Marc Bessler, M.D.*, Amna Daud, M.D., M.P.H., Teresa Kim, M.D., Mary DiGiorgi, M.P.H.

Transcription

1 Surgery for Obesity and Related Diseases 3 (2007) Original article Prospective randomized trial of banded versus nonbanded gastric bypass for the super obese: early results Marc Bessler, M.D.*, Amna Daud, M.D., M.P.H., Teresa Kim, M.D., Mary DiGiorgi, M.P.H. Center for Obesity Surgery, Columbia University, New York-Presbyterian Hospital, New York, New York Received May 30, 2006; revised January 17, 2007; accepted January 21, 2007 Abstract Keywords: Background: Banded gastric bypass has been reported to result in superior weight loss compared with standard nonbanded gastric bypass. However, an adequate comparison of these procedures has not yet been reported. Methods: A total of 90 patients were enrolled in this prospective randomized double-blind trial comparing banded and nonbanded open gastric bypass for the treatment of super obesity. The banding technique involved placement of a cm polypropylene band around the proximal gastric pouch of a standard gastric bypass procedure using the technique of Capella. Chi-square testing and analysis of variance were performed to find any differences in patient characteristics (gender, age, and initial body mass index), percentage of excess weight lost at 6, 12, 24, and 36 months postoperatively, improvement or resolution of co-morbidities, and complications in the banded versus nonbanded gastric bypass groups. Results: As expected, no differences were present in the patient characteristics or incidence of co-morbidities between the banded (n 46) and nonbanded (n 44) groups. The body mass index, percentage of women, and mean age was 59.5 and 56.5 kg/m 2, 64% and 73.8% (P.09), and and years for the banded and nonbanded groups, respectively; all differences were nonsignificant. No significant differences were found in the resolution of co-morbidities. No significant difference was present in the percentage of excess weight loss at 6, 12, and 24 months (43.1% versus 24.7%, 64.0% versus 57.4%, and 64.2% versus 57.2%, respectively) postoperatively; however, the banded patients had achieved a significantly greater percentage of excess weight loss at 36 months (73.4% versus 57.7%; P.05). The incidence of intolerance to meat and bread was greater in the banded patients. The overall number of complications was 12 (26%) in the banded and 13 (29.5%) in the nonbanded group, a nonsignficant difference. No band erosions had occurred at the last follow-up visit, and no patients in either group died. Conclusion: These results suggest that although the initial weight loss was not significantly different between the 2 groups, the banded patients continued to lose weight for 3 years. The polypropylene band appeared to be well tolerated. We plan longer follow-up to confirm the possibility of additional weight loss and the prevention of weight regain in the banded group, as well as to document any long-term band complications. (Surg Obes Relat Dis 2007;3: ) 2007 American Society for Bariatric Surgery. All rights reserved. Super obesity; Morbid obesity; Gastric bypass; Banded gastric bypass Presented at the 2006 Plenary Session of American Society for Bariatric Surgery, San Francisco, California, June July 2006 *Reprint requests: Marc Bessler, M.D., Center for Obesity Surgery, New York-Presbyterian Hospital, Columbia University Medical Center, 161 Fort Washington Avenue, 6th Floor, Room 620, New York, NY mb28@columbia.edu Surgical treatment for morbid obesity has gained increasing popularity during the past years and is considered the most effective treatment, resulting in sustained weight loss and resolution of obesity-related co-morbidities [1]. The Roux-en-Y gastric bypass procedure has become the standard bariatric procedure performed in the United States. This procedure has replaced vertical banded gastroplasty, /07/$ see front matter 2007 American Society for Bariatric Surgery. All rights reserved. doi: /j.soard

