Prognosis

Prognosis is essentially a medical forecast of the baby's health. It plays an important role in developing a long term management strategy for the baby and provides valuable information for parents.

We are investigating the use of EEG and other physiological signals for predicting the future neurological  health of the baby. In particular we are looking at babies with HIE (hypoxic-ischaemic encephalopathy). This is an injury that is caused by a lack of oxygen in the baby’s brain. The initial injury occurs around the time of  birth. However, after a relatively short period of primary recovery, a rapid increase in cerebral blood flow might be responsible for a period of secondary damage to the newborn brain. It is, therefore, important to monitor the EEG, and any other physiological signals, through all these stages when attempting to make an accurate prognosis.

The challenge of determining prognosis from the EEG is shown in Figure 1.

Figure 1: Determining prognosis from EEG monitoring.

We have looked at how certain EEG features in babies with HIE correlate with the neurodevelopmental outcome at 2 years of age. Our classification system grades the EEG as normal, mildly abnormal, moderately abnormal, severely abnormal or inactive using a rule based system. This results in definitions of each class as;

1. Normal EEG: Continuous background pattern with normal physiological features such as anterior slow waves.
2. Normal/mild abnormalities: Continuous background pattern , including slightly abnormal activity, e.g.  mild asymmetries, mild voltage depression, poorly defined sleep-wake cycles.
3. Moderate abnormalities: Discontinuous activity with IBI of <10s, no clear sleep-wake cycles, clear asymmetry or asynchrony.
4. Severe abnormalities: Discontinuous activity with  IBI of 10-60s, severe attenuation of background patterns, no sleep-wake cycles.
5. Inactive EEG: Background activity of < 10 mV or severe discontinuity with IBI of >60s.

 (IBI = interburst interval, GA = gestational age)

An example of the discontinuous activity mentioned in the EEG classification is shown in Figure 2. This discontinuous EEG pattern is typically termed burst-suppression.

Figure 2: EEG showing a discontinuous or burst suppression pattern with an interburst interval of approximately 4 seconds.

Sleep wake cycles are characteristic EEG patterns that correspond to active sleep, quiet sleep, intermediate sleep and wakefulness in the infant. Asymmetry and asynchrony are manifested on the EEG as relative latencies (time lags) and attenuations between the hemispheres of the brain.

We studied a group of 44 infants and found that a simple EEG classification system is a reliable predictor of normal outcome, particularly when the EEG is measured at 6 hours after birth (before the secondary injury occurs), that is babies with a normal EEG in the first 6 hours always have a normal outcome. Infants who displayed continued major and severe abnormality on the EEG tended to have a poorer outcome (such as movement disorders and developmental delay). This is shown in Figure 3.

Figure 3: The change of EEG classification over time for infants with normal and poor outcome.

We are also developing new classifications of EEG that are capable of monitoring long term changes in the EEG for prognosis.

Further information regarding classification of EEG for prognosis can be found in,

1. R.M. Pressler, G.B. Boylan, M. Morton, C.D. Binnie, J.M. Rennie, "Early serial EEG in hypoxic ischaemic encephalopathy", Clinical Neurophysiology, vol. 112, pp. 31-37, 2001.
2. D.M. Murray, C.A. Ryan, G.B. Boylan, A.P. Fitzgerald and S. Connolly, "Prediction of seizures in asphyxiated neonates: correlation with continuous video-electroencephalographic monitoring", Pediatrics, vol. 118, pp. 41-46 2006.
3. D.M. Murray, G.B. Boylan, C.A. Ryan, and S. Connolly, "Early EEG findings in hypoxic-ischaemic encephalopathy predict outcome at 2 years", Pediatrics, vol. 124, pp. e449-e467, September 2009.