Pulmonary aspiration of gastric contents remains a risk in modern anaesthesia. A recent survey of anaesthetic practice in New Zealand indicated that over 71% of all respondents to a postal survey had had at least one case of aspiration in their careers, with some indicating up to 10 . Greater understanding of the pathophysiology of gastric motility and factors which influence normal function has allowed reduction of this complication in recent years. In addition, fasting guidelines are being relaxed to allow patients free access to fluids closer to the induction of anaesthesia . Surgical procedures which have previously mandated airway protection (tonsillectomy, laparoscopy) are now being undertaken with laryngeal mask airway (LMA), with minimal evidence of significant complications . Yet anaesthetists continue to devote considerable time and research to reducing the incidence of aspiration pneumonitis in anaesthetic practice. A MedlineTM search using the key words; anesthesia, anaesthesia, aspiration, pneumonitis, pneumonia and regurgitation revealed at least 469 publications in the last 10 years on this subject. The Anaesthetic Incident Monitoring Study (AIMS) now has over 5000 anaesthetic incidents reported to its database. It is an anonymous self-reporting study which allows collation of anaesthetic incidents . A recent review of the first 196 incidents of regurgitation/vomiting/aspiration allowed a management protocol based on COVER ABCD crisis management algorithm  to be evaluated (Australian Patient Safety Foundation, unpublished observations). In light of recent changes in fasting practice, the increasing use of the LMA and the call for the development of protocols and/or guidelines for prophylaxis against aspiration in anaesthesia, data from AIMS which involved incidents of aspiration, regurgitation and vomiting during anaesthesia were comprehensively reviewed.
Two-hundred and forty incidents of vomiting/regurgitation and aspiration were reported to the Anaesthetic Incident Monitoring Study database consisting of 5000 reports. Of these, 133 cases of aspiration were recorded. Passive regurgitation occurred three times more commonly than active vomiting. Aspiration was reported twice as often in elective compared with emergency surgery, with 56% of incidents taking place during induction of anaesthesia. Anti-aspiration prophylaxis was prescribed in 14% of patients who subsequently aspirated; however, the majority of cases had at least one predisposing factor for regurgitation, vomiting or aspiration evident peri-operatively. While a major immediate physiological disturbance was common, long-term morbidity was not. Death ensued in five cases, all of whom had significant co-morbidities. Factors reported as contributing to the incident included error of judgement and fault of technique, while clinical experience and anaesthetic assistance tended to minimise the incident. Aspiration remains an important anaesthetic-related morbidity. The application of simple guidelines may have prevented the incident in 60% of all cases of aspiration. Ensuring airway security may be as important as chemoprophylaxis in its prevention.
Of the first 5000 incidents reported to AIMS, those which made reference to ‘aspiration, regurgitation and/or vomiting’ were extracted. Each incident form was reviewed and relevant factors entered into a Microsoft Excel spreadsheet (Microsoft Corporation). Data were entered as originally recorded on the individual AIMS reports. Data were analysed for the following fields: age, sex, ASA status, co-morbidities, type of anaesthetic, evidence of active vomiting or passive regurgitation, type of material aspirated, when the incident occurred, anti-aspiration prophylaxis, fasting status, effects of aspiration, type of airway in use, anaesthetic management after the incident, factors promoting and reducing severity of incident and potential alleviating measures. Risk factors were evaluated by interpretation of the narrative and grouping the most obvious factors into recognised categories. These included: emergency cases, obesity, inadequate anaesthesia, opioid medication, lithotomy position [increased intra-abdominal pressure (IAP)], gastro-oesophageal reflux, hiatus hernia, tracheal tube problems, difficult intubation/airway management, gastrointestinal pathology (acute and/or chronic), anxiety, trauma, pregnancy, neurological deficit (pathology or sedation), dyspepsia, inadequate reversal, diabetes, sepsis and renal failure. Some patients had more than one risk factor.
Aspiration was considered to have occurred if: any obvious nonrespiratory secretions were suctioned via a tracheal tube, there was chest X-ray evidence of new pathology after an incident and/or there were signs of new wheeze or crackles after an episode of regurgitation or vomiting.
