Equal and opposite expert opinion. Airway obstruction caused by a retrosternal thyroid mass: management and prospective international expert opinion


  • Presented in part at the Difficult Airway Society Annual Meeting, Perth, Scotland, 30 November 2009

Correspondence to: Dr T. M. Cook
Email: timcook007@googlemail.com


Anticipated problems with airway management during anaesthesia require careful planning, particularly when they involve a risk of airway obstruction. Advice may be sought from published literature (usually written by experts) or through direct communication with experts. More frequently, expert involvement is through retrospective review following patient harm. We present the case of a patient suffering from a retrosternal thyroid mass that was compressing her trachea in the midline and dividing it into two 2 × 3 mm lumens. After local discussion the patient underwent thyroidectomy. We invited international experts in anaesthetic airway management to review her case and submit their opinions regarding the best airway management for this difficult case. Opinions differed markedly, mirroring the published literature. Some experts specifically criticised techniques that, unbeknown to them, were proposed by others. The case raises issues about the nature of expert opinion that extends beyond this particular case. The nature and implications of expert opinion, when evidence is absent or conflicting, are discussed.

Anaesthetic airway management requires planning, particularly when pre-operative assessment indicates the likelihood of difficulty. If the initial plan fails, a ‘plan B’ (C and D etc) is required [1]. Optimal planning involves choosing the best plan A and the next best option as plan B. Determining the best plan for a given clinical circumstance involves synthesis of information regarding clinician experience, available assistance, available equipment and ‘back-up’ facilities. External ‘advice’ may be sought, traditionally from textbooks and published literature and more recently by use of the internet, to access academic and non-academic literature (e.g. departmental guidance published online) or by direct communication. We describe a patient in whom we anticipated airway difficulty due to the presence of a large thyroid mass extending retrosternally that caused dramatic tracheal compression of her thoracic inlet. The patient required urgent total thyroidectomy. We present our management of the case and the prospective opinion of several international experts. The nature of expert opinion, particularly where evidence is absent or conflicting, and the implications of this, are discussed.

Case report

An otherwise healthy 59-year-old woman presented to our hospital’s emergency department with recent-onset inspiratory stridor and voice change on a background of 6 weeks of ‘chestiness’. Examination revealed an anxious lady with a body mass index of approximately 30 kg.m−2. She adopted a forward sitting posture and had a hoarse voice with audible inspiratory stridor and expiratory wheeze. She had only mild respiratory distress and was able to complete short sentences. She stated that 3 days earlier, she had been able to cycle slowly for 2 h without great difficulty. Her breathing was deep, at a rate of 12–16 breaths.min−1 with minimal accessory muscle use. Auscultation revealed only a markedly prolonged expiratory phase without wheeze. Her lower neck showed fullness anteriorly in the midline and no other abnormalities (Appendix Fig. 1). Nasendoscopy performed with the patient awake showed normal anatomy as far as the glottis and normally functioning vocal cords. A computerised tomography scan (performed supine and awake soon after admission) demonstrated a pretracheal mass extending from the level of the 5th cervical to the 2nd thoracic vertebra (Appendix Fig. 4). The tracheal compression resulted in an unusual appearance with the trachea divided into two discrete lumens (Fig. 1). At its narrowest, at the level of 1st thoracic vertebra, the lumens measured a little less than 2 × 3 mm. The luminal narrowing started close to the thoracic inlet and extended 2 cm retrosternally. An urgent thyroidectomy was planned for the following day. Nebulised adrenaline and intravenous dexamethasone were administered overnight to minimise any airway swelling.

Figure 1.

 Transverse computerised tomogram showing the thyroid mass compressing the patient’s trachea and dividing it into two very narrow lumens at the level of the thoracic inlet.

Our assessment was that the airway stenosis might cause problems after general anaesthesia with: (i) facemask ventilation, (ii) ventilation via a supraglottic airway, (iii) tracheal intubation and (iv) potential impediment of airway rescue using transtracheal access [2]. Our primary plan was to undertake fibreoptic inspection of the patient’s airway with her awake, followed by tracheal intubation beyond the stenosis with the narrowest lumen tube practicable. If encountering difficulty, we planned to abandon this approach and embark on an alternative: intravenous induction of general anaesthesia and rapid neuromuscular blockade with tracheal intubation either with a small standard tracheal tube (plan B) or with a rigid bronchoscope (plan C). Plan D was to perform airway rescue with a supraglottic airway and/or direct tracheal access, both of which might be highly problematic.

