Half (72) of all Certain/Probable or Possible reports to NAP5 involved induction or transfer to theatre.

Neuromuscular blockers were used in 93% of these cases and in one third of cases the induction dose was judged to be inappropriately low.

Half the cases in this category were an emergency and 35% of patients were obese. Factors that contributed to underdosing included reduction of doses to compensate for inadequate optimisation in the sick and not contributing for increased volume of distribution and cardiac output in the obese.

A number of cases showed a trend to dosing that equated to one ampoule of a drug rather than to patient need and in others underdosing of NMB increased airway management difficulty.

Choose induction doses with care especially in the obese

Assess loss of consciousness logically (AVPU)

Assess the airway of every patient and develop a stratagem for management

51 of 141 (36%) reports of AAGA evolved during the maintenance phase (four of these also involved AAGA at induction and one at emergence).

The duration varied greatly, from less than an estimated minute in one third up to over sixty minutes. A short experience did not diminish distress or morbidity. Pain and paralysis was the most common experience during maintenance, and in 40% of cases the time of commencement of surgery was recalled.

Just over a third of the cases were an emergency.

Some 20% of cases involved a vaporiser mishap; in a few omission to switch it on when arriving in theatre.

In 3 out of 4 cases the provoking cause of AAGA was identified, and judged as preventable. But in a quarter the cause was uncertain. Some may represent a resistance to the effect of anaesthetic agents. A few cases occurred with concurrent BIS monitoring. 83% of cases using volatile agent anaesthesia employed and recorded end-tidal agent monitoring.

A checklist should be conducted before the start of surgery to confirm delivery of adequate anaesthesia; maybe incorporated into the WHO checklist.

The surgical team should formally confirm with the anaesthetist that it is appropriate to start a procedure before doing so.

If AAGA is suspected during a painful procedure, then prompt delivery of analgesia, as well as deepening the level of anaesthesia, and giving verbal reassurance to the patient is needed at this time.

Anaesthetic rooms are essentially a UK-only idea (78.7% of all GA cases in the UK are induced there). Six reports were made as a consequence of failure to turn on a vapouriser after induction or in theatre.

Use of volatiles at too low an inital level, at too low initial flows or the use of an unchecked or empty device were causative.

Rapid rescue efforts mitigated the severity of patient experience.

A simple checklist could prevent such problems (see below)

Almost a fifth of the reports received involved emergence after the procedure.

The vast majority of patient reports described experience of distress from paralysis while awakening. As elsewhere, these cases highlight that adverse outcomes were almost exclusively associated with the use of neuromuscular blocking drugs without a nerve stimulator.

In a third of cases communication failure within the team highlighted led to the ill-judged selection, dose, or timing of neuromuscular drug use.

35% of cases were an emergency and 59% of patients were overweight or obese.

The management of awake extubation techniques may require discussion with the patient in advance, but certainly require appreciation of the speed of offset of anaesthetic drugs.

Patients reporting awareness with paralysis during this phase of anaesthesia should be managed in the same manner as for the other phases.

There should be formal confirmation from the surgeon to the anaesthetist and other theatre staff when surgery has finished.

Patients who have less than full motor capacity as a result of pharmacological neuromuscular blockade should remain anaesthetised.

A nerve stimulator should be used to establish motor capacity. Consider including a nerve stimulator as 'essential' in monitoring guidelines when neuromuscular blocking drugs are used.

There were 14 cases of AAGA during obstetric general anaesthesia reported to NAP5. Obstetric cases account for 0.8% of general anaesthetics in the NAP5 Activity Survey but ~10% of reports of AAGA to NAP5, making it the most over-represented of all surgical specialties.

Most reports of AAGA occurred after Caesarean section, but a number of reports followed obstetric anaesthesia for other procedures.

Obstetric general anaesthesia includes most of the risk factors for AAGA; rapid sequence induction with thiopental, neuromuscular blockade during maintenance, a population with a relatively high incidence of obesity and difficult airway management. The urgency of the situation necessitates surgery beginning within moments of induction.

Thiopental was used for <3% and RSI for <8% of anaesthetic inductions, but both were used for >90% of Caesarean sections. An opioid was used in 25%. Nitrous oxide was used for >70% of Caesarean sections but <30% non-obstetric cases.

