There is conflicting evidence concerning the role and safety of prehospital intubation, and which providers should deliver it. Success rates for physician-performed rapid sequence induction are reported to be 97–100%, with limited evidence of improved survival in some patient groups. However, there is also evidence that prehospital intubation and ventilation can do harm. Prospective data were recorded on the success of intubation, the quality of the laryngeal view obtained and the number of attempts at intubation. These data were then analysed by the grade of intubating doctor and whether the intubating doctor had a background in anaesthesia or emergency medicine. All groups had a similar success rate after two attempts at intubation. Doctors with a background in anaesthesia and consultant emergency physicians had a significantly better first-pass intubation rate than emergency medicine trainees. The quality of laryngeal view was significantly better in doctors with an anaesthetics background.
- medical training
- prehospital care
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Several studies examining prehospital trauma deaths have suggested that airway compromise is a common contributory factor.1 2 However, there is ongoing debate on the merits of prehospital intubation, prehospital anaesthesia and the type of provider that should perform it. Early evidence suggested that intubation in the field could be performed safely3 and may improve survival.4 There is recent evidence to suggest that targeted prehospital ventilation (to a partial pressure carbon dioxide of 30–35 mm Hg/4–4.7 kPa on arrival at hospital) is associated with lower mortality after severe traumatic brain injury.5 Equally, there is evidence that (paramedic) rapid sequence intubation (RSI) can do more harm than good.6–10 The development of good practice is hampered by the lack of robust evidence and the fact that the patient case mix, the use of drugs and the skills of providers are often unclear in prehospital studies. Guidelines have been produced in the USA,11 Scandinavia12 and the UK.13 The standard of care provided during prehospital intubation should be similar to that provided in an emergency department (ED).14
Success rates for physician-performed prehospital RSI are usually high and have been reported at between 97% and 100%.15–17 In the UK prehospital intubation requiring drugs is performed by doctors from a variety of backgrounds including anaesthesia, emergency medicine and, rarely, general practice.
In the UK ED intubations are usually performed by anaesthetists, intensivists and emergency physicians. The proportion performed by emergency physicians is small but increasing with support from both colleges.18–21 Several authors have compared the success rates of ED intubation between these groups.21–23
There is a paucity of data concerning prehospital intubation in the UK. We set out to evaluate our performance prospectively in prehospital anaesthesia and record overall success rates, success rates between different physician groups, the effectiveness of cricoid pressure and the presence of airway contamination. This paper examines the success rates in terms of the number of intubation attempts and the effect of base speciality and the length of anaesthetic training on intubation performance.
Ethical approval was sought and the investigators were advised that, as there were no deviations from normal practice and no interventions, the project constituted a service evaluation. The study was therefore registered as such according to institutional guidelines. All participating doctors were asked if they would be willing to participate in the study. All data were stored on a password-protected NHS computer.
London helicopter emergency medical services (HEMS) is a prehospital trauma service, which aims to deliver critical care skills to trauma casualties at incident scenes within the London area. It serves a population of up to 10 million. It operates by helicopter during the day and fast response car at night. It is activated by a flight paramedic screening emergency services calls and delivers the doctor–paramedic team to the incident scene either before or soon after the arrival of the ground ambulance response.
London HEMS is staffed by a paramedic and a doctor trained in prehospital medicine. There are six permanent consultant staff (five emergency medicine, one anaesthesia/intensive care) and five trainees who are seconded to the service for 6 months. A number of ex-trainees remain an integral part of the service and continue to work part time once they have completed their attachment. The trainees are drawn mainly from emergency medicine, anaesthesia and intensive care medicine. Most are senior trainees (with five or more years of postgraduate experience) or have completed training in their primary speciality but wish to develop a subspeciality interest in prehospital care. All trainees undergo a 6-week induction during which their clinical work is supervised, audited and debriefed. For this period of training the training doctor is accompanied by a supervising doctor for all clinical shifts.
During the study period clinical shifts were medically staffed by four consultant anaesthetists, two anaesthetics trainees, 12 consultant emergency physicians, (five permanent staff and seven working as trainees on secondment), 12 trainees in emergency medicine, one neurosurgical trainee and one principal in general practice.