Standard monitoring was applied. The nasotracheal route was prepared by applying topical anaesthesia and ‘conscious sedation’ was achieved using a target controlled infusion of remifentanil (0.5–2.0 ng.ml−1). Oxygen was administered via nasal cannulae and a facemask. In the operating room, with the patient sitting upright and breathing spontaneously, nasal fibreoptic inspection was performed and lidocaine 4% applied to her trachea. The fibrescope was preloaded using an Aintree Intubating Catheter (AIC; Cook Critical Care, Letchworth, UK) and a 6.5-mm internal diameter intubating LMA tracheal tube (ILMA tracheal tube; Intavent Direct, Maidenhead, UK). The ILMA tracheal tube was chosen because of its soft bullet tip and excellent, non-traumatic railroading characteristics. The fibrescope and AIC were passed though the vocal cords and beyond the tracheal narrowing (Fig. 2). The fibrescope was removed and the presence of carbon dioxide expired via the AIC was confirmed using capnography. Sedation was increased and the ILMA tracheal tube was passed over the AIC to lie below the obstruction. The AIC was removed and the correct position of the ILMA tracheal tube was confirmed by capnography and direct inspection with the fibrescope reinserted. General anaesthesia was then induced. At the end of an uneventful 90-min total thyroidectomy, spontaneous respiration was re-established and the tracheal tube was removed and replaced with a size 4 ProSeal LMA (Intavent Direct). The absence of tracheomalacia and normal vocal cord movement were confirmed by fibroscopy. The patient made an uncomplicated recovery and was discharged 48 h later. Subsequent histology confirmed a benign multinodular goitre.

Figure 2.

 (a) Intubation of the sedated patient's trachea with an Aintree Intubation Catheter, (b) followed by the railroading of a small intubating laryngeal mask airway tracheal tube.

Expert review

With permission, we sent detailed information on this case, including images, but not our management, to 14 prominent airway management experts. The case was also posted by one expert on the Society for Airway Management (SAM) secure online forum. Experts were specifically chosen because of their high national or international profiles and included several former presidents of SAM, and past presidents of the equivalent UK society, the Difficult Airway Society (DAS). The experts included co-authors of national guidance on management of the difficult airway in at least three countries. Experts were not made aware of one another’s opinions. The information provided to the experts and questions asked are detailed in the Appendix.

Nine experts from Britain and North America generously offered their opinions. All were acknowledged experts, with many years of independent practice and in current full-time practice. One expert offered more than one option for the primary plan and therefore was excluded. Four expert opinions from Britain and four from North America were therefore included in this analysis.

Expert opinions for the primary management plan differed considerably and are tabulated in Table 1.

Table 1.  Summary of expert opinions on case management.
Expert Induction and intubation strategy Plan B Extubation strategy
  1. TT, tracheal tube; AEC, airway exchange catheter; ILMA, intubating laryngeal mask airway; LMA, laryngeal mask airway.

1Intravenous induction, neuromuscular blockade with non-depolarising agents, direct laryngoscopy, intubation with a 4.0-mm microlaryngoscopy TTRigid bronchoscopyInsertion of ProSeal LMA size 4 and fibroscopy to exclude tracheomalacia
2Intravenous induction, neuromuscular blockade, fibreoptic intubation with a small TT or Aintree Intubation CatheterRigid bronchoscopyFibroscopy at time of extubation. Tracheostomy if laryngeal or tracheal abnormalities
3Inhalational induction with sevoflurane, spontaneous ventilation, direct laryngoscopy, intubation with one TT or two small tubes (one either side of obstruction)Rigid bronchoscopy or cardiopulmonary bypassNo inspection. Consider leaving AEC after extubation
4Inhalational induction with sevoflurane, maintain spontaneous ventilation. Asleep fibreoptic (paediatric scope) intubation with a small TTRigid bronchoscopy or cardiopulmonary bypassInsertion of LMA classic and fibroscopy to exclude tracheomalacia
5Awake fibreoptic intubationNone offeredInsertion of LMA classic and fibroscopy to exclude tracheomalacia
Intubation with a reinforced 7.0-mm TT
6Awake fibreoptic intubationRigid bronchoscopyExtubate with the patient awake
7Awake fibreoptic intubationIntravenous induction and standard intubation or rigid bronchoscopyAwake extubation without inspection unless problems developed
8Sedated intubation using a rigid videolaryngoscope (spontaneous ventilation) and passage of AEC. Jet ventilationRigid bronchoscopyNone specified
ActualAwake fibreoptic intubation with Aintree Intubation Catheter and ILMA TTIntravenous induction. neuromuscular blockade and intubation. Rigid bronchoscopyInsertion of ProSeal LMA and fibreoptic inspection to exclude tracheomalacia

Of note, several experts (unaware of other opinions) mentioned specific avoidance of techniques that others advocated, sometimes describing these alternatives as dangerous. Several views were strongly expressed.