Anaesthesia for Category 1–2 Caesarean section was less often performed by consultants than other (non-obstetric) emergencies and more often by quite junior trainees. Comparison of management of obstetric and non-obstetric emergencies should be approached with caution, as it should be borne in mind that the NCEPOD classification excludes obstetric cases, and an NCEPOD ‘emergency’ should not be regarded as synonymous with a ‘category 1’ Caesarean section.

There were 14 cases of AAGA in obstetric patients: 13 in Class A (Certain/probable) and one in Class B (Possible). In addition there were two cases involving drug errors (Class G) and 12 ‘Statement Only’ cases. The obstetric cases thus represent 14/141 (~10%) of the total number of Certain/probable and Possible AAGA cases.

In the NAP5 Activity Survey obstetric general anaesthetics constituted 0.8% of the total general anaesthesia cases for the UK. Thus, obstetrics is over-represented in AAGA cases by a factor of >10.

All the experiences in obstetric cases concerned awareness at induction or soon after. All except one case involved use of neuromuscular blockade; in none was the use of a specific depth of anaesthesia monitor recorded.

Anaesthetists should regard obstetric patients, particularly those undergoing Caesarean Section, as being at increased risk for AAGA. This risk should be communicated appropriately to patients as part of the consent process.

Consideration should be given to reducing the risk of AAGA in healthy parturients by (a) the use of increased doses of induction agents (b) rapidly attaining adequate end-tidal volatile levels after induction without delay (c) use of nitrous oxide in adequate concentrations (d) appropriate use of opoids (e) maintaining uterine tone with uterotonic agents to allow adequate concentrations of volatile agents to be used.

Before induction, the anaesthetist should have decided what steps to take if airway management proves difficult, with maternal wellbeing being the paramount consideration, notwithstanding the presence of fetal compromise. An additional syringe of intravenous hypnotic agent should be immediately available to maintain anaesthesia in the event of airway difficulties, when it is in the mother’s interest to continue with delivery rather than allow return of consciousness.

Anaesthetists should regard failed regional technique leading to the need for general anaesthesia for obstetric surgery to be an additional risk for AAGA (and for other complications).

Anaesthetists should regard the presence of antibiotic syringes during obstetric induction as a latent risk for drug error leading to AAGA. The risk can be mitigated by physical separation, labelling or administration of antibiotics by non-anaesthetists. Using propofol for induction mitigates the risk of this drug error.

NAP5 received four reports of AAGA during cardiac surgical procedures and four during thoracic surgery.

Based on the Activity Survey data this gives an incidence of reports of AAGA in cardiac and thoracic surgery of 1 in 10,000 and 1 in 7,000 respectively: both higher than the overall incidence of reports.

Most reports in this field involved either brief interruption of drug delivery (caused by human error or technical problems) or use of intentionally low doses of anaesthetic drugs in high-risk patients.

There are too few cardiothoracic cases of AAGA reported to NAP5 to make robust recommendations.

Cardiothoracic anaesthesia would seem to lend itself well to research questions relating to whether EEG-based monitoring helps achieve the optimum balance between too light and too deep levels of anaesthesia.

Spontaneous reporting of AAGA in children is very rare and may be delayed until adulthood.

Children's reports can be as reliable as those from adults.

Children should be believed and treated sympathetically.

Serious long-term psychological harm an anxiety states are rare but do occur.

The paucity of AAGA cases involving children reported to NAP5 means it is unjustified to make specific recommendations about prevention and management of AAGA in this group.

All cases where AAGA was reported from the ICU/ED involved critically ill patients: concerns about the adverse effects of induction of anaesthesia would have been justified.

The performed procedures were appropriate and RSI was used appropriately.

In common with the vast majority of reports, all cases of AAGA from ICU involved patients who had received a neuromuscular blocking drug, so when used, the risk of AAGA should reasonably be considered higher.

All ICU reports were associated with distress and the majority with subsequent psychological harm. This should guide a supportive approach to an ICU patient who reports AAGA (see Psychological Support Pathway, Chapter 7, Patient Experience).

