ORIGINAL ARTICLES ALIMENTARY TRACT. Esophageal Eosinophilic Infiltration Responds to Proton Pump Inhibition in Most Adults

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2011;9: ORIGINAL ARTICLES ALIMENTARY TRACT Esophageal Eosinophilic Infiltration Responds to Proton Pump Inhibition in Most Adults JAVIER MOLINA INFANTE,* LUCIA FERRANDO LAMANA, CRISTINA RIPOLL, MOISES HERNANDEZ ALONSO,* JOSE M. MATEOS,* MIGUEL FERNANDEZ BERMEJO,* CARMEN DUEÑAS,* NURIA FERNANDEZ GONZALEZ, EVA M. QUINTANA, and MARIA ANGELES GONZALEZ NUÑEZ Departments of *Gastroenterology and Pathology, Hospital San Pedro de Alcantara, Caceres; and Department of Digestive Diseases, Hospital Gregorio Marañon, Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD), Madrid, Spain See Editorial on page 99. BACKGROUND & AIMS: Despite consensus recommendations, eosinophilic esophagitis (EoE) is commonly diagnosed upon esophageal eosinophilic infiltration (EEI; based on 15 eosinophils per high power field; eo/hpf). We evaluated the prevalence of EEI before and after proton pump inhibitor (PPI) therapy and assessed the accuracy of EEI and ph monitoring analyses. METHODS: Biopsies were taken from the upper-middle esophagus of 712 adults with upper gastrointestinal symptoms who were referred for endoscopy due to upper gastrointestinal symptoms. Patients with EEI were treated with rabeprazole (20 mg, twice daily) for 2 months. EoE was defined by persistent symptoms and 15 eo/hpf following PPI therapy. RESULTS: Thirty-five patients (4.9%) had EEI, of whom 55% had a history of allergies, and 70% had food impaction or dysphagia as their primary complaint. Twenty-six EEI patients (75%) achieved clinicopathological remission with PPI therapy; of these, 17 had GERD-like profile (EEI 35 eo/hpf and objective evidence of reflux, based on endoscopy or ph monitoring), and 9 had EoE-like profile (EEI eo/hpf, typical EoE symptoms and endoscopic findings). The PPI response was 50% in the EoE-like profile patients. The PPI-response was 50% in EoE-like profile patients. Likewise, PPIresponsive EEI occurred with normal (33%) and pathologic (80%) ph monitoring. Higher histologic cut-off values improved specificity and positive predictive for EoE (35% 35% for 20 eo/hpf; 46% 39% for 24 eo/hpf; 65% 50% for 35 eo/hpf). CONCLU- SIONS: In adults with EEI, 75% of unselected patients and 50% with an EoE phenotype respond to PPI therapy; ph monitoring is poorly predictive of response. Patients with PPI-responsive EEI >35 eo/hpf are phenotypically undistinguishable from EoE patients. EoE might be overestimated without clinical and pathologic follow-up of patient response to PPI. Keywords: Allergy; Eosinophil; Eosinophilic Esophagitis; Esophageal ph Monitoring; Gastroesophageal Reflux Disease; Proton Pump Inhibitors; Sensitivity; Specificity. Eosinophilic esophagitis (EoE) is an emerging inflammatory esophageal disorder in adults which usually presents in young allergic males with dysphagia and food impaction. Current standard for diagnosis remains esophageal biopsies revealing esophageal eosinophilic infiltration (EEI) with a peak count over 15 eosinophils per high power field (eo/hpf), correlated with a compatible clinical context. 1 Gastroesophageal reflux disease (GERD) constitutes its main differential diagnosis, although eosinophilic infiltration in this condition is believed to be minor ( 7 eo/hpf) and confined to the distal esophagus. 2,3 According to the current consensus statement, 1 histological suspicion of EoE should always be confirmed by unresponsiveness to high dose proton pump inhibitor (PPI) therapy or a normal ph esophageal monitoring. Nonetheless, the vast majority of studies concerning EoE have based the diagnosis exclusively on EEI, 4 including the largest series recently reported. 3,5 Indeed, few prospective studies have systematically ruled out GERD in EEI patients, using either a ph monitoring or a PPI trial. 2,6,7 On the other hand, recent small series or case reports in children and adults have concurrently highlighted the existence of symptomatic patients who met clinical, endoscopic, and histological criteria for EoE with complete clinicopathologic remission on PPI therapy Additionally, no differences in therapeutic response have been recently observed when comparing esomeprazole to topical steroids in adults with a diagnosis of EoE based on EEI. 13 Therefore, the major aims of this study were: (1) to determine the prevalence of EEI, before and after PPI therapy, both in the overall sample and in high risk subpopulations; (2) to evaluate the operating characteristics of current diagnostic parameters for EoE, comparing the accuracy of histology and ph esophageal monitoring to a systematic therapeutic trial of a PPI; and (3) to examine the distinguishing features among EEI responsive and nonresponsive to PPI therapy. Abbreviations used in this paper: CI, confidence interval; EEI, esophageal eosinophilic infiltration; eo/hpf, eosinophils per high power field; EoE, eosinophilic esophagitis; GERD, gastroesophageal reflux disease; GI, gastrointestinal; HPF, high power field; IgE, immunoglobulin E; IL, interleukin; OR, odds ratio; PPI, proton pump inhibitor; PPI-R, PPI responsive by the AGA Institute /$36.00 doi: /j.cgh