2 M. Bessler et al. / Surgery for Obesity and Related Diseases 3 (2007) which resulted in a high failure rate for weight loss and a high reoperative rate [2]. However, even after early success with Roux-en-Y gastric bypass, a degree of recidivism exists. Patients may regain significant weight 3 5 years after surgery [3]. The preoperative predictors of weight loss outcomes have not been well delineated. However, several reports have documented poorer weight loss after gastric bypass in the super-obese (body mass index [BMI] 50 kg/m 2 ) population [4]. Mason et al [5] introduced the concept of super obesity in 1987 and observed inadequate weight loss in this particular group of patients after vertical banded gastroplasty. Since then, other studies have also reported a lower percentage of excess weight loss (%EWL) and fewer patients reaching a 50% EWL in the super-obese population compared with the less obese population after a gastric bypass procedure [6,7]. Studies on distal bypass and biliopancreatic diversion have shown that patients experience significant metabolic sequelae [8]. In an effort to improve weight loss outcomes, several surgeons have been performing banding of the gastric pouch during gastric bypass. The gastric bypass with its small gastric pouch and narrow stoma is initially restrictive but, with time, patients are able to eat larger volumes of solid food because of dilation of the stoma. Theoretically, banding of the pouch causes a fixed narrowing and maintains the restriction for a longer period. Continued restriction is believed to result in better weight loss and less weight regain over time. Capella and Capella [9] and Fobi [10] have reported on large series of banded bypass patients and have claimed that these patients lost more weight than patients with a nonbanded bypass. However, no studies have evaluated these claims. Furthermore, the presence of a foreign object around the stomach has been associated with complications such as stenosis, erosion, and infection, leading many to oppose placement of a band without evidence of benefit. Vertical banded gastroplasty with placement of a 1.5-cm wide polypropylene mesh band around a vertical lesser curve pouch was associated with a 2.8% incidence of band erosion [11]. The silastic ring vertical gastroplasty was similarly associated with a 1 2% incidence of erosion [12,13]. Because of the poor relative weight loss in the super obese with conventional long-limb gastric bypass and the possibility of improving this with banding of the gastric pouch, we undertook this prospective, randomized, doubleblind study to compare the outcomes of super-obese patients who underwent long-limb gastric bypass by way of laparotomy and were randomly assigned to undergo banded or nonbanded long-limb gastric bypass. Methods The Columbia University institutional review board approved the enrollment of human subjects. A total of 278 patients with a body mass index 50 kg/m 2 underwent surgery by one surgeon from June 2001 and July All these patients were offered adjustable gastric banding, longlimb gastric bypass and enrollment into the study with possibility of randomization to banded or nonbanded longlimb gastric bypass. Patients who had undergone previous gastric surgery were excluded from the study, as were minors. A total of 90 patients were enrolled in this study and are the subject of this report. All patients provided informed study consent before surgery. All patients had a BMI of 50 kg/m 2, were counseled extensively, and had undergone a psychological evaluation. Pulmonary, cardiac, and other specialty evaluations were obtained as clinically indicated. Randomization was performed after induction of anesthesia by opening 1 of 100 sequentially numbered sealed envelopes containing the words banded or nonbanded (50 each in random order) on a folded piece of paper. Postoperatively, the patients were treated identically and underwent contrast swallow studies on postoperative day 1 before diet advancement. Clear liquids were allowed if no radiographic or clinical signs of a leak were found. If liquids were tolerated, the diet was advanced to purees on postoperative day 2, and discharge was on postoperative day 2 or 3, unless clinically contraindicated. Data on food intolerance to bread, meat, and pasta; gastrointestinal (GI) symptoms, including emesis, dumping, diarrhea, constipation, flatulence, and abdominal pain; resolution of co-morbidities (i.e., diabetes mellitus, hypertension, hyperlipidemia, osteoarthritis, gastroesophageal reflux disease, and stress incontinence), and weight loss were prospectively collected at 6, 12, 24, and 36 months after surgery. We scored the postoperative GI symptoms according to a subjective scale of 0 4 (0, none; 1, mild; 2, moderate; and 3, severe). Upper GI series were requested at 6, 12, and 24 months postoperatively. Statistical analysis Chi-square tests and analysis of variance were performed to find any differences in patient characteristics (gender, age, and initial BMI); percentage of excess weight (%EWL) lost at 6, 12, 24, and 36 months postoperatively; improvement or resolution of co-morbidities; and the incidence of complications in the banded versus nonbanded groups. P.05 was considered significant. Surgical technique Preoperative prophylaxis, including antibiotics and lowdose heparin, was administered, and pneumatic compression devices were placed. Under general endotracheal anesthesia, access to the abdomen was obtained by way of an upper midline abdominal incision. The biliopancreatic limb was 75 cm distal to the ligament of Treitz. The enterolenterostomy was fashioned with a 150-cm Roux limb. The Roux limb was mobilized in a retrocolic and retrogastric position. A cm 3 divided gastric pouch was created

