Background:
Recurrent stroke has higher morbidity and mortality than incident stroke. We evaluated hemodynamic risk factors for multiple recurrent strokes.
Methods:
We included patients in the SPS3 trial. The primary predictor was the top tertile, compared to the bottom tertile, of the mean systolic blood pressure (mSBP) and blood pressure variability represented as standard deviation (sdSBP) using blood pressures from day 30 of the trial to the end of follow-up. We excluded blood pressures from the first 30 days to reduce confounding from the trial’s intervention. We fit a logistic regression model to ≥2 recurrent strokes from day 30 to the end of follow-up and, to accurately analyze the multiple failure-time data, we ordered the multiple failure events to the Prentice, Williams and Peterson extension of the Cox proportional-hazards model.
Results:
We included 2,882 patients, of which 223 had a recurrent stroke and 41/223 had ≥2 recurrent strokes for a total of 272 strokes. The mean (SD) number of blood pressure readings was 78.0 (37.4). The etiology of the 272 strokes was 161 (59.2%) lacunar, 22 (8.1%) intracranial atherosclerosis, 10 (3.7%) extracranial atherosclerosis, 24 (8.8%) cardioembolic, and 55 (20.2%) cryptogenic or other. In both unadjusted and adjusted logistic regression models and PWP Cox models, the top tertile of sdSBP was consistently predictive of multiple recurrent strokes, while mSBP was not (Tables 1/2).
Conclusions:
We found that in patients with an index lacunar stroke, higher SBP variability, but not mean SBP, was predictive of multiple recurrent strokes of varying mechanisms.