2 February 2011 ESOPHAGEAL EOSINOPHILIC INFILTRATION AND PPI 111 Methods Patients This prospective study was done in a secondary referral hospital in 2008, between January 1 and September 31. All consecutive adult patients (18 93 years of age) referred to the Endoscopy Unit at our center due to upper gastrointestinal symptoms were considered for enrollment in the study. All symptomatic patients who gave their consent and had no exclusion criteria were included. Exclusion criteria were: (1) different indications than upper gastrointestinal (GI) symptoms, such as anemia, weight loss, or diarrhea; (2) prior diagnosis of eosinophilic gastrointestinal disorder; (3) acid suppression therapy in the previous week before the endoscopic procedure. An empiric PPI trial before this time period was not an exclusion criteria. In clinical practice, we systematically request withdrawal from PPI at least 1 week previous to initial esophago-gastro-duodenoscopy (EGD); or (4) inability to take biopsies due to the presence of esophageal varices or active anticoagulant therapy. Immediately before the endoscopy, patients completed a symptom questionnaire assessing the presence of dysphagia, food impaction, heartburn, regurgitation, chest pain, nausea, or abdominal pain for the last 2 months, including the PPI withdrawal period. Patients also indicated the main symptom for endoscopic referral. When more than 1 symptom was present, the patients were requested to state which was predominant. The questionnaire (see Supplementary Figure 1) was developed by the investigators based on previous validated questionnaires for upper GI symptoms, owing to the lack of a uniform validated tool for symptoms in EoE. Anthropometric data, allergic history, and laboratory counts including leukocytes and percentage of peripheral eosinophils were also recorded before the procedure. Allergy was defined by the presence of atopic diathesis (rhinoconjunctivitis, asthma, atopic dermatitis) or known presence of antigen-specific immunoglobulin E (IgE) for foods or aeroallergens by skin prick test, atopic patch test, or plasma-specific IgE testing. 14 Upper GI Endoscopy Endoscopic explorations were carried out with topical pharyngeal anesthesia by 5 experienced endoscopists (JMI, MHA, JMM, MFB, and CD) in upper GI endoscopy. In the year previous to the initiation of the study, all endoscopists reviewed images from patients with classical findings of EoE and/or in whom histology revealed EEI. Furthermore, during the study period, JMI reviewed the literature on a bimonthly basis in order to identify new reports concerning the endoscopic diagnosis of EoE. Endoscopic diagnosis of EoE was considered with the presence of at least 1 of the following features: multiringed esophagus, linear furrows, whitish exudates, undulations, 1 or more strictures with either rings, furrows, or exudates, or blocking the passage of the endoscope with linear shearing or mucosal fragility ( crepe paper ) after contact with endoscopic instrumental and papular elevations. 2,15 Duodenal and antrum biopsies were taken in all patients to rule out other eosinophilic gastroenteritis. A minimum of 4 esophageal endoscopic biopsies were taken with a standard forceps. Esophageal samples were obtained from upper-mid esophagus 10 cm above the esophagogastric junction to avoid the theoretical GERD-related eosinophilic infiltration area. GERDrelated esophageal endoscopic injury was considered when reflux esophagitis (graded according to Los Angeles classification), peptic stricture (a single concentric and fibrotic distal narrowing), Barrett s esophagus, or esophageal adenocarcinoma were present. 16 Histology Mucosal biopsy specimens were fixed in formalin, embedded in paraffin, and stained with hematoxylin and eosin (H&E) for pathologic examination. All esophageal biopsies were blindly examined by 3 staff pathologists (NFG, EMQ, MAG) with special attention regarding eosinophil esophageal detection. Then, a senior gastrointestinal pathologist (LFL) carried out a meticulous examination of esophageal samples in a blinded fashion. One high power field (HPF) had an area of 0.24 mm 2. The peak count of intraepithelial eosinophils/hpf ( 400) was determined in the area of highest density of eosinophils by the most densely populated HPF and histological suspicion of EoE, defined by EEI, was established upon the presence of 15 eo/hpf in at least 1 field. 1 Major histopathologic criteria, which were degranulating eosinophils, superficial layering of eosinophils (preferential location of eosinophils in the upper-luminal surface of the epithelium), eosinophils microabscesses (foci of 4 or more aggregated eosinophils), and minor histological features, such as lamina propia fibrosis (marked thickening of subepithelial collagen layer), the detection of 5 eo/hpf in the fibrotic lamina propia, basal zone hyperplasia ( 30%), papillary elongation ( 70% of epithelium height), and intercellular edema, were also evaluated. 