Introduction

Morbid obesity is an accumulation of adipose cells such that there are significant negative health implications [1–3]. It is increasingly prevalent in Australia and worldwide and is a major cost burden due to the morbidity and mortality associated with diabetes, cardiovascular disease, hypertension, and sleep apnea [1–4]. Prevention has been suggested to be the only cost-effective method to manage this epidemic; however, effective preventative methods are not yet available [1, 5].

Currently, surgical therapy as part of a multi-disciplinary approach to obesity, offers the only practical means of maintaining long-term weight loss and a reduction in obesity-related comorbidities [1–3, 5, 6]. With the success of bariatric surgery, the number of primary obesity procedures in Australia has increased by a factor of 10 between the years 2000 (1,259 procedures) and 2010 (11,015 procedures) (Australian Government, Medicare data). Over the same time however, the number of reoperations has increased from 207 to 2,084 (Australian Government, Medicare data).

The laparoscopic adjustable gastric banding (LAGB) is the most common bariatric procedure in Australia [1, 7]. It is a relatively simple operation with most reports identifying a perioperative morbidity rate of 1% to 5% and mortality of 0% to 0.05% [1, 8–11]. It is not resource intensive; most patients only require a 23-h hospital stay. In terms of function, it is simple to adjust and causes early satiety after small volume meals [1]. Multiple series confirm sustained excess weight loss (EWL) over 5 to 10 years of 42.8–73.2% [1, 12–16]. If there are complications or band failure the LAGB may be surgically removed. While removal does not leave the stomach completely unscarred, there is no permanent anatomic alteration [9].

The relative disadvantages of the LAGB include the need for frequent follow-up and band adjustment, variable weight loss, and the potential for a number of adverse band-related symptoms; specifically, dysphagia, odynophagia, vomiting, and gastro-esophageal reflux [2, 9, 17]. In addition some patients, even without symptoms, are found to have radiologically demonstrable esophageal dilatation (pseudo achalasia) or Barrett’s metaplasia after long-term band use [9, 17].

Complication rates after all bariatric procedures increase with time [11, 18, 19]. The common complications seen after LAGB are band slippage, band erosion, band infection, or pouch dilatation and port damage or infection [1, 17, 20–22]. Failure to lose weight or weight regain is also a problem for some patients [2, 18, 19]. For these reasons, the reported LAGB surgical revision rate is between 3.3% and 59.3% [11, 14–16, 20, 21, 23–29], and in one report as high as 70% [30].

Surgical options in patients with band complications or failure to lose weight include band removal, band repositioning (where appropriate), band replacement, or conversion to an alternate procedure. Band removal without further intervention results in a very high rate of weight regain [31, 32]. Although LAGB repositioning or replacement are technically possible, it is not clear if this is superior to conversion to an alternate procedure. The short-term requirement for further reoperation is up to 45% [23, 32] and there is less convincing support for further weight loss and comorbidity resolution [17, 23, 28, 31, 33, 34].

Many groups have reported conversion to Roux-en-Y gastric bypass (RYGB) as the procedure of choice in these patients [25, 31, 32, 35–37]. Patients describe improved food tolerance and a better quality of life [22, 23, 31, 33, 38], and it is likely that this procedure has the most predictable further weight loss and comorbidity resolution [22, 23, 31, 33, 38].

We have used RYGB as the salvage procedure for complicated or failing LAGB. The purpose of this non-randomized, prospective case series was to assess the safety of single-stage conversion to RYGB, and to assess the efficacy of this procedure.

Methods

A prospective database is maintained for all bariatric procedures. Data was extracted for patients converted from LAGB to RYGB from December 2007 to April 2011. Patients were recruited from our own clinic but also referred from other centers. Up to April 2011, our unit had implanted 2,210 gastric bands. The senior surgeon (GH) performed or directly supervised all procedures.

Conversion to RYGB is offered to patients between the ages of 18 and 65 after band failure or complication. Band failure is defined as inadequate weight loss (EWL of less than 50% of pre-operative weight or a BMI of greater than 35 kg/m2) or band intolerance. Band intolerance is defined as dysphagia, odynophagia, volume reflux (including nocturnal aspiration), regurgitation, or vomiting which were unable to be managed with band volume adjustment. LAGB complications included erosion (part of band visible from within the stomach), band slippage, band sepsis, proximal pouch dilatation, or esophageal dilatation.

