Emergency Department Blood Pressure Management in Type B Aortic Dissection: An Analysis with Machine Learning

TL;DR

This study uses machine learning to analyze blood pressure and heart rate management in type B aortic dissection patients in the emergency department.

Contribution

The study introduces a machine learning approach to identify predictors of successful ED management of type B aortic dissection.

Findings

Higher triage SBP and intravenous labetalol treatment predicted unmet SBP goals at ED departure.

Pain management with higher morphine equivalents and faster CT scans predicted met HR goals.

Many patients did not meet AHA guidelines for hemodynamic control before transfer.

Abstract

Acute aortic dissections (AAD) have a high morbidity and mortality rate. Treatment for type B aortic dissection includes strict systolic blood pressure (SBP) and heart rate (HR) control per the American Heart Association (AHA) guidelines. However, predictors of successful emergency department (ED) management of SBP have not been well studied.

We retrospectively analyzed the records of adult patients presenting to any regional ED with type B AAD between 2017–2020 with initial SBP >120 mmHg and HR >60 beats per minute (bpm) and were subsequently transferred to our quaternary center. Primary outcome was SBP <120 mmHg based on both the 2010 and 2022 AHA guidelines and HR <60 bpm (based on the 2010 guideline), or HR <80 (2022 guideline). We used random forest (RF) algorithms, a machine-learning tool that uses clusters of decision trees to predict a categorical outcome, to identify predictors of achieving HR and SBP goals prior to ED departure, defined as the time point at which patients left the referring ED to come to our institution.

The analysis included 134 patients. At the time of ED departure, 26 (19%) had SBP <120 mmHg, 96 (67%) received anti-impulse therapy, and 40 (28%) received beta-blocker or vasodilator infusions specifically. The RF algorithm identified higher triage SBP and treatment with intravenous labetalol as the top predictors for SBP >120 mmHg at ED departure, contrary to AHA guidelines. Pain management with higher total morphine equivalent unit, as well as shorter time to computed tomography as predictors for HR <60 bpm and <80 bpm, were in concert with AHA guidelines.

Many patients with type B AAD did not achieve hemodynamic parameters in line with 2010 or 2022 AHA guidelines while being in the ED prior to transferring to a quaternary care center for further evaluation and management. Patients with higher heart rate and systolic blood pressure on ED arrival were less likely to achieve goals at the time of departure from the referring EDs. Those receiving more pain medications prior to transfer were more likely to meet certain AHA goals.

Figures