Abstract
BACKGROUND: Subtle biological changes related to frailty may be undetected by standard clinical methods, and reliable biomarkers for frailty are still under investigation. This study was conducted to profile plasma metabolite patterns associated with frailty and validate the most significant metabolite for identifying and predicting frailty in cross-sectional and longitudinal analyses. METHODS: The "Fujian Prospective Aging Cohort" (ChiCTR 2000032949) enrolled 2,265 community-dwelling individuals aged 60 and above in 2020. Plasma metabolites were analyzed using ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Frailty was assessed using Fried's phenotype and the Frailty Index. RESULTS: Widely targeted metabolomic analysis identified 889 metabolites. GAA was identified as the top frailty-associated candidate by ROC analysis and validated in a large cross-sectional cohort (AUC = 0.670). This cohort (N = 1,972) confirmed that subjects with lower GAA levels had a higher prevalence of frailty (P < .001). Multinomial logistic regression showed that higher GAA levels were significantly associated with lower odds of prefrailty and frailty; the ORs were 0.46 (95% CI: 0.32-0.66), and 0.15 (95% CI: 0.07-0.33) in the highest quartile, both P < .001). Over a 3-year follow-up period, a group-based trajectory model identified three Frailty Index trajectories: low-elevated (59.6%), moderate-elevated (34.1%), and high-elevated (6.3%). Subjects in the highest GAA quartile had a 36% and 66% lower likelihood of following moderate-elevated and high-elevated Frailty Index trajectories (P = .016 and P = .022). CONCLUSIONS: This study identifies GAA as a potential metabolic biomarker for frailty. Higher GAA levels are associated with lower frailty odds and provide predictive value for a lower likelihood of frailty progression.