Abstract

Background

This project was devised to estimate the incidence of major complications of airway management during anaesthesia in the UK and to study these events.

Methods

Reports of major airway management complications during anaesthesia (death, brain damage, emergency surgical airway, unanticipated intensive care unit admission) were collected from all National Health Service hospitals for 1 yr. An expert panel assessed inclusion criteria, outcome, and airway management. A matched concurrent census estimated a denominator of 2.9 million general anaesthetics annually.

Results

Of 184 reports meeting inclusion criteria, 133 related to general anaesthesia: 46 events per million general anaesthetics [95% confidence interval (CI) 38–54] or one per 22 000 (95% CI 1 per 26–18 000). Anaesthesia events led to 16 deaths and three episodes of persistent brain damage: a mortality rate of 5.6 per million general anaesthetics (95% CI 2.8–8.3): one per 180 000 (95% CI 1 per 352–120 000). These estimates assume that all such cases were captured. Rates of death and brain damage for different airway devices (facemask, supraglottic airway, tracheal tube) varied little. Airway management was considered good in 19% of assessable anaesthesia cases. Elements of care were judged poor in three-quarters: in only three deaths was airway management considered exclusively good.

Conclusions

Although these data suggest the incidence of death and brain damage from airway management during general anaesthesia is low, statistical analysis of the distribution of reports suggests as few as 25% of relevant incidents may have been reported. It therefore provides an indication of the lower limit for incidence of such complications. The review of airway management indicates that in a majority of cases, there is ‘room for improvement’.

Airway management is fundamental to safe anaesthetic practice and in most circumstances is uncomplicated, but it has been recognized for many years that complications of airway management occur with serious consequences.1,2 Good-quality information on the frequency and nature of major adverse events related to anaesthetic airway management is incomplete. Litigation-based analyses add some insight into the severity of such events and have driven changes in practice.3–6 These indicate that airway and respiratory complications leading to litigation are a small proportion of all claims against anaesthetists but are associated with notably high rates of death and brain damage, high rates of ‘less than appropriate care', and high costs. Owing to the complexity of the relationship between complications and litigation, and the lack of denominators, they do not add information about prevalence or incidence of complications.7,8 Analyses of critical incident reports in the UK have also added useful information, but these reports largely focus on minor incidents and are likely to miss a considerable proportion of major events.9

Knowledge of the incidence of such complications should be an important component of clinical decision-making, risk management, and the consent processes. Information on serious and common complications should guide the specialty into appropriate areas for research by demonstrating areas in which our current practice or performance can improve.

The Fourth National Audit Project of the Royal College of Anaesthetists (RCoA) and the Difficult Airway Society (DAS) (NAP4) was established to estimate the incidence of major complications of airway management in NHS hospitals in the UK and to perform a quantitative and qualitative analysis. Three areas of clinical practice were identified and considered separately: This paper, which reports complications of airway management during anaesthesia, and the accompanying paper, which reports on complications during airway management in ICU and the emergency department, present the major results of the project.10 For reasons of space, this paper is limited to an overview of events that were reported to the project and their quantitative analysis. It should be read in conjunction with the full report of the project available on http://www.rcoa.ac.uk/index.asp?PageID=1089.

  • airway management during anaesthesia;

  • airway management in the intensive care unit (ICU);

  • airway management in the emergency department.

Methods

A two-part project was devised using methods based on the Third National Audit Project of the RCoA.11 First, a census of airway management techniques used in the UK National Health Service (NHS) provided information on anaesthetic activity and airway management techniques in current use (for denominator information); secondly, a registry of the major complications of airway management over a 12 month period recorded details of serious adverse events (for numerator information). Discussions with the National Research Ethics Service indicated that ethical approval was not required. The project was examined by the Patient Information Advisory Group of the Department of Health and the project design was assessed to ensure current standards of patient confidentiality were met. There was wide consultation with other specialist societies and organizations with an interest in this area of clinical care.

Using surface mail, e-mail, and telephone, the anaesthetic department in every NHS hospital in the UK was contacted and invited to participate in the project and to nominate a local reporter who would act as the point of contact for the audit, co-ordinate the census of current activity, and assist with the second phase during which reports of individual serious complications were to be submitted. Data were not sought from private hospitals or Independent Sector Treatment Centres. However, data were collected from treatment centres attached to NHS hospitals.

A detailed written explanation of the NAP4 project and the purpose of the census were placed on both the DAS and RCoA websites. Data collection forms and information sheets were also made available for downloading. The project was very widely advertised in UK journals of anaesthesia, by specialist societies (see Supplementary Appendix) and by a poster campaign to promote awareness and encourage participation. Reminders were sent to hospital local reporters approximately every 6–8 weeks throughout the data collection period.

Part 1: census of clinical activity (denominator data)

A detailed description of the census phase has been published,12 but a brief summary is appropriate here. Each local reporter was asked to return data for a 2-week period in September 2008 on the number of anaesthetics performed in the hospital other than in the ICU and emergency department. For each general anaesthetic, detailed information on the primary airway management technique, defined as that ‘used for maintenance of anaesthesia’ (facemask, supraglottic airway device, or tracheal tube), was requested. Tracheal intubation included all forms of intubation of the trachea, that is, single- and double-lumen tubes, tracheostomy, surgical bronchoscopy, transglottic, and trans-tracheal techniques. The decision on how to collect these data was left at the discretion of the local reporter. Local data were summed to give cumulative totals and submitted to the project team. After collating all returns, the project team used the submitted data to estimate national annual activity and primary airway techniques used.

Part 2: event reporting (numerator data)

Inclusion criteria

Triggers for inclusion and notification to the project were complications of airway management that led to: death, brain damage, the need for an emergency surgical airway, unanticipated ICU admission, or prolongation of ICU stay.

Reports of events occurring in the ICU, in the emergency department, or during transfer were also requested, but these were not used for the calculation of incidence of complications associated with anaesthesia and are the subject of a separate publication.10 The project did not collect data on events occurring out of hospital or on hospital wards.

Definitions

Brain damage was available as an inclusion criterion. Although this was not defined in detail, the manifestations of central nervous system injury and deficit at 1 month were requested.

Emergency surgical airway was taken to include all forms of emergency access to the upper trachea as part of airway management (i.e. surgical tracheostomy, surgical cricothyroidotomy, needle or cannula cricothyroidotomy, or tracheotomy). Emergency surgical airway was an inclusion criterion only when it did not form part of the primary airway management plan. Thus, if a patient presented with critical airway obstruction and required a surgical airway which was planned and performed successfully either after tracheal intubation or without attempting intubation, the case did not meet inclusion criteria. Where the primary airway management plan failed and a needle/cannula or a surgical airway was performed, this was deemed to meet inclusion criteria.

ICU admission that was required as a result of an airway problem was an indication for inclusion. For patients on the ICU, an airway event which would have led to admission to ICU or which led to prolongation of ICU treatment was an inclusion criterion.

Obesity. Reporters were asked to indicate the patient's weight and height and body habitus. Obesity was defined as a body mass index (BMI) of >30 kg m−2 or obese body habitus.

Notification of events

The RCoA-lead (T.M.C.) was notified of events meeting inclusion criteria by e-mail. Local reporters or clinicians involved in the event usually informed the RCoA-lead of an event, but notifications were accepted from any source. The notifier was required to provide their name, the date of the event, the hospital name, and the location of the event. No other identifying data were accepted, including patient or clinician details. The RCoA-lead then e-mailed the local reporter for that hospital, specifying the project inclusion criteria, and requesting confirmation that the case met the criteria and was not a duplicate notification.

Moderator

A moderator was available who was able to discuss the case and offer a confidential opinion on inclusion/exclusion. The moderator was not part of the case review process and could be contacted directly rather than via the RCoA-lead. Cases deemed not to meet the inclusion criteria were withdrawn from the project before being submitted for panel review.

Secure website

For cases meeting criteria, the local reporter was issued with a unique identifying number and website access password using a remote process enabling a secure connection to the project website for on-line data submission. The RCoA-lead had no access to the password but was aware of the unique identification number, which was used to ‘track’ the case.

Data submission

Data were submitted by the local reporter or the clinician involved in the case according to the local preference. After logging on for the first time, a mandatory change of access password was required before proceeding to the reporting forms. The website directed the person submitting data to specific submission forms for reporting of events during anaesthesia, in ICU, or the emergency department. The clinician submitting data could make multiple visits to the website to enter additional data as more information became available. When a report was complete, it was closed and submitted electronically, after which no further changes could be made. The RCoA-lead was unable to view the submitted data but could follow the progress of cases on-line by using the unique identifier to note whether the case was recorded as ‘password unchanged’, ‘password changed', or ‘form closed’. Regular review of the website enabled the RCoA-lead to identify where there were delays in data submission and to encourage submission by direct contact with the local reporter. When a file was completed and submitted, this was notified automatically to the DAS-lead (N.W.). Files were downloaded by the DAS-lead and saved in Word and Excel format for review. If more information was needed, files could be re-opened and a message sent to the local reporter through the project website by a remote process. The DAS-lead was able to access all submitted files but had no knowledge of their origin. In contrast, the RCoA-lead knew event locations but had no access to any files. It was a pre-condition of the project imposed by the Patient Information Advisory Committee of the Department of Health that these two pieces of data could not be linked. Identifying numbers were not present on any information reviewed by the review panel.

Events were included in NAP4 from September 1, 2008, to August 31, 2009: notifications were accepted until June 2010, after which the identification numbers issued to local reporters were destroyed by the RCoA-lead.

Case review panel

Each clinical report was reviewed by a panel of representatives from all the parties involved in the project: the RCoA, DAS, the Association of Anaesthetists of Great Britain and Ireland, the Association of Paediatric Anaesthetists, the Association for Peri-operative Practice, British Association of Otorhinolaryngologists (ENT-UK), the College of Emergency Medicine, the College of Operating Department Practitioners, the Intensive Care Society, the National Patient Safety Agency, the Obstetric Anaesthetists Association, and the Patient Liaison Group of the RCoA.

Case review process

Each clinical case was reviewed at least twice. At each review meeting, the reviewers were in two equal groups (at least five members with differing clinical backgrounds). Each group reviewed half of the cases and when these had been reviewed, the two groups re-joined. Each case was then presented and re-reviewed by the whole panel. If a report was unclear, more information was sought using the process outlined previously. The case was first reviewed to determine whether it met inclusion criteria and to identify duplicate reports. Cases meeting inclusion criteria were included and reviewed, those which did not were removed. The review panel indicated if the event showed underlying contributory, causal, or positive factors (Table 1). Causal factors were those that were considered directly linked to the event whereas contributory factors were those with evidence of impact on the event without being causal. Positive factors indicated areas judged to be of notably good management. The degree of harm attributable to the event was graded using the National Patient Safety Agency (NPSA) severity of outcome scale for patient safety incidents (Table 2).13 Cases with an outcome of death and persisting brain damage were also extracted. Cases were analysed for learning points and some were selected to act as illustrations of clinical care for inclusion in a detailed report of the project. Airway management was classified as good, poor, mixed (elements of both good and poor management), or unclassifiable, reviewers were reminded of likely outcome14 and hindsight bias.15 Reviewers were instructed on the strict confidentiality of the process and if a reviewer was aware of a case (e.g. the case came from their hospital), external knowledge was not admissible in the review process. Clear errors in submitted data (e.g. a fatal outcome not being recorded) were corrected at this time.

Table 1

Categories of incident contributory factors. Each case was examined for causal, contributory or positive factors in these categories. Categories are taken from the National Patient Safety Agency document Seven Steps to Patient Safety: A Guide for NHS Staff13

Factors 
Communication (includes verbal, written, and non-verbal: between individuals, teams, and/or organizations) 
Education and training (e.g. availability of training) 
Equipment/resource factors (e.g. clear machine displays, poor working order, size, placement, ease of use) 
Medication (where one or more drugs directly contributed to the incident) 
Organization and strategic (e.g. organizational structure, contractor/agency use, culture) 
Patient (e.g. clinical condition, social/physical/psychological factors, relationships) 
Task (includes work guidelines/procedures/policies, availability of decision-making aids) 
Team and social (includes role definitions, leadership, support, and cultural factors) 
Work and environment (e.g. poor/excess administration, physical environment, work load and hours of work, time pressures) 
Other 
Factors 
Communication (includes verbal, written, and non-verbal: between individuals, teams, and/or organizations) 
Education and training (e.g. availability of training) 
Equipment/resource factors (e.g. clear machine displays, poor working order, size, placement, ease of use) 
Medication (where one or more drugs directly contributed to the incident) 
Organization and strategic (e.g. organizational structure, contractor/agency use, culture) 
Patient (e.g. clinical condition, social/physical/psychological factors, relationships) 
Task (includes work guidelines/procedures/policies, availability of decision-making aids) 
Team and social (includes role definitions, leadership, support, and cultural factors) 
Work and environment (e.g. poor/excess administration, physical environment, work load and hours of work, time pressures) 
Other 
Table 2

Severity of outcome scale. Categories are taken from the National Patient Safety Agency document Seven Steps to Patient Safety: A Guide for NHS Staff.13 *First aid, additional therapy, or additional medication. Excludes extra stay in hospital, return to surgery or readmission. **Return to surgery, unplanned re-admission, prolonged episode of care as in- or out-patient or transfer to another area such as intensive care. ***Permanent lessening of bodily functions, sensory, motor, physiological, or intellectual

Grade of severity Description 
None No harm (whether lack of harm was due to prevention or not) 
Low Minimal harm but necessitating extra observation or minor treatment* 
Moderate Significant, but not permanent harm, or moderate increase in treatment** 
Severe Permanent harm due to the incident*** 
Death Death due to the incident 
Grade of severity Description 
None No harm (whether lack of harm was due to prevention or not) 
Low Minimal harm but necessitating extra observation or minor treatment* 
Moderate Significant, but not permanent harm, or moderate increase in treatment** 
Severe Permanent harm due to the incident*** 
Death Death due to the incident 

Incidence calculations

Cases were included in the numerator where an airway complication of anaesthesia met inclusion criteria and had been performed within the data collection period in an NHS hospital. Data were collected on events in the ICU and emergency departments but were not used in calculation of the incidence of complications during anaesthesia.

