Accidental awareness during general anaesthesia

Clinical vignette

A 40-year-old female patient was distressed upon emerging from anaesthesia after open myomectomy surgery. She immediately reported to her recovery nurse that she could remember aspects of what she thought might have been surgery occurring. She had a medical history of fibroids, a body mass index of 38, but was otherwise healthy. This was the first time she had undergone general anaesthesia.

She recalled hearing voices asking for numbered objects, feeling her neck being extended further back, then a plugging sensation in her mouth and a sense of suffocation. The patient tried to cry and shout out to indicate that she was still awake, but was unable to move or communicate. The patient recalled being moved and feeling a sharp pain in her lower abdomen, after which her memories became blank. The patient remembers being scared about how this would end and if she would survive.

Despite a detailed debrief and explanation provided by her anaesthetist, the patient later developed post-traumatic stress disorder with intrusive flashbacks, panic attacks and fear of future anaesthesia, and received regular counselling from a psychologist.

Clinical question

How can accidental awareness during general anaesthesia (AAGA) be prevented and managed to minimise psychological harm in patients undergoing surgery?

Discussion

Patients expect general anaesthesia to be associated with unconsciousness and an absence of memory for the duration that anaesthesia is provided. Failure of both these processes is termed accidental awareness during general anaesthesia (AAGA), which describes an unintended return of consciousness between the planned induction and emergence periods of general anaesthesia.

AAGA remains a fundamental complication of anaesthetic practice and one of the few complications experienced by surgical patients that anaesthetists have sole, undivided responsibility for. Even though AAGA is rare, the vast number of surgical procedures carried out worldwide means that the odds are high that someone, somewhere, experiences awareness each day.

Understanding AAGA relies upon an appreciation of the underlying complexity of the apparently simple goal of anaesthesia: to safely take a patient from a conscious to an unconscious state, then back again. Both consciousness and general anaesthesia are complex phenomena, with multiple theories and levels of understanding regarding mechanisms of action. Activation of depressant protein channel receptors, GABA_A, is a key feature of general anaesthetic drugs, leading to disrupting communication between key brain regions and thalamocortical circuits. Higher doses of general anaesthetic drugs preferentially reduce brain activation in higher order regions, responsible for information integration, over primary sensory areas [1, 2]. This conceptual model has implications for understanding how awareness is perceived and how other higher order cortical processing, such as memory, is affected during general anaesthesia.

The easiest to detect (and arguably most important) form of AAGA is with declarative memory or explicit recall in which the patient can remember and report, in detail, the events that occurred during the awareness episode. But return of consciousness may also occur without a clear memory of the experience being encoded or able to be retrieved later. This form of awareness is known as implicit recall, and may manifest as behavioural changes, depression or sometimes not at all.

The process of dimming, then loss, of consciousness is experienced by most patients with the same clear pattern: one moment they remember being in the operating or anaesthetic room, lying on their back, with a mask delivering oxygen over their mouth and an anaesthetist beside them. The next moment, they wake with a total ‘blackout’ in between, feeling like no time has passed and with no memory of intraoperative events. This experience is exactly our intention as anaesthetists: to generate hypnosis, analgesia and amnesia, and therefore make the intolerable process of surgery tolerable and possible.

As a frightening iatrogenic complication, AAGA has a high public profile, increases patients’ apprehension of surgery, and affects the medical–legal risks [3] associated with anaesthesia. The root cause is always the inadequate delivery of an effective dose of an anaesthetic drug to a patient’s brain, which can occur for a variety of reasons (e.g. situational challenges to the delivery of anaesthesia for emergency patients, variability in patient responses to anaesthesia drugs or equipment malfunction). Consciousness is multidimensional and has several components, including felt experience, self-awareness, perception and attention. Our window of measurement of consciousness is limited, as experience is inherently first person, highly subjective, multifaceted and reliant on communication to share understanding. For this reason, it is better to think about the interaction of general anaesthesia with consciousness as less of a binary ‘on–off’ switch, but rather as dynamic effects on a panel of dimmer switches, each manipulating different neural bases that contribute to conscious experience.

By far the most common AAGA experiences are trivial, consisting only of brief, hazy recollections. Less commonly, patients recall more specific details of sensory input from their surroundings during surgery. This may be a feeling of something in a patient’s mouth, hearing noises or feeling touch. Most patients who report awareness during surgery do not feel pain and report brief experiences, lasting less than 5 minutes. Common experiences include: hearing voices or the noise of equipment; trying to move to alert staff and being unable to; feeling anxious that something has gone wrong with the operation and being powerless to do anything; or feeling frightened that things are going to get worse. Half of patients reporting memories of awareness find the experience distressing, with pain and paralysis being strongly associated with distress.