3 482 M. Bessler et al. / Surgery for Obesity and Related Diseases 3 (2007) Table 1 Comparison of study and nonstudy patients Variable Study (n 90) Nonstudy (n 188) P value Initial BMI (kg/m 2 ) Age (y) NS Women (%) BMI body mass index; NS not significant. starting 4 cm distal to the gastroesophageal junction. An anastomosis was then fashioned between the proximal gastric pouch and the Roux limb with a 2-layer, hand-sewn, running suture over a 32F intragastric tube. For the patients randomized to the banded group, a 1.5-cm by 7.0-cm Marlex band was placed around the proximal gastric pouch and was sutured to create a 5.5-cm circumference ring. The band was placed at least cm above the point of the anastomosis. Anastomotic integrity was confirmed by air insufflation under saline. All mesenteric defects were sutured closed. Results Patient characteristics From June 2001 through July 2005, 278 patients with a BMI 50 kg/m 2 underwent surgery for morbid obesity by a single surgeon. Of the 278 patients, 90 were enrolled in this study. The demographics of the 90 patients enrolled in the study were compared with those of the 188 nonstudy patients, who also had a BMI 50 kg/m 2, to assess for selection bias. The study group included more male patients than did the nonstudy group (Table 1). Also, the initial BMI of the study patients was significantly greater than that of the nonstudy population. Patient age was not significantly different in the 2 groups (Table 1). A comparison between the 2 groups of study patients (banded versus nonbanded) demonstrated no significant differences in age, gender, or initial BMI (Table 2). No significant differences were present in the prevalence of co-morbidities, with the exception of arthritis; 91% of the nonbanded group had arthritis compared with 72% of the banded group (Table 3). Weight loss and resolution of co-morbidities The %EWL in the nonbanded patients was 25%, 57.4%, 57.2%, and 57.7% at 6, 12, 24, and 36 months, respectively. Table 3 Preoperative co-morbidities Co-morbidity Banded (%) Nonbanded (%) P value Hypertension NS Diabetes mellitus NS Hyperlipidemia NS Arthritis GERD NS Stress urinary incontinence NS GERD gastroesophageal reflux disease; NS not significant. In contrast, the corresponding %EWL in the banded patients was 43.1%, 64%, 64.2%, and 73.4% for the same periods (Fig. 1). These data show that the banded patients lost more weight than did the nonbanded patients. Although the %EWL was not significantly different at 6, 12, and 24 months, at 36 months, the banded patients had lost significantly more weight than had the nonbanded patients (P.05). However, this was calculated from the small number of patients reaching the 36-month follow-up period. The percentage of patients achieving a BMI of 35 kg/m 2 is reported in Table 4. Among the banded group, 48% and 53% of patients had achieved a BMI of 35 kg/m 2 at 12 and 24 months, respectively. In contrast, 41% and 38% of patients in the nonbanded group had achieved a BMI of 35 kg/m 2 at 12 and 24 months, respectively. The study patients were also compared with all patients who had undergone surgery during the same period who had had an initial BMI of 50 kg/m 2 or a BMI of 50 kg/m 2 who had not participated in the study (Table 4). Of the patients with an initial BMI of 50 kg/m 2, 92% had achieved a BMI of 35 kg/m 2 at 12 months of follow-up. Also, the nonstudy group with an initial BMI 50 kg/m 2 who had undergone standard gastric bypass had a percentage of patients who had attained a BMI of 35 kg/m 2 similar to that of the nonbanded study patients (38% versus 44% at 24 months). Table 2 Comparison of banded and nonbanded patients Variable Banded (n 46) Nonbanded (n 44) P value Initial BMI (kg/m 2 ) NS Age (y) NS Women (%) NS (.09) Abbreviations as in Table 1. Fig. 1. Comparison of %EWL between banded and nonbanded patients during 36 months of follow-up.

