To evaluate alterations in blood pressure as compared to a baseline as an indicator of ischemic events in the setting of inpatient code strokes.

Inpatient strokes account for 6.5-15% of all stroke presentations. Early identification, diagnosis, and treatment of stroke in inpatients should be optimal given the close monitoring provided in this setting; however, in prior studies between 26% to 67% of patients in whom an inpatient code stroke was called suffered from actual stroke compared to 51% of patients presenting to the ED with suspicion of stroke. In the ED setting, elevations in systolic blood pressure [SBP] at presentation have been associated with cerebral ischemia in patients presenting with suspicion of stroke. Relative elevations in BP have yet to be evaluated and are likely more sensitive.

This is a retrospective study evaluating inpatients of Rhode Island Hospital in whom a code stroke was called over a period of 24months. Systolic, diastolic [DBP] and mean arterial blood pressures were recorded within the first 24hrs, 48hrs, and 72hrs prior to activation of the code stroke.  These data were analyzed using multivariate fractional logistic regression models to evaluate changes in blood pressures as predictive of ischemic events.

A model using BPs at periods of 24, 48 and 72hours referenced to measures at times of the code were weakly predictive of ischemia with an area under the ROC curve of 0.77 (95% CI 0.71-0.91).

Increase in BP can be used to discriminate between ischemic and non-ischemic events with better than chance accuracy. Using SBP and DBP together are more predictive particularly when using a baseline comprised of mean measurements up to 96hrs. Results are not as pronounced as would be expected which may be a function of confounding comorbid active medical issues and small sample size.