The most common serious long-term consequence of neonatal endotracheal intubation is acquired subglottic stenosis (SGS). The incidence in intubated infants is variously reported as between one and eleven per cent but this may be a relatively high range, reflecting the interest of the reporting centres and the various methods of diagnosis. (Fearon et al., 1966; Parkin et al., 1916; Strong and Passy, 1977; Papsidero and Pashley, 1980; Jones et al., 1981; Sherman et al., 1986).

In one study conducted at University College Hospital London (UCH) we reported an incidence of 1.8 per cent of SGS in all intubated neonates in a three year period (1981-83); if only survivors were considered the incidence was 2.6 per cent (Quiney and Gould, 1985). The relatively high incidence during this particular time focussed our attention on SGS and the factors which might precipitate its occurrence.

Mechanical ventilation is now widely used in neonatal intensive care. Methods and materials for endotracheal intubation vary between intensive care units. Up to eight per cent of ventilated infants may have subglottic stenosis after discharge from hospital (Jones et al., 1981).

Historically, tracheostomy has been performed from about the first century, and instruments were devised to carry out such a procedure. Relief of airway obstruction was later facilitated by instruments to perforate the cricothyroid membrane and at the turn of the century, O'Dwyer introduced an instrument to allow blind intubation of the larynx with a metal cannula. Direct visualization of the larynx for the purposes of diagnosis and treatment was effectively delayed until general anaesthesia became available, and early laryngoscopes and bronchoscopes were hampered by poor illumination and difficulty in maintaining oxygenation. The invention of fibreoptic lighting overcame the former problem to a large degree, but easy maintenance of oxygenation and anesthesia during bronchoscopy has been revolutionized by the invention of rigid ventilating bronchoscopes.

This can conveniently be divided into two parts. First, anaesthesia for diagnostic endoscopy and further assessment of subglottic stenosis and, second, anaesthesia for laryngeal reconstructive surgery.

In this paper, I shall mention a survey I undertook in September 1986 on the prevalence and aetiology of tracheostomy among children in the U.K. (Jennings, 1987). Four hundred and fifty-five members of the British Association of Otolaryngologists were circulated with a questionnaire and 58 per cent responded.

Apart from the neonatal period acquired subglottic stenosis is most often seen in infancy and early childhood. It remains the commonest indication for tracheostomy in the paediatric age group, (Jennings, 1987). The aetiology is usually multiple and may include major contributory factors from the airway itself, the presence of an intercurrent illness and the complications of intubation (Table I). The airway may be intrinsically normal or abnormal at birth but may subsequently be stressed by various events such as infection either in the neonatal period or in the first few months of life. Extrinsic compressions of the airway, for example by a vascular ring (Westaby et al., 1984) may produce tracheomalacia so increasing the risks of respiratory obstruction during intercurrent infection. Intubation is the preferred method of airway support for those with associated severe respiratory failure but brings with it a number of potential complications which may eventually lead to significant subglottic narrowing.

Most children with acquired subglottic stenosis require a tracheostomy. This paper describes the communication skills of children with tracheostomies before and after decannulation.

There are also a small number of children with acquired subglottic stenosis who do not require a tracheostomy. They are rarely referred for speech therapy. Quiney et al. (1986) in a retrospective review described this group as all having normal speech development. It should be noted, however, that the lack of comment of a speech delay in the medical notes is inconclusive. Unless there is an ongoing study with a speech therapist routinely assessing the speech and language development of this group, one cannot conclude categorically that this group is always free of problems in their speech development.

Surgical management includes the initial evaluation, tracheostomy and subsequent monitoring of the child requiring surgical reconstruction, although most of this presentation concerns reconstructive techniques and results. Previous presentations have dealt with the initial endoscopic and radiological evaluation of children who present with stridor due to acquired subglottic stenosis: this includes microlaryngoscopy and bronchoscopy in order to determine the anatomical and functional situation in the larynx and tracheobronchial tree. Mr. Bull has discussed the techniques involved and Dr. Facer the anaesthetic aspects.