Bronchial balloon occlusion in children with complex pulmonary air leaks
- Division of Respiratory Medicine, British Columbia Children's Hospital, Vancouver, British Columbia, Canada
- Correspondence to Dr Michael Seear, Room 1C31, British Columbia Children's Hospital, 4480 Oak Street, Vancouver, BC, V6H 3V4, Canada;
- Received 19 June 2012
- Accepted 9 November 2012
- Published Online First 6 December 2012
Pulmonary air leaks in children are most commonly due to infection or barotrauma. While cases of severe barotrauma are falling because of advances in neonatal care, the incidence of necrotising pneumonia is rising. The majority of air leaks can be managed conservatively, but more severe cases pose a significant challenge to the clinician. The use of occlusive endobronchial balloons is an established anaesthetic technique for a number of indications, but is not widely used in children. We conducted a review over a 12-year period, and report six cases of complex air leaks in which balloon occlusion was used. Balloon occlusion was successful in both cases of bronchopleural fistulae (secondary to severe necrotising pneumonia) and half of the cases of intrapulmonary air leak (due to barotrauma). In the other two cases (due to barotrauma and filamin A deficiency), it was transiently effective. No serious adverse effects or complications were encountered. In selected cases, endobronchial balloons are a useful adjunct in the management of life-threatening bronchopleural fistulae and cystic lung disease. The procedure is non-operative, minimally invasive and reversible. With the increasing incidence of bronchopleural fistulae, this may become an increasingly important therapy.
The use of various forms of endobronchial occlusion, in combination with single-lung ventilation, is an infrequently used but well-established technique in paediatric anaesthesia.1 The usual indication is the isolation of a single diseased lung during surgery so that blood or infected material does not contaminate the relatively healthy ventilated lung.2 In keeping with many anaesthetic techniques, the method has slowly moved into the world of intensive care. Possibly because of the limited range of indications and the lack of small double-lumen tubes, it has been used more commonly in adult rather than paediatric patients.3
Apart from the use of fetal tracheal occlusion in specialised centres for the treatment of congenital diaphragmatic hernias,4 the principal paediatric indication has been its use in the management of complicated pulmonary air leaks. The limited paediatric literature consists almost entirely of case reports. A few describe endobronchial occlusion as a treatment for bronchopleural fistulae (BPF),5 almost all of which were caused by necrotising pneumonia.6 A slightly larger neonatal literature describes the use of the technique in managing pulmonary interstitial emphysema (PIE) secondary to barotrauma.7
While modern ventilatory techniques have reduced the complications of barotrauma in infants,8 the incidence of necrotising pneumonia in children,9 and associated BPF,10 ,11 are both increasing.12 As a result, we believe indications for the use of endobronchial balloon occlusion will also increase over the next few years. Since we have found balloon occlusion to be a useful treatment in selected cases of pulmonary air leak, we describe our institution's experience with the technique over the last 12 years, including our indications, outcomes and side effects.
We performed a review of all cases where an occlusive bronchial balloon was used in British Columbia's Children's Hospital over a 12-year period (1999–2011). They were all managed by one of two paediatric respirologists. The occlusive balloons were inserted by one of two paediatric radiologists. In each case, the procedure was explained carefully to parents, and informed consent was obtained.
A previously well 6-year-old girl presented with multiorgan failure secondary to haemolytic uraemic syndrome triggered by serotype 19A pneumococcal pneumonia.13 She required significant ventilatory and inotropic support, as well as continuous renal replacement therapy. Over the next few days, she developed progressive complications of barotrauma including severe subcutaneous emphysema, pneumomediastinum and acute tension pneumothorax (figure 1, left). The source was probably a necrotic area in the right upper lobe. Despite chest tube drainage and maximal ventilation, the child continued to deteriorate to the point that extracorporeal life support (ECLS) was considered. She was too unstable for surgical lobectomy, so an attempt was made to occlude the main air leak with a balloon. Under image guidance, a pigtail guidewire was passed alongside the endotracheal tube. A 6 Fr balloon angiography catheter was threaded over it and positioned in the right upper lobe bronchus using fluoroscopy guidance. The fistula was located by inflating the balloon in different locations until the pleural chest tube stopped bubbling. A small volume of diluted contrast was used to ensure that the balloon was completely occluding the airway. The patient improved rapidly after successful occlusion of the leak (figure 1, middle). The balloon was left in situ for 9 days to prevent further air leak and to allow the fistula to heal. It was deflated after 4 days but required re-inflation because of recurrence of the air leak. The balloon required repositioning on one occasion following accidental displacement, but no other complications were encountered. After balloon deflation and chest tube removal, the patient was extubated on day 24 and made a good recovery (figure 1, right).