When interpretation of risk reduction was made, the following criteria were applied. Cricoid pressure would only have been of potential benefit if the airway was planned to be intubated; antacids, histamine type 2 receptor antagonists (H2 blockers) would have been potentially beneficial in cases of documented ‘clear fluid, vomit or gastric contents’ aspiration. Prokinetic agents were deemed useful in this group and also with solid food particles. The presence of bile-coloured vomit, when not validated by pH measurement, would only be potentially alleviated by prokinetic agents and cricoid pressure (where applicable). Antacids and H2 blockers were not considered of proven benefit in this latter group of patients. Prokinetic agents were not used in suspected cases of bowel obstruction. Finally, all reports were reviewed retrospectively and simple guidelines were applied (Table 1). Because data were qualitative, no statistical analysis was performed.
On review of the data extracted from the AIMS database, 255 cases had the keywords, ‘aspiration, regurgitation or vomiting’ in the text. Fifteen were discarded owing to inadequacy of data or not being relevant to the study, leaving 240 incidents for analysis.
Demographic data are shown in Table 2. Aspiration, regurgitation and vomiting incidents were reported twice as commonly following elective surgery in comparison with emergency surgery (Fig. 1). The majority of incidents occurred during general anaesthesia, however, seven cases occurred during local anaesthesia plus sedation (1), sedation only (4) and spinal anaesthesia plus sedation (2).
There were 133 cases of aspiration according to our criteria. In 17 reports, aspiration definitely did not occur, while in 90 cases, it was not clear whether significant aspiration had occurred following a vomiting or regurgitation episode. The majority of patients were reported to have fasted prior to their surgery. Only three patients who regurgitated/vomited but did not aspirate and 17 in the aspiration group were considered to be incompletely fasted by the reporters.
Passive regurgitation occurred three times more frequently than active vomiting. The type of material involved in the incident is shown in Table 3. The majority of aspiration events (56%) occurred at induction of anaesthesia, with the remainder occurring during maintenance, emergence and recovery periods (Fig. 2).
Anti-aspiration prophylaxis was prescribed in 10% of the whole group and 14% of the aspiration group. Cricoid pressure was used most frequently [14 (6%), 11 (8%)], with infrequent use of antacids [3 (1%), 3 (2%)], H2 blockers [5 (2%), 3 (2%)] and metoclopramide [3 (1%), 2 (2%)] in the whole group and aspiration subgroups, respectively.
The majority of cases had at least one predisposing factor for aspiration (Fig. 3). The 10 most common predisposing factors for vomiting/regurgitation and aspiration are shown in Table 4. The majority of aspiration events were associated with an immediate major physiological disturbance, while death ensued in five cases (Figs 4 and 5). There were no deaths in ASA grade 1 or 2 patients. (ASA 3; two deaths, ASA 4; two deaths and ASA 5; one death).
The majority of cases occurred with either face mask or LMA anaesthesia. In a few cases, an incident occurred with a tracheal tube in situ (Table 5). Chest X-ray evidence of aspiration pneumonitis was demonstrated in 38% of the aspiration group. Right-sided pathology was recorded more often than the left. In a large number of reports, information was missing regarding X-ray results (Table 6). Just under one-third (30%) of all patients who had evidence of aspiration were admitted to an intensive care/high dependency unit for follow-up care. Adjunctive therapy for the aspiration event is shown in Table 7.
Each respondent indicated up to three factors which contributed to and minimised each incident. These were taken from the reports unaltered, with no interpretation made of the incident by the investigators (Tables 8 and 9). When the data were analysed for factors which might have reduced the chance of this incident, it appeared that prokinetic agents might have been of greatest benefit, followed by H2 blockers, antacids and cricoid pressure. The potential value of naso/orogastric intubation was difficult to assess (Table 10). The application of simple management guidelines (Table 1) might have prevented 92 incidents overall (38%) or 79 cases of documented aspiration (60%).
Data from AIMS has shown that aspiration during anaesthesia remains an important cause of anaesthetic morbidity. As a denominator is not available, such retrospective analyses have important limitations for interpretation and the conclusions that can be drawn. Nonuniform reporting, incomplete follow up of patients, along with subjective interpretation of reporting contribute to this variability. Despite these limitations, this is one of the largest series of aspiration incidents reported and does have important implications for anaesthetic safety. Results from past surveys indicate that aspiration is more common in emergency than elective surgery [6, 7], yet elective cases in the present study were reported twice as frequently as those from emergency surgery. This may represent reporting bias from ‘low-risk’ patient groups, or may indeed reflect a real pattern of presentation. Approximately 20% of aspirations were in the population aged under 14 years. In the study by Olsson et al. , the incidence of aspiration in the 0–9 years age group was three times that for the 20–49-year-old group. The preponderance of females for aspiration seen in Olsson et al.'s study was not evident in the AIMS database. It would appear therefore that all patient groups are at risk from this complication.