AAGA in the critically ill may occur despite cardiovascular instability. Early support of the cardiovascular system that then enables increased doses of anaesthetic agents is likely to reduce distressing AAGA.

Critical illness, leading to an obtunded mental state also does not guarantee absence of consciousness that retention of a memory for events. This implies the pathological brain state preventing spontaneous or reflex movement does not inevitably prevent perception. Even patients with lowered conscious levels should receive adequate anaesthesia for intubation and surgical procedures where this is safe.

Where critical illness demands a significant reduction in the doses of anaesthetic agents that can be safely administered, the possibility of wakefulness should be considered. Patient explanation and reassurance are likely to be of benefit to patients experiencing AAGA.

Notwithstanding these comments, the Panel noted that AAGA during anaesthesia in the critically ill may not be completely avoidable without putting patients at risk of major harm from the cardiovascular complications of anaesthetic agents.

In several cases AAGA arose soon after intubation and involved infusions of propofol (without opioids) in patients who had received neuromuscular blocking drugs. Delay in starting infusions and use of very low dose infusions contributed. Patients receiving low dose non-TCI infusions of propofol while paralysed are likely to be at increased risk of AAGA (see Chapter 18, TIVA). Use of TCI infusions might lead to more appropriate doses of anaesthetic agent being administered. Using a checklist prior to intubation (such as that described in NAP4, (Cook et al., 2011), or a checklist as suggested in Chapter 8 (Induction), should reduce the risk of delays in initiating appropriate anaesthetic/sedative (and vasopressor/inotrope) infusions.

There is scope for research in validating the use of DOA monitoring in the critically ill (Nasraway et al., 2002; Vivien et al., 2003). In the Activity Survey, only three patients out of 29,000 undergoing general anaesthesia in ICU or ED received DOA monitoring (one BIS, one entropy and one other; ~1:10,000). It is not known how many UK or Irish ICUs have immediate access to DOA monitoring.

Delays in starting infusions of anaesthetic agents were on more than one occasion, attributed to distraction. In one case, difficult airway management and a failure to administer extra induction agent, likely contributed to AAGA. Management of critical illness is inevitably complex and is a rich potential source of human factors affecting delivery of reliable safe care.

It is notable that all ICU AAGA reports were made by patients who had short ICU stays with a short period of intubation, and that the interval to reporting the episodes was also consistently short. This raises several questions, including the possibility that episodes of AAGA may occur but be forgotten when critical illness or sedation is prolonged.

Overall the data reported here raise concerns about a higher incidence of AAGA during anaesthesia in patients from ICU than in other settings. However our methodology, which focussed primarily on theatre practice, means we cannot confirm this. Relevant national organisations could usefully consider whether further research should be commissioned to study this important area and whether our learning points could drive recommendations for practice.

Awareness (AAGA) encompasses a wide range of experiences (from the trivial to something akin to feelings of torture) and a wide range of psychological consequences (from none to life-changing).

In about half of NAP5 reports recall was expressed expressed in a neutral way, focused on a few isolated aspects of the experience.

In about half of cases there was distress at the time of the experience ; distress was particularly likely with sensations of paralysis or pain , but could also occur when only isolated sounds or tactile sensations were experienced.

Distress during AAGA was strongly associated with subsequent psychological sequelae.

Understanding what was happening, or what had happened, seemed to mitigate immediate and longer-term psychological distress.

Active early support may offer the best prospect of mitigating the impact of AAGA, and a structured pathway to achieve this is proposed.

If AAGA is suspected intra- or peri-operatively, anaesthetists should speak to patients to reassure them that they know of their predicament and are doing something about it.

Conversation and behaviour in theatres should remain professional, especially where there is a situation or concern that AAGA is a risk (e.g. RSI, prolonged intubation, transfer). Adverse impact of any recall may be mitigated where the patient is reassured by memories of high quality care.

All reports of AAGA should be treated seriously, even when sparse or delayed, as they may have serious psychological impact. If reported to someone else, every attempt should be made to refer the case to the anaesthetist responsible.

The anaesthetist who provided the anaesthesia care at the time of a report of AAGA should respond promptly and sympathetically to the patient, to help mitigate adverse impacts.