The data were entered into a Microsoft Excel 2007 spreadsheet (Microsoft Corporation, USA) and incidences were calculated (by dividing the numerator for a given group by the relevant denominator). Confidence intervals (CIs) were derived using binomial probability tests with the stat-conf programme (Handbook of Biological Statistics 2008, http://udel.edu/~mcdonald/statconf.html).

Missing reports

Although the individual case reports were anonymous, the RCoA-lead retained the date and source of individual reports. Data on the number and source hospital of events were examined for evidence of clustering by time and place in an attempt to assess the completeness of data collection. Reports from local reporters (i.e. in which the local reporter was also the anaesthetist) were identified. It was assumed that all local reporters would return all cases meeting inclusion criteria and therefore that this small highly motivated group could be used to create an upper estimate for the number of cases that might have been reported if all anaesthetists acted as local reporters did.

Results

Agreement to participate and appointment of a local reporter was established in all 309 NHS hospitals by September 2008. In total, 286 local reporters were appointed with some representing more than one hospital.

Numerator data (complications reported)

A total of 286 cases were reported to the RCoA-lead or discussed with the moderator. Seventy-nine reports were withdrawn after discussion with the moderator or the reporter reviewed the inclusion criteria sent by the RCoA-lead: 207 cases were reviewed by the review panel. During the review process, additional information, using the methods described above, was requested from the reporters of 12 of the cases. After final review, 184 reports met the inclusion criteria. Of the 184 reports, 133 complicated the management of anaesthesia, 36 occurred in patients on ICU, and 15 in the emergency department.

Capture of cases

Hospital clustering

Reports were received from 42% of hospitals and a minority of hospitals accounted for disproportionately high percentages of reported cases (Table 3). Four per cent of hospitals reported 23% of cases, 6% reported 34%, and 15% reported 59% of the cases. An analysis of the distribution of reports suggested that they did fit a Poisson distribution, consistent with complete data capture, but not confirming it.

Table 3

Clustering of cases by hospital. Analysis of 207 reviewed cases

Number of cases reported Number of hospitals Per cent of hospitals Per cent of all cases 
0.3 3.3 
0.0 0.0 
0.3 2.4 
2.9 17.2 
2.6 11.5 
26 8.5 24.9 
85 27.7 40.7 
177 57.7 0.0 
Sum 307 100 100 
Number of cases reported Number of hospitals Per cent of hospitals Per cent of all cases 
0.3 3.3 
0.0 0.0 
0.3 2.4 
2.9 17.2 
2.6 11.5 
26 8.5 24.9 
85 27.7 40.7 
177 57.7 0.0 
Sum 307 100 100 

Person clustering

Local reporters reported 19 anaesthesia-related events (i.e. the local reporter was also the anaesthetist) out of 130 where this information was provided. There were 286 local reporters and the 2007 RCoA census identified 6233 consultant anaesthetists16 (i.e. 4.6% of all consultant anaesthetists). If all consultant anaesthetists behaved as local reporters, we might anticipate 19×6233/286=414 reports from consultants. As 36% of cases occurred in the absence of a consultant, this figure for all anaesthetists might increase to 414×100/(100−36)=414×1.56=646. As this figure is based on only 130 of the 133 anaesthesia cases, our upper limit of cases is 646×133/130=661. This figure suggests that, at worst, we captured approximately one in five of relevant cases. It is likely that this figure should be adjusted further: part-time consultants account for 10% of the consultant workforce and up to one-third of departmental ‘consultant anaesthetist’ activity is delivered in ICU, pain clinics, management, and academia. Further adjustments might be made that are almost limitless and increasingly speculative, but we conclude that we may only have captured one in three or one in four cases that occurred.

Patient characteristics

There were a total of 113 males and 71 females, including 82 male and 51 female anaesthesia cases (Table 4). The majority of anaesthesia cases were ASA I or II (56%), males (62%), and age <60 yr (61%). A BMI of >30 kg m−2 or obesity was recorded in 40% and a BMI of <20 kg m−2 or cachexia in 11%. The majority (54%) of the procedures were elective or scheduled. The event occurred during normal working hours (08:01–18:00) in 69%, out of hours before midnight (18:01–24:00) in 17% and out of hours after midnight (00:01–08:00) in 14%. The anaesthesia events occurred in the operating theatre (47%), anaesthetic room (37%), and recovery unit (14%). The phase of anaesthesia was induction (52%), maintenance (20%), emergence (16%), and in the recovery phase (12%). In 63% of anaesthesia cases, the most senior anaesthetist present at the start of the event was a consultant. A locum anaesthetist was the main anaesthetist in 5% of cases. A request for help around the time of an anaesthetic airway event was recorded in 95 (70%) cases and assistance arrived without request in a further four. The time to arrival of assistance was recorded in 99 cases: 32 in <1 min, 43 in 1–4 min, 21 in 5–30 min, and three after >30 min. Of 97 identified responders, 69 were consultants in anaesthesia/intensive care medicine, 13 consultant surgeons, 11 senior anaesthesia trainees, two anaesthetic non-consultant career grades, and two surgical trainees. Of 70 requests for help made during the airway event, in 21 the response time was <1 min, in 36 was 1–4 min, in 11 was 5–30 min, and in two was >30 min: five of the 13 events with a response time >5 min occurred out of hours.

Table 4

Incident reports classified: by ASA grade and type of event; by age and type of event; and by inclusion criteria provided by the reporter. More than one inclusion criterion could be chosen. Note that some deaths were considered by the review panel not to be causally related to the event, in other cases patients reported with an inclusion criterion of brain damage either made a full recovery at the time of reporting or died. Therefore, figures in this table do not exactly match final outcomes in Table 5. *Prolongation of stay in the case of patients already in ICU

 All cases (n=184) Anaesthesia (n=133) 
ASA 
 I 26 23 
 II 62 51 
 III 59 40 
 IV 29 13 
 V 
Not recorded 
Age 
 <10 10 
 11–20 
 21–40 39 26 
 41–60 56 41 
 61–80 60 44 
 >80 10 
Not recorded 
Reporter provided inclusion criteria 
 Death 33 14 
 Brain damage 13 
 ESA 75 54 
 ICU admission* 122 100 
 Sum 243 174 
 All cases (n=184) Anaesthesia (n=133) 
ASA 
 I 26 23 
 II 62 51 
 III 59 40 
 IV 29 13 
 V 
Not recorded 
Age 
 <10 10 
 11–20 
 21–40 39 26 
 41–60 56 41 
 61–80 60 44 
 >80 10 
Not recorded 
Reporter provided inclusion criteria 
 Death 33 14 
 Brain damage 13 
 ESA 75 54 
 ICU admission* 122 100 
 Sum 243 174 
Table 5

Final outcome: narrative outcome and NPSA classification (Table 2)

 All cases (n=184) Anaesthesia (n=133) 
Final outcome (narrative) 
 Death 38 16 
 Brain damage 
 Other partial recovery 10 
 Full recovery 124 106 
 Unrelated death 
Final outcome (NPSA definitions) 
 Death 38 16 
 Severe 10 
 Moderate 126 103 
 Low 
 None 
 All cases (n=184) Anaesthesia (n=133) 
Final outcome (narrative) 
 Death 38 16 
 Brain damage 
 Other partial recovery 10 
 Full recovery 124 106 
 Unrelated death 
Final outcome (NPSA definitions) 
 Death 38 16 
 Severe 10 
 Moderate 126 103 
 Low 
 None 

Inclusion criteria and event outcomes

Death

Death resulting from an airway problem was the inclusion criterion for 33 reports (Table 5), of which 14 occurred during anaesthesia, 16 in ICU, and three in the emergency department (Table 4). In 10 further cases, the reporter indicated a lower severity inclusion criterion but also that the patient died before the report was submitted. Of these 10 ‘late deaths', the airway event was judged causal in three, contributory in two, and unrelated in five. In total, there were therefore 38 deaths attributable to an airway event: 16 during anaesthesia, 18 on ICU, and four in the emergency department. Hypoxia was the common theme in deaths caused by an airway problem, though in several late deaths, sepsis and single or multi-organ failure was recorded. Death rate for all cases was 38/184 (20.7%) and for events during anaesthesia 16/133 (12.0%).

Brain damage

In 13 patients, brain damage was provided as an inclusion criterion (Table 5), and three other cases were identified during case review. Six of these patients died and two made a full recovery (e.g. post-event fitting or depressed level of consciousness that fully resolved). Eight cases of persistent non-fatal brain damage were identified: three events occurred during anaesthesia, four in ICU, and one in the emergency department. Reported outcomes included permanent low conscious level, neuro-behavioural deficit, or ‘persistent vegetative state’ (recorded after 1 month, although it would require a year to elapse before this diagnosis could be made). The combined rate of death and brain damage for all cases was 46/184 (25.0%) and for events during anaesthesia 19/133 (14.3%).

Emergency surgical airway

An attempt at emergency surgical airway was reported in 80 of 184 reported cases (43%) with only 75 being recorded as indications for inclusion. An emergency surgical airway was attempted in 58 (43%) of the 133 anaesthesia-related reports.

In 29 anaesthesia cases, the first choice for emergency surgical airway was tracheostomy: 18 in semi-controlled circumstances where intubation had failed or not been attempted, but the airway could be maintained on a facemask or laryngeal mask and in 11 cases as a true emergency rescue technique for a patient in extremis. All emergency tracheostomies were successful, although not always without difficulty or delay. Two patients in this group died, one because the tracheostomy was not able to bypass a low-lying obstructing tracheal tumour and one died later due to severe hypoxia occurring before the tracheostomy was performed. Cricothyroidotomy was the first approach in 29 cases: 19 with a narrow-bore (≤2 mm) cannula, seven with a wide-bore cannula, and three with a surgical approach. Twelve of 19 narrow-bore cannula cricothyroidotomy failed with rescue achieved by surgical tracheostomy in seven, surgical cricothyroidotomy in two, wide-bore cannula in one, and successful oral intubation in two. Three out of seven wide-bore cannulae failed and were rescued with tracheostomy, surgical cricothyroidotomy, or tracheal intubation. The three first-choice surgical cricothyroidotomies were all successful. Of 58 attempts at emergency surgical airway, nine (16%) failed to rescue the airway: 51 (88%) patients made a full recovery from the incident, three (5%) a partial recovery, and four (7%) died: two after successful surgical airway and two after failure.

Of the 58 cases requiring emergency surgical airway, this was performed by a surgeon in 33 cases (mostly head and neck surgeons during relevant cases) and by an anaesthetist in 25. Only nine of these 25 anaesthetic attempts were successful in rescuing the airway; 11 failures were rescued by a surgeon-performed tracheostomy, one by percutaneous tracheostomy placed by a colleague, three by tracheal intubation, and one patient died.

ICU admission

ICU admission (or prolongation of stay) was reported as an inclusion criterion in 122 cases, including 100 patients following an airway event during anaesthesia. Reported indications for admission to ICU following anaesthesia-related events were to manage airway swelling or trauma in 38 patients, aspiration of gastric contents or blood in 32, hypoxia due to post-obstructive pulmonary oedema in 13, failure to awaken after surgery in 13, or myocardial ischaemia or cardiac arrest in four. Of the 100 admitted to ICU after an anaesthesia-related airway event, 12 died, seven made a partial recovery, and 81 were reported to have made a full recovery. Of the 29 patients admitted to ICU with aspiration of gastric contents, aspiration during anaesthesia was the primary airway event in 23, while in six it complicated another primary event: eight of these patients died and two suffered brain damage.

Primary airway problem during anaesthesia

Problems with tracheal intubation were the most frequently recorded primary airway problem (Fig. 1). Difficult or delayed intubation, failed intubation, and ‘can't intubate can't ventilate’ (CICV) accounted for 39% of all events and events during anaesthesia. Aspiration then extubation problems followed tracheal intubation in frequency of reported complications. For anaesthesia events, aspiration, CICV, and problems during use of a supraglottic airway, iatrogenic airway trauma, and failed mask ventilation were the next most prominent complications.

Fig 1

Primary airway problem

Fig 1

Primary airway problem

Primary airway device during anaesthesia

For anaesthesia events, the airway in use or intended for maintenance was: tracheal tube of any sort (91), supraglottic airway device (35), and facemask (7) (Table 6).

