Background Weight loss‐induced metabolic adaptation refers to the decrease in thermogenesis to an extent greater than predicted based on the change in body weight and composition 1 (Figure 1). Teleologically, this phenomenon might arise as a way to reduce the gap between energy intake and expenditure, so further weight loss is prevented and chance of survival is enhanced. Such adaptation could be of benefit in times of famine; however, in an obesogenic environment, it dampens the rate of weight loss in individuals attempting to lose it. Many strategies have been pursued to maximize weight loss including the use of thermogenic boosters. In general, none has proved to be successful. Comprehensive understanding of the underlying mechanisms of metabolic adaptation remains a challenge that deserves further assessment. In this Perspective, critical conceptual and analytical aspects for studying metabolic adaptation are discussed. Figure 1 Open in figure viewerPowerPoint Relevance of the method to estimate weight loss‐induced metabolic adaptation. The figure shows the change in resting metabolic rate as a function of the decrease in fat‐free mass. Three types of individuals are represented by letters A, B, and C. Subject A shows a drop in thermogenesis; however, by considering only the residual value after weight loss, this individual will be considered as a non‐metabolically adapted subject. In turn, subjects B and C will be considered as having a similar degree of metabolic adaptation based on the negative residual value after weight loss; however, only subject B is metabolically adapted when considering the residual value at baseline. This hypothetical scenario clearly shows the need for considering the difference between residual values (after—baseline) to define weight loss‐induced metabolic adaptation.

How Should Metabolic Adaptation Be Estimated? Metabolic rate is dependent on the size of fat‐free and fat masses, which explain 60‐70% and ∼5% of the variance in resting thermogenesis in humans, respectively 2. Therefore, the definition of metabolic adaptation must take into account body composition, particularly when body composition is modified. Based on the fact that thermogenesis does not change in direct proportion with body size (i.e., allometric relationship) and the intercept of the regression line relating these two variables is not zero, using mathematical ratios is inappropriate to control for intra‐ and inter‐individual differences in body size. An accepted method to control thermogenesis for body size (and its composition) is through residual analysis (i.e., comparison of observed relative to predicted values from regression analysis). Through this approach, weight loss‐induced metabolic adaptation is often defined by the extent of the negative residual value at follow‐up 3-7. Thus, the lower the negative residual value, the higher the extent of metabolic adaptation. However, this method does not entirely account for the essence of this phenomenon, because it does not consider the residual value at baseline. As a consequence, the trajectory of the change in body size and thermogenesis is dismissed. Figure 1 shows hypothetical examples highlighting the need for an alternative approach to define metabolic adaptation. In this regard, Doucet et al. 8 defined metabolic adaptation as the difference between residual values at follow‐up relative to baseline. Non‐published analysis of our data set including 58 individuals with typ 2 diabetes before and after a 1‐year weight loss intervention 9 showed only a moderate association between methods (r = 0.50; P < 0.01). Therefore, the method chosen may influence the outcome, particularly when correlation analysis of individual values of metabolic adaptation is conducted. Meanwhile, it should have little effect on the outcome considering that the mean baseline residual value should be zero.

Enhanced Energy Efficiency as an Explanation of Metabolic Adaptation ATP synthesis at a lower oxygen need is proposed as a putative mechanism accounting for metabolic adaptation. On the one hand, circulating triiodothyronine concentration decreases in response to energy restriction, and this hormone has been shown to induce in vivo skeletal muscle mitochondrial uncoupling, i.e., decreases energy efficiency. On the other hand, sympathetic tone is reduced while parasympathetic tone is increased after weight loss. These neuroendocrine changes might enhance skeletal muscle contractile energy efficiency after weight loss 10. Alternatively, brown adipose tissue in adult humans and the known influence of thyroid hormones and catecholamine in its metabolic activity could also contribute to decrease facultative thermogenesis.

Metabolic Adaptation in Total and Resting Thermogenesis Body weight loss depends on the extent of the negative energy balance. The degree of metabolic adaptation in resting and total thermogenesis will partially determine how severe negative energy balance will be. Metabolic adaptation of resting and total thermogenesis may not necessarily be related. Eventually, compensation between each other may exist. For instance, a decrease in physical activity after weight loss may reduce the extent of metabolic adaptation in resting thermogenesis. If any compensation exists, it would challenge the notion that metabolic adaptation prevents further weight loss. Indeed, non‐published analysis of our data set 9 found a weak association between weight loss and metabolic adaptation (r = 0.25; P = 0.07), which is in line with earlier studies 5, 6. Somehow, these observations are against the prediction that greater weight loss will be found in non‐metabolically vs. metabolically adapted individuals. Alternatively, the extent of the metabolic adaptation may be mostly determined by the energy deficit, and this adaptation can only partially compensate energy restriction. In any case, testing this hypothesis requires comprehensive analysis of energy intake and physical activity.

Role of Organ Size on Metabolic Adaptation As already mentioned, adjustment for body composition is the conventional way of analysis of metabolic adaptation. Individuals with obesity have larger high‐metabolic‐rate organs and skeletal muscle mass than lean individuals. After weight loss, the relative contribution of these organs to thermogenesis decreases, leading to a lower specific metabolic rate of fat‐free mass. Thus, a greater than predicted reduction in thermogenesis may be attributed to a change in body composition rather than enhanced energy efficiency. Therefore, when organ size is taken into account, metabolic adaptation is decreased at a level close to day‐to‐day variation in resting metabolic rate 11.

Some Statistical Remarks A statistical decision rule to define metabolic adaptation is desirable at an individual level beyond the comparison of observed vs. predicted values. A prediction interval predicts a y‐value for a certain x‐value based on the regression line, while a confidence interval refers to the estimation of the mean on the y‐axis at a given x‐value. Prediction intervals are wider than confidence intervals. Thus, for single measurements, prediction limits may identify metabolic adaptation when the value falls outside the prediction interval. However, regression lines based on small sample sizes have the disadvantage of generating prediction intervals that are often too wide to be useful for this purpose. Alternative statistical decision rules may be developed for repeated measurements obtained before and after weight loss.

Concluding Remarks Better understanding of metabolic adaptation is limited due to our incipient access to high‐technology instruments that allow us to measure volume and mass of organs from the whole human body. Comprehensive analysis of ATP synthesis and oxygen consumption at the tissue/organ level will account for the heterogeneity in metabolic activity of the components of fat‐free mass. It is also critical to standardize the way metabolic adaptation is estimated. It appears inappropriate to do so only by the extent of the negative residual value of metabolic rate after weight loss. Identifying mechanisms and predictors of the extent of metabolic adaptation may contribute to better treatment of obesity.