While the majority of incidents occurred at induction of anaesthesia, a significant proportion presented during the maintenance and recovery periods. Induction problems can be compounded by airway difficulties and/or inadequate anaesthesia. Difficulty with intubation can be predicted in only around two-thirds of cases and regurgitation has been reported to occur in up to 7% of these . Aspiration can also occur despite cricoid pressure being applied, or resulting from its premature release. While there is an almost unerring faith in the efficacy of this manoeuvre, some workers are now beginning to question its role, safety and efficacy in anaesthesia , and indeed there were at least four incidents where the adverse event occurred while cricoid pressure was being applied. Incidents arising during maintenance were usually multifactorial in origin. Inadequate anaesthesia with an unprotected airway plus another risk factor often contributed to the adverse event. Aspiration in the recovery period is of concern, with implications for recovery room staffing, discharge criteria and the use of intra-operative methods to reduce this complication. At least two incidents involved an aspiration event where there was a specific order by the surgical team to place the patient in a supine position for a surgical consideration, while the patient had risk factors for regurgitation.
It is always assumed that gastric contents have an acidic pH. Limited data from animal and human studies have shown that bile retrograde reflux can be an important cause of oesophagitis and mucosal irritation . Interestingly, bile aspiration has been associated with a worse acute lung injury picture than that of acid aspiration . It is therefore not conclusive that pure antacid therapy will reduce the risk and improve outcome from aspiration. Moreover, as there were several causes of other aspirated material (e.g. solids, coffee grounds, milk, faecal or juice), simple reduction of pH would not necessarily have prevented and/or optimised the incident. There is great variation in the prescription of antacid and histamine receptor antagonists. While some workers in the UK  and New Zealand  report a high use of these agents in high-risk patients, other studies from France  and Australia  report much lower rates of administration. Despite widely differing frequencies of prophylaxis, obvious differences in the incidence of aspiration-related morbidity and/or mortality are not clear. The role of these agents requires further investigation, with risk–benefit analyses designed for their use. Retrospective analysis of these incidents to evaluate potentially alleviating strategies is difficult. According to our data, it would appear that prokinetic agents should be more effective than antacid therapy in reducing adverse outcomes, being based on the fact that particulate, solid food and bile aspirate would not be affected by antacids. However, there is at present no drug with a parenteral formulation that is reliable, efficacious and safe in all patient groups.
Data from this study indicate that the majority of patients who had an incident of aspiration, had one or more predisposing factors for this complication. While some factors are controversial as to their significance in the overall risk (e.g. anxiety) others seem clearly to require securing of the airway (e.g. morbid obesity, in lithotomy with history of reflux). A recurring theme in many incidents was one of inadequate anaesthesia leading to coughing/straining and subsequent regurgitation/vomiting. Obese patients continue to be over-represented in these incidents, possibly due to their coincidental airway problems and co-morbidities, including diabetes, impaired gastrointestinal transit time and increased intra-abdominal pressure. Gastrointestinal pathology, either acute or chronic may impair normal transit times. Oesophageal disease, gastric/duodenal ulceration can also impair normal gastric emptying. This, in combination with opioid therapy to control pain, proves an important risk factor. It would appear therefore that inadequate anaesthesia depth contributed to several incidents. While this factor cannot be predicted, the fact that this often arose in combination with another predisposing factor, e.g. obesity, hiatus hernia or gastrointestinal pathology indicates that a high index of suspicion for this potential complication should be raised and airway security prioritised.
Most incidents involving aspiration caused major immediate physiological effects as reported by the attending anaesthetist. This contrasts to the nonaspiration group which accounted mainly for minor effects. These observations, however, are purely subjective and are open to bias from the reporting anaesthetist. The deaths following aspiration events occurred in sicker patients (ASA grades 3 and 5), a factor consistent with other prospective surveys , although mortality can occur in healthy younger patients (e.g. obstetric patients undergoing Caesarean section) . Interestingly, nonaspiration incidents, while accounting for several major effects in the immediate period, were not associated with a major adverse outcome.
