Predictors of Response to Proton Pump Inhibitor Therapy among Children with Significant Esophageal Eosinophilia
Article Outline
Objectives
To determine predictors of histological response to proton pump inhibitor (PPI) therapy among children with significant esophageal eosinophilia (SEE), defined as ≥15 eosinophils per high powered field (eos/hpf) on esophageal mucosal biopsy (EMB).
Study design
Response to PPI therapy among children with SEE treated with PPI who underwent repeat EMB was studied retrospectively. Response was defined as <5 eos/hpf on repeat EMB. Characteristics of responders and nonresponders were analyzed.
Results
Of 326 patients (ages 1 through 18 years) diagnosed with SEE over a 7-year period, 43 (mean age, 8.5 years; 67% males) met inclusion criteria. After PPI therapy, 17 patients (40%) were responders. There were no significant differences in demographics, presenting symptoms, endoscopic, or histological findings between responders and nonresponders. Among patients with 15 to 20 eos/hpf on EMB, 50% were responders; among patients with >20 eos/hpf on EMB, 29% were responders. Seven of 17 (41%) patients with abnormal pH monitoring and 5 of 11 (45%) patients with normal monitoring were responders.
Conclusions
Forty percent of patients with SEE demonstrated histological response to PPI therapy. None of the clinical characteristics evaluated predicted response, and response was not dependent on results of pH study. The role of PPI therapy in treating SEE warrants further prospective investigation.
Abbreviations: BCH, Basal cell hyperplasia, EGD, Esophagogastroduodenoscopy, eos/hpf, Eosinophils per high powered field, GERD, Gastroesophageal reflux disease, PPI, Proton pump inhibitor, SEE, Significant esophageal eosinophilia
The normal human esophageal epithelium is devoid of eosinophils.1 Early investigations suggested that esophageal eosinophilia is related to exposure of esophageal epithelium to gastric acid; most of their patients had mild esophageal eosinophilia.2 Subsequent studies have included patients with symptoms of gastroesophageal reflux disease (GERD) and esophageal eosinophilia who failed conventional anti-GERD treatment.3, 4, 5 These patients often had significant esophageal eosinophilia (SEE) (≥15 eos/hpf), and, when performed, normal 24-hour continuous monitoring of intra-esophageal pH.3, 4, 5 This observation led to alternative theories of the cause of esophageal eosinophilia. In 1995, Kelly et al6 proposed an allergic etiology to SEE. Use of an elemental diet led to resolution of SEE in a group of 10 children who had symptoms of GERD but failed to respond to anti-GERD treatment. Subsequent reports have supported an allergic cause for SEE by demonstrating successful treatment with systemic and topical corticosteroids, elimination and elemental diets, and biologic agents such as anti-interleukin 5.5, 6, 7, 8, 9, 10, 11, 12
In clinical practice, the treatment of SEE has trended away from anti-GERD therapy to anti-allergy therapy, though questions regarding the role of GERD in the pathogenesis of SEE remain.5, 6, 7, 8, 9, 10, 11, 13 Baxi et al14 classified patients with ≤5 eos/hpf on esophageal mucosal biopsies (EMB) as having GERD, patients with ≥15 eos/hpf as having eosinophilic esophagitis, and patients with 6 to 14 eos/hpf as a “gray zone” which might represent underlying GERD, eosinophilic esophagitis, or a combination of the 2 disorders. Ngo et al15 reported 3 patients with SEE who responded, both symptomatically and histologically, to proton pump inhibitor (PPI) therapy. They recommended a trial of PPI as initial therapy in all patients with SEE. This raises important questions about the role of PPI in managing SEE and highlights a need for further investigation.
We wished to further investigate the response of patients with SEE to PPI therapy and attempt to identify factors that might predict response to PPI therapy.
