The impact of glycated hemoglobin trajectories on hypertension risk: a retrospective cohort study

Figures

• 

  Figure 1. Study design and participant flow diagram.

• 

  Figure 2. Patterns of HbA1c across three trajectory groups during follow-up. (A) Line graphs showing mean HbA1c levels across 7 sequential measurements. Trajectory 1 (red, n = 7,176) maintained a low level of ~5.5%; Trajectory 2 (green, n = 2,405) maintained a medium level of ~6.5%; and Trajectory 3 (blue, n = 557) maintained a high level of ~8.3%. (B) Box plots illustrating the distribution of HbA1c across the three trajectory groups, confirming that the groups maintained stable and distinct HbA1c distributions throughout the follow-up period.

• 

  Figure 3. Kaplan–Meier survival curve analysis for hypertension incidence across different tertile groups (T1–T3). The Kaplan–Meier curves demonstrate significant differences in cumulative hypertension incidence among the three groups (log-rank p < 0.001). Cumulative incidence rates were: T1 group (red line, n = 2,628): 21.84%; T2 group (green line, n = 3,870): 28.42%; T3 group (blue line, n = 3,640): 48.85%. Over the follow-up period (0–70 months), the survival curves progressively diverged, indicating persistent differences between groups over time. The inset displays a comparison of glycated hemoglobin (HbA1c) levels across groups.

• 

  Figure 4. Kaplan–Meier survival curve analysis for hypertension incidence across different trajectory groups (trajectory 1–3). The Kaplan–Meier curves demonstrate significant differences in cumulative hypertension incidence among the three trajectory groups (log-rank p < 0.001). Cumulative incidence rates were: Trajectory 1 (red line, n = 7,176): 26.77%; Trajectory 2 (green line, n = 2,405): 48.36%; Trajectory 3 (blue line, n = 557): 66.07%. Over the follow-up period (0–70 months), the survival curves progressively diverged, indicating persistent differences between groups over time. The inset displays a comparison of glycated hemoglobin (HbA1c) levels across trajectory groups.

• 

  Figure 5. Restricted cubic spline analysis of the non-linear dose–response relationship between baseline HbA1c levels and risk of hypertension. This figure demonstrates the non-linear association between HbA1c levels and the risk of incident hypertension after full adjustment for confounders. The solid line represents the adjusted hazard ratio (HR) estimates, with the gray-shaded area indicating the 95% confidence interval. The analysis identified a significant non-linear inflection point at HbA1c = 5.70% (marked by the blue vertical dashed line). Below this threshold, the relationship with hypertension risk remained relatively stable, while above 5.70%, the risk of hypertension increased significantly with rising HbA1c levels (non-linear p < 0.001). The red horizontal dashed line represents the reference risk level (HR = 1.0).

• 

  Figure 6. Association between HbA1c levels and risk of incident hypertension by trajectorys. The plot shows log-transformed hazard ratios (with 95% confidence intervals) for incident hypertension according to HbA1c levels, stratified by baseline HbA1c trajectory groups: low-stable (trajectory 1, red, n = 7,176, mean HbA1c = 5.57 ± 0.36%), medium-stable (trajectory 2, green, n = 2,405, mean HbA1c = 6.45 ± 0.59%), and high-stable (trajectory 3, blue, n = 557, mean HbA1c = 8.42 ± 1.39%). Estimates were derived using Cox restricted cubic spline model, adjusted for sex, age, ethnic group, marriage status, current drinking, current smoking, antihyperlipidemic agents, lipid-lowering medications, BMI, BUN, and eGFR, lymphocyte, neutrophil, LDL-C, TG, and HDL-C. Scatter points represent individual observations.