Table 6

Primary airway used or intended for maintenance of anaesthesia

Airway  
Tracheal intubation (including fibreoptic intubation) 82 
Laryngeal mask airway 32 
Hudson mask/nasal cannulae 
Rigid bronchoscopy 
Another supraglottic device 
Anaesthetic facemask±oropharyngeal airway 
Tracheostomy 
New tracheostomy or cricothyroidotomy 
Total 133 
Airway  
Tracheal intubation (including fibreoptic intubation) 82 
Laryngeal mask airway 32 
Hudson mask/nasal cannulae 
Rigid bronchoscopy 
Another supraglottic device 
Anaesthetic facemask±oropharyngeal airway 
Tracheostomy 
New tracheostomy or cricothyroidotomy 
Total 133 

Incidence of incidents

The total number of events reported in relation to anaesthesia was 133. The number of anaesthetics administered in the same period derived from the census phase of NAP4 was 2.9 million (2872600),12 giving a minimum incidence (point estimate) of 133/2872600: i.e. 46 per million or approximately one per 22 000 general anaesthetics. Using binomial statistics, we can estimate an upper 95% confidence limit of 54 per million and a lower CI of 38 per million (although as the actual event rate in our population cannot be lower than that we observed, some might omit this value).

Using the same methodology, we can calculate the point estimate and CIs for incidence of death (or death and brain damage) from an airway event during general anaesthesia (Table 7). The census data also provided estimates of frequency of use of airway devices (tracheal tube, supraglottic airway device, and facemask) and estimates of the risk of events and poor outcomes with these devices can be derived (Table 7).

Table 7

Incidence estimates of major airway complications by airway type for events and death/brain damage: expressed as events per million cases and fractions (one in n cases). The denominator for each calculation is from the Fourth National Audit project Census.15 For each, point estimate and lower and upper confidence limits (CL) are presented

Type of event Numerator Denominator Events per million cases
 
Events as fractions one in n cases
 
   Point estimate Lower CL Upper CL Point estimate Lower CL Upper CL 
Events 133 2 872 600 46.3 38.4 54.2 21 598 26 021 18 461 
Deaths 16 2 872 600 5.6 2.8 8.3 179 538 352 033 120 495 
Death/brain damage 19 2 872 600 6.6 3.6 9.6 151 189 274 717 104 294 
Tracheal tube events 91 1 102 900 82.5 65.6 99.5 12 120 15 254 10 054 
Tracheal tube death/brain damage 10 1 102 900 9.1 3.4 14.7 110 290 290 087 68 089 
SAD events 35 1 616 100 21.7 14.5 28.8 46 174 69 051 34 684 
SAD death/brain damage 1 616 100 5.0 1.5 8.4 202 013 657 942 119 325 
FM event 154 200 45.4 11.8 79.0 22 029 84 985 12 654 
FM death/brain damage 154 200 6.5 0.0 19.2 154 200 52 095 
Type of event Numerator Denominator Events per million cases
 
Events as fractions one in n cases
 
   Point estimate Lower CL Upper CL Point estimate Lower CL Upper CL 
Events 133 2 872 600 46.3 38.4 54.2 21 598 26 021 18 461 
Deaths 16 2 872 600 5.6 2.8 8.3 179 538 352 033 120 495 
Death/brain damage 19 2 872 600 6.6 3.6 9.6 151 189 274 717 104 294 
Tracheal tube events 91 1 102 900 82.5 65.6 99.5 12 120 15 254 10 054 
Tracheal tube death/brain damage 10 1 102 900 9.1 3.4 14.7 110 290 290 087 68 089 
SAD events 35 1 616 100 21.7 14.5 28.8 46 174 69 051 34 684 
SAD death/brain damage 1 616 100 5.0 1.5 8.4 202 013 657 942 119 325 
FM event 154 200 45.4 11.8 79.0 22 029 84 985 12 654 
FM death/brain damage 154 200 6.5 0.0 19.2 154 200 52 095 

Case-mix

Aspiration of gastric contents

Aspiration of gastric contents was the primary event in 23 anaesthesia cases, two emergency department cases, and no ICU cases. It was the most common cause of death in the anaesthesia group accounting for eight deaths and two cases of brain damage. Aspiration occurred most frequently in patients with risk factors (>90%), at induction of anaesthesia or during airway instrumentation (61%). Planned airway management was as follows: laryngeal mask 13, i-gel 1, tracheal tube 8, and none 1. Aspiration occurred before airway instrumentation in five cases and during airway placement in two. Two cases had clear indications for rapid sequence induction (RSI) and in several others, its use could be argued, one case occurred during RSI laryngoscopy. Management of the cases was judged good in four, mixed in seven, and poor in eight, with management judged poor in four deaths. Aspiration also complicated other primary events (secondary aspiration), most frequently difficult or failed intubation. There were six such events in anaesthesia cases. Aspiration of blood was the primary event in five anaesthesia cases, one of which led to death.

Head and neck cases

Seventy-two reported cases (39%) involved an airway problem in association with an acute or chronic disease process in the head, neck, or trachea. Approximately 70% of these reports were associated with obstructive lesions within the airway. The qualifying airway event was death or brain damage in 13 cases, emergency surgical airway in 50, and unexpected ICU admission in 27. The outcome at the time of form completion (if recorded) was death in 17, partial recovery in two, and full recovery in 51 cases. These cases included 55 anaesthesia cases. Forty-two involved anaesthesia for diagnostic or resection surgery, with problems occurring at induction in 21 cases, during maintenance in eight and during extubation or recovery in 13. In 10 patients, complications arose during induction of anaesthesia primarily to secure a critical airway. Three complications were reported in patients after elective head and neck surgery, who returned to theatre from wards for urgent reoperation. The reviewers assessed airway management as poor in nearly one-third of reported cases. Issues of assessment, planning, and communication within teams were prominent in these cases.

Obstetrics

There were four reported events in pregnant women: all involved emergency Caesarean section and problems at the time of intubation. All took place out of hours and involved complex patients (two of whom had a BMI >35 kg m−2) and were managed by senior anaesthetists: in two, a consultant was present throughout; in one, a staff grade; and in one, a year 6 specialist trainee. Consultants attended in all cases. Two cases occurred during an operation where anaesthesia was induced for failed regional anaesthesia. One patient had a secondary aspiration (i.e. aspiration complicated another primary airway event), one had a failed cricothyroidotomy attempt, and one a successful surgical airway. All were admitted to ICU and made a full recovery.

Paediatrics

There were 10 events in children under the age of 10 yr: eight during anaesthesia, and one each in ICU and in the emergency department. Five cases were infants and nine were children aged <4. Outcomes included three deaths. Of the eight anaesthetic complications, there were four cases of difficult intubation (two due to subglottic narrowing), two aspirations (one of blood after tonsillectomy), one due to tracheal tube blockage by secretions, and one patient required an emergency tracheostomy during the removal of a foreign body. One child died, one had persistent stridor, and six recovered fully. All patients were anaesthetized in the presence of a consultant. The review panel considered airway management to be good in two cases, mixed in four cases, poor in one, and had inadequate information to comment in one case.

Obesity

Seventy-seven of 184 patients (42%) were obese; of whom, 19 (25%) suffered death or brain damage, the same rate as the non-obese population. Of 53 events during anaesthesia in obese patients, four resulted in death and one persistent neurological deficit: a rate of 9%, lower than the rate in non-obese anaesthesia cases, 18%.

In anaesthesia cases, some form of airway assessment was recorded in 36 and difficulty was anticipated in 25. The proportion of primary airway problems related to tracheal intubation was similar in obese and non-obese patients (23 of 53 vs 33 of 80). Eight reports described aspiration, seven extubation problems, and four airway trauma. Airway management was assessed as good in 12 cases, mixed in 23, poor in 15, and unassessable in three. The most frequently cited causal or contributory factors were patient in 42 cases, judgement in 29, and education/training in 20. Several patients experienced complications of airway management during general anaesthesia when regional anaesthesia would have been a suitable alternative for surgery, but of note five obese patients also developed airway complications after requiring general anaesthesia when a regional anaesthetic technique or sedation failed: a situation observed in only one non-obese patient.

Events at the end of anaesthesia and in recovery

There were 38 events at the end of anaesthesia or during the recovery period; 20 in the operating theatre, 16 in the recovery room, and two occurred in transit. Airway obstruction was the most common problem: causes included laryngospasm, complete occlusion of an airway device by patient biting, blood in the airway or airway swelling (in three patients, this followed surgery in the Trendelenburg position). Diagnosis of airway obstruction was not always prompt, particularly in recovery. Two patients died following events occurring in the recovery room. In one case, an inhaled blood clot after tonsillectomy produced total tracheal obstruction which was initially attributed to asthma and led to fatal cardiac arrest. In the other, airway obstruction resulted in pulmonary oedema and severe hypoxia requiring cardiopulmonary resuscitation (CPR). The patient subsequently died in ICU. In total, five patients developed severe hypoxia requiring CPR. Negative pressure pulmonary oedema was seen frequently after these obstructive events and required admission to ICU in 13 cases, 12 of whom made a full recovery. Several cases of laryngeal mask occlusion were deemed preventable by the use of a bite block. Sixteen of the 38 events followed surgery within the airway and in this group, the reviewers noted evidence of poor anticipation and planning for management after extubation in the face of known problems.

Capnography and monitoring

Monitoring was used in all anaesthesia cases. In contrast to cases reported from the ICU and emergency departments, capnography appeared to be used universally for intubation and in the operating theatre. Reviewers judged that the use of capnography in the recovery area (and its appropriate interpretation) would have led to earlier identification of airway obstruction in several cases. There were three anaesthesia-related cases, including two deaths in which optimal interpretation of capnography might have altered the clinical course. In one case, described above, prolonged airway obstruction in recovery due to an aspirated blood clot was diagnosed as asthma for an extended period. It was not stated whether capnography was used. In the second case, laryngeal mask misplacement in an ASA II patient led to severe hypoxia; intubation was performed while the patient was peri-arrest. Intubation was difficult, as was ventilation and the capnograph showed ‘minimal CO2’. Capnography was ‘flat’ during prolonged cardiac arrest and this appeared to be a case of unrecognized oesophageal intubation. In the third case, a healthy patient was intubated and transferred into theatre but became hypoxic with a flat capnography trace. Anaphylaxis was suspected but senior anaesthetic help promptly diagnosed the tracheal tube in the oesophagus: the patient was transferred to ICU and made a full recovery. In total there were three cases of unrecognized oesophageal intubation during anaesthesia leading to one death and one case of brain damage.

Review panel analysis

Degree of harm

The review panel ascribed outcomes to all 184 cases (Table 5).

Causal, contributory and positive aspects of care

All reports were assessed to identify causal and contributory factors (Table 8). Of all 184 cases, the most frequent causal and contributory factors were the patient (77% of cases), followed by judgement (59%) and education/training (49%). Equipment/resource and communication factors were causal or contributory in more than one-quarter of cases. Medication and work/environment were the least frequently cited factors. Positive factors were identified in 91 cases (49%): the most frequent positive factors being communication (22% of cases) and organization/strategic (19%).

Table 8

Factors assessed by review panel to contribute or cause events and factors indicating good practice. For definitions of factors listed, see Table 2

Factors ALL cases (n=184)
 
Anaesthesia (n=133)
 
 Causal Contributory Positive Causal Contributory Positive 
Communication 38 40 26 20 
Education and training 12 77 17 10 52 13 
Equipment and resources 46 21 30 16 
Medicines 31 21 
Organization and strategic 42 35 35 28 
Patient 37 103 28 76 
Task 31 22 
Team and social 36 22 26 20 
Work and environment 14 
Judgement 19 90 23 16 67 18 
Other 
Factors ALL cases (n=184)
 
Anaesthesia (n=133)
 
 Causal Contributory Positive Causal Contributory Positive 
Communication 38 40 26 20 
Education and training 12 77 17 10 52 13 
Equipment and resources 46 21 30 16 
Medicines 31 21 
Organization and strategic 42 35 35 28 
Patient 37 103 28 76 
Task 31 22 
Team and social 36 22 26 20 
Work and environment 14 
Judgement 19 90 23 16 67 18 
Other 

In the anaesthesia-related cases, similar patterns were observed (Table 8). The patient was considered causal in one-fifth of cases and causal or contributory factors included patient (79% of cases), followed by judgement (62%) and education/training (47%). Organization/strategic factors were also causal or contributory in more than one-quarter of cases. Positive factors were identified in 65 cases (49%): the most frequent positive factors were organization/strategic (21% of cases) and team/social and communication (each 15%).

Quality of airway management conduct

Of 184 airway events, the review panel assessed the airway management as good in 16% cases, mixed in 43%, and poor in 35% (9). In only three of 46 events leading to death or brain damage, did the reviewers assess airway management as good and in 25 (54%), it was assessed as poor.

Of 133 airway events during anaesthesia, airway management was assessed as good in 18% cases, mixed in 41%, and poor in 34% (Table 9).

