Few studies estimated the impact of antihypertensive adherence on cardiovascular diseases (CVD) in a longitudinal cohort with presence of time-dependent confounders. This study aims to assess the association between antihypertensive adherence and CVD using marginal structural Cox model (MSM-Cox) and to characterize blood pressure (BP) trajectories of patients with different adherence.
This longitudinal study included 16,896 hypertensive patients receiving antihypertensive medication. The median follow-up time was 3.5 years (25th to 75th, 1.75-4.75 years). BP and medication adherence were measured four times every year. We used MSM-Cox and Cox model to assess association between antihypertensive adherence and CVD events. The linear mixed-effects model was used to characterize BP trajectories of patients with different adherence, and the area under curves (AUC) was calculated as BP burden.
We documented 4735 CVD events, crude incidence of CVD was 80.8 (95% CI, 78.1-83.4) and 112.6 (95% CI, 107.2-118.0) per 1000 person-years for baseline high-adherence and low-adherence, respectively. Compared with high adherence, the adjusted hazard ratio (HR) for association between low adherence with CVD was 1.75 (95%CI, 1.62-1.89) and 1.34 (95%CI, 1.26-1.42) based on the MSM-Cox and the Cox model, respectively. The BP burden and fluctuation range of BP trajectory in low-adherence patients were larger than those of high-adherence patients. Patients with high adherence got 28% greater reduction of BP burden than low-adherence patients.
Antihypertensive adherence was more strongly associated with the risk of CVD than conventional regression analyses based on a single adherence measurement.