Back-up plans showed greater consensus opinion, with several experts identifying the use of rigid bronchoscopy as part of their plan B. Some suggested the use of cardiopulmonary bypass as a plan B or plan C.

Of the plans for managing tracheal extubation, there was disagreement as to the likelihood of tracheomalacia, with some regarding it as highly likely (‘almost certain’) and others as very unlikely (‘extremely rare....so much so as to not to be worth predicting’). Most experts stated that the best management of the patient’s airway at the end of her surgery was to extubate her trachea under deep anaesthesia, insert a supraglottic airway device and perform fibreoptic inspection to exclude vocal cord palsy, tracheomalacia or oedema. A minority suggested tracheal extubation with the patient awake; one advised extubation over an airway exchange catheter which would remain in situ as a route for re-intubation should this be required. Two experts suggested tracheostomy may be needed postoperatively.


The case we present is unusual due to the division of the trachea into dual lumens and the patient’s surprising paucity of symptoms. However, it is, in essence, a case of central airway obstruction requiring anaesthesia: an unusual and challenging, but not unique, circumstance. Our management technique may be criticised or agreed with, but that is not the main reason for this report: rather we wish to examine the nature of expert opinion in this case and more broadly.

In our case, the patient required a total thyroidectomy to relieve airway obstruction. There was a significant risk that complete airway occlusion could occur during induction of anaesthesia, during surgical resection (if the airway was not secured beyond the obstruction), or at the time of extubation (if there was tracheomalacia). Facemask ventilation and ventilation via a supraglottic airway device might fail due to worsening tracheal obstruction either consequent on induction of anaesthesia, or on initiating neuromuscular blockade or controlled ventilation. Tracheal intubation might be difficult or impossible due to the tracheal narrowing, and cricothyroid puncture techniques (the standard rescue technique for the ‘can’t intubate, can’t ventilate’ scenario) would be significantly impeded by the anterior neck mass and intratracheal compression. Thus, the patient was clearly at risk of complications from airway management; if managed poorly death might well have resulted.

The experts whose advice we sought did not have the benefit of examining the patient. Had they done so, this may have influenced their opinion on subsequent management. However, we were struck by the diversity of opinions and techniques suggested as their first (and presumably best) choice: none matched our actual management. Inadequate expertise is an unlikely cause for this variation as the experts were selected for their known national and international profiles in airway management. They included past presidents of national airway societies, members of airway taskforces and authors of numerous major airway publications, standard textbooks and national airway guidelines. Several have published specifically on the management of the obstructed airway and stated in their replies that they drew on personal experience of ‘similar cases’. Yet despite this, opinions on the best primary management for this patient differed; importantly, these differences were not subtle, but implied inherently different beliefs about the greatest risks to which the patient would be exposed, and methods for mitigating these risks. Some experts wished to secure the patient’s airway with her awake and some with her anaesthetised, some opted to maintain spontaneous ventilation and some to induce neuromuscular blockade, some to use conventional tracheal tubes and ventilation and others advanced techniques such as ‘jet ventilation’. This suggests fundamental divergences in opinion, an impression borne out by the fact that several experts went so far as to specifically condemn as being unsafe, techniques independently proposed by other experts. Of note, for many of the techniques advocated, some support can be found in the published literature and some of the responding experts quoted this evidence [3–6].

Many factors influence a clinician’s decisions on clinical management. For an anaesthetist presented with the patient described here, these will include patient factors (the specific airway problem, patient comorbidity, body habitus, planned surgical intervention and surgical urgency), anaesthetist factors (personal training and specific knowledge, prior experience and perhaps special expertise) and organisational factors (the environment, available assistance, equipment including back-ups, local policies and routines). Patients' preferences may, for many cases, have an important influence, but in complex emergencies, choices are likely to be offered or requested less often. During patient evaluation (history, examination and investigations), these numerous factors are assimilated and synthesised into a plan or plans (i.e. a strategy). This process may include a period of careful reflection, further discussions with colleagues, or the use of external educational resources. How these many factors may influence local practice, personal preference and the approach to an individual patient are elegantly set out in a recent editorial that introduces the term ‘context sensitive airway management’ [7]. The responses of the experts in this case perhaps act as a practical illustration of Hung and Murphy’s proposal.

As expert opinion may be influenced by personal experience, local practices and national guidelines, we acknowledge that by seeking opinions from disparate experts from several countries, it may be inevitable that more than one opinion will be offered. However, most doctors in most countries now have access to the internet, and the medical literature can be considered globalised. Practice in one country is now likely to be influenced by science, experience and expert opinion from around the world. We therefore consider our approach to be acceptable. Of note, conflicting and opposing opinions were offered from experts within one geographical area, suggesting that our results were not simply due to geographical variations in approach.

