# Emergency Department Vital Sign Variability Is Associated with Hematoma Progression in Spontaneous Intracerebral Hemorrhage

## TL;DR

Variability in vital signs in the emergency department is linked to increased brain bleeding in stroke patients.

## Contribution

Identifies emergency department vital sign variability as a predictor of hematoma progression in spontaneous intracerebral hemorrhage.

## Key findings

- 41 out of 142 patients (29%) experienced hematoma progression within 24 hours.
- Machine learning identified blood pressure and heart rate variability as impactful predictors of hematoma progression.
- Algorithms achieved 92.59% accuracy in predicting hematoma progression using vital sign variability.

## Abstract

Background/Objectives: Spontaneous intraparenchymal hemorrhage (sIPH) accounts for a significant proportion of strokes and is associated with an estimated 30-day mortality between 35 and 52%. Subsequent hematoma progression (HP) occurs in up to 30% of patients and is associated with blood pressure variability, increasing poor outcomes. This study evaluates systolic blood pressure and heart rate variability in the emergency department (ED) and HP in the first 24 h of admission. Methods: This retrospective study analyzed patients with sIPH presenting to the ED and transferred to a resuscitation unit between 2017 and 2020. Outcomes included the occurrence of HP. Variables included blood pressure variability as measured by the standard deviation in systolic blood pressure (SBP-SD), successive variation of systolic blood pressure (SBP-SV), standard deviation of heart rate (HR-SD), and successive variation of heart rate (HR-SV). Bivariate analysis and machine learning algorithms were used to identify ED predictors for HP. Results: Of the 142 records analyzed, 41 (29%) patients experienced HP. The medians [interquartile (IQR)] for baseline characteristics were similar between groups. In the group with no HP (control), the median [IQR] for SBP-SD was 17.6 [11–26] compared with 20.5 [13.9–26.1, p = 0.25]. The median [IQR] for standard deviation in SBP-SV was 18 [11.4–25.4] for the control group and 19.8 [15.2–27.3, p = 0.19] for the HP group. While bivariate analysis did not show statistical difference for SBP-SD, SBP-SV, HR-SD, or HR-SV, machine learning algorithms identified SBP-SD, HR-SD, and HR-SV as clinically impactful on HP with good accuracy (92.59% and 79.31%). Conclusions: This study suggests that there are factors in hyperacute hemodynamic management in the ED associated with HP among patients with sIPH.