345 / 412
© 2013 by the American College of Gastroenterology
The
American Journal
of
GASTROENTEROLOGY
Fluticasone vs. Esomeprazole for Esophageal Eosinophilia
Proton pump inhibitors (PPIs) are also used to control symptoms
associated with EoE and to treat gastroesophageal reflux disease
(GERD), the main differential diagnosis of EoE, which can present
with similar clinical symptoms and histopathology. According
to the AGA consensus statement published in 2007, administra-
tion of PPIs to presumed EoE patients was part of the diagnostic
evaluation, primarily to exclude GERD as a cause of esophageal
eosinophilia (3). If dense eosinophilia persisted following therapy,
then a diagnosis of EoE is made. However, some patients with a
phenotypic appearance more suggestive of EoE rather than GERD
(i.e. young atopic patient presenting with food impaction with
concentric rings on endoscopy and having elevated eosinophils on
esophageal biopsies) can respond to PPI therapy (4,5). This phe-
nomenon now recognized by the more recent and updated EoE
consensus statement has been termed PPI-responsive esophageal
eosinophilia or PPI-responsive EoE (2,6).
The aimof this studywas to performa randomized controlled trial
to compare the efficacy of fluticasone propionate (FP) to esomepra-
zole (ESO) in patients with esophageal eosinophilia. A secondary
aim of this study was to determine whether the presence of GERD
impacted the response to therapy in each treatment group.
METHODS
Study design and patient population
This is a prospective investigator-blinded randomized study.
Adult patients (age
≥
18 years) seen at Walter Reed Army Medical
Center (WRAMC) with esophageal eosinophilia were enrolled
from April 2008 to October 2010. All patients had at least one
clinical symptom of esophageal dysfunction (dysphagia, food
impaction, heartburn) with
≥
15 eosinophils/hpf (eos/hpf; high
power field) on index endoscopy. Patients who had a history of
secondary hypereosinophilic disorders, severe coagulopathy, or
who were pregnant were excluded from the study. Patients who
were dilated at index endoscopy were not excluded from the study.
Baseline demographic data, history of coexisting atopic diathesis
(seasonal allergies, food allergies, asthma, and eczema), and data
from index endoscopy (concentric rings, longitudinal furrows,
white plaques, mucosal tearing/friability, strictures, Schatzki
rings, erosive esophagitis) were collected. All patients completed
a validated dysphagia questionnaire, known as the 2-week Mayo
Dysphagia Questionnaire (MDQ), following index endoscopy
once eosinophilic infiltration was established on biopsies (7). This
29-item instrument is scored from 0 to 100 based on the pres-
ence and severity of dysphagia and whether patients avoided or
had trouble swallowing different foods (oatmeal, banana, apple,
ground meat, bread, and fibrous meat) (
Supplementary file
).
Informed consent was obtained from all patients. This study was
approved by the WRAMC Institutional Review Board (Work Unit
number: 08-14045) and the study was registered at www.clinical-
trials.gov (NCT00895817).
GERD diagnosis
Upon enrollment into the study, all patients underwent 24-h pH
with impedance studies. Locationof the lower esophageal sphincter
was determined by esophageal manometry utilizing a Solar Sta-
tionary GI motility system (Medical Measurement Systems USA,
Dover, NH) and an electrically powered water perfusion pump
(Mui Scientific, Ontario, CA). A 24-h pH/impedance catheter
was then placed 5 cm above the proximal location of the lower
esophageal sphincter. The catheter was connected to a ZepHr-
reflux recording system (Sandhill Scientific, Highlands Ranch,
CO) to capture pH/impedance, as well as symptom data. Subjects
returned to our clinic the following day for analysis of the study.
Data was analyzed with Bioview Analysis software (Sandhill Sci-
entific). GERD was defined by the validated Johnson-DeMeester
score (8,9). This scoring method takes into account six param-
eters, which include: total % time pH below 4, % time pH below
4 in the upright position, % time pH below 4 in the supine posi-
tion, the total number of reflux episodes within a 24-h period,
the number of reflux episodes longer than 5min, and the longest
reflux episodes in minutes. A composite score is then calculated
with a score of greater than 22 being indicative of GERD. The pH
drops without accompanying reflux events on impedance and
reflux events during meals were excluded from analysis.
Randomization and drug administration
A computer-generated list of random numbers was used to sepa-
rate patients into two equal treatment groups (esomeprazole and
fluticasone proprionate). Concealed allocation using a sealed
opaque envelope containing data on the sequence of randomi-
zation was maintained by a research pharmacist. Following data
from the 24-h pH study, patients were stratified into GERD-
negative or GERD-positive groups. Within each group, sub-
jects were randomized to receive either 40mg of ESO once daily
or 440mcg of FP twice daily. Patients randomized to ESO were
instructed to take the medicine 30–60min before their first meal.
Patients randomized to FP were educated by the research pharma-
cist on correct delivery of the medication using an inhaler with-
out the use of a spacer and instructed not to drink or eat 30min
following administration. The research pharmacist observed the
patients priming the metered dose inhaler and administering at
least one puff to ensure correct delivery. Adherence was assessed
in the ESO arm by counting the number of pills at the end of the
treatment study. For the FP arm, the number of puffs was counted
using a specially designed metered dose inhaler, which recorded
the number of puffs administered. Patients were considered
adherent to treatment if
≥
80% of the medication was taken during
the study period.
Follow-up
Following 8 weeks of treatment, patients underwent repeat
upper endoscopy with esophageal biopsies. A total of eight sam-
ples using standard biopsy forceps (Boston Scientific, Natick,
MA) were taken from all patients, four from the proximal
esophagus, ~15 cm from the gastroesophageal junction, and four
from the distal esophagus, ~3 cm above the gastroesophageal
junction. All endoscopies were performed with Olympus P160
or 180 endoscopes (Olympus, Tokyo, Japan). Endoscopic data
was collected including concentric rings, longitudinal furrows,
120