Methods
All patients between the ages of 1 and 18 years who underwent esophagogastroduodenoscopy (EGD) between 1999 and 2006 by the Division of Pediatric Gastroenterology, Hepatology, and Nutrition at the James Whitcomb Riley Hospital for Children, Indiana University School of Medicine, Indianapolis, Indiana, and demonstrated SEE (defined as ≥15 eos/hpf on distal and/or mid-esophageal biopsy) were identified. The indications for EGD included regurgitation, vomiting, heartburn, abdominal pain, dysphagia, and sensation of food impaction. Patients with SEE were eligible for this study if they were treated with PPI (omeprazole, esomeprazole, or lansoprazole) and underwent repeat EGD to assess response to PPI. Patients were excluded if they were receiving corticosteroid, dietary elimination, or montelukast therapy for any indication. Patient demographics, past medical and surgical history, endoscopy results, histological data, and results of allergy testing were accessed using medical records. The Institutional Review Board of Indiana University School of Medicine approved this study.
Endoscopy and Histology
Endoscopic findings were recorded from dictated reports with particular attention to the presence of vertical lines, ulcers, white specks, erosions, hiatal hernia, and esophageal stricture. Endoscopic photographs were reviewed by one of the authors (J.E.D.) and blindly by another author (S.K.G.). Esophageal mucosal biopsies were obtained from the distal esophagus (3 cm cephalad to the esophagogastric junction) in all patients and the mid-esophagus (8 cm cephalad to the esophagogastric junction) in some patients. Biopsy specimens were fixed in 10% buffered formalin and stained with hematoxylin and eosin. Pathology reports were reviewed for documentation of basal cell hyperplasia (BCH) and eosinophil concentration. Basal cell hyperplasia was documented as the percentage of the epithelial cell layer occupied by basal cells. The maximal numbers of esophageal epithelial eosinophils were counted in a single hpf (×40 magnification, field area of 0.4 mm2).
Definition of Response
Response to PPI therapy was defined as ≤5 eos/hpf on all esophageal biopsies obtained during repeat EGD. Nonresponse to PPI therapy was defined as >5 eos/hpf on any esophageal biopsy (distal or mid-esophagus) obtained during repeat EGD.
pH Study
Twenty-four to 48-hour continuous monitoring of intraesophageal pH (pH study) using standard catheter (Sandhill Scientific, Highlands Ranch, Colorado) or wireless technique (Bravo System, Medtronic, Minneapolis, Minnesota) was performed. pH study results were included if the monitoring was performed on the same day as initial EGD showing SEE, or at a time after initial EGD but before commencing treatment with PPI. Histamine-2 receptor antagonists and prokinetic medications were discontinued for 72 hours before pH study, and PPI therapy was stopped for 7 days prior. A reflux index of >6.0%, indicating an esophageal pH of <4.0 for >6% of the time monitored, was considered abnormal.16
Statistical Analysis
Two-sample t tests and χ2 tests were used to analyze the data. The significance level for all data was set at P < .05.
Results
Patients with SEE
A total of 326 patients between ages 1 and 18 years were diagnosed with SEE during the study period. Of these, 43 patients (mean age, 8.5 years; median age, 10 years; 67% males) met the study inclusion criteria, with the oldest child age 17 years. After PPI therapy, 17 of the 43 patients (40%) were responders and 26 patients (60%) were nonresponders (Table I). There were no significant differences in the demographics or presenting symptoms between responders and nonresponders. The vast majority of patients (86%), irrespective of histologic response, reported symptom improvement at follow-up EGD. There was no seasonal variation in response. PPI doses were similar between responders and nonresponders.
