Automated Detection And Prediction of Preterm and Term newborn Seizure

• Research type

  Research Study

• Full title

  Automated Detection and Prediction of Preterm and Term newborn Seizures (ADAPTS ) study

• IRAS ID

  284449

• Contact name

  Divyen Shah

• Contact email

  divyen.shah@nhs.net

• Sponsor organisation

  Barts Health NHS Trust,

• Duration of Study in the UK

  3 years, 0 months, 1 days

• Research summary

  Background

  Seizures, also known as fits, affect about 1500 newborns annually in the UK and may indicate brain injury. Most clinicians believe seizures in the newborn worsen underlying brain injury and constitute a neurologic emergency. Seizures are detected using EEG (Electroencephalography) monitoring in which small electrodes are placed on the baby’s scalp to measure tiny voltages and is therefore sensitive to electrical interference and baby movement. Seizures observed visually are often subtle in newborns and difficult to detect and their relationship with EEG seizures is poor with many EEG seizures having no clinical manifestation. Also, even among experts, agreement for what constitutes EEG seizures is variable. Hence seizure detection and management remain a major problem in the care of sick babies.

  
  Aims

  1)Investigate if there are concomitant changes in physiological measurements during neonatal EEG seizures from simultaneous observational recordings.
  2) Describe the antenatal, intrapartum and postnatal factors associated with seizures in newborns, and describe outcomes of seizures in newborn period in terms of changes seen on MRI brain scan and later neurodevelopmental assessment.

  Workplan
  This study will be conducted at a single tertiary neonatal centre for a period of three years. We aim to recruit 45 babies with seizures and 45 babies without seizures but with a compatible clinical presentation.
  Routinely collected physiological parameters like heart rate, blood pressure, oxygen levels and breathing rates will be recorded along with video EEG. Using machine learning, these datasets will be analysed to devise a novel seizure detection algorithm. The algorithm will be further adapted to develop an integrated approach to seizure management by incorporating antenatal and intrapartum data and comparing with outcome measures including neuro-imaging and neuro-developmental outcome.

  Expected outcomes – Accurate seizure detection will enable clinicians to initiate treatment so that the right patient group receives timely anticonvulsant therapy.
  

• REC name

  London - Fulham Research Ethics Committee

• REC reference

  20/PR/0969

• Date of REC Opinion

  12 Feb 2021

• REC opinion

  Further Information Favourable Opinion