Intragastric manometric measurements of patients with hiatal hernia ...
Clinical relevance of laparoscopically diagnosed hiatal hernia
-
Upload
yves-van-nieuwenhove -
Category
Documents
-
view
219 -
download
7
Transcript of Clinical relevance of laparoscopically diagnosed hiatal hernia
Clinical relevance of laparoscopically diagnosed hiatal hernia
Yves Van Nieuwenhove Æ Jeroen Sonck Æ Boudewijn De Waele ÆPeter Potvlieghe Æ Georges Delvaux Æ Patrick Haentjens
Received: 3 February 2008 / Accepted: 25 April 2008 / Published online: 20 May 2008
� Springer Science+Business Media, LLC 2008
Abstract
Background To determine the clinical relevance of a la-
paroscopically diagnosed hiatal hernia.
Methods Consecutive patients undergoing an elective
laparoscopy were prospectively recruited. We assessed
preoperative gastroesophageal reflux symptoms using a
validated score, and documented the presence or absence of
a hiatal hernia during laparoscopy.
Results Of the 95 evaluable patients, 42 (44%) had a
hiatal hernia. The mean age was 49.8 years. Logistic
regression analysis indicated that three features were sig-
nificantly and independently associated with hiatal hernia:
a higher reflux score (odds ratio [OR] 2.44; 95% confi-
dence interval [CI] 1.48-4.05; p \ 0.001), low body mass
index (BMI) (OR 0.83; 95% CI 0.70–0.98; p = 0.029), and
type of surgery (OR 0.34; 95% CI 0.14–0.92; p = 0.033).
The diagnostic accuracy of a reflux score of more than 2
was 81%, with a sensitivity, specificity, positive predictive
value, and negative predictive value of 76%, 85%, 80%,
and 82%, respectively. The likelihood ratio for a positive
result was 5.05.
Conclusion Hiatal hernia is common in this population of
surgical patients undergoing an elective laparoscopy.
Patients with reflux symptoms or a low BMI were more
likely to have a hiatal hernia. With a reflux score of more
than 2, the probability of finding a hiatal hernia during
laparoscopy is 80%.
Keywords Laparoscopy � Hiatal hernia �Gastroesopheageal reflux � Diagnosis
The yearly budget spent on proton pump inhibitors (PPIs)
in Belgium is as high as €153 million and consumes about
0.20% of Belgium’s gross domestic product. The larger
part of this medication is prescribed to patients with
chronic gastroesophageal reflux disease (GERD), which is
estimated to affect about 15–40% of the Western popula-
tion [1–3].
GERD is a multifactorial disease resulting from an
unbalance between lower esophageal sphincter (LES)
pressure, esophageal peristalsis, transient relaxations of the
LES, and gastric acid production. This can lead to the
inflammation of the esophageal mucosa caused by chronic
contact with aggressive gastric refluxates. During the 1960s
GERD was considered synonymous with hiatal hernia.
Thanks to technological innovations in the following years,
elegant instruments were developed to measure the pres-
sure in the different segments of the esophagus and it was
demonstrated that many GERD patients had a hypotensive
LES [4]. Thereafter, transient relaxations of the LES could
be measured and were held responsible for most of the
gastroesophageal reflux events [5]. The importance of the
hiatal hernia was recently rediscovered by the findings of
two independent papers indicating that a hiatal hernia not
only increases the frequency of having transient relaxations
of the LOS [6] but also increases the severity of reflux [7].