Expert involvement in such cases may arise for a number of reasons. First, complex patients may more commonly present to experts as it is through such contact that expertise is achieved, and experts may choose to work in environments where their expertise is relevant. Second, rather rarely, an expert’s view may be sought in the acute situation. Third, an expert may be asked to comment on the management of a case, most frequently when an untoward event has occurred leading to a retrospective analysis as part of a local investigation, during a medico legal claim, or as part of a closed claims analysis.

It has previously been shown that expert opinion may differ. In 1996, Posner et al. wrote “Expert opinion in medical malpractice is a form of implicit assessment, based on unstated individual opinion” [8]. Fifteen pairs of anaesthetist reviewers independently assessed the appropriateness of care in 103 anaesthesia malpractice claims. There was agreement on appropriateness of care in 62% of claims and disagreement in 38% (kappa = 0.37; 95% CI 0.23–0.51). Such opinions may be significantly influenced by outcome: the same authors asked 112 practising anaesthetists to judge care in 21 cases as ‘appropriate’, ‘less than appropriate’, or ‘impossible to judge’. Each case was sent to several reviewers, but with the outcome varied and reported to some reviewers as ‘temporary harm’ and to others as ‘permanent harm’ [9]. In more than 70% of cases, judgments of the appropriateness of care varied inversely with the allocated outcome. ‘Appropriate care’ ratings decreased by 31% when outcome was changed from temporary to permanent and increased by 28% when outcome was changed from permanent to temporary. Experts may also fall prey to hindsight bias: an exaggerated belief that a poor outcome could and should have been predicted, and thus avoided [10]. Crosby, reviewing the topic, commented on the lack of peer or professional review of experts and concluded that “Bias is pervasive in the analysis of medical occurrences and may result in findings against caregivers which are unfair” [11].

In this case, we neither asked for comments on the appropriateness of our care (experts were not informed of our management) nor did we describe an outcome; indeed, the patient was presented as a prospective case: ‘How would you anaesthetise this patient’. As such, the above biases are unlikely to have markedly affected the opinions we received. Nevertheless, the one thing that was consistent about the expert opinions was their lack of consensus for a primary plan. Although we did not ask experts for retrospective judgments, we speculate that management that varied considerably from their proposed plans would more likely be judged by that expert to be ‘less than appropriate’. As such, several experts might judge a number of different plans as being ‘less than appropriate’.

It is interesting therefore to speculate on what might have happened if our management of this difficult case had not been successful and the patient had come to harm. Our treatment choices might have been reviewed by these or other experts and, depending on who was selected and their opinion, we might have been found wanting. The consequences for the clinician might be significant.

While this problem may apply to many aspects of medicine, it is likely that it is accentuated in management of the difficult airway. This is an area of medical practice that is almost bereft of any high quality evidence to support particular management approaches, due to the lack of controlled studies. Practice is influenced by cases studies, occasional cohort studies and by expert opinion. Expert opinion (particularly in a medicolegal context) may be based on, or even cite, adherence to nationally agreed guidelines or protocols, deemed to represent ‘best practice’. In airway management, the only established national management guidelines in the UK and US are for the unanticipated difficult intubation and the ‘can’t intubate can’t ventilate’ situation [1, 2]. Even these are largely (and necessarily) based on ‘expert or consensus opinion, anecdote and literature review… (with) few validated prospectively’ [12], rather than on robust evidence. Such evidence and derived recommendations are graded by the Scottish Intercollegiate Guidelines Network (SIGN) and the US Government Agency for Health Care Policy and Research (AHCPR) at the weakest possible level [13, 14]. There are no national guidelines that we are aware of, in our country or internationally, that are applicable to the management of a complex airway obstruction such as that presented here. Perhaps an anonymous registry of cases of airway obstruction, their management and outcome would be a useful way in which to explore best practice, but such an undertaking would be considerable both in practical and legal terms.

Difficult airway management is one of the ‘highest stakes’ areas in anaesthetic, and perhaps medical, practice. Failure to oxygenate or ventilate the patient’s lungs during anaesthesia will quite rapidly and predictably lead to brain damage or death. In a recent analysis of medicolegal claims against anaesthetists in the National Health Service of the UK, claims were classified into 14 different clinical categories [15]: airway-related claims were amongst the three categories with the highest proportion of severe or fatal outcomes, the highest median cost per claim and the highest total cost of claims.