17 Follow-Up According to the endoscopic and histological findings, patients were subdivided into 3 groups: (A) patients without EEI on esophageal biopsies, which were considered the control population; (B) patients with proximal EEI and distal GERD features on endoscopy; and (C) patients having proximal EEI without GERD endoscopic findings. Patients with GERD on endoscopy (group B) did not undergo ph monitoring because the expectancy of pathologic esophageal acid contact time was very high (75% 90%), 18 whereas esophageal acid exposure was documented in patients without GERD stigmata (group C). PPI therapy was discontinued during 2 weeks prior to the esophageal ph recording. Prior stationary esophageal manometry was performed to define the upper limit of the lower esophageal sphincter. An antimonium single probe ph catheter was then positioned 5 cm above it. The results were interpreted in accordance with classical DeMeester and Johnson criteria, considering abnormal 24-hour esophageal acid exposure time 4% or DeMeester score Afterward, all patients with 15 eo/hpf were treated with rabeprazole 20 mg twice a day for 2 months and subsequently underwent follow-up endoscopy using the same protocol in clinical and histological data recording. Clinical response was defined by complete resolution of symptoms. Endoscopic remission was defined by disappearance of previous endoscopic findings. Histological response for patients with EEI was defined by an eosinophil count 5 eo/hpf. According to current guidelines, EoE was defined by persistent symptoms and 15 eo/hpf following PPI therapy. Statistical Analysis The SPSS 16.0 statistical analysis package (SPSS Inc, Chicago, IL) was used. Categorical variables are described with per-

3 112 MOLINA-INFANTE ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 2 Figure 1. Flow chart of patients, depending on histological, encoscopic, and ph monitoring data, summarizing the results of the study. b.i.d., twice daily. centages and continuous variables are described with mean (SD) or median (range) as appropriate. Associations between categorical variables were tested with the 2 test (with Fisher s correction when necessary) and continuous data were assessed using the 2 sample t test or Mann-Whitney U, for parametric and nonparametric data, respectively. Univariate and multivariate logistic regression analyses were performed to evaluate independent predictive variables for finding 15 eo/hpf in esophageal biopsies. The magnitude of the effect is described with the odds ratios (ORs) and 95% confidence intervals (CIs). Variables chosen to be introduced in predictive models depended on clinical relevance and statistical significance on univariate analysis. The gold standard for diagnosis of EoE was the lack of clinical or pathological response in those patients who had received high-dose acid suppression. The sensitivity, specificity, positive, and negative predictive values to predict this final diagnosis from the baseline histological data were determined. All patients gave their consent to participate in the study. Approval by the ethics committee of San Pedro de Alcantara Hospital was obtained. The manuscript was elaborated following the recommendations of the Standards for Reporting of Diagnostic Accuracy (STARD). 20 Results Baseline Characteristics and Prevalence of EEI During the study period, 1249 upper endoscopies were performed due to the presence of upper GI symptoms. Five hundred thirty-seven patients refused or were not eligible for enrollment and 712 patients (57%) were finally recruited (see Figure 1). Multiple biopsies (median 6.5 [range 4 9]) were taken while progressively extracting the endoscope along the proximal esophagus. Overall, 35 patients (4.9%) were found to have EEI. The baseline characteristics of the study population and patients after subclassification depending on 15 eo/hpf or 15 eo/hpf are summarized in Table 1. On the univariate regression analysis, male gender, age under 50 years, peripheral eosinophilia, allergy, dysphagia, or food impaction, emergency endoscopy due to bolus obstruction, endoscopic classical features of EoE, and all major and minor histopathological features of EoE were associated to the detection of EEI. On multivariate regression analysis, age under 50 years: OR, 5.9 (95% CI, ; P.01], allergic history (OR, 12.9; 95% CI, ; P.01), and endoscopic features of EoE (OR, 34.76; 95% CI, ; P.001) remained independent predictive factors of 15 eo/hpf on histopathological examination. The prevalence of EEI was 4.9% in the study population. Among patients under 50 years in the total cohort (n 333), EEI was found in 50% with food bolus obstruction (19/38) and 35% with dysphagia (21/57). Interestingly, it was detected in 3.5% and 8.7% of patients with clinical or endoscopic GERD, but notably increased in these subsets of patients when an allergic feature was present (25% and 36% of the cases, respectively). The low prevalence of EEI in patients with endoscopic findings of EoE (21/49, 43%) was mainly due to patients with multiringed esophagus