Table I. Demographics, symptoms, and endoscopic findings
| Responders (n = 17) | Nonresponders (n = 26) | P value | |
|---|---|---|---|
| Males | 10 (59%) | 19(73%) | NS |
| Mean age (range), y | 8.41(1-16) | 8.58(1-17) | NS |
| Asthma | 4(24%) | 2(8%) | NS |
| Positive SPT | 1/5(20%) | 4/13(31%) | NS |
| Presenting symptoms⁎ | |||
| Vomiting/regurgitation | 12(71%) | 17(65%) | NS |
| Abdominal pain | 11(65%) | 12(46%) | NS |
| Heartburn | 2(12%) | 2(8%) | NS |
| Dysphagia | 2(12%) | 4(15%) | NS |
| Improved or resolved | 16(94%) | 21(81%) | NS |
| Endoscopic findings | |||
| Initial EGD abnormal | 16(94%) | 26(100%) | NS |
| Vertical lines | 9(53%) | 19(73%) | NS |
| Stricture | 0(0%) | 2(8%) | NS |
| Erosions | 2(12%) | 3(12%) | NS |
| Hiatal hernia | 2(12%) | 2(8%) | NS |
| Follow-up EGD normal | 6(35%) | 2(8%) | 0.024 |
| Mean PPI dose (range), mg/kg/d | 1.06(0.5-1.88) | 0.99(0.47-1.78) | NS |
| Mean interval between EGDs (range), mo | 5(2-9) | 4.5(1-9) | NS |
| Season of presentation: | |||
| Winter | 6(35%) | 10(38%) | NS |
| Spring | 4(24%) | 9(35%) | NS |
| Summer | 3(18%) | 4(15%) | NS |
| Fall | 4(24%) | 3(12%) | NS |
⁎Some patients had >1 presenting symptom. |
Endoscopy
Gross endoscopic esophageal abnormalities were noted at baseline in 94% of responders and 100% of nonresponders (Table I). After PPI treatment, 35% of responders and 7.7% of nonresponders had improvement to normal gross endoscopic findings on follow-up EGD (P = .024). The presence of vertical lines, erosions, or hiatal hernia on initial EGD did not predict histologic response to PPI treatment.
Histology
Maximum eos/hpf and degree of BCH on initial EGD were similar between responders and nonresponders (Table II). As expected, there was significant improvement in eosinophil concentration in responders as compared with baseline, but not in nonresponders.
Table II. Histological findings on initial EGD
| Responders (n = 17) | Nonresponders (n = 26) | P value | |
|---|---|---|---|
| Mean eos/hpf (range) | 28.6(15-80) | 31.0(15-57) | NS |
| Mean eos/hpf (range) mid-esophagus | 19.3(0-65) | 30.4(7-57) | NS |
| Mean eos/hpf (range) distal esophagus | 28.2(15-80) | 25(10-40) | NS |
| BCH (range) mid-esophagus, % | 31(0-80) | 51.7(0-100) | NS |
| BCH (range) distal esophagus, % | 63.4(25-100) | 72.5(25-100) | NS |
| No. of patients with 15-20 eos/hpf | 11(65%) | 11(42%) | NS |
| No. of patients with >20 eos/hpf | 6(35%) | 15(58%) | NS |
| Mean eos/hpf (range) on follow-up EGD | 1.2(0-4) | 26.6(7-60) | <.0001 |
A trend toward response was seen in patients with maximum 15 to 20 eos/hpf on EMB (mid and/or distal), of whom 50% were responders, as compared with patients with >20 eos/hpf on any EMB, of whom 29% were responders (P = .130) (Table II). Demographics and clinical characteristics of patients with 15 to 20 and >20 eos/hpf were not statistically different (Table III). Within the 15- to 20-eos/hpf group, responders had significantly less BCH in mid-esophageal biopsies than did nonresponders (mean, 22.9% vs 58.4%) (P = .028). Within the >20 eos/hpf group, responders had a significantly higher distal esophageal eosinophil concentration as compared with nonresponders (mean, 48.3 vs 30.1 eos/hpf) (P = .007). No other significant differences between responders and nonresponders were noted among patients with 15 to 20 eos/hpf and >20 eos/hpf.
