• infant, newborn
• intubation
• endotracheal tube

Trumping even ‘the threshold of viability’ and tongue tie, intubation may be the most emotive subject in neonatology. Insertion of an endotracheal tube (ETT) for breathing support proved life saving for many apnoeic term babies and many preterm babies with respiratory distress. It was pivotal in defining the specialty. It provided a raison d’être to the buccaneering neonatologists who supplanted fatalistic paediatricians in the 1980s. When I began looking after babies as one of five senior house officers working long hours at a stand-alone maternity hospital in Dublin in the late 1990s, intubation was all the rage. We completed the Neonatal Resuscitation Program1 in our first week, intubating stiff plastic mannequins with gusto. We were taught that all babies born through meconium-stained liquor required intubation for tracheal suction.1 This should mean that about 20% of our babies were intubated; however, many of them were not. More often, a nervy junior doctor put a laryngoscope blade in the mouth of a conscious term baby who was neither sedated nor ‘muscle-relaxed’ and a fierce struggle ensued. I diligently suctioned many the oropharynx, oesophagus and stomach until one day, somewhat surprisingly, the larynx and vocal cords magically popped into view. I learnt how to acquire this view reliably and then learnt how to intubate. In those days, we routinely intubated preterm infants in the delivery room for surfactant administration and continued ventilation. Speed was of the essence; we were taught that intubation attempts should be successful within 20 seconds1 (though, unsurprisingly, it often took longer2). Proficiency at intubation was the yardstick by which we measured ourselves. My generation learnt to intubate while arm wrestling term babies, refined our skills on preterm babies, had punishingly long working hours in which to practise, and so became pretty slick at intubation.

But life evolves. A large randomised trial demonstrated no benefit of intubating vigorous infants born through meconium,3 so it is no longer recommended. Better awareness of the hazards of prolonged intubation and ventilation resulted in increased use of non-invasive breathing support. For infants who are ventilated, better humidification of gases has all but consigned blocked ETTs to history. The numbers of babies being intubated has fallen dramatically. Also, twice as many junior doctors now work at the hospital where I started, and mercifully, their working hours are much shorter. So, nowadays, dwindling opportunities to intubate babies are spread among increasing numbers of trainees. The reduction in the prevalence of intubation is likely good for babies; however, concerns about limited opportunities to learn and deteriorating intubation skills among paediatric and neonatal trainees have been voiced with increasing frequency for more than 15 years now.4–9

In this context, clinicians at 10 academic centres in the USA, Canada, Singapore and Germany formed the NEAR 4NEOS collaboration to study intubation in newborns. Based on similar ventures in adult and paediatrics, the group prospectively collects data about intubation attempts in babies. They recently reported on 2607 intubations performed over 2.5 years (2014–2017) at their hospitals.10 Sawyer et al report on a subset of these intubations that were performed in the neonatal intensive care unit (NICU) and proved difficult.11 They defined an intubation as ‘difficult’ when three or more attempts were taken by a senior clinician (a senior neonatal trainee [fellow] or a trained neonatologist) to insert the ETT. Difficult intubations were common (276/2009, 14% of attempts) and occurred more frequently in very preterm, very low birthweight infants. Adverse effects were common—falls in oxygen saturations of >20% occurred in 75% of difficult intubations. Eight bedside assessments, adapted from adult and paediatric anaesthesia and performed before intubation was attempted, were poorly predictive of intubation difficulty. This information is timely and thought provoking. The investigators deserve great credit for gathering data on a large number of intubations from multiple sites in a standardised fashion in the modern era. It is worth noting that the data represent discrete intubations, not individual babies. Babies who were repeatedly intubated contribute more than once to the data. This does not invalidate the data; it frequently took senior clinicians three or more attempts to intubate a baby. However, small babies who were repeatedly intubated may have contributed disproportionately to the difficult intubations (eg, the 276 difficult intubations could have occurred in 90 babies). This may have affected the assessment of the utility of bedside tests. That said, even allowing for the effect of repeated observations, the sensitivity and positive predictive value of each of the tests appears very poor. Perhaps it is unsurprising, as the tests do not appear that useful in the adult population where their use is more widespread and generally accepted.12

I find many aspects of this study intriguing. Medications were not given before 35% of intubation attempts. Analgesia/sedation was given before 58% of difficult and 65% of non-difficult intubation attempts. Paralytics were used before less than half of all attempts and less frequently in difficult (23%) compared with non-difficult (50%) attempts. This is surprising as all of these intubations occurred in NICU, where drugs were presumably available. Premedication has been demonstrated in randomised studies to increase success rates,13 particularly for more inexperienced clinicians. The prevalence of difficult intubation may be at least partly due to inadequate use of premedication, paralytics in particular. It is also interesting to consider the generalisability of this study. A large number (809/2009, 40%) of intubation attempts were made by ‘advanced practice providers;, while 15%, 30% and 7% of intubation attempts were made by junior paediatric trainee doctors, senior neonatal trainees and attending neonatologists, respectively. This contrasts with many settings where doctors in training or trained neonatologists perform the majority of intubations. For example, in a recently completed randomised trial performed at my hospital,14 63% and 37% of intubations were performed by trainees and consultants, respectively. My impression that many of these centres are large and well-resourced is reinforced by the report that at one participating centre, two failed intubation attempts prompts the attendance of a respiratory therapist, anaesthesiologist and ear, nose and throat surgeon. I am envious. These resources are not readily available at the stand-alone maternity hospital where I currently work, nor will they be when we relocate to a nearby adult hospital in the coming years.

