ASA PAPER
Gastroesophageal Reflux Disease Outcomes After Vertical Sleeve
Gastrectomy and Gastric Bypass
Daniel Leslie, MD, Y Eric Wise, MD, MA, Adam Sheka, MD, Hisham Abdelwahab, Ryan Irey, MA, MS,y
Ashley Benner, MPH,z and Sayeed Ikramuddin, MD, MHA
Objective: The objective of this study is to assess whether vertical sleeve
gastrectomy (VSG) increases the incidence of gastroesophageal reflux disease
(GERD), esophagitis and Barrett esophagus (BE) relative to patients undergoing Roux-en-Y gastric bypass (RYGB) in patients with and without
preoperative GERD.
Summary of Background Data: Concerns for potentiation of GERD, supported by multiple high-quality retrospective studies, have hindered greater
adoption of the VSG.
Methods: From the OptumLabs Data Warehouse, VSG and RYGB patients
with 2 years enrollment were identified and matched by follow-up time.
GERD [reflux esophagitis, prescription for acid reducing medication (Rx)
and/or diagnosis of BE], upper endoscopy (UE), and re-admissions were
evaluated beyond 90 days.
Results: A total of 8362 patients undergoing VSG were matched 1:1 to patients
undergoing RYGB, on the basis of post-operative follow-up interval. Age, sex,
and follow-up time were similar between the 2 groups (P > 0.05). Among all
patients, postoperative GERD was more frequently observed in VSG patients
relative to RYGB patients (60.2% vs 55.6%, respectively; P < 0.001), whereas
BE was more prevalent in RYGB patients (0.7% vs 1.1%; P ¼ 0.007).
Postoperatively, de novo esophageal reflux symptomatology was more common
in VSG patients (39.3% vs 35.3%; P < 0.001), although there was no difference
in development of the histologic diagnoses reflux esophagitis and BE. Furthermore, postoperative re-admission was higher in the RYGB cohort (38.9% vs
28.9%; P < 0.001).
Conclusions: Compared to RYGB, VSG may not have inferior long-term
GERD outcomes, while also leading to fewer re-hospitalizations. These data
challenge the prevailing opinion that patients with GERD should undergo
RYGB instead of VSG.
Keywords: acid reflux, Barrett’s esophagus, esophagitis, gastroesophageal
reflux disease, GERD, heartburn, Rouxen-Y gastric bypass, RYGB, vertical
sleeve gastrectomy, VSG
(Ann Surg 2021;274:646–653)
S
ince its endorsement as a stand-alone primary bariatric procedure
by the American Society of Metabolic and Bariatric Surgery in
2012,1 the vertical sleeve gastrectomy (VSG) has surpassed the Rouxen-Y gastric bypass (RYGB) as the most frequently performed bariatric operation. The VSG is uniformly performed in a minimally invasive
From the Department of Surgery, University of Minnesota, Minneapolis, MN;
yInstitute for Healthcare Informatics, University of Minnesota, Minneapolis,
MN; and zClinical & Translational Science Institute, University of Minnesota,
Minneapolis, MN.
lesli002@umn.edu.
Reprints will not be available from the author.
The authors report no conflict of interests.
Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions of
this article on the journal’s Web site (www.annalsofsurgery.com).
Copyright ß 2021 Wolters Kluwer Health, Inc. All rights reserved.
ISSN: 0003-4932/21/27404-0646
DOI: 10.1097/SLA.0000000000005061
646 | www.annalsofsurgery.com
fashion, with both laparoscopic and robotic-assist approaches
employed. Advantages of the VSG are avoidance of anatomic rearrangement, relative technical simplicity and noninferior remission of
diabetes mellitus relative to RYGB.2 A putative disadvantage is VSG’s
propensity for non-resolution or potentiation of symptomatic gastroesophageal reflux disease (GERD) and sequelae.
Mechanisms of de novo GERD or worsening of pre-existing
GERD after VSG have been proposed, and these include lack of
gastric compliance, increased intraluminal and lower esophageal
sphincter pressures, narrowing at the hiatus, persistence of hiatal
hernia, and residual fundus.3,4 In contrast, the RYGB is often
regarded as a gold-standard operation for GERD due to a decrease
in parietal cell mass exposed to the esophagus.3 The weight loss
inherent in both operations reduces intra-abdominal pressure, alleviating a major contributing factor to GERD. Nonetheless, the RYGB
is still considered by many the bariatric operation of choice for obese
patients with GERD.
Barrett esophagus (BE), the metaplastic change from squamous to columnar epithelium at the gastroesophageal junction
induced by prolonged acid exposure, is a consequential risk of
sustained and unmanaged GERD.5 Although only routinely screened
for among high-risk groups, its prevalence among adults in the
United States is estimated to be 5.6%.6 Further dysplasia of these
cells can lead to esophageal adenocarcinoma if untreated. A recent
meta-analysis commissioned by the International Federation for the
Surgery of Obesity notes the benefits of RYGB in patients with
preoperative BE, including regression of disease, but does not make a
recommendation against VSG due to the paucity of data with respect
to outcomes of these patients. Moreover, data on de novo BE after
RYGB and VSG is even less well characterized.7 In this study, it is
hypothesized that clinically germane GERD-related outcomes after
VSG and RYGB can be determined and contrasted by analyzing a
comprehensive, real-world insurance claims database.