In the American Society of Anesthesiologists Closed Claims Project (ASACCP), claims relating to airway management are a major part of ‘respiratory claims’. Compared with non-respiratory claims, ASACCP airway and respiratory claims are associated with a notably higher proportion of poor outcomes (death or brain damage), rates of ‘less than appropriate management’ and cost of settlements. In 1990, Caplan et al. reported that 34% of all claims related to respiratory events: of these, 85% led to death or brain damage, 76% were considered to represent substandard care and 72% were considered preventable with improved monitoring. In contrast, 30% of non-respiratory claims led to death or brain damage, 30% were judged to represent substandard care and 11% were considered preventable with improved monitoring [16]. Respiratory claims led to payment to claimants in 69% of cases as opposed to 48% for non-respiratory claims, while average payments in settled respiratory claims were more than threefold higher than those following non-respiratory claims. In 2006, Cheney et al. reported a reduction in respiratory claims associated with death and brain damage, but care was still judged as being less than adequate in 64% of the cases, still more than twice the rate as that found in non-respiratory claims [17, 18]. In 2005, Peterson et al. reported on a subset of 179 claims relating to difficult airway management made between 1983 and 1999: death or brain damage occurred in 63%, anaesthetic care was judged ‘less than appropriate’ in 42%, and in 62%, payments were made to the claimant [19]. It is particularly notable that the ASACCP analysis reports that airway management is consistently rated by assessors as having higher rates of ‘substandard’ or ‘less than appropriate care’ than other sub-specialties [16–19]. Our case report illustrates that this might be influenced by variation in expert opinion as well as actual management.

In conclusion, we have shown, in accordance with the findings of previous retrospective analyses, that expert opinion on the management of difficult airway cases varies considerably. It is likely that this arises from a combination of lack of robust evidence, lack of consensus guidance, the unique nature of individual cases and personal preferences in practice. Whatever the cause, the important observation is that experts may differ fundamentally in their opinion of what is ideal practice. This observation is relevant to all organisations where expert opinion may be used to judge others’ performances. This includes internal and external serious untoward incident review, medicolegal opinion and closed claim analysis.


Written consent was obtained from the patient to publish details of her case, including photographs.

The authors thank each of the experts who freely gave their opinions on how this case might be managed. Their names are not included, as the agreement was that their opinions would be offered anonymously.

Competing interests

The authors are not aware of any potential conflicts (financial or intellectual) relating to this case. Dr Cook has been paid by Intavent Orthofix and the LMA company for lecturing in the past. That company had no involvement in this case (although some of the devices used in the case are distributed by them). Dr Cook has no financial interests in this or in any similar company. Dr Morgan and Dr Hersch declare no competing interest.


Information on the case provided to the experts:


Previously fit 59-year-old woman who presented with stridor (of moderate intensity during inspiration and expiration when deep-breathing) of short duration on a background of 6 weeks of ‘chestiness’. She had light deepening of her voice. There was no other relevant history. She had been able to cycle for 2 h, 3 days earlier without great difficulty.


She was stoic, with almost no respiratory distress, loquacious and able to speak in full (short) sentences with only mild shortness of breath. She adopted a forward sitting posture but her respiratory rate and chest excursion were normal. Auscultation revealed a notably prolonged expiratory phase. There was no use of her accessory muscles. She found lying flat anxiety provoking, but she was able to tolerate it well. She was slightly overweight, with a body mass index of approximately 30 kg.m−2. Nasendoscopy by me in the Emergency Department was poorly tolerated, but showed a normal glottis and functioning cords.

Investigations (The photographs sent to the experts were much larger but are included to illustrate the information that they received):

inline image Figure 1 Photograph of the patient’s neck.

inline image Figure 2 Photograph demonstrating the patient’s mouth opening.

inline image Figure 3 The patient’s chest radiograph.

Computed tomography scan images:

inline image Figure 4 Saggital image in the midline.

inline image Figure 5 Transverse image at the level of the cords.

inline image Figure 6 Transverse image above the sternum.

inline image Figure 7 Transverse image from just above the sternum: two lumens of 2 × 3 mm can be seen (this narrowing continued ∼2 cm below the level of the thoracic inlet).

inline image Figure 8 Transverse image from above the carina.

inline image Figure 9 Image at the level of the carina.


The surgeons plan a thyroidectomy.

  • 1. What would be your approach to anaesthetising the patient (e.g. GA vs LA, gas vs iv induction, spontaneous vs paralysed ventilation, special equipment or technique, method of securing the airway, precautions etc)?

  • 2. What would be your back-up plan?

  • 3. Assuming that airway management and surgery went well, how would you approach extubation?