The authors report an association between a fellow being first clinician to attempt intubation and difficult intubation. I agree that this is probably explained by experience. Fellows, particularly early in their training, may have little experience of intubating babies. Many years ago, an adult anaesthesia colleague opined to me that you needed to successfully intubate 40 times to achieve competence. Twenty years ago, you could expect to clock these up pretty quickly. Nowadays, doctors have not intubated anything like this number of babies when they embark on training in neonatology and some may well not have on completion. Fellows and newly minted consultants today are no less ‘senior’ clinicians than they were when I was in their shoes; their level of responsibility is the same. I do not believe that current trainees are any less intelligent, talented or dextrous than we were. They do, however, have less experience of intubating babies, like I have less experience of doing exchange transfusions or inserting chest drains than my mentors. The crucial problem with the loss of intubation skills is that if you cannot intubate a baby when you really need to, it very quickly becomes a very big problem.

Deterioration in intubation skills poses difficult and important questions for training and organisation of care. The ability to intubate a baby is not some innate, otherworldly skill, bestowed only on the gifted. It is a learned physical skill. When performing any physical skill—riding a bicycle, playing guitar or hurling15—the aptitude of individuals varies and has some bearing on performance; however, practice has a much greater influence. The opportunities to intubate babies in real life are limited, and the stressful circumstances in which these limited opportunities occur are often not conducive to learning. As a result, simulation has become more widespread. While they have evolved from the stiff, inflexible models that I first learnt on, newer mannequins do not (yet) faithfully recreate the real experience. High-fidelity simulation is costly and resource intensive. While it does increase an individual’s confidence, the evidence that it improves performance is weak. I think simulation is useful for learning about team dynamics and the importance of communication; I am less convinced of its utility in learning how to intubate a baby. As in hurling, it is doubtful that there is a satisfactory substitute for championship (big match) experience. In this context, videolaryngoscopy—where the view obtained by an inexperienced operator on direct laryngoscopy is shared on a screen with an experienced operator who can advise them how to improve it, if needed—increases the success rates of inexperienced operators.16 While this likely helps individual trainees to learn more efficiently, the presence of an experienced operator is still required, which has implications for how care is delivered.

Supraglottic airways, such as the laryngeal mask airway (LMA), have been suggested as an alternative to endotracheal intubation in newborns. These devices were developed for short-term use in unconscious adults who mostly have compliant aerated lungs. The attraction of the LMA is that can be inserted quickly by inexperienced operators with minimal training. An important drawback is that supraglottic airways are not available for babies who weigh less than 1.5 kg, the babies who are most frequently intubated and identified as more likely to be difficult to intubate in this study. LMAs have been used in newborns for resuscitation, short-term manual ventilation and surfactant administration. It is not clear how well babies, particularly those with poorly compliant lungs, can be ventilated through an LMA for prolonged periods. While supraglottic airways have a role, they are unlikely to replace ETTs and do away with the need for intubation any time soon.

For many years, competence at neonatal endotracheal intubation was a requirement for doctors in basic paediatric training in many countries, including Ireland, the UK and the USA. It was also expected that wherever a baby was born, someone capable of intubating a baby should be immediately on hand at all times.1 In my opinion, neither is now a realistic expectation. Given the experience of paediatric trainees over recent years, I think it is also unrealistic to expect all senior (non-neonatologist) paediatricians who are on call remotely to delivery rooms that are outside tertiary centres to be competent at endotracheal intubation. This has implications for the provision of maternity services more generally. It needs to be acknowledged that a person competent at intubating babies cannot be on hand everywhere that babies are born. If this is not acceptable, then hard decisions may need to be made. Should all babies deliver at tertiary centres? For many reasons, I think this is neither feasible nor desirable.

It is difficult to learn how to intubate newborn babies. We have not yet figured out how to achieve and maintain competence. Given its importance, we have to try harder. Improved and refined simulation training may offer some hope. However, the real-life opportunities to intubate—where the real learning happens—will remain limited. At present, it seems prudent to give these limited opportunities to trainees who need the skills most (ie, neonatologists in training), rather than to trainees who are unlikely to use them in the future (eg, community paediatricians in training), to increase their chances of achieving proficiency.