Bronchoscopy in paediatric intensive care

Endobronchial toilet 

Lobar and segmental collapse should first be treated with conventional physiotherapy. Inhaled and intratracheal recombinant human DNAase (rhDNase) has been reported anecdotally to be of benefit in segmental collapse even in children who do not have cystic fibrosis.19., 20., 21. If this fails, bronchoscopy is performed. First, unexpected airway pathology may be detected: the collapse may, for example, actually be caused not by mucus plugging but by vascular compression, which is not treatable by physiotherapy. The instrument with the largest suction channel should be used provided safe ventilation can be maintained.22 Samples should be sent for bacteriology and cytology and, if the plugs are very tenacious, the bronchoscopist should consider instilling rhDNase under direct vision.

Tube patency and position 

Even if a suction catheter passes easily, the tube may be blocked; this can be seen at bronchoscopy and remedied (Fig. 1). If there is any doubt, it is usually simpler to change the tube, but there may be reluctance to do this if the intubation was a difficult one. Tube position may be determined endoscopically, thus saving a chest radiograph, which is obviously attractive. Note that it is easy to confuse the subcarina between the right upper lobe and the right bronchus intermedius with the main carina; there is a learning curve to this examination and at least 20 examinations may be needed before the operator is confident.23 If in doubt, the tube should be pulled back 1–2cm under endoscopic control.

View Large Image Download to PowerPoint

Figure 1. Extensive mucus plugging of a tracheostomy tube which was said to suction easily (Photograph courtesy Dr Ian Balfour-Lynn).

Difficult intubation 

This is a rare indication since conventional intubation is usually easily accomplished.24 If, however, the child has an unstable cervical spine, midfacial disease (for example one of the craniofacial syndromes, such as Apert’s syndrome) or mandibular hypoplasia (Treacher Collins syndrome or Pierre Robin sequence), bronchoscopic intubation may be indicated. The child should be lightly anaesthetised and pre-oxygenated. The bronchoscope is threaded through an appropriately sized endotracheal tube, passed rapidly through the nose and into the trachea. Note should be taken of any unexpected pathology, for example subglottic stenosis. The endotracheal tube is then advanced over the bronchoscope into position. The use of MaGill’s forceps to pass the tube over the bronchoscope and beyond the vocal cords may be helpful. The bronchoscopist should then swiftly check the final position of the endotracheal tube. This procedure may be extremely difficult, and if a problematic intubation is anticipated, facilities for emergency tracheostomy should be available.

A similar technique may be used for an elective tube change; the bronchoscope is threaded through the new tube, and positioned just above the larynx. The old tube is removed, the bronchoscope passed through the vocal cords and the new tube railroaded as above.

Selective intubation 

This may be indicated if there is unilateral pathology necessitating very different ventilator strategies, for example a unilateral uncontrolled air leak caused by a bronchopulmonary fistula or pulmonary interstitial emphysema. The procedure is technically much easier if the long left main bronchus is to be intubated; selective right-sided intubation risks occluding the right upper lobe bronchus or leaving an unstable tube position. The anaesthetist electively changes the endotracheal tube for one of an appropriately smaller size to fit into a main airway. This in itself may destabilise a precarious child. The bronchoscope is then passed rapidly down the inside of the endotracheal tube and advanced into the target main bronchus. The endotracheal tube is then advanced over into position, the end result being checked first bronchoscopically and second with a chest radiograph.

Ventilator associated pneumonia 

This is one of the most difficult situations in the paediatric intensive care unit. Clinical and chest X-ray signs are non-specific and there are many other causes of infiltrates, for example pulmonary oedema, aspiration and segmental mucus plugging.25., 26., 27. Furthermore, the trachea is inevitably colonised by Gram-negative rods within a few days of intubation, with a risk of contaminating cultures taken from lower down the airways. There is no agreed gold standard for diagnosis. Diagnostic methods include blind tracheal aspirates, blind airway lavage, blind protected specimen brush, and bronchoscopy with lavage, protected specimen brush or protected lavage.

An important recent randomised controlled trial in adults28 reported on the use of invasive techniques to diagnose ventilator-acquired pneumonia. A total of 413 adult patients from 31 intensive care units were studied; 209 were treated on the basis of an examination of tracheal aspirates, and 204 had immediate bronchoscopy with either a lavage or a protected specimen brush, on which the treatment decisions were made. Of these, 86% (not particularly surprisingly) had positive tracheal aspirates, but only 43% had a positive bronchoscopy. The bronchoscopy group received significantly fewer antibiotics (mean number of antibiotic-free days 11.5±9.0 vs. 7.5±7.6; P<0.001) but had a significantly reduced mortality (hazard ratio 1.54, 95% CI 1.10–2.16, P=0.01). A further point made in this study is that a negative bronchoscopy may lead to an accelerated search for foci of infection elsewhere, with possible benefit from an earlier diagnosis. This study has clearly established that invasive bronchoscopic diagnosis is worthwhile at least in the context of adult intensive care; it seems unlikely that a similar paediatric study will ever be mounted, and we should consider making greater use of bronchoscopy in suspected paediatric cases of ventilator-acquired pneumonia.

