Journal of Gastroenterology and Hepatology
556–562 DOI: 10.1111/j.1440-1746.2005.04003.x
Blackwell Science, LtdOxford, UKJGHJournal of Gastroenterology and Hepatology0815-93192005 Blackwell Publishing Asia Pty Ltdunknown 200521unknown556562Original Article
Gastroesophageal flap valve and GERDB-R Lin et al.
Correspondence: Dr Teh-Hong Wang, Department of Internal Medicine, National Taiwan University Hospital, no. 7, Chung-
Shan South Road, Taipei, 10016, Taiwan. Email: email@example.com
Accepted for publication 6 March 2005.
Abnormal gastroesophageal flap valve is highly associated with
gastroesophageal reflux disease among subjects undergoing routine
endoscopy in Taiwan
WAN-RU LIAO** AND TEH-HONG WANG
Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University, College
of Medicine, Taipei,
Biomedical Science Team, Chang-Gung Institute of Technology, Taoyuan,
Gastroenterology, Department of Internal Medicine, Cathay General Hospital, Taipei, Departments of
Nursing, National Taiwan University Hospital and National Taiwan University, College of Medicine,
Division of Endoscopy, Department of Integrated Diagnostics & Therapeutics,
Division of Gastroenterology,
in Caucasians, but its role in other populations is unclear. This study evaluated the significance of endo-
scopic grading of the GEFV in Taiwanese subjects.
: Five hundred and six consecutive patients undergoing routine check-ups at the National Tai-
wan University Hospital were enrolled. Symptoms of upper gastrointestinal disease and endoscopic
severity of esophageal mucosal injury were correlated to GEFV grades according to the Hill
: The frequency of abnormal valves (Hill grades III or IV) was 27.3%. Of these, 42.7% had ero-
sive esophagitis (EE). The majority of patients with EE were classified as Los Angeles grades A and B
(79.7 and 16.9%, respectively). The prevalence of EE, hiatal hernia and, to a lesser degree, non-erosive
reflux disease, increased with altered GEFV. Patients with abnormal valves were younger and more likely
to be male, overweight, and to have atypical and extraesophageal symptoms.
: Taiwanese patients with abnormal GEFVs share similar characteristics and risk factors
with the patients who have EE. Endoscopic grading of the GEFV is highly associated with GERD, and
in particular EE, in subjects undergoing routine endoscopy.
© 2005 Blackwell Publishing Asia Pty Ltd
: Gastroesophageal flap valve (GEFV) grade predicts severe gastroesophageal reflux disease
: endoscopy, erosive esophagitis, gastroesophageal flap valve, gastroesophageal reflux disease.
Evidence of a flap valve at the gastroesophageal junction
was first noted in cadavers.
strated that a gradient is present across the gastroesoph-
ageal valve in the absence of hiatal hernia. Loss of this
flap valve is associated with the deterioration of lower
esophageal sphincter length and pressure, increasing
the likelihood of a mechanically defective sphincter.
Moreover, the gradient can be increased by surgically
accentuating the valve.
The clinical importance of the
. later demon-
gastroesophageal flap valve (GEFV) is reflected in the
fact that esophageal acid exposure increases proportion-
ately to the degree of valve dysfunction.
speculated that reinforcement of the GEFV is achiev-
able by endoluminal gastroplication using endoscopic
suturing as therapy for gastroesophageal reflux disease;
this endotherapy reinforced valve is comparable to a
grade I valve achieved by Nissen fundoplication.
grade has been found to predict gastroesophageal reflux
better than measured lower esophageal sphincter pres-
An endoscopic grading system to assess the
It has been
Gastroesophageal flap valve and GERD
GEFV is a simple and reproducible predictor of gas-
troesophageal reflux disease (GERD).
