GLM7 – A Novel Composite Glycolipid Index Derived from Routine Health Indicators for Enhanced Diagnosis and Prediction of Multimorbidity
Zhihua Wang, Shuo Chen, Xiaojun Feng, Xi Chen, Paul C. Evans, Hans Strijdom, Yu Ding, Jianping Weng, Suowen Xu
Abstract
Routine health examinations for healthy adults typically involve measurements such as height, weight, blood biochemistry, complete blood count, and urinalysis. However, the current scope of physical examinations has expanded to include numerous tests, some of which have questionable insight into underlying pathology. In this study, we analyzed 26,289 samples from the NHANES (National Health and Nutrition Examination Survey) database, along with 49 included indicators, to systematically explore the correlation between conventional indicators and various diseases. Our aim was to establish new diagnostic and predictive indicators. Initially, the top 10 diagnostic and predictive indicators for five disease categories, namely cardiovascular diseases, diabetes, liver diseases, cancer, and comorbidities, are identified, and the reliability of the routine test indicators is emphasized. Moreover, GLM7 (glycolipid metabolism 7 factors), a novel indicator integrating seven routine factors, has been developed. Restricted cubic spline (RCS) analysis and forest plot evaluations reveal its relationships and risk thresholds across diseases. An extreme gradient boosting (XGBoost) model using these factors exhibits excellent predictive performance in both the NHANES discovery and CHARLS (China Health and Retirement Longitudinal Study) validation cohorts. This study confirms conventional indicators' efficacy and introduces GLM7 as a tool for disease diagnosis/prediction, providing new insights into precise disease management.