Dear Editor

Rennie highlighted the problems of using the Lectromed cerebral function monitor (CFM) for identifying seizures, when compared with a video EEG system. However, few if any neonatal units have the facilities and expertise for performing video EEG monitoring and a single standard EEG is not helpful for assessing seizures.

Previous studies have confirmed the value of the amplitude integrated EEG (aEEG) recorded with the Lectromed CFM for assessing the severity of neonatal encephalopathy and predicting neurological outcome. This has led to the development of digital CFM devices that include the ability to display the raw EEG as well as the aEEG, making it possible to confirm seizure discharges and exclude artefacts. These digital CFM devices are user-friendly enough for nurses and medical staff to quickly initiate display and recording in a busy neonatal unit. Because the raw EEG is recorded, the presence or absence of electrical seizure activity can be reviewed at the time or later, depending on the availability of staff able to interpret neonatal EEG seizure activity. These devices are likely to have a useful role in modern neonatal intensive care for monitoring neonatal encephalopathy, quantifying seizure episodes and assessing the response to treatment.

Dear Editor

We read, with interest, the paper by Rennie et al.[1] comparing seizure detection from simultansous EEG and cerebral function monitor (CFM) recordings. Their group is known to have a specific interest in neonatal neurophysiology and they have an enviable video-EEG set-up. Data from their experience with this technique is available in previous publications.[2,3]

Sadly, in most neonatal intensive care units, these facilities are not available. There are still many units where getting an EEG is difficult and getting one at night or the weekend would be impossible. The introduction of CFM has allowed, for the first time, electrophysiologial monitoring on the NICU.

In this study Rennie and colleagues asked four neonatologists with 3- 5 hours of training in CFM interpretation to assess CFM (Lectromed) traces from 19 infants with EEG seizures and 21 infants without EEG seizures. The EEG and CFM recordings were made simultaneously. The cohort was a mixture of preterm infants with a gestational age as low as 24 weeks and fullterm infants. Five infants had focal seizures only, 6 a mixture of focal and generalised seizures and 8 either generalised seizures or multifocal seizures. The trainees were able to detect only half of the seizures reasons being that the seizures were short lived, localised or of low amplitude.

We have been been using the CFM for more than a decade and were disappointed to read these negative findings which has not been our experience.We feel that there are several points which need to be addressed.

The amount of teaching for those analysing the CFM traces was very limited especially considering that the trainees were asked to analyse traces from preterm and full term infants and traces obtained at different speeds. Analysis of the CFM is based on pattern recognition which takes considerable time to learn especially for the wide range of gestational ages of the infants in this study. Using four different speeds does not help this process of pattern recognition. Running the CFM at a different speeds is not most peoples clinical practice and is confusing for those with a lot of CFM experience.

Most of the papers dealing with the CFM have dealt with fullterm infants [4,5] and only 7 of the infants in this study were term. No information is given as to whether seizure recognition from the CFM was easier in this group. Although we agree that the preterm population is a very important population in rapidly advancing neonatal intensive care, it has been stated in the past that CFM may not be the best modality and other techniques, like spectral edge frequency are under investigation and the first published data are promising.[6]

Addtionally, prognostication is more reliably done from the background activity than the detection of brief seizures and missing a few brief seizures is probably not of great clinical importance.

In our own EEG-CFM comparison paper [7] we have also missed some seizures in 2 of 10 infants with electrical EEG discharges. The reasons were similar to the ones stated in this paper: low amplitude seizures, very short discharges and focal discharges. In the present study 11 of the 19 infants with seizures had only or partly focal discharges, which appears to be an overrepresentation of what we normally see and perhaps reflects the wide range of gestational age of the infants

As we have stated many times in the past, the CFM is a monitoring device and not an EEG. All our infants also therefore have an EEG during the time they have CFM monitoring. The new digital CFM devices also have the advantage of the simultaneous raw EEG and may therefore overcome some of the problems mentioned above.

Most important and not mentioned in this study, is that an EEG, if available at short notice, will usually be obtained for a period of only 30-45 minutes. During this time the information obtained is detailed and we do not miss any focal or short discharges. However, what we do miss is all the subclinical abnormality occurring during the remaining 23 hours of the day. As they have shown themselves,[3] more than half of the electrical discharges are subclinical and will go unnoticed. At present we do not yet know whether these subclinical seizures will adversely affect the brain and we are performing a multicentre randomised study at present in the Netherlands and Belgium to answer this important question. From our data the majority of these subclinical episodes are detected on the continuous recordings possible using the CFM and therefore overall the CFM is giving a far superior rate of seizure detection.

We hope that neonatologists reading the paper by Rennie et al. [1] will not be put off using the CFM. We hope they will still see that the advantages of long term continuous information will in most cases outweigh detailed information obtained over a short period of time. We are sure that they see that proper teaching is essential,[8] just as we are sure that they would not like to board a plane, where the pilot only had 3 -5 hours of flying experience.

References

1. Rennie JM, Chorley G, Boylan GB et al. Non-expert use of the cerebral function monitor for neonatal seizure detection. Arch Dis Child Fetal Neonatal Ed 2004; 89:F37-F40.

2. Boylan GB, Pressler RM, Rennie JM et al. Outcome of electroclinical, electrographic and clinical seizures in the newborn infant. Dev Med Child Neurol 1999; 41:819-25.

3. Boylan GB, Rennie JM, Pressler RM et al. Phenobarbitone, neonatal seizures and video-EEG. Arch Dis Child Fetal Neonatal Ed 2002; 86: F165-70

4 Hellström-Westas L, Rosen I, Svenningsen N.W. Predictive value of early continuous amplitude integrated EEG recording on outcome after birth asphyxia in full term infants. Arch Dis Child Fetal Neonatal Ed, 1995; 72: F34-F38.

5 Toet MC, Hellström-Westas L, Groenendaal F et al. Amplitude Integrated EEG at 3 and 6 hours after birth in fullterm neonates with hypoxic ischaemic encephalopathy. Arch Dis Child Fetal Neonatal Ed 1999 81:F19-F23.

6 Inder TE, Buckland L, Williams CE et al. Lowered electroencephalographic spectral edge frequency predicts the presence of cerebral white matter injury in premature infants. Pediatrics 2003; 111: 27-33

7 Toet MC, van der Meij W, de Vries LS et al. Comparison between simultaneously recorded amplitude integrated EEG (cerebral function monitor (CFM)) and standard EEG in neonates. Pediatrics 2002; 109:772-779

8 Hellstrom-Westas L, de Vries LS, Rosen I. An atlas of amplitude- integrated EEGs in the newborn. Parthenon Publishing London; 2003