Patients with band erosion are considered a special case. If erosion is diagnosed pre-operatively, then band removal followed by conversion to RYGB is performed in a staged fashion. If the diagnosis of erosion is made intra-operatively, then conversion to RYGB proceeds only if the portion of stomach, which has been perforated by the band, can be excised and the anastomosis made above this. If not, the band-related gastrotomy is closed and the bypass completed in a second operation.

Band position and pouch size are assessed pre-operatively with a liquid radio-contrast meal. Selective referral for pre-operative anesthetic and medical optimization is made. Dietetic consultation with 2 weeks of very low calorie diet is used in patients with a BMI > 45 for pre-operative weight loss and reduction of liver volume. All bands are deflated for at least the week preceding surgery. Operative technique is described below.

Post-operatively patients are seen at 4 weeks and then 3 monthly for the first year. This is extended to 6 monthly for the second and third years. Lifelong multivitamin supplements are commenced at the first post-operative review. Surveillance of electrolytes, liver function, vitamins A, B1, B6, B12, D, E, folate, zinc, iron studies, and lipid profile occurs twice yearly.

At each post-operative review patients were asked about dietary tolerance, vomiting, and reflux. Dietary tolerance is classified as solid diet consisting of all foods, soft diet with complete variety, pureed diet, or liquid diet. Vomiting was defined as that symptom-related to specific food intake and not to other unrelated illness. Reflux was defined as the requirement for daily proton pump inhibition with recurrence of symptoms with medication cessation. These definitions are used for concordance with those of the Swedish Obese Subjects Study [3].

Demographic data, indications for surgery, 30-day morbidity and mortality, and post-operative weight loss and symptoms are presented. Major complications are defined as those requiring re-admission to hospital or surgical intervention. Weight loss is described as percentage of excess BMI lost (%EBL) and change in BMI [39].

All data is presented as median and (range). Nominal data are presented as fractions. Statistical analysis was performed using the Chi square test with Yates correction to compare frequencies between two or more groups.

Operative Technique

Surgery is performed under general anesthesia with a single dose of prophylactic antibiotics and perioperative deep venous thrombosis prophylaxis. Access to the abdomen is obtained using a 12-mm bladeless optical trocar (Ethicon Endo Surgery, Cincinnati, Ohio). Adhesions are divided as necessary to allow placement of two further 12-mm ports and a 5-mm port such that they lie roughly transverse across the abdomen in the transpyloric plane. A liver retractor is inserted through an epigastric incision.

Adhesiolysis and band mobilization are performed first. Mobilization of the hepato–gastric adhesions allows retraction of the left lobe of liver from the proximal stomach. The band tubing is then followed cranially dividing the capsule around the prosthesis. This leads to the gastro–gastric sutures, which are incised, and any remaining capsule removed. This step is easier with the prosthesis in situ as diathermy to tissue overlying the band does not injure the underlying stomach [40]. After removing the capsule, the gastric band can be removed by unbuckling or cutting. Further exposure of the angle of His and separation from the left crus of the diaphragm is possible after band removal. It is important to mobilize the stomach from anteriorly as close to the angle of His as possible. This improves access for the subsequent creation of the gastric pouch.

After band removal, the size of the gastric pouch and the thickness of the infra-band capsule are assessed. The volume of the new gastric pouch should be less than 20–30, and this requires transverse stapling approximately 6 cm below the gastro-esophageal junction. This point is usually below the level of the previous band. Using this unscarred plane minimizes the risk of staple line insufficiency. If it is necessary to divide the stomach at the level of the band capsule, then it is important to remove this prior to stapling. The capsule is usually easy to peel from the underlying stomach.

The lesser sac is then approached from lesser curve of the stomach. A window is created using an ultrasonic dissector (Ethicon Endo Surgery, Cincinnati, Ohio), staying within the vascular arcade of the descending branch of the left gastric artery. This allows the passage of a linear stapler (Ethicon Endo Surgery, Cincinnati, Ohio), which is fired transversely against a 46-Fr bougie, using 3.5 or 4.1 mm staples depending on the thickness of the tissue. After the first firing, further clearance of posterior gastric adhesions is possible, and this allows exposure of the angle of His from the anterior and posterior. It is important to mobilize this fully in order to minimize pouch size and prevent the retention of a large redundant segment of gastric cardia. The endoscopic stapler is then fired towards the angle of His, while abutting vertically against the previously inserted bougie. In revision surgery, because of the thickness of the tissue, we are careful not to staple too tightly against the bougie. The staple line on the remnant stomach is oversewn to reduce the risk of staple line bleeding.

