Blood pressure variability and cardiovascular events in high-risk individuals: individual patient data analysis of SPRINT and ACCORD trials

Abstract Background and Aims: Blood pressure variability (BPV) is associated with cardiovascular risk and has been shown to confer prognostic information independent of mean blood pressure (BP). However, the consistency of its incremental predictive value across different clinical settings and populations warrants further investigation. A patient-level pooled analysis of two large randomized trials (SPRINT and ACCORD) was conducted to clarify the association between BPV and major cardiovascular events (MCEs). Methods: Visit-to-visit BPV was calculated from month 3 onwards using multiple metrics (including variation independent of mean, VIM) in participants with ≥3 visits. Associations between BPV and MCEs (myocardial infarction, stroke, or cardiovascular death) were assessed using Cox regression and restricted cubic splines. Results: Among 18,415 participants (median 12 BP measurements; 3.6-year follow-up), 1244 (6.8%) experienced MCEs. After multivariable adjustment, higher SBP-VIM (highest versus lowest tertile) was associated with greater risk of MCEs (hazard ratio 1.15, 95% confidence interval 1.00-1.32), with similar associations for myocardial infarction and cardiovascular death. Restricted cubic spline analyses revealed a J-shaped relationship between SBP-VIM and cardiovascular outcomes (all p<0.05). The prognostic value of SBP-VIM was comparable to mean SBP. These findings were consistent across alternative BPV metrics and intensified with extended follow-up. Conclusions: Visit-to-visit BPV was independently associated with the risk of MCEs, particularly myocardial infarction and cardiovascular death, with a J-shaped relationship indicating that both low and high BPV may be harmful. This association was independent of mean BP and comparable in prognostic value, underscoring the need to determine optimal BPV targets and explore potential BPV-modulating interventions.