Few published studies have addressed the relation-
ship of esophagitis to GEFV grade.
patients have been symptomatic or had severe
esophagitis. The relationship of GEFV grade with
GERD is not well examined in non-Caucasian popula-
tions. We hypothesized that GEFV grade is associated
with the severity of erosive esophagitis (EE) and con-
ducted a prospective study in subjects undergoing rou-
In all but one
PATIENTS AND METHODS
From July to November 2002, 506 consecutive patients
visiting the National Taiwan University Hospital for a
health check-up received a panendoscopy and com-
pleted a self-administered questionnaire regarding
One day prior to the panendoscopy, patients filled out a
self-administered questionnaire regarding the presence
and frequency of typical symptoms of GERD (heart-
burn and acid regurgitation), atypical symptoms, and
symptoms (heartburn and/or acid regurgitation) in the
current month was the basis for classification into three
groups: (i) one or more times per week; (ii) less than
once per week, but more than once per month; or (iii)
previous symptoms, but none during the current
month. The current symptoms of the remaining patients
were the basis for classification into four other groups:
(iv) atypical symptoms, including epigastric burning,
chest pain or discomfort, food regurgitation, dysphagia
and odynophagia; (v) extraesophageal symptoms,
including globus sensation over throat and neck, sore
throat, hoarseness, chronic cough and asthma; (vi)
simultaneous atypical and extraesophageal symptoms;
and (vii) lack of any symptoms in groups IV through VI.
The frequency of typical
Panendoscopic examination and
classification of esophageal mucosal injury
Flexible video endoscopy was performed in each patient
under premedication with intramuscular injection of
20 mg of hyoscine butylbromide and topical anesthesia
of the pharynx with xylocaine gel. The lower esophagus,
squamocolumnar junction, gastroesophageal junction,
and diaphragmatic hiatus were thoroughly inspected
under adequate endoscopic inflation to maintain opti-
mal visualization without overinflation or masking due
to tight tonic contraction of the lower esophageal
The presence and degree of esophageal
mucosal injury were graded according to the Los Ange-
Patients with findings of non-specific
changes (e.g. friability, edema, granularity, congestion,
erythema or hyperemia) were excluded.
an isolated salmon-colored mucosal island in their
lower esophagus and no squamocolumnar junction
involvement were also excluded.
Patients were classified into four groups on the basis
of upper endoscopic findings: erosive esophagitis (EE);
minute esophagitis (ME) signifying minimal change at
the gastroesophageal junction, but absence of typical
symptoms such as dyspepsia; non-erosive reflux disease
(NERD), signifying normal endoscopy but typical
symptoms (symptomatic GERD);
(NE), signifying normal endoscopy and absent symp-
toms. GERD was defined as EE plus NERD. Hiatal
hernia (HH) was defined as a circular extension of the
gastric mucosa above the diaphragmatic hiatus greater
than 2 cm in axial length.
Panendoscopy was performed by an experienced
endoscopist, and photographs were reviewed retrospec-
tively by a second endoscopist.
or no esophagitis
Classification of the gastroesophageal flap
The GEFV was inspected with a retroflexed endoscope
and graded I to IV according to the Hill classification as
grade I, prominent fold of tissue along the
lesser curvature closely apposed to the endoscope;
grade II, fold present, but slightly less well defined than
in grade I and it opens rarely with respiration and closes
promptly; (i) an incomplete fold; (ii) a non-contact fold
remote to the endoscope; grade III, fold not prominent;
the endoscope was not gripped tightly by the tissues and
often failed to close around the endoscope; grade IV,
fold absent with the lumen of the esophagus gaping
open, allowing squamous epithelium to be viewed from
below. Grades I and II were classified as normal valves
and grades III and IV as abnormal valves (Figs 1–4).
The prevalence rates of EE and hiatal hernia and the
frequency of abnormal GEFV were calculated. Age,
male to female ratio, body mass index (BMI), presence
of hiatal hernia, and symptom category were compared
between groups with normal and abnormal grades of
GEFV, using unpaired
-tests. The chi-squared test with
2 table and multiple comparative test were per-
formed to evaluate normal and abnormal grades of
GEFV among the four groups with esophageal mucosal
injury or symptoms. A
be statistically significant.
0.05 was considered to
The prevalence rates of GERD and EE were 41.3%
209) and 24.7% (
125), respectively. As shown
in Table 1, more than one-quarter of patients had
abnormal valves. Approximately 5% of patients were
classified as having grade IV valves. Patients with abnor-
mal valves were significantly younger than those with
normal valves. Male patients had a significantly higher
prevalence of abnormal valves than female patients, and
the prevalence correlated linearly to increasing valve
grade. Similarly, BMI correlated linearly with increas-
ing grade of GEFV.
