Saliva, plasma, and multi-fluid metabolomic profiles of excess adiposity and their associations with diabetes progression among Puerto Ricans

<title>Abstract</title> Introduction: Plasma metabolomic profiles of BMI are associated with type 2 diabetes (T2D) risk, but none have examined saliva or combined information from multiple biofluids. Our aim was to characterize saliva, plasma, and plasma-saliva multi-fluid metabolomic profiles of BMI and waist circumference (WC) and examine their associations with diabetes progression. Methods We included 911 participants from the San Juan Overweight Adult Longitudinal Study, a 3-year prospective cohort of overweight Puerto Ricans. At baseline, using LC-MS, we quantified metabolites from saliva (n = 635) and plasma (n = 1,051). We used elastic net regression with 10-fold cross-validation to identify features from saliva, plasma, and saliva and plasma (multi-fluid) that were predictive of BMI and WC. We used Cox-proportional hazard models to evaluate associations between metabolomic profiles and diabetes progression, adjusting for age, sex, socioeconomic factors, lifestyle, and medication use. Results For BMI metabolomic profiles, we identified 207 metabolites in plasma, 118 metabolites in saliva, and 225 in the multi-fluid profile. For WC, we identified 157, 89, and 210 metabolites for saliva, plasma, and multi-fluid profiles, respectively. Highly positively weighted metabolites across all BMI and WC metabolomic profiles included those in pathways of alanine and aspartate metabolism, purine metabolism, and sphingomyelins. Each SD increase in saliva, but not plasma or multi-fluid, metabolic profile of BMI was significantly associated with all stages of diabetes progression. Saliva, but not plasma or multi-fluid, metabolomic profile of WC was significantly associated with progression from pre-diabetes to T2D. All associations became stronger after further adjustment for anthropometric measures of BMI and WC. Conclusion Saliva is an underexplored and easily accessible biofluid to measure metabolites that are reflective of adiposity measures. Validation in future studies will confirm our findings that saliva, plasma, and multi-fluid metabolites can supplement measures of BMI and WC to better predict diabetes progression among overweight and obese Puerto Rican adults.