AIM-WEIGHT: AI-Guided Microbiome-Targeted Nutritional Intervention for Weight Loss

AIM-WEIGHT is a randomized, masked, placebo-controlled, parallel-group proof-of-concept study designed to obtain an initial controlled estimate of the efficacy and safety of a fixed microbiome-targeted nutritional intervention for weight management. The study is intended to determine whether the observed treatment signal is sufficiently promising to support a larger confirmatory trial and to provide a more precise estimate of the treatment effect for future study planning. The investigational formulation was selected before participant enrollment using an artificial-intelligence-guided formulation-development process. The AI methodology was used to inform the selection and dosing of candidate microbiome-active ingredients based on prior analyses of relationships between the gut microbiome and metabolic health. The resulting formulation is fixed and identical for all participants assigned to the active group. AI will not be used during the study for participant-level personalization, treatment assignment, dose modification, prediction, or clinical decision-making. Consequently, the study evaluates the clinical performance of the fixed formulation and is not designed as an independent validation of the AI algorithm. The study includes a 12-week primary weight-loss period followed by continued assigned intervention through Week 24. The Week 24 assessments are intended to provide supportive evidence regarding maintenance of the Week 12 effect during continued treatment. They should not be interpreted as evaluating maintenance after treatment withdrawal. The sample size is based on the primary endpoint of percentage change in body weight from baseline to Week 12. Assuming a between-group difference of 3.5 percentage points, a common standard deviation of 5.0 percentage points, and a two-sided significance level of 0.05, 34 evaluable participants per group provide approximately 81% statistical power. Allowing for approximately 15% attrition by Week 12, 80 participants will be randomized in total, with 40 assigned to each group. The primary analysis will compare randomized groups using analysis of covariance adjusted for baseline body weight and the prespecified randomization stratification factors.