Healthcare or managerial staff receiving a report of AAGA should inform the anaesthetist who provided the care and institute the NAP5 Psychological Support Pathway (or similar system) to provide patient follow up and support.

Almost one in five (32 patients) reports to NAP5 did not follow a GA and, as in AAGA, women predominated, here in a ratio of 2:1

The majority of reports related to orthopaedic or endoscopic procedures.

Sedation was administered by anaesthetists in 2/3 so because sedation is administered by a wide range of medical and paramedical staff in the UK total estimates of incidence are difficult. The activity survey data for anaesthetist delivered sedation calculated ~310,000 cases per year suggesting a figure of ~1:15,000.

The type of experience did not correlate with the degree of long term harm. In two thirds of cases the experience was auditory or tactile, in one third pain. Only one patient experienced paralysis and pain and was distressed at the time. Despite these apparently limited sensations, longer term harm (NPSA score) was moderate or severe in about half.

In 81% of reports communication or expectation management were contributory or causal. Some reports specifically mentioned the nature of explanations including the terms 'sleep' or 'light anaesthesia', in others there was documented evidence that sedation was the planned care and had been explained to the patient.

Written information should be provided well in advance of the procedure and reinforced verbally on the day.

Sedation should be described using patient-centred terms as in the table below.

NAP5 identified human factors (HF) contributors in the majority of reports of AAGA.

The commonest contributory factors, using the National Patient Safety Agency (NPSA) classifications (see Classification > Cause page) were: medication, patient, education/training and task.

All anaesthetists should be educated in human factors, so they can understand their potential impact on patient care and how environments, equipment and systems of work might impact on the risk of, amongst other things, AAGA.

Investigation of and responses to episodes of AAGA – especially those involving drug error – should consider not only action errors, but also the broader threats and latent factors that made such an event more or less likely.

RSI is only used in 7.4% of all GAs administered in the UK but an RSI was used in 36% of all certain/probable reports .

Sodium thiopental (STP) is used in <3% of inductions in the UK and 87% of these inductions are RSI. In NAP5 92% of reports involved an RSI used STP.

Opioids are used in 75.8% of all non-LSCS RSIs and 23.4% of LSCS RSIs in the UK. Only 39% of reports of AAGA during RSI included use of an opioid .

STP in conventional doses (4mg/kg) has a wide range of BIS values including many >70 and duration of action is frequently shorter than the duration of paralysis from suxamethonium.

RSI increases the risk of difficult/failed intubation and failure to plan whether to continue anaesthesia or to wake the patient was evident in several reports, as was relative underdosing of NMBs which may contribute to difficult airway management.

Reference to standard doses should guide intravenous induction agent doses

Caution should be exercised using thiopental for RSI

Obesity and difficult airway management are risk factors during RSI

Prolonged/difficult airway management will require continued (logically intravenous) anaesthesia

Specific depth of anaesthesia (DOA) monitors are rarely used in UK anaesthetic practice: in only 2.8% of general anaesthetics for processed EEG (pEEG) and 0.03% for the isolated forearm technique (IFT).

Of the 141 reports to NAP5 judged to be Certain/probable/possible AAGA, six (4.3%) occurred despite use of a pEEG monitor. However lack of detail means appropriate and continuous use cannot be confirmed. These monitors appeared to be used in a ‘targeted fashion’: for instance, in the Activity Survey, whereas pEEGs were used in only 3.5% of cases where volatile and neuromuscular blockade (NMB) was used, they were used in 23.4% of cases with total intravenous anaesthesia (TIVA) and NMB. A crude analysis of the cases of AAGA in which pEEG was used or omitted was not able to detect whether there was a marked protective effect of its use.

Only one report of AAGA in association with DOA monitoring was followed by adverse psychological sequelae. The possibility of a more subtle benefit of DOA monitors in protecting against ‘AAGA with sequelae’ merits investigation.

Although end-tidal anaesthetic gas monitoring is an alternative to DOA monitoring, in ~75% of reports to NAP5 it would probably have been impractical or ineffective at preventing AAGA.

The overall findings are supportive of the use of DOA monitoring in selected circumstances, but provide no support in others.