Table III. Comparison of patients with 15-20 eos/hpf vs >20 eos/hpf
| 15-20 Responders (n = 11) | 15-20 Non-responders (n = 11) | >20 Responders (n = 6) | >20 Nonresponders (n = 15) | |
|---|---|---|---|---|
| Males | 7(64%) | 7(64%) | 3(50%) | 12(80%) |
| Mean age (range), y | 8.27(1-16) | 8.91(1-17) | 8.67(2-13) | 8.93(3-14) |
| Presenting symptoms⁎ | ||||
| Vomiting/regurgitation | 9(82%) | 6(55%) | 3(50%) | 11(73%) |
| Abdominal pain | 8(73%) | 6(55%) | 3(50%) | 6(40%) |
| Dysphagia | 2(18%) | 1(9%) | 0 | 3(20%) |
| Heartburn | 1(9%) | 0 | 1(17%) | 1(7%) |
| Mean eos/hpf on initial EGD (range) | ||||
| Mid-esophageal | 9.29(0-20) | 13.4(7-20) | 42.7(25-65) | 39.1(30-57) |
| Distal esophageal | 17.3(15-20) | 18.2(10-20) | 48.3(30-80)† | 30.1(10-40) |
| BCH on initial EGD, % (range) | ||||
| Mid-esophageal | 22.9(0-70)‡ | 58.4(27-100) | 50(0-80) | 46.9(0-100) |
| Distal esophageal | 55.3(25-100) | 71.6(33-100) | 78.3(30-100) | 73.6(25-100) |
| Endoscopic findings | ||||
| Vertical lines | 6(55%) | 9(82%) | 4(67%) | 9(60%) |
| Stricture | 0(0%) | 0(0%) | 0(0%) | 2(13%) |
| Erosions | 2(18%) | 0(0%) | 0(0%) | 3(20%) |
| Hiatal hernia | 1(9%) | 0(0%) | 0(0%) | 1(7%) |
| Asthma | 2(18%) | 0(0%) | 2(33%) | 2(33%) |
| Positive SPT | 1/3(33%) | 1/3(33%) | 0/2(0%) | 3/11(27%) |
⁎Some patients had >1 presenting symptom. |
†>20 eos/hpf distal esophagus eos/hpf R vs NR, P = .007. |
‡15-20 eos/hpf mid-esophagus R vs NR BCH, P = .028. |
pH Study Results
Twenty-eight of 43 patients underwent pH study per inclusion criteria. Overall, 7 of 17 (41%) patients with abnormal pH study and 5 of 11 (45%) patients with normal pH study responded to PPI treatment. Among patients with an abnormal pH study, a trend in response was noted among those with 15 to 20 eos/hpf (5/7) as compared with those with >20 eos/hpf (2/10) (P = .052). Among patients with a normal pH study, there was no significant difference in response among those with 15 to 20 eos/hpf (3/8) vs those with >20 eos/hpf (2/3).
Discussion
Our study examines the response to PPI therapy in a large cohort of children with SEE. The use of PPI as initial therapy for SEE has recently resurfaced on the basis of the report by Ngo et al15 of 3 patients with SEE who responded symptomatically and histologically to PPI therapy. This, coupled with the recent recommendation from the First International Gastrointestinal Eosinophil Research Symposium17 to treat SEE with a PPI in lieu of a pH study, prompted a review of our experience in treating SEE with PPI. Our goal was to identify factors that may contribute to successful histological treatment of SEE with PPI.17 Although no statistically significant predictors of response were identified, in univariate analysis, an eosinophil count of 15 to 20 eos/hpf trended toward predicting response, particularly when in association with an abnormal pH study, indicating that this subgroup of patients might suffer primarily from GERD and therefore benefit from treatment with PPI, as opposed to patients with >20 eos/hpf. This is important because a specific recommendation to first expose the majority of patients with SEE to PPI, before other therapies, has significant implications on patient management, efficient health care delivery, and associated costs.
Our study supports the notion that symptom resolution may not reflect objective response to PPI therapy in patients with SEE. A subset of patients with SEE experience symptomatic relief with PPI.18, 19 In our study, 91% of patients with SEE obtained symptomatic relief with PPI, though only 40% demonstrated an objective histological response (Table I).
