Five-year outcomes of transoral outlet reduction for the treatment of weight regain after Roux-en-Y gastric bypass

Abstract

Background and Aims:

Transoral outlet reduction (TORe) is a common endoscopic treatment for patients with weight regain after Roux-en-Y gastric bypass (RYGB) with a dilated gastrojejunal anastomosis (GJA). This study aims to assess long-term efficacy of TORe.

Methods:

This was a retrospective review of prospectively collected data on RYGB patients who underwent TORe for weight regain or inadequate weight loss. The primary outcome was efficacy of TORe at 1, 3, and 5 years. Secondary outcomes were procedure details, safety profile, and predictors of long-term weight loss after TORe.

Results:

A total of 331 RYGB patients underwent 342 TORe procedures and met inclusion criteria. Of these, 331, 258, and 123 patients were eligible for 1-, 3- and 5-year follow-ups, respectively. Mean body mass index (BMI) was 40 ± 9 kg/m². Pre-TORe GJA size was 23.4 ± 6.0 mm, which decreased to 8.4 ± 1.6 mm after TORe. Patients experienced 8.5 ± 8.5%, 6.9 ± 10.1%, and 8.8 ± 12.5% total weight loss (TWL) at 1, 3, and 5 years with follow-up rates of 83.3%, 81.8%, and 82.9%, respectively. Of 342 TORe procedures, 76%, 17.5%, 4.4%, and 2.1% were performed using single pursestring, interrupted, double-pursestring, and running suture patterns, respectively, with an average of 9 ± 4 stitches per GJA. Pouch reinforcement suturing was performed in 57.3%, with an average of 3 ± 2 stitches per pouch. There were no severe adverse events. A total of 39.3% had additional weight loss therapy (pharmacotherapy or procedure), with 3.6% getting repeat TORe. Amount of weight loss at 1 year (β=0.43, p=0.01) and an additional endoscopic weight loss procedure (β=8.52, p=0.01) were predictors of %TWL at 5 years.

Conclusions:

TORe appears to be safe, effective, and durable at treating weight regain after RYGB.

INTRODUCTION

Roux-en-Y gastric bypass (RYGB) is a common bariatric surgery that is performed worldwide. Despite the efficacy and durability of RYGB, weight regain after initial weight loss occurs in many patients. Specifically, it has been demonstrated that after 10 years after RYGB, patients regain on average 20% to 30% of the weight they initially lost, with approximately one-third of the patient population gaining back almost all of their lost weight^1,2. Etiologies are likely multifactorial with possible contributors being behavioral, medical, and anatomical factors. From an anatomical standpoint, presence of gastrogastric fistula (GGF) and dilation of the gastrojejunal anastomosis (GJA) have been shown to be correlated with weight regain^3–6. Therefore, treatments focusing on reversing these 2 factors have been developed to induce weight loss in this patient population.

Transoral outlet reduction (TORe) is an endoscopic procedure focusing on reducing the size of the GJA. In 2013, a prospective, multicenter, randomized, blinded, sham-controlled trial (RESTORe trial) was conducted to evaluate the safety and efficacy of the procedure. At 1 year, patients who underwent TORe experienced 3.5% total weight loss (TWL), which was statistically significantly higher than those who underwent a sham procedure who experienced 0.4% TWL. This study therefore established level I evidence for TORe at treating weight regain after RYGB⁷. In this study, TORe was performed using a partial-thickness endoscopic suturing device. Since the study, a new full-thickness suturing device has become available and has been shown to be superior at performing TORe compared with the older partial-thickness system⁸. In 2013, we reported our initial experience with TORe using this newer device (Overstitch, Apollo EndoSurgery, Austin, Tex, USA)⁹. A few years later, a newer-generation of this device became available that simplified the suturing steps and allowed placement of a variety of suture patterns. Technical details of the procedure, such as mucosal ablation before suturing, suture pattern, and reinforcement suturing in the pouch, have also been further refined in order to optimize weight loss outcomes^10–13. With the most-common current technique, which includes mucosal ablation with argon plasma coagulation (APC) around the GJA followed by suturing of the GJA with or without reinforcement suturing in the distal pouch, patients experience approximately 8.4% TWL at 1 year with significant improvement in obesity-related comorbidities¹⁴. A medium-term study demonstrated that patients who underwent TORe were able to maintain significant weight loss up to at least 3 years¹⁵.