The duodenal–jejunal flexure is then identified and a 50-cm biliopancreatic limb measured. A side-to-side stapled anastomosis is created to a 100-cm alimentary limb, which is brought ante-colic and ante-gastric to the gastric pouch. In patients with a BMI greater than 50 kg/m2, a 150-cm alimentary limb is created. The mesenteric defect between the two limbs is closed. Our preference is to create a 12-mm hand-sewn, end-to-side gastro-jejunostomy; however, linear or circular stapling devices can be used. After completing the anastomosis, a nasogastric tube is placed and the anastomosis tested with air insufflation. The band port is removed as the final stage of the procedure.

Post-operatively, patients are commenced on proton pump inhibition for 6 weeks, to reduce the risk of stomal ulceration. The nasogastric is left for 24 h and a radio-contrast study ordered on post-operative day 2. Patients were commenced on a fluid diet after this study and usually discharged on day 3. Fluid diet is continued for 2 weeks.

Results

Eighty-two patients in total were treated by LAGB removal and conversion to RYGB. Median age was 49 years (range 18–66), and 70 patients were female. Pre-operative data and indications for surgery are shown in Tables 1 and 2.

Table 1 Pre-operative data
Full size table
Table 2 Primary indications for conversion to RYGBa
Full size table

A summary of operative data is shown in Table 3.

Table 3 Operative data
Full size table

Median post-operative length of stay was 4.5 days [3–39]. There was no 30-day mortality. One patient died from a myocardial infarct 60 days post-operatively, this was unrelated to surgery. Total post-operative morbidity was 46.3%. Major and minor complications are listed in Table 4.

Table 4 Post-operative complications
Full size table

Patients with wound sepsis were stratified by method of gastro-jejunostomy formation. Anastomosis with a circular stapler was associated with a 9/34 (26.4%) wound sepsis rate, with a linear stapler, 1/26 (3.8%), and hand-sewn anastomosis 0/22 (0%). When anastomosis with circular stapler was compared against both other anastomoses, this difference was found to be significant (p = 0.00286).

Anastomotic strictures occurred in 2/18 (11.1%) patients having two-stage conversion and 14/64 (21.9%) patients having single-stage conversion. This difference was not significant (p = 0.4957). Anastomotic strictures occurred in 3/13 patients when a 21-mm circular stapler was used compared with 1/21 patients when a 25-mm stapler was used (p = 0.2740). Anastomotic strictures occurred in 9/22 patients in whom a hand-sewn anastomosis was performed. When compared with strictures by all other methods of anastomosis (8/60), this difference was statistically significant (p = 0.012)

Post-operative weight loss data is shown in Tables 5 and 6. Table 6 shows only those patients treated for failed weight loss after gastric banding.

Table 5 Post-operative weight loss for entire patient group
Full size table
Table 6 Post-operative weight loss for patients with poor weight loss after LAGB
Full size table

Discussion

LAGB is the most common bariatric surgery in Australia because of good sustained weight loss and very low morbidity and mortality [41]. Like all bariatric procedures, gastric banding has a long-term complication and failure rate, and many of these patients require further treatment [11, 14–16, 20, 21, 23–29]. The aim of revision procedures is to eliminate adverse symptoms from the primary procedure, to treat complications and usually to provide further weight loss and comorbidity control. Many authors have reported that conversion of failed or complicated LAGB to RYGB is the most effective means of achieving these goals (Table 7) [17, 22, 24, 28, 37, 40, 42].

Table 7 Conversion of LAGB to RYGB [17, 22, 24, 28, 40, 43]
Full size table

Reoperative bariatric surgery is considerably more challenging than primary surgery due to post-operative adhesions, distorted tissue planes, and altered anatomy. Unsuspected band erosion must be recognized, and may require conversion in two stages [36]. In addition stapled division through scarred and thickened tissue may not be as secure. The morbidity with revision surgery is therefore much higher (Table 7) than for primary bypass procedures, which is usually reported as 6–9% [10, 13]. Recommending a patient undergo conversion from LAGB to RYGB therefore requires thorough understanding of the patient’s goals and tolerance of operative risk. Although there is no specific volume data, consideration of referral of this difficult procedure to subspecialty bariatric centers is likely to be beneficial.

The results from the 82 patients treated in this series confirm that conversion to RYGB remedies band-related symptoms and provides further weight loss after failed LAGB. The rate of perioperative morbidity was high in comparison with reported results for primary bypass but consistent with other reported studies for revision (Table 7) [10].

