Long‐term impact of gastropexy on use of acid‐reducing medication, second operations for gastroesophageal reflux and subjective reflux symptoms after sleeve gastrectomy

We investigated whether adding gastropexy to sleeve gastrectomy (SG) reduced gastroesophageal reflux disease (GERD) in patients operated for severe obesity, assessed mainly by use of anti‐reflux medication (ARM) and second operations due to GERD worsening. In a prospective non‐randomized study, patients undergoing SG at two Norwegian hospitals were included from 2011 to 2015 and followed for 7 years. GERD was defined by regular use of ARM, and epigastric pain and heartburn were measured by the Rome II questionnaire. Gastropexy was done by suturing the gastrocolic ligament to the staple line. Patients undergoing SG only, mainly before gastropexia was introduced in 2013, were compared to those with additional gastropexy from 2013 onwards. Of 376 included patients (75% females, mean age 42.6 years and BMI 42.9 kg/m2), 350 (93%) and 232 (62%) were available for evaluation after 1 and 7 years, respectively. Baseline characteristics in the no‐gastropexy (n = 235) and gastropexy groups (n = 141) were similar. In patients without ARM use before surgery, the use increased and in those that used ARM at baseline, the proportion decreased, with no difference in the no‐gastropexy and gastropexy groups. With a combined endpoint of ARM use and/or second operation for GERD, there was no difference during follow‐up between the two groups. With time, adding gastropexy did not reduce symptoms of GERD significantly. In this population, adding gastropexy to SG did not reduce use of ARM and/or second operation for uncontrolled GERD, epigastric pain or heartburn during the first 7 postoperative years.


Key points
What is already known about this subject?
• De novo, or worsening of GERD are common complications after sleeve gastrectomy for severe obesity.
• Gastropexy has been proposed as a surgical procedure to prevent GERD after sleeve gastrectomy.
What this study adds?
• Adding gastropexy to sleeve gastrectomy was associated with longer operation time, but not with an increased rate of short-term complications.
• Adding gastropexy to the sleeve gastrectomy did not reduce GERD after surgery when assessed by use of anti-reflux medication, second operations for severe GERD, epigastric pain or heartburn.

| INTRODUCTION
Obesity is one of the major risk factors for gastroesophageal reflux disease (GERD), a condition where reflux of gastric contents causes symptoms and/or complications, and GERD is common in patients seeking bariatric surgery. 1,2 GERD outcomes after sleeve gastrectomy (SG) are significantly worse than after Roux-en-Y gastric bypass (RYGB), and de novo GERD and Barrett's esophagus have been reported after SG. Technical modifications introduced to ameliorate GERD symptoms after SG are currently being investigated. [3][4][5][6] SG, a preferred bariatric procedure globally, has the advantages of preserving the normal continuity of the gastrointestinal tract with no anastomoses, and fewer metabolic disruptions. 6 However, in randomized controlled trials (RCTs), SG is associated with a higher rate of GERD symptoms and GERD-related complications in the years following surgery compared to RYGB. [7][8][9][10] For instance, in the SM-BOSS trial, the rates of GERD worsening and de novo GERD were both 32% at 5 years after SG, compared to 6% and 11% after RYGB. 7 In the SLEEVEPASS trial, with 10 years follow-up available, the prevalence of esophagitis was 31% after SG compared to 7% after RYGB. 8 The combined analysis of these two landmark RCTs further showed that surgical reintervention for severe GERD symptoms was performed in 16 of 228 patients after SG compared to none of 229 after RYGB. This is concerning as acid reflux into the esophagus increases the risk of complications such as Barrett's esophagus, stenosis and/or esophageal cancer. 11,12 The presence of GERD may also impair patients' quality of life and social functioning. 8,[13][14][15] It has been proposed that the increased incidence of GERD and related complications after SG is caused by loss of gastric fixation, e.g., by disrupting the phrenoesophageal ligament, leading to improper positioning of the sleeved stomach with intrathoracic migration of the gastroesophageal junction and remaining ventricle. 16 Furthermore, prevention of strictures, kinks or twists of the gastric remnant is important as these may increase intragastric pressure and cause reflux. 17 These possible mechanisms have motivated gastropexy or omentopexy as means to stabilize the position of the gastric remnant by suturing the gastrocolic ligament, separated from the gastric wall during the SG procedure, back onto the staple line. Other changes introduced by surgery, such as possible damage to the sling fibres during SG, may not be alleviated by such fixation. 18 Gastropexy was pioneered by Lucius D. Hill as a surgical treatment for hiatal hernia, but the efficacy of several modified techniques of gastric fixation to abdominal structures in alleviating GERD after SG is still unclear. 19,20 An RCT from Egypt with 200 patients undergoing SG showed a lower incidence of reflux symptoms during the first three postoperative months after the addition of gastropexy, as measured by dose and duration of ARM usage. 21 However, in another smaller double-blinded RCT from the United States, adding gastropexy did not significantly improve symptoms from GERD 1 year after surgery. 22 A prospective study from one Norwegian hospital evaluating the effect of adding gastropexy to SG showed a clear reduction in use of anti-reflux medication (ARM) at 2 years compared to a historical cohort operated with SG alone. 17 To expand the knowledge of how gastropexy may affect GERDrelated outcomes when added to SG, we prospectively recorded changes in ARM use, second operations for severe reflux symptoms not adequately controlled by ARM, and symptoms of epigastric pain and heartburn up to 7 years after SG. We compared two cohorts before and after the introduction of gastropexy as a routine adjunct to the SG operation. Our objective was to determine whether adding gastropexy to the SG procedure was associated with a decline in these GERD-related outcomes in a long-term follow-up study.