Despite its efficacy at 1 and 3 years, the longer-term effect of TORe remains unknown. This study aims to assess the safety, efficacy and durability of TORe at 5 years. Additionally, predictors of long-term weight loss are determined.

METHODS

Study Design and Patient Selection

This study was a retrospective study of prospectively collected data. The study was conducted at a single tertiary referral center with the bariatric center of excellence from October 2010 to May 2018. All patients with inadequate weight loss (defined as loss of less than 50% excess weight loss (EWL) after bariatric surgery) or weight regain (defined as gaining of at least 15% of maximal weight initially lost) who underwent TORe using the Overstitch device (Apollo EndoSurgery) with APC before suturing were included. At our institution, patients with increased tissue compliance or a dilated GJA were considered for TORe provided that they adhered to lifestyle modification and had inadequate weight loss after 6 months of a medical supervised program. Of note, subgroups of this patient cohort have previously been included in our group’s prior studies looking at shorter follow-up or different outcomes. All suture patterns were included, including interrupted, pursestring and running patterns (Figure 1). Of note, the majority of patients undergoing the pursestring pattern had this performed with a single suture as detailed in Figure 1. However, in some patients, a double pursestring pattern was used, which refers to when 2 separate sutures are used to place consecutive pursestrings, and being individually cinched. Regarding pouch suturing, from 2010 to 2012, with the first-generation device, pouch suturing was performed with an intention to reduce pouch volume. This was performed in patients with a pouch length of at least 4 cm. Data analysis for a related procedure suggested that GJA reduction and GJA aperture was more important than pouch volume reduction¹⁶. Therefore, since 2012, the primary focus of pouch suturing has been to reinforce the outlet sutures. In general, pouch suturing is almost always performed unless the pouch anatomy does not allow it, such as when the pouch length is shorter than 2 cm or it is severely angulated.

Figure 1.

Suture patterns used for transoral outlet reduction (TORe). A, Interrupted suture pattern: each suture is used to place 2 stitches across from each other. Several sutures may be used. B, Pursestring suture pattern: one suture is used to place multiple stitches (usually 8–14) around the GJA in a continuous circumferential fashion before cinching over a hydrostatic balloon (usually 7–12 mm). C, Running suture pattern: each suture is used to place greater than 2 stitches in a continuous noncircumferential fashion. Several sutures may be used.

Our clinic first started using weight loss medications as an adjunctive therapy to TORe in late 2016. In this study, patients who underwent adjunctive therapy for inadequate weight loss after initial TORe were included. These adjunctive therapies included pharmacotherapy (Phentermine, Topiramax, Phentermine/Topiramax, Bupropion/Naltrexone, Lorcaserin or Liraglutide), endoscopy (APC alone, additional TORe or gastric plication (restorative obesity surgery endoluminal (ROSE)), and surgery (GJA reconstruction or limb distalization). The decision to add an adjunctive therapy depended on a combination of patient’s clinical course and anatomy. Specifically, at 3 months, if patients had had less than 3% TWL or experienced diminished satiety or satiation, an adjunctive therapy was offered. Both medical and endoscopic adjunctive options were discussed with the patients along with their benefit and risk profile. If the patients did not have a contraindication to weight loss medications and wished to pursue, pharmacotherapy was added. For patients who had a contraindication to weight loss medications or elected to pursue an endoscopic adjunctive therapy, a diagnostic endoscopy was performed. If GJA was larger than 15 mm or had increased compliance, an endoscopic adjunctive therapy was performed. A surgical therapy was usually preserved for patients who did not respond or did not qualify for medical and/or endoscopic treatment options. Patients who underwent gastric plication as the initial therapy or procedures involving tissue resection during TORe were excluded. At our institution, patients with a smoking history within the past 6 months and those with marginal ulceration or GGF as demonstrated on diagnostic EGD were not offered the procedure. Details of the procedure and possible benefits and adverse events were discussed with the patients as per the standard protocol. Patients gave written informed consent before undergoing the procedure. After the procedure, patients were followed in the bariatric endoscopy clinic at months 1, 3, 6, and 12, and then every 6 months thereafter.