4 February 2011 ESOPHAGEAL EOSINOPHILIC INFILTRATION AND PPI 113 Table 1. Baseline Characteristics of the Total Cohort (n 712) and Patients With 15 eo/hpf and 15 eo/hpf Total 15 eo/hpf 15 eo/hpf P a Number of patients Demographics Male/female 366/346 (51%/49%) 287/296 (42%/58%) 25/10 (71%/29%).02 Age, y 52 (18 93) 53 (18 93) 35 (18 67).01 Body mass index (BMI) 26.4 (5.1) 26.4 (4.12) 26.3 (3.88).530 Peripheral eosinophilia 2.29 (1.8) 2.18 (1.75) 5.49 (2.39).01 Allergic history 76 (10.7) 57 (8.4) 19 (54.3).01 Rhinoconjunctivitis 28 (4) 19 (2.8) 9 (25).01 Asthma 35 (5) 25 (3.6) 10 (28.5).01 Aeroallergen sensitization 55 (7.7) 41 (6) 14 (40).01 Emergency endoscopy due to food bolus 71 (10) 56 (8.2) 15 (43).01 obstruction Heartburn history 428 (60) 404 (59) 24 (68).300 Prior PPI intake 462 (65) 448 (66) 14 (40).060 PPI in the month before withdrawal period 101 (14.1) 95 (14) 6 (17).470 Indication for endoscopy, n (%) Dysphagia 100 (14) 89 (13.1) 11 (28.5).01 Food impaction 81 (11.4) 67 (9.9) 14 (40).01 Heartburn 198 (27.8) 191 (28.2) 7 (14.2).290 Dyspepsia/abdominal pain 291 (40.9) 289 (42.7) 2 (7.40).01 Chest pain 6 (0.8) 5 (0.7) 1 (2.8).180 Nausea/vomiting 30 (4.2) 30 (4.4) Endoscopic findings, n (%) Normal endoscopy 347 (48.7) 341 (50.3) 6 (17.1).01 Endoscopic pattern of EoE 30 (4.2) 28 (4) 21 (60).01 Reflux distal injury 161 (22.6) 147 (21.71) 14 (40).110 Hiatal hernia/schatzki ring 163 (22.8) 159 (23.4) 4 (11.4).090 Histopathologic findings, n (%) Eosinophil superficial distribution 20 (2.8) 4 (0.5) 16 (45.7).01 Degranulating eosinophils 34 (4.7) 8 (1.1) 26 (74.3).01 Eosinophils microabscesses 10 (1.4) 0 10 (28.5).01 Basal cell hyperplasia 123 (17.2) 94 (13.9) 29 (82.9).01 Papillae elongation 90 (12.6) 64 (9.4) 26 (74.3).01 Intercellular edema 176 (14.6) 145 (22.74) 31 (88.5).01 Presence of lamina propia 364 (51.1) 344 (50.8) 20 (57.1).400 Lamina propia fibrosis 25 (6.8) 8 (2.3) 17 (85).01 5 eosinophils in lamina propia 16 (4.4) 3 (0.8) 13 (65).01 NOTE. Qualitative variables are expressed as absolute values and proportions. Continuous variables are expressed as mean (SD) or median (range) as appropriate. a Comparing patients with 15 eo/hpf and those with 15 eo/hpf. (10/27, 37%), whereas it was 90% (9/10) in patients with either furrows or exudates. Prevalence of EoE After universal PPI therapy in the 35 patients with proximal EEI at baseline histological examination, only 9 out of 35 (25%) had EoE according to the current standards (Figure 1). Twenty-six patients (75%) achieved clinicopathological remission with PPI therapy. Of note, when considering a more selected population with clinical and endoscopical findings compatible with EoE, only 50% of patients achieved remission using PPI therapy. These results are summarized in Table 2. Symptoms and Endoscopic Findings Before and After PPI Therapy The assessment of symptoms and endoscopic findings, correlated with histological remission, in patients with EEI, before and after PPI therapy, is available online as a supplementary file (see supplementary material online at org). Clinical remission was achieved in 75% of patients complaining of food impaction or dysphagia, albeit histological lesions persisted in 10% of them. Likewise, 60% (n 21) had endoscopic findings suggestive of EoE, but 70% (n 15) of them achieved clinicopathologic remission using PPI therapy. All endoscopic features of EoE disappeared after response to PPIs excepting rings, which were persistent in half of the cases despite clinicopathological remission. Of note, clinical remission occurred despite persistent eosinophilia in 5 out of 9 EoE patients. ph Esophageal Monitoring The global rate of documented pathologic acid exposure (n 15) or GERD-endoscopic damage (n 14) in patients with EEI was high (29/35, 82%). Two out of 6 patients with EEI (43 eo/hpf and 103 eo/hpf, respectively) and normal ph monitoring (acid contact time 2.7% and 3.7%, DeMeester Score 11.7 and 14,1, respectively; see Figure 1) also showed clinical

5 114 MOLINA-INFANTE ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 2 Table 2. Assessment of the Prevalence of EEI in High-Risk Groups Within the Total Cohort, Before and After PPI Therapy n Prevalence of EEI before PPI therapy Prevalence of EEI after PPI therapy Total cohort (upper GI symptoms), n (%) 712 (100) 35/712 (4.9) 9/712 (1.2) Food bolus obstruction 50 yrs, n (%) 38 (5.3) 19/38 (50) 7/38 (18) Dysphagia 50 yrs, n (%) 57 (8) 21/57 (36) 7/57 (12) Allergic history, n (%) 76 (10.7) 19/76 (25) 7/76 (9) Endoscopic signs of EoE, n (%) 49 (6.8) 21/49 (43) 6/49 (12) Dysphagia/food impaction and endoscopic signs of EoE, n (%) 35 (4.9) 15/35 (43) 6/35 (17) Dysphagia/food impaction, allergic history and endoscopic signs of EoE, n (%) 14 (1.9) 10/14 (71) 5/14 (35) and pathological remission after PPI therapy. In those with abnormal ph monitoring, complete remission was achieved in 12/15 (80%). Accuracy