Patient distribution into the above four classifications
were: EE (
125; 24.7%), ME (
84; 16.6%), and NE (
greater proportion of patients with EE had abnormal
valves than did those with NE (47.2
0.05). The proportions of patients with EE accord-
ing to Los Angeles classification A to D were 72.8%
91), 24.8% (
31), 1.6% (
1), respectively. The ratio of abnormal to normal
valves increased in a linear fashion in association with
NE, NERD, ME and EE, with the highest prevalence of
abnormal valves in the patients with EE.
Table 2 shows that the prevalence of EE increased
with increasing valve grade; similarly, a lesser degree
was in that of NERD. The prevalence of EE was mark-
edly higher in patients with abnormal valves than in
43; 8.5%), NERD
254; 50.2%). A much
2) and 0.8%
the lesser curvature that was closely apposed to the endoscope.
Grade I signified a prominent fold of tissue along
slightly less well defined than in Grade I and it opens rarely
with respiration and closes promptly; (a) an incomplete fold;
(b) a non-contact fold remote to the endoscope.
Grade II signified that a fold was present, but
Demographic data and gastroesophageal flap valve grade of study subjects
Gastroesophageal flap valve
I IIIII IVI
Male : Female
(%) 175 (34.6)
0.05. BMI, body mass index.
Gastroesophageal flap valve and GERD
those with normal valves (42.7
contrast, the prevalence of NE was inversely correlated
to the GEFV grade. NE was present in 58.2% of
patients with normal valves (66.9% with grade I), while
only 17.9% of patients with normal valves had EE. Of
the patients with abnormal valves, 42.7% had EE and
only 29.0% had NE. The ratio of the prevalence of
abnormal to normal grade GEFV was highest in EE,
followed by ME, NERD and NE. A chi-squared test
with a 4
2 table demonstrated that there were differ-
ent GEFV grades among the four groups classified by
esophageal mucosal injury and symptoms (
Furthermore, a multiple comparative test showed there
was a significantly higher proportion of abnormal
GEFVs in subjects with EE than in those with NERD
and NE (
0.0001); in subjects with ME
as compared with those with NE (
0.003); and in
subjects with NERD than in those with NE
Table 3 shows that most of EE in the patients with an
abnormal valve were in low-grade esophagitis.
The prevalence of hiatal hernia was 7.7% (
shown in Table 4, the prevalence significantly increased
with increasing alteration of the GEFV. Up to 80% of
cases of hiatal hernia occurred in patients with abnor-
As shown in Table 5, patients with abnormal valves
had a significantly higher incidence of current typical
symptoms of GERD at a frequency of once weekly or
greater. In addition, among those lacking typical symp-
toms, patients with abnormal valves had a higher prev-
alence of atypical or extraesophageal symptoms (45.1
0.05; the data were calculated in these study
populations but are not shown in Tables 1–5); and a
the endoscope was not gripped tightly by the tissues and often
failed to close around the endoscope.
Grade III signified that a fold was not prominent;
the lumen of the esophagus gaped open, allowing the squa-
mous epithelium to be viewed from below.
Grade IV signified that there was no fold and that
Correlation of the prevalence of endoscopic esophageal mucosal injury with gastroesophageal flap valve grade
Gastroesophageal flap valve grade
RI IIIIIIV I + IIIII + IV
Classification of esophageal mucosal injury
NERD 27 (15.4)
R, ratio of abnormal to normal grades of GEFV = case number of grade III + IV/I + II. The chi-squared test with a 4 × 2 table
and multiple comparative test showed there was a significantly higher proportion of abnormal GEFVs in subjects with EE, ME
and NERD than in subjects with NE. EE, erosive esophagitis; ME, minute esophagitis; NE, no esophagitis; NERD, non-erosive
B-R Lin et al.
significantly higher prevalence of current simultaneous
atypical and extraesophageal symptoms (17.0 vs 5.4%,
P < 0.05; the data were calculated in these study popu-
lations but are not shown in Tables 1–5), in comparison
to those with normal valves. Patients with normal valves
had a slightly higher incidence of being free of any
The present study demonstrates a close association
between increasing deformation of the GEFV and
endoscopically demonstrated GERD. In contrast to
previous studies of patients with symptomatic or severe
esophagitis in a Caucasian population,2–4 we sought to
survey the prevalence of abnormal valves and the dis-
tribution of various grades of GEFV in relation to the
degree of esophageal injury by simultaneous endoscopy
and symptom analysis in Taiwanese subjects presenting
for routine health check-ups.