The problem with the diagnosis of hiatal hernia is that
there is no real standard diagnostic test to assess its
Y. Van Nieuwenhove (&) � J. Sonck � B. De Waele �G. Delvaux � P. Haentjens
Department of Surgery, Universitair Ziekenhuis Brussels, Vrije
Universiteit Brussel, Laarbeeklaan 101, 1090 Brussels, Belgium
e-mail: [email protected]
P. Potvlieghe
Department of Surgery, Aalsters Stedelijk Ziekenhuis,
Merestraat 80, 9300 Aalst, Belgium
123
Surg Endosc (2009) 23:1093–1098
DOI 10.1007/s00464-008-9970-4
presence. Endoscopy is the most accurate means for the
diagnosis of esophagitis, but it significantly underestimates
the presence of a hiatal hernia when compared with radi-
ology [8]. Barium contrast studies are still considered as
the most accurate method of diagnosing a hiatal hernia, but
it remains virtually impossible to compare this test with in
vivo findings. During laparoscopic surgery, the surgeon is
in the position to assess the esophageal hiatus from within
the peritoneal cavity, and this with an increased intra-
abdominal pressure. In this study we examined the rele-
vance of a laparoscopically detected hiatal hernia in the
light of GERD symptoms in a population of patients
undergoing a laparoscopy for various reasons.
Patients and methods
Between April and July 2003, all consecutive patients who
were to undergo an elective laparoscopic intervention for
various reasons in our department of abdominal surgery
were recruited to participate in this prospective study. The
laparoscopic interventions were: cholecystectomies for
symptomatic gallstones, diagnostic explorations for right
quadrant pain with subsequent appendectomy, sigmoidec-
tomies for recurrent diverticulitis, and adjustable gastric
bandings for morbid obesity. The gastric bandings were
performed in patients between 18 and 65 years old with a
body mass index (BMI) of more than 35 kg/m2. We
excluded patients under 18 years old and pregnant women,
as well as patients with a history of gastroesophageal sur-
gery, gastroesophageal cancer or esophageal achalasia.
From 103 eligible patients, 8 patients were excluded during
the planned intervention because they had adhesions in the
upper abdomen. The study protocol stipulated that patients
were to be excluded when it was impossible to visualize the
hiatal region without adhesiolysis. Eventually, in 95
patients it was possible to evaluate the hiatal region.
Bodyweight and height were measured on admission.
Body mass index (BMI) was calculated by dividing the
patient’s weight (kg) by the square of the patient’s height
(m). All patients were subjected to a GERD Symptom
Score questionnaire by an independent interviewer (J.S.),
not being part of the surgical team. Permission to use the
GERD questionnaire was obtained from M. Anvari who
has validated the questionnaire for the assessment of
GERD symptoms previously [9]. This questionnaire scores
six different GERD symptoms based on their frequency (0,
never to 4, daily) and severity (0, none to 3, severe). The
evaluated symptoms were heartburn, retrosternal/epigastric
pain, regurgitation, dysphagia, epigastric fullness, and
coughing. The total score was the sum of the products of
frequency and severity scores of each individual symptom
and ranged between 0 and 72. If patients were on PPIs or
H2 blockers, the questions were asked referring to the
situation prior to the start of medication.
The study was approved by the local ethical committee
and informed consent was obtained from each patient
before entering the study.
Laparoscopic diagnosis of hiatal hernia
Anaesthesia was obtained following a standardized proto-
col. After induction with intravenous propofol (Diprivan�,
TCI) orotracheal intubation was carried out and inhalation
with desflurane in combination with a 50% oxygen–air
mixture was maintained. Muscle relaxation was induced
with succinylcholine followed by cisatracurium. Intermit-
tent doses of sufentanil or remifentanil were given when
needed in function of pain control.
The surgical team was blinded for the results of the GERD
score questionnaire. During laparoscopy and after insertion
of one 10-mm trocar to allow insertion of a 10-mm 25�angled laparoscope and at least one 5- or 10-mm trocar for an
extra grasper, a 15-mmHg pneumoperitoneum was deployed
and the patient was put in a 30� anti-Trendelenburg position.
Before the eventual insertion of a nasogastric tube, the left
lobe of the liver was gently retracted ventrally. It was then
possible to visualize the hiatal region and the presence or
absence of a hiatal hernia. It was agreed in the study protocol
that patients were to be excluded when it was impossible to
visualize the hiatal region without adhesiolysis.
