You can respond to this article at http://www.anaesthesiacorrespondence.com
Monitoring the brain in cardiac surgery – an evolving area for research
Article first published online: 9 FEB 2012
Anaesthesia © 2012 The Association of Anaesthetists of Great Britain and Ireland
Volume 67, Issue 3, pages 216–219, March 2012
How to Cite
Grocott, H. P. (2012), Monitoring the brain in cardiac surgery – an evolving area for research. Anaesthesia, 67: 216–219. doi: 10.1111/j.1365-2044.2012.07076.x
- Issue published online: 9 FEB 2012
- Article first published online: 9 FEB 2012
Millions of patients worldwide undergo procedures requiring anaesthesia every year, and although anaesthetic safety has increased substantially over recent decades, significant peri-operative morbidity remains. Although many of the major causes of mortality have been addressed and many of the significant morbidity issues have received considerable attention, the bar continues to be appropriately raised with respect to what we consider a good peri-operative outcome. No longer is simply emerging from anaesthesia relatively pain-free and haemodynamically stable considered sufficient when we say the patient ‘did well’. Long-term survival with freedom from significant morbidity is increasingly the focus of peri-operative outcomes studies. One of these morbidities relates to postoperative cognitive dysfunction (POCD), which is addressed by van Harten et al. in this issue of Anaesthesia . Less dramatic adverse outcomes (compared with stroke), such as POCD and delirium, are clearly no less important and have themselves been associated with significant long-term morbidity [2, 3].
Multiple approaches have been utilised to add information aiding our understanding of the risk factors, pathophysiology, early detection and potential therapy for these adverse cerebral outcomes. Specific monitoring of the brain is an important aspect that has added information to each of these considerations. However, the brain, arguably our most important of organs and despite being an obvious site of anaesthetic action, ironically is inconsistently the focus of targeted monitoring during anaesthesia. There have been several notable attempts to open a window to the ‘black box’ represented by the brain, but compared with other physiologic parameters related to the respiratory and cardiovascular system, it lags far behind. van Harten et al. address several cerebral monitoring modalities, and in doing so, highlight several of the issues and limitations related to the currently available techniques .
Transcranial Doppler (TCD), in addition to its ability to interrogate various cerebral vessels (most commonly the middle cerebral artery) to determine blood flow velocity (an indirect measure of cerebral blood flow), has frequently also been used for detection of emboli, thus elucidating their potential relationship to POCD. Although historically emboli detected by TCD have been shown to be associated with POCD , this has recently been questioned [5, 6]. One of the limitations with the data linking TCD emboli to postoperative neurological outcome is the inability of current Doppler technology to differentiate reliably between particulate and gaseous emboli, despite several notable attempts [5, 7]. This poor signal-to-noise ratio is often further impaired by gaps in monitoring because of loss of signal due to patient movement. Logistical limitations, coupled with the cumbersome nature of TCD signal acquisition and probe stabilisation, have substantially limited its usefulness in the peri-operative setting . With a few exceptions in the postoperative period, such as for vasospasm monitoring in patients with subarachnoid haemorrhage , it is a cumbersome cerebral monitor whose use is difficult to recommend confidently.
In contrast to TCD, electroencephalography (EEG) has made major inroads into the peri-operative arena. It has been widely available for clinical use for almost two decades, though is still inconsistently used. Although multi-channel monitoring has only select demonstrable peri-operative benefit , single (or dual) channel EEG monitoring has had substantive demonstrable utility  as processed EEG depth of anaesthesia monitors, such as the bispectral index (BIS). Initially intended as a monitor to gauge hypnotic depth in order to facilitate reducing the incidence of intra-operative awareness, early BIS studies focused on other more germane aims. For example, its utility in titrating emergence, thereby reducing operative times, as well as decreasing the incidence of postoperative nausea and vomiting, has been demonstrated . Studies supporting the BIS as a specific intra-operative awareness monitor initially consisted of observational studies . More carefully designed studies did not appear until the BIS had already been widely clinically available, despite only being sporadically used. The B-Aware trial, a large prospective, randomised trial by Myles et al. that studied patients at increased risk of intra-operative awareness, demonstrated clear utility in reducing this complication . However, criticism related to issues such as study design raised questions regarding the generalisability of BIS.
The controversy surrounding the impact of BIS in reducing intra-operative awareness was highlighted in the B-Unaware  and recent BAG-RECALL  trials. Both of these trials failed to replicate the beneficial awareness results initially reported by Myles et al.. If anything, the more recent trials demonstrated a potential awareness liability with the use of these monitors. Part of this may have been related to their own intrinsic study design. An alternative explanation for the apparent failure of a BIS-guided protocol  to reduce intra-operative awareness could be related to the BIS value chosen and not the actual BIS technology itself. In titrating anaesthesia to a BIS < 60, a conventional target threshold that is arguably at the ‘edge of wakefulness’ , one may risk brief excursions above this level during the normally fluctuating anaesthetic course that could result in brief periods of awareness. The incidence of awareness in these studies might have been lower had a more conservative upper BIS limit (e.g. 45–50) been used. This might have allowed a sufficient anaesthetic buffer below the actual awareness danger zone of 60 BIS units. Accordingly, Avidan et al. may have actually reported a failure of the 40–60 BIS target as opposed to a failure of this depth of anaesthesia technology itself. Interpretation of these results warrants caution so that practitioners do not prematurely risk throwing out the baby (i.e. a more appropriate BIS target) with the bathwater (i.e. BIS technology) . These studies serve to highlight the limitations of processed EEG as a depth of anaesthesia monitor and are a call to action for increased study on cerebral monitoring.
