Multiomics Outperform Phenotype in Predicting Post-Diet Weight Regain and Body Composition Dynamics

Highlights

• Multiomic models, incorporating gut microbiome and fecal metabolome data, significantly outperform traditional phenotypic models in predicting weight loss and subsequent regain.
• The multiomic approach achieved an Area Under the Curve (AUC) of 0.95 for predicting clinically meaningful weight loss (≥5%) during the calorie-restricted phase.
• Key baseline predictors, including N-acetyl-L-aspartic acid and specific gut microbes like Bifidobacterium adolescentis, were identified as shared markers for weight and body composition changes across both phases.
• These findings suggest that baseline biological signatures can serve as a foundation for personalized weight management strategies to mitigate the common challenge of weight regain.

The Challenge of Weight Recidivism in Clinical Practice

Obesity remains one of the most significant public health challenges globally, yet the clinical response to it is often frustrated by the high rate of weight regain. While many patients can successfully lose weight through calorie restriction, maintaining that loss is notoriously difficult. This biological resistance—often referred to as metabolic adaptation or the “yo-yo effect”—suggests that weight regulation is a complex interplay of genetics, behavior, and physiological factors that are not fully captured by clinical phenotype alone.

Traditional predictors of weight loss, such as baseline BMI, age, or metabolic rate, have historically shown limited predictive power in individual outcomes. As we transition toward an era of precision medicine, there is an urgent need to identify robust biomarkers that can forecast how an individual will respond to dietary interventions and, perhaps more importantly, who is at the highest risk for weight regain after the intervention ends.

Study Design and Methodology: The LEAN-TIME Trial

A post hoc analysis of the Low-Carbohydrate Diet and Time-Restricted Eating (LEAN-TIME) feeding trial, recently published in Diabetes Care, sought to address this gap. Researchers evaluated 88 adults with overweight or obesity who completed a 12-week calorie-restricted weight-loss phase. Of these, 79 individuals continued through a 28-week weight-regain phase, providing a unique longitudinal dataset of metabolic flux.

The study was comprehensive in its data collection. At baseline, researchers gathered dietary intake data, metabolic markers, fecal metabolome profiles, and gut microbiome sequences. These were used as candidate predictors for changes in three primary outcomes: total body weight, Body Fat Mass (BFM), and Soft Lean Mass (SLM). To develop the prediction models, the team utilized multivariable regression and the least absolute shrinkage and selection operator (LASSO) model to identify the most significant predictors and create weighted-sum models.

Key Findings: Superiority of Multiomic Integration

The results of the analysis provide compelling evidence for the value of multiomic data in metabolic forecasting. In both the weight-loss and weight-regain phases, models that integrated multiomic data with phenotypic features significantly outperformed those relying solely on clinical phenotypes (P < 0.05).

Weight Loss Phase (Weeks 0-12)

During the initial 12-week calorie restriction, the multiomic and phenotypic model demonstrated strong predictive performance. The R2 values—representing the proportion of variance explained by the model—were 0.49 for weight change, 0.61 for BFM, and 0.54 for SLM. The corresponding root mean square errors (RMSEs) were 1.59 kg, 1.41 kg, and 0.98 kg, respectively. Notably, for binary classification of whether a patient would achieve a clinically meaningful weight loss of 5% or more, the model achieved an impressive AUC of 0.95, with a sensitivity of 94.12% and a specificity of 86.79%.

Weight Regain Phase (Weeks 12-40)

Predictive accuracy was even higher during the regain phase. The R2 values reached 0.72 for weight change, 0.73 for BFM, and 0.66 for SLM. The RMSEs remained low (1.40 kg for weight), suggesting that baseline multiomic signatures can accurately forecast an individual’s tendency to regain weight nearly seven months after the initial intervention. This high predictive power for weight regain is particularly valuable, as it identifies high-risk individuals before they begin a weight-loss program.

Shared Biomarkers: The Role of the Microbiome and Metabolome

One of the most clinically relevant aspects of this study is the identification of specific baseline predictors that were shared across both the loss and regain phases. These predictors were primarily derived from the gut microbiome and fecal metabolome, reinforcing the theory that the gut environment plays a central role in energy homeostasis.

Specifically, the study highlighted Ruminococcus callidus and Bifidobacterium adolescentis as key microbial predictors. Bifidobacterium species are often associated with gut health and have been linked to improved metabolic outcomes in previous literature. On the metabolic side, N-acetyl-L-aspartic acid emerged as a significant predictor. The presence of these shared markers suggests that an individual’s baseline metabolic and microbial “milieu” sets a trajectory for how their body will manage weight fluctuations over the long term.

Expert Commentary: Mechanistic Insights and Clinical Utility

The findings from the LEAN-TIME post hoc analysis represent a significant step toward precision nutrition. The ability to predict weight regain with over 70% accuracy using baseline data is a breakthrough. From a clinical perspective, this could allow physicians to tailor the intensity of follow-up care. For example, a patient whose baseline multiomic profile suggests a high risk of regain might be prioritized for more frequent dietary counseling, pharmacotherapy (such as GLP-1 receptor agonists), or metabolic monitoring after the initial weight-loss phase.

However, while the statistical performance is robust, experts note that the implementation of such models in routine clinical practice faces hurdles. Multiomic sequencing—particularly metabolomics and metagenomics—is currently expensive and requires specialized bioinformatics support. Furthermore, while these baseline features are predictive, the biological mechanisms through which N-acetyl-L-aspartic acid or specific Ruminococcus species influence weight regain remain to be fully elucidated. It is possible these markers are indicative of underlying systemic inflammation, insulin sensitivity levels, or specific dietary habits that persist beyond the intervention period.

Conclusion

The LEAN-TIME trial analysis demonstrates that baseline multiomic and phenotypic data are highly effective at predicting weight and body composition dynamics. By identifying individuals who are biologically predisposed to weight regain, clinicians can move away from one-size-fits-all dietary recommendations toward targeted, evidence-based interventions. Future research should focus on validating these models in larger, more diverse cohorts and investigating whether modifying these baseline microbial or metabolic factors can alter weight loss trajectories.

References

1. Li L, Li R, Qiu Z, et al. Prediction of Weight Loss and Regain Based on Multiomic and Phenotypic Features: Results From a Calorie-Restricted Feeding Trial. Diabetes Care. 2026;49(1):68-77. doi:10.2337/dc25-0728.
2. Hall KD, Kahan S. Maintenance of Lost Weight and Predictions of Weight Regain: A Review. Med Clin North Am. 2018;102(1):183-197.
3. Zmora N, et al. Personalized Stepping-Stone Approach to Weight Loss Maintenance. Nature. 2018;560(7719):494-498.