A previously well 2-year-old girl had a month-long admission for pneumonia complicated by empyema and a BPF. She was managed with intravenous antibiotics and two chest tubes. Unfortunately, shortly after discharge, she contracted influenza A with reaccumulation of air and fluid in the right pleural space. After insertion of a chest tube under anaesthesia, she deteriorated rapidly, probably as a result of aspiration of pleural fluid via a significant right BPF (figure 2, left). She required venovenous ECLS to support her respiratory function. Despite this, she remained unstable because of a persistent large air leak from the right lung.
On day 3 of ECLS, a 5 Fr balloon catheter was passed into the right main bronchus under radiological guidance, using the technique described above. Once inflated, the air leak stopped and the patient's condition improved (figure 2, middle). The balloon required repositioning on two occasions, but no other complications were encountered. On day 5 of endobronchial blockade, bronchography identified two air leaks in the right upper lobe, which were embolised with cyanoacrylate adhesive. A second balloon was inflated after this procedure and left in situ for a further 3 days to protect the repaired fistulae. The patient made a full recovery and was discharged home 3 weeks after the discontinuation of ECLS (figure 2, right).
A 3-month-old term female infant had a brief admission for bronchiolitis. At routine follow-up a month later, she was found to be hypoxic (saturations 80%) and failing to thrive. A chest x-ray showed marked hyperinflation, particularly of the left upper lobe (figure 3, left). She was admitted for further investigation and found to have widespread emphysematous changes on chest CT and lung biopsy. Additional investigations revealed right ventricular hypertrophy, an atrial septal defect, and bilateral subependymal nodular heterotropia. A provisional diagnosis of filamin A deficiency14 was made and the patient was discharged home on oxygen.
Two weeks later she presented with marked respiratory distress and desaturations necessitating ventilatory support. She underwent endobronchial occlusion of the left upper lobe bronchus using a 5 Fr balloon catheter. This produced clear radiological improvement but little clinical change. Hyperinflation recurred when the balloon was deflated. Because of significant neurological involvement, the child was not considered a candidate for surgical lobectomy. The baby deteriorated further and subsequently died at 8 months of age. Filamin A deficiency was confirmed by genetic testing.
A preterm (33-week gestation) male infant (birth weight 2390 g) required ventilation for apnoeas and surfactant deficiency. On day 1 of life, he developed a right-sided tension pneumothorax and, subsequently, right lower lobe PIE secondary to barotrauma (figure 4, left). In order to protect the localised area of barotrauma from further injury, a 4 Fr balloon catheter was inserted and inflated in the right bronchus intermedius. After the procedure, the right middle and lower lobes collapsed completely (figure 4, middle), and it became possible to wean the patient to lower ventilatory pressures. Following deflation of the balloon after 4 days, the right middle and lower lobes re-aerated and the area of PIE had improved considerably. In fact, the child was extubated to room air on the day of balloon deflation. He was transferred back to his local hospital a week later and discharged home soon afterwards (figure 4, right).
A term female infant with trisomy 21 and complex congenital heart disease developed a tension pneumothorax while mechanically ventilated for acute respiratory distress. She required two chest tubes and went on to develop severe PIE of the left lung (figure 5, left). In an attempt to reduce damage to the left lung, a 5 Fr balloon catheter was placed in the left main bronchus on day 26. The balloon was inflated and left in situ for 8 days (figure 5, middle). Following balloon removal, the left lung re-inflated with apparent improvement in the severe barotrauma (figure 5, right). The child was extubated to continuous positive airway pressure (CPAP)2 weeks later, but died from cardiac causes a month afterwards. During that period, the cystic areas in the left lung had slowly reaccumulated. Autopsy was refused so it was not possible to determine if any of the cystic areas were due to a congenital cystic malformation.
A 27-week gestation male triplet (birth weight 1200 g) developed increasingly severe barotrauma to his left lung. By day 26, he had become difficult to ventilate because of severe hyperinflation of the left lung with mediastinal shift and compression of the right lung (figure 6, left). In an attempt to improve his respiratory mechanics, a 4 Fr balloon catheter was placed in the left main bronchus and inflated under fluoroscopic guidance. Subsequent collapse of the left lung was followed by rapid clinical improvement (figure 6, right). The balloon was left inflated for 7 days with no side effects except that it needed repositioning twice because of migration. The cystic changes in the left lung improved markedly after balloon deflation. Hyperinflation slowly recurred 3 weeks later. It was decided not to repeat the balloon occlusion, so the child underwent an uneventful left upper lobectomy. He made a good recovery and was discharged home after a month.
The six cases we describe can be broadly split into two groups: two children had BPF secondary to severe necrotising pneumonia, and four infants had significant intrapulmonary air leaks. Three of these were the result of barotrauma, while one was due to an intrinsic deficiency in the lung microstructure caused by filamin A deficiency.