Fujiwara et al. previously described the association of
GEFV grade and GERD in a Japanese population.8 Our
findings clarify the spectrum of GERD in Taiwan and
confirm the concept that deformation of the GEFV is
closely associated with GERD in Asians, who are tradi-
tionally thought to be less predisposed to GERD than
Caucasians.15 This implies that an altered GEFV is a
common pathogenetic factor in GERD, unrelated to
Correlation of grades of erosive esophagitis and gastroesophageal flap valves
Gastroesophageal flap valve grade
I II IIIIV I + IIIII + IV
Grade of EE†
†According to the Los Angeles classification criteria. Data are expressed as n (%).
Prevalence of hiatal hernia in subjects with differing grades of gastroesophageal flap valve
Gastroesophageal flap valve grade
I IIIII IVI + IIIII + IV
*P < 0.05. Data are expressed as n (%). HH, hiatal hernia, defined as the axial length between the squamocolumnar junction
and the gastroesophageal junction >2 cm.
Distribution of symptom groups in patients with differing grades of gastroesophageal flap valve
Gastroesophageal flap valve grade
I II III IV I + IIIII + IV
a + b + c
d + e + f
175 193 113 25 368138
*P < 0.05. Data are expressed as n (%).
Gastroesophageal flap valve and GERD
In the present study, the prevalence of EE, and to a
lesser degree NERD, increased with increasing alter-
ation of the GEFV, which is shown by an increasing
abnormal to normal valve ratio in NE, NERD, ME and
EE. Approximately one-quarter of ME patients in each
group had past typical, current atypical, and extrae-
sophageal symptoms, respectively, implying that abnor-
mal valves are a more sensitive index of esophageal
mucosal injury than typical symptoms are. Previous
reports have found that abnormal esophageal acid expo-
sure is a better indicator of endoscopic esophagitis than
symptoms are.3,16 Together with our findings, it is rea-
sonable to believe that the pathology in ME patients is
most likely related to acid reflux in a mildest form of
EE. Accordingly, it is appropriate to delineate ME as a
group separate from NERD.
Explanations have been proposed towards why
NERD patients have a lower ratio of abnormal valves
than patients with EE or ME, but have a ratio compa-
rable to that of NE patients. Patients with NERD may
have transient lower esophageal sphincter relaxation;17
weakly acidic or non-acidic refluxates;18,19 a hypersen-
sitive esophagus;20 or other more complex mecha-
nisms.21 In contrast, patients with EE are more likely
to have a pre-existing structurally defective gastro-
esophageal junction with incompetent lower esoph-
ageal sphincter.22,23 In the present study, NERD
patients had a prevalence of atypical (55.4%) and
extraesophageal symptoms (56.6%), which was much
higher than that of patients with NE; in addition, only
one-fifth of the patients with typical symptoms but
negative endoscopic findings had acid exposure (Lin
BR and Liao CC, unpublished observations, 2004).
Endoscopic esophagitis has been found to be a better
indicator of GERD than typical symptoms in Taiwan-
ese patients.24 Thus, in accordance with the study
conducted by Contractor et al.,4 the present study con-
firmed that GERD patients with endoscopic esophagi-
tis tended to have pre-existing abnormal valves, whose
incidence is correlated with severity of esophageal
The use of the GERD symptom questionnaire may
have resulted in potential bias, leading to under- or
overestimation of the prevalence of NERD, because a
wide spectrum of GERD symptoms may depend on dif-
ferent thresholds of visceral sensitivity,16 and patients
may also have underreported symptoms in the self-
It has been shown that EE can accurately predict the
presence of acid reflux-related GERD, with modest sen-
sitivity.3 Öberg et al. demonstrated that more than half
of GERD patients have normal valves, indicating that
more than half of GERD patients might actually have
NERD.3 In the present study, 40% of patients with
GERD had NERD, suggesting that abnormal valves
would not be a sensitive predictor of NERD among all
GERD patients. In the present study, 29% of those with
abnormal valves had NE and, conversely, 15.7% of NE
subjects had abnormal valves. This indicates that some
subpopulations of NE subjects have no GERD, but
could be potentially predisposed to NERD or EE
because of abnormal valves. This hypothesis deserves
In the present study, a grade IV valve only predicted
52% of cases of EE, which is lower than that predicted
by esophageal acid exposure (75%).3 Endoscopic grad-
ing of the GEFV seems to provide useful information
about reflux status, but it is less useful as an indicator of
the presence of esophageal mucosal injury.3 Ismail et al.
hypothesized that the geometry of the GEFV is not the
sole factor influencing abnormal acid exposure, but that
the loss of the GEFV combined with other factors such
as a decrease in LES length and pressure and an
increasing prevalence of mechanically defective sphinc-
ters, might lead to a higher likelihood of gastroesoph-
ageal reflux.26 This might be an explanation for why
abnormal valves had only a modest sensitivity for low-
grade EE in our patients.