Service delivery is defined by standard operating procedures (SOP), which guide the practice of all team members. The RSI SOP highlights the importance of obtaining the best possible intubating conditions for the initial intubation attempt. Before intubation the patient is placed in a position allowing 360° access and the best lighting available. Etomidate (B-Braun, Melsungen AG, Germany) is used for induction of anaesthesia and suxamethonium (Anectine; Glaxo SmithKline, Uxbridge, UK) (1.5 mg/kg) for neuromuscular blockade. Sedation is maintained with morphine and midazolam (Hameln Pharmaceuticals, Gloucester, UK) and neuromuscular blockade is maintained with pancuronium (Hospital UK Ltd, Warwickshire, UK). The training emphasises rapid intubation as soon as the vocal cords are visualised rather than attempting to achieve a perfect view. All intubations are performed using an intubating bougie. If the cords are not seen or the bougie cannot be passed between the cords a series of ‘30 s drills’ are instituted. These include suctioning, repositioning of the operator and/or patient, reinsertion of the laryngoscope blade to a maximum then slow withdrawal, removal of cricoid pressure, laryngeal manipulation and the use of an alternative laryngoscope blade.
If these manoeuvres are not successful then the patient's position or intubator is changed. If two doctors are present at the scene, in addition to the two intubation attempts detailed in the SOP, subsequent attempts may occur. If the oxygen saturation falls below 92% the attempt at intubation is abandoned and the patient re-oxygenated using bag mask ventilation. If there are two doctors on scene the intubating doctor may have a second attempt at intubation only providing there is something that may be changed to facilitate a success. Otherwise the intubator is changed. Failure to ventilate the patient adequately mandates the use of a laryngeal mask or surgical airway.
We developed a database to examine the aspects of advanced airway care of interest. The database was maintained for 16 months from 1 January 2006 to 31 May 2007. Each doctor completed the database on returning to the helipad after each mission. Physiological observations were obtained using the printout from the monitor (Propaq, Encore; Welch Allyn, Protocol Inc., Beaverton, Oregon, USA). Data collected included information on whether intubation difficulty was predicted, the need for the 30 s drills, the number of attempts at intubation, the presence of airway contamination, the quality of laryngeal view24 and complications of the procedure. Diagrams of the Cormack and Lehane (CL) laryngeal view grading were provided on the data collection sheet.
An attempt at intubation is defined as the initial attempt at laryngoscopy combined (if required) with a 30 s drill. Removal of the laryngoscope, further pre-oxygenation and a repeat attempt at laryngoscopy was defined as a subsequent attempt.
All analysis was carried out with SAS version 9.2. Fisher's exact tests were used when frequency counts were less than five, otherwise a standard χ2 test was used. Overall tests were computed using the Mantel–Haenszel test statistic, which provided a test of overall statistical significance between any two variables. The 95% CI for the OR as estimates of RR (when appropriate) are presented along with the point estimates. Multiplicity adjustments were not carried out.
In the 16-month study period there were 481 intubations. Four hundred and two (76.1% male) patients' tracheas were intubated by London HEMS and are included in this study (see figure 1). Forty-two (10.4%) were in cardiac arrest and were intubated without drugs. The remaining 79 patients' tracheas were intubated by members of the ambulance service without drugs and are not considered further. These patients were already intubated before the arrival of HEMS, the position of the tracheal tube was confirmed and then transported by HEMS.
The results are summarised in figure 1. Four hundred patients' tracheas were orotracheally intubated and two primary surgical airways were performed. The rates of successful tracheal intubation were 87.5% on the first attempt, 98.8% on the first or second attempt and 99.8% by the third attempt. Four patients (1%) required three attempts at tracheal intubation and one patient had a rescue surgical airway performed after two attempts at orotracheal intubation. The success rates for tracheal intubaton were analysed by grade (consultant vs specialist registrar) and speciality—anaesthetics, emergency medicine and ‘other’ (general practice and neurosurgery). The results are presented in table 1 with statistical analysis in table 2. The number of patients' tracheas intubated at the first attempt was significantly lower in the group intubated by trainees in emergency medicine when compared with the other groups. There was no significant difference in the success rate of first-pass orotracheal intubation between emergency consultants, anaesthetic consultants or anaesthetic trainees. There was also no difference in any of the groups when looking at the success rates at intubation after two attempts.
