Systematic Review With Meta-analysis

Endoscopic Dilation Is Highly Effective and Safe in Children and Adults With Eosinophilic Oesophagitis

F. J. Moawad; J. Molina-Infante; A. J. Lucendo; S. E. Cantrell; L. Tmanova; K. M. Douglas


Aliment Pharmacol Ther. 2017;46(2):96-105. 

In This Article


After removal of 989 duplicates, a total of 3495 potential articles were identified. Of these studies, 3435 number were excluded by title and abstract review. The full text of the remaining 60 articles was retrieved and reviewed, upon which it was determined that 27 studies met inclusion criteria (Figure 1).

Figure 1.

Flow chart for the systematic review

Of these 27 studies, one was a randomised controlled trial,[25] two were prospective cohorts,[37,38] 16 were retrospective cohort studies,[17,21,23,24,26–29,39–47] two were case series and six were case reports.[19,46,48–53] All of the studies were single-center, except for one which included two centers.[21] Details are presented in Table 1 .

Overall, there were 2873 EoE patients, of which 1112 were children (<18 years). The mean age of patients in the studies was 32.5 years (SD: 11.8) with a range from 4 to 83 years. The mean percentage of male patients in the studies was 75.5% with a range of 65% to 90%.

Rings were the most common endoscopic feature, reported in a mean of 73% of patients per study (range: 32%-100%, 18 studies) followed by furrows with a mean of 60% (range: 11%-93%, 12 studies) and white plaques with a mean of 47% (range: 3%-73%, 14 studies).

All studies reported a frequency of dysphagia, with 19 studies reporting dysphagia in 100% of patients. The mean percentage of dysphagia in the remaining studies was 82% (range: 29%-96%, eight studies).[17,23,24,29,40–43] Food impaction was reported in 59% of patients (range: 9%-100%, 20 studies), heartburn in 22% (range: 7%-56%, 15 studies), chest pain in 10% (range: 0%-33%, 13 studies).

The most common medical treatment used was topical steroids (mean: 58%, range: 6%-100%, 15 studies), followed by PPI (mean: 56%, range: 12%-100%, 16 studies), and diet (mean: 12%, range: 0%-23%, 8 studies).

Among the studies, 845 EoE patients (87 children) underwent a total of 1820 oesophageal dilations with a median of three dilations per patient with a range from 1 to a maximum of 35 dilations per patient.[26] The method of dilation included 110 Maloney, 454 Savary, 768 TTS balloon, 20 Celestin, two EndoFLIP[49,52] and one with an upper endoscope.[17] Some studies reported the minimum target diameter achieved, which ranged from 15 to 20 mm.[21,26,28,29,37,40,41,45–48,50–53] Among the eight studies that reported oesophageal diameter before and after dilation, the mean pre-dilation lumen diameter was 9.9 mm (SD: 2.0) and the mean post-dilation diameter was 16.1 mm (SD: 2.8). Of all 27 studies, only two did not explicitly report the presence of stricture.[26,37] Eight studies described the oesophageal stricture location, which was most commonly found distally (73.6%), followed by proximally (14.6%), and then (11.6%) in the mid-oesophagus.[25,27,38,43,45,49,51,53]

Clinical improvement from dilation occurred in 95% of EoE patients following dilation (95% CI: 90%-98%, I2: 10%). Clinical response was similar between children (95% CI: 0.83–1.00, I2: 8.6%, 3 studies) and adults (95% CI: 0.89–0.99, I2: 15, 14 studies) (Figure 2). The duration of improvement was reported in 13 studies and with a median of 12 months and a range from 1 week to 36 months.[21,24,26–28,37,39,40,46,49–51,53] Follow-up rates exceeded 80% in all these studies except for 1, in which follow-up was 53%.[40] Sensitivity analysis conducted by excluding case series or case reports did not significantly alter the results (Figure 2).

Figure 2.

Forest plot for clinical outcomes including all studies. With inclusion of all studies, the summary effect for clinical improvement was 95%, I 2: 10%. After excluding case report and case series, summary effect was 90%, I 2: 32% (Figure S1)

Complications after dilation were rare. Perforations occurred in 0.38% (7/1831) (95% CI: 0.18–0.85 I 2: 27%, 27 studies), haemorrhage in 0.05% (1/1746) (95% CI: 0–0.3%, I 2: 0, 18 studies), and hospitalisation in 0.67% (12/1777) (95% CI: 0.3%-1.1%, I 2: 44%, 24 studies). There were no deaths reported in the studies (0/1831) (95% CI: 0–0.002, I 2: 0, 27 studies) (Figure 3). Significant heterogeneity was found for perforation (P=.097) and for hospitalisation (P=.01). Due to the event outcomes rarely occurring, stratified analysis could not be performed for major complication events between children and adults.

Figure 3.

Forest plot for the frequency of the four major complications with dilation

Chest pain not requiring hospitalisation was reported in the majority of studies and occurred in 9.3% of patients (142/1513) with a wide variation between studies ranging from 0.63% to 100%. When stratified by age, chest pain following dilation occurred in similar frequency between children (95% CI: 0.00–0.43, I2: 76%, four studies) and adults (95% CI: 0.00–0.22, I2: 94%, 15 studies) (Figure S2). In the Schoepfer study, chest pain was self-reported in 74% and considered mild, but noted in only 7% of their existing medical records.[21] There was a wide variation in the definition of laceration in the studies from mucosal disruptions following dilation to deep mucosal tears and therefore a summary estimate could not be calculated. Mucosal laceration following dilation was reported in as low as 0.6% of patients[26] and up to 100% of patients.[27]

Due to the low overall complications, data were insufficient to compare the frequency of major complication rates between the three major types of dilators (Maloney, Savary, TTS).