AT in adults after a period of negative energy balance (EB) with or without a weight stabilisation phase. Studies published until 15 May 2020 were identified from PubMed, Cochrane Library, EMBASE, MEDLINE, SCOPUS and Web of Science. Inclusion criteria included statistically significant WL, observational with follow-up or experimental studies, age > 18y, sample size ≥10 participants, intervention period ≥ 1week, published in English, objective measures of total daily energy expenditure (EE) (TDEE), resting EE (REE) and sleeping EE(SEE). The systematic review was registered at PROSPERO (2020 CRD42020165348). A total of thirty-three studies comprising 2528 participants were included. AT was observed in twenty-seven studies. Twenty-three studies showed significant values for AT for REE (82·8 %), four for TDEE (80·0 %) and two for SEE (100 %). A large heterogeneity in the methods used to quantify AT and between subjects and among studies regarding the magnitude of WL and/or of AT was reported. Well-designed studies reported lower or non-significant values for AT. These findings suggest that although WL may lead to AT in some of the EE components, these values may be small or non-statistically significant when higher-quality methodological designs are used. Furthermore, AT seems to be attenuated, or non-existent, after periods of weight stabilisation/neutral EB. More high-quality studies are warranted not only to disclose the existence of AT but also to understand its clinical implications on weight management outcomes.

Diet-induced weight loss (WL) is associated with reduced resting and non-resting energy expenditure (EE), driven not only by changes in body composition but also potentially by adaptive thermogenesis (AT). When exactly this happens, during progressive WL, remains unknown. The aim of this study was to determine the timeline of changes in RMR and exercise-induced EE (EIEE), stemming from changes in body composition v. the presence of AT, during WL with a very-low-energy diet (VLED). In all, thirty-one adults (eighteen men) with obesity (BMI: 37 (sem 4·5) kg/m2; age: 43 (sem 10) years) underwent 8 weeks of a VLED, followed by 4 weeks of weight maintenance. Body weight and composition, RMR, net EIEE (10, 25 and 50 W) and AT (for RMR (ATRMR) and EIEE (ATEIEE)) were measured at baseline, day 3 (2 (sem 1) % WL), after 5 and 10 % WL and at weeks 9 (16 (sem 2) %) and 13 (16 (sem 1) %). RMR and fat mass were significantly reduced for the first time at 5 % WL (12 (sem 8) d) (P<0·01 and P<0·001, respectively) and EIEE at 10 % WL (32 (sem 8) d), for all levels of power (P<0·05), and sustained up to week 13. ATRMR was transiently present at 10 % WL (−460 (sem 690) kJ/d, P<0·01). A fall in RMR should be anticipated at ≥5 % WL and a reduction in EIEE at ≥10 % WL. Transient ATRMR can be expected at 10 % WL. These physiological adaptations may make progressive WL difficult and will probably contribute to relapse.

Dieting makes you fat – the title of a book published in 1983 – embodies the notion that dieting to control body weight predisposes the individual to acquire even more body fat. While this notion is controversial, its debate underscores the large gap that exists in our understanding of basic physiological laws that govern the regulation of human body composition. A striking example is the key role attributed to adipokines as feedback signals between adipose tissue depletion and compensatory increases in food intake. Yet, the relative importance of fat depletion per se as a determinant of post-dieting hyperphagia is unknown. On the other hand, the question of whether the depletion of lean tissues can provide feedback signals on the hunger–appetite drive is rarely invoked, despite evidence that food intake during growth is dominated by the impetus for lean tissue deposition, amidst proposals for the existence of protein–static mechanisms for the regulation of growth and maintenance of lean body mass. In fact, a feedback loop between fat depletion and food intake cannot explain why human subjects recovering from starvation continue to overeat well after body fat has been restored to pre-starvation values, thereby contributing to ‘fat overshooting’. In addressing the plausibility and mechanistic basis by which dieting may predispose to increased fatness, this paper integrates the results derived from re-analysis of classic longitudinal studies of human starvation and refeeding. These suggest that feedback signals from both fat and lean tissues contribute to recovering body weight through effects on energy intake and thermogenesis, and that a faster rate of fat recovery relative to lean tissue recovery is a central outcome of body composition autoregulation that drives fat overshooting. A main implication of these findings is that the risk of becoming fatter in response to dieting is greater in lean than in obese individuals.

Adaptive thermogenesis is defined as a greater than predicted change in energy expenditure in response to changes in energy balance. This issue is particularly relevant in the context of a weight-reducing programme in which diminished thermogenesis can be sufficient to compensate for a prescribed decrease in daily energy intake. In the present pilot study, we investigated the adaptive reduction in thermogenesis in resting state that appears to favour resistance to further weight loss. Eight obese men (mean BMI: 33·4 kg/m2, mean age: 38 years) participated in this repeated-measures, within-subject, clinical intervention. They were subjected to a weight-loss programme that consisted of a supervised diet ( − 2930 kJ/d) and exercise clinical intervention. The phases investigated were as follows: (i) baseline, (ii) after 5 (se 1) kg loss of body weight (phase 1), (iii) after 10 (se 1) kg weight loss (phase 2) and (iv) at resistance to further weight loss (plateau). At each phase of the weight-reducing programme, body weight and composition as well as RMR were measured. A regression equation was established in a control population of the same age to predict RMR in obese men at each phase of the weight-loss programme. We observed that body weight and fat mass (FM) were significantly reduced (P < 0·05), while fat-free mass remained unchanged throughout the programme. In phase 1, measured RMR had fallen by 418 kJ/d, more than predicted (P < 0·05), and this difference reached 706 kJ/d at plateau (P < 0·05 v. phase 1). A positive association (r 0·64, P < 0·05) was observed between the reduction in thermogenesis and the degree of FM depletion at plateau. The adaptive reduction in thermogenesis at plateau was substantial and represented 30·9 % of the compensation in energy balance that led to resistance to further lose body weight. In conclusion, these results show that adaptive reduction in thermogenesis may contribute to the occurrence of resistance to lose fat in obese men subjected to a weight-reducing programme.

The present study was performed to further investigate the adaptive component of thermogenesis that appears during prolonged energy restriction. Fifteen obese men and twenty obese women underwent a 15-week weight-loss programme. During this programme, body weight and composition as well as resting energy expenditure (REE) were measured at baseline, after 2 and 8 weeks of energy restriction (-2929 kJ/d) and drug therapy (or placebo), and finally 2–4 weeks after the end of the 15-week drug therapy and energy restriction intervention, when subjects were weight stable. Regression equations were established in a control population of the same age. These equations were then used to predict REE in obese men and women at baseline, after 2 and 8 weeks, as well as after the completion of the programme. In both men and women body weight and fat mass were significantly reduced (P < 0.05 in all cases) while fat-free mass remained unchanged throughout the programme. At baseline, REE predicted from the regression equation was not significantly different from the measured REE in men, while in women the measured REE was 13 % greater than predicted. After 2 weeks of energy restriction, measured REE had fallen by 469 and 635 kJ/d more than predicted and this difference reached 963 and 614 kJ/d by week 8 of treatment in men and women respectively. Once body-weight stability was recovered at the end of the programme, changes in REE remained below predicted changes in men (-622 kJ/d). However, in women changes in predicted and measured REE were no longer different at this time, even if the women were maintaining a reduced body weight. In summary, the present results confirm the existence of adaptive thermogenesis and give objective measurements of this component during weight loss in obese men and women, while they also emphasize that in women this component seems to be essentially explained by the energy restriction.