There is no precise definition of a hiatal hernia, but for
the sake of standardization we chose to adapt the com-
monly used definition of a hiatal hernia as the sliding of the
gastric cardia in a caudo-cranial direction for more than
2 cm beyond the diaphragmatic crura [10]. This measure-
ment was possible by inserting the 2-cm-long legs of a
laparoscopic forceps along the cardia through the esopha-
geal hiatus. If no hernia was observed spontaneously, an
extra pressure was exerted on the abdominal wall during 5–
10 s to reach an intraperitoneal pressure of 30 mmHg.
After visualization of the hiatal region, the rest of the
procedure was carried out as planned.
Statistical analysis
Patient data were prospectively entered in a computerized
database and statistical analyses were performed using
SPSS 11 for MacOSX.
Clinical characteristics and GERD score were summa-
rized descriptively for the whole group, and for patients
stratified according to the presence or absence of a hiatal
hernia.
Differences between groups were assessed by the chi-
square test and Fisher exact test for categorical data and the
Student’s t test for continuous data.
1094 Surg Endosc (2009) 23:1093–1098
123
To determine the independent contribution of each
variable to the presence of a hiatal hernia, logistic regres-
sion analysis was used. The status of the esophageal hiatus
(presence or absence of a hiatal hernia) was used as the
dependent variable, and the clinical characteristics (age,
gender, BMI, type of surgery, use of PPIs, and GERD
score) as the independent variables.
The diagnostic performance of GERD score to dis-
criminate patients with hiatal hernia from patients without
hernia was explored by receiver operating characteristic
(ROC) curve analysis. The area under the ROC curve
(AUC) represents the probability that a test correctly dis-
criminates between patients with and those without hiatal
hernia, where 0.5 is chance discrimination and 1.0 is per-
fect discrimination. The optimal cut-off is chosen from the
graph. The best cut-off is that which maximizes the sum of
the sensitivity and specificity, which is the point nearest to
the top left-hand corner. Formally, this point corresponds
to the maximal value of the Youden index, which is simply
the sum of sensitivity + specificity – 1. Statistical indices
of diagnostic performance (sensitivity, specificity, accu-
racy, positive and negative likelihood ratios, and positive
and negative predictive values) and their corresponding
95% confidence intervals (CI) were computed with the
two-by-two table method using standard formulae.
Results
The overall mean age ± standard deviation (SD) was
49.8 ± 17.6 years and there were 25 men and 70 women
(Table 1). The mean overall GERD score was 2.64 (95%
CI 1.99–3.65).
Univariate analyses indicated that subjects with a lapa-
roscopically diagnosed hiatal hernia had a higher GERD
score (p \ 0.01), a higher BMI (p = 0.015), and were
older (p \ 0.01) than those where a hernia was not seen
(Table 1). Furthermore, the proportion of patients on PPIs
(p \ 0.01) and the types of interventions (p = 0.001) were
significantly different between subjects with and without a
hiatal hernia (Table 1).
Multivariate (logistic) regression analysis indicated that
only three features were significantly and independently
associated with hiatal hernia: a higher reflux score
(p \ 0.001), a low BMI (p = 0.029), and type of surgery
(p = 0.033). Hiatal hernia was not significantly associated
with either gender, age or the use of PPIs (Table 2).