Table 9

Reviewers’ assessment of quality of airway management and degree of harm. Mixed refers to an assessment of both good and poor elements

Clinical area Airway management
 
 Good Mixed Poor Not classified Sum 
Anaesthesia (n=133) 24 55 45 133 
Anaesthesia death (n=16) 16 
Anaesthesia death and brain damage (n=19) 10 19 
All (n=184) 30 79 65 10 184 
All deaths (n=33) 14 20 38 
All death and brain damage (n=46) 16 25 46 
Clinical area Airway management
 
 Good Mixed Poor Not classified Sum 
Anaesthesia (n=133) 24 55 45 133 
Anaesthesia death (n=16) 16 
Anaesthesia death and brain damage (n=19) 10 19 
All (n=184) 30 79 65 10 184 
All deaths (n=33) 14 20 38 
All death and brain damage (n=46) 16 25 46 

Discussion

This is the first prospective study of all major airway events occurring throughout the UK during anaesthesia, in ICU and the emergency department. It has identified a cohort of patients, a minimum prevalence, and enabled calculation of a minimum incidence of such events. This paper focuses on quantitative data relating to events during anaesthesia collected during the project. Combined with data from the matched anaesthesia census,12 we are able to estimate an incidence of such complications occurring during anaesthesia. The incidence calculations have limitations and these are discussed below. Of equal importance, the project enables comparisons between rates of major complications when different airways (tracheal tube, supraglottic airway device, facemask) are used for anaesthesia. Finally, and perhaps most importantly, the project offers the opportunity to learn from review of a large series of such sentinel events and analysis of emerging themes. A complete report of this project with expanded clinical details and analysis to identify clinical learning points and recommendations has been compiled and this will be made available on the RCoA website (http://www.rcoa.ac.uk/index.asp?PageID=1089). A detailed analysis of events which occurred in ICUs and in emergency departments is presented in an accompanying paper.10

While the ideal solution for identifying the incidence of rare complications is a continuous process of notification of critical incidents and their analysis, this is currently impracticable. Alternatives require study of a very large population or a prolonged period of assessment. The current project has observed complications in the whole of the UK over a period of 1 yr. A similar study of deaths related to airway complications performed in France during 199917 analysed death certificates to identify cases, a questionnaire was then sent to the certifiers. In the USA, Li and colleagues18 collected reports by using the International Classification of Diseases (ICD-10) codes to identify anaesthesia-related complications. Deficiencies with death certification in the UK have been highlighted previously in the earliest confidential enquiry into perioperative deaths and problems remain.19 The use of death certification is retrospective, identifies mortality but not morbidity, relies on accurate certification data, and analysis of individual cases is problematic. In this project, we chose a prospective methodology with a system of local reporters to identify cases. This enabled us to identify those cases that we believe most would classify as major complications, even when the degree of harm was temporary. In addition to the NPSA classification of severity, we also assessed frequency of death and death/brain damage as this is clinically relevant and is the outcome used by several litigation-based-analyses.3,4

This study identified 33 deaths and 46 cases of death or brain damage as a result of airway complications during anaesthesia, in ICU and the emergency department over a 1 yr period. We calculate the incidence of serious airway complications during general anaesthesia to be (at least) 133 per 2.9 million or one per 22 000 general anaesthetics, death and brain damage (at least) one in 180 000 anaesthetics, ICU admission (at least) one in 29 000, and emergency surgical airway (at least) one in 50 000 general anaesthetics. Since the reports represent a timed sample, it is possible that the true incidence could be higher or lower than this figure; therefore, 95% confidence limits are provided (Table 7).

An important finding is the relative frequency of major airway events occurring with different airway devices. Comparisons between these groups are likely to be robust as reporting rates are likely to be equal. Categorizing devices as broadly as possible, it is notable that while airway events are more frequent during anaesthesia with a tracheal tube (point estimate 83 per million) than with, for instance, a supraglottic airway device (22 per million), the range of incidences is not extreme and this is even more evident if only deaths and brain damage are included: tracheal tube 9.1 per million, facemask 6.6 per million, supraglottic airway 5 per million. It is not surprising that events are more frequent for tracheal tubes as these cases include the vast majority of higher risk cases and also the group includes intrinsically more complicated techniques (e.g. tracheostomies, trans-tracheal ventilation, etc.). While some might argue that the rates of complications of the simpler techniques should be considerably lower, the fact that we have not demonstrated markedly higher rates of the most severe outcomes in one particular group is reassuring in terms of the airway techniques chosen ‘en masse’ in UK anaesthetic practice.

Aspiration was the single most common primary cause of fatality (primary event in 50% of deaths) in anaesthesia events. Aspiration is the cause of litigation in about 10–15% of anaesthesia airway-related claims in America20 and the UK3 and of about one-third of cases where litigation is related to death. In the French study, aspiration was the cause of death in 83 of 131 deaths (63%).17 While the absolute incidence of such events is rare, these data emphasize the importance of aspiration as a major contributor to airway-related morbidity and mortality in anaesthetic practice. Case review identified several cases where airway management was with a laryngeal mask, despite clear evidence of risk factors for aspiration and also cases where RSI was not performed in patients with bowel obstruction. Various strategies are available to reduce the risk of aspiration in low- and high-risk patients: in NAP4 some deaths occurred without these precautions being used.

Approximately 42% of anaesthesia events reported had a primary airway event indication intubation difficulty. Many of these cases involved patients with head and neck cancer and airway obstruction, with emergency surgical airway being necessary in 43% of anaesthesia cases. Poor planning of airway strategies and failure to change routine plans despite evidence of likely difficulty or when that plan failed were identified problems. In both the French study17 and this project, 13% of airway deaths were associated with difficult tracheal intubation. Put another way, 87% of deaths were not associated with difficult intubation. The French study's point estimate for deaths related to difficult intubation is 21 per million with a very wide CI of 3–77. In the US study18 failed, difficult intubation or wrongly placed tracheal tubes accounted for 2.3% of all anaesthesia-related deaths. As the majority of airway events occurred in elective surgery, in ASA I–II patients aged <60, this project acts as a reminder that a major airway complication can occur during complex and also apparently ‘straightforward’ routine anaesthesia.

When emergency surgical airway was required, this was performed most frequently by head and neck surgeons performing a rescue tracheostomy, all of which were successful. Cricothyroidotomy was the rescue technique of choice for anaesthetists but ∼65% of these attempts failed to secure the airway. As two-thirds of emergency tracheostomies were performed in semi-controlled conditions, the cricothyroidotomies likely did represent a greater proportion of ‘in extremis’ cases. As NAP4 studied events with poor outcomes, it is possible that a disproportionate number of successful rescue cannula cricothyroidotomies were not reported. Even accepting these caveats, the high failure rate of this technique is a cause for concern. Whether this is due to failures of training, use of inappropriate equipment, equipment design problems, or technical failures during use requires further exploration and research. Anaesthetists might usefully study this area and ensure their competence with both cannula and surgical techniques.

Forty-two per cent of all patients notified to NAP4 were obese and 11% cachectic. The incidence of adult obesity in the UK in 2008 was reported to be 24.5%,21 and although we do not know the incidence of obesity or cachexia in the surgical population both groups are likely over-represented. An excess of cachectic patients is accounted for by a significant number of events occurring in patients with recurrent (sometimes pre-terminal) head and neck cancers. In contrast, the excess of obese patients underscores the fact that obese patients are at increased risk of an adverse airway event. Reasons for this include mechanical difficulty in securing the airway (mask ventilation,22 tracheal intubation,23 and emergency surgical airway), increased risk of aspiration, increased risk of airway obstruction during difficulty, and accelerated speed and extent of oxygen desaturation during airway obstruction.24 Of the 53 anaesthesia-related cases reported, mechanisms of injury and outcomes were notably similar to the non-obese reports. The fact that airway events occurred in obese patients who might have had their surgery performed under regional anaesthesia, but also after attempted regional anaesthesia or sedation failed, illustrates that these patients are a major challenge for all anaesthetic techniques and anaesthetists. In view of the trends in population obesity in developed countries, the number of patients at risk of such events due to obesity is almost certain to increase.

It was notable that events occurred at all phases of the anaesthetic process. While induction was the phase when most (52%) events occurred, a significant minority occurred during emergence (16%) and in (or during transfer to) the recovery area (14%). The latter phase being particularly dangerous as the anaesthetist may be neither present nor immediately available to respond to an emergency.

In the cases of tracheal obstruction or tube misplacement, capnography and correct interpretation may have led to a change in clinical management and outcome. Each of the cases serves to remind us that the absence of expired carbon dioxide indicates lack of ventilation. When this occurs in an intubated patient, even during cardiac arrest, the possibility of tracheal tube occlusion, tracheal obstruction, or oesophageal intubation must be excluded before treating other causes. The capnograph trace is not flat in a correctly intubated patient during CPR and this is discussed in depth in the companion paper.10

Cases of high airway pressure and ineffective ventilation with inadequate capnograph trace were erroneously attributed to asthma or anaphylaxis. Endoscopic examination of the tracheal tube would have assisted earlier diagnosis of intraluminal obstruction or oesophageal intubation.

The AAGBI recently published a statement recommending that ‘Continuous capnography should be used in the following patients, regardless of location within the hospital: Those whose tracheas are intubated and those whose airways are being maintained with supraglottic or other similar airway devices'.25

The statement specifically includes recovery rooms. Capnography in recovery would likely have mitigated several events reported to NAP4. Other potential methods of improving diagnosis of airway obstruction in recovery include nursing education, observation of ‘t-bag’ movement to monitor respiration, and the presence of an anaesthetist in the recovery area.

Analysis of reviewer's opinions indicates that intrinsic patient features contributed to the airway event in more than three-quarters of anaesthesia events. The most common extrinsic contributory factors were judgement and training. After excluding the patient as a contributory/causal factor, the ratio of contributory/causal factors to positive factors was ∼2.5 for all cases and for anaesthesia cases. This reinforces the finding that reviewers assessed airway management to have elements that were poor in three-quarters of anaesthesia events and in more than 80% of deaths. A caveat is that the NAP4 process was good at identifying procedural and narrative events but was not, because of its design, suited for in-depth analysis of human factors. Despite this, and limitations described below, the assessment was that in many cases better planning, better knowledge, better judgement, or better communication, among other factors, would likely have mitigated the events or even prevented some. Among the human factors most frequently identified were elements of poor communication, poor teamwork, poor leadership, and task fixation.

There are numerous positive aspects to the findings in this report and space only allows a brief comment. Perhaps most important is that all UK NHS hospitals took part and individual anaesthetists were willing to report these high impact events. It is also notable that most anaesthesia cases were managed in the presence of a consultant anaesthetist and often by several senior anaesthetists working together. When problems arose a call for assistance was usual (73%), the person responding to the request was a consultant in 85% of cases, and assistance arrived in <4 min in 79% of cases. These findings suggest that appropriately senior anaesthetists manage many difficult cases and that anaesthetic departments in UK NHS hospitals generally have a culture of colleague assistance and structures that enable prompt assistance in the event of a crisis. This is reinforced by the reviewers’ analysis of cases which indicated that the factors most commonly identified as ‘positive’ in anaesthesia cases were organization/strategic followed by communication and team/social. This report has necessarily focused on deaths and brain damage but each of the non-fatal cases reported to NAP4 can be considered a near death. The 133 reports of events during anaesthesia may well be a significant underestimate. As more than one anaesthetist is generally involved in each case, as many as 1000 anaesthetists may be involved with such events each year (approximately one event for a consultant every 6 yr). It is a tribute to the specialty that so few patients came to serious harm and few died, but these were still very serious events and to individual anaesthetists these will probably be events that they will never forget.

One of the aims of this project was to determine the incidence of major complications of airway management in anaesthesia. This has been challenging, both in determining an accurate denominator and in establishing a numerator, because we know there will have been cases that were not reported. We identified 133 major events including 16 deaths and three cases of brain damage related to airway complication of anaesthesia. Accepting the limitations, we are able to calculate a point estimate of this incidence and a CI surrounding it. Our estimate is of 46 events per million (95% CI 38–54) and with 12% of these leading to death, a fatality rate of 5.6 per million (CI 2.8–8.3). The French study identified ‘airway deaths’ of 20 per million (CI 7–36), and while these confidence limits overlap, they are wide and suggest a higher rate of complications than the current study.17

Limitations

The project has several limitations. It is likely that not all cases were reported but we cannot know how many, or indeed if any were missed. We tried to maximize reporting but acknowledge that many factors may have contributed to under-reporting. There may be a personal or organizational reluctance to release information if there is an ongoing investigation or if litigation is anticipated. Cases took up to a year after the event to be fully reported. Our analyses of reporting patterns by institution and by time are compatible with complete reporting but do not guarantee it. Our incidence calculations are based on reported cases; however, statistical advice and analysis indicated the true incidence may be up to four-fold higher. In this project, aspiration of gastric contents was the cause of death in eight patients giving an incidence of 1 in 360 000 anaesthetics (95% CI 1 in 212 000–1.1 million). Other large studies have reported rates of fatal aspiration associated with anaesthesia from 1 in 45 00026 to 1 in 240 00027 with one study identifying no cases in 198 000 paediatric anaesthetics.28 These data suggest under-reporting to the NAP4 project, but cannot confirm or quantify it. Comparisons between NAP4 data and those from studies performed in other countries, several decades ago, with different methodology should be treated with caution.

We are not aware of any better estimates of anaesthesia airway-related morbidity by other researchers. As we recruited local reporters in 100% of NHS hospitals in the UK and all local reporters returned data to the project, we believe our effort approaches the best achievable with current methods. Our explicit description of how many cases we estimate may have been missed enables readers to interpret the data in the knowledge of these limitations.