| METHODS
This two-center observational study is part of the project 'Bariatric Surgery on The West Coast of Norway', approved by the Regional The study design has been described in detail previously. 23 In brief, patients were included at the community hospitals in Voss and Haugesund that serve patients from non-overlapping geographical regions. 23 Eligible patients (BMI ≥ 40 kg/m 2 or ≥ 35 kg/m 2 with obesityrelated comorbidities, age 18-70 years, no alcohol or drug abuse and no active psychosis) scheduled for bariatric surgery were invited to participate. 24 The present analysis includes patients undergoing SG at either hospital. We collected demographic, clinical and biochemical data using standardized checklists 2-3 months before surgery, and at routine outpatient visits 3 months, 1, 2 and 5 years postoperatively, all detailed in the protocol. Five-year data were supplemented with an electronically administered survey on average 7 years after surgery, specifically also capturing symptoms related to GERD and ARM use.
Hospital records were reviewed to ensure consistent recording of per-operative gastropexy and/or performance of hiatal repair, use of ARM and reoperations performed for GERD symptoms not sufficiently controlled by medication. Routine evaluation for GERD by endoscopy, esophageal manometry or pH monitoring was not done, but could be part of the evaluation of selected patients during follow up, e.g., before reoperation. Written informed consent was obtained from all patients prior to inclusion.

| Surgical procedures
All patients were part of a comparative study of SG and RYGB, allocated to the preferred procedure at their respective hospital. In a limited number of cases, an individual decision as to the surgical procedure was allowed. Pre-and postoperative care were similar at both hospitals and included prescription of a low-calorie diet (<1000 kcal per day) 3-4 weeks prior to surgery. SG was performed laparoscopically with a gastric resection using a 32 French tube, start-

| Outcome definitions
The primary endpoints for our analysis were use of ARM or undergoing a secondary operation for GERD symptoms not adequately controlled by medication. As prespecified in the protocol, ARM use as Stepwise illustration of the gastropexy surgical procedure. The gastrocolic ligament is sutured to the sleeve at the oral end of the stapler line (A), continuous suture-line with V-Loc™ to ensure inclusion of sufficient tissue of both the gastrocolic ligament and stapler-line of the sleeve (B). In this case, the suture includes the full length of the stapler-line (C). proton pump inhibitors with or without additional medication was recorded for each timepoint. Patients who underwent a secondary operation for GERD symptoms with or without other simultaneous indications for reoperation, such as inadequate weight loss, were recorded to have reached the endpoint for all following visits, irrespective of ARM use. Patients undergoing a second bariatric procedure for inadequate weight loss or other complications not related to GERD were excluded at the time of operation.
Before surgery, and at 1, 5 and 7 years, we obtained patient reports of epigastric pain and heartburn by the following two questions derived from the Rome II questionnaire for functional esophageal disorders: In the last 3 months, did you often have pain in the middle of your chest? and In the last 3 months, did you often have heartburn, a burning pain or discomfort in your chest? 25,26 Response categories were yes or no.
Weight was assessed according to international guidelines. 27 Baseline weight (in light clothing without shoes to the nearest 0.1 kg), height (in a standing position without shoes to the nearest 1 cm) and BMI were recorded at the first preoperative visit and at all follow-up visits.
Early major postoperative complications within 30 days and late major complications were classified as Clavien-Dindo ≥ 3b. 28 Length of hospital stay was counted from day of operation to discharge from hospital to home, excluding intermittent days outside of hospital care.

