Campos GM, Rabl C, Roll GR, Peeva S, Prado K, Smith J, Vittinghoff E. Better Weight Loss, Resolution of Diabetes, and Quality of Life for Laparoscopic Gastric Bypass vs BandingResults of a 2-Cohort Pair-Matched Study. Arch Surg. 2011;146(2):149-155. doi:10.1001/archsurg.2010.316
Laparoscopic Roux-en-Y gastric bypass (RYGB) and laparoscopic gastric banding (LB) are the 2 most common operations used to treat morbid obesity, but few controlled comparative studies have reported perioperative and long-term outcomes.
Two-cohort pair-matched study.
Academic tertiary referral center.
One hundred consecutive morbidly obese patients treated with LB were pair-matched by sex, race, age, initial body mass index, and presence of type 2 diabetes mellitus with 100 patients who were treated with RYGB.
Main Outcome Measures
Perioperative and postoperative complications, reoperations, and 1-year outcomes, including weight loss, type 2 diabetes resolution, and quality of life.
The RYGB and LB groups had similar characteristics. One-year outcomes were available for 93 patients in the LB group and 92 in the RYGB group. The overall rate of complications was similar in both groups (11 patients in the LB group [12%] vs 14 in the RYGB group [15%]; P = .83), with a higher rate of early complications (≤30 days) after RYGB (11 patients [11%] vs 2 [2%] for LB; P = .01) and a higher rate of reoperations after LB (12 patients [13%] vs 2 for RYGB [2%]; P = .009). No deaths occurred. Excess weight loss (36% vs 64%; P < .01), resolution of diabetes (17 patients [50%] vs 26 [76%]; P = .04), and quality-of-life measures were better in the RYGB group.
When performed in high-volume centers by expert surgeons, RYGB has a similar rate of overall complications and lower rate of reoperations than LB. With the benefit of greater weight loss, increased resolution of diabetes, and improved quality of life, RYGB, in these circumstances, has a better risk-benefit profile than LB.
The most common surgical procedure for treating obesity in the United States is the laparoscopic Roux-en-Y gastric bypass (RYGB), with an estimated 180 000 operations performed each year.1- 3 Roux-en-Y gastric bypass results in loss of about 60% of excess weight in the first year after surgery, and most patients maintain long-term weight loss.1- 3 Of obese patients with type 2 diabetes mellitus (T2DM), 86% experience significant improvement in glucose control after RYGB and 78% have long-term remission of T2DM.4 However, RYGB has a steep learning curve5- 11 and even in high-volume centers is associated with a 10% to 18% morbidity rate and a 0.1% to 0.3% mortality rate.3,12,13
The laparoscopic adjustable gastric band (LB), approved by the US Food and Drug Administration in 2001 as an alternative to RYGB for treatment of morbid obesity in the United States, has been touted as a less invasive and safer surgical therapy than RYGB.14- 16 The number of LBs implanted has grown exponentially in the past few years in the United States, with an estimated increase of 329% from 2004 to 2007. The LB is currently estimated to be used in about 23% of bariatric surgeries in specialized academic US medical centers.17
Although some centers have reported good results with LB,15,16,18- 20 long-term results seem less impressive than those for RYGB in most series.3- 8,21,22 However, the utility of these findings is limited by methodological shortcomings, notably significant differences in baseline patient characteristics and long-term follow-up of only 50% to 70% of the sample. To our knowledge, only 2 prospective randomized trials have compared LB with RYGB, one of them small23 and the other with baseline imbalances.24 Two other studies retrospectively evaluated prospectively collected data using a pair-matched design. One had only a 76% follow-up rate,21 but, in the other, 94% of 206 patients completed long-term follow-up and those who had RYGB had fewer complications and better weight loss and resolution of comorbidities.22 To add to the information available and help resolve the debate over the relative risks and benefits of these competing bariatric procedures, we conducted a pair-matched study of complications and outcomes 1 year after surgery among morbidly obese patients undergoing RYGB and LB at a large academic medical center.
We identified morbidly obese patients who underwent primary RYGB or LB from January 1, 2004, through January 31, 2008, in the Bariatric Surgery Program at the University of California, San Francisco (UCSF), Medical Center using a clinical database prospectively maintained by a trained research coordinator (S.P.). The database includes demographic information, preoperative clinical measurements, and perioperative and long-term outcomes of all patients who underwent bariatric surgery in the program. All patients filled out a structured questionnaire (at the clinic visit or by telephone interview). In addition, we examined all patients' electronic medical records, operative reports, anesthesia records, discharge summaries, and follow-up clinic notes to search for possible missing outcomes. The study was approved by the UCSF institutional review board, and informed consent was obtained from all patients.