There were several cases where the decision to include or exclude was not clear-cut. One case of fatal aspiration which occurred while an anaesthetist who had sedated a patient performed a spinal anaesthetic was excluded; the level of sedation was unknown and the primary aim of the project was not to study complications of sedation. In contrast, two cases that initially took place under local anaesthesia or sedation were included. In one, an anaesthetist administered sedation for endoscopy including oesophageal and pyloric dilation before aspiration occurred, the patient died. In the other, tonsillar biopsy under local anaesthesia with ‘deep sedation’ was complicated by profuse bleeding. The anaesthetist attempted to rescue the airway but intubation failed and an emergency airway was required, this patient made a full recovery. These cases likely fall under the umbrella of ‘managed anaesthesia care’. They were considered to be consistent with the sorts of cases the project was designed to study.

A final limitation is inherent when expert panel review is used to ‘judge cases’. We relied on submitted questionnaires and did not have access to case-notes nor the facility to speak to the clinicians involved. Despite this, we believe that our review process was robust. It can be summarized as a structured implicit review performed in teams. Pitfalls of retrospective case review include variation in reviewer opinion, outcome bias,14 hindsight bias,15 and ‘consensus bias’. The latter bias occurs because teams reviewing cases often reach internal agreement but disagree with other teams.29 While it is impossible to overcome all these biases, we made the following efforts to do so. The review panel was educated in hindsight and outcome bias and at each meeting the reviewers were reminded of these biases, definitions of which appeared on the sheets categorizing outcomes. Each case was reviewed by two teams enabling an exploration of ‘between group disagreement’ to balance the tendency for ‘within group agreement’. Guidelines and recommendations published by other organizations were used in the review process where considered appropriate. When judging case conduct against guidelines, the review panel attempted to ensure they were applicable, based on high-quality evidence, up-to-date and specific to the individual case.

In conclusion, airway management during anaesthesia is associated with serious complications, but these are rare. Optimistically, the incidence of complications resulting in death is 16 in 2.9 million, an incidence of one death per 180 000 general anaesthetics. Pessimistically, based on the assumptions discussed if only 25% of reports have been received, this figure could increase to one death per 45 000 general anaesthetics.

Important findings related to anaesthesia cases in this project include: (i) more than half of the patients were male, ASA I–II, aged <60, and most events occurred during elective surgery under the care of anaesthetic consultants. (ii) Aspiration was the most frequent cause of anaesthesia airway-related mortality. (iii) Obese patients were disproportionately represented. (iv) Obstructing airway lesions generated a large number of complications, many reports showed evidence of poor planning of primary and rescue techniques. (v) Cricothyroidotomy by anaesthetists was associated with a high rate of failure. (vi) One in four events occurred at the end of anaesthesia or in the early recovery room. (vii) Omission or incorrect interpretation of capnography led to undiagnosed oesophageal intubation. (viii) Elements of poor management were observed in the majority of airway complications and most deaths.

Detailed analysis of the reports of individual airway events during anaesthesia will contribute to our understanding of events causing patient harm and should enable improvements in the quality of care delivered.

Supplementary material

Supplementary material is available at British Journal of Anaesthesia online.

Conflict of interest

T.M.C. has been paid by Intavent Orthofix and the LMA Company (manufacturers of laryngeal mask airways) for lecturing. He has never had and has no financial interest in these or any anaesthetic equipment companies.

Funding

The project was funded by three organizations: the Royal College of Anaesthetists, the Difficult Airway Society, and the National Patient Safety Agency.

Acknowledgements

The authors would like to express their thanks to the network of local reporters who were responsible for collecting and supplying data. Their role was difficult and demanding, this report would not have been possible without their hard work, persistence, and diligence. In addition to the Royal College of Anaesthetists and the Difficult Airway Society, a number of organizations contributed to the development of the project in various ways, including being represented on its Working Group. These include the Association for Peri-operative Practice, Association of Anaesthetists of Great Britain and Ireland, Association of Paediatric Anaesthetists, College of Emergency Medicine, College of Operating Department Practitioners, Intensive Care Society, National Patient Safety Agency, Obstetric Anaesthetists Association. We are also indebted to the President, Council, and the Head of Professional Standards (Mr Charlie McLaughlan) at the Royal College of Anaesthetists. We would also like to acknowledge the advice of Mrs Karen Thomson, Patient Information Advisory Group at the Department of Health. Dr Ian Calder performed an essential role by acting as a moderator outside of the running of the project. He used his extensive background knowledge and understanding of airway management and of the project to aid the local reporters in discussions about inclusion criteria. We thank the following statisticians for advice: Rosemary Greenwood and Hazel Taylor (Research Design Service, South West, Bristol), Gordon Taylor (University of Bath). Finally, we are particularly indebted to Ms Shirani Nadarajah at the Royal College of Anaesthetists for her major contribution to the prompt collection and accurate collation of census data and for project administration. The project was also endorsed/supported by the Chief Medical Officers of England (Sir Liam Donaldson), Northern Ireland (Dr Michael McBride), Scotland (Dr Harry Burns), and Wales (Dr Tony Jewell), the Medical Protection Society, and Medical Defence Union.

References

1
Utting
JE
Pitfalls in anaesthetic practice
Anaesthesia
 , 
1987
, vol. 
59
 (pg. 
877
-
90
)
2
Gannon
K
Mortality associated with anaesthesia. A case review study
Anaesthesia
 , 
1991
, vol. 
46
 (pg. 
962
-
6
)
3
Cook
TM
Scott
S
Mihai
R
Litigation related to airway and respiratory complications of anaesthesia: an analysis of claims against the NHS in England 1995–2007
Anaesthesia
 , 
2010
, vol. 
65
 (pg. 
556
-
63
)
4
Caplan
RA
Posner
KL
Ward
RJ
Cheney
FW
Adverse respiratory events in anesthesia: a closed claims analysis
Anesthesiology
 , 
1990
, vol. 
72
 (pg. 
828
-
33
)
5
Gene
N
Peterson
GN
Domino
KB
Lee
MLA
Cheney
FW
Management of the difficult airway. A closed claims analysis
Anesthesiology
 , 
2005
, vol. 
103
 (pg. 
33
-
9
)
6
Special Article. Practice guidelines for management of the difficult airway. An updated report by the American Society of Anesthesiologists Task Force on the Management of the Difficult Airway
Anesthesiology
 , 
2003
, vol. 
98
 (pg. 
1269
-
77
)
7
Brennan
TA
Leape
LL
Laird
NM
, et al.  . 
Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I
New Engl J Med
 , 
1991
, vol. 
324
 (pg. 
370
-
6
)
8
Studdert
DM
Thomas
EJ
Burstin
HR
Zbar
BI
Orav
EJ
Brennan
TA
Negligent care and malpractice claiming behavior in Utah and Colorado
Med Care
 , 
2000
, vol. 
38
 (pg. 
250
-
60
)
9
Smith
AF
Mahajan
RP
National critical incident reporting: improving patient safety
Br J Anaesth
 , 
2009
, vol. 
103
 (pg. 
623
-
5
)
10
Cook
TM
Woodall
N
Harper
J
Benger
J
Major complications of airway management in the UK: results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 2: intensive care and emergency departments
Br J Anaesth
 , 
2011
, vol. 
106
 (pg. 
632
-
42
)
11
Cook
TM
Counsell
D
Wildsmith
JAW
On behalf of the Royal College of Anaesthetists Third National Audit Project. Major complications of central neuraxial block: report on the 3rd National Audit Project of the Royal College of Anaesthetists
Br J Anaesth
 , 
2009
, vol. 
102
 (pg. 
179
-
90
)
12
Woodall
NM
Cook
TM
A national census of airway management techniques employed during anaesthesia in the UK: results of the first phase of the 4th National Audit Project at the Royal College of Anaesthetists
Br J Anaesth
 , 
2011
, vol. 
106
 (pg. 
266
-
71
)
13
Seven Steps to Patient Safety: A Guide for NHS Staff
 , 
2004
National Patient Safety Agency
 
Available from http://www.nrls.npsa.nhs.uk/resources/?entryid45=59787, accessed December 26, 2010)
14
Caplan
RA
Posner
KL
Cheney
FW
Effect of outcome on physician judgements of appropriateness of care
J Am Med Assoc
 , 
1991
, vol. 
265
 (pg. 
1957
-
60
)
15
Henriksen
K
Kaplan
H
Hindsight bias: outcome knowledge and adaptive learning
Qual Saf Health Care
 , 
2003
, vol. 
122
 
Suppl. 2
(pg. 
ii46
-
50
)
16
McLaughlan
C
Royal College of Anaesthetists census report 2007
Bull R Coll Anaesth
 , 
2008
, vol. 
50
 (pg. 
2577
-
8
17
Auroy
A
Benhamou
D
Péquignot
F
Bovet
M
Jougla
E
Lienhart
A
Mortality related to anaesthesia in France: analysis of deaths related to airway complications
Anaesthesia
 , 
2009
, vol. 
64
 (pg. 
366
-
70
)
18
Li
G
Warner
M
Lang
BH
Lin
H
Sun
LS
Epidemiology of anesthesia-related mortality in the United States, 1999–2005
Anesthesiology
 , 
2009
, vol. 
110
 (pg. 
759
-
65
)
19
Tuffin
R
Quinn
A
Ali
F
Cramp
P
A review of the accuracy of death certification on the intensive care unit and the proposed reforms to the Coroner's system
J Intensive Care Soc
 , 
2009
, vol. 
10
 (pg. 
134
-
7
)
20
Domino
KB
Posner
KL
Caplan
RA
Cheney
FW
Airway injury during anaesthesia
Anesthesiology
 , 
1999
, vol. 
91
 (pg. 
1703
-
11
)
21
Statistics on obesity physical activity and diet: England, 2010.
2010
The Health and Social Care Information Centre
 
22
Kheterpal
S
Han
R
Tremper
KK
, et al.  . 
Incidence and predictors of difficult and impossible mask ventilation
Anesthesiology
 , 
2006
, vol. 
105
 (pg. 
885
-
91
)
23
Kristensen
MS
Airway management and morbid obesity
Eur J Anaesthesiol
 , 
2010
, vol. 
27
 (pg. 
923
-
7
)
24
Farmery
AD
Roe
PG
A model to describe the rate of oxyhaemoglobin desaturation during apnoea
Br J Anaesth
 , 
1996
, vol. 
76
 (pg. 
284
-
91
)
25
AAGBI Safety Statement Capnography outside the Operating Theatre
  
26
Olsson
GL
Hallen
B
Hambraeus-Jonzon
K
Aspiration during anaesthesia: a computer-aided study of 185,358 anaesthetics
Acta Anaesthesiol Scand
 , 
1986
, vol. 
30
 (pg. 
84
-
92
)
27
Harrison
GG
Death attributable to anaesthesia. A 10-year survey (1967–1976)
Br J Anaesth
 , 
1978
, vol. 
50
 (pg. 
1041
-
6
)
28
Tiret
L
Nivoche
Y
Hatton
F
Desmonts
JM
Vourc'h
G
Complications related to anaesthesia in infants and children. A prospective survey of 40 240 anaesthetics
Br J Anaesth
 , 
1988
, vol. 
61
 (pg. 
263
-
9
)
29
Crosby
E
Medical malpractice and anesthesiology: literature review and role of the expert witness
Can J Anaesth
 , 
2007
, vol. 
54
 (pg. 
227
-
41
)

Author notes

This article is accompanied by the Editorial.

Supplementary data

Comments

12 Comments
The ARNDT airway can contribute to the elective management of difficult airway cases and should be further evaluated as a potential emergency standard for 'can't intubate can't ventilate' situations in light of NAP 4 findings
15 May 2011
Jim Roberts (with Lynne Barrass, Annie Hunningher)

Cannula cricothyroidotomy is well recognised as a rescue tool for the 'can't intubate, can't ventilate' (CICV) scenario. However, the NAP 4 report dramatically highlights that this emergency technique commonly fails. Of the 27 uses of cannula cricothyroidotomy in head and neck patients, 12 were successful and 15 failed [1]. In contrast, formal surgical techniques were consistently successful in accessing the airway in CICV situations.

One such device, the Ravussin needle (VBM Medizntechnik GmbH, Sulz, Germany) is 18 gauge, 60mm long and was originally introduced for elective trans-tracheal jet ventilation ('Plan A airway'), in cases where the upper airway was known to be patent to allow unimpeded trans-oral surgical access to the larynx. However, it has been co-opted for emergency use ('Plan B airway') in the CICV setting and as a further development some anaesthetists now advocate 'prophylactic' insertion of a Ravussin catheter prior to inducing anaesthesia in patients with a critical airway.

However, there are several major disadvantages of the Ravussin catheter, whether used in emergency or elective scenarios:

* Does not provide an expiratory pathway. In patients where there is upper airway obstruction, the very cases where this device may be used, the high volumes of gas injected cannot adequately equilibrate with atmospheric pressure through the device or through the glottis. With a manujet breath-stacking occurs resulting in the high intra-thoracic pressures that are associated with poor venous return, V/Q mismatch and hypotension. With automatic jet ventilators the high pressures trigger the pause pressure alarm and the ventilator will simply stop.

* Requires a high pressure gas source to achieve oxygenation, with the attendant issues of barotrauma, mucosal damage, surgical emphysema, familiarity and availability.

* Susceptible to kinking and blocking by secretions or blood.