We looked at the time all non-anaesthetists had spent in dedicated anaesthetic–intensive care training posts and found that those who had spent less than 6 months in such posts had significantly lower first-pass intubation rates than those who had spent more than 6 months. However, there was no significant difference between 6–12 months and over 12 months training (see table 3 for results and table 4 for statistical analysis).
There was trend to or a significant difference in the proportion of intubations in which a good (CL grade I or II) laryngeal view was obtained in favour of anaesthetists (both consultants and trainees) when compared with other groups (see table 5). All groups of doctors used the 30 s drills to facilitate intubation (see table 5).
Two patients were observed to aspirate after removal of cricoid pressure, both were difficult to intubate and had prolonged periods of bag mask ventilation.
Overall Mantel–Haenszel χ2 test indicates differences between groups (p=0.0004)—that is at least one pair of groups are different in terms of first attempt at intubation success rates. The subsequent analysis showed specific pair-wise differences as outlined in table 2.
Overall, the χ2 Mantel–Haenszel test provided evidence of differences in succesful first-pass intubation rates (p=0.0003). Specific comparisons are shown in table 4.
The grade of laryngeal view listed is for the initial view obtained, before any attempts at improving it.
The effect of medical specialty background and grade on obtaining a CL grade I or II view
Overall, the χ2 Mantel–Haenszel test provided evidence of differences between training backgrounds over all grades, ie, there is a relationship between grading (all grades) and training background (p=0.0077). See table 6.
More specifically for those graded CL I or II, there were also differences between training backgrounds (p=0.00044). The specific details of where these differences are outlined in table 6.
In this prospective observational study of intubation we found that 399/400 (99.8%) of patients' tracheas were successfully orotracheally intubated, with 87.5% intubated on the first attempt and 98.8% after the second attempt. A previous retrospective study of our service17 on 359 patients demonstrated six failed intubations (1.7%) with four requiring surgical airway (1.1%); all for failed oral intubations. In this study period, eight (2.2%) surgical airways were performed in total compared with three (0.75%) in the current study. Two cases were selected for primary surgical airway as they had facial burns and injuries that were judged to render orotracheal intubation impossible. One patient had an emergency cricothyroidotomy for failure to oxygenate or place a tracheal tube. The higher success rate in the recent study probably reflects the changes made to the intubation SOP, which structures the roles for the intubator and assistant, and mandates the use of intubation aids including a bougie.
In this study we found that anaesthetics trainees, anaesthetics consultants and emergency physician consultants all achieved significantly more intubations on the first attempt than trainees from an emergency medicine background. There were no significant differences between the anaesthetics trainees, anaesthetics consultants or emergency physician consultants. However, all groups achieved a high success rate (98.3–100%) for correct placement of the tracheal tube in the first two attempts.
Our figures compare favourably with data reported elsewhere in the UK. The largest published UK series of ED intubations (735 patients, 48.5% trauma) demonstrated a first-pass intubation rate of 94% for anaesthetists and 84% (p=0.001) for emergency physisians.22 Stevenson et al21 looked at 234 ED intubations, and found emergency physisians intubated 82% and anaesthetists 91% (p=0.056) on the first pass, with both groups intubating 97% successfully. Reid23 found junior doctors (SHO grade) with 6 months anaesthetic experience achieved significantly more first-pass successful intubations than doctors with no anaesthetic training (92.2% vs 86.6%, p=0.007), in a study of emergency intubations outside the operating theatre. Breckwodlt et al25 compared anaesthetists and internists working in the prehospital environment for a composite endpoint of the difficult intubation of a grade III/IV laryngeal view or more than three attempts at intubation. This occurred in 17.7% internist-performed intubations (130 internists) compared with 8.9% when intubation was performed by anaesthetists.