Patients with previous bariatric operations were excluded from consideration. All patients met the following 1991 National Institutes of Health Consensus Development Conference25 and the UCSF bariatric surgery program criteria for bariatric operation: body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared) of greater than 40 or ranging from 35 to 40 with high-risk obesity-related comorbid conditions for at least 5 years, a documented attempt of medically supervised therapy for weight loss for at least 6 months, and participation in a 1-hour educational and screening session with a dietitian and a bariatric surgeon. All patients also underwent preoperative psychological, nutritional, and comprehensive medical evaluations. Procedure selection was determined by patient preference and occurred after an educational session with a bariatric surgeon and dietitian and medical and psychological clearances. We were able to pair-match 100 consecutive patients treated with LB to 100 patients treated with RYGB on the basis of sex, age (±5 years), race (black or other), BMI (±3), and diagnosis of T2DM, defined as a clinical diagnosis of adult-onset diabetes mellitus in patients who were receiving oral hypoglycemic agents with or without added insulin replacement therapy.
We used a commercially available LB system (Lap Band; Allergan, Inc, Irvine, California). A 10-cm (n = 34), 11-cm (n = 46), or 12-cm or larger (Vanguard; Allergan, Inc) (n = 20) band system was used at the discretion of the attending surgeon. The band was inserted using the pars flaccida technique according to the manufacturer's guidelines. In brief, the operation consisted of adequately exposing the His angle by removal of the covering fat pad while retracting the gastric fundus inferiorly. Dissection was continued until the left crus of the diaphragm was completely exposed. A small incision in the avascular aspect of the gastrohepatic ligament was created, and care was taken to identify and preserve the hepatic branch of the vagus nerve. Blunt dissection was used to create a space between the base of the right crus and its overlying peritoneum. A long grasper was then gently passed above the right crus, underneath the gastroesophageal junction, toward the His angle. The LB was passed around the gastroesophageal junction, and a 30-mL balloon introduced through the patient's mouth helped determine the proper band position. The balloon was deflated and removed from the stomach. The band was snapped in place and secured with 4 or 5 gastrogastric stitches, with the initial throw of each gastric stitch in a position as far as needed in the outer aspect of the anterior gastric fundus to prevent a tight tunnel. The band reservoir was not filled until the sixth postoperative week.
The technique for RYGB has also been described previously.26 In brief, it was performed laparoscopically with 6 or 7 ports. A 3.5-mm linear stapler (United States Surgical Corporation, Norwalk, Connecticut) transected the stomach to create a 15- to 30-mL gastric pouch. An antecolic gastrojejunostomy route was always used. A circular anastomosis with a 25-mm stapler (United States Surgical Corporation) was preferred. A Roux limb of 100 cm (n = 92) or 150 cm (n = 8) was measured, and a completely stapled side-to-side jejunojejunostomy was created. The mesenteric defect at the jejunojejunostomy was routinely closed.
Operating room times collected and available for analyses include the combined duration of anesthesia induction, surgical procedure, and anesthesia recovery.
Patients were given clear liquids on the first postoperative day and then advanced to a full-liquid diet on the second postoperative day. Patients undergoing RYGB were seen in the clinic 2 weeks postoperatively, then every 3 months for the first year, and then annually thereafter. Patients undergoing LB were seen 2 and 6 weeks postoperatively, then every 3 months for the first year, and annually thereafter. Adjustments to LB were performed according to manufacturer guidelines. More frequent clinic visits were arranged if needed. After the operation, patients received prenatal vitamins, calcium supplements, cholecalciferol (vitamin D), and acid suppression medication. Nutritional supplements were prescribed as needed for documented deficiencies. All weight measurements were obtained at the UCSF Bariatric Surgery Center.
At each postoperative visit, all patients were asked to complete the Moorehead-Ardelt Quality of Life Questionnaire II.27 This questionnaire is specifically designed to measure subjective quality of life in obese subjects in the following 5 key areas: self-esteem, physical well-being, social relationships, work, and sexuality. The results of this questionnaire were combined with scores for weight loss and improvement of medical conditions in the Bariatric Analysis and Reporting Outcome System (BAROS). This is a standardized tool designed specifically for assessing psychosocial outcomes after bariatric surgery.27
Six outcomes were analyzed. Perioperative (≤30 days) and long-term (>30 days) complications included use of unexpected drug therapy or imaging, total parenteral nutrition, a bedside procedure, blood transfusion, a hospital stay longer than twice the median stay, diagnostic or therapeutic endoscopy, reoperation (with or without organ resection or anastomotic revision), or death. We also recorded the number of therapeutic interventions and reoperations and the frequency of postoperative clinic visits and LB adjustments.
