Development of a Sex-Specific Prevalent Hypertension Discrimination Model in Korean Adults Using Genetic Risk Scores and Clinical Biomarkers: A Cross-Sectional Study

Figures

• 

  Figure 1. Association of genetic variants and genetic risk scores (GRSs) with hypertension risk. Odds ratios (ORs) and 95% confidence intervals (CIs) were derived from logistic regression analyses. Green dots indicate the OR point estimates, and the horizontal lines indicate the 95% CIs. a Unadjusted estimates are shown as reported in the original analysis. b Estimates adjusted for age and BMI are shown. The vertical dashed line indicates OR = 1 (no association). The GRS was constructed using prespecified selection criteria, including statistical significance (p < 0.05) in both the crude and age- and BMI-adjusted single-SNP logistic regression models, alignment of effect direction across genotype-model checks. SNP weights (β coefficients) were obtained from the corresponding unadjusted single-SNP logistic regression models. Genotype coding followed Table S5 (carrier coding for rs117559502 and rs62099117; additive coding for rs13175330 and rs6001482). Chr, chromosome; SNP, single-nucleotide polymorphism.

• 

  Figure 2. ROC curves for prevalent hypertension discrimination. (A) ROC curve for the total study population. (B) ROC curve for the male group. (C) ROC curve for the female group. Single-marker curves are labeled by predictor name; Model 1–3 correspond to the multivariable models defined in Table 3 (Model 1: BMI + ba-PWV; Model 2: BMI + ba-PWV + 8-epi-PGF2α; Model 3: BMI + ba-PWV + 8-epi-PGF2α + GRS, where available). The diagonal reference line indicates no discrimination (AUC = 0.5). 8-epi-PGF2α, 8-epi-prostaglandin F2α; ba-PWV, Brachial–ankle pulse wave velocity; BMI, Body mass index; GRS, Genetic risk score; ROC, Receiver operating characteristic.