Enhancing early detection of acute brain injury using heart rate variability in patients on venoarterial extracorporeal membrane oxygenation

Abstract

Introduction

Early detection of acute brain injury (ABI) in patients on venoarterial extracorporeal membrane oxygenation (VA ECMO) is challenging. We aimed to investigate if heart rate variability (HRV) in the first 24 h of ECMO cannulation correlated with ABI in VA ECMO patients.

Methods

We retrospectively analyzed adults (≥18 years) on VA ECMO support in our center (June 2016-April 2022). ABI included ischemic stroke, intracranial hemorrhage (ICH), and hypoxic ischemic brain injury (HIBI) as diagnosed on neuroimaging performed within 48 to 72 h of ECMO cannulation. The primary outcome was ABI, and secondary outcomes were mortality and neurological outcomes at discharge assessed by modified Rankin Score (mRS), poor (mRS>3) and good (mRS≤3). Continuous electrocardiographic recordings during ECMO support were used to compute sample entropy (HRV-SE), a measure of HRV. Logistic regression models were used to evaluate the association between early HRV-SE and primary and secondary outcomes.

Results

Among 139 patients (median age 56 years; mean ECMO duration 6 days; 66% male), 61 (44%) developed ABI, most commonly ischemic stroke (54%) followed by ICH (30%). HRV-SE in the first 24 h was lower in ABI versus no-ABI (0.29 ± 0.22 vs 0.41 ± 0.24; p = 0.009) and lower with central versus peripheral cannulation (0.31 ± 0.25 vs 0.41 ± 0.22; p = 0.02). Furthermore, ABI occurred slightly more frequently in patients with central compared to peripheral cannulation (23% vs 20%). Reduced HRV-SE within the first 24 h demonstrated strong diagnostic performance for ABI (90% accuracy, 100% sensitivity, 82% specificity; AUC = 0.86) and was also associated with increased mortality (0.30 ± 0.23 vs 0.42 ± 0.23; p = 0.002 and poor neurological outcomes (0.35 ± 0.25 vs 0.44 ± 0.16; p = 0.05).

Conclusions

Lower HRV-SE within 24 h of VA ECMO cannulation was associated with ABI, mortality, and poor neurological outcomes at discharge, supporting HRV-SE as a potential adjunctive early bedside screening marker that warrants prospective validation.