The area under the ROC curve for the GERD score was
0.84 (95% CI 0.75–0.92, p \ 0.001, Fig. 1). The Youden test
resulted in a cutoff value of 2 for dichotomizing the contin-
uous GERD score, allowing the statistical indices of
diagnostic performance to be calculated (Table 3). A GERD
score value of more than 2 identified about 76% of patients
with hiatal hernia (sensitivity) and about 85% of subjects
without hiatal hernia (specificity). Of those subjects with a
GERD score value of more than 2, about 80% had hiatal
hernia (positive predictive value), whereas 82% of patients
with a GERD score value of more than 2 had no hiatal hernia
(negative predictive value). The likelihood ratio for a
Table 1 Clinical and operative characteristics for the whole group and for patients stratified for the presence or absence of a hiatal hernia
Total group (n = 95) No hernia (n = 53) Hernia (n = 42) p value*
Age in years (mean ± SD) 49.8 ± 17 42.9 ± 15.9 58.5 ± 15.8 \0.001
Gender (M/F) 25/70 11/42 14/28 0.203
BMI in kg/m2 (mean ± SD) 27.7 ± 7.2 29.0 ± 8.7 25.9 ± 4.3 0.015
PPI use (yes/no) 12/83 0/53 12/30 \0.001
GERD score (mean ± SD) 2.6 ± 3.8 0.5 ± 1.2 5.3 ± 4.4 \0.001
Laparoscopic interventions
Cholecystectomy, n (%) 46 (48.4) 19 (35.8) 27 (64.3) 0.001
Appendectomy, n (%) 24 (25.3) 16 (30.2) 8 (19.0)
Sigmoidectomy, n (%) 9 (9.5) 3 (5.7) 6 (20.0)
Adjustable gastric banding, n (%) 16 (16.8) 15 (28.3) 1 (2.4)
* Univariate analysis, no hernia group versus hernia group
BMI, body mass index; PPI, proton pump imhibitor; GERD, gastrooesophageal reflux disease score [9]
Table 2 Logistic regression analysis for the association between
clinical variables and the presence of a hiatal hernia
Variable Odds ratio 95% confidence interval p value
Age (decades) 1.22 0.81–1.84 0.34
Female gender 0.95 0.20–4.46 0.95
BMI 0.83 0.70–0.98 0.029
PPI use 52.7 0–7.19 e24 0.88
GERD score 2.44 1.48–4.05 0.001
Surgical procedure 0.34 0.14–0.92 0.033
BMI, body mass index; PPI, proton pump inhibitor; GERD, gas-
trooesophageal reflux disease score [9]
Surg Endosc (2009) 23:1093–1098 1095
123
positive result of 5.05 indicates that a GERD score value of
more than 2 is five times more likely in an individual with
hiatal hernia than in one without hiatal hernia.
Discussion
In this study, we observed a hiatal hernia during laparos-
copy in 44% of our patients, and we found that three
features were significantly and independently associated
with hiatal hernia: a higher reflux score, a low BMI, and the
type of surgery.
Our prevalence of hiatal hernia during planned laparo-
scopic surgery is much higher than in an autopsy study
showing only 8 hiatal hernias in 55 cadavers [11]. During
autopsy, an underestimation is likely to occur as it might
prove difficult to detect a small hiatal hernia when the
intra-abdominal pressure is neutralized after opening the
peritoneum. Our prevalence of hiatal hernia is also higher
than the 239 hiatal hernias detected in 1000 subjects during
endoscopy in a large population-based Swedish study [3].
Several reports, however, provide evidence that endoscopy
does not give a correct estimation of the real incidence of
hiatal hernia [8, 12].
We determined the discriminatory value of the GERD
reflux score for differentiating hiatal hernia based on ROC
curve analysis. An area under the curve of more than 0.80
indicates a diagnostic test to be effective. Our findings also
indicate that, above a defined cutoff value of 2 of the
GERD reflux score, patients have a fivefold increased
likelihood of having a hiatal hernia. In other words, 80% of
patients suffering from mild heartburn more than once a
week are likely to have a hiatal hernia.
Traditionally, the association of a hiatal hernia and
obesity has led to the advice of weight reduction for
patients suffering from GERD. Several recent meta-anal-
yses have demonstrated that there is a significant
association between obesity and GERD symptoms [13, 14].