Doctors with an anaesthetist background achieved a better view of the larynx than those with a non-anaesthetic background, with anaesthetics consultants/trainees reporting a grade I or II view in 93.7% and 94.1% (94% combined) of their intubations, respectively. Figures for ED consultants/trainees were 76.3%/78.6% (78.3% combined; see table 5). This figure differs from previous work with up to 8.5% of elective26 and 12% of emergency intubations23 being graded CL III or IV. However, most of our intubations were in blunt trauma patients with an immobilised cervical spine, whereas the studies above report a mixture of trauma and non-trauma patients. This finding may be due to familiarity of the CL grading system; however, before data collection teaching was provided on the use of the CL scale and a diagram with text explanation was provided on the data collection sheet to minimise this effect. It may also be that the non-anaesthetists chose to attempt intubation at a poorer laryngeal view rather than optimising it. However, it is most likely to be due to non-anaesthetists being less familiar and thus less dextrous at laryngoscopy. Similar findings have been reported by Graham et al,22 with anaesthetists obtaining a grade I or II view in 94% and emergency physisians in 89.3%, (p=0.039) and by Stevenson et al,21 who found anaesthetists achieved a significantly better view than emergency physisians (p=0.032), but contrasts with the study of Reid et al,23 in which the laryngeal views were similar. The doctors in the groups of Graham et al22 and Stevenson et al21 and our work were of all grades, whereas those in the study by Reid et al23 focused on juniors. This suggests that the higher intubation success rates at first pass achieved by doctors with an anaesthetic training were consequent on a more complete view of the larynx.
These data are based on self-reported scores using the CL grading system. They are thus subject to skill and familiarity with this tool and are reliant on honest reporting. Therefore, there may inevitably be inaccuracies between the actual and logged values.
In a prehospital study with intubations performed by senior anaesthetists Helm et al16 reported 84.8% CL grade I or II laryngeal views in patients intubated first pass and 57.9% in those intubated on the second attempt, which is more in keeping with our figures.
We also looked at the length of time spent in dedicated anaesthetic training for the emergency trainees. The group who had spent less than 6 months in anaesthesia/intensive care had a significantly lower rate of successful first-pass intubations compared with those who had 6–12 months or those with over 1 year's experience. There was no difference between those who had 6–12 or more than 12 months experience. Information on the exact numbers of intubations performed by each practitioner was not available.
Given the risks of failed or oesophageal intubation, the frequency with which intubation is undertaken and the diversity of practitioners who perform it, surprisingly little research has been dedicated to defining the learning curve. Konrad et al27 found that a 90% success rate (intubation at first or second attempt) was achieved after 57 attempts, with 18% of anaesthetics residents still needing assistance after 80 intubations. However Wang and Yealy28 suggested that paramedics required 15–20 intubations to achieve a predicted 90% success rate. Mulcaster et al29 suggested that an 80% probability of performing a ‘good’ intubation occurred after 35 attempts in a group of medical students, respiratory therapists and paramedics. The term ‘good’ referred to technique rather than successful intubation. We were unable to identify specific studies on intubation training in the ED; however, the North American NEAR database indicates that trainees have a 10% intubation failure and 0.9% cricothyroidotomy rate.30 As important as the numbers of intubations required to achieve competency is the frequency with which the procedure should be performed to maintain competence. This question remains unanswered.
This study demonstrates good tracheal intubation success rates in a physician-led trauma system. The introduction of a prescriptive SOP defining the roles of the intubator/assistant, the equipment used and the planning of the procedure has been associated with an improvement in the rates of successful intubation for our service. We have shown a high rate at successful intubation on the first two attempts regardless of training background; however, those with an anaesthetics background or non-anaesthetists with longer anaesthetics training obtained better views of the larynx than doctors from other training backgrounds, and this was associated with a higher rate of first-pass successful intubation.
The authors would like to thank Ms Liz Foster for her valuable assistance with data collection.
A summary of this paper was presented at the stand-up science session of the London Trauma Conference, London, December 2009. An abstract was submitted to the Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine, which published all abstracts from this conference.
Funding The funding for the statistical analysis was provided by the London air ambulance charitable foundation. The statistical analysis was funded by London HEMS.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.