[Development and validation of a prediction model for dyslipidemia in adults based on gut microbiota features of 4 provinces].

OBJECTIVE: To develop and validate a 5-year dyslipidemia prediction model for adults aged ≥18 years based on gut microbiota data. METHODS: Data were derived from the China Development and Nutrition Health Impact Cohort Study(formerly the China Health and Nutrition Survey, 2018-2023). We included 650 adults aged ≥18 years with baseline gut microbiota data and non-dyslipidemia. Key predictors of dyslipidemia were selected using LASSO regression, and a Cox proportional hazards regression model was constructed to assess dyslipidemia risk. Internal validation was performed using 1000 bootstrap resamples. Model performance was evaluated by the area under the receiver operating characteristic curve(AUC-ROC), Harrell's concordance index(C-index), and calibration curves. RESULTS: The median(Q1, Q3) age of participants was 55.2(46.5, 65.9) years old. Over the 5-year follow-up, 138 individuals(21.2%) developed dyslipidemia. LASSO regression identified 30 key predictors, with 23 variables(e. g. , body mass index, Eubacterium xylanophilum group, and Ruminococcus gauvreauii group) retained in the final Cox model. The model demonstrated strong discriminative ability for 5-year dyslipidemia risk(AUC-ROC: 0.7843) and a bootstrap-corrected C-index of 0.6944. Calibration curves indicated high agreement between predicted and observed probabilities of dyslipidemia. CONCLUSION: This gut microbiota-based prediction model, developed from a nutrition-health cohort, exhibits robust performance and may provide basis for personalized and precise intervention.