These meta-analyses make several reservations when it
concerns the association of obesity and the presence of a
hiatal hernia. Only four out of the nine individual studies
included in these meta-analyses specifically examined the
relation between esophagitis, hiatal hernia, and BMI. Two
individual studies were able to show a distinct association
between BMI and the combination of reflux symptoms with
a hiatal hernia [15, 16]. One individual study [17] dem-
onstrated a significant relation between BMI and the
isolated presence of a hiatal hernia, while another failed to
demonstrate such a relation [18]. In all these individual
studies, however, a hiatal hernia was not identified and, in
those that did, the criteria for identifying and measuring
hiatal hernia may not have been used uniformly.
Our own data even suggest a decreased risk of hiatal
hernia in obese subjects. This remarkable finding should be
handled with caution, mainly because the relation between
BMI and hiatal hernia was not the primary objective of our
study. Furthermore, given the small sample size of our
study, we acknowledge the potential for selection bias. For
instance, the presence of GERD in morbidly obese subjects
has long been a relative contraindication for performing a
laparoscopic adjustable gastric banding, which can account
for the a priori but unintended exclusion of patients with a
hiatal hernia in this subset of patients [19].
The type of surgery was a significant and independent
factor for the presence of a hiatal hernia, even though, for
Fig. 1 Receiver operating characteristic (ROC) curve of the GERD
score in patients with and without a laparoscopically diagnosed hiatal
hernia (AUC = 0.84, 95% CI = 0.75–0.92, p \ 0.001). GERD,
gastro-oesophageal reflux disease; ROC, receiver operating charac-
teristic curve; AUC, area under the curve
Table 3 Statistical indices of diagnostic performance for GERD
score (cutoff value = 2 based on ROC) and the presence of a hiatal
hernia
Point estimate 95% confidence interval
Sensitivity 0.76 0.63–0.89
Specificity 0.85 0.75–0.95
Positive likelihood ratio 5.05 2.61–9.77
Negative likelihood ratio 0.28 0.16–0.49
Positive predictive value 0.80 0.68–0.92
Negative predictive value 0.82 0.70–0.92
Diagnostic accuracy 0.81 0.70–0.92
GERD, gastrooesophageal reflux disease [9]; ROC, receiver operating
charcteristic curve
1096 Surg Endosc (2009) 23:1093–1098
123
example, patients undergoing gastric banding were
younger than those undergoing sigmoidectomy or chole-
cystectomy, confirming the usefulness of an approach
based on multivariate (logistic) regression analysis. The
type of surgery has an impact on the presence of a hiatal
hernia presumably because it already characterizes a spe-
cific patient profile. Our findings support the assumption of
‘‘Saint’s triad,’’ according to which cholelithiasic patients
would have an increased prevalence of hiatal hernia [20].
Methodological strengths of our study include its pro-
spective design, the consecutive recruitment of unselected
patients planned to undergo laparoscopic surgery (not
excluding younger adults or very old individuals), a high
proportion of eligible patients being enrolled, and the use
of a well-validated instrument to measure gastroesophageal
reflux symptoms [9].
The problem with the diagnosis of hiatal hernia is that
there is no real standard diagnostic test to assess its pres-
ence. Barium contrast studies are still considered as the
most accurate method of diagnosing a hiatal hernia, but it
remains virtually impossible to compare this test with in
vivo findings. During laparoscopic surgery, on the other
hand, the surgeon is in the position to assess the esophageal
hiatus from within the peritoneal cavity, helped by an
increased intra-abdominal pressure.
