Success in bariatric surgery is classically defined as greater than 50% of initial excess body weight loss (EWL). Average percent EWL show greatest initial and long-term weight loss with biliopancreatic diversion (BPD) and the duodenal switch (DS) (range, 75–80%), followed by Roux-en-Y gastric bypass (RYGB) (range, 60–85%) and adjustable gastric banding (AGB) has the lowest average percentage EWL with a range of 23–70% [1]. Laparoscopic adjustable gastric banding (LAGB) has a failure rate in the US approaching 40–50% [6]. This may be evident only if follow-up is stringently high and longer than five years, preferably 10 years. Multiple articles have debated the merits and disadvantages of each procedure as the initial form of surgical intervention for the management of morbid obesity, but few touch upon the issues regarding the best choice for revisional surgery after failed bariatric surgery with inadequate weight loss or significant weight regain [3, 4].

The main indications for revisional surgery are multiple, and specific to the type of operation performed, but inadequate weight loss is common to most bariatric procedures. Problems specific to placement of AGB/LAGB include hardware problems with the band itself, motility problems, and psychological intolerance to the band [5, 811, 14, 15, 17, 2022, 2427, 29, 30].

This review of conversions of gastric banding to RYGB, sleeve, or DS attempts to determine which revisional procedure best enhances weight loss.

Revisional Options after Adjustable Gastric Banding (AGB/LAGB)

Indications for revision after AGB or LAGB are numerous and include problems with the hardware of the band itself, such as band slippage, tubing leakage or breakage, motility problems caused by the band such as proximal pouch and esophageal dilatation, and dysmotility along with problems seen in other bariatric procedures such as inadequate weight loss and wound infection (Table 1) [11, 14, 17, 2022, 24]. Inadequate weight loss is considered <25% EWL as defined by the Reinhold criteria [1]. Band complications are reported in 15–58% of patients [14]. Band removal is the most commonly performed revisional operation for AGB, however, because of the dramatic recurrence of morbid obesity, other options have been explored [14]. The most frequent revisional strategies are currently re-banding or band removal and revision to RYGB either with an open or laparoscopic technique (Table 2) [11, 14, 17, 2022, 24, 27].

Table 1 Indications for revision/conversion after AGB/LAGB
Table 2 Strategies for conversions in AGB/LAGB

In a study of four cohorts, with a total of 193 patients that underwent re-banding, after their first banding, patients’ BMI was 32–38 kg/m2 compared to an initial BMI of 45–48 kg/m2 [11, 14, 20, 22]. Only one published study presents a reasonable follow-up, and BMI was unchanged after re-banding and remained at 37 kg/m2 after 12 months compared to 38 kg/m2 originally [14]. Complication rates were 0–19% in this patient population. This latter study also looked at laparoscopic RYGB (LRYGB) as a revisional therapy in an additional 32 patients. BMI was found to decrease to 32 kg/m2 after a follow-up period of 12 months with a similar complication rate to the rebanding group at 19% [14]. The authors concluded that RYGB was a superior revisional operation as compared to AGB. They recommended that AGB be placed only in patients with adequate weight loss after their initial band placement and with no evidence of esophageal dysmotility. Roux-en-Y gastric bypass was recommended in those with evidence of esophageal dysmotility [14].

Studies including a total of 214 patients that had undergone RYGB or LRYGB after a failed placement of a band were found in the literature [14, 17, 21, 24]. BMI prior to their revisional procedures was 38–45.8 kg/m2 and dropped to 26–38 kg/m2 after a follow-up of 12–18 months. Complication rates were 0–19%, again similar to that reported for the rebanded patients. In one study comparing the revision of LAGB to RYGB and LRYGB, postoperative BMI was noted to decrease similarly to 38 kg/m2 from 45.8 kg/m2 in both groups, but operative time was significantly higher in the laparoscopic group, 120 versus 194 minutes, respectively [24]. This observation was noted in other studies [17]. Although the complication rate was higher in the laparoscopic group (10 vs. 23%), this was not statistically significant and the estimated blood loss was also similar in both groups [24].

Revision to RYGB after failed AGB from the four largest series show a complication rate of 7% (Table 3). Complications include wound infection (2%); intestinal hemorrhage, fever, and small-bowel obstruction (SBO) requiring surgery (1%). Staple-line leaks, intra-abdominal abscesses, and pneumonia occurred in less than 1% [14, 17, 21, 24]. Rutledge has also published on converting gastric band to mini-gastric bypass. Even though this operation appears technically simpler, it is not popular in USA as it causes symptomatic bile reflux, which requires conversion to a RYGB [26].

Table 3 Complications after LAGB conversions to different bariatric surgical strategies
Table 4 Cohorts of patients with band revisions/conversions

Discussion

Revisional bariatric surgery is becoming more common due to the rapid increase of patients undergoing surgery as treatment for morbid obesity. Unfortunately, there is a paucity of significant data to help the surgeon decide which revisional procedure to choose based on the patient’s previous bariatric procedure. In general, patients that require revision for several broad categories of reasons: inadequate weight loss or weight regain, motility problems, problems of acid secretion, and in the case of bands, problems with the hardware itself.

