Was NAP 5 ‘ NICE ’ enough; where next for depth of anaesthesia monitors?

I do not fear computers. I fear the lack of them —Isaac Asimov Since the publication of the fifth National Audit Project (NAP5) regarding accidental awareness during general anaesthesia (AAGA) 1, there have been a number of editorials and commentaries on the implication of the study's results for specific areas of practice, such as cardiac anaesthesia 2, obstetric anaesthesia 3, and even the administration of sedation 4. One of the authors (JCA) became interested in awareness and depth of anaesthesia monitoring after being involved in some low-level research as a senior registrar in the last century! He was excited when NAP5 was announced and eagerly volunteered to be a local coordinator. This enthusiasm was tempered somewhat by the methodology of the ensuing study. The NAP5 investigators reassured us that NAP5 would not make any strong statements about the incidence of awareness. Unfortunately, upon its release, the headline-grabbing news in the national media was the 1:19 000 ‘incidence’ 5, an order of magnitude lower than that previously reported 6, 7. The methodology and this interpretation of NAP5's results have been analysed and critiqued in some detail in previous editorials 8, 9. The NAP5 authors have attempted to clarify this figure by pointing out that it is merely the incidence of reported AAGA during the study period, rather than the total incidence 10 – which was previously accepted by the international anaesthetic community as being approximately 1:1000 7. This somewhat higher incidence is arrived at using ‘Brice’ methodology (post-anaesthesia interview) and the NAP5 authors, whilst acknowledging the differences between the two methodologies, have also pointed out potential problems in deriving an incidence of awareness solely from Brice-type questionnaires 10. Why does the incidence of AAGA matter? Firstly, when the initial flurry of excitement following the publication of NAP5 has died down, a newly perceived low incidence of AAGA may unintentionally make anaesthetists complacent regarding this complication, and they may labour under a false sense of security. Secondly, should an anaesthetist be unfortunate enough to have a complaint of awareness made against him/her, there may be a greater likelihood of assignment of blame to the individual anaesthetist, either by colleagues or by the patient, due to lack of recognition that AAGA remains relatively common. It is of note that NAP5 reported a low rate (2.8%) of the routine use of depth of anaesthesia monitors, despite guidance from the National Institute of Health and Care Excellence (NICE) that such devices should be considered in certain cases 11. In light of this new reported low incidence of AAGA with infrequent use of depth of anaesthesia monitors, it would seem unlikely that anaesthetic departments and hospital managers will now be looking to invest time and money increasing the utilisation of this new technology, given the ongoing financial constraints facing the NHS. Part of this is driven by the confusion as to how the information obtained from a so-called ‘depth monitor’ can be interpreted. Table 1 addresses some of the common misconceptions surrounding the output of processed EEG (pEEG) monitors. These correlate primarily with the probability of explicit recall following general anaesthesia rather than depth of anaesthesia, although pEEG output measurements do correlate with cerebral metabolic rate 15. The NICE report DG6 (perhaps in retrospect unhelpfully entitled Depth of Anaesthesia Monitors) 11 was hampered by the very reason for its existence, namely finance. The NICE Medical Technologies Advisory Committee could only recommend technology if its use added up in terms of benefits in health economics. On one side of the equation were the costs of using pEEG monitoring, and on the other side were the estimated savings of this technology. The main saving was thought to be avoidance of the psychological consequences of post-traumatic stress disorder, as well as smaller sums from minor savings in drug expenditure costs. The theatre efficiency savings that were identified (of up to 30 minutes in emergence and length of stay in the post-anaesthesia care unit) were deemed too small to be included, despite contrary arguments by the clinicians on the Specialist Committee that advised the NICE Advisory Committee. Consequently, even when using the higher reported incidence of AAGA at 1:1000, it was concluded that it was difficult to justify using pEEG monitors in anything other than total intravenous (TIVA)-based anaesthetics, since the financial balance relied on the assumption that these monitors would decrease the incidence of awareness. It should be noted, however, that the bispectral index (BIS) is the only pEEG monitor that has been shown to decrease the incidence of awareness 16, 17. The NICE report ultimately concluded that the recommendations could also be extended to include both E-Entropy and Narcotrend since they were ‘broadly similar’ to BIS monitoring. It is perhaps fortunate that this conclusion was made, since their considerably lower costs helped to balance the incremental cost-effectiveness ratio versus quality-adjusted life year (QALY) weighing scale 18. Furthermore, if the calculations had been made on the lower incidence of AAGA from the NAP5 report, it is unlikely that any pEEG monitor would have been recommended on financial grounds. The findings of NAP5 have, therefore, added weight to the recommendations of the earlier NICE report, by advising pEEG monitoring in patients who are undergoing TIVA with neuromuscular blockade. The analysis of causes of AAGA from NAP5 will rightly result in some long-term changes in the utilisation of some anaesthetic monitoring devices such as peripheral nerve stimulators. However, the impact of the perceived low rate of AAGA described by NAP5 may result in a reluctance to implement pEEG monitoring techniques, thus missing the other emerging potential benefits of the routine use of such devices: pEEG monitoring is not just about reducing AAGA. The section of the NAP5 report detailing the small number of awareness reports in which a pEEG monitor was used is of particular interest, but with only six such events it is difficult to apply any meaningful statistical analysis. The attempt to do so detracts from some otherwise excellent points made in interpreting