| Statistical analysis
Categorical and continuous variables are presented as percentages and mean values with standard deviations (SD) or 95% confidence intervals (CI). Groups of patients at defined timepoints were compared using chi-square and two sample t-tests as appropriate.

| RESULTS
Of 376 SG patients operated between September 2011 and February 2015 (75% females, mean age 42.6 years, mean baseline BMI 42.9 kg/m 2 ), 350 (93%) and 232 (62%) were evaluable after 1 and 7 years, respectively ( Figure 2). No-gastropexy was performed before F I G U R E 2 Flow-chart of included patients. Merged short-and long-term time points include patients with data available from 1 and/or 2 years or 5 and/or 7 years follow-up. n, number of patients with data registration at respective timepoint.
T A B L E 1 Patient baseline characteristics according to type of surgery. Baseline patient characteristics in the no-gastropexy (n = 235) and gastropexy (n = 141) groups were similar (Table 1). For an attrition analysis, we compared patients available for follow-up at 7 years to those who were not. No significant differences at baseline in any of the groups were seen, except more patients reporting heartburn at baseline in the no-gastropexy group were lost to follow-up (  group. The rate of major early complications was similar in both groups ( Table 3). The proportion of patients with major late complications was 12.8% in the no-gastropexy group and 7.1% in the gastropexy group ( p = .08). Most major late complications were second operations because of GERD with or without inadequate weight loss, done in 11.9% and 6.4% of the cases in the respective groups ( p = .08). Reoperations solely due to GERD symptoms not controlled by ARM were done in 5.5% and 3.5% of the patients in the nogastropexy and gastropexy groups, respectively ( p = .38).
There was a significant association between ARM use and symptoms of heartburn at all timepoints after surgery. At 1 and 7 years, 44.7% and 63% of those with heartburn used ARM, respectively, compared to 12.0% and 37.2% of those without (p < .001). No such difference was seen between reports of epigastric pain and ARM use (data not shown).
In patients not reporting ARM use prior to surgery, the use increased significantly and at similar rates in the no-gastropexy and gastropexy groups; from 0 at baseline to 40.4% and 44.5% at 7 years after SG, respectively (Table 4). In patients who used ARM prior to surgery, the proportion decreased to 60.7% and 65% at 1 year in the no-gastropexy and gastropexy groups and was found to be 66.7% and 81% at 7 years. In patients who did not use ARM before surgery, rates of ARM use were generally lower at all timepoints after surgery compared to those with pre-operative use of ARM, but again there was no difference between the no-gastropexy and gastropexy groups at any timepoint during follow-up (Table 4).
In mixed effect analysis, there was no difference in the combined endpoint, the proportion of patients with ARM use and/or second operation for GERD symptoms not adequately controlled by medication over the study period (Table 5, Figure 3).
Similarly, the number of patients reporting epigastric pain increased significantly from baseline to 7 years after SG, with no difference between the no-gastropexy and gastropexy groups. Over time, BMI was similar for patients who did not undergo gastropexy as compared to those who did (Table 5, Figure 3).
Patients reporting heartburn were differently distributed between the no-gastropexy and gastropexy groups at several timepoints. At baseline, more patients reported heartburn in the no-gastropexy group than in the gastropexy group, but the difference was not statistically F I G U R E 3 Changes from surgery to 7 years after sleeve gastrectomy in predicted probability for use of anti-reflux medication (ARM) and/or surgery for inadequately controlled gastro-esophageal reflux disease (GERD) (A), epigastric pain (B) or heartburn (C) and body mass index (D). significant. With time, more patients suffered from heartburn after gastropexy, and in the mixed effect model, the interaction between time and surgical technique reached significance (p = .01; Table 5, Figure 3).