Outcomes

The primary outcome was amount of weight loss at 1, 3, and 5 years after the initial TORe. This was reported using absolute weight loss (AWL) and percent total weight loss (%TWL). Additionally, the proportion of patients who stopped gaining weight, achieved clinically significant weight loss, defined as %TWL of at least 5%, and the number needed to treat (NNT) to prevent weight regain at each time point were assessed. Secondary outcomes included procedural details, serious adverse event (SAE) rate and predictors of %TWL at 5 years.

Definitions

Suturing Terminology:

The suture used in this procedure is made of a 2–0 nonabsorbable polypropylene material. Each suture may be used to place multiple stitches via a variety of patterns (Figure 1). A suture comes attached to a needle, which also functions as a t-tag tissue anchor and remains in the tissue. A stitch is defined as a one in-and-out movement of a threaded needle through tissue (ie, a bite) during suturing. A cinch is defined as an automated suture fastening system that secures and cuts the suture.

Weight-related Terminology:

TWL was calculated using the formula: (TORe weight – follow-up weight) / TORe weight x 100 %. AWL was calculated using the formula: TORe weight – follow-up weight. EWL was calculated using the formula: (TORe weight – follow-up weight) / (TORe weight – ideal weight) with ideal weight being defined by the weight corresponding to a BMI of 25 kg/m². The amount of weight regain at index TORe was calculated using the formula: (TORe weight – nadir weight after RYGB) / (pre-RYGB weight – nadir weight after RYGB) x 100%. For patients who underwent more than one TORe procedures, the TORe weight at the time of the index TORe was used to calculate the amount of weight regain.

Adverse Event Terminology:

Adverse events were categorized using the ASGE lexicon as mild, moderate, severe and fatal¹⁷. Specifically, adverse events (AEs) were considered mild when the procedure was aborted due to the event or when there was an unplanned admission of 3 nights or less. Moderate AEs referred to when there was an unplanned admission of 4 to 10 nights, intensive care unit (ICU) admission for 1 night, or when transfusion, repeat endoscopy, or interventional radiology procedure was required. Severe AEs were when there was an unplanned admission of greater than 10 nights, ICU admission for greater than 1 night and when a surgery was required to address the AE. Fatal AEs referred to when death occurred due to the event.

Statistical Analysis

All continuous variables were expressed as mean ± standard deviation unless otherwise specified. Categorical variables were expressed as proportions (%). Paired Student t-test and chi-squared test were used to compare continuous and categorical variables before and after TORe, respectively. Univariable and multivariable linear regression analyses were used to determine predictors of weight loss at 5 years after TORe. Standardized β coefficients were reported and a significant 2-sided P value was set at 0.05 or less. All statistical modeling was performed using SAS version 9.4 software (Cary, NC, USA). The study was approved by the Institutional Review Board.

RESULTS

From October 2010 to May 2018, a total of 331 RYGB patients with weight regain (87%) or inadequate weight loss (13%) met the inclusion criteria. Baseline characteristics are shown in Table 1.

Table 1.

Baseline characteristics of the 331 patients included in the study.

Characteristics	N = 331
Age (years)	50 ± 11
Sex (female, %)	284 (86)
Pre-RYGB weight (kg)	139.3 ± 31.4
Nadir weight (kg)	84.1 ± 21.3
Weight at initial TORe (kg)	110.0 ± 26.3
BMI at initial TORe (kg/m2)	40.1 ± 9.1
Amount of weight regain (% from maximal lost weight)	51.0 ± 45.4
Duration from RYGB to initial TORe (years)	9.3 ± 4.7
Pre-TORe GJA diameter (mm)	23.4 ± 6.0
Pre-TORe pouch length (cm)	4.4 ± 1.8

Primary Outcomes

Out of 331 patients, 331, 258, and 123 patients were due for a 1-, 3-, and 5-year followup, respectively. Of these, data were available on 276, 211, and 102 patients, representing a follow-up rate of 83.3%, 81.8%, and 82.9%, respectively.

At 1 year, TORe was successful at preventing weight gain in 87.4% of the patient cohort, with the number needed to treat of 1.1. On average, patients lost 9.4±12.3 kg, which corresponded to 8.5±8.5% TWL (p<0.0001). Sixty-five percent of the cohort experienced at least 5% TWL.

