Abstract / Summary
Abstract Background Longitudinal glycemic and blood-pressure profiles may provide information not captured by single measurements, but sparse biomarker waves complicate their interpretation in older adults. Methods We conducted a landmark analysis of adults aged 60 years or older with dysglycemia in CHARLS, ELSA, and HRS who had two paired HbA1c and systolic blood pressure (SBP) assessments. Cohort-specific Cox models related pre-landmark interpolated joint favorable-range time to post-landmark survey-reported heart disease or stroke, and estimates were pooled using random-effects meta-analysis. Results The analytic sample included 5,662 participants and 1,330 post-landmark composite events. In the primary overall-association Model 2, each 10-percentage-point higher interpolated glycemic–blood-pressure favorable-range time (GB-TTR) was associated with a pooled hazard ratio (HR) of 0.987 (95% CI 0.972–1.001; p = 0.075). The secondary component-adjusted incremental Model 3 estimate was 0.993 (95% CI 0.976–1.010; p = 0.403). Conclusions Higher pre-landmark interpolated GB-TTR was not robustly associated with lower post-landmark survey-reported cardiovascular disease. This two-wave interpolated phenotype is descriptive and should not be interpreted as a treatment target, causal effect, or validated risk-stratification measure.