Obesity is a major health problem globally, and while social and environmental factors contribute, the proximate cause of weight gain remains an energy imbalance between intake and expenditure. Accurate TEE measurement is crucial for research in this area. TEE encompasses energy used for basal metabolism, thermoregulation, digestion, physical activity, and other physiological processes. Low TEE has been suggested as an obesity risk factor, but research has yielded mixed results. Some studies indicate a link between low TEE and increased body fat in children, while others show no association between TEE and changes in body fat percentage in adults and children. Methodological limitations, such as small sample sizes and the reliability of TEE measurements, may contribute to conflicting findings. Changes in weight and adiposity under normal conditions are often slow and difficult to detect over short periods; large samples are needed to discern significant associations. The reliability of TEE measurements is another critical factor, as fluctuations over time due to physiological or behavioral changes or measurement error can confound results. Prior studies have shown repeatability in basal metabolic rates and 24-h energy expenditures, but less is known about TEE's repeatability, especially regarding the time interval between measurements and age differences. This study aimed to determine the influence of age on TEE repeatability and examine the association between TEE and changes in weight or body composition utilizing the IAEA DLW database.

Existing research on the relationship between total energy expenditure (TEE) and body composition or weight change presents conflicting results. Early studies using doubly labeled water (DLW) showed similar TEE in obese and normal-weight women after adjustment for body weight and composition. However, other research suggested that low TEE in infants and children predicted greater body fat gain, while high TEE was associated with higher body fat gain in preadolescent girls. Conversely, some longitudinal studies found no predictive relationship between TEE and body fat changes in infants, children, or adult women. A study using heart rate monitoring reported an inverse association between energy expenditure and fat mass in younger participants, but a positive association in older individuals. These discrepancies highlight the need for larger sample sizes and reliable TEE measurements to draw robust conclusions. The repeatability of TEE measurements has been examined in limited studies, showing some level of repeatability, particularly in adults. However, the impact of the time interval between measurements and age differences on repeatability is unclear. These inconsistencies in the literature underscore the need for further investigation into TEE's relationship with weight and body composition changes across different age groups.

This study utilized the International Atomic Energy Agency (IAEA) DLW database (version 3.1.2), which comprises 6,787 TEE measurements from individuals across multiple studies. A common calculation method was employed to standardize TEE estimations. The study included individuals over 1 year old with at least two TEE measurements (348 adults and 47 children, totaling 696 TEE measurements). Repeatability, or intraclass correlation coefficient (ICC), 'R', was calculated using a mixed-effects model. The model incorporated FFM, FM, sex, and age as fixed factors to control for their influence on TEE variance. Two approaches examined the association between TEE and body composition changes in adults (20-60 years old). First, a multi-response model decomposed the covariance between TEE and FM (fat mass) and TEE and body fat percentage into between-individual (r_ind) and within-individual (r_e) components. r_ind reflects individual mean values correlation and r_e represents the correlation of changes in traits over time within individuals. Second, an adjusted TEE was calculated, controlling for FFM, FM, age, and sex using a multiple regression model; The adjusted TEE was then used to test whether it was associated with changes in body weight and body fat percentage, both in the full dataset and in a subset where the time interval between measurements exceeded 4 weeks. This subset helps mitigate measurement errors' impact on calculated changes in FFM and FM. Statistical analysis employed R version 3.6.2. Repeatabilities were tested using likelihood ratio tests (LRT) and permutation tests, with parametric bootstrapping generating confidence intervals. Associations between adjusted TEE and weight/body composition changes were assessed using linear models. Multivariate Bayesian mixed models using the MCMCglmm package assessed covariations between TEE and FM or body fat percentage, partitioning variance into within- and between-individual components. All analyses used In-transformed values for TEE, FFM, and FM to improve model assumptions.

Repeatability analysis revealed that adjusted TEE, controlling for FFM, FM, sex, and age, was significantly repeatable (R = 0.54) for all individuals (adults and children combined). However, a marked difference was observed when analyzing adults and children separately. Adjusted TEE showed high repeatability in adults (R = 0.64), but not in children (R = 0.00). Body mass, adjusted for sex and age, displayed higher repeatability than TEE in both adults and children (R > 0.9 for adults, and 0.38 for children). Analyses examining the association between TEE and body composition changes used two approaches. Firstly, multi-response mixed models showed no within-individual (r_e), among-individual (r_ind), or phenotypic (r_p) correlations between TEE and body fat, regardless of representing body fat as fat mass or body fat percentage. Secondly, adjusted TEE (accounting for FFM, FM, age, and sex) demonstrated no significant correlation with short-term changes in body composition (body fat percentage). However, a positive association was found between the difference in adjusted TEE between measurements and changes in body weight. This indicates individuals with a larger increase in adjusted TEE over time tend to weigh more. Neither average adjusted TEE nor the difference in adjusted TEE between measurements showed a significant association with body fat percentage changes. These findings are consistent in both the full dataset and the subset with a longer time interval between measurements (over 4 weeks), indicating robustness and consistency over longer timescales. The observed positive relationship between changes in TEE and weight gain challenges previous hypotheses that lower TEE would lead to weight gain.

The study's key finding—the high repeatability of TEE in adults—indicates that individual differences in metabolic rate are stable over time. This suggests that the concept of a "fast" or "slow" metabolism in adults holds merit. Yet, despite this repeatability, the study found no evidence linking variations in TEE to changes in body weight or fat mass over the observed short-to-medium-term periods. The lack of association between TEE and body fat changes contradicts previous hypotheses and suggests that variations in metabolic rate alone are not sufficient to explain individual differences in weight or fat accumulation. The positive relationship between changes in adjusted TEE and weight change is intriguing and may reflect the complex interplay of energy intake and expenditure. Further research into specific factors influencing energy intake may shed light on this. The lack of correlation in the long-term interval between measurements suggests that it is likely that even over longer timescales, no association exists.

This study using the extensive IAEA DLW database demonstrates high repeatability in total energy expenditure (TEE) among adults, confirming 'fast' and 'slow' metabolisms as durable traits. However, it challenges the long-held assumption that low TEE is a risk factor for weight gain or that high TEE protects against it. No correlation was observed between TEE and short-term changes in body composition in adults. Future research should incorporate physical activity measures and investigate the effect of organ size and aging-related metabolic changes. Further exploration of the interaction between TEE and energy intake is needed to better understand individual weight variation.

The study is limited by the lack of data on physical activity levels, organ size, and resting energy expenditure. The absence of these variables hinders a complete understanding of the factors contributing to TEE variability and its relationship with body composition. Additionally, the use of short-term measurements may limit the detection of subtle but significant relationships between TEE and longer-term body composition changes. Finally, the data's reliance on a database compiled from multiple studies with potentially varying methodological approaches may introduce some level of heterogeneity. The study is also limited to healthy individuals and excludes athletes or those undergoing interventions, potentially limiting generalizability to these populations.