METHODS
Data Source
This study utilized administrative insurance claims data from
the OptumLabs Data Warehouse (OLDW), which contains de-identified, longitudinal health information on enrollees and patients,
representing a mixture of ages, ethnicities, and geographical regions
across the United States. The data in OLDW include medical claims,
pharmacy claims, laboratory results, and enrollment records for
commercial and Medicare Advantage enrollees.8 Because the data
analyzed here are preexisting and deidentified, this study was
deemed exempt from the University of Minnesota Institutional
Review Board review process.
Patients
This study used a retrospective cohort model to examine the
prevalence of postoperative GERD in patients having had a RYGB or
VSG (Fig. 1). Patients undergoing these procedures were identified
Annals of Surgery Volume 274, Number 4, October 2021
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 274, Number 4, October 2021
GERD Outcomes After Bariatric Surgery
FIGURE 1. CONSORT flow diagram of included and excluded patients undergoing bariatric surgery along with final matched cohort.
ß
2021 Wolters Kluwer Health, Inc. All rights reserved.
www.annalsofsurgery.com | 647
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 274, Number 4, October 2021
Leslie et al
and retained for inclusion in the cohort if they were aged 18 years
or older at time of surgery, had their operation between 2004 and
2018, had at least 1 year of preoperative continuous enrollment,
and had at least 2 years of postoperative continuous enrollment.
International Classification of Diseases, Ninth Revision (ICD-9)
and Tenth Revision (ICD-10) codes, as well as Current Procedural
Terminology (CPT) codes, were used to identify VSG procedures
(43.82, 0DB64Z3, and 43775, respectively) and RYGB procedures
(43644; CPT only). Data were not analyzed beyond 5 years
postoperatively.
Cohort Matching
Patients having had a RYGB procedure were matched 1:1 to
patients having had a VSG procedure. RYGB patients were exact
matched to VSG patients based on the number of days of postoperative continuous enrollment. When exact matching was not possible,
nearest neighbor matching was used.
Preoperative Comorbidities
A panel of known diseases and conditions associated with
obesity were collected from the OLDW claims data, including
diabetes mellitus, hyperlipidemia, hypertension, sleep apnea, and
osteoarthrosis/itis. ICD-9 and ICD-10 codes were used for attribution
of these conditions (Supplementary Table 1, http://links.lww.com/
SLA/D261).
Pre- and Postoperative Procedures
In addition to these comorbidities, data on other procedures
used to evaluate the foregut were collected both preoperatively and
postoperatively, including esophagogastroduodenoscopy, esophagoscopy, pH monitoring study, and manometry. Finally, mucosal diseases diagnosed from these procedures were collected, including
peptic/esophageal ulcers and esophageal cancer.
Acid-Treating Prescriptions
One set of features within the OLDW dataset describes
information surrounding the administration of prescription medication. From these available data, inferences can be drawn about the use
of acid-treating prescriptions before bariatric surgery and beyond the
window of typical postoperative care (91 days post-surgery, up to
1825 days or 5 years). Here, the term ‘‘new’’ is used to describe
instances of a prescription, procedure, or diagnosis that is observed in
the postoperative analysis window (91 to 1825 days post-operation)
that is not otherwise observed before surgery. Prescription information was obtained by searching via specific therapeutic class, generic
name, and/or brand name. Due to variations in abbreviation and
spelling of generic names and brand names found in the data, the
codes for specific therapeutic class were utilized. Specifically, drugs
from the classes of gastroprotective agents (GPA; D4B, D4E),
proton-pump inhibitors (PPI; D4J), and H2 Blockers (H2B; Z2D,
Z2T) were queried. Detailed information on the generic drugs found
in the data from these classes can be found in Supplemental Table 2,
http://links.lww.com/SLA/D261.
GERD
In this study, GERD was taken to represent a composite
diagnosis resulting from any of the following conditions: esophageal
reflux, reflux esophagitis, and BE. When studying the underlying
diagnoses comprising a patient’s GERD status, all present conditions
were counted, but in terms of analytical consideration, only the
patient’s most severe condition (BE > reflux esophagitis > esophageal reflux) was counted. Additionally, the prefix de novo has been
used to indicate the presence of GERD in a patient who otherwise did
not have GERD before bariatric surgery.
648 | www.annalsofsurgery.com
Re-Hospitalization
Another set of features within the OLDW dataset describes
information relating to hospital stays, including admission dates,
lengths of stay, and admission diagnosis codes. From these available
data, we were able to discern patient re-hospitalizations following
their bariatric surgery. Patient re-hospitalization data was collected as
a binary variable (ie, whether a patient had any postoperative rehospitalization), the total number of days a patient spent admitted to a
hospital following their procedure, the number of distinct hospital
stays following their procedure (ie, distinct, nonoverlapping rehospitalization events), and the patient’s average length of stay
(eg, a patient with 3 postoperative rehospitalizations of 1 day, 4
days, and 1 day, respectively, has an average length of stay of 2 days).