* Can be difficult to insert, for example due to anatomical issues, false end point due to separation of needle and cannula or can result in kinking on insertion if the tracheal wall is abutted, which necessitates insertion of a new device.

We propose that the Arndt cricothyroidotomy catheter set (Cook Medical) can provide an alternative approach. This device is also designed for emergency cricothyroid airway access in a CICV scenario. However, in our institution, we have also found it helpful in elective difficult airway cases. Airway access is achieved utilizing a percutaneous entry (Seldinger) technique. The introducer needle is fine bore (18 gauge), with a cannula that used as a conduit for the guidewire. The device itself consists of a reinforced curved cannula 60mm long with a 3mm internal diameter. It is inserted with an integrated dilator over the guidewire through the skin, which has had a small incision made in it, into the trachea. The wire and dilator are removed together following insertion and positioning is checked. It has both a female Luer-Lok connector and a standard 15mm connector. Being neither a small cannula over a small needle nor a large cannula over a large needle this system affords several major advantages:

* The dimensions of the cannula, (60mm x 3mm) allows positive pressure ventilation with a familiar, available low pressure gas source (such as the anaesthetic circuit or ambubag) and provides a much more adequate expiratory pathway.

* The absence of a cuff does not prevent ventilation of the lungs. The upper airway obstruction ensures lung ventilation/oxygenation occurs.

* The presence of the Luer-Lok means the device can be used with jet ventilators with the advantage of an expiratory pathway independent of the larynx.

* Plan B use. The 18 gauge cannula can be inserted prior to induction of anaesthesia and then rapidly converted to the full device if CICV ensues. If it is not required the 18 gauge cannula can simply be removed.

* The full device can be left in situ in cases where post op oedema or bleeding are a concern again simply being removed if not required the following morning.

* Much less prone to kinking and blockage than a Ravussin catheter

* A familiar ( Seldinger/low pressure gas source) and less traumatic (cf surgical cricothyroidotomy) technique that may lead to a lower threshold for its use when it's really needed in a CICV scenario.

In conclusion, elective experience with the Arndt airway in the management of difficult airway cases has lead us to believe that its unique dimensions and insertion technique may mean that it provides a superior alternative to cannula cricothyroidotomy in the emergency CICV scenario so clearly highlighted as a problem area in NAP4. This device warrants further evaluation as a potential standard for the management of CICV.

1. Cook T, Woodall N, Frerk C. Major complications of airway management in the UK. NAP4 report and findings.

Conflict of Interest:

None declared

Submitted on 15/05/2011 8:00 PM GMT
A comment on NAP4 from The Society of Bariatric Anaesthetists (SOBA)
15 May 2011
Claire E Nightingale (with Jonathan Cousins, William Fox, Nick Kennedy, Mike Margarson and Euan Shearer)

We would like to congratulate Drs. Cook, Woodall and Frerk on the 4th National Audit Project NAP41. We feel the recommendations in Annex 5 section 13 on obesity merit some additional comment. We agree with the recommendations made but would like to add the following:

As stated in chapters 12 and 20, pre-oxygenation and intubation in the obese patient should be performed with the patient in the head-up or ramped position, with the tragus of the ear level with the sternum. This position improves the efficacy of pre-oxygenation2 and thus maximises the time before desaturation3, reduces the risk of reflux and most importantly reduces the incidence of difficult intubation closer to that of the non- obese population 4,5. Obesity is a weak risk factor for difficult intubation and the predictors of difficulty are generally the same as for normal weight patients.

Increased work of breathing and early airway closure occurring during tidal ventilation would suggest that obese patients should not be allowed to breathe spontaneously for anything other than the shortest procedure. These patients will desaturate rapidly, so the time interval from induction of anaesthesia to assisted ventilation of the lungs should be minimised. Morbidly obese patients can be at increased risk of regurgitation and aspiration, which was implicated in 50% of deaths in the NAP-4 report. As most will require assistance with ventilation, we would advise great caution with the use of supraglottic airway devices (SADs) in patients with a body mass index (BMI) >35kg/m2. All these factors point towards a requirement for a secure airway and SOBA recommends tracheal intubation for the vast majority of general anaesthetics in the obese population. We believe a tracheal tube should be the default airway, with justification required for the use of a SAD.

Whenever possible, tracheal extubation in an obese patient should likewise be performed in the head-up position with the patient awake.

In the event of a failed intubation during rapid sequence induction, the advice is to follow the Difficult Airway Society (DAS) guidelines by allowing the patient to wake up. By its nature NAP4 was not able to address the number of cases where an airway complication was avoided by following these guidelines. Traditional UK practice has been to teach our trainees to put the patient in the left lateral position as part of the failed Rapid Sequence Intubation drill. However, this is not part of the DAS guidelines.6 Our view is that in the event of a failed airway in a morbidly obese patient, it is safer to maintain the patient in the semi-upright position rather than move to the left lateral. Given the increasing incidence of obesity in the population, we suggest it might be time to review traditional teaching on this topic. Details of the body position of the obese patients discussed in NAP-4 might be helpful to inform this debate. Moving the morbidly obese patient is extremely difficult, and one aspect of anaesthetic management is to have appropriate equipment and staff numbers to enable timely changes of position if these are required.

We agree that the presence of obstructive sleep apnea or obesity hypoventilation syndrome should be actively sought. We would recommend the STOP-BANG screening tool for this purpose7. In patients with obstructive sleep apnea or obesity hypoventilation syndrome, the use of sedatives or long acting opioids should be avoided to minimise the risk of delayed respiratory depression and hypercapnic respiratory arrest in the post- operative period.

Problems occur because of poor choice of technique, inadequate preparation and lack of anticipation or plans for managing airway problems. We believe that morbidly obese patients should only be managed with experienced anaesthetists present, as part of an experienced team - in the same manner as other high-risk patients.

We agree that better training in the management of the obese patient is vital, particularly the recognition of potential comorbidities and airway difficulty. The points raised by NAP4 in this respect are well made. SOBA suggests that management of the obese patient should become a training module in the intermediate part of the anaesthetic training curriculum.

C. Nightingale J. Cousins W. Fox N. Kennedy M. Margarson E. Shearer Committee of The Society of Bariatric Anaesthetists

References

1. Cook TM, Woodall N, Frerk C. Major Complications of Airway Management in the UK. NAP4 Br J Anaesth 2011; 106: 617-31 2. Dixon BJ, Dixon JB, Carden JR, Burn AJ, Schacheter LM, Playfair JM, Laurie CP, O'Brien P. Pre-oxygenation is more effective in the 25 degrees head-up position in severely obese patients: a randomized controlled study. Anesthesiology 2005; 102: 1110-5 3. Berthoud MC, Peacock JE, Reilly CS. Effectiveness of pre-oxygenation in morbidly obese patients. Br J Anaesth 2004; 67: 464-66. 4. Collins JS, Harry Lemmens J.M., Brodsky JB, Brock-Utne JG, Levitan, RM. Laryngoscopy and Morbid Obesity: a Comparison of the "Sniff" and "Ramped" Positions. Obesity Surgery, 14, 1171-1175 5. Fox W, Harris S, Kennedy N. Prevalence of difficult intubation in a bariatric population, using the beach chair position. Anaesthesia 2008; 63: 1339-42. 6. J. J. Henderson et al. Management of unanticipated difficult intubation guidelines. Anaesthesia, 2004, 59, pages 675-694 7. Abrishami A, Khajehdehi A, Chung F. A systematic review of screening questionnaires for obstructive sleep apnea. Can J Anesth 2010; 57: 423-438

Conflict of Interest:

None declared

Submitted on 15/05/2011 8:00 PM GMT
Risk on the chin
26 May 2011
Mark W. Davies (with Andrew Bonnett)

We note with interest that serious airway difficulties were encountered more frequently with male patients, but are left wondering about the extent to which beards contributed to this sex difference. NAP 4 undoubtedly helps us to shave risks: should it encourage us to shave chins?

Conflict of Interest:

None declared

Submitted on 26/05/2011 8:00 PM GMT
Algorithm for management of vomiting following induction
26 May 2011
Annie Hunningher

The NAP4 report [1] demonstrates that aspiration was the most common cause of death and brain injury in the anaesthesia group (8 deaths/16, 2 brain injury/3). In 1848 aspiration under anaesthesia was described as being lethal [2] and in 1946 Mendelson [3] famously went on to describe the aspiration syndrome that results. Today it clearly remains a significant cause of airway related morbidity and mortality. This highlights the importance of adherence to starvation guidelines, risk assessment for aspiration, appropriate use of rapid sequence induction but also of checking suction before the start of each case. Although the incidence of an aspiration event is infrequent, it clearly can be catastrophic and can occur in starved patients without risk factors. Restriction of view from vomit in the upper airway along with subsequent sub-glottic obstruction from these solids or liquids can make an easy airway difficult and in these circumstances following the DAS algorithms is not the appropriate management. The use of an LMA or a needle cricothyroid cannula is unlikely to allow ventilation or adequate suction. We therefore propose an algorithm for management of such events be devised as part of the NAP4 and DAS collaboration/review process.

Algorithm for the management of vomiting following induction- draft outline

Supraglottic obstruction and aspiration of gastric contents at induction of anaesthesia. CALL FOR HELP-->

Head down tilt, suctioning including use of suction catheters, apply cricoid pressure, RSI if not paralysed, 100% 02-->

Intubation and suction down ETT before ventilation-->

Can intubate

CXR, Bronchoscopy. Bronchodilators. Antibiotics- no evidence. Steroids- no evidence. ITU, Protective ventilation strategy.

Cannot intubate

Proseal LMA. Surgical airway.

1.Cook T, Woodall N, Frerk C. Major complications of airway management in the UK. NAP4 report and findings.

2.Simpson JY. Remarks on the alleged case of death from the action of chloroform. Lancet 1848; 1:175-176.

3.Mendelson CL. The aspiration of stomach contents into the lungs during obstetric anaesthesia. American journal of Obstetrics and Gynaecology 1946; 52: 191-205

4.Beck-Schimmer B, Bonvini, B. Bronchoaspiration: incidence, consequences and management. European Journal of Anaesthesiology 2011; 28(2) : 78-84

Conflict of Interest:

None declared

Submitted on 26/05/2011 8:00 PM GMT
Re:A comment on NAP4 from The Society of Bariatric Anaesthetists (SOBA)
1 June 2011
Nicholas Woodall (with T Cook, C Frerk.)

We would like to thank Dr Nightingale et al. (A comment on NAP4 from The Society of Bariatric Anaesthetists [SOBA]) for their very positive comments on NAP4, in addition to these they suggest some lengthy additional recommendations. Whilst much of the information presented by Dr Nightingale is included in the NAP 4 report this was not intended to be text-book of anaesthesia and should not be regarded as a primary source of information on specialist topics within anaesthesia. The NAP4 report represents a description of the events reported to the project and analysis of issues pertinent to these cases. Analysis was performed by a broad selection of health-care professionals, drawn from a spectrum of disciplines. SOBA is a relatively young specialist society and was formed after the start of the NAP4 project and therefore, unfortunately this society was not represented in the process. Some of their comments, particularly about airway device selection and spontaneous ventilation though reasonable might not be supported by the broader anaesthetic community and need to be bolstered by clear scientific evidence. No doubt in due course SOBA will produce and justify their own guidelines and recommendations on the management of this challenging group of patients. N.Woodall, T.Cook and C.Frerk.

Conflict of Interest:

None declared

Submitted on 01/06/2011 8:00 PM GMT
Re:Risk on the chin
13 June 2011
Tim M Cook (with Nick Woodall, Chris Frerk)

Dear Editor

We thank Dr Davies and Bonnett1 for their interest in the work of the NAP4 project team.2 Patients with beards are recognised to have an increased risk of difficult mask ventilation.2,3 Male patients are also recognised to be at increased risk of both difficult mask ventilation and difficult laryngoscopy.2,5

NAP4 identified an excess of males in reported cases of major airway complications: 67% of all patients reported to NAP4, 62% of anaesthesia reports and 58% of the cohort in Intensive Care Units were male. We did not enquire about facial hair for all reported cases, however regarding mask ventilation we asked if this was difficult and if it was whether the cause was a beard interfering with mask seal. Of 131 anaesthesia reports which answered the question about mask ventilation 77 (59%) reported it was difficult. In only two cases was this reported to be due a beard preventing a good seal. These two cases were 1) a case of aspiration likely due to light anaesthesia in an obese patient managed with a laryngeal mask and 2) post-operative airway obstruction in an obese patient emerging from laryngeal mask anaesthesia for perineal surgery. There were therefore no cases reported ot NAP4 where the primary airway event arose from difficult mask ventilation caused by a beard.

The male gender appears to increase the risk of difficult mask ventilation, difficult laryngoscopy and major airway complications. Having a beard also appears to increase the risk of difficult mask ventilation and perhaps difficult laryngoscopy. However on the basis of NAP4 we conclude that it is not beardedness per se that leads to most major airway complications in males.

Tim Cook

Nick Woodall

Chris Frerk 1. Davies MW, Bonnett. Risk on the chin. E-correspondence. http://bja.oxfordjournals.org/content/106/5/617/reply#brjana_el_7385

2. Major complications of airway management in the UK: results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 1: Anaesthesia" Cook TM, Woodall N, Ferk C. Br J Anaesth 2011; 106: 617-631 doi:10.1093/bja/aer058

3. Langeron O, Masso E, Huraux C, Guggiari M, Bianchi A, Coriat P, Riou B. Prediction of difficult mask ventilation. Anesthesiology. 2000; 92: 1229-36.

