Results

All 245 units listed were contacted. Two declined to participate and three others had relocated their services. One further unit did not accept any premature deliveries. The sample therefore comprised 239 units, of whom 190 were in the routine group.

Discussion

Neonates respond adversely to various neonatal intensive care practices11 ,12 and several investigators have studied the physiological responses of neonates to awake intubation,2 ,5 ,8 ,9 reporting significant rises in systemic arterial blood pressure and intracranial pressure, as well as significant falls in heart rate and transcutaneous oxygen tensions or saturations. Concern has also been raised about the contribution of raised intracranial pressure4 ,13-15 to the aetiology of intraventricular haemorrhage.

Premedication can improve or abolish these potentially deleterious responses. Bradycardia can be abolished by the use of atropine2 ,6or glycopyrrolate,7 while rises in intracranial pressure have been attenuated by using muscle relaxants alone2 ,6 ,9 and combined with intravenous anaesthetics.4 ,5 Friesen et al 4 reported a mean increase in anterior fontanelle pressure of 197% in one group of premature neonates who were intubated awake compared with 25% in a matched second group, who received premedication before intubation. The use of suxamethonium can shorten the procedure of intubating neonates,6 ,7 thus reducing the incidence and duration of the hypoxaemia that commonly occurs.2 ,3

This study is an accurate representation of the current state of UK practice on premedication for intubation. The questionnaire was quick and easy to administer, with unambiguous questions, each addressing a single point, and requiring simply a yes or no response in most cases. Only two people made calls, thus achieving a high level of consistency in the application of the questionnaire, and a 99% response rate. The weaknesses of the study design lie in the potential for inaccurate information to be given over the telephone or recorded afterwards, and for the subject matter of the study to bias the responses. However, such a bias should increase positive responses. The brevity and closed design of the questionnaire sought to minimise all these problems.

These results show that most UK units do not premedicate before intubation. Our findings are similar to those of Ziegler and Todres.10 Reasons were not formally documented, but many commented that there wasn't time, particularly in an emergency reintubation. This has not been the experience in our unit and only in exceptional cases has it not been possible to ventilate using a bag–valve–mask or T piece while premedication is prepared. All doctors should be able to maintain an airway and hand ventilate after appropriate training. The use of local guidelines and protocols should mitigate any anxiety over the use of potent sedative and anaesthetic drugs by junior doctors. In practice, elective intubations by inexperienced juniors should be supervised and emergency intubations are likely to be performed by a more experienced operator. Barrington and Byrne16 have recently reported on the safety and practicability of premedicating 253 neonates for intubation, using a regimen of atropine, fentanyl, and suxamethonium, in a unit with the variable skill mix and experience that is inevitable in daily practice.

In those units that do premedicate, morphine remains the most popular choice. This may be due to familiarity, although the wide range of doses of morphine recorded in table 4 suggests that some operators use homeopathy to avoid prolonged respiratory depression while others use much larger doses to try to accelerate the slow onset of effect. However, there are several drugs that are better than morphine in terms of speed of onset, duration of effect, and side effect profile. Fentanyl and alfentanil are potent opioids with a rapid onset and short duration of action when given as a bolus. Barrington's regimen,16 using fentanyl 3–4 mcg/kg produced chest wall rigidity as the only side effect in just 4/253 neonates and three resolved promptly with the administration of suxamethonium. This side effect seems to be related to peak plasma concentration and can be minimised by slow bolus administration.17

Diazepam has a rapid onset of action and causes little haemodynamic or respiratory depression in isolation, but it and its active metabolites have long half lives. Midazolam is less problematic in this respect, but has been reported to be associated with myoclonic activity when given as a bolus18 and possibly with fits in preterm neonates when given too rapidly.19 ,20

Of the induction agents, propofol is widely used before intubation in children and infants, although there are no published data on experience in neonatal units. Ketamine offers cardiorespiratory stability but may raise intracranial pressure, which makes it less suitable. Barbiturates are not widely used, although methohexitone has been studied.21 They are rapidly effective but have prolonged elimination half lives.

Only 22% of units using any sedation also used neuromuscular blockade before intubating. Encouragingly, all muscle relaxants were accompanied by sedation. Most units used a short acting agent. Suxamethonium is still the most rapidly acting relaxant available and has important safety benefits in clinical practice where variations in skill and experience are inevitable. Half the units using suxamethonium did not coadminister a vagolytic drug to pre-empt the potential profound bradycardia that may be associated with suxamethonium, especially if given more than once. We did not enquire whether a vagolytic was drawn up ready in these nine units.

Currently 63% of UK neonatal units give no sedation before intubation on their unit. Only 14% have a protocol for premedication before intubation and only 8% use it for elective and emergency reintubations on the unit. This is despite an increasing body of research which supports this practice, on the grounds that it helps to maintain physiological homeostasis, may mitigate against risk factors for intraventricular haemorrhage, and reduces the time taken to intubate. There is also no evidence that sedation before intubation is less appropriate in neonates than in adults and children, in whom awake intubation would not be tolerated. While further studies are necessary to establish the optimal drug regimen and doses, there is sufficient evidence to support the routine practice of neonatal premedication prior to elective intubation.