of Different Eosinophil Count Threshold for the Diagnosis of EoE From the histopathological point of view, acid suppression led to a statistically significant decrease of eosinophil peak count in patients with EEI (see Figure 2). All of the 17 patients between 15 and 35 eo/hpf were responsive to PPIs. EoE was diagnosed in 4/11 (36%) between 35 eo/hpf and 72 eo/hpf, whereas 5/6 (83%) between 72 eo/hpf and 165 eo/hpf had definite EoE. The specificity and predictive positive value of the 15 eo/hpf cutoff point and higher thresholds is listed on Table 3. Figure 2. Outcome of peak eosinophil count in esophagus biopsies in patients with EEI (n 35), before and after acid suppressive therapy. Eosinophil count (median and range) before proton pump inhibitor (PPI) therapy was 36 eo/hpf (15 165) and after PPI therapy was 3 eo/hpf (0 116) (P.01). Comparison of EEI Responsive and Nonresponsive to PPI Therapy The study population with 15 eo/hpf (n 35) were subclassified according to eosinophilic infiltration degree and response to acid suppression into 3 subgroups to analyze their characteristics (PPI-responsive [R] EEI eo/hpf [n 17], PPI-R EEI 35 eo/hpf [n 9] and EoE [n 9]). The results are shown in Supplementary Table 1. No pathognomonic features that distinguished these entities were identified, including major histopathologic criteria. However, 2 different subgroups were identified: (1) GERD-like with a low peak eosinophil count (15 35 eo/hpf) and heartburn as a major complaint (35%) with either endoscopic reflux distal injury in 58% and pathologic ph monitoring in the remaining 42%; and (2) EoE-like with high peak eosinophil count ( eo/hpf) and symptom and endoscopic profile identical to EoE (Table 4). Unfortunately, no statistical comparison could be done due to the small sample size. Discussion The present study is the first prospective investigation that evaluates the prevalence of EEI after systematic high dose acid suppression. The most noteworthy result is that 75% of unselected adults referred for endoscopy due to upper GI symptoms with proximal EEI achieved clinicopathological remission with PPIs and, therefore, the specificity of esophageal eosinophilia in the diagnosis of EoE is questionable. Interestingly, a recent retrospective study concluded that a similar 70% of patients with 20 eo/hpf were considered to have GERD according to clinical, endoscopic, and ph monitoring data. 11 The PPI response rate in patients with a more typical EoE phenotype (allergy, dysphagia, or food impaction and characteristic endoscopic findings) was 50%. EoE is an emerging novel allergic disease which has gained interest for adult gastroenterologists within the last decade. The prevalence of EoE in recent adult studies has been reported as 1.1% in an asymptomatic population, % in patients with upper gastrointestinal symptoms undergoing endoscopy, 22 10% in patients presenting with nonobstructive dysphagia 23 and even 50% in patients with food bolus obstruction. 24 All of these studies, as noted in the largest recent series, 3,5 established the diagnosis of EoE upon EEI, but no ph monitoring or PPI trials were performed according to consensus recommendations. We found similar prevalence of baseline EEI in the total cohort (4.9%), there was a notable 70% decrease after acid suppressive therapy (see Table 2). This observation raises the possibility that the prevalence of EoE might have been overestimated if based exclusively upon histological criteria. In our series, interestingly, EEI was more prevalent than expected in atopic patients with either GERD endoscopic injury (36%) or heartburn (25%). These patients did not have strictly clinical or endoscopic indications for esophageal biopsies. The diagnostic importance of EEI in this setting, irrespective of responsiveness to PPI, remains unclear. Regarding diagnostic criterion for EoE, our findings of PPI-R EEI in 33% and 80% of patients with normal and

6 February 2011 ESOPHAGEAL EOSINOPHILIC INFILTRATION AND PPI 115 Table 3. Estimated Accuracy of Different Eosinophil Cutoff Points for the Diagnosis of EoE n Sensitivity % (95% CI) Specificity % (95% CI) PPV % (95% CI) PNV % (95% CI) 15 eo/hpf (70 100) 0 (0 13) 26 (14 42) 20 eo/hpf (70 100) 35 (19 54) 35 (19 54) 100 (70 100) 24 eo/hpf (70 100) 46 (29 65) 39 (22 59) 100 (76 100) 35 eo/hpf (70 100) 65 (46 81) 50 (29 71) 100 (82 100) 50 eo/hpf 9 67 (35 88) 90 (74 96) 67 (35 88) 90 (74 96) PNV, predictive negative value; PPV, predictive positive value. pathologic ph monitoring, respectively, are similar to those obtained in a recent retrospective large cohort of children. 25 Therefore, a normal ph monitoring may not preclude GERD in patients with EEI, as stated in current guidelines. 1 As a matter of fact, up to 50% of nonerosive GERD and 10% 25% of erosive GERD may have a normal ph recording and dilation of intercellular spaces in GERD has been shown reversible using PPIs with physiological acid exposure. 18 On the other hand, EoE can coexist with GERD as evidenced by cohorts with persistent symptoms and esophageal eosinophilia with PPI therapy in spite of abnormal ph testing. 