4. Kheterpal S, Han R, Tremper KK, Shanks A, Tait AR, O'Reilly M, Ludwig TA. Incidence and predictors of difficult and impossible mask ventilation. Anesthesiology. 2006; 105: 885-91.

5. Lundstr?m LH, M?ller AM, Rosenstock C, Astrup G, G?tke MR, Wetterslev J; Avoidance of neuromuscular blocking agents may increase the risk of difficult tracheal intubation: a cohort study of 103,812 consecutive adult patients recorded in the Danish Anaesthesia Database. Danish Anaesthesia Database. Br J Anaesth 2009; 103: 283-90.

Conflict of Interest:

None declared

Submitted on 13/06/2011 8:00 PM GMT
Re:Algorithm for management of vomiting following induction
15 June 2011
Tim M Cook (with Nick Woodall, Chris Frerk)

Dear Editor We thank Dr Hunningher and colleagues for their interest1 in the work of the NAP4 project team.2 We agree that the NAP4 project has re-emphasised the importance of aspiration as a complication that anaesthetists should be at all times wary of and do their utmost to prevent. This is made clear in the project report both in the published paper2 and in the fuller report.3 In the report a full chapter is devoted to the topic of aspiration and it is also discussed at length in the chapter on supraglottic airway devices: interested readers are referred to these more in depth analyses of the topic. The various preventative strategies described by Dr Hunningher are also emphasised in the NAP4 report. Amongst our nine recommendations regarding aspiration we include

* Anaesthetists must assess all patients for risk of aspiration prior to anaesthesia. This applies particularly to urgent and emergency surgery. Where significant doubt exists, the higher risk should be assumed.

* The airway management strategy should be consistent with the identified risk of aspiration. Where reasonable doubt exists it is likely to be safer to assume increased risk and plan accordingly.

* No matter how low the perceived risk of aspiration, when anaesthesia is induced, the equipment and skills should exist to detect, and promptly manage, regurgitation and aspiration.

Hunningher states that aspiration may occur in patients with no risk factors for aspiration. While we agree with this, it was an important finding of NAP4 that many cases of aspiration occurred in patients who had clearly identifiable risk factors for aspiration but in whom no change in airway management plan had been made to account for this. Many cases were patients with significant risk factors managed with standard laryngeal masks.

The algorithm proposed by the authors appears to have a limited application but may be a useful starting point towards a more comprehensive guideline that would need to cover active vomiting and regurgitation with or without aspiration, both prior to and following placement of airway maintenance devices, and possibly at emergence from anaesthesia. Further discussion of such a protocol is clearly necessary before it can be recommended in a wider context and this might usefully take place with the Difficult Airway Society, one of the co-sponsors of the NAP4 project.

Tim Cook, Nick Woodall, Chris Frerk

References

1. Hunningher A, Saunders P. Algorithm for managing vomiting following induction. http://bja.oxfordjournals.org/content/106/5/617/reply#brjana_el_7385

2. Major complications of airway management in the UK: results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 1: Anaesthesia" Cook TM, Woodall N, Ferk C. Br J Anaesth 2011; 106: 617-631 doi:10.1093/bja/aer058

3. The 4th National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society: Report and Findings. Editors Cook TM, Woodall N, Frerk N. Royal College of Anaesthetists. Available at http://www.rcoa.ac.uk/index.asp?PageID=1089

Conflict of Interest:

None declared

Submitted on 15/06/2011 8:00 PM GMT
Re:The ARNDT airway can contribute to the elective management of difficult airway cases and should be further evaluated as a potential emergency standard for 'can't intubate can't ventilate' situations in light of NAP 4 findings
15 June 2011
Tim M Cook (with Nick Woodall, Chris Frerk)

The ARNDT airway can contribute to the elective management of difficult airway cases and should be further evaluated as a potential emergency standard for 'can't intubate can't ventilate' situations in light of NAP 4 findings

Dear Editor

We thank Dr Roberts and colleagues for their interest1 in the work of the NAP4 project team.2 The authors suggest that a 3mm internal diameter (Arndt Airway) cricothyroidotomy cannula placed using a Seldinger technique might offer 'the best of both worlds' in terms of ease of access (compared to a surgical approach) and ease of ventilation (compared to a narrow ?2mm id cannula). The list of advantages and disadvantages of small and large bore devices have been described elsewhere.3

It is worth clarifying several points. First the authors perhaps over emphasise the difficulty of expiration when using a narrow bore cannula and high pressure source ventilation. While breath-stacking can occur this is not inevitable: in an estimated 86% the upper airway provides an expiratory route4 enabling safe repetitive high pressure source inspirations. In those cases where the upper airway does not provide a clear airway the cannula can be used for low flow oxygen insufflation and a larger airway inserted. Second, the Ravussin cannula is 13 Gauge not 18 Gauge. Third, the authors suggest expiration may take place via the 3mm internal diameter of the Arndt airway: previous evidence suggests that the smallest airway that expiration can be achieved through is 4.0 mm and this is discussed in detail in a recent publication.5,6 Finally the authors state that even without a cuff the Arndt airway enables effective ventilation of the lungs: in Craven and Vanner's model use of a larger uncuffed deice was associated with effective ventilation only if there was no expiratory route via the upper airway.7 If upper airway resistance was low ventilation was preferentially to the upper airway rather than the lungs, this effective ventilation via an uncuffed tube may require active closure of the partially obstructed upper airway.

Despite these comments the Arndt airway may well be worth investigating further. It is recognised that most UK anaesthetists (particularly trainees) have a preference for Seldinger over surgical techniques during emergency access to the trachea8 and while there other Seldinger devices are available (for instance the Cuffed Melker Emergency cricothyrotomy device which is inserted similarly to the Arndt airway and is available with internal diameters of 3.5mm, 4.0 mm, 4.0 mm (uncuffed) and 5.0mm id (cuffed) all cook Medical, Bloomingtom, IN, USA) there are few adult length devices in the 2-4mm id rage.

NAP4 has identified a high failure rate of emergency cannula cricothyroidotomy amongst anaesthetists dealing with high risk cases. The topic is discussed in detail in the full report of the project in the chapter Management of the 'can't intubate can't ventilate' situation and emergency surgical airway.9 The reasons for failure were numerous and disparate: we cannot be sure whether this is because of poorly designed equipment, malfunctioning equipment, use of equipment not designed for this purpose, poorly used equipment, poor insertion technique or use of the wrong method of ventilation after insertion: there were examples of each reported to NAP4. It is worth re-emphasising the almost all the emergency cannula cricothyroidotomies were done in extremis, while many of the surgical airways were performed by specialist surgeons in a more controlled setting while the anaesthetists maintained an adequate airway. We do not know how well the anaesthetists who failed at cricothyroidotomy would fare when performing an open surgical airway in the patient in extremis.

Optimum training in cricothyroidotomy insertion and ventilation, use of equipment specifically designed for this use and regular practice with such equipment seem logical initial goals to pursue in trying to minimise this failure rate. Further examination of the equipment used for cricothyroidotomy is also required and we hope Roberts and colleagues will examine the performance of the Arndt cannula in more detail. Other novel development such as 'active expiration' via a narrow bore cannula (e.g. the Ventrain system, Dolphys, Eindhoven, Netherlands) might also be explored.10

Tim Cook, Nick Woodall, Chris Frerk

References 1. Roberts J, Barrass L, Hunningher A. The Arndt airway can contribute to the elective management of difficult airway cases and should be further evaluated as a potential emergency standard for 'can't intubate can't ventilate' situations in the light of NAP4 findings. http://bja.oxfordjournals.org/content/106/5/632/reply#brjana_el_7385

2. Cook TM, Woodall N, Frerk C. Major complications of airway management in the UK: results of the 4th National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 1 Anaesthesia. British Journal of Anaesthesia 2011; 106: 617-31

3. Cook TM, Nolan JP, Cranshaw J, Magee P Needle Cricothyroidotomy. Anaesthesia 2007; 62: 289-90

4. Smith RB, Babinski M, Klain M, Pfaeffle H. Percutaneous transtracheal ventilation Journal of American College of Emergency Physicians 1976; 5: 765-70.

5. Nunn JF, Ezi-Ashi TI. The respiratory effects of resistance to breathing in anesthetized man. Anesthesiology 1961; 22: 174-85.

6. Isono S, Kitamura Y, Asai T, Cook TM. Perioperative Airway Management of a Patient with Tracheal Stenosis: expert opinions and actual clinical management . Anesthesiology. 2010; 112: 970-8.

7. Craven RM, Vanner RG. Ventilation of a model lung using various cricothyrotomy devices. Anaesthesia 2004; 59: 595-9.

8. Sulaiman L, Tighe SQ, Nelson RA. Surgical vs wire-guided cricothyroidotomy: a randomised crossover study of cuffed and uncuffed tracheal tube insertion. Anaesthesia 2006; 61: 565-70.

9. The 4th National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society: Report and Findings. Editors Cook TM, Woodall N, Frerk N. Royal College of Anaesthetists. Available at http://www.rcoa.ac.uk/index.asp?PageID=1089 10. Hamaekers AEW, Borg PAJ, G?tz T, Enk D. Ventilation through a small- bore catheter: optimizing expiratory ventilation assistance. Br J Anaesth. 2011; 106: 403-9.

Conflict of Interest:

None declared

Submitted on 15/06/2011 8:00 PM GMT
"The Royal Perth Hospital solution for managing the ""can't intubate, can't ventilate"" crisis"
19 June 2011
Simon R. Massey (with Dane Blackford, Danielle Murray)

Dear Editor,

We commend the Royal College of Anaesthetists and Difficult Airway Society (DAS) in producing the Fourth National Audit Project "Major complications of airway management" [1].

Practicing anaesthetists should be alarmed at how ineffective traditional techniques are in managing the 'can't intubate, can't ventilate' (CICV) crisis. Over 65% (16 of 25) cricothyroidotomy attempts by anaesthetists failed to secure the airway [1].

The authors estimate anaesthetists in the UK are likely to experience a CICV scenario once every six years. Poor management was thought to be an aetiological factor in the majority of airway related deaths in their series prompting the authors to conclude "research is needed to identify equipment and techniques most likely to be successful for direct tracheal access."

We would like to draw attention to excellent work already published in the British literature that addresses a number of these questions. The Royal Perth Hospital has published a "can't intubate, can't ventilate" algorithm, the result of a number of years of airway crisis simulation and training using anaesthetized sheep [2]. The Royal Perth group identified that in a CICV situation, the DAS Difficult Intubation algorithm leaves one at a difficult junction: should you perform a cannula or a surgical cricothyroidotomy? What if you fail with your first choice? What if you cannot clearly identify the midline structures? Without a clear management plan, this incredibly stressful situation becomes even more challenging.

The Royal Perth algorithm outlines a logical progression of techniques from the least invasive through to the most successful, but most invasive. The algorithm emphasizes rapid oxygenation and favours the skill complement of the anaesthetist, who may be more reluctant than a surgeon to reach for a scalpel.

The algorithm initially recommends attempting cannula cricothyroidotomy or tracheotomy. If successful, jet oxygenation can stabilise the patient. The algorithm next suggests either waking the patient, considering further upper airway techniques or placing a COOK Melker Seldinger guided 5.0 tube.

If the initial cannulation fails, the next step depends on having palpable neck airway anatomy. If palpable, then a "scalpel bougie" technique is recommended. The COOK Frova Bougie has a lumen to facilitate jet ventilation. This is introduced into the trachea through a surgical incision and guided into the trachea against the scalpel blade. This allows jet ventilation to oxygenate and stabilize, prior to railroading a 6.0 tube.

If the "scalpel bougie" technique fails or the airway anatomy is not palpable at the outset, the Royal Perth algorithm recommends a 6cm longitudinal scalpel incision and finger dissection until the neck anatomy is palpable. A cannula tracheotomy then allows jet oxygenation.

Crisis algorithms successfully facilitate logical task progression in a stressful environment. The current DAS algorithm is excellent until arriving at the CICV scenario. We believe the Royal Perth CICV guideline complements the current DAS algorithm and importantly provides guidance at the CICV juncture.

At our institution, we have initiated anaesthetic CICV training adopting the Royal Perth algorithm and techniques. The YouTube videos produced by the authors greatly facilitated this process [3]. The algorithm was easily learned and understood. The practical techniques were quickly mastered using mannequin and pig specimens. We are now better prepared for a CICV crisis.

Simon Massey, MRCP, FRCA, FRCPC Dane Blackford, FANZCA Danielle Murray, BA, Research Fellow

References:

1. Cook TM, Woodall N, Frerk C. Major complications of airway management in the UK: results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 1: Anaesthesia and Part 2: intensive care and emergency departments. Br J Anaesth 2011; 106: 617-42

2. Heard AMB, Green RJ, Eakins P. The formulation and introduction of a "can't intubate, can't ventilate' algorithm into clinical practice. Anaesthesia 2009; 64: 601-8

3. DrAMBHeardAirway. DrAMBHeardAirway's Channel. YouTube. Available from http://www.youtube.com/user/DrAMBHeardAirway (accessed on 15 June 2011)

Conflict of Interest:

None declared

Submitted on 19/06/2011 8:00 PM GMT
Better airway skills make for better anaesthetists
19 June 2011
Nick Barnett

I read the recent national airway audit and subsequent recommendations with some interest. I feel however that it does not go far enough, beyond certain generic propositions for improved training and education, in pushing the cause of improved airway management skills in anaesthesia.

