The effect of aryepiglottoplasty for laryngomalacia on gastro-oesophageal reflux

1. Introduction 

Stridor is noisy breathing caused by turbulent airflow in a partially obstructed airway. Gastro-oesophageal reflux (GOR) is the retrograde movement of stomach contents through the lower oesophageal sphincter into the oesophagus. GOR and airway obstruction, characterised by stridor, often occur together in children [1], [2], [3]. GOR is thought to be an inflammatory co-factor in various laryngeal disorders [4], [5]. Despite attempts to identify the mechanism responsible, no explanation for the link between GOR and stridor has been confirmed [4]. It is not certain whether GOR causes airway symptoms or vice versa. Possible explanations for the former theory include: induction of inflammatory supraglottic oedema, aspiration of gastric acid into the airway or reflex bronchospasm due to acid in the oesophagus stimulating an oesophagobronchial nerve reflex [2]. Alternatively, the increased negative intrathoracic pressure during inspiration with an obstructed laryngeal airway could predispose to GOR by overcoming the lower oesophageal sphincter [6]. Previous studies have shown that successful surgical treatment of GOR often fails to relieve respiratory problems [7]. This study attempts to discover whether the converse is true: does relief of airway obstruction lead to reduction of GOR?

Laryngomalacia is the commonest congenital laryngeal anomaly and the commonest cause of stridor in infants. Onset is usually shortly after birth but may be delayed for several weeks or months [8], [9], [10]; resolution often occurs by 18–24 months, but can be significantly later [11]. Laryngomalacia is characterised by inspiratory stridor, worse on agitation and lying supine. Endoscopy reveals an omega-shaped, tall, posteriorly placed epiglottis with short aryepiglottic folds. The mucosa and submucosa over the arytenoids is often excessive and this, together with the epiglottis, collapses inwards over the glottis on inspiration leading to partial airway obstruction. Most cases can be managed conservatively, but if symptoms are severe apnoea, hypoxia, failure to thrive and pectus excavatum may occur. In this situation aryepiglottoplasty (supraglottoplasty) is helpful. This involves surgical division of the short aryepiglottic folds [12], [13], [14], [15] to release the epiglottis and/or reduction of excess aryepiglottic fold mucosa [16] to prevent it from prolapsing over the glottis. Resection of the lateral edge of the epiglottis has also been described [17] but was not performed in any of these cases.

Several authors have reported that laryngomalacia is often associated with reflux, the incidence ranging between 23 and 80% [9], [18], [19], [20], [21]. Earlier studies based the diagnosis of GOR on contrast meal, oesophageal manometry, oesophageal biopsy or scintigraphy. Although contrast meal is a useful investigation in the diagnosis of GOR, it is limited by the short observation period imposed by radiation exposure, resulting in a poor sensitivity. All these techniques have now been superseded by continuous intra-oesophageal pH probe monitoring. The stomach normally secretes acid at a pH of 1.5–2.0, which contrasts with the luminal environment of the oesophagus where pH is almost neutral (pH 6.0–7.0). Distal oesophageal pH will thus decrease dramatically when GOR occurs. Continuous pH probe monitoring measures GOR quantitatively over a 24-h period [20]. This test is considered the gold standard in the diagnosis of GOR with a sensitivity of 87–93% and specificity of 93–97% [21]. Normal reflux values for children and infants have been established [22], [23], allowing pathological reflux to be separated from physiological reflux.

By performing 24-h pH studies pre- and post-operatively in children with laryngomalacia undergoing aryepiglottoplasty for relief of stridor, we attempted to ascertain whether eradication of stridor with relief of airway obstruction also results in reduction of reflux.

2. Methods 

Having obtained ethical approval for the study from the Great Ormond Street Hospital for Children NHS Trust/Institute of Child Health Research Ethics Committee, written parental consent was obtained for all participating children. Infants with stridor and a clinical diagnosis of severe laryngomalacia, due to undergo diagnostic microlaryngobronchoscopy and possible aryepiglottoplasty were recruited to the study. None of the subjects had GOR secondary to intestinal obstruction or dysmotility. They all had a pre-operative 24-h oesophageal pH study and if the diagnosis of laryngomalacia was confirmed by microlaryngobronchoscopy and an aryepiglottoplasty was performed, a second pH test was carried out between 2 and 4 weeks post-operatively.

The pH study was a 24-h recording using a Zinetics 24-h multi-use pH catheter (Synetics Medical Ltd., Middlesex, UK).

No infants were being treated with anti-reflux medication prior to the pre-operative pH test. In those with confirmed reflux post-operative medication was given. This was ranitidine and domperidone, which were discontinued prior to the post-operative pH test for 24 and 48 h, respectively.

