Interaction of Mean Arterial Pressure and Blood Pressure Variability in Critically Ill Brain Injured Patients

Introduction

Increased blood pressure variability (BPV) is associated with worse outcome in critically ill patients.^1–3 Because brain injury may impair cerebral autoregulation and predispose to the harmful effects of BPV,⁴ we explored the interaction of BPV and mean arterial pressure (MAP) in brain-injured versus other critically ill patients.

Methods

We conducted a cross-sectional analysis of deidentified data from MIMIC-III, a single-center study of adult patients in a critical care unit at Beth Israel Deaconess Medical Center from 2001–2012 that is publicly available at https://physionet.org/content/mimiciii/1.4/. We included patients with a primary discharge diagnosis of brain injury, defined as acute ischemic stroke (AIS), intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), or traumatic brain injury (TBI), and a control group with acute myocardial infarction (AMI). All patients had at least ten blood pressures in the 72 hours after admission. With logistic regression and marginal effects, we examined the impact of BPV (standard deviation of systolic blood pressure), mean MAP, and their interaction on the primary outcome of in-hospital death.

Results

We included 5,546 patients, of which 2,749 had brain injury (AIS n=989, ICH n=959, SAH n=504, TBI n=297) and 2,797 had AMI, with a mean ± SD age of 68.0±15.2 years, 57.4% male, 68.8% non-Hispanic white, and 15.6% died in-hospital. There were 650,005 BP measurements (399,838 non-invasive and 255,167 invasive, with a median (IQR) of 50 (30–72) measurements per patient, and 53.4% of patients had all non-invasive measurements, 20.9% had all invasive, and 25.7% had a mix of invasive and non-invasive).” Histograms of BPV and MAP after stratification by the primary outcome are in Figure 1.

Figure 1.

Histograms of BPV (A) and MAP (B) after stratification by in-hospital death in all patients. Reference line at 15 mm Hg for BPV and 80 mm Hg for MAP. Adjusted predicted probability of in-hospital death, shown for the lowest tertile (blue) versus highest (red) of BPV (C) and MAP (D), and for the interaction between BPV and MAP in brain-injured (E) versus control AMI (F) patients.

The adjusted predicted probability of in-hospital death for the lowest compared to the highest tertile of BPV and MAP is panels C and D of Figure 1. While a significant association is seen for higher BPV across all diagnoses, for MAP it is significant for ICH and AMI only. The interaction between BPV and MAP in all brain-injured patients compared to AMI controls was significant (p=0.011). In Figure 1 panels E-F, both subgroups show the synergistic adverse effect of low MAP and high BPV, but in brain-injured patients the effect of BPV is more important than MAP.

Conclusion

Consistent with prior research, we found an association between higher BPV and in-hospital death in brain-injured patients.^1–3 In addition, low MAP and high BPV has a synergistic effect that increases the probability of in-hospital death. In brain-injured patients, BPV was a more important modifier of death risk than MAP. Limitations of this analysis include its post-hoc retrospective nature, the absence of data on vasoactive and antihypertensive medications, and the lack of long-term functional outcomes after hospital discharge.

ABBREVIATIONS:

AIS

acute ischemic stroke

AMI

acute myocardial infarction

BPV

blood pressure variability

ICH

intracerebral hemorrhage

MAP

mean arterial pressure

TBI

traumatic brain injury

SAH

subarachnoid hemorrhage