Extended Data Fig. 3: Comparison of EMRAge and PhenoAge in association with aging-related health outcomes in the All of Us cohort.
From: OMICmAge quantifies biological age by integrating multi-omics with electronic medical records
The analysis compares associations from individual (a) and joint (b) perspectives. Error bars show hazard/odds ratios and 95% confidence intervals per one-standard deviation change from the estimated mean value. For individual analysis, each aging metric (EMRAge, PhenoAge, or chronological age) was tested separately, adjusting for sex, race, BMI, smoking status, and alcohol consumption. Chronological age was also adjusted for when testing EMRAge or PhenoAge. For joint analysis, both EMRAge and PhenoAge were included as predictors in a single model, without additional adjustment. The number of cases (n) and sample sizes (N) for each phenotype are as follows: Incident Cases: all-cause mortality, N = 10,769 (n = 378); stroke, N = 9,285 (n = 548); type-2 diabetes, N = 6,946 (n = 503); COPD, N = 8,878 (n = 402); Depression, N = 6,226 (n = 639); CVD (excl. stroke), N = 4,734 (n = 658); cancer, N = 8,942 (n = 380). Prevalent Cases: The total sample size for prevalent diseases is 10,769, with the following case numbers: stroke, n = 1,484; type-2 diabetes, n = 3,823; COPD, n = 1,891; Depression, n = 4,543; CVD (excl. stroke), n = 6,035; cancer, n = 1,827.
