The present findings represent the first study examining if metabolic adaptation, at the level of RMR, is associated with the magnitude of weight and FM loss in response to LED. We found that the larger the metabolic adaptation (RMRm-RMRp), immediately after weight loss (under negative EB), the lower the weight and FM loss seen, even after adjusting for variables known to modulate weight loss responses, namely adherence to the diet (27), average PAL (31), sex and baseline body composition (32). This suggests that metabolic adaptation may worsen weight loss outcomes during LEDs.
In the present analysis, individuals with obesity who had lost an average of 14 ± 4 kg (13%) of body weight, over 8 weeks on a LED, presented with a metabolic adaptation of approximately − 90 kcal/day at week 9. Our regression model showed that even after adjusting for adherence to the diet, average PAL, sex, baseline body composition, and randomisation group, metabolic adaptation was still a significant predictor of both weight and FM loss. On average, for each 50 kcal/day increase in metabolic adaptation, weight and FM loss were reduced by 0.5 kg. This might not seem of clinical relevance, given that the average metabolic adaptation was only approximately − 90kcal/day at week 9. However, in face of the large inter-individual variation in metabolic adaptation seen in the present analysis, ranging from − 337 to + 352 kcal/day, that would mean that those with the largest metabolic adaptation (RMRm-RMRp=-337 kcal/day) would lose 3 kg less of body weight and 2.7 kg of FM, compared with those with no metabolic adaptation (RMRm-RMRp = 0 kcal/day). This probably helps to explain some of the variation in weight (-28 to -7 kg) and FM (-19 to -6) loss seen in response to the LED.
It needs to be taken into consideration that we have only looked at metabolic adaptation at the level of RMR and several studies have shown that metabolic adaptation might in fact be of a larger magnitude at the level of non-resting energy expenditure (15, 16). This suggests that overall, metabolic adaptation (regarding TEE) might have an even larger contribution to weight and FM loss in response to LEDs.
From our knowledge only one study has previously reported an association between metabolic adaptation and weight loss outcomes in response to energy restricted diets. Goele and colleagues (33) reported that metabolic adaptation, at the level of RMR, explained 38% of the difference between measured and predicted weight loss in 22 out of the 48 women with overweight and obesity who experience metabolic adaptation after a 1000 kcal/day diet (33). However, this study suffers from several important methodological limitations. First, the association was only seen in a subgroup who experienced metabolic adaptation. Second, metabolic adaptation was defined as a reduction in RMR/ kg of FFM. Despite FFM being the main determinant of RMR, FM also contributes to RMR and should be included in the prediction model. Third, no adjustments were done for dietary adherence or PAL of the participants, both important determinants of weight loss outcomes in response to dietary interventions. The present findings confirm the preliminary findings by Goele and colleagues (33) and expand them further, by showing that metabolic adaptation modulates weight loss outcomes in both men and women with obesity, even after adjusting for dietary adherence and other important confounders. Moreover, we have recently shown that metabolic adaptation, at the level of RMR, increases the length of time necessary to achieve weight loss goals (BMI ≤ 25kg/m2), in premenopausal women with overweight (21).
The evidence previously discussed, together with the present findings, suggest that metabolic adaptation might be of clinical relevance. Despite the lack of association between metabolic adaptation and increased risk for weight regain (8, 9, 18), this phenomenon seems to be of clinical relevance in modulating weight loss in the short-term, in response to lifestyle interventions. Clinicians need to be aware of inter-individual variations in metabolic adaptation in response to negative EB when evaluating success to weight loss interventions and not assume that differences between measured and predict outcomes result only from “cheating” (reduced compliance to the intervention).
Our study has both strengths and limitations. The main strength is the fact that we have adjusted for important variables known to modulate weight loss outcomes following energy restricted diets, namely dietary adherence (27), PAL (31), as well as sex and body composition at baseline (32). Second, this analysis includes a heterogeneous sample of both males and females with obesity, with a wide range of BMI (30–43 kg/m2) and age (26–62 years), which is important for generalization purposes. The main limitation of our study is the fact that our estimation of dietary adherence was based on the assumption that energy needs equal RMR x PAL, which is not 100% acurate. However, the error associated with not having taken into account exercise economy is likely minor, as the SD of exercise economy in sedentary individuals has been shown to be around 12% of the mean value (34), meaning that in the present study differences in economy would only account for an average difference of 40kcal/day in the estimated TEE. Future studies should use TEE data from doubly labeled water to provide a more accurate estimate of energy needs.