MINERVA - Neuroprognostication using heart rate variability (v1.1)

  • Research type

    Research Study

  • Full title

    Machine Intelligence for NEuroprognostication using heart Rate Variability in out-of-hospital cardiac Arrest patients (MINERVA)

  • IRAS ID

    354160

  • Contact name

    Paige Smith

  • Contact email

    paige.smith16@nhs.net

  • Sponsor organisation

    Royal Papworth Hospital NHS Foundation Trust

  • Duration of Study in the UK

    1 years, 4 months, 25 days

  • Research summary

    This study explores whether heart rate variability (HRV) - how the time between heartbeats varies - can predict recovery from brain injury after an out-of-hospital cardiac arrest (OHCA). Patients who survive cardiac arrest experience reduced blood flow to the brain which can cause brain injury from a lack of oxygen. This is often catastrophic and irreversible, so life-sustaining treatment is withdrawn in ICU. Decisions about continuing treatment are often based on the likelihood of neurological recovery, but existing tests, such as brain scans or electrical brain activity monitoring may not be immediately available to some ICUs, delaying decisions around treatment.

    HRV is derived from continuous electrocardiogram (ECG) recordings, which are routinely collected in intensive care units (ICUs) to monitor heart activity. Recent research suggests HRV may provide early clues about brain injury and recovery, making it a promising and widely accessible tool to complement existing methods of predicting outcomes.

    The study will include adults admitted to Royal Papworth Hospital, a tertiary cardiothoracic centre, following OHCA. Patients will be eligible if they have sufficient ECG recordings and satisfy one of three outcomes: a good recovery with little to no brain injury, a protracted recovery with moderate brain injury, or no recovery, where the patient's treatment is withdrawn because tests show their brain injury is unsurvivable. Data will be collected for past, present and future patients from the earliest available historical ECG signal data to 31st December 2025. Participants will not need to undergo any new procedures or interventions, as the study uses data already collected during routine care.

    The study will run until April 2026 and aims to develop a machine learning model to analyse HRV and predict neurological outcomes. This research could improve decision-making for patients and families by providing earlier, more accessible insights neurological recovery, ultimately enhancing decisionmaking and reducing unnecessary treatments.

  • REC name

    Wales REC 1

  • REC reference

    25/WA/0130

  • Date of REC Opinion

    15 May 2025

  • REC opinion

    Further Information Favourable Opinion