Waist-to-Height Ratio Identifies Cardiometabolic Risk Not Captured by Body Mass Index: The All of Us Research Program

Background: Body mass index (BMI) does not characterize central adiposity and may obscure cardiometabolic risk among adults classified as normal weight. We evaluated associations between waist-to-height ratio (WHtR) and incident disease across the BMI spectrum and among adults with normal BMI. Methods: We studied 308,600 adults in the All of Us Research Program. Participants with the corresponding prevalent disease were excluded separately for eight cardiometabolic outcomes. Cause-specific Cox models adjusted for demographic, socioeconomic, smoking, and BMI-related factors. We evaluated continuous and nonlinear WHtR associations, elevated WHtR among normal-BMI adults, and incremental discrimination. Additional analyses tested WHtR-by-BMI-category interactions. Results: Among 77,684 adults with normal BMI, 24,301 (31.3%) had WHtR 0.50. In whole-cohort analyses, adjusted hazard ratios (HRs) per 0.05 higher WHtR ranged from 1.07 to 1.31 across the eight outcomes. In BMI-stratified interaction models, corresponding diabetes HRs were 1.52 (95% CI, 1.43-1.61) in normal BMI, 1.42 (1.36-1.49) in overweight, and 1.22 (1.20-1.25) in obesity (P for interaction <0.001). Hypertension associations also differed across BMI groups (P for interaction <0.001). Among normal-BMI adults, WHtR 0.50 versus <0.50 was associated with diabetes (HR, 1.74), myocardial infarction (1.49), and ischemic stroke (1.52). Associations with all eight outcomes persisted in normal-BMI landmark analyses excluding the first year. Flexible age adjustment preserved the principal findings. Adding nonlinear WHtR improved cross-validated discrimination, most notably for diabetes (C-index=0.0134). Conclusions: WHtR identified cardiometabolic risk beyond BMI classification. Associations with diabetes and hypertension were particularly pronounced among normal-BMI adults, supporting assessment of central adiposity alongside BMI.