Despite the controversies surrounding BIS, it and other similar depth of anaesthesia monitors have considerable potential to impact our practice. Irrespective of the issues related to their potential utility in reducing intra-operative awareness, the potential non-awareness capabilities of the BIS are still being elucidated and may have significant utility in the peri-operative setting. For example, an association between deeper hypnotic levels (i.e. prolonged reductions in BIS) and overall patient mortality was first reported by Monk et al. . In a post hoc analysis of their large trial examining POCD, cumulative deep hypnotic time (i.e. BIS < 45) was one of the multivariate predictors of one-year postoperative mortality. This introduced the concept of a potential risk to relative anaesthesia ‘overdose’. Although regarded as controversial when this analysis was published [19, 20], other studies corroborated this relationship between a prolonged low BIS and adverse outcomes. For example, in a post hoc analysis of the B-Aware trial, not only was a cumulative low BIS time (< 45 for > 5 min) associated with an increased incidence of myocardial infarction and stroke, but mortality was also significantly increased . Importantly, these were association studies and it is uncertain whether the low BIS (i.e. a relative anaesthetic overdose) resulted in organ dysfunction (perhaps from anaesthetic mediated haemodynamic suppression) and subsequent increased morbidity and mortality, or whether organ dysfunction lead to an increased sensitivity to anaesthetics and subsequently a low BIS, with the tenuous baseline state of their organ function being the mortality culprit . Using cerebral monitoring to determine which patients are at risk and how to intervene to reduce this risk, though not clearly established, remains an intriguing possibility and a focus of future studies.
Irrespective of the ability of BIS to demonstrate any relationship to postoperative outcomes, it may have significant utility in facilitating titration of the haemodynamic management of these patients. For example, a patient whose BIS is well below what would be considered a threshold of ‘light’ anaesthesia may exhibit significant hypertension; this is frequently seen in cardiac anaesthesia settings. Knowing that the patient’s plane of anaesthesia is adequate (i.e. the BIS is ≤ 40) points to the need for a vasoactive agent to be administered, as opposed to any further anaesthesia. Coupled with the data that increasingly link anaesthetic agents to potential neurotoxicity , astute intra-operative management in this situation would be to use a directly acting vasodilator as opposed to exploiting the haemodynamic side effects (vasodilation and haemodynamic suppression) of a potentially neurotoxic substance.
Although the above discussion of monitors that reflect neuronal function (i.e. EEG signals) have strengths, they are not without several previously stated limitations. As such, other modalities that do not directly monitor brain function, but indicate oxygen delivery and consumption parameters, have unique application. Cerebral oximetry employing near-infrared spectroscopy (NIRS) is a non-invasive modality used to estimate regional cerebral oxygen saturation (rSo2) . Cerebral haemoglobin oxygen saturation can be determined by measuring the differential absorption of various wavelengths of light when haemoglobin is either oxidised or reduced . Near-infrared spectroscopy has increasingly been used in the peri-operative setting and many studies have outlined an increased incidence of adverse postoperative outcomes in patients demonstrating significant intra-operative reductions in rSo2. These outcomes include POCD , prolonged hospital length of stay  and other major organ morbidity and mortality . Recently, low baseline measurements of rSo2 have also been demonstrated to identify an increased risk of adverse outcomes [28–30]. Although cerebral oximetry has seen increasing use in several clinical settings [11, 31], it has yet to be recognised as a clinical standard of care.
Despite most of the randomised trials’ being limited by study design issues and/or small sample size, they do point to a possible role in guiding specific interventions. One of the potential utilities to cerebral oximetry is in individualising the care of patients . An emerging field of research was highlighted recently by Joshi et al. in a study that used NIRS to determine the optimal position on the autoregulatory curve for these patients . This and other work has lead to renewed interest in the importance of careful consideration of the cerebral autoregulatory curves . This may have major impact on the outcome of patients, particularly those undergoing cardiac surgery, as it has been previously been demonstrated that patients having impaired autoregulation have an increased incidence of adverse outcomes .
In summary, despite the huge strides that have already been made in peri-operative safety, many opportunities exist for further improvement. The brain will continue to be a major focus, with studies directed at the application and refinement of cerebral monitoring modalities playing an important role. These investigations will be aimed at furthering our understanding of adverse cerebral outcomes, as a means of both determining when the brain is at risk, and providing real-time information to optimise the peri-operative haemodynamic and anaesthetic conditions for these patients. The limitations seen in previous cerebral monitoring investigations identify specific opportunities for further refinements that should be expected to continue the improvement in peri-operative cerebral and related outcomes.
No external funding and no competing interests declared.
- 7Neurosonology. St. Louis: Mosby, 1996., , .
- 28Noninvasive cerebral oxygenation may predict outcome in patients undergoing aortic arch surgery. Journal of Cardiothoracic and Vascular Surgery 2011; 141: 815–21., , , et al.
- 31A proposed algorithm for the intraoperative use of cerebral near-infrared spectroscopy. Seminars in Cardiothoracic and Vascular Anesthesia 2007; 11: 274–81., , .
- 33Predicting the limits of cerebral autoregulation during cardiopulmonary bypass. Anesthesia and Analgesia 2011 Nov 21; doi:10.1213/ANE.0b013e31823d292a., , , et al.