Long-term blood pressure variability and risk of cardiovascular diseases in populations with different blood pressure status: an ambispective cohort study

Abstract

Objective We aimed to investigate the correlation between long-term blood pressure variability (BPV) and the risk of cardiovascular diseases (CVDs) among population with different blood pressure statuses (normotension, well-controlled hypertension, and uncontrolled hypertension). Methods In this ambispective cohort study, CVD-free residents aged over 50 years were consecutively enrolled from two community health service centers (CHCs) in Tianjin, China from April 2017 to May 2017. Information on blood pressure was retrospectively extracted from electronic medical records of CHCs between January 2010 and May 2017, and the occurrence of new-onset CVDs was prospectively observed during follow-up until September 2019. Long-term variation of SBP and DBP was assessed using four indicators: SD, coefficient of variation (CV), average successive variability (ASV), and average real variability (ARV). Cox proportional hazards regression model was developed to identify the potential impact of BPV on the incidence of CVDs. The receiver operating characteristic curve (ROC) was utilized to evaluate the predictive value of BPV indicators for the occurrence of CVDs. Results Of 1275 participants included, 412 (32.3%) experienced new CVD events during the median 7.7 years of follow-up, with an incidence density of 499/10 000 person-year in the whole cohort. Cox regression analysis revealed that almost all SBP and DBP variability indicators (except for SBP-SD) were significantly related to the risk of CVDs, especially among individuals with well-controlled hypertension. A trend toward an increased risk of CVDs across BPV quartiles was also observed. Moderate predictive abilities of BPV were observed, with the area under ROC curves ranging between 0.649 and 0.736. For SBP variability, SD had the lowest predictive ability, whereas for DBP variability, ARV had the lowest predictive ability. No significant association of CVDs with SBP-SD was found in all analyses, no matter as a continuous or categorical variable. Conclusion Elevated long-term BPV is associated with an increased risk of CVDs, especially among individuals with well-controlled hypertension. CV and ASV had higher predictive values than SD and ARV.