Statistical Analysis
Statistical analyses and modeling were performed using
Python version 3.5.2 (Python Core Team, 2018) and R version
3.5.0 (R Core Team, 2018). A detailed listing of software names
and versions can be found in Supplemental Table 3, http://links.lww.com/SLA/D261. Where cell sizes were appropriately large,
cohort characteristics were compared using the Chi-square test for
categorical variables; Fisher exact test was used for comparisons
involving small cell counts of 15. For comparison of continuous
variables, the Mann-Whitney U test was used. A significance level
(a) of 0.05 was used for all analyses, although all P values were
reported in accordance with recommendations from the American
Statistical Association.9 Additionally, the association between bariatric procedure type and postoperative GERD was evaluated using
the Kaplan–Meier method.
RESULTS
Preoperative Patient Characteristics: Within 1 Year
Before Surgery
A total of 54,695 patients having had a RYGB or VSG
procedure were identified. After applying continuous enrollment,
age, and minimum index date requirements, 19,721 patients were
available for matching: 11,359 RYGB patients and 8362 VSG
patients. To control for differences in postoperative follow-up time
between surgery groups, each VSG patient was matched to a RYGB
patient (without replacement) based on follow-up time (U ¼
35,240,898.0, P ¼ 0.37). Thus, the final cohort under study comprised 8362 RYGB patients and 8362 VSG patients (n ¼ 16,724;
Fig. 1). Based on this matched cohort, approximately 79% of each
group were females and 21% males with median age 47 years at the
time of the bariatric procedure. Mean follow-up time was 4.1 years,
ranging from 2 to 5 years in each group. The VSG group was
observed to have lower rates of T2DM (Chi-squared ¼ 111.4,
P < 0.001), hyperlipidemia (Chi-squared ¼ 49.5, P < 0.001),
hypertension (Chi-squared ¼ 52.3, P < 0.001), and osteoarthrosis/osteoarthritis (Chi-squared ¼ 19.23, P < 0.001). Conversely, the
baseline VSG group was observed to have higher rates of sleep apnea
(Chi-squared ¼ 20.1, P < 0.001), EGD (Chi-squared ¼ 131.4, P <
0.001), esophagoscopy (Chi-squared ¼ 26.7, P < 0.001), peptic/
esophageal ulcers (Chi-squared ¼ 54.3, P < 0.001), and acidreducing prescriptions (Chi-squared ¼ 21.1, P < 0.001). Rates of
pH monitoring studies (Chi-squared ¼ 0.3, P ¼ 0.57), manometry
(Chi-squared ¼ 2.1, P 1.0), and esophageal cancer (odds ratio ¼
0.9, P 1.0) were rare and statistically similar between the 2
surgery groups.
Preoperative GERD
Preoperative GERD was common in both surgery groups,
affecting approximately 77% of RYGB patients and 80% of VSG
ß
2021 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 274, Number 4, October 2021
GERD Outcomes After Bariatric Surgery
TABLE 1. Baseline Demographics, Follow-Up, Comorbidities, Procedures, Diagnoses, Prescriptions, and GERD
Summary of Cohort
-365 to -1 days, Pre-Op, n ¼ 16,724
RYGB
N
Sex
Female
Male
Index age
Mean
SD
Median
Follow-up time, days
Mean
SD
Median
Preoperative comorbidities
Diabetes mellitus
Hyperlipidemia
Hypertension
Sleep apnea
Osteoarthrosis/it is
Procedures
Esophagogastroduodenoscopy
Esophagoscopy
pH monitoring study
Manometry
Diagnoses
Peptic or esophageal ulcer
Esophageal cancer
Acid-treating prescriptions
GERD
Barrett esophagus
Reflux esophagitis
Esophageal reflux
VSG
8362
50.0%
8362
50.0%
Ps
6637
1725
79.4%
20.6%
6576
1786
78.6%
21.4%
0.26
46.8
10.8
47.1
46.7
10.8
46.6
0.21
1502.9
541.8
1421
1490.0
521.8
1421
0.37
3424
5011
5860
3433
1514
40.9%
59.9%
70.1%
41.1%
18.1%
2764
4560
5421
3816
1296
33.1%
54.5%
64.8%
45.6%
15.5%
<0.001
<0.001
<0.001
<0.001
<0.001
3717
29
59
99
44.5%
0.3%
0.7%
1.2%
4459
85
52
79
53.3%
1.0%
0.6%
0.9%
<0.001
<0.001
0.57
1.0
2681
<11
1013
6456
70
897
5489
32.1%
<0.1%
12.1%
77.2%
1.1%
13.9%
85.0%
3136
<11
1216
6697
57
810
5830
37.5%
<0.1%
14.5%
80.1%
0.9%
12.1%
87.1%
<0.001
1.0
<0.001
<0.001
0.29
0.64
<0.001
SD indicates standard deviation.
patients (Chi-squared ¼ 20.5, P < 0.001). The vast majority of
preoperative GERD was based on the diagnosis of esophageal reflux;
approximately 85% in the RYGB group and 90% in the VSG group.