the more relevant findings related to this monitoring modality. The six cases are briefly described and in half of them, problems with the way in which BIS was used or interpreted are elaborated upon. In one case, the BIS was not used at induction of anaesthesia when awareness occurred; in a second case, it was likely to have been removed before inadequate reversal of neuromuscular blockade; and in a third case, the BIS reading was above the recommended figure of 60. The remaining three cases are hard to interpret but lead to some well-constructed ideas about the way in which pEEG is used and some excellent recommendations for future research. Most of these are logical and centred upon the practicalities of pEEG monitoring such as using the monitoring at induction of anaesthesia and leaving it on until adequate reversal of neuromuscular blockade has been confirmed with a peripheral nerve stimulator. Other recommendations for research into cohorts at high risk of AAGA are also welcome. The NAP5 authors also state that the interpretation of the output of such monitoring is currently ‘an intangible art form’ and call for practical or pragmatic guidelines before such monitoring is widely adopted. This argument is somewhat of a tautology since the correct interpretation of the output from these monitors requires experience and it is only by their more widespread adoption that this experience and training can be obtained. At a personal level, we find it hard to understand the anaesthetic community's ambivalence to a monitor that can assess the blood flow and ongoing metabolic consumption of the very organ at which the anaesthetic is being targeted 15. The reticence to adopt ‘depth of anaesthesia’ monitoring may, in part, be due to a lack of understanding of what such monitors can tell us. Many studies (including NAP5) criticise BIS (the most validated pEEG monitor) since it does not truly predict or detect awareness (this is only really possible with the isolated forearm technique 19), thus overlooking other potentially valuable information on brain metabolism that it can provide. What does the future hold for pEEG monitoring? In light of the NAP5 results, it seems unlikely to us that the UK anaesthetic community will adopt the routine use of such monitors to help reduce AAGA, given the reported low incidence. There is, however, accumulating evidence that pEEG monitors may be of value at the opposite end of the spectrum, namely, the delivery of excessive amounts of anaesthetic agents. The recognition in observational studies that the so-called ‘triple low’ of mean arterial pressure < 75 mmHg, BIS < 45 and end-tidal (i.e. alveolar) concentration of inhalational anaesthetic agent < 0.7 MAC may be associated with an increase in postoperative mortality has increased interest in pEEG monitoring 20. It remains unclear as to whether this observation is purely a marker of increased risk or whether the use of pEEG monitors may be able to modify this risk. The BALANCED study (see http://balancedstudy.org.nz/) will examine one-year mortality in 6500 high-risk patients randomly assigned to either deep (BIS target 35) or light (BIS target 50) anaesthesia, whilst a US study (see http://clinicaltrials.gov/ct2/show/NCT00998894) aims to establish whether the automatic alerting of the attending anaesthetist of the occurrence of a ‘triple low’ will decrease mortality. Should either of these studies show a reduction in mortality, then pEEG monitoring in the high-risk patient may become routine. There will always be an argument that the evidence underpinning the adoption of new technologies and monitoring modalities may be relatively weak, and cannot justify the associated costs. For example, the case against the routine use of oesophageal Doppler monitoring has previously been eloquently argued within this journal 21. It should be remembered, however, that the evidence underpinning the use of some of the monitoring now regarded as routine and mandatory is also sparse. There is no single monitor that has been shown to improve peri-operative outcomes 22, and some widely used monitoring strategies, such as central venous pressure measurement for the assessment of volaemic status, have been shown to be of little value in clinical studies yet are still widely used 23. Proponents of central venous pressure monitoring would argue that they are able to interpret the output measurement in conjunction with other clinical factors, thereby correcting for some of the potential errors. We would take a similar view with pEEG monitoring, and believe that the information gained from their use, allowing for the lack of evidence from randomised controlled trials and inherent weakness and inaccuracies of the devices, still provides the clinician with useful additional information. Most well-trained anaesthetists can administer a safe anaesthetic, according to an individual patient's requirements, without supplementary monitoring – but we believe that a more tailored anaesthetic is possible with the extra information gained from a processed EEG signal 24. What is clear is that this technology should be applied and interpreted with the same rigour and background knowledge as any other monitoring device. Processed EEG monitoring should be initiated from the beginning of the case; applying it after a critical incident has occurred is like shutting the proverbial stable door after the horse has bolted. Application from the start of the case allows calibration at baseline values and any subsequent intra-operative changes can be followed and acted upon appropriately. The authors of NAP5 stressed the need for education in the use of these monitors; hearing about them at an occasional annual meeting is not enough. Like all monitors, they will sporadically throw up random data and patients will behave in unexpected ways. It is only by applying them regularly in low-risk cases that their utility in high-risk cases or during unexpected critical incidents will be truly appreciated. This is a form of on-going ‘fuzzy learning’ that can only be gained from experience. JCA was a member of the Specialist Committee that advised the NICE Advisory Committee that produced its guidance on depth of anaesthesia monitors. He has received a grant from the Neuroanaesthetic Society of Great Britain and Ireland to carry out research into BIS monitoring. No external funding and no other competing interests declared.