Laparoscopic conversion was possible in 96% of patients. This is consistent with most recent reported series (Table 7). It is expected that the laparoscopic approach for conversion to RYGB will have the same benefits as laparoscopic approach for other procedures namely, less pain, earlier mobilization, quicker return to work, and less wound complications, although there is no data specifically pertaining to this procedure [43, 44]. Vision and exposure of the gastric band, cardia, and angle of His is improved with laparoscopic access. Laparoscopic approach also allows for multipoint retraction aiding in dissection. Calmes et al. reported lower complication rates using laparoscopic conversion compared to open conversion [36].

Single-stage conversion to RYGB was possible in 78% of the patients in this group. The advantages of single-stage conversion are single anesthetic, reduced patient waiting time, better distribution of resources, and access to services, which can be very difficult for publically funded patients in our system. Technically, we have found that removal of the gastro–gastric plication, and dissection at the angle of His can be assisted by the gastric band which provides counter tension, acts as a target for dissection, and as a guide to the anatomy of the proximal stomach [17]. Van Niewenhove and co-workers reported a lower risk of gastro-jejunostomy strictures in two-stage conversions [45]. Although there was a trend toward a higher rate of strictures in single-stage conversion in our group, this was not a statistically significant difference. The main disadvantage of two-stage procedures in our practice is the limitation to access imposed by a publicly funded system. Patients usually have marked weight gain during the waiting period for the second stage of conversion, which can be extensive.

Despite our preference for single-stage conversion, the two-stage procedure should always be maintained as a surgical option. We use this if technical difficulties are found. In the setting of unexpected band erosion, unless the pouch can be created above the defect, the band should be removed and the defect repaired as the first stage [31]. In the setting of marked pouch dilatation, delaying RYGB allows some resolution of the dilatation and prevents stapling through effaced tissue [29]. In the setting of band sepsis, we prefer not to create an anastomosis in a contaminated field.

In the group of patients converted for poor weight loss, median %EBL was 58% at 12 months. This is in the upper part of the range reported for revision bypass surgery (Table 7), and in contrast to the study by Zingg et al. who found that weight loss after revision RYGB is probably less than for primary RYGB (but better than revision banding) [46]. Our results add to the body of evidence supporting good weight loss after conversion.

The ten patients with adverse band-related symptoms had complete resolution of these after conversion to gastric bypass. Other series have reported improved quality of life after conversion to bypass [37, 47]. Patients report less vomiting, no choking sensation when they eat, and better satiety [37]. There is no requirement for frequent visitation and band adjustment [37]. One shortcoming of this report has been a failure to objectively measure this symptomatic improvement with quality of life questionnaires.

In this series, the 12% of patients with port site infections was unacceptably high. In almost all of these patients wound sepsis occurred at the port site used for the introduction of the circular stapler for intra-corporeal docking with the anvil (Ethicon Endo Surgery, Cincinnati, Ohio). Nine of the ten wound infections in this series were in patients in whom a circular stapler was used. When compared to patients in whom an alternate anastomotic method was used this difference was statistically significant. It is felt that significant trauma to the skin edges and subcutaneous tissue is caused by introduction of the stapler and that this predisposes to sepsis. We have since ceased using this device and have had a marked reduction in the incidence of infection.

In this series there was also a high rate of anastomotic stricture, which was particularly high in the group with a hand-sewn anastomosis. This may be attributed to the fact that the gastro-jejunostomy is in, or near a previously operated field; however, other authors have not reported rates as high [45]. We surmise the high rate of strictures in our series is related to our attempts to maximize weight loss in this difficult to treat group of patients. Since the assessment of this data we have marginally increased our anastomosis width to compensate for this. In our entire patient series, we have not seen a substantial nor significant difference in anastomotic stricture rate between hand-sewn and stapled anastomosis (unpublished data). As such, we have until now continued with the hand-sewn method.

Conclusion

As the number of LAGB procedures continues to increase as will the revision rate and requirement for conversion to other procedures. We use laparoscopic RYGB as the standard revision procedure after failed or complicated LAGB. Laparoscopic conversion is technically challenging but can be performed, in most circumstances, as a single-stage procedure. Although mortality is low, there is substantially more morbidity when revision bypass is compared to primary bypass. Conversion results in effective symptom resolution and further weight loss in post-band patients but referral to subspecialty units should be considered due to the operative difficulties and the increased morbidity.