4 M. Bessler et al. / Surgery for Obesity and Related Diseases 3 (2007) Table 4 Percentage of patients attaining BMI 35 kg/m 2 Follow-up time (mo) In contrast, at 24 months, 53% of the banded patients had attained a BMI of 35 kg/m 2. Both the banded and the nonbanded patients had significant resolution of co-morbidities. No significant differences were found in the resolution of the listed co-morbidities (Table 5). Of those in the nonbanded group with arthritis, 91% had resolution compared with 76% of those in the banded group. Postoperative GI symptoms The data collected on GI symptoms, including dumping, emesis, diarrhea, constipation, flatulence, and abdominal pain, demonstrated that the banded patients experienced a significantly greater rate of emesis. No significant differences were present for the other GI symptoms (Table 6). Food intolerance to meat, bread, rice, and pasta was also recorded at each follow-up visit. The banded patients experienced significantly greater food intolerance than did the nonbanded patients (79% versus 33%, P.05). Complications No patient in either study group died. No significant differences were found in the postoperative complications between the banded and nonbanded groups (Table 7). Discussion Banded Nonbanded ibmi 50 kg/m BMI body mass index; ibmi initial BMI. ibmi 50 kg/m 2 The treatment of the super-obese patient is challenging because of inadequate weight loss after standard bariatric surgery. The addition of a band to the standard gastric bypass procedure has been reported to improve the %EWL. However, arguments have been made against this type of procedure. The presence of a foreign body could potentially increase the rate of infection. The band erosion rate has also Table 5 Resolution of co-morbidities Co-morbidity Banded (%) Nonbanded (%) P value Hypertension NS Diabetes mellitus NS Hyperlipidemia NS Arthritis NS GERD NS Stress urinary incontinence NS GERD gastroesophageal reflux disease; NS not significant. Table 6 Average scores of gastrointestinal symptoms Symptom Banded Nonbanded P value Dumping NS (.06) Emesis Diarrhea NS Constipation NS Flatulence NS Abdominal pain NS NS nonsignificant. been reported to be 1 6% [11]. Other potential complications include slippage and stenosis. The choice of band material in this study was Marlex. Silastic bands have also been used in other restrictive procedures. The silastic ring has been shown to have an erosion rate of 1 2% [12,13]. This report is the first to prospectively study banded versus nonbanded gastric bypass. We found a greater %EWL for the banded cohort than for the nonbanded gastric bypass cohort, with a statistically significant difference at 36 months. We had no incidences of band erosion or slippage. Though the rate of wound infection was greater in the banded group, the difference was not statistically significant. The banded group reported more postoperative emesis and food intolerance; however, all were treated conservatively and did not require invasive intervention. Although Fobi [10] and Capella and Capella [9] have previously reported on the banded gastric bypass procedure, the strength of the present study was that it was a prospective, randomized study. Our study was also performed by a single surgeon, minimizing the variability in surgical technique. We have reported not only the %EWL, but also the data on patients tolerance to the procedure. In addition, our total rate of follow-up was close to 90%. The weaknesses of this study were that it included a limited number of patients, reducing the power of our statistical analysis. Also, the follow-up at 36 months included a limited number of patients. Further follow-up of a larger cohort is being conducted and should address these problems. Table 7 Complications Complication Banded (%) Nonbanded (%) Wound infection 7 (15.2) 5 (11) Anastomotic leak 0 2 (4.8) Pneumonia 1 (2.1) 1 (2.2) Pulmonary embolism 0 0 Small bowel obstruction 1 (2.1) 1 (2.2) Band erosion/slippage/removal 0 NA Other 3 (6.5) 4 (9.0) Mortality 0 0 Total 12 (26) 13 (29.5) Data in parentheses are percentages.