Lack of consciousness is generally inferred from external signs, such as a lack of response or measurement of brain electrical activity, using tools such as processed electroencephalograms (pEEG). These limited diagnostic tools and observational methods do not fully capture internal awareness across the continuous spectrum of consciousness, including hidden or partial conscious states. One example of how our simple monitoring tools are limited is in the common situation of patients receiving drugs used to paralyse skeletal muscles (called neuromuscular blocking drugs) as part of general anaesthesia. These drugs are administered to facilitate the placement of airway support devices and maintain appropriate conditions for safe surgical operations. However, a patient accidentally provided with neuromuscular blocking drugs with inadequate dosing of hypnotic anaesthetic drug may have incomplete ablation of consciousness, despite appearing completely unresponsive. Neuromuscular blocking drugs do not affect consciousness, instead they block the window of observation used to measure the conscious state [4]. In this situation, a patient may still perceive experiences, but be unable to signal a response due to the inability to trigger muscle activation. To complicate matters further, pEEG monitors, which are now frequently used to monitor the effects of anaesthesia on brain function, are subject to potential error and misinterpretation in the absence of electrical signals from muscle activity [5]. As such, pEEG can provide false reassurance to anaesthetists providing care to patients receiving paralysis drugs, and introducing potential risk of AAGA without appropriate interpretation.

Furthermore, due to the differential effects of general anaesthetic drugs on cortical activity, sensory stimuli can sometimes still be detected during general anaesthesia. Playing a piece of music to a patient receiving anaesthesia drugs can activate the auditory cortex, although this information is not necessarily transmitted to higher order centres responsible for consciousness processing [6]. In a sense, the music is heard but not perceived. Another vivid example comes from an infrequently used anaesthetic monitoring approach called the ‘isolated forearm technique’ (IFT) [7, 8]. IFT consists of the inflation of a blood pressure cuff on a patient’s arm to above systolic blood pressure during the induction of general anaesthesia. Neuromuscular blocking drugs are injected intravenously; however, because the arm is vascularly isolated from the rest of the circulation the distal hand remains unparalysed. Surgery can proceed (with the cuff deflated after 20 minutes) if the patient is asleep. If awake, the patient could use this unparalysed arm to communicate. It is also possible to issue verbal instructions to the anaesthetised patient to test for appropriate motor reactions, too.

Although studies of response during IFT are small in sample size and few in number, remarkably up to 5% of tested patients can respond to verbal commands by contracting with their hands; almost all report no memory of doing so afterwards [9, 10]. This is despite no otherwise apparent clinical signs of awareness and provision of general anaesthesia sufficient to obtund a response to surgical stimulation, and no spontaneous movement in the absence of command.

The state of wakefulness without recall that is detected by IFT may be very similar to that of intended sedation with amnesia (e.g. by use of drugs like midazolam). In contrast, the responses may also be explained by a distinct brain state in which conscious perceptions are uncoupled from surgical stimulation, but responses are not uncoupled from verbal stimulation. This state has been termed ‘dysanaesthesia’ [7].

Understanding how the neurobiology of memory is modulated by both the pharmacology and process of anaesthesia is important when interpreting AAGA reporting. Not only do most anaesthetic drugs have amnesic properties at sub-anaesthetic threshold, but all memories are vulnerable to bias or being overwritten by later similar events, which can impair their subsequent retrieval. When patients regain consciousness during emergence at the end of surgery, the memory of this episode of waking may be more easily understood and recalled, suppressing any intraoperative memory of wakefulness.

Reported patient experiences of AAGA can therefore be thought of in concentric circle terms, with the most frightening complications occurring only to a small minority [11]. Most patients undergoing surgery experience no awareness whatsoever (the outermost ring). However, a smaller proportion of patients might experience a period of return of consciousness during surgery but have no memory of that process (these patients may be those represented by a positive response to verbal command during IFT). An even smaller proportion will remember being awake, but variations in the encoding of memories mean that they may require prompting or direct questioning to retrieve the memory. A tiny minority experience awareness and can immediately and spontaneously describe their experience after surgery (the innermost circle).

The largest ever study of AAGA was The Fifth National Audit Project (NAP5) of the Royal College of Anaesthetists and the Association of Anaesthetists of Great Britain and Ireland ‘Accidental Awareness during General Anaesthesia in the United Kingdom and Ireland’, published in 2014 [12]. In contrast with previous research involving direct questioning of patients about potential memories shortly after general anaesthesia, NAP5 collected only reports made spontaneously by patients to clinical teams. Although reports were only admissible if they were collected during a hospital admission in the UK and Ireland between June 2012 and May 2013, the awareness event could occur at any time in the past.

