Abstract MP28: The Association Between the Saliva Microbiome and Biomarkers of Cardiometabolic Disease in ORIGINS

Introduction: The present study sought to characterize taxonomic and functional characteristics of the saliva microbiome and their cross-sectional association with cardiometabolic biomarkers. Hypothesis: Features of the saliva microbiome will be associated with cardiometabolic risk. Methods: Saliva microbiome composition from 265 ORIGINS participants was defined with 16S rRNA sequencing (Illumina, DADA2). Functional composition (KEGG orthologs; KOs) was inferred with PICRUSt2. Alpha Diversity was computed with the Shannon Index. Blood pressure was defined as normal (SBP<120, DBP<80, and no antihypertensive medication) vs elevated/high (hypertension: SBP≥120, DBP≥80 and/or antihypertensive medication). Prediabetes was defined as 5.7≤HbA1≤6.4% or 100≤fasting plasma glucose≤125 mg/dL. HOMA-IR was calculated from insulin and glucose. HOMA-IR and CRP were dichotomized as high/low using thresholds of 3 mg/L and 1.9, respectively. Multivariable multinomial regression with cross-validation was conducted for each feature (species, KOs) and outcome, ranking regression coefficients from most positively to most inversely associated features. Microbiota and KO ratios (Micro-R, KO-R) were computed as the log-ratio of the sum of abundances of taxa/KOs in the top vs bottom tertiles of cross-validated ranks, such that higher ratios are potentially adverse. Adjusted multivariable logistic regressions produced ORs summarizing the association between Shannon, Micro-R, and KO-R and each cardiometabolic outcome. Results: Participants were 71% female, 22% Black, 40% Hispanic, and 18% white, with mean age of 33±10 years. Prevalence of hypertension was 36%, prediabetes was 17%, high CRP was 23%, and high HOMA-IR was 25%. Estimated ORs for the association between microbiome metrics and cardiometabolic outcomes are summarized in Figure 1. Conclusion: Microbiome scores based on specific taxa or KOs are associated with hypertension, prediabetes, and increased inflammation and insulin resistance.