Presented at the Neonatal Society 2009 Autumn Meeting.
Thoresen M1, Hellström-Westas L2, Liu X1, de Vries LS3
1 University of Bristol and University Hospital Bristol, UK
2 Uppsala University, Uppsala, Sweden
3 University Medical Centre, Utrecht, Netherland
Introduction: Amplitude integrated electroencephalography (aEEG) at <6h of age is the best single predictor of outcome in term infants at normothermia (NT) with perinatal asphyxia.
Objectives: This study investigates the validity of the predictive value of the initial aEEG background activity in asphyxiated infants receiving intensive care both at NT and HT.
Patients and Methods: A single centre recruited 74 infants to NT or HT within ongoing studies and aEEG was recorded for the whole treatment period and analysed by LHW and LSV comparing the pattern recognition and voltage method. Outcome was assessed using the Bayley-II at 18-months.
Results: Death and disability occurred in 40% HT and 65% NT infants (p=0.04). A “poor” aEEG pattern at 3-6h correctly predicted poor outcome in 87% of NT infants with odds ratio (OR) 10.7 (95%CI 1.9-59.6) p=0.007, while OR for HT was 1.7 (95%CI 0.56-3.78) p=0.18. Lack of normalisation of the background trace by 24h for the NT group and by 48h for the HT group was strongly associated with a poor outcome. Sleep wake cycling (SWC) did not develop in 60% of NT and 35% of HT and its absence predicted poor outcome. Combining time to regain a normal trace and time to SWC predicted outcome correctly in >95% of infants regardless of treatment. Correct prediction was significantly better using pattern recognition than voltage criteria for both NT (p=0.03) and HT (p=0.05) in the 58% of infants with moderately abnormal voltage traces.
Conclusion: The pattern recognition method is superior to the voltage method for correct prediction in infants with a moderately abnormal trace. Both methods predict outcome well for those with normal voltage/good trace or severely abnormal voltage/poor trace. Lack of normalisation of the background trace by 24h for the NT group and by 48h for the HT group was strongly associated with a poor outcome. Hypothermia treatment has changed the “cut-off” values for prediction of outcome in infants undergoing hypothermia using aEEG voltage patterns and time at onset of SWC.