We found that the best way to approach these patients is to first perform a full history and physical examination and assess their BMI. We then divided patients into those that have inadequate weight loss or significant weight regain from those with or without gastrointestinal motility problems. We then reviewed the operative notes from their previous surgery, looking for the type of band inserted, size, sutures used and location, specific passage of the band itself (retrogastric below or above the peritoneal reflection), medial or lateral to vagus nerves, and location of the tubing/port. All patients with a history suggestive of maladaptive eating disorders because of their bariatric surgery underwent further psychological evaluation and were treated prior to consideration for surgical revision.

If patients had symptoms of gastroesophageal reflux disease (GERD) they would also be treated with an esophagogastroduodenoscopy (EGD) to look for evidence of marginal ulceration and its sequelae. All marginal ulcers were treated with up to six months of proton-pump inhibition; if their issues did not resolve after this point they were offered revisional surgery. The next step was to document their anatomy with a barium swallow to look for evidence of slippage, stenosis, hiatal hernia, erosion, reflux, and whether a gastric pouch was present above the band.

Band revision is a good option in patients who have hardware problems from their bands, particularly port-site complications [20, 27]. Nonetheless, in the setting of inadequate weight loss after band placement, RYGB appears to be a superior procedure [8, 14, 17, 19, 21, 24, 25]. Of four studies analyzing the results after rebanding only one reported postoperative BMI, and it was unchanged after 12 months [11, 14, 20, 22]. Because of the lack of data showing adequate weight loss after rebanding we do not perform this procedure. RYGB on the other hand appears to have adequate EWL, decreasing to 32 kg/m2 from 45 kg/m2 at 12 months in one study [17]. Furthermore, complication rates are similar to rebanding rates [11, 14, 17, 2022, 24]. When revisional surgery is performed by experienced hands, creation of an LRYGB after failed AGB can have similar complication rates as primary LRYGB surgery [8, 14, 17, 19, 21, 24, 25]. The recent series by Spivak, in which 33 patients underwent a conversion of band to gastric bypass, confirmed the results of previous series without mortality, with two severe complications (one splenectomy and one internal hernia), which can be considered a technical error [29]. Similarly Calmes reported on 32 patients without conversion or mortality and 20% morbidity. Up to 75% of patients ended with a BMI less than 35 kg/m2 (good) [19].

DS is another interesting option in this patient population, but larger series are still needed [7, 10, 12, 16, 18, 30]. Because of the superior weight loss seen with the DS when compared to other bariatric procedures, interest has grown in using this procedure in the treatment of morbidly obese patients who fail in other surgical therapy [16, 18, 23]. The DS was first used as a revisional therapy in 2001 to treat a patient who had undergone VBG in the past, and was still suffering from diabetes mellitus and obesity. At one-year follow-up, after open DS, the patient’s diabetes had resolved [7]. The use of DS as a revisional therapy has become even more widespread since it was carried out laparoscopically [2]. Our group subsequently reported two laparoscopic duodenal switches in patients that had undergone LAGB in the past [10]. Gagner has also used DS laparoscopically to convert gastric bypass successfully [18]. Also, DS alone with the band in place can be used in selected patients, but ultimately, the band may fail and be removed [12]. In the largest series to date, 46 patients who had undergone other bariatric surgeries in the past underwent open DS. No mortalities were reported and 69% EWL was noted after a mean follow-up of 30 months [16]. Laparoscopic BPD can be done but appears to be an inferior operation when compared to DS, due to higher rates of malnutrition [15].

Another option is to remove the band and proceed with a sleeve gastrectomy without the DS. The re-operative sleeve has already been described as a re-operation/revision of DS itself, but when the concept has been used for the failed gastric band [13]. Bernante has described a small cohort of eight patients who underwent a laparoscopic sleeve gastrectomy I, in three for esophageal dilatation and in five for unsuccessful weight loss. Although the operation only took an average of 90 minutes, without conversions and complications, the follow-up was too short to conclude on the final results [28]. It is believed that, for patients in the range of 40 kg/m2, a drop of 10 BMI points might be expected, as this has been seen in revised DS patients [13].

Conclusion

Although the data is still lacking, revisional surgery should be done laparoscopically in experienced hands only.

When behavioral or anatomic issues are not present, revisional surgery should be approached with the goal of treating malnutrition or enhancing percentage EWL. Revision of bands to another band should only be undertaken in the setting of port and hardware complications. In the superobese, we may leave the AGB in place or more frequently convert to DS with band removal. In patients with failed RYGB we convert our patients to DS, but, not discussed in this paper, placement of an AGB may be an acceptable option in the near future.

To date, no prospective trial has been done to adequately determine which revisional bariatric procedure should be carried out in the setting of inadequate weight loss or excessive weight regain in the refractory bariatric patient.