In adults, SEE unresponsive to conventional GERD therapy has classically been described in patients presenting with dysphagia.20, 21 In children, however, presenting symptoms appear to be less predictive of response to GERD treatment. In 2 studies by Liacouras et al,5, 22 the majority of children with SEE who were nonresponsive to conventional GERD therapy did not have isolated dysphagia. They had symptoms that overlapped with classic GERD, including vomiting, regurgitation, and abdominal pain in up to 82% of patients. Our study focuses on the use of PPI and reports that the response of SEE to PPI therapy in children is not predicted by presenting symptoms.
A number of endoscopic findings are reported in SEE including vertical lines/furrows, white specks, and rings.19, 23, 24 Overall, we did not observe any differences in endoscopic appearance of esophageal mucosa between responders and nonresponders. Esophageal mucosal histological findings were similar between all responders and nonresponders (Table II). As such, we are not able to set an absolute threshold eosinophil concentration or degree of BCH for success of PPI treatment among patients with SEE.
Studies vary in their definition of SEE from ≥10 eos/hpf to ≥25 eos/hpf,10, 25 with ≥15 and ≥20 the most commonly used cutoff values.17 We arbitrarily studied the 15- to 20-eos/hpf range as a possible “gray zone” in which acid reflux might play a primary role in the esophageal eosinophilic inflammation. A trend in response to PPI therapy among children with 15 to 20 eos/hpf as compared with >20 eos/hpf (50% vs 29%) was noted. These findings indicate that the use of PPI, especially in patients with 15 to 20 eos/hpf, deserves further prospective evaluation, as this subgroup may represent a separate phenotype of SEE.
An extensive review of published literature demonstrates that 90% of patients (children and adults) with SEE have a normal pH study.17 We analyzed the association of response of SEE to PPI and results of pH monitoring, expecting that an abnormal pH study would predict favorable response to PPI. We observed that only 41% of those with abnormal pH study were responders, and 45% of those with a normal pH study were responders, indicating that an abnormal pH study does not appear to predict response to PPI and a normal pH study does not preclude response to PPI. Subanalysis revealed that among patients with 15 to 20 eos/hpf, 71% of patients with an abnormal pH study were responders as opposed to only 38% with normal pH study. Although not statistically significant, it appears that GERD is more likely to be the primary cause of SEE among patients with 15 to 20 eos/hpf, and the pH study may be helpful in predicting response to PPI among this group.
As with any retrospective study, this project carries some limitations. Treatment and evaluation of patients with SEE at our institution during the study period was not by standard protocol. Not all patients were treated with PPI, doses were not uniform, and not all patients treated with PPI underwent repeat endoscopy. Subsequently, only 12% of patients with SEE (43/326) were eligible for analysis. We cannot exclude a selection bias in the presentation of our data. It is conceivable that some of the treating physicians were biased toward either an acid-related or an allergic etiology for their endoscopic/histologic observations, and treated accordingly. Konikoff et al10 report a 10% placebo response in children with SEE in a study comparing placebo with swallowed fluticasone. Therefore, a possible significant placebo response to PPI can not be ignored. White specks identified on endoscopy, now an accepted manifestation of SEE, probably were under-reported or under-recognized, especially in the early years of study. A uniform time to follow-up endoscopy was not followed, and this could confound the response rates, considering that the natural history of remission and relapse of SEE has yet to be described in detail.
In summary, we extensively analyzed a large cohort of children with SEE who were treated with PPI in an effort to identify any predictors of response. No statistically significant factors predicting response were identified between responders and nonresponders. On univariate analysis, we noted a trend in response to PPI therapy in patients with 15 to 20 eos/hpf, particularly those with an abnormal pH study. Response to PPI therapy was poor among patients with >20 eos/hpf, including those having an abnormal pH study. Prospective controlled studies examining the role of PPI in patients with SEE are needed, as arbitrary initial treatment of all patients with SEE patients with PPI with a subsequent endoscopy to assess response will impose a burden on patient health and healthcare costs.
The authors thank Joseph Fitzgerald, MD, and Jean Molleston, MD, for their expert advice on preparation of this manuscript, and Sharon McPheeters for her excellent administrative support.