Among GERD diagnoses that required a biopsy (reflux esophagitis and
BE), rates between the 2 groups were statistically similar (Chi-squared
¼ 0.6, P ¼ 0.46). These data are summarized in Table 1.
Postoperative Patient Characteristics: 91 Days to
5 Years Following Surgery
The VSG group was observed to have lower postoperative rates
of EGD (Chi-squared ¼ 130.6, P < 0.001), peptic/esophageal ulcer
(Chi-squared ¼ 53.4, P < 0.001), and new acid-reducing prescriptions
(Chi-squared ¼ 24.7, P < 0.001). Conversely, the VSG group was
observed to have higher rates of pH monitoring studies (Chi-squared ¼
7.5, P ¼ 0.006), and manometry (Chi-squared ¼ 4.5, P ¼ 0.03). Rates
of esophagoscopy (Chi-squared 0.0, P 1.0), esophageal cancer
(Chi-squared ¼ 0.5, P ¼ 0.48), and acid-reducing medications lasting
>90 days (Chi-squared ¼ 0.2, P ¼ 0.68) were rare and statistically
similar between the 2 surgery groups.
Re-Hospitalization
Over the assessed follow-up period, rates of re-hospitalization
were greater in RYGB patients relative to VSG patients (Chi-squared
¼ 187.8, P < 0.001). Furthermore, VSG patients had a shorter mean
length of stay than RYGB patients (1.3 days vs 1.5 days, respectively;
P < 0.001).
ß
2021 Wolters Kluwer Health, Inc. All rights reserved.
Postoperative GERD
Postoperative GERD was assessed from a nested perspective:
both with and without consideration for whether participants were
considered to have had GERD preoperatively. Such de novo GERD
implies that a patient did not have GERD preoperatively and that
diagnoses comprising the GERD composite diagnosis occurring in
the first 90 days following surgery were not noted as new, postoperative GERD. Postoperative GERD—regardless of preoperative
GERD status—was observed at a higher rate for patients having
had a VSG procedure (Chi-squared ¼ 36.8, P < 0.001). However, a
more granular inspection of the 3 GERD components revealed that
there was no difference in reflux esophagitis (Chi-squared ¼ 0.29,
P ¼ 0.59), whereas VSG had a greater prevalence of esophageal
reflux (Chi-squared ¼ 45.9, P < 0.001) and RYGB had a greater
prevalence of BE (Chi-squared ¼ 7.2, P ¼ 0.007). Additionally,
among these patients, RYGB recipients were observed to have
higher rates of new acid-reducing prescriptions (Chi-squared ¼
17.2, P < 0.001). Although the rate of de novo esophageal reflux was
higher in the VSG cohort (Chi-squared ¼ 6.3, P ¼ 0.01), there was
no difference in the rates of BE (p 1.0) and reflux esophagitis (Chisquared ¼ 0.02, P ¼ 0.88). These postoperative outcomes for all
patients are summarized in Table 2. As use of acid-reducing
medication can be due to institution-specific postoperative protocols, rates of use after 5 years of follow-up were assessed. VSG
patients (269/2,152 with five year follow-up) were on medication at
rates no different from RYGB patients (309/2345 with 5-year
www.annalsofsurgery.com | 649
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 274, Number 4, October 2021
Leslie et al
TABLE 2. Postoperative Procedures, Diagnoses, Prescriptions, Re-Hospitalizations, and GERD
Summary of Cohort
91 to 1825 Days, Post-Op
n ¼ 16,724
RYGB
N
Procedures
Esophagogastroduodenoscopy
Esophagoscopy
pH monitoring study
Manometry
Diagnoses
Peptic or esophageal ulcer
Gastrojejunal ulcer
Esophageal cancer
New acid-treating prescription(s)
Re-hospitalization
Re-hospitalization; count
Mean
SD
Median
Re-hospitalization; average length of stay
Mean
SD
Median
GERD
New acid-treating prescription(s)
Barrett esophagus
Reflux esophagitis
Esophageal reflux
De novo GERD
Patients with no preoperative GERD
New acid-treating prescription(s)
Barrett esophagus
Reflux esophagitis
Esophageal reflux
VSG
8362
50.0%
8362
50.0%
Ps
2442
31
32
73
29.2%
0.4%
0.4%
0.9%
1798
31
59
102
21.5%
0.4%
0.7%
1.2%
<0.001
1.0
0.006
0.03
1990
525
33
677
3254
23.8%
6.3%
0.4%
8.1%
38.9%
1601
37
40
511
2415
19.1%
0.4%
0.5%
6.1%
28.9%
<0.001
<0.001
0.48
< 0.001
<0.001
0.8
1.9
0
0.6
1.5
0
<0.001
1.5
3.6
0
4650
519
93
542
4015
<0.001
55.6%
6.2%
1.1%
6.5%
48.0%
1.3
6.9
0
5038
397
60
525
4453
60.2%
4.7%
0.7%
6.3%
53.3%
<0.001
<0.001
0.007
0.59
<0.001
1906
113
<11
67
672
22.8%
5.9%
<0.6%
3.5%
35.3%
1665
76
<11
57
655
19.9%
4.6%
<0.7%
3.4%
39.3%
<0.001
0.07
1.0
0.88
0.01
SD indicates standard deviation.
