Incidence rate ratios (IRR) were calculated using survey-weighted generalized linear regressions for each pair of cardiometabolic outcome (i.e., incident diabetes, incident hypertension) and SDB related phenotypes as the predictor in the combined dataset (batch 1 and batch 2), and are presented as effect estimates with 95% confidence intervals (CIs). p values were derived from a one degree-of-freedom Wald test. Exact IRR, 95% CIs and p values are provided in Supplementary Data 14 . Model 1 adjusted for demographic variables, including age, sex, field center, Hispanic/Latino background (Mexican, Puerto Rican, Cuban, Central American, Dominican, and South American and other/multi) and body mass index (BMI). Model 2 adjusted for all model 1 covariates and lifestyle variables alcohol use, cigarette use, total physical activity (MET-min/day), and diet (Alternative Healthy Eating Index 2010) in addition to demographic variables. REI3 Respiratory Event Index (REI) computed over all respiratory events, defined as apneas or hypopneas with at least 50% cannula flow reduction for a minimum duration of 10 s with >=3% oxygen desaturation; HB hypoxic burden; SDB PC1 MRS: metabolite risk score calculated based on the coefficients from LASSO regression trained in both sexes combined to predict SDB PC1 in the discovery dataset (batch 1); SDB PC2 MRS metabolite risk score calculated based on the coefficients from LASSO regression trained in both sexes combined to predict SDB PC2 in the discovery dataset (batch 1); OSA LASSO MRS: metabolite risk score calculated based on coefficients from LASSO regression trained to predict OSA in previous publication 30 ; incident_dm: Incident diabetes (fasting glucose >=126 mg/dL, or post-OGTT glucose >=200 mg/dL or A1C >= 6.5%, or self-report of diabetes), n = 2908 samples for Model 1 and n = 2874 samples for Model 2; incident_htn: incident hypertension (systolic or diastolic blood pressure is greater than or equal to 140/90 or participant self-reported as currently taking antihypertensive mediations), n = 2388 samples for Model 1 and n = 2360 samples for Model 2. Source data are provided in Supplementary Data 14 . * indicates p < 0.05. ** indicates p < 0.01. *** indicates p < 0.001.