5 484 M. Fobi / Surgery for Obesity and Related Diseases 3 (2007) Conclusion The results of the present study have shown that banded, long-limb gastric bypass is feasible, with only a small increase in minor complications and no difference in the incidence of major complications compared with standard gastric bypass. The patients were able to tolerate the procedure. Banded gastric bypass provides another therapeutic option for the super-obese population and may result in greater weight loss for a longer period. This procedure is also possible using the laparoscopic approach and is currently being performed in a larger cohort of patients. Additional follow-up is needed to assess the long-term outcomes of banded versus nonbanded gastric bypass procedures. Disclosures M. Bessler is a consultant to, and receives research support from, Ethicon Endo-Surgery; is a consultant to, and receives marketing support from, Inamed (LapBand); and is a consultant to USGI, Intrapace, Metacure, and Bard/ Davol. References [1] National Institutes of Health. Gastrointestinal surgery for severe obesity: National Institutes of Health Consensus Development Conference Statement. Am J Clin Nutr 1992;55:615S 19S. [2] Ortega J, Sala C, Flor B, et al. Vertical banded gastroplasty converted to Roux-en-Y gastric bypass: little impact on nutritional status after 5-year follow-up. Obes Surg 2004;14: [3] Schwartz RW, Strodel WE, Simpson WS, et al. Gastric bypass revision: lessons learned from 920 cases. Surgery 1998;104: [4] Brolin RE. Bariatric surgery and long term control of morbid obesity. JAMA 2002;288: [5] Mason E, Doherty E, Maher C, et al. Super obesity and gastric reduction procedures. Gastrointest Clin North Am 1987;16: [6] Brolin RE, Kenler HA, Gorman B, et al. Long limb gastric bypass in the super obese: a prospective randomized study. Ann Surg 1992; 215: [7] Choban PS, Flancbaum L. The effect of Roux limb lengths on outcome after Roux-en-Y gastric bypass: a prospective, randomized clinical trial. Obes Surg 2002;12: [8] Sugerman HJ, Kellum JM, DeMaria EJ. Conversion of proximal to distal gastric bypass for failed gastric bypass for super obesity. J Gastrointest Surg 1997;1: [9] Capella JE, Capella RF. An assessment of vertical banded gastroplasty-roux-en-y gastric bypass for the treatment of morbid obesity. Am J Surg 2002;183: [10] Fobi MA. Vertical banded gastroplasty vs. gastric bypass: 10 years follow up. Obes Surg 1993;3: [11] Moreno P, Alastrue A, Rull M, et al. Band erosion in patients who have undergone vertical banded gastroplasty: incidence and technical solutions. Arch Surg 1998;133: [12] Fobi ML, Igwe H, Delahy D, et al. Band erosion: incidence, etiology, management and outcome after banded vertical gastric bypass. Obes Surg 2001;11: [13] Gavert N, Szold A, Abu-Abeid S. Safety and feasibility of revisional laparoscopic surgery for morbid obesity: conversion of open silastic vertical banded gastroplasty to laparoscopic adjustable gastric banding. Surg Endosc 2004;18: Editorial comment The birth of, and most of the advances in, bariatric surgery have came from the observations of astute surgeons in their daily practice of surgery. The jejunoileal (JI) bypass evolved from the observation that the short gut syndrome resulted from excising long segments of the small bowel because of vascular injury, cancer, or trauma [1]. Theshortenedgutresultedinweightlossand then weight maintenance in many of the affected patients. Similarly, the gastric bypass operation developed from the observation that the Billroth-2 operation also caused weight loss and then weight loss maintenance in patients with cancer or ulcer disease [2]. Itwasonlyyearslater that prospective and randomized studies validated the information reported in large series of these operations. The JI bypass and gastric bypass procedures have undergone several modifications over time. In this issue of the Journal, Bessler et al. have presented a prospective, randomized study comparing nonbanded gastric bypass and banded gastric bypass. Banded gastric bypass is an operation that, similar to the JI and gastric bypass procedures, evolved from the observation that converting a vertical banded gastroplasty or silastic ring gastroplasty to gastric bypass, with the gastroenterostomy distal to the band or ring, resulted in more weight loss and better weight loss maintenance than observed after nonbanded gastric bypass [3]. As with the JI and gastric bypass, the surgeons who made these observations have performed thousands of banded gastric bypass operations and published series that have confirmed the observations of greater weight loss in more patients and maintained longer than reported with nonbanded gastric bypass [4 6]. The results of the study by Bessler et al. are thus a confirmation and an objective documentation of the effects of banding the pouch in the gastric bypass operation. These results duplicate the findings of previously published large series. A couple of presentations and abstracts of prospective studies comparing banded and nonbanded gastric bypass with a similar outcome have been presented [7,8]. However,thestudybyBessleretal. is the only full report that has been submitted, reviewed, and published. The problems with emesis and food tolerance in this study are to be expected, because these were also documented in the initial use of the banded gastric bypass pro-