Condition is a subset of the first row in the section only.
follow-up) after 5 years (12.5% vs 13.2%, Chi-squared ¼ 0.46, P ¼
0.50).
DISCUSSION
This study represents a retrospective analysis of a comprehensive
administrative claims database, examining GERD-related sequelae of
bariatric surgery. Patients with obesity who underwent VSG were
matched to those who underwent RYGB. Matching was performed
based on length of follow-up, as opposed to demographics, initial body
mass index or preoperative comorbidities. Furthermore, only patients
with continuous enrollment for at least 1 year before surgery and 2 to
5 years after surgery were included. This was critical to ensure equal
capture of the spectrum of GERD, reflux esophagitis, and BE, which
may be picked up on an initial postoperative endoscopy or in the context
of a surveillance protocol.10 Furthermore, there is a highly regimented
postoperative visit schedule recommendation by the American Society
for Metabolic and Bariatric Surgery, making matching on the basis of
follow-up interval more critical.11 Expectedly, the burden of comorbid
illness was higher in the RYGB group, as these patients may have both
higher initial body-mass indices and severity of diabetes mellitus.
Nonetheless, it was deemed that the matching algorithm on followup only was most critical to ensure adequate comparisons during which
symptomatic GERD, reflux esophagitis, and BE could
ostensibly develop.
650 | www.annalsofsurgery.com
Mechanistically, VSG is expected to have a propensity for
increased post-operative persistent and de novo GERD, and subsequently, need for GERD medications, incidence of reflux esophagitis, and Barrett metaplastic and dysplastic changes.12 This has further
been suggested in more recent smaller studies, which conclude that
GERD symptoms are either improved or not exacerbated by VSG,
although de novo incidence of GERD is greater.13,14 The VSG is
associated with a decreased lower esophageal sphincter pressure and
increased total and recumbent acid exposure times, in comparison to
RYGB.12 Furthermore, a wrap operation is not feasible after a VSG,
as these patients are sometimes converted to RYGB for refractory
reflux or even BE.15,16 Using a comprehensive real world dataset,
this study examined the prevalence of GERD-related outcomes after
both procedures, but also specifically examined them in the cohort
that did not have GERD preoperatively.
The primary findings of this study are twofold. Among all
patients, incidence of postoperative GERD was significantly greater
after VSG relative to RYGB (60.2% vs 55.6% for VSG and RYGB,
respectively; P < 0.001), and there was no difference in reflux
esophagitis (6.3% vs 6.5%; P ¼ 0.59), whereas BE was actually more
prevalent in the RYGB cohort (1.1% vs 0.7%; p ¼ 0.007). Second,
although de novo esophageal reflux symptomatology occurred less
frequently in RYGB patients relative to VSG patients (39.3% vs
35.3%; P ¼ 0.01), there was no difference in the frequency of de novo
ß
2021 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 274, Number 4, October 2021
reflux esophagitis (3.4% vs 3.5%; P ¼ 0.88) and BE (<0.7% vs
<0.6%; P ¼ 1.0).
While conferring little benefit with respect to GERD-related
outcomes, RYGB patients in this dataset did experience more frequent
re-hospitalizations and more re-hospitalized days reflecting the
increased risk of associated complications, at a rate commensurate
with findings from van Olst et al in a recent meta-analysis.17 Postoperative re-hospitalizations were specifically chosen for concurrent
analysis due to facility of accurate data capture, prevalence and
importance of, as reflected by its use as a benchmark of quality.
Thirty-day readmissions are known to be more frequent after RYGB
relative to VSG, according to the MBSAQIP national data registry.18
Findings both from Garg et al, and Gero et al suggest that a very high
percentage of re-hospitalizations in the first post-operative year are
surgically related, typically due to abdominal pain or dietary issues;
reported data beyond 1 year are lacking.19,20 This finding is plausibly
attributable to both the higher burden of comorbidities in RYGB
patients, and the increased risk of digestive and pain syndromes as
well as higher prevalence of short- and long-term complications.17
This study definitively shows the persistence of this trend in the longterm postoperative period; however, assessments of longer-term reasons for readmission after bariatric surgery, beyond 1 year, are lacking.
Preoperative BE was found in 0.8% of all patients who
underwent RYGB, and 0.7% of patients who underwent VSG, data
consistent with the preoperative prevalence reported in a metaanalysis by Qumseya et al in 2020.21 The absolute rates of postoperative BE in this study were low, at 1.1% in the RYGB cohort and
0.7% in the VSG cohort, similar to the rates reported in other large
studies worldwide.22,23 In contrast, when followed to 5 years, other
estimates of BE after VSG vary widely, from between 11% and
19%.24,25 Recently, the International Federation for the Surgery of
Obesity and Metabolic Disorders estimated an incidence of BE after
bariatric surgery. Derived from meta-analyses, the incidence of BE
after VSG was estimated at 4.6% in patients with at least 2 years of
follow-up, comparable to the 4.2% rate estimated in similarly
followed patients irrespective of procedure type.7 Widely disparate
reported rates of BE prevalence can be attributed to lack of routine
endoscopic surveillance protocols, reporting bias and patient selection bias, however, of greatest clinical relevance is the lack of
increased risk of BE development in patients who underwent
VSG relative to RYGB.