Another situation in which an invasive diagnosis may need to be pursued is in the ventilated immunocompromised host. Blind airway lavage is appropriate as a first investigation as many of the probable pathogens (for example, Pneumocystis carinii, Nocardia, cytomegalovirus and Candida) are unlikely to be tracheal commensals. Pulmonary haemorrhage can be easily identified by the blind technique. If, however, a Gram-negative pneumonia is suspected, bronchoscopic diagnosis is preferable. A further advantage of bronchoscopy is that unexpected airway pathology may be seen and biopsied, for example the ulcerating lesions of invasive aspergillosis (Fig. 2) or tuberculous lymph nodes invading the airway. Furthermore, there is evidence from adult studies that bilateral lavage may increase the diagnostic yield;29 this is unlikely to be successful using a blind technique. The increased disturbance in gas exchange from multiple-site lavage should be borne in mind, although in this paper there were no adverse consequences even in ventilated patients. Overall, this author would prefer bilateral lavage to a single blind aspirate if both are available in an intubated, immunocompromised child. If a good-quality lavage is negative and the child is deteriorating, it is better to move to an open lung biopsy rather than repeat the lavage.30

View Large Image Download to PowerPoint

Figure 2. Sloughing endobronchial ulcers in a child with leukaemia complicated by invasive aspergillosis.

Assessment of lobar collapse 

The differential diagnosis of lobar collapse is wide and includes vascular compression (especially by the left atrium of left main bronchus), mucus plugging, diaphragmatic palsy in the spontaneously breathing child, consolidation, pleural effusion and a subphrenic abscess. Bronchoscopy should be considered after, for example, abdominal and pleural space ultrasound in the investigation of these children. Any airway mucus is dealt with by suction, aspiration and rhDNase instillation, as above.

The investigation of possible airway malacia requires particular anaesthetic expertise (Table 2). The child should be studied breathing spontaneously through a face mask, with minimal added PEEP, to determine the baseline airway calibre. Next, the highest acceptable level of PEEP should be applied, usually via a laryngeal mask or endotracheal tube. If the airway opens to a normal calibre, the problem is one of malacia; if the lumen remains impaired, extrinsic compression is the problem. Airway malacia and calibre may also be assessed with a low-contrast volume bronchogram (Fig. 3),31 which, in the intubated child, does not even require a bronchoscopy. A related indication is the investigation of localised hyperlucency. The pathology may be similar to lobar collapse, with a mucus plug acting as a ball valve, or unexpected airway pathology may be seen, for example an endobronchial polyp (Fig. 4).32

Table 2. Summary of investigation and management of suspected airway malacia in the intensive care unit.

	Problem	Clinical features	Management options
	Airway malacia	Reversible with PEEP	Long-term tracheostomy ventilation
			Stenting
			(Aortopexy)
	Airway compression	Not reversible with PEEP	Surgical relief of compression
			Stenting

PEEP, positive end-expiratory pressure.

View Large Image Download to PowerPoint

Figure 3. Low contrast volume bronchogram, digital subtraction technique, in a child with a pulmonary artery sling. Note the narrowed segment of trachea (arrowed) due to complete cartilage rings.

View Large Image Download to PowerPoint

Figure 4. Endobronchial polyp (arrowed) in right bronchus intermedius. This caused localised hyperinflation on the chest radiograph and computed tomogram.

Airway stent assessment 

First, the cause and severity of airway compression should be determined (see above). Stenting in children is still an experimental procedure, although in selected cases it may offer a good alternative to prolonged tracheostomy ventilation. Stent placement is normally a surgical procedure using a rigid bronchoscope. After the stent has been inserted, the position can be checked and the state of any distal malacia assessed, using fibre-optic bronchoscopy. Post-stent endobronchial toilet can be performed but usually needs to be repeated; the stent severely interferes with the normal mucociliary escalator. Anecdotally, regular nebulised rhDNase may be helpful in children with a regular accumulation of secretions.

Stridor on extubation 

Bronchoscopy is indicated if there has been at least one failed extubation apparently caused by upper airway obstruction (Fig. 5) and with apparently good weaning parameters and a trial of dexamethasone.9 The differential diagnosis includes airway malacia (either acquired as a complication of prolonged ventilation or resulting from pre-existing disease) and respiratory muscle disease (acquired as a result of phrenic nerve trauma during cardiac surgery or caused by a pre-existing and previously unrecognised myopathic process). Respiratory muscle dysfunction will not of course result in stridor.

View Large Image Download to PowerPoint

Figure 5. Severe subglottic stenosis after prolonged intubation in a neonate.

The clinical situation will determine whether a subsequent trial of extubation should be undertaken, with pre-treatment using dexamethasone. If bronchoscopy is requested, there is a strong temptation to ‘have a quick look’ while the child is extubated in the paediatric intensive care unit; this should almost always be resisted. A little thought reveals that the intensivists will wish the child to be vigorous and ready to breathe; the bronchoscopist will want a child to be heavily sedated or even anaethetised if good-quality information is to be obtained. In small children, the bronchoscope may nearly totally occlude the airway (above) and if the child is struggling against a closed airway, there is a real risk of pneumothorax or, in the preterm neonate, a cerebral haemorrhage. Furthermore, the author’s personal experience is that the endotracheal tube almost invariably contains thick strands of mucus despite attempts at airway suction. These take time to clear even with a bronchoscope, during which time the chances of obtaining useful information are very small. It is strongly recommended that such procedures be undertaken in an anaesthetic room with the help of an experienced consultant anaesthetist.

The baby is given a formal general anaesthetic with a short-acting agent and is extubated to face mask ventilation. The bronchoscope is passed through the face mask into a nostril. The airway is gently but thoroughly cleared of secretions and is inspected. The subglottic region is first examined from above the cords to avoid trauma from the bronchoscope to what may be a critical airway. Airway malacia can be documented in the spontaneously breathing infant, as can the effect of positive airway pressure applied by the anaesthetist. The child is then either allowed to wake up from anaesthesia while breathing spontaneously or, if the airway is judged to be critical, is reintubated either with a bladed laryngoscope or under direct vision using the bronchoscope.