BiHiVE2 Study. The Investigation and Validation of Predictive Biomarkers in Hypoxic-ischaemic Encephalopathy.

Background and current knowledge in the area:

Perinatal asphyxia is one of the commonest causes of neonatal death and long term disability, occurring in 20 per 1000 live births. Of these, approximately 2-3 per 1000 will go on to develop hypoxic-ischaemic encephalopathy (HIE) (1). In Ireland and the UK HIE is the third commonest cause of neonatal mortality, accounting for 9% of all deaths, and 21% of term deaths, while globally it is estimated to cause over 1 million neonatal deaths each year. For the survivors there is significant secondary morbidity associated, including cerebral palsy (15%), severe cognitive delay (11%), seizure disorders (8%), sensori-neural deafness (7%), and visual loss (3%)(2).

HIE is an evolving process, which can be simplified into two insults. The initial insult is significant cerebral hypoxia and/or ischaemia which is followed by energy depletion, from anaerobic metabolism leading to the accumulation of lactate, and loss of cellular integrity. This sets in motion a cascade of events with the resultant formation of free radicals, proteases and caspases leading to cell damage and apoptosis, the secondary insult(3). The secondary energy failure has been shown to occur between 6-48 hours after the initial insult.

Until recently no therapeutic intervention which could reduce the associated morbidity or mortality existed. However the recent publications of several international trials, and their subsequent meta-analysis has shown that early induced hypothermia is beneficial in HIE, improving survival and reducing neurological disability (4-5). Induced hypothermia has now become a standard of care in moderate/severe HIE. However to be effective it must be commenced within 6 hours of delivery, before the secondary energy failure occurs. In this narrow window of time the population who would benefit from treatment must be identified, resuscitated, stabilised and cooled.

Unfortunately the ability to recognise and diagnose those who would benefit from therapy is not always possible within this time period. We have shown that the current standard methods to identify those at risk of HIE such as acid-base balance, lactate and Apgar score and initial neurological assessment are unreliable (6). Neonatal electroencephalography (EEG) is useful for the grading of encephalopathy and is an excellent predictor of long term outcome. In particular its evolution over the first 24 hours following delivery strongly correlates with outcome (7). However it requires a highly specialised skill set that is rarely available to the neonatologist in the acute setting. Amplitude integrated EEG is more widely used by clinicians, but its interpretation is user dependent, and the majority of neonatologists report they are not confident in their own interpretive ability (8). The need for early and reliable prediction of outcome in HIE has never been greater. A quick, cheap, reproducible, non-user dependent method for quantifying severity of HIE and likely prognosis is an urgent need. Many biochemical and haematological markers have been proposed in perinatal asphyxia, but none have led to clinical advances, or commercially viable bedside tests (9).

Over the last 2 years, here in the Cork University Maternity Hospital, through the BIHIVE study (Biomarkers in Hypoxic-Ischaemic Encephalopathy study) we have identified potential biochemical markers in umbilical cord blood which can accurately identify those infants who will progress to moderate/severe encephalopathy (10,11). The BIHIVE study has adopted a multifaceted approach to identifying blood-borne biomarkers in HIE through 1) targeted proteomic profiling using Luminex technology 2) miRNA profiling 3) comprehensive mapping of the blood metabolome using both Mass Spectrometry and Nuclear Magnetic Resonance techniques.

Our aim is to recruit a cohort of infants with clearly defined perinatal asphyxia and HIE, classified using both repeated neurological assessment and multichannel EEG. At the time of birth, umbilical cord blood will be drawn, processed and banked, using strict standard operating procedures (SOPs) which we have developed for this purpose. These procedures will ensure that all specimens will be suitable for later proteomic, metabolomic, transcriptomic and genomic analysis. Using a systems biology approach we have studied multiple proposed markers in the stored umbilical cord blood from our discovery cohort, initially focusing on the proteome and metabolome. Through this work we have already identified a number of biomarkers present in umbilical cord blood with the potential to predict the development of moderate/severe HIE. Our early data is extremely promising. Our initial metabolomic analysis has shown that we can differentiate between HIE and controls with an area under the curve (AUROC) of 0.92. Further analysis will be required to identify and quantify the exact metabolite peaks which produce this differentiation. We now wish to validate these findings in a further cohort of infants with perinatal asphyxia and HIE, and to validate our ability to predict longer term neurological outcome. Our pilot work examining microRNA changes in the cord blood of infants with moderate/severe HIE is also very encouraging with significant upregulation of a multiple studied miRNAs in both asphyxia and HIE. The most promising miRNAs are now being quantified using polymerase chain reaction (PCR) amplication and will be validated in the discovery cohort.

Future cohort-BIHIVE2: To develop these markers further to the point where we have validated a marker with >90% sensitivity and specificity for moderate/severe HIE we will need to recruit a further cohort of carefully recruited infants with high quality biobanking at birth and detailed clinical phenotyping. This is the main work of this current application. We will require a further cohort of 300 infants with perinatal asphyxia, of whom 60 will develop HIE and 30 will develop moderate to severe HIE. We will also need to recruit 300 contemporary control infants. To recruit this number of infants with perinatal asphyxia we will need to study 30,000 live births. We have therefore expanded our study centres to include Karolinska University Hospital, Stockholm, a large state of the art maternity hospital with 10,000 live births per annum.

Further validation will allow us to develop an algorithm of clinical and biochemical markers which will provide a sensitive and specific predictive test for moderate/severe HIE. Neurodevelopmental follow up of both cohorts will allow us to assess our ability to predict neurodevelopmental and behavioural outcome at 2 years.

Lastly we will examine the exact timing of biomarker alteration using a timed animal model of hypoxic-ischaemic injury (HI). In clinical studies hypoxic-ischaemic injury is unpredictable and its timing is often impossible to estimate. Dr Tracey Bjorkmann in Queen's University, Brisbane has developed a piglet model of HI which produces HI induces seizures and a level of injury equating to moderate to severe HIE (12). She has agreed to collaborate with our programme and will provide piglet samples of serum taken at 0,1,2 and 12 hours following HI. These samples will be invaluable in examining the progression of biomarker derangement and will help us to time neonatal brain injury after delivery.