This study did not examine progression of BE, and only
conclusions about BE prevalence before and after bariatric surgery
can be derived. However, the natural history of progression of known
BE is of critical importance when deciding the appropriateness of
bariatric procedure. Recently, Mahawar et al reported the first
modified Delphi consensus statement on VSG. Notably, there was
near universal agreement that in patients with BE, VSG was not a
suitable primary surgical weight loss option.26 Data on the natural
history of patients with BE after RYGB are even more scarce.
Goonawardena and Ward recently examined 5 publications reporting
the nature of BE progression after RYGB. Regression rates varied
widely, from 36% to 62%, although it was not evident whether the BE
regression persisted, there was poor compliance with diagnostic and
reporting protocols suggested by the American College of Gastroenterology, and there were only 58 patients examined across the five
studies.27 In our data, there were very few patients who had BE postoperatively than pre-operatively in the VSG cohort, and even fewer
additional patients with BE after RYGB, relative to before. This
study, more than all else, demonstrates the need for a prospective,
long-term study to determine the natural history of both development
and progression of BE and subsequent esophageal adenocarcinoma
with routine and long-term endoscopic surveillance after both VSG
and RYGB, as present studies rely on retrospective data with
ß
2021 Wolters Kluwer Health, Inc. All rights reserved.
GERD Outcomes After Bariatric Surgery
inherently unstandardized methods of follow-up and endoscopy.
Although difficult to assess and detect due to low incidence, longterm development of esophageal cancer from BE does not appear to
be increased after VSG, relative to RYGB.28,29
There were multiple limitations to this study within which the
context of our findings must be considered. Although expansive, this
study is subject to selection and reporting bias inherent in retrospective data collection. Furthermore, as a claims-based dataset, attribution of comorbidities, procedures, and diagnoses is necessarily
dependent on accurate and fully captured coding.30 Systematic
miscoding errors are undetectable using this methodology and
may introduce error. Nonetheless, procedural attribution via CPT
code use is quite common and mandated with studies considering
Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program database, whereas comorbidity attribution via presence of relevant ICD-9 or ICD-10 codes has been used in myriad
well-designed analyses whose outcomes have been verified.31–33 As
a retrospective dataset, inferences about temporal causality cannot be
ascertained. Specifically, inferences on reasons for initiation of antireflux medications and indications for EGDs cannot be determined
with the dataset. Next, as discussed previously, matching was
performed on the basis of follow-up interval, meaning differences
in demographics or comorbidities may have unduly influenced
development of GERD-related outcomes. Specifically, RYGB
patients expectedly had a higher burden of comorbidities including
diabetes mellitus and hypertension, whereas VSG patents were more
frequently on acid-reducing medications preoperatively.
Despite the limitations, this large dataset demonstrates that
although GERD may be slightly more prevalent after VSG than
RYGB, differences in postoperative reflux esophagitis and BE are not
observed, challenging the notion that RYGB is superior to VSG in
patients with pre-existing GERD or even BE.
REFERENCES
1. Committee ACI. Updated position statement on sleeve gastrectomy as a
bariatric procedure. Surg Obes Relat Dis. 2012;8:e21–e26.
2. Wolnerhanssen BK, Peterli R, Hurme S, et al. Laparoscopic Roux-en-Y gastric
bypass versus laparoscopic sleeve gastrectomy: 5-year outcomes of merged
data from two randomized clinical trials (SLEEVEPASS and SM-BOSS). Br J
Surg. 2021;108:49–57.
3. El-Hadi M, Birch DW, Gill RS, et al. The effect of bariatric surgery on
gastroesophageal reflux disease. Can J Surg. 2014;57:139–144.
4. Stenard F, Iannelli A. Laparoscopic sleeve gastrectomy and gastroesophageal
reflux. World J Gastroenterol. 2015;21:10348–10357.
5. Rees JR, Lao-Sirieix P, Wong A, et al. Treatment for Barrett’s oesophagus.
Cochrane Database Syst Rev. 2021;3:CD004060.
6. Maret-Ouda J, Markar SR, Lagergren J. Gastroesophageal reflux disease:
a review. JAMA. 2020;324:2536–2547.
7. Fisher OM, Chan DL, Talbot ML, et al. Barrett’s Oesophagus and Bariatric/
Metabolic Surgery-IFSO 2020 Position Statement. Obes Surg. 2021.
8. OptumLabs. OptumLabs and OptumLabs Data Warehouse (OLDW) Descriptions and Citation. Eden Prairie, MN: n.p., July 2020. PDF. Reproduced with
permission from OptumLabs.
9. Laber EB, Shedden K. Statistical significance and the dichotomization of
evidence: the relevance of the ASA statement on statistical significance and pvalues for statisticians. J Am Stat Assoc. 2017;112:902–904.