In accordance with the study by Contractor et al., we
found that approximately 13% of those with grade I
valves had EE.4 This contrasts with the belief that
esophagitis does not occur with a true grade I valve.2,3
Furthermore, 29% of patients with abnormal valves did
not have any esophageal mucosal lesions or symptoms
comparable to the findings that 29 and 54% of those
with abnormal valves did not have reflux esophagitis or
symptoms of GERD, respectively, in other series,3,4 but
these findings were in contrast to the near-universal
association of abnormal valves with esophagitis in Hill’s
group.2 This discrepancy could be explained by two
possibilities. First, false-positive patients with minimally
affected squamocolumnar junction might be classified
as grade A esophagitis. This confounder was avoided in
the present study by using well-accepted diagnostic and
classification systems.12,13,27,28 Abnormal esophageal
acid exposure has been documented in 10.6% of
patients with grade I valves.4 Second, the GEFV may
disappear or appear during respiration or postural
change, resulting in false negatives.29 We inspected the
gastroesophageal junction by retroflexed endoscope and
recorded GEFV grade only when the esophagus was
steady without contraction or shortening. Therefore, it
is conceivable that EE can develop in patients with nor-
mal valves. Compatible with the findings by Oberg
et al., we demonstrated that the geometry of the valve
was a less strong predictor of NERD than EE was.3
More well-designed studies are needed to elucidate
Patients with an abnormal GEFV share similar charac-
teristics and risk factors with patients who have EE. The
endoscopic grading of the GEFV is highly associated
with GERD, and in particular with EE, in subjects
undergoing routine endoscopy in Taiwan.
This study was supported by grants from the National
Taiwan University Hospital (NTUH.91S032) and
National Science Council (NSC90-2314-B-002-217
and NSC91-2315-B-002-008), Taiwan.
562 Download full-text
B-R Lin et al.
1 Thor KBA, Hill LD, Mercer CD. Reappraisal of the
flap valve: In vitro and in vivo mechanisms in the gas-
troesophageal junction. Acta Chir. Scand. 1987; 153:
2 Hill LD, Kozarek RA, Kraemer SJ et al. The gastroesoph-
ageal flap valve: In vitro and in vivo observations. Gas-
trointest. Endosc. 1996; 44: 541–7.
3 Öberg S, Peters JH, DeMeester TR et al. Endoscopic
grading of the gastroesophageal valve in patients with
symptoms of gastroesophageal reflux disease (GERD).
Surg. Endosc. 1999; 13: 1184–8.
4 Contractor QQ, Akhtar SS, Contractor TQ. Endoscopic
esophagitis and gastroesophageal flap valve. J. Clin. Gas-
troenterol. 1999; 28: 233–7.
5 Mahmood Z, McMahon BP, Arfin Q et al. Endocinch
therapy for gastro-oesophageal reflux disease: A one year
prospective follow up. Gut 2003; 52: 34–9.
6 Mason RJ, DeMeester TR, Lund RJ et al. Nissen fun-
doplication prevents shortening of the sphincter during
gastric distension. Arch. Surg. 1997; 132: 719–24.
7 Jobe BA, Kahrilas PJ, Vernon AH et al. Endoscopic
appraisal of the gastroesophageal valve after antireflux sur-
gery. Am. J. Gastroenterol. 2004; 99: 233–43.
8 Fujiwara Y, Higuchi K, Shiba M et al. Association
between gastroesophageal flap valve, reflux esophagitis,
Barrett’s epithelium, and atrophic gastritis assessed by
endoscopy in Japanese patients. J. Gastroenterol. 2003; 38:
9 Richter JE. Typical and atypical presentations of gastroe-
sophageal reflux disease – the role of esophageal testing in
the diagnosis and management. Gastroenterol. Clin. North
Am. 1996; 25: 75–102.