26 Other authors and our data have found an increased prevalence of GERD in adult patients with EEI (38% 56% 13,27 ). However, the high rate in our series was possibly due to the Table 4. Baseline Characteristics of Patients With EEI Subclassified According to Eosinophilic Infiltration Degree and Response to Acid Suppression PPI-R EEI eo/hpf 35 eo/hpf EoE Number of patients Demographics, n (%) Male/female 12/5 (70/30) 7/2 (77/23) 6/3 (66/33) Age, y 40 (18 67) 39 (21 61) 39 (24 57) Body mass index (BMI) Peripheral eosinophilia Allergic history 5 (29) 7 (77) 7 (77) Rhinoconjunctivitis 3 (17) 2 (22) 4 (44) Asthma 2 (11) 4 (44) 4 (44) Aeroalllergen sensitization 3 (17) 4 (44) 7 (77) Emergency endoscopy due to food bolus obstruction 6 (35) 7 (77) 4 (44) Indication for endoscopy, n (%) Dysphagia 3 (17) 4 (44) 4 (44) Food impaction 6 (35) 4 (44) 4 (44) Heartburn 6 (35) 0 1 (11) Dyspepsia/abdominal pain 1 (0.05) 1 (11) 0 Chest pain 1 (0.05) 0 0 Others Endoscopic findings, n (%) Normal endoscopy 3 (17) 1 0 Endoscopic pattern of EoE 2 (11) 5 (55) 5 (55) Reflux distal injury 10 (58) 2 (22) 2 (22) Hiatal hernia / Schatzki ring Others ph monitoring Normal/pathologic 0/7 2/5 4/3 Histopathologic findings, n (%) Superficial distribution 5 (29) 6 (66) 5 (55) Degranulating eosinophils 12 (70) 7 (77) 7 (77) Eosinophils microabscesses 2 (11) 3 (33) 5 (55) Basal cell hyperplasia 12 (70) 9 (100) 8 (88) Papillae elongation 12 (70) 7 (77) 7 (77) Intercellular edema 15 (88) 8 (88) 8 (88) Presence of lamina propia 9 (52) 4 (44) 7 (77) Lamina propia fibrosis 6 (35) 4 (44) 7 (77) 5 eosinophils in lamina propia 3 (17) 3 (33) 6 (66) NOTE. Qualitative variables are expressed as absolute values and proportions. Continuous variables are expressed as mean (SD) or median (range) as appropriate.

7 116 MOLINA-INFANTE ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 2 referral population characteristics, as approximately 15% 30% would be expected to have GERD. From a histopathological point of view, neither increasing eosinophils count thresholds (20 eo/hpf, 24 eo/hpf, 35 eo/ HPF) nor the presence of major histological criteria helped to distinguish accurately between GERD and EoE. Eosinophil microabscesses were not pathognomonic for EoE, as has been suggested in adults. 28,29 Thus, EoE cannot be exclusively diagnosed by histological criterion. Clinical remission with persistent eosinophilia was detected in half of the EoE cohort after PPI therapy, similarly to that reported in children after PPIs, steroids, or dietary modification. 25,30 Overall, these results suggest that a rigid distinction between GERD and EoE based on clinical, endoscopic, ph monitoring, or histopathological data may be too simplistic, 31 while a thoughtful clinicopathological correlation after systematic high dose PPIs seems the most accurate diagnostic tool in clinical practice. Emerging literature is consistent with the proposal of a novel eosinophilic type of GERD, which was initially described with eosinophils ranging between 7 and 24 eo/hpf. 32 The prevalence of proximal EEI in our study increased 4 6 fold in patients with heartburn or GERD-related endoscopic damage when allergic background was present. Therefore, GERD may cause high density eosinophil infiltration even in proximal esophagus. Similarly, recent immunohistochemical studies in children 33 and adults 29 have shown that 20% 50% of GERD biopsies may have 15 eo/hpf and increased expression of IgEbearing cells and activated mast cells. We failed to distinguish PPI-R EEI and EoE on the basis of an eosinophil count but identified a GERD-like, PPI-R EEI group (n 17) with low eosinophil count (15 35 eo/hpf). These patients had a high rate of clinical symptoms of GERD (35% heartburn as major complaint), endoscopic GERD (60%) and pathological ph monitoring in the remaining 40%. Similarly, this GERD-like profile (heartburn as presenting symptom in 7/8 patients and erosive esophagitis in 5/8 patients) has been reported for EoE with an eosinophil count below 25 eo/hpf. 34 A second PPI-R EEI group (n 9) with high eosinophil count ( eo/hpf) or EoElike, was indistinguishable from EoE in clinical, endoscopic, ph monitoring, and histological terms. Whether patients with PPI-R EEI are suffering from acid reflux-mediated esophagitis (GERD) or immune-mediated esophagitis (EoE) is still unknown. Responsiveness to PPI therapy, according to current guidelines, rules out EoE by definition. However, this statement is being questioned, 31 since emerging works are reporting in vitro anti-inflammatory effects of PPIs, independent of acid suppression. Our group demonstrated that omeprazole improved murine asthma by downregulating interleukin (IL)-4, IL-13 and Signal Transducer and Activator of Transcription Furthermore, omeprazole has been shown able to block in vitro IL-13 stimulated secretion of eotaxin-3 in esophageal line cells from EoE patients. 