Of the 300 reported cases of awareness, a committee judged 141 to be certain or probable cases of awareness. Based on the results of a previous activity survey, the incidence of spontaneously reported awareness was estimated as 1:19,600. This proportion is so markedly low compared with outcomes reported by studies in which patients are directly questioned about their memories of anaesthesia (e.g. 1:600 in the Sandin et al. study of 12,000 prospectively recruited patients in Sweden from 2000 [13]) that its value cannot be interpreted without critical discussion of the discrepancy. In this context, given relatively similar patient cohorts and anaesthetic techniques to those in other UK and USA studies, there must be a difference in outcomes detected. The explanation returns to the concentric circle model for awareness reporting; suggesting that some awareness occurs, but is not remembered by patients without direct prompting, and a smaller proportion of cases are always recalled.

One conclusion of these findings is that direct questioning detection is weighted towards sensitivity, whereas measuring only spontaneous reports of AAGA may be more specific and reliable. However, sensitivity may be compromised by factors that interfere with decisions or the capacity of patients to file a report. For these reasons, it is reasonable to consider NAP5 as the most powerful study for describing the relative risks of AAGA between different patient populations; however, the absolute risk values are subject to variable interpretation.

Therefore, there is no simple answer to exactly how rare AAGA is, as it depends on how you detect, categorise and measure the outcome. However, most estimates using direct questioning consistently put the risk of recalling some memories attributable to the intraoperative period at approximately 1–2:1000 [13, 14, 15, 16, 17, 18, 19].

Several major findings on relative risk are clearly evident. Especially that obstetric, emergency and cardiac surgery patients, obese patients or those with difficult airway management are all at elevated risk of experiencing AAGA. Obstetric patients were markedly over-represented in reported cases of AAGA in NAP5, accounting for 0.8% of all general anaesthetics in the NAP5 activity survey, but almost 10% of AAGA reports. This translated into an incidence of awareness of approximately 1:670 general anaesthetics (versus ~1:8000 in non-pregnant patients receiving neuromuscular blockade drugs during anaesthesia). A more recent study by Odor et al. used direct question screening to detect AAGA in 3000 UK patients following obstetric surgery with general anaesthesia, estimating the risk to be as high as 1:256 [20]. One of the major findings from this study was an association between AAGA and use of the older anaesthetic hypnotic drug thiopental, which has persisted in obstetric anaesthetic practice despite declining use elsewhere. Appreciation of AAGA risks has catalysed the removal of thiopental from routine use in the UK, meaning that the current AAGA risk for patients is likely to be much lower than the values reported in the Odor et al. study from 2018.

Interestingly, there are two other important trends for the relative risk of AAGA. The first is that AAGA is an overwhelming complication that occurs in patients receiving neuromuscular blockade drugs (~1:8000 versus ~1:136,000 in patients with no paralysis in the NAP5 study). This is because paralysis removes one of the most vital and immediate signs of inadequate anaesthesia, that of patient movement in response to painful stimuli. The second is that young, female patients (characterised exclusively by the obstetric patient cohort) are most resistant to the effects of general anaesthetic hypnotic drugs, requiring higher doses, emerging more rapidly after anaesthesia and having, by far, the highest response proportion during IFT.

The skill and experience of a trained anaesthetist are vital in carefully selecting and adjusting the dose of hypnotic agent, which is likely to be sufficient to prevent awareness in the majority of patients. The use of depth of anaesthesia monitoring devices, alongside recognition of their inherent limitations, helps to guide the continual titration of anaesthetic dosing throughout surgery. As a consequence, general anaesthesia is remarkably safe and, mostly, uneventful despite the complexity of the interactions that occur.

Nevertheless, incidences of AAGA persist, with many being potentially avoidable. Improving anaesthetist education, drawing attention to the issue of AAGA, developing more reliable monitoring techniques and establishing best organisational practice for vigilance of drug errors and pre-operative assessment of risk factors are all vital steps in preventing awareness. Providing a vigilant and safe environment for delivery of anaesthetic care, with particular focus on continuous and adequate hypnotic drug delivery, especially when risk factors for AAGA are present, is essential [21].

Should AAGA occur, then early identification and supportive intervention may reduce the risk of developing psychological harm. Aftercare focuses on providing prompt, sympathetic communication alongside an investigation of the root cause of the event and shared learning [21]. In some cases, a duty of candour response may be required. The presence of signs of anxiety state or depression is an indicator for formal psychological assessment and ongoing support (e.g. cognitive behaviour therapy).

Disclaimer

This Narrative article is adapted with permission from Case Studies in Perioperative Medicine, a UCL Press open access educational resource, available from https://doi.org/10.14324/111.444.9781787356917.01

Declarations and conflicts of interest