权威代谢学刊发声：瘦身产业迎巨变！定制型疗法已掀起新浪潮

Core Viewpoint - Obesity has become a global health challenge affecting over 650 million adults, and traditional weight loss interventions show significant variability in effectiveness among individuals. A new predictive tool combining genetic information and physiological parameters has been developed to enhance personalized obesity treatment [6][9]. Research Background - Traditional BMI-based obesity classification fails to reveal individual differences in appetite regulation and energy metabolism, leading to inconsistent drug treatment outcomes. Current FDA-approved weight loss medications show up to threefold differences in effectiveness among individuals, highlighting the need for more precise efficacy prediction methods [9]. - The study focuses on the quantifiable physiological indicator of "satiety" through deep phenotyping and polygenic risk scoring, aiming to establish a personalized treatment prediction model [9]. Research Methodology - The study included 717 obese patients (BMI ≥ 30 kg/m²), with an average age of 41.1 years and average BMI of 37.0 kg/m². A standardized breakfast was consumed after an overnight fast, followed by a test to measure total caloric intake at maximum satiety [10]. - Genetic analysis involved extracting DNA from leukocytes and using the OmniExome v2.5 chip to detect 2637 loci, focusing on 41 obesity-related genes. Machine learning methods were employed to construct a satiety prediction model [11]. - Two randomized controlled trials were designed to validate the model, assessing weight changes in response to medications over specified periods [11]. Research Results 1. Drivers of Satiety Variability: Significant differences in caloric intake for satiety were observed, with males requiring more calories than females. Traditional body composition and metabolic rate indicators had limited explanatory power [12]. 2. Establishment and Validation of Genetic Risk Score: The genetic risk score (CTSGRS) showed a strong correlation with average satiety levels in training and validation cohorts, with specific gene variants contributing significantly to predictions [13]. 3. Predictive Ability of Genetic Risk Score: The machine learning-derived CTSGRS demonstrated excellent predictive performance, with AUC values indicating strong reliability in both training and validation phases [14]. 4. Individualized Drug Response Prediction: The study revealed significant individual differences in drug response, with specific patient profiles responding better to certain medications based on their CTS and CTSGRS levels [15]. Research Conclusion - The study successfully integrated genetic and physiological data to create a predictive model for obesity treatment. Key findings include the genetic basis of satiety differences, the strong predictive capability of the CTSGRS, and the variability in drug response among individuals [16]. Clinical Significance - This research marks a significant advancement in obesity precision classification, with the CTSGRS aiding clinical decision-making and potentially improving the efficacy of personalized medication strategies. Future work may optimize the model for broader weight loss interventions and explore gene-drug interactions for comprehensive obesity management [18].