At 3 years, TORe was successful at preventing weight gain in 79% of the patient cohort, with the number needed to treat of 1.3. On average, patients lost 8.7±13.8 kg, which corresponded to 6.9±10.1% TWL (p<0.0001). Fifty-four percent of the cohort experienced at least 5% TWL.

At 5 years, TORe was successful at preventing weight gain in 77% of the patient cohort, with the number needed to treat of 1.3. On average, patients lost 10.3±14.6 kg, which corresponded to 8.8±12.5% TWL (p<0.0001). Sixty-two percent of the cohort experienced at least 5% TWL (Figure 2).

Figure 2.

Long-term efficacy of transoral outlet reduction (TORe) at treating weight regain after Roux-en-Y gastric bypass (RYGB). A, Weight in kilograms (kg) before and at 1, 3, and 5 years after TORe. B, Percent total weight loss (%TWL) at 1, 3, and 5 years after TORe. C, Proportion of patients who achieved at least 5% TWL at 1, 3, and 5 years after TORe. Data presented as means ± S.E.

Secondary Outcomes

Procedure Details

A total of 331 patients underwent 342 TORe procedures. All procedures were performed using an endoscopic suturing device with the patient under general anesthesia. Before suturing, APC was performed on the gastric side of the GJA in all patients (100%). Subsequently, an average of 9±4 stitches were placed per GJA. Of the 342 TORe procedures, 260 (76.0%), 60 (17.5%), 15 (4.4%) and 7 (2.1%) were performed using single pursestring, interrupted, double pursestring and running suture patterns, respectively. The final GJA size was 8.4±1.6 mm, which represented a 64% reduction from the baseline diameter. Reinforcement suturing was performed in the distal pouch in 196 procedures (57.3%) with an average of 3±2 stitches per pouch.

After the procedure, all patients (100%) were placed on liquid diet for 45 days to optimize tissue healing¹⁸ and were prescribed open proton pump inhibitor (PPI) and sucralfate. Additionally, all patients (100%) were prescribed a soluble form of their home medications during the 45-day period. Out of 342 cases, 210 (61.4%) were admitted after the procedure with 9 cases (2.6%) being admitted due to nausea and vomiting and the remaining 201 cases (58.8%) being admitted per protocol, which had been initiated since August 2014.

Safety Profile

There were no severe or fatal AEs as graded by the ASGE lexicon. Moderate AEs occurred in 11 out of 342 cases (3.2%). These included abdominal pain due to clean based marginal ulceration on the gastric side of the GJA at the prior APC site (5), melena due to marginal ulceration at the prior APC site requiring medical management (3), GJA stenosis responding to balloon dilation (2) and GJA erosion (1). Other AEs were graded as mild (24 out of 342 cases; 7.0%). These included submucosal esophageal tear requiring clipping or suturing (5), superficial mucosal esophageal abrasion not requiring an intervention (5), oozing from a suture site that stopped spontaneously (4), bleeding at a suture site requiring either epinephrine injection or early cinching to tamponade (3), suture breakage (1), abdominal pain (3), nausea (1), non-cardiac chest pain (1) and severe constipation (1).

Adjunctive Therapies for Weight Regain

After the initial TORe, 130 of 331 (39.3%) patients received an additional therapy. Specifically, 95 (28.7%), 62 (18.7%), and 4 (1.2%) patients underwent additional endoscopic, medical and surgical therapy for weight regain (some patients received multimodality treatment).

Of the 95 patients who underwent additional endoscopic therapy for weight regain, 89.1%, 7.9%, and 3.0% of the procedures performed were APC (setting: flow of 0.8 L/min and power of 30–80 watts), suturing TORe and plication procedures, respectively. The majority of patients required an additional procedure early with 8 patients requiring an adjunctive procedure between 3 and 5 years from the initial TORe.

Of the 62 patients who required adjunctive pharmacologic therapy, prescribed weight loss medications included Phentermine/Topiramate (41.3%), Topiramate (22.2%), Phentermine (19%), Liraglutide (7.9%), Lorcaserin (6.3%) and Bupropion/Naltrexone (3.2%).