One year after surgery, we analyzed weight loss, percentage of excess weight loss (EWL), current BMI, and change in BMI from baseline. Improvement and resolution of T2DM were combined. Resolution was defined as glycemic control without any medications, and improvement was defined as a decrease in more than 50% of preoperative antidiabetic medication doses. We also analyzed the Moorehead-Ardelt Quality of Life Questionnaire II overall and subscale scores.
To account for the pair-wise matching of patients in the LB and RYGB groups, we used the Wilcoxon signed rank test to compare differences in continuous outcomes and the McNemar test to compare binary outcomes. To compare resolution of diabetes between the groups, we used normal-logistic models with a random effect for matched pair. To evaluate mediation of the effect of procedure type, we added EWL to the initial model. In the full model, the adjusted odds ratio represents an estimate of the direct effect of procedure type, if any, that is not explained by within-pair differences in EWL. Analyses were performed using Stata statistical software, version 10 (StataCorp, College Station, Texas).
Of the 741 morbidly obese patients who underwent a primary LB or RYGB from January 1, 2004, through January 31, 2008, 148 (20%) underwent LB, 557 (75%) underwent RYGB (of whom 10 [2%] underwent conversion to open gastric bypass), and 36 (5%) underwent open gastric bypass. Patients in each group had similar baseline characteristics (Table 1). Matching was incomplete for a single pair consisting of 1 black and 1 white patient. In each group, 34 patients (34% per group) had T2DM.
Operating room time and length of hospital stay were significantly shorter for patients in the LB group, but they had more postoperative clinic visits (Table 1). One-year outcomes were available in 93 patients in the LB group and 92 patients in the RYGB group. Patients in the LB group had an average of 3.2 adjustments in the first year (range, 0-12). Both groups had similar rates of overall (early and late) (11 [12%] for LB vs 14 [15%] for RYGB; P = .83) and late complications, but patients in the RYGB group had significantly more early complications, and those in the LB group had more reoperations and therapeutic interventions after 30 days (Table 2 and Table 3). All 3 conversions from LB to RYGB were because of failure to lose more than 10% of excess weight despite multiple LB adjustments, radiological confirmation of appropriate LB position, and nutritional counseling. Removal of the LB was performed in one patient for gastric slippage and in another for LB erosion in the stomach. Late complications after RYGB occurred in 3 patients; 2 of these had gastrojejunal ulcers identified by upper gastrointestinal tract endoscopy and treated with acid suppression therapy, and 1 had small-bowel obstruction due to an internal hernia and was treated with a laparoscopic reoperation. No deaths occurred in either group.
All weight loss outcomes were significantly better for RYGB (Table 4). The average EWL was 64% for patients in the RYGB group but only 36% for those in the LB group. In addition, 86 patients in the RYGB group (93%) successfully lost more than 40% of their excess weight compared with 29 (31%) after LB (P < .01).
Each group had 34 patients with T2DM. Most patients were receiving oral medications only; 6 patients in the LB group and 8 in the RYGB group were receiving insulin in addition to other oral medications (P = .77). One-year outcomes were available for 32 of 34 patients in the LB group and 33 of 34 in the RYGB group (Table 4). Resolution or improvement of T2DM was significantly better after RYGB (26 patients [76%]) than after LB (17 [50%]) (P = .04). At 1 year, 6 patients (75%) in the RYGB group who were using insulin had discontinued its use, but only 1 (17%) in the LB group had done so (P = .10). In addition, logistic regression showed that greater EWL explained the increase in resolution or improvement in T2DM in the RYGB group (odds ratio for resolution/improvement, 0.95 [95% confidence interval, 0.23-3.86]; P = .95 after adjustment for EWL).
The BAROS postoperative quality of life score was significantly better for the RYGB group. The individual self-esteem, physical, and social measures were significantly better for the RYGB group, but labor and sexual measures did not differ between the groups (Table 5).
Our study compared complications and 1-year outcomes of RYGB and LB in 200 pair-matched morbidly obese patients, with more than 90% follow-up. We found that RYGB had a higher rate of perioperative complications, a similar rate of overall complications, and a lower rate of reoperations compared with LB. At the end of the first year, patients who had undergone RYGB had greater weight loss, more resolution of diabetes, and better quality of life than those who underwent LB.
