Background and purpose: This study aimed to determine whether variability of day-by-day blood pressure (BP) during the subacute stage of acute ischemic stroke is predictive of long-term stroke recurrence.
Methods: We analyzed 7665 patients (mean±SD age: 72.9±13.1 years; women: 42.4%) hospitalized for first-ever ischemic stroke in 7 stroke centers in Fukuoka, Japan, from June 2007 to November 2018. BP was measured daily during the subacute stage (4-10 days after onset). Its mean and coefficient of variation (CV) values were calculated and divided into 4 groups according to the quartiles of these BP parameters. Patients were prospectively followed up for recurrent stroke or all-cause death. The cumulative event rate was calculated with the Kaplan-Meier method. We estimated the hazard ratios and 95% confidence intervals of the events of interest after adjusting for potential confounders and mean BP values using Cox proportional hazards models. The Fine-Gray model was also used to account for the competing risk of death.
Results: With a mean (±SD) follow-up duration of 3.9±3.2 years, the rates of recurrent stroke and all-cause death were 3.9 and 9.9 per 100 patient-years, respectively. The cumulative event rates of recurrent stroke and all-cause death increased with increasing CVs of systolic BP and diastolic BP. The systolic BP CV was significantly associated with an increased risk of recurrent stroke after adjusting for multiple confounders and mean BP (hazard ratio [95% CI] for fourth quartile versus first quartile, 1.26 [1.05-1.50]); the risk of recurrent stroke also increased with an increasing systolic BP CV for nonfatal strokes (1.26 [1.05-1.51]) and when death was regarded as a competing risk (1.21 [1.02-1.45]). Similar associations were observed for the diastolic BP CV.
Conclusions: Day-by-day variability of BP during the subacute stage of acute ischemic stroke was associated with an increased long-term risk of recurrent stroke.
Keywords: Japan; blood pressure; cardiovascular diseases; ischemic stroke; recurrence.