HRV-Derived Autonomic Responses to Noradrenaline Differentiate ICU Mortality Outcomes

Mateusz Solinski1, Teemu Pukkila2, Aliisa Lönnrot3, Esa Rasanen2
1Computational Physics Laboratory, Tampere University, 2Tampere University, 3Faculty of Medicine and Health Technology, Tampere University, Arvo Ylpön katu 34, 33540 Tampere, Finland


Abstract

Critically ill patients in the Intensive Care Unit (ICU) often require noradrenaline (NE) to maintain hemodynamic stability in states of vasodilatory hypotension and circulatory failure. This study investigates whether autonomic response phenotypes, quantified via heart rate variability (HRV), are associated with ICU mortality. We extracted RR intervals from ECG signals recorded by ICU monitors of patients receiving NE therapy and combined them with clinical data from the MIMIC-III database. HRV was analysed across four temporal stages relative to infusion: pre-infusion (-70 to -10 min), early infusion (+10 to +70 min), late infusion (-70 to -10 min before cessation), and post-infusion (+10 to +70 min). We stratified patients into two groups: (1) those who died in the ICU after the post-infusion stage, and (2) those discharged after this stage. We included 243 patients (24% mortality) in the analysis. Significant differences in HRV features between groups were observed at all stages except pre-infusion. The largest divergences occurred during late and post-infusion periods. Survivors exhibited a significantly higher level of HRV features related to complexity, particularly in detrended fluctuation analysis metrics such as DFA alpha2 (p<0.01 for late and post-infusion stages) and scale-dependent DFA exponents across a wide range of scales (16–100 for late infusion, and 20-50 for post-infusion). These findings suggest that late-stage and recovery-phase autonomic dynamics reflect physiological resilience and the capacity to restore coordinated control following hemodynamic stress. HRV-derived response phenotypes to NE may provide clinically relevant prognostic markers and support more personalised management strategies in critically ill patients.