36 All of these questions should be addressed soon in order to refine the current definition of EoE. This study has several shortcomings. The small size of the sample of patients with EoE is a major drawback, which limits the statistical interpretation of the results. A relevant concern is that the cohort examined included all patients with upper GI symptoms and not only those with a typical symptom profile for EoE (dysphagia, food bolus obstruction, or chest pain usually on PPI therapy). Furthermore, patients using PPI therapy were excluded. Both issues could have led certainly to an underestimation of the prevalence of EoE. On the other hand, the inclusion of patients in the total cohort (101/712 [14%]) having PPI therapy in the month before the 1-week withdrawal period may have influenced the prevalence of GERD-like esophageal eosinophilia among the cohort. The high response rate, mainly in clinical terms, to PPI in our study should be viewed with caution because patients included sometimes had symptom profile and endoscopic features more characteristic of GERD than EoE. Nonetheless, most prevalence studies concerning EoE also included a significant percentage of patients with heartburn and acid regurgitation (47%, 3 40%, 22 and 54% 23 ), GERD as primary indication for endoscopy (27%, 5 and 23% 22 ), or reflux endoscopic injury (15%,,21 12%, 22 14%,,23 and 40% 24 ), but gave a diagnosis of EoE upon EEI without precluding GERD. Overall, our results are only generalizable to a population with similar characteristics as the study population, particularly regarding the prevalence. The potential mechanisms involved in the response to PPI therapy in patients with EEI (placebo effect, concurrent GERD, acid hypersensitivity, anti-inflammatory effect of PPIs) 26,30,36,37 are matters of ongoing research. The distinction between PPI-R EEI and EoE could have been substantiated by the use of novel biomarkers such as genetic targets (eotaxin-3, transcriptome analysis) and immunohistochemistry (anti eosinophil peroxidase). 38 Indeed, the results of gene expression profiling have been proposed recently for inclusion in the definition of EoE, instead of a PPI trial. 39 On the other hand, our biopsy protocol may have lead to an underestimation of the prevalence of EoE as a higher eosinophilia has been described for distal esophagus. However, no significant differences have been reported in diagnostic accuracy for EoE comparing biopsies in proximal and distal esophagus. 40 In conclusion, 75% of symptomatic adults undergoing esophago-gastro-duodenoscopy (EGD) with proximal EEI achieve complete clinicopathologic remission using acid suppressive therapy. The PPI response was 50% in the patients with an EoE-like profile. Diagnosis of EoE based only on histopathology and normal ph monitoring has a poor accuracy when compared with a PPI trial. Further studies evaluating the clinical, genetic, immunohistochemical, and ultrastructural features among EEI patients, responsive and nonresponsive to PPI therapy, are needed. Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at and at doi: / j.cgh References 1. Furuta GT, Liacouras CA, Collins MH, et al. Eosinophilic esophagitis in children and adults: a systematic review and consensus recommendations for diagnosis and treatment. Gastroenterology 2007;133: Lucendo AJ, Pascual-Turrión JM, Navarro M, et al. Endoscopic, bioptic, and manometric findings in eosinophilic esophagitis before and after steroid therapy: a case series. Endoscopy 2007; 39: Müller S, Pühl S, Vieth M, et al. Analysis of symptoms and endoscopic findings in 117 patients with histological diagnoses of eosinophilic esophagitis. Endoscopy 2007;39: Dellon ES, Aderoju A, Woosley JT, et al. Variability in diagnostic criteria for eosinophilic esophagitis: a systematic review. Am J Gastroenterol 2007;102:1 14.

8 February 2011 ESOPHAGEAL EOSINOPHILIC INFILTRATION AND PPI Kapel RC, Miller JK, Torres C, et al. Eosinophilic esophagitis: a prevalent disease in the United States that affects all age groups. Gastroenterology 2008;134: Attwood SE, Lewis CJ, Bronder CS, et al. Eosinophilic oesophagitis: a novel treatment using montelukast. Gut 2003;52: Arora AS, Perrault J, Smyrk TC. Topical corticosteroid treatment of dysphagia due to eosinophilic esophagitis in adults. Mayo Clin Proc 2003;78: Morrow JB, Vargo JJ, Goldblum JR, et al. The ringed esophagus: histological features of GERD. Am J Gastroenterol 2001;96: Desai TK, Stecevic V, Chang CH, et al. Association of eosinophilic inflammation with esophageal food impaction in adults. Gastrointest Endosc 2005;61: Ngo P, Furuta GT, Antonioli DA, et al. Eosinophils in the esophagus: peptic or allergic eosinophilic esophagitis? Case series of three patients with esophageal eosinophilia. Am J Gastroenterol 2006;101: Rodrigo S, Abboud G, Oh D, et al. High intraepithelial eosinophil counts in esophageal squamous epithelium are not specific for eosinophilic esophagitis in adults. Am J Gastroenterol 2008; 103: Sayej WN, Patel RA, Baker RD, et al. Treatment with high-dose proton pump inhibitors helps distinguish eosinophilic esophagitis from