Surgical success in relieving the airway obstruction was ascertained post-operatively in all infants who underwent aryepiglottoplasty by the cessation of stridor, satisfactory oxygen saturation and normal feeding.

The 24-h oesophageal pH study recorded: the reflux index (RI), the number of reflux episodes and the duration of the longest reflux episode before and after airway surgery.

For each child the RI was calculated pre- and post-operatively. RI is the percentage of time that the pH is below 4 during the whole investigation, as defined by The European Society for Pediatric Gastroenterology and Nutrition Working Group (ESPGAN) [24]. Normal reflux index results according to age have been described by Vandenplas [23], [25]. We recorded the change in RI, as a result of surgery, for each child.

3. Results 

Nine infants entered the study with ages ranging from three days to six months. All had laryngomalacia confirmed on microlaryngobronchoscopy. Their pre-operative reflux indices ranged from 2.4 to 20.9 with a mean of 11.6. According to the normal age ranges described by Vandenplas [24], six had age-significant GOR (Table 1). This incidence of 66% GOR in patients with laryngomalacia is comprable with previous series which found incidences of 23–80% [9], [18], [19], [20], [21].

Table 1. The prevalence and post-operative change of age-significant dastro-oesophageal reflux

	Child	Age	Reflux index pre-op	Reflux index post-op	Significant reflux index for age?	Reduction in reflux index post-op
	C.S.	6 months	20.9	11.6	Yes	Significant
	S.H.	6 months	18.4	10.2	Yes	Significant
	O.T.	9 weeks	17.7	4.5	Yes	Significant
	A.L.	5 months	19.3	no consent	Yes	N/A
	B.A.	3 days	11.2	no op	Yes	N/A
	S.L.	3 months	7.6	no op	Yes	N/A
	T.M.	5 months	4.2	1.8	No	Non-significant
	C.T.	9 weeks	3.3	2.5	No	Non-significant
	K.D.	4 months	2.4	3.4	No	Non-significant

Of the nine children who entered the study, six underwent aryepiglottoplasty and post-operative pH recording. The other three children did not complete the study. In two cases this was because although the diagnosis of laryngomalacia was confirmed by microlaryngobronchoscopy, the clinical condition was not severe enough to justify aryepiglottoplasty and the reflux was treated medically; in the other case it was because the parents were unwilling to give consent to a second, post-operative pH study. All three of these children were in the age-significant reflux group.

Of the six children who did proceed with the study, three had significant pre-operative reflux according to age [24]. Each child acted as its own control, pre- and post-operatively.

To analyse the data, which was normally distributed, the paired t-test was used to compare the RI pre- and post-operatively in each child. Two groups of patients became apparent, those with an age-significant pre-operative RI and those with non-significant RI for age (Table 1). In the group with age-significant reflux, the RI fell markedly as a result of surgery and this was statistically significant (paired t-test, p=0.02). In the group without age-significant reflux, there was no significant change in the RI (paired t-test, p=0.53) (Fig. 1). In all cases the stridor resolved after airway surgery.

View Large Image Download to PowerPoint

Fig. 1. In children with age-significant gastro-oesophageal reflux, the Reflux Index fell post-operatively. In those children without age-significant rewflux it did not change.

4. Discussion 

The majority of infants with laryngomalacia (in this series 66%) have significant co-existing GOR. The mechanisms causing GOR and laryngomalacia are incompletely understood. GOR has previously been thought of as an important inflammatory co-factor in laryngomalacia [4], [5], causing or exacerbating airway symptoms. However, relief of GOR by medical or even surgical treatment, does not always lead to improvement of respiratory symptoms[7]. This study approached the problem from the opposite direction, investigating whether GOR improves after surgical relief of partial airway obstruction. One theory is that partial airway obstruction causes powerful thoraco-abdominal pressure gradients, overcoming the usual anti-reflux barrier function of the lower oesophageal sphincter, leading to GOR. Research in animal models has shown this to be the case [6].

5. Conclusions 

This study shows that when partial airway obstruction due to laryngomalacia co-exists with GOR, treatment of the airway problem improves respiratory symptoms in all cases but only reduces GOR in patients with age-significant reflux. This suggests that there are two clinical groups, those with severe, age-significant reflux, possibly caused by airway obstruction, whose GOR benefits from aryepiglottoplasty; and those with physiological reflux which is not influenced by aryepiglottoplasty. If this is the case, and further larger studies are needed for confirmation, aryepiglottoplasty can be expected to reduce GOR in those infants with laryngomalacia who have age-significant reflux.