Many studies have investigated the safety of LMA with respect to airway protection and the risk of aspiration, most of which indicate that this is an infrequent occurrence in both adult and paediatric patients [17, 18]. In a meta-analysis of 12 901 patients only three cases of aspiration were described . The number of cases involving LMA anaesthesia is probably underestimated in this study as many cases where ‘mask’ was mentioned may have been a laryngeal mask. Nonetheless, there appears to be a divergence of opinion relating to the perceived risk of aspiration. While meticulous patient selection, anaesthesia and application for suitable surgical procedures has been advocated , it is clear that this does not always occur. In addition, several reports from paediatric cases demonstrated the relative lack of airway security with uncuffed tracheal tubes, where gastric contents passed the tube in situ, into the lungs.
It is recognised that human error contributes to up to 90% of all system failures . Knowledge-based error (error of judgement), technical error (fault of technique) and rule-based error (inadequate patient preparation) were the most common contributing factors identified in this study. This pattern is consistent with previous reports from the AIMS database reviewing the human failure component of incidents . The high status of knowledge-based errors emphasises the ongoing requirement for various quality assurance and continuing medical education programmes, while rule-based errors can be reduced by the use of protocols and crisis management algorithms. Finally, the elimination of technical errors can be aided by the use of simulators, anatomical models and other applied educational models. The factors minimising the incident can also be analysed according to category of error involved. Anaesthetic assistants (technicians and/or nurses) are coming under threat and the generation of ‘multiskilled’ rather than dedicated anaesthetic assistance is becoming common practice in some areas. While inadequate assistance contributed to 14 cases of aspiration, experienced assistance was deemed a minimising strategy in 39 cases. Overall, therefore, in just under half of all aspiration events, the input by the anaesthetic assistant was deemed to have a major bearing on outcome by minimising one of the system-based problems. The importance of this cannot be underestimated.
Aspiration had financial implications due to a significant number of patients requiring admission to an intensive care or high dependency unit (ICU/HDU). The reasons for admission were not clear from all the reports. Warner's large retrospective study suggests that any patient with symptoms and/or signs following an aspiration event that last for 2 h following admission to recovery should be admitted to an ICU/HDU . Current practice would suggest that corticosteroids are of no benefit and may in fact worsen outcome from particulate aspiration . The role of ‘prophylactic’ antibiotics is debated and ranges from an expectant approach to ‘routine’ administration following all aspiration events. A middle ground for antibiotic administration is suggested by some authorities for patients who have aspirated obviously infected or particulate material .
As with all studies, there are more resulting questions than answers. Can we answer any of the following? Is aspiration still a clinical problem? Yes, with at least 133 documented cases in the AIMS database. Are clinical risk factors relevant in predicting problems? This is unclear. What can be stated is that there is a complex interplay between patient risk factors, anaesthetic technique and surgical procedure. Recognition of risk factors and securing the airway or ensuring a meticulous anaesthetic technique may prevent this complication. Are current guidelines effective? At present there are few guidelines for this area and even greater discussion as to their relevance, importance and necessity. There is even considerable variation and debate as to appropriate fasting times for elective and emergency surgery . However, it is interesting that by the application of simple guidelines for the prevention of aspiration, nearly two-thirds of all cases might have been prevented. There is no doubt that further research is needed in such basic areas as indications for tracheal intubation in certain procedures, the role of cricoid pressure, the benefit of chemoprophylaxis and the safety of LMAs in all branches of anaesthesia and surgery.
In conclusion, this retrospective review of 240 incidents has revealed that peri-operative regurgitation and/or vomiting is associated with important morbidity. Mortality, although infrequent, is usually associated with the patient in ASA class 3, 4 or 5 with significant co-morbidities. Importantly, risk factors are usually self-evident, although not deemed as such by the attending anaesthetist. Factors which minimise this problem include skilled anaesthetic assistance, securing the airway in high-risk groups (obese patients, emergency, patients who have received opioids) as well as ensuring adequate anaesthetic depth in those patients in whom the airway is not protected.
Many thanks must go to Mrs Monika Bullock from the APSF who retrieved the reports for subsequent analysis. In addition, the contribution from Prof. W. B. Runciman in reviewing the manuscript is appreciated.