| DISCUSSION
In the present non-randomized comparison of two cohorts, both part of the same prospective study investigating effects of bariatric surgery, we found no effect of adding gastropexy to SG in terms of preventing ARM use, second operations for GERD symptoms not adequately controlled by mediation, or self-reported symptoms of epigastric pain or heartburn during 7 years of follow-up.
Our study was motivated by the limited data to support routine use of gastropexy as an adjunct to SG in patients undergoing surgery for severe obesity. Two smaller randomized trials were previously published, both with 1-year follow-up, but with conflicting conclusions. 21 to detect a 50% difference in food intolerance symptoms, i.e., it was not focusing primarily on GERD. Therefore, even meaningful differences in GERD symptoms at 1 year may have been missed by this smaller study. 22 Another explanation for differences in outcome may involve surgical technique, i.e., alternative ways of which the gastrocolic ligament is fixed to the gastric remnant. All three reports cited here state that the gastropexy procedure involved suturing the omentum back to the staple line or greater curvature, with the length of the fixations being clearly defined. In our cohort, the proximal extension of the fixation appears to have varied depending on individual choices of the surgeon intraoperatively. If fixation of the gastric remnant to prevent torsion, kinks or intrathoracic displacement is a mechanistic determinant of success, one may assume that the length of fixation plays a role for outcome.
With 38% of patients developing de novo GERD and only 5% of those with pre-existing GERD entering remission at 7 years, patients in our prospective study have a high symptom burden of GERD and high rate of ARM use, even despite adding gastropexy. 23 In a recent systematic review, it was estimated that up to 30% of patients may experience some GERD symptoms after SG, but most do not require operative therapy and can be treated successfully with medication. 8 Unfortunately, we did not perform pre-and postoperative gastroscopy as a routine, and postoperative gastroscopy was only performed in a small minority of patients with severe symptoms of reflux, vomiting or if a leak was suspected. The presence of symptoms cannot be considered a reliable indicator of higher-grade acid reflux or endoscopic mucosal changes as many of these may be asymptomatic. 29,30 Furthermore, mucosal damage due to acid reflux may be asymptomatic. 30 Nevertheless, common reporting criteria for outcomes after bariatric surgery acknowledge use of medication and changes thereof as indicators also for GERD, and ARM use was the main outcome in one of the studies that has shown benefit of adding gastropexy to SG. 17,27 Furthermore, in the 10-year follow-up of the SLEEVEPASS study, where 73% of all patients volunteered to a second gastroscopy as part of their 10-year evaluation, the rate of objective esophagitis correlated with the worsening of reported symptoms, rate of ARM use and reduced GERD health-related quality of life. 8 Other late complications of GERD, such as stenosis and Barrett's esophagus, were rare both after SG and RYGB and no differences were detected. However, such a correlative analysis of patients after SG supports the notion that gastropexy may not reduce rates of more objective GERD findings in our cohort of patients.
The main limitation of our study is the non-randomized design comparing patients operated on before and after the introduction of gastropexy as an adjunct to SG and the lack of objective measures of GERD. However, the large number of patients compared to other studies, the use of gastropexy as part of routine bariatric surgery, the long-term follow-up with fair retention rates add valuable information to the field.
Taken together with the available evidence, our data do not support routine use of gastropexy to prevent or ameliorate GERD after SG. The conflicting results as to the efficacy of gastropexy warrant a RCT, which ideally should include pre-and postoperative endoscopy, use of validated patient-reported outcome measures and need for ARM, and with a clear and uniform surgical technique.

| CONCLUSION
In this prospective non-randomized cohort study, addition of gastropexy to SG did not significantly reduce the use of ARM, risk of any secondary operation for GERD symptoms not adequately controlled by medication, symptoms of heartburn or epigastric pain at any point during a 7-year trajectory after surgery. Definition of the optimal surgical technique and evaluation in an RCT are warranted.