Of the 4 patients who underwent surgical revision for weight regain after initial TORe, 2 (50%) underwent GJA reconstruction and 2 (50%) underwent limb distalization.

Predictors of Weight Loss at 5 Years

On univariable linear regression analyses, the amount of weight regain at index TORe and %TWL at 1 year were significant predictors of %TWL at 5 years (β = 0.05, p = 0.03 and β = 0.32 and p = 0.04, respectively). Age, sex, pre-TORe GJA size, post-TORe GJA size, number of stitches around the GJA, number of stitches in the pouch and addition of adjunctive endoscopic, medical or surgical therapy for weight regain were not associated with %TWL at 5 years (Table 2).

Table 2.

Linear regression analysis of predictors of %TWL at 5 years after TORe.

Univariable				Multivariable
Variables	β	S.E.	P-value	Variables	β	S.E.	P value
Age	0.06	0.11	0.58	Age	−0.02	0.13	0.88
Male sex	0.69	3.05	0.82	Male sex	0.97	3.79	0.80
Weight regain	0.06	0.02	0.03	Weight regain	0.05	0.03	0.11
Pre-GJA size	0.08	0.18	0.68	Pre-GJA size	0.14	0.22	0.51
GJA stitches	0.32	0.31	0.30	GJA stitches	−0.24	0.43	0.58
Pouch stitches	0.18	0.39	0.65	Pouch stitches	−0.12	0.47	0.80
Post-GJA size	0.43	0.50	0.39	Post-GJA size	0.30	0.65	0.65
Endo	3.73	2.57	0.15	Endo	8.52	3.20	0.01
Med	−3.93	3.22	0.23	Med	−7.90	3.99	0.05
Surg	1.13	7.15	0.87	Surg	3.37	7.37	0.65
%TWL at 1 year	0.34	0.15	0.03	%TWL at 1 year	0.43	0.17	0.01

On a multivariable linear regression analysis, %TWL at 1 year remained a significant predictor of %TWL at 5 years after controlling for age, sex, amount of weight regain, pre-TORe GJA size, post-TORe GJA size, number of stitches around the GJA, number of stitches in the pouch and whether or not an adjunctive therapy for weight regain was performed (β = 0.38, p = 0.03). Additionally, performing an additional endoscopic procedure for weight regain in addition to the initial TORe was associated with a greater %TWL at 5 years (β = 7.64, p = 0.02), whereas adding weight loss medication(s) was associated with a smaller %TWL at 5 years (β = −9.37, p = 0.02) (Table 2).

DISCUSSION

This study demonstrates the safety, efficacy and 5-year durability of TORe in patients with weight regain or inadequate weight loss after RYGB. TORe appears to be effective at arresting weight regain in the majority of this patient population. Additionally, the amount of weight loss at one year and a combination of TORe with an additional endoscopic weight loss procedure predict greater weight loss at 5 years.

As far as we know, this is the first study to demonstrate long-term efficacy of up to 5 years of an endoscopic bariatric procedure. Although endoscopic bariatric procedures have previously been shown to be safe, minimally invasive, and effective during a short-term followup, no studies have established their long-term efficacy. In this study, patients who underwent TORe were able to maintain significant weight loss of 8.6% of their initial weight at 5 years. This amount of weight loss is considered clinically significant, as studies have previously shown that improvement in several obesity-related comorbidities including hypertension (systolic and diastolic blood pressure), hyperlipidemia (triglycerides and HDL) and diabetes (hemoglobin A1c) occur once patients achieve at least 5% TWL^19–21. Additionally, at 5 years, almost two-thirds of the patients in our cohort were able to maintain this threshold of clinically significant weight loss (≥5% TWL). Of note, given the retrospective nature of this study, comorbidities were not assessed or collected routinely unless clinically indicated, which prevented us from being able to perform a direct analysis of changes in comorbidities at this time. Furthermore, the number needed to treat in order to halt weight regain at 5 years was 1.3. This suggests a possible greater health benefit of TORe as these patients would have been at a higher weight if they did not undergo TORe. Therefore, the health benefit is in fact derived not only from the amount of weight loss they achieved from the baseline weight (i.e. TORe weight), but also from the differential between the higher weight they might have reached (if they continued to gain weight without TORe) and the baseline weight at the time of TORe.

