Visit-to-visit systolic blood pressure variability and risk of severe hypertensive episodes in a multicenter cohort

BACKGROUND: Visit-to-visit blood pressure variability has been associated with cardiovascular outcomes; however, its role in predicting severe blood pressure (BP) elevation requiring emergency department care remains uncertain. We investigated whether long-term systolic blood pressure variability independently predicts severe BP elevation and evaluated externally validated prediction models. METHODS: We analyzed electronic health records from three tertiary hospitals. After excluding patients with cardiovascular events during a 2-year monitoring window, systolic blood pressure variability, quantified by standard deviation (SD), was assessed for incident severe BP elevation (systolic BP ≥ 180 mmHg or diastolic BP ≥ 120 mmHg at emergency department visits) during 5-year follow-up using landmark Cox models. Risk prediction models incorporating systolic blood pressure variability were developed and externally validated across institutions. RESULTS: Among 20,090 patients, 2,128 (10.6%) developed severe BP elevation. Each 1-SD increase in systolic blood pressure variability was independently associated with severe BP elevation (adjusted hazard ratio 1.21, 95% confidence interval 1.16-1.26; P < .001), irrespective of mean blood pressure and clinical covariates. The association was consistent across major subgroups. Prediction model incorporating blood pressure variability achieved good discrimination in external validation cohorts (area under the curve 0.763 [Guro] and 0.765 [Anam]). CONCLUSIONS: Visit-to-visit blood pressure variability independently predicts severe hypertensive destabilization beyond mean blood pressure. These findings suggest that incorporating visit-to-visit blood pressure variability into clinical assessment may improve identification of patients at risk for severe hypertensive episodes.