My own experience, coming to the end of my training, are that airway management and knowledge of how to actualise this in real-world scenarios is inadequate amongst many anaesthetists. For example, I have recently received two identical trainee audits examining the state of our general knowledge concerning difficult airway scenarios, with the general consensus being that our knowledge and skill base was insufficient. I also struggle to recall how many times I have received any indication of where the difficult airway trolley was and what it contained. Indeed where are the posters of say a difficult airway algorithm in our anaesthetic rooms ?

Thankfully major airway complications are rare but it is precisely their rarity that demands a coordinated approach as well as the the vague feeling despite numerous guidelines of not feeling entirely confident about what to do in a tricky elective or trickier still, emergent situation.

One solution, triggered by the audit, is elaboration of a mandatory national airway management course for anaesthetists and other interested professionals. This would be similar to an ALS or ATLS course in its format, contain a course manual, be evidence-based (or consensus based at the very least), enact airway scenarios and moulages, elective and non- elective, involve simulation or mannequin work, review of airway devices and a structured ability to deal with the difficult airway under a series of guises. I don't think a course like this would aim to make experts of us but merely to lift the general confidence and standard of care around airway issues and perhaps remove or make us aware of some of the cognitive biases that arise when we encounter these situations. I also don't think this skill base should remain the preserve of enthusiasts or those who have an interest in 'difficult airways'. This is of interest, I believe, to all anaesthetists and all the patients for whom we care. Simply put - better airway skills make for better anaesthetists. Success of a pilot scheme might see interest from overseas and other anaesthetic societies and wider dissemination of such an initiative.

Might I suggest that we use the national airway audit as a powerful lever to improve the level of airway skills amongst anaesthetists and give serious consideration to developing a college-led, mandatory course to enhance these skills.

Conflict of Interest:

None declared

Submitted on 19/06/2011 8:00 PM GMT
Re:Better airway skills make for better anaesthetists
5 July 2011
Tim M Cook (with Nick Woodall, Chris Frerk)

Better airway skills make for better anaesthetists: a reply

Tim Cook, Nick Woodall, Chris Frerk

We thank Dr Barnett for her interest1 in the NAP4 project.2 It is particularly important to hear opinion from trainees about the NAP4 report as many of the findings are directly relevant to trainees: in time these trainees as consultants will be required to drive necessary change. In future projects, consideration might usefully be given to appointing a trainee to the review panel.

The title of Dr Barnett's letter is difficult to argue with and resonates with the two quotations that are used in the introduction of the report. Rosen and Latto wrote in 1984 that 'There is one skill above all that an anaesthetist is expected to exhibit and that is to maintain the airway impeccably' and in 1995 Benumof stated that 'The most compelling educational effort for the anaesthesia community should be to reduce the frequency and severity of complications related to managing the airway'.3 These two statements, and Barnett's headline, are at the core of what NAP4 is about.

NAP4, in addition to identifying learning points has made approximately 170 recommendations: while some are generic many are specific and direct. Each recommendation aims to improve the 'fit' of one piece of the airway management 'jigsaw' and thus improve the overall picture.

NAP4 does not identify a 'crisis in airway management', indeed it has confirmed that airway management during anaesthesia is very safe, but this is no reason to be complacent. The report does however shine a light on complications of airway management and has identified that in the majority of cases in anaesthesia (as well as in the Intensive Care Unit and Emergency Department) when major airway complications arise the quality of management has either been poor or has had elements of poor practice.

It is likely that the emphasis on (and training in) airway management and the experience of difficult airway management varies from region to region and also within regions. While numerous 'airway courses' exist around the country it is highly likely these also lack an overall coherent approach. There are several airway courses that, in addition to technical skills, do include an element of human factors training using simulated scenarios but it is certainly true that many of the 'advanced provider courses' focus solely on skills and settings that many anaesthetist rarely find themselves in. The idea of a nationally agreed airway course, along the lines of an Advanced Life Support Course has some appeal and was mooted during NAP4 review panel discussions. A standardised advanced airway course might educate and update both inexperienced and experienced anaesthetists and potentially improve overall standards. Dr Barnett's descriptions of what a course might achieve and its limitations are well described. Many challenges would need to be overcome to establish such a course: as a start the challenges of gathering the necessary high quality evidence4 (or even consensus5) are not minor.

However, the proposal for a training course is a good one but for a training course to be transferable and useful in all parts of the country, it would likely to be necessary to first define the equipment needed, including the contents of a standard airway trolley and the core skills required by all anaesthetists. These topics in themselves are a challenging area beyond the immediate remit of the NAP4 project. The major partners of NAP4, the Royal College of Anaesthestist, the Difficult Airway Society, could together drive forward such an endeavour. NAP4 may provide the speciality with a suitable tipping point for encouraging these developments.

Tim Cook Nick Woodall Chris Frerk

Conflict of Interest: None declared

References 1. Barnett N. Better airway skills make for better anaesthetists http://bja.oxfordjournals.org/content/106/5/617/reply#brjana_el_7545

2. Major complications of airway management in the UK: results of the 4th National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 1 Anaesthesia. Cook TM, Woodall N, Frerk C. British Journal of Anaesthesia 2011; 106: 617-31

3. Forth National Audit project of the Royal College of anaesthetists Major Complications of Airway management in the United Kingdom. Report and Findings. Editors Cook TM, Woodall N, Frerk C. Royal College of Anaesthetists, London 2011. Introduction pp13-15 (http://www.rcoa.ac.uk/index.asp?PageID=1089)

4. Forth National Audit project of the Royal College of anaesthetists Major Complications of Airway management in the United Kingdom. Report and Findings. Editors Cook TM, Woodall N, Frerk C. Royal College of Anaesthetists, London 2011. Chapter 2, Evidence-based medicine and airway management are they incompatible. pp 16-19 (http://www.rcoa.ac.uk/index.asp?PageID=1089)

5. More than one way to skin a cat. Airway obstruction caused by retrosternal thyroid: management and expert opinion on management. Morgan P, Hersch P, Cook TM. Anaesthesia 2011 e-published 12 April 2011 doi 10- 1111/j.1365-2044.2011.06660.

Conflict of Interest:

None declared

Submitted on 05/07/2011 8:00 PM GMT
"Respopnse to The Royal Perth Hospital solution for managing the ""can't intubate, can't ventilate"" crisis"
15 July 2011
Tim M Cook (with Nick Woodal, Chris Frerk)

The Royal Perth Hospital solution for managing the "can't intubate, can't ventilate" crisis

We thank Dr Massey and colleagues1 for their interest in the NAP4 publications.2-4 NAP4 showed a high failure rate for anaesthetist- performed needle cricothyroidotomy and this raises many issues about training in management of CICV, the best equipment to use and the best techniques.

We are aware of the excellent work done by Heard and colleagues in the 'wet lab' at Royal Perth Hospital, Western Australia.5 The set-up in that hospital is perhaps unique and we are certainly not aware of a similar set-up in the UK. In the article cited by Dr Massey the authors describe a number of techniques they report to have found effective in training a large cohort of anaesthetists and present their institutional algorithm which is based on their local experience and also analysis of the literature. There is a considerable amount to learn from the paper and it is well worth reading.

NAP4 has already led to considerable discussion about the present recommendations that are embedded in the DAS guidelines for failed intubation and failed ventilation.6 Since the publication of NAP4, at meetings presenting the results, it has been common for audience members to suggest that anaesthetists should abandon cannula cricothyroidotomy in favour of a surgical approach. In the NAP4 report we intentionally stopped short of this recommendation. NAP4 highlighted a high failure rate for cannula cricothyroidotomy. The reasons were numerous and included poor technique, poor choice of equipment, equipment failure, and failures of both cannula insertion and of safe and effective ventilation. There is no useful data from NAP4 on the success rate of anaesthetists with a surgical technique, but it is perhaps relevant of itself that anaesthetists did not use this technique, which suggests it is currently not popular with anaesthetists. Although the success rates for surgeons using a surgical technique to rescue the airway was almost 100% it should be noted that many of these occurred while the anaesthetist maintained a clear or adequate airway and oxygenation; in contrast cannula cricothyroidotomies were typically performed in extremis and with the anaesthetist having to abandon any attempt at oxygenation while attempting the procedure. They are therefore not directly comparable. As a final piece to the jigsaw we have very little idea how anaesthetists perform with a scalpel, or indeed whether they are even willing to do so. NAP4 does not answer all the questions. It has identified a high failure rate of cannula cricothyroidotomy but does not reveal whether better equipment choices and better training would improve this, nor whether surgical techniques are more appropriate for us.

Our recommendations regarding CICV include the following * All anaesthetists must be trained in emergency cricothyroidotomy and keep their skills up to date. * Surgical cricothyroidotomy should be taught alongside cannula cricothyroidotomy, including to anaesthetists. * Further research focussed at identifying the success rates and optimal techniques of cannula cricothyroidotomy is required. * Anaesthetists should understand that the decision to perform an emergency surgical airway is commonly inappropriately delayed. The importance of early, clear decision-making should be highlighted during training in cricothyroidotomy.

The Perth algorithm has much to recommend it. It describes an initially minimally invasive technique of small cannula over needle insertion, progressing as needed to a larger cannula inserted with a Seldinger technique and thence as indicated to a scalpel bougie or more surgical technique. Some aspects of the described techniques require examination. The algorithm advocates use of a standard intravenous cannula for tracheal access during the surgical technique rather than a device such as the Ravussin cannula which is designed for this role. 'Jet ventilation' (high pressure source ventilation) via a hollow bougie is also advocated to re-oxygenate before completing a surgical airway. The depth to which the bougie is inserted is not stated. Several case reports (we are aware of four at least four cases) have reported significant barotrauma, including death, from high pressure source ventilation, oxygen insufflation and even lower pressure ventilation via a blindly inserted narrow bore airway exchange catheter or bougie.7-9 We suggest that, in all but the most exceptional circumstances, the priority after established the airway with a narrow bore exchange device should be to complete the exchange to a larger tube and then ventilate conventionally.

We are pleased that NAP4 has initiated considerable discussion on management of CICV. We believe there needs to be more discussion by the anaesthetic community, further review of the literature and of recent innovations (e.g. the Ventrain device which claims to enable active exhalation via a narrow cannula10) and possibly further research. This might then lead to development of updated national guidelines on management of CICV.

As a final point we suggest that these guidelines might usefully explore the human factors associated with CICV that lead many anaesthetists to delay when action is required.11 Rescue techniques need to be technically appropriate, reliable and effective but also need to be deployed in a timely manner when needed.

Tim Cook

Nick Woodall

Chris Frerk

1. Massey S, Blackford D, Murray D. The Royal Perth Hospital solution for managing the "can't intubate, can't ventilate" crisis. (http://bja.oxfordjournals.org/content/106/5/617/reply#brjana_el_7617 )

2. Major complications of airway management in the UK: results of the 4th National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 1 Anaesthesia. Cook TM, Woodall N, Frerk C. British Journal of Anaesthesia 2011; 106: 617-31

3. Major complications of airway management in the UK: results of the 4th National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 2 Intensive Care and Emergency Department. Cook TM, Woodall N, Harper J, Benger J. British Journal of Anaesthesia 2011; 106: 632-42

4. Forth National Audit Project of the Royal College of Anaesthetists and Difficult Airway Society. Major Complications of Airway management in the United Kingdom. Report and Findings. Editors Cook TM, Woodall N, Frerk C. March 2011. ISBN 978-1-9000936-03-3 Royal College of Anaesthetists. London (http://www.rcoa.ac.uk/index.asp?PageID=1089)

5. Heard AMB, Green RJ, Eakins P. The formulation and introduction of a "can't intubate, can't ventilate' algorithm into clinical practice. Anaesthesia 2009; 64: 601-8

6. Henderson J, Popat M, Latto P, Pearce A. Difficult airway Society Guidelines. Anaesthesia 2004; 9: 675-94.

7. Duggan LV, Law JA, Murphy MF. Brief review: Supplementing oxygen through an airway exchange catheter: efficacy, complications, and recommendations. Can J Anaesth. 2011; 58: 560-8.

8. Ruxton L. Fatal Accident enquiry 15 into the death of Mr Gordon Ewing. Glasgow April 2010. (http://www.scotcourts.gov.uk/opinions/2010FAI15.html)

9. Rao A, Webster L, Turner M. Bougie induced airway trauma. Presented at the Difficult Airway Society annual congress Cheltenham 2010. Poster 83C

10. Hamaekers AEW, Borg PAJ, G?tz T, Enk D. Ventilation through a small-bore catheter: optimizing expiratory ventilation assistance. Br J Anaesth. 2011; 106: 403-9.

11. Peterson GN, Domino KB, Caplan RA, Posner KL, Lee LA, Cheney FW. Management of the Difficult Airway. Anesthesiology 2005; 103: 33-9

Conflict of Interest:

None declared

Submitted on 15/07/2011 8:00 PM GMT