In addition to durability, this study also demonstrates repeatability of endoscopic bariatric procedures. Specifically, in our cohort, 28.9% of the patients underwent an additional endoscopic procedure after the initial TORe. Of these, the majority (nearly 90%) were APC of the GJA alone, which is a relatively simple, outpatient procedure that can be performed with the patient under conscious sedation^22–25. The remainder was repeat TORe or plication procedure. Therefore, this suggests that it is technically feasible and safe to perform other available endoscopic procedures for weight regain after the initial TORe^{14,16,22,23,26,27}. On the regression analysis, adding an adjunctive endoscopic procedure to the initial TORe was associated with greater weight loss at 5 years. Therefore, for a subgroup of patients who achieve inadequate weight loss with the initial TORe, combining TORe with another endoscopic procedure may be considered as this approach appears effective at inducing further weight loss and at preventing patients from undergoing surgical revision (only 1 out of 100 patients who underwent an additional endoscopic procedure underwent surgical revision; data not shown).

This study also demonstrates the long-term safety of TORe. Specifically, the SAE rate of TORe at 5 years was 3.9%, with the majority being intraprocedural and device-related. Specifically, these included submucosal esophageal tear due to overtube placement and bleeding at a suture site. All of these were treated successfully endoscopically during the index procedure. The remaining postprocedural SAEs included bleeding from marginal ulceration and GJA stenosis, which may have been caused and/or exacerbated by APC performed before suturing to enhance tissue apposition. These SAEs all occurred shortly after the procedure (ie, within the first few months) and were treated successfully via an endoscopic approach. Furthermore, in our cohort, 4 patients underwent surgical revision for weight regain (2 surgical reconstructions of the GJA and 2 limb distalizations) within 5 years after the initial TORe without any intraoperative adverse events or technical challenges. Therefore, this suggests that TORe does not prevent subsequent surgical revision of RYGB if one is needed in the future.

This study had a few limitations. First, the study was conducted at a single bariatric center of excellence. Although this may affect the generalizability of our findings, most cases were performed with the participation of trainees under the supervision of an expert bariatric endoscopist. Additionally, all consecutive TORe cases that met the inclusion and exclusion criteria from 2010 to 2018 were included in the analysis. Throughout this period, techniques at our institution continued to evolve, such as APC settings, suture pattern, and final GJA size. Therefore, we suspect that the heterogeneity of experience levels and techniques would reflect real life experience of TORe. Another limitation was a retrospective design without a control group, which may have introduced bias. Furthermore, it is possible that patients who were willing to undergo TORe were more ready to adhere to lifestyle modification compared to the general weight regain population, leading to selection bias. Additionally, in our study, about a third of the patients received adjunctive therapy after the initial TORe. Nevertheless, the majority were APC alone, which was performed as a reinforcing procedure, with a small number of patients undergoing repeat TORe. This report likely reflected the real-life experience where an adjunctive weight loss procedure may be added to enhance and maintain the long-term outcome. In the regression analysis, any patient who received at least one prescription for any of the Food and Drug Administration (FDA)-approved medications for obesity between TORe and 5-year follow-up were included regardless of the duration of medication usage and/or early discontinuation due to intolerance and/or adverse events. As most patients were prescribed medications when they experienced early weight plateau or inadequate weight loss after initial TORe and duration of usage was unclear, the negative correlation between adjunctive medication usage and amount of weight loss at 5 years must be interpreted with caution.

In summary, TORe appears to be safe, effective, and durable at treating weight regain after RYGB. At 5 years after TORe, nearly all patients have cessation of weight gain with the majority experiencing clinically significant weight loss.

Acronyms

TORe

Transoral Outlet Reduction

RYGB

Roux-en-Y gastric bypass

GGF

gastrogastric fistula

GJA

gastrojejunal anastomosis

TWL

total weight loss

APC

argon plasma coagulation

IRB

Institutional Review Board

EWL

excess weight loss

ROSE

Restorative Obesity Surgery Endoluminal

AWL

absolute weight loss

%TWL

percent total weight loss

NNT

number needed to treat

SAE

serious adverse event

PPI

proton pump inhibitor