noneosinophilic esophagitis. J Pediatr Gastroenterol Nutr 2009;49: Peterson KA, Thomas KL, Hilden K, et al. Comparison of esomeprazole to aerosolized, swallowed fluticasone for eosinophilic esophagitis. Dig Dis Sci 2010;55: Assáad A. Eosinophilic esophagitis: association with allergic disorders. Gastrointest Endosc Clin N Am 2008;18: Fox VL. Eosinophilic esophagitis: endoscopic findings. Gastrointest Endosc Clin N Am 2008;18: Vakil N, van Zanten SV, Kahrilas P, et al. The Montreal definition and classification of gastroesophageal reflux disease: a global evidencebased consensus. Am J Gastroenterol 2006;101: Odze RD. Pathology of eosinophilic esophagitis: what the clinician needs to know. Am J Gastroenterol 2009;104: Long JD, Orlando RC. Non erosive reflux disease: a pathophysiologic perspective. Curr Gastroenterol Rep 2008;10: Hirano I, Richter JE, Practice Parameters Committee of the American College of Gastroenterology. ACG. Practice guidelines: esophageal reflux testing. Am J Gastroenterol 2007;102: Bossuyt PM, Reitsma JB, Bruns DE, et al. The STARD statement for reporting studies of diagnostic accuracy: explanation and elaboration. Ann Intern Med 2003;138:W1 W Ronkainen J, Talley NJ, Aro P, et al. Prevalence of oesophageal eosinophils and eosinophilic oesophagitis in adults: the population-based Kalixanda study. Gut 2007;56: Veerappan GR, Perry JL, Duncan TJ, et al. Prevalence of eosinophilic esophagitis in an adult population undergoing upper endoscopy: a prospective study. Clin Gastroenterol Hepatol 2009; 7: Prasad GA, Talley NJ, Romero Y, et al. Prevalence and predictive factors of eosinophilic esophagitis in patients presenting with dysphagia: a prospective study. Am J Gastroenterol 2007;102: Kerlin P, Jones D, Remedios M, et al. Prevalence of eosinophilic esophagitis in adults with food bolus obstruction of the esophagus. J Clin Gastroenterol 2007;41: Dranove JE, Horn DS, Davis MA, et al. Predictors of response to proton pump inhibitor therapy among children with significant esophageal eosinophilia. J Pediatr 2009;154: Shah A, Kagalwalla AF, Gonsalves N, et al. Histopathologic variability in children with eosinophilic esophagitis. Am J Gastroenterol 2009;104: Remedios M, Campbell C, Jones DM, et al. Eosinophilic esophagitis in adults: clinical, endoscopic, histologic findings, and response to treatment with fluticasone propionate. Gastrointest Endosc 2006;63: Parfitt JR, Gregor JC, Suskin NG, et al. Eosinophilic esophagitis in adults: distinguishing features from gastroesophageal reflux disease: a study of 41 patients. Mod Pathol 2006;19: Mueller S, Neureiter D, Aigner T, et al. Comparison of histological parameters for the diagnosis of eosinophilic oesophagitis versus gastro-oesophageal reflux disease on oesophageal biopsy material. Histopathology 2008;53: Pentiuk S, Putnam PE, Collins MH, et al. Dissociation between symptoms and histological severity in pediatric eosinophilic esophagitis. J Pediatr Gastroenterol Nutr 2009;48: Spechler SJ, Genta RM, Souza RF. Thoughts on the complex relationship between gastroesophageal reflux disease and eosinophilic esophagitis. Am J Gastroenterol 2007;102: Rothenberg ME, Mishra A, Collins MH, et al. Pathogenesis and clinical features of eosinophilic esophagitis. J Allergy Clin Immunol 2001;108: Kirsch R, Bokhary R, Marcon MA, et al. Activated mucosal mast cells differentiate eosinophilic (allergic) esophagitis from gastroesophageal reflux disease. J Pediatr Gastroenterol Nutr 2007; 44: Ramakrishnan R, Chong H. Eosinophilic oesophagitis in adults. Histopathology 2008;52: Cortes JR, Rivas MD, Molina-Infante J, et al. Omeprazole inhibits IL-4 and IL-13 signaling signal transducer and activator of transcription 6 activation and reduces lung inflammation in murine asthma. J Allergy Clin Immunol 2009;124: Zhang X, Cheng E, Huo X, et al. In esophageal squamous epithelial cell lines from patients with eosinophilic esophagitis (EoE), omeprazole blocks the stimulated secretion of ootaxin-3: a potential antiinflammatory effect of omeprazole in EoE that is independent of acid inhibition. Gastroenterology 2010;138;Suppl 1:S Krarup AL, Villadsen GE, Mejlgaard E, et al. Acid hypersensitivity in patients with eosinophilic esophagitis. Scand J Gastroenterol 2010;45: Protheroe C, Woodruff SA, de Petris G, et al. A novel histologic scoring system to evaluate mucosal biopsies from patients with eosinophilic esophagitis. Clin Gastroenterol Hepatol 2009;7: Rothenberg ME. Biology and treatment of eosinophilic esophagitis. Gastroenterology 2009;137: Gonsalves N, Policarpio-Nicolas M, Zhang Q, et al. Histopathologic variability and endoscopic correlates in adults with eosinophilic esophagitis. Gastrointest Endosc 2006;64: Reprint requests Address requests for reprints to: Javier Molina-Infante, MD, Department of Gastroenterology, Hospital San Pedro de Alcantara, C/ Pablo Naranjo s/n Caceres, Caceres, Spain. xavi_molina@ hotmail.com; fax: Conflicts of interest The authors disclose no conflicts. Funding CIBEREHD is funded by the Instituto de Salud Carlos III.