The procedure was successful in both cases of BPF, reducing the length of ECLS course in one child and avoiding ECLS altogether in the other. In the second group, it was successful in two children. For the remaining two children, the endobronchial balloon was transiently effective in reducing lung hyperinflation, but unfortunately the cystic changes recurred when the balloons were deflated. In both of these cases, the failure of the balloon did not alter the overall outcome of the child.
No serious adverse effects or complications were encountered in any of the patients. For each patient, the balloon was inflated carefully with the minimum volume of full- or half-strength contrast required to occlude the airway. There was no evidence of bleeding or bronchial injury noted in any of the cases. Three of the balloons migrated resulting in a recurrence of the air leak, but were easily repositioned in the interventional radiology department. One balloon deflated while in situ, without adverse effect, and a second balloon was used to replace it.
We performed a retrospective study of the use of endobronchial occlusion balloon catheters in our paediatric institution over a 12-year period. While this will never be a common procedure, we found that there were clear clinical indications for its use outside its more traditional application as an anaesthetic technique during thoracic surgery.1 When used for a few specific indications, by appropriately trained staff, the procedure has clinical benefits with minimal side effects. Although our sample was small, the indications for endobronchial balloon occlusion broadly fell into two groups: BPF secondary to necrotising pneumonia, and intrapulmonary air leaks secondary to barotrauma. There is surprisingly little published research in the paediatric literature regarding the use of endobronchial occlusion for either of these indications.
There are reliable data reporting an increase in paediatric necrotising pneumonia9 with BPF10 ,11 in recent years. McKee et al noted an increase from 1% among a cohort in 2002–2007 to 33% in patients studied in 2008–2009.10 Hsieh et al found that the incidence of BPF was 16% in cases of culture-proven pneumococcal pneumonia within the last decade.11 Based on the literature and our own local experience, we suspect that post-infectious BPF will become an increasingly common paediatric problem in the coming years. While conservative management with antibiotics and chest tube drainage is sufficient for most cases, there will always be severe cases of persistent fistulae that require further intervention. In the two cases we have described, endobronchial occlusion was a successful and safe treatment that allowed far riskier therapies to be avoided, such as surgery in one case and ECLS in the other. On the basis of this experience, we would suggest a trial of endobronchial balloon occlusion, before advancing to surgical lobectomy for a persistent BPF in a child.
Advances in the management of preterm infants have meant that pulmonary air leaks secondary to barotrauma have become much less common in recent years.8 However, preterm birth is so common that it is likely that complications such as PIE and cystic air leaks will continue to be a management issue. The limited available literature would suggest that balloon occlusion and one-lung ventilation have a part to play in the management of significant PIE.7 Our experience would certainly confirm that observation. In our cases, the use of endobronchial occlusion produced benefit with minimal risk. Half the children had permanent resolution of their air leaks while the remainder at least had a transient improvement. Again, there were no significant side effects.
The potential adverse effects of bronchial occlusion include retained secretions with distal infection, necrosis, erosion and/or stenosis of the bronchial wall, airway perforation during wire-guided catheter insertion, and migration of the balloon with major airway occlusion.5 ,15 ,16 Most of these can be avoided with careful management of the balloon. Care must, of course, be taken to avoid overdistension during inflation. In cases of BPF, it is possible to titrate the balloon volume by observing the point where chest tube air leak ceases. In small infants with PIE, the best guide is caution. We usually inflate with half-strength contrast to improve balloon visibility on x-ray. Others inflate with air because of concerns about leakage of contrast in the event of balloon rupture.5 Once in place, the balloon should be regularly deflated for short periods of time to reduce pressure on the bronchial walls. The catheter must be fixed carefully to the patient's face and the endotracheal tube. Regular observations of balloon position using a mark on the catheter must be made. The impressive list of potential problems clearly dictates the need for caution. However, with access to good standards of interventional radiology, backed by meticulous paediatric intensive care nursing, we have found endobronchial balloon occlusion to be a safe procedure.
This is the first published case series in which endobronchial balloon catheters have been used across the full range of paediatric indications. Even in a busy referral centre, it is unlikely that there will be more than 1–2 cases per year, so this first review is necessarily small. However, we hope our work acts as a stimulus for further examination of this technique. In carefully selected cases, we believe endobronchial balloon occlusion has a definite place in the management of children with severe complications of barotrauma or necrotic lung infections. We found it to be minimally invasive, safe and effective when used for clear indications. We hope our experience will be of value to other centres considering the use of this procedure.
Dr Ashley Robinson and Dr Gordon Culham, Interventional Radiologists at British Columbia Children's Hospital, Vancouver, who performed all the balloon placement procedures.
Contributors MS and DW were the paediatric respirologists responsible for the care of all six patients. CH and NA collated the clinical information and images for all the cases. All authors contributed to the writing and revision of the paper. MS is the guarantor for the paper.
Competing interests None.
Provenance and peer review Commissioned; externally peer reviewed.