In the United States, RYGB and LB are the most commonly performed bariatric surgery procedures. The choice between these bariatric procedures is currently based mainly on patient and surgeon preference but should take into account the available evidence regarding the risks and benefits of each procedure. However, current data about the efficacy and safety of these procedures is inconsistent. This is likely because of differences in patient characteristics, surgeon competence, and institutional settings. In addition, some cohort studies have suggested that weight loss results with LB may improve with longer-term follow-up (≥2 years) to similar levels obtained with RYGB, but most published reports have long-term follow-up data for only 50% to 70% of patients; therefore, conclusions drawn from these studies may be of limited value.15,16,20,28
To our knowledge, only 2 prospective randomized comparisons of LB to RYGB have been performed. Of these, the first included only 51 patients23 and found better weight loss and a lower number of weight loss failures after RYGB. However, the sample was not large enough to provide definitive conclusions, especially about uncommon adverse events. The only other randomized trial was larger and well conducted, but it had baseline imbalances (significant differences in baseline BMI and age), a poor ratio of recruitment to enrollment, and an unusual gastrojejunostomy stricture rate of 14%.24 Most series have stricture rates ranging from 4% to 8% at most. Nevertheless, this trial followed up 198 patients 4 years postoperatively. Patients in the RYGB group had a higher rate of perioperative complications (likely due to the unusually high stricture rate) but better weight loss, and resolution of comorbidities was significantly superior with RYGB.
Two other studies have used a pair-matched design similar to ours. A 2004 Swiss study of 206 subjects treated in an academic medical center reported that 94% of patients completed long-term follow-up.22 In that study, RYGB had a lower rate of late complications (14% for RYGB vs 44% for LB; P < .001), higher proportions of patients losing at least 40% of excess body weight at 2 years (54% for RYGB vs 42% for LB; P < .05), and increased resolution of diabetes (from 44% to 18% for LB and from 37% to 6% for RYGB; P = .007). The percentages of patients experiencing early and late complications, and the severity of these occurrences, were comparable to our findings. The second study matched 203 patients who underwent LB to 181 patients who underwent RYGB, but only 76% of patients completed follow-up.21 That study also showed that RYGB achieved superior EWL at 24 months (80% for the RYGB group vs 59% for the LB group; P < .001) and increased resolution of T2DM (78% for the RYGB group vs 50% for the LB group; P = .01). Similar to those studies, the weight loss outcomes at 2 years in the 79% of both of our study groups who had completed 2-year follow-up were also better after RYGB (EWL at 2 years: 66.3% for 67 of 85 patients in the RYGB group vs 40.0% for 63 of 80 patients in the LB group; P < .01). In addition, we showed that greater EWL explained the higher rates of T2DM resolution in the RYGB group.
Numerous reports have been published about the safety of RYGB and LB, notably a recent systematic review of 14 comparative studies29 and the report from the Longitudinal Assessment of Bariatric Surgery (LABS) Consortium.30 In the LABS Study, in which only major adverse outcomes were reported (death; venous thromboembolism; percutaneous, endoscopic, or operative reintervention; and failure to be discharged from the hospital), patients in the RYGB group had higher BMI and more comorbidities than patients in the LB group. The composite rate of complications and death was very low for all groups studied, and the rates were higher for laparoscopic RYGB (4.8% and 0.2%, respectively) than for LB (1% and 0%, respectively). In another systematic review,31 the rate of complications varied widely across these studies, likely also related to differences in patient baseline characteristics and differences in the definitions of adverse events and how they were identified. Nonetheless, it is clear that the types of complications differ. Specifically, the most common complications after RYGB are wound infection (6%-13%), gastrojejunal anastomotic stricture (3%-8%), postoperative bleeding (1%-4%), readmission for dehydration (1%), small-bowel obstruction (1%), enteric leak (0.3%-5%), pulmonary embolism (0.3%-1%), and death (0.1%-1%).3,13,16,32- 34 On the other hand, the most common complications after LB are device-related complications (10%), esophageal or gastric pouch dilation (2%-5%), band erosion into the stomach (2%), gastric slippage (4%-6%), pulmonary embolism (0.1%), and death (0.1%-0.4%).16 Although RYGB is considered by most experts to be a more complex procedure, many have demonstrated that, when performed on properly selected and prepared patients by properly trained surgeons at high-volume medical centers, RYGB offers a low complication rate that is similar to that for LB.2,13,14,16,29,30,35,36 Our study sample is not large enough to distinguish differences in rare events such as mortality; however, the different rates of death reported in the LABS Consortium30 and others31 for laparoscopic RYGB and LB may not be directly comparable because there are significant differences between groups for patient baseline characteristics and because selection bias and confounding by severity likely affected these results.37- 39
The expected rate of resolution or improvement of T2DM is an important end point when choosing bariatric surgery. A recent systematic review and meta-analysis4 for 135 246 patients showed that EWL and diabetes