The question remains of whether a laparoscopic visu-
alization of a hiatal hernia is the most precise way to
diagnose this. In other words: is the laparoscopic diagnosis
of a hiatal hernia the gold standard? The effects of muscle
relaxants, positive ventilation, and pneumoperitoneum
might be confounding factors. Hernias, however, be they
hiatal, ventral or inguinal, are not a problem of muscles but
are breaches in the abdominal wall. The main difference is
that a ventral or inguinal hernia can be diagnosed by
clinical body surface examination, which cannot be done
for a hiatal hernia. Moreover, an inguinal or ventral hernia
will slide through the abdominal wall during a Valsalva
maneuver pushed by the increased intraperitoneal pressure,
while the sliding of a hiatal hernia from abdomen to thorax
will be neutralized by the equally high intrathoracic pres-
sure. Muscle relaxants probably do not have any effect on
the presence of a hernia. Just as in laparoscopic inguinal
and ventral hernia repair, the hernias remain unchanged
whether the patient is curarized or not. In our series, all
patients were routinely curarized and more than half of
patients did not have a hiatal hernia, leaving no grounds to
assume its effect on the hiatal hernia. Normal intra-
abdominal pressures lie between 0 and 5 mmHg. The
working pressure commonly used in laparoscopy is
15 mmHg, but is still less than the intra-abdominal pres-
sures of more than 60 mmHg which can be measured
during laughing, coughing or sneezing. In this study we
have increased the intra-abdominal pressures to 30 mmHg
to detect the hiatal hernias, but for security reasons we did
not pass beyond this level. Overall, we do acknowledge
that the laparoscopic inspection of the hiatal region is too
invasive to be used as a diagnostic tool in routine day-to-
day clinical practice, but it is certainly a physiological
means of having an in vivo diagnosis of a hiatal hernia and
might thus be used as the gold standard for clinical research
purposes in patients undergoing a laparoscopy.
Despite its prospective design, the current study also has
some limitations related to its observational nature, in par-
ticular the potential for selection bias. The current study
included consecutive patients for which a laparoscopic sur-
gical procedure was needed. In doing so, there was, for
example, an imbalance in the proportion of patients scheduled
for laparoscopic gastric banding. As already indicated,
patients undergoing gastric banding were younger than those
undergoing sigmoidectomy or cholecystectomy, which might
also have confounded the results. We therefore conducted a
logistic regression analysis to determine the independent
contribution of each characteristic to the presence of a hiatal
hernia. Using this approach the type of surgery was confirmed
to act as an independent factor on the presence of a hiatal
hernia, while age was not. More specifically, there was a much
higher proportion of patients undergoing a cholecystectomy in
the group of patients with a hiatal hernia, a feature which
actually confirms the existence of ‘‘Saint’s triad’’ in which
gallstones, hiatal hernia, and diverticular disease share a
common predisposing factor [20]. Finally, laparoscopy may
underestimate the presence of hiatal hernia due to the peri-
esophageal and epiphrenic fat concealing a hernia in obese
subjects. In the current study such an underestimation is very
unlikely, as in our hospital, during placement of a gastric
banding, the hepatogastric ligament is routinely opened,
allowing direct visualization of the base of the crus and the
more retro-esophageal hiatal defect.
In conclusion, a hiatal hernia was observed in almost
half of a population of surgical patients scheduled to
undergo a laparoscopic procedure. Patients with reflux
symptoms or a low BMI were significantly more likely to
present with a hiatal hernia on laparoscopy. When a patient
has a reflux score of more than 2, there is an 80% proba-
bility of finding a hiatal hernia during laparoscopy. To our
knowledge, this is the first study examining the clinical
relevance of a hiatal hernia diagnosed in vivo by laparos-
copy. Furthermore, this study adds basic knowledge to
recent insights that the hiatal hernia is actually of far
greater importance in GERD than we assume [10].
References
1. Locke GR III, Talley NJ, Fett SL, Zinsmeister AR, Melton LJ III
(1997) Prevalence and clinical spectrum of gastroesophageal
Surg Endosc (2009) 23:1093–1098 1097
123
reflux: a population-based study in Olmsted County, Minnesota.
Gastroenterology 112:1448–1456
2. Nebel OT, Fornes MF, Castell DO (1976) Symptomatic gastro-
esophageal reflux: incidence and precipitating factors. Am J Dig
Dis 21:953–956
3. Ronkainen J, Aro P, Storskrubb T, Johansson SE, Lind T, Bol-
ling-Sternevald E, Graffner H, Vieth M, Stolte M, Engstrand L,
Talley NJ, Agreus L (2005) High prevalence of gastroesophageal
reflux symptoms and esophagitis with or without symptoms in the
general adult Swedish population: a Kalixanda study report.