10. Andrew B, Alley JB, Aguilar CE, et al. Barrett’s esophagus before and after
Roux-en-Y gastric bypass for severe obesity. Surg Endosc. 2018;32:930–936.
11. Mechanick JI, Apovian C, Brethauer S, et al. Clinical Practice Guidelines for
the Perioperative Nutrition, Metabolic, and Nonsurgical Support of Patients
Undergoing Bariatric Procedures - 2019 Update: Cosponsored by American
Association of Clinical Endocrinologists/American College of Endocrinology,
the Obesity Society, American Society for Metabolic & Bariatric Surgery,
Obesity Medicine Association, and American Society of Anesthesiologists Executive Summary. Endocr Pract. 2019;25:1346–1359.
12. Jaruvongvanich V, Matar R, Ravi K, et al. Esophageal pathophysiologic
changes and adenocarcinoma after bariatric surgery: a systematic review
and meta-analysis. Clin Transl Gastroenterol. 2020;11:e00225.
www.annalsofsurgery.com | 651
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 274, Number 4, October 2021
Leslie et al
13. Silveira FC, Poa-Li C, Pergamo M, et al. The effect of laparoscopic
sleeve gastrectomy on gastroesophageal reflux disease. Obes Surg.
2021;31:1139–1146.
14. Balla A, Quaresima S, Palmieri L, et al. Effects of laparoscopic sleeve
gastrectomy on quality of life related to gastroesophageal reflux disease.
J Laparoendosc Adv Surg Tech A. 2019;29:1532–1538.
15. de Tomas J. Conversion to gastric bypass in patients with Barret’s esophagus
after sleeve gastrectomy. Cir Esp. 2021.
16. Parmar CD, Mahawar KK, Boyle M, et al. Conversion of sleeve gastrectomy to
Roux-en-Y gastric bypass is effective for gastro-oesophageal reflux disease
but not for further weight loss. Obes Surg. 2017;27:1651–1658.
17. van Olst N, van Rijswijk AS, Mikdad S, et al. Long-term emergency department visits and readmissions after laparoscopic Roux-en-Y gastric bypass: a
systematic review. Obes Surg. 2021;31:2380–2390.
18. Kapur A, Thodiyil P. Primary laparoscopic sleeve gastrectomy versus gastric
bypass: a propensity-matched comparison of 30-day outcomes. Surg Obes
Relat Dis. 2021.
19. Garg T, Rosas U, Rogan D, et al. Characterizing readmissions after bariatric
surgery. J Gastrointest Surg. 2016;20:1797–1801.
20. Gero D, Raptis DA, Vleeschouwers W, et al. Defining global benchmarks in
bariatric surgery: a retrospective multicenter analysis of minimally invasive
Roux-en-Y gastric bypass and sleeve gastrectomy. Ann Surg. 2019;270:859–
867.
21. Qumseya B, Gendy S, Wallace A, et al. Prevalence of Barrett’s esophagus in
obese patients undergoing pre-bariatric surgery evaluation: a systematic
review and meta-analysis. Endoscopy. 2020;52:537–547.
22. Zacharakis G, Almasoud A, AlZahrani J, et al. Upper gastrointestinal tract
involvement in the management of bariatric patients in the Kingdom of Saudi
Arabia. Ann Gastroenterol. 2021;34:177–182.
23. Al Sabah S, AlWazzan A, AlGhanim K, et al. Does laparoscopic sleeve
gastrectomy lead to Barrett’s esophagus, 5-year esophagogastroduodenoscopy findings: a retrospective cohort study. Ann Med Surg (Lond).
2021;62:446–449.
24. Sebastianelli L, Benois M, Vanbiervliet G, et al. Systematic endoscopy 5 years
after sleeve gastrectomy results in a high rate of Barrett’s esophagus: results of
a multicenter study. Obes Surg. 2019;29:1462–1469.
25. Qumseya BJ, Qumsiyeh Y, Ponniah SA, et al. Barrett’s esophagus after sleeve
gastrectomy: a systematic review and meta-analysis. Gastrointest Endosc.
2021;93:343–352. e2.
26. Mahawar KK, Omar I, Singhal R, et al. The first modified Delphi consensus
statement on sleeve gastrectomy. Surg Endosc. 2021.
27. Goonawardena J, Ward S. Effect of Roux-en-Y gastric bypass on Barrett’s
esophagus: a systematic review. Surg Obes Relat Dis. 2021;17:221–230.
28. Andalib A, Bouchard P, Demyttenaere S, et al. Esophageal cancer after sleeve
gastrectomy: a population-based comparative cohort study. Surg Obes Relat
Dis. 2021;17:879–887.
29. Bevilacqua LA, Obeid NR, Yang J, et al. Incidence of GERD, esophagitis,
Barrett’s esophagus, and esophageal adenocarcinoma after bariatric surgery.
Surg Obes Relat Dis. 2020;16:1828–1836.
30. Wirth K, Kizy S, Abdelwahab H, et al. Bariatric surgery outcomes in Medicare
beneficiaries. Obes Sci Pract. 2021;7:176–191.
31. Wise ES, Amateau SK, Ikramuddin S, et al. Prediction of thirty-day morbidity
and mortality after laparoscopic sleeve gastrectomy: data from an artificial
neural network. Surg Endosc. 2020;34:3590–3596.