References
- . Oesophagitis: epidemiology, clinical aspects, and association to allergy. J Pediatr Gastroenterol Nutr. 2007;45:281–289
- . Intraepithelial eosinophils: a new diagnostic criterion for reflux esophagitis. Gastroenterology. 1982;83:818–823
- . Severity of esophageal eosinophilia predicts response to conventional gastroesophageal reflux therapy. Pediatr Develop Pathol. 1999;2:15–18
- . Esophageal eosinophilia with dysphagia: a distinct clinicopathologic syndrome. Dig Dis Sci. 1993;39:109–116
- . Primary eosinophilic esophagitis in children: successful treatment with oral corticosteroids. J Pediatr Gastroenterol Nutr. 1998;26:380–385
- . Eosinophilic esophagitis attributed to gastroesophageal reflux: improvement with an amino acid based formula. Gastroenterology. 1995;109:1503–1512
- . Treatment of eosinophilic esophagitis with inhaled corticosteroids. J Pediatr Gastroenterol Nutr. 1998;27:90–93
- Comparison of prednisone and fluticasone in the treatment of eosinophilic esophagitis: a randomized trial in children. Clin Gastroenterol Hepatol. 2008;6:165–173
- . Review article: the pathogenesis and management of eosinophilic oesophagitis. Aliment Pharmacol Ther. 2006;24:173–182
- A randomized, double blind, placebo controlled trial of fluticasone propionate for pediatric eosinophilic esophagitis. Gastroenterology. 2006;131:1381–1391
- . Elemental diet is an effective treatment for eosinophilic esophagitis in children and adolescents. Am J Gastroenterol. 2003;98:777–782
- . Anti-IL-5 (mepolizumab) therapy for eosinophilic esophagitis. J Allergy Clin Immunol. 2006;118:1312–1319
- High intraepithelial eosinophil counts in esophageal squamous epithelium are not specific for eosinophilic esophagitis. Am J Gastroenterol. 2008;103:435–442
- . Clinical presentation of patients with eosinophilic inflammation of the esophagus. Gastrointest Endosc. 2006;64:473–478
- . Eosinophils in the esophagus: peptic or allergic eosinophilic esophagitis? (Case series of three patients with esophageal eosinophilia). Am J Gastroenterol. 2006;101:1666–1670
- Guidelines for evaluation and treatment of gastroesophageal reflux in infants and children: recommendations of the North American Society for Pediatric Gastroenterology and Nutrition. J Pediatr Gastroenterol and Nutr. 2001;32:S1–S31
- Eosinophilic esophagitis in children and adults: a systematic review and consensus recommendations for diagnosis and treatment. Gastroenterology. 2007;133:1342–1363
- . The spectrum of pediatric eosinophilic esophagitis beyond infancy: a clinical series of 30 children. Am J Gastroenterol. 2000;95:1422–1430
- . The ringed esophagus: histological features of GERD. Am J Gastroenterol. 2001;96:984–989
- . Eosinophilic oesophagitis: a novel treatment using Montelukast. Gut. 2003;78:830–835
- . Eosinophilic esophagitis in adults: clinical, endoscopic histologic findings, and response to treatment with fluticasone propionate. Gastrointest Endosc. 2006;63:3–12
- Eosinophilic esophagitis: a 10-year experience in 381 children. Clin Gastroenterol Hepatol. 2005;3:1198–1206
- . Vertical lines in distal esophageal mucosa (VLEM): a true endoscopic manifestation of esophagitis in children?. Gastrointest Endosc. 1997;45:485–489
- White specks in the esophageal mucosa: an endoscopic manifestation of non-reflux eosinophilic esophagitis in children. Gastrointest Endosc. 2004;59:835–838
- . Marked eosinophilia in esophageal mucosal biopsies. Am J Surg Pathol. 1985;9:475–479
The authors declare no conflicts of interest or grant support.
PII: S0022-3476(08)00605-7
doi:10.1016/j.jpeds.2008.07.042
© 2009 Mosby, Inc. All rights reserved.