Scand J Gastroenterol 40:275–285
4. Cohen S, Harris LD (1971) Does hiatus hernia affect competence
of the gastroesophageal sphincter? N Engl J Med 284:1053–1056
5. Dodds WJ, Dent J, Hogan WJ, Helm JF, Hauser R, Patel GK,
Egide MS (1982) Mechanisms of gastroesophageal reflux in
patients with reflux esophagitis. N Engl J Med 307:1547–1552
6. Kahrilas PJ, Shi G, Manka M, Joehl RJ (2000) Increased fre-
quency of transient lower esophageal sphincter relaxation
induced by gastric distention in reflux patients with hiatal hernia.
Gastroenterology 118:688–695
7. van Herwaarden MA, Samsom M, Smout AJ (2000) Excess
gastroesophageal reflux in patients with hiatus hernia is caused by
mechanisms other than transient LES relaxations. Gastroenter-
ology 119:1439–1446
8. Panzuto F, Di Giulio E, Capurso G, Baccini F, D’Ambra G, Delle
Fave G, Annibale B (2004) Large hiatal hernia in patients with
iron deficiency anaemia: a prospective study on prevalence and
treatment. Aliment Pharmacol Ther 19:663–670
9. Allen CJ, Parameswaran K, Belda J, Anvari M (2000) Repro-
ducibility, validity, and responsiveness of a disease-specific
symptom questionnaire for gastroesophageal reflux disease. Dis
Esophagus 13:265–270
10. Gordon C, Kang JY, Neild PJ, Maxwell JD (2004) The role of the
hiatus hernia in gastro-oesophageal reflux disease. Aliment
Pharmacol Ther 20:719–732
11. Bombeck CT, Dillard DH, Nyhus LM (1966) Muscular anatomy
of the gastroesophageal junction and role of phrenoesophageal
ligament; autopsy study of sphincter mechanism. Ann Surg
164:643–654
12. Sloan S, Rademaker AW, Kahrilas PJ (1992) Determinants of
gastroesophageal junction incompetence: hiatal hernia, lower
esophageal sphincter, or both? Ann Intern Med 117:977–982
13. Corley DA, Kubo A (2006) Body mass index and gastroesoph-
ageal reflux disease: a systematic review and meta-analysis. Am J
Gastroenterol 101:2619–2628
14. Hampel H, Abraham NS, El-Serag HB (2005) Meta-analysis:
obesity and the risk for gastroesophageal reflux disease and its
complications. Ann Intern Med 143:199–211
15. El-Serag HB, Johanson SF (2002) Risk factors for the severity of
erosive esophagitis in Helicobacter pylori-negative patients with
gastroesophageal reflux disease. Scand J Gastroenterol 37:899–
904
16. Stene-Larsen G, Weberg R, Froyshov Larsen I, Bjortuft O, Hoel
B, Berstad A (1988) Relationship of overweight to hiatus hernia
and reflux oesophagitis. Scand J Gastroenterol 23:427–432
17. Wilson LJ, Ma W, Hirschowitz BI (1999) Association of obesity
with hiatal hernia and esophagitis. Am J Gastroenterol 94:2840–
2844
18. Wu AH, Tseng CC, Bernstein L (2003) Hiatal hernia, reflux
symptoms, body size, and risk of esophageal and gastric adeno-
carcinoma. Cancer 98:940–948
19. Ovrebo KK, Hatlebakk JG, Viste A, Bassoe HH, Svanes K (1998)
Gastroesophageal reflux in morbidly obese patients treated with
gastric banding or vertical banded gastroplasty. Ann Surg
228:51–58
20. Capron JP, Payenneville H, Dumont M, Dupas JL, Lorriaux A
(1978) Evidence for an association between cholelithiasis and
hiatus hernia. Lancet 2:329–331
1098 Surg Endosc (2009) 23:1093–1098
123