32. Grieco A, Huffman KM, Cohen ME, et al. The Metabolic and Bariatric
Surgery Accreditation and Quality Improvement Program bariatric surgical
risk/benefit calculator: 30-day risk. Surg Obes Relat Dis. 2021.
33. Skulsky SL, Dang JT, Battiston A, et al. Higher Edmonton Obesity Staging
System scores are associated with complications following laparoscopic
Roux-en-Y gastric bypass. Surg Endosc. 2020;34:3102–3109.
DISCUSSANT
Dr. Aurora D. Pryor
I am Aurora Pryor from Stony Brook, New York, and I would
like to thank the American Surgical Association for the invitation to
discuss this paper. I would also like to congratulate Daniel Leslie
from the University of Minnesota for this outstanding article evaluating GERD after vertical sleeve gastrectomy and RYGB. Dr. Leslie
652 | www.annalsofsurgery.com
used an insurance claims database to evaluate matched patients
undergoing RYGB and sleeve gastrectomy for a claims-based diagnosis of gastroesophageal reflux disease before and after surgery. The
groups were well matched. There was 77% GERD in the preop
bypass patient group and 80% in sleeve. Interestingly, there was no
significant difference in baseline BE. Following surgery at 91 days to
5 years, post-op GERD was seen in 56% or bypass patients and 60%
of VSG, with denovo GERD in 12 or 11.7%, respectively. Surprisingly, postoperative BE and do novo GERD were significantly less in
the sleeve population.
My questions to the author are the following:
1. As this cohort is based on an insurance claims database, is this
study possibly confounded by patients carrying a GERD diagnosis for empiric use of PPI beyond 90 days post-op? Or, is it
possible the PPI use was for an alternative diagnosis, such as
marginal ulcer? If so, did you look at alternative diagnoses or
break down the presence of GERD into timepoints further out
from surgery?
2. This study also noted a higher number of hospitalizations following gastric bypass. What were the reasons for those visits?
3. Did you look for any diagnoses of esophageal cancer in either
group, potentially as a sequelae of BE?
4. Following this study, for which patient populations would you
consider RYGB the preferred operation?
Thank you again for the opportunity to discuss this paper.
Response Dr. Daniel Leslie
We would like to thank the moderators, and the American
Surgical Association and its membership for the opportunity to
present our data; thank you, Dr. Pryor, for your thorough and
insightful review. Your questions significantly improved our
submitted manuscript.
To answer your first question, we attempted to control for early
postoperative confounding by only counting diagnoses beyond 91 days
through up to 5 years after surgery. We view the first 90 days as a
somewhat chaotic and choppy period of diagnosis and prescribing is
more empiric and protocol-based; for most patients, surgical healing has
completed by the 3 month point and longer-term GERD was analyzed
after that. Beyond 90 days, all bariatric teams prescribe empirically at
one point or another, and the study cannot decipher this intent. Typically,
a prescription requires a diagnosis and symptoms of esophageal reflux
are likely the most common for this. Endoscopies are not always done,
unfortunately. We also identified marginal ulcers in 6.4% of patients
who underwent gastric bypass and this needs to be part of a balanced
discussion. Reporting requirements for the OptumLabs Datawarehouse
mandate that >10 patients carry a specific diagnosis, so further breakdowns into windows of time and comparisons within those windows
become less useful, even with such a large data set.
Regarding your second question about hospitalizations, many
were for nausea, vomiting or abdominal pain and were brief;
however, the reporting is for all-cause hospitalizations and could
include those for orthopedic or obstetric reasons as well; we view
these data as a project for a different analysis.
To address your 3rd question, there were 73 cases of esophageal cancer post-bariatric surgery, 33 in gastric bypass and 40 in
sleeve, not significantly different and are in our submitted manuscript. Interestingly, none of these patients had antecedent Barrett’s.
Moreover, there were no esophageal cancers in the gastric bypass
population post-surgery and <11 in the sleeve population; the
difference was not significant and the exact number NOT reportable.
Finally, your last question regarded procedural selection and
our use of the gastric bypass. In our practice at the University of
ß
2021 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 274, Number 4, October 2021
Minnesota, the RYGB is the preferred operation in very few patients
currently; we exclude consideration of gastric bypass in patients with
a history of large abdominal wall hernia or prior small bowel or colon
resection, immune disorders, major psychological disorders, chronic
opiate requirement, long-term need for NSAID therapy, and history
of smoking or addiction. In some of these cases, we strongly support
medical weight management without bariatric surgery. For patients
ß
2021 Wolters Kluwer Health, Inc. All rights reserved.
GERD Outcomes After Bariatric Surgery
who don’t have these histories, a gastric bypass is considered for
patients who understand risks and prefer this approach.
The data I presented don’t support using the presence of GERD
as a primary reason for selecting a gastric bypass compared to a sleeve
gastrectomy. However, they certainly will inform our preoperative
discussion with patients about short- and long-term risks of the
operations and side effects in the first 5 years after surgery.
www.annalsofsurgery.com | 653
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.