10 Liebermann-Meffert D, Allgöwer M, Schmid P, Math D,
Blum AL. Muscular equivalent of the lower esophageal
sphincter. Gastroenterology 1978; 76: 31–8.
11 Boyce HW. Endoscopic definitions of esophagogastric
junction regional anatomy. Gastrointest. Endosc. 2000; 51:
12 Armstrong D, Bennett JR, Blum AL et al. The endoscopic
assessment of esophagitis: A progress report on observer
agreement. Gastroenterology 1996; 111: 85–92.
13 Boyce HW. Endoscopic diagnosis and classification or
reflux esophagitis: Are we there yet? Gastroenterology 2002;
14 Dent J, Brun J, Fendrick AM et al. On behalf of the Gen-
val Workshop Group. An evidence-based appraisal of
reflux disease management: The Genval Workshop
Report. Gut 1999; 44 (Suppl. 2): S1–16.
15 Kang JY, Tay HH, Yap I, Guan R, Lim KP, Math MV.
Low frequency of endoscopic esophagitis in Asian sub-
jects. J. Clin. Gastroenterol. 1993; 16: 70–3.
16 Johnson DA, Winters C, Spurling TJ, Chobanian SJ,
Cattau Jr EL. Esophageal acid sensitivity in Barrett’s
esophagus. J. Clin. Gastroenterol. 1987; 9: 23–7.
17 Mittal RK, Holloway RH, Penagini R, Blackshaw LA,
Dent J. Transient lower esophageal sphincter relaxation.
Gastroenterology 1995; 109: 601–10.
18 Sifrim D, Holloway R, Silny J et al. Acid, non-acid, and
gas reflux in patients with gastroesophageal reflux disease
during ambulatory 24-hour pH-impedance recordings.
Gastroenterology 2001; 120: 1588–98.
19 Shay SS, Johnson LF, Richter JE. Acid rereflux: A review,
emphasizing detection by impedance, manometry, and
scintigraphy, and the impact on acid clearing pathophys-
iology as well as interpreting the pH record. Dig. Dis. Sci.
2003; 48: 1–9.
20 Achem. SR. Endoscopy-negative gastroesophageal reflux
disease. The hypersensitive esophagus. Gastroenterol. Clin.
North Am. 1999; 28: 893–904.
21 Fass R, Ofman JJ. Gastroesophageal reflux disease –
should we adopt a new conceptual framework? Am. J.
Gastroenterol. 2002; 97: 1901–9.
22 van Herwaarden MA, Samsom M, Smout AJ. Excess gas-
troesophageal reflux in patients with hiatus hernia is
caused by mechanisms other than transient LES relax-
ations. Gastroenterology 2000; 119: 1439–46.
23 Pandolfino JE, Shi G, Trueworthy B, Kahrilas PJ. Esoph-
agogastric junction opening during relaxation distin-
guishes nonhernia reflux patients, hernia patients, and
normal subjects. Gastroenterology 2003; 125: 1018–24.
24 Lin BR, Wong JM, Yang JC, Wang JT, Lin JT, Wang TH.
Limited value of typical gastroesophageal reflux disease
symptoms to screen for erosive esophagitis in Taiwanese.
J. Formos. Med. Assoc. 2003; 102: 299–304.
25 Carlsson R, Dent J, Bolling-Sternevald E et al. The use-
fulness of a structured questionnaire in the assessment of
symptomatic gastroesophageal reflux disease. Scand. J.
Gastroenterol. 1998; 33: 1023–9.
26 Ismail T, Bancewicz J, Barlow J. Yield pressure, anatomy
of the cardia and gastro-esophageal reflux. Br. J. Surg.
1995; 82: 943–7.
27 Amstrong D, Monnier P, Nicolet M, Blum AL, Savary
M. Endoscopic assessment of esophagitis. Gullet 1991; 1:
28 Pandolfino JE, Nimish B, Kahrilas PJ. Comparison of
inter-intraobserver consistency for grading of esophagitis
by expert and trainee endoscopists. Gastrointest. Endosc.
2003; 56: 639–43.
29 Hill LD, Kozarek RA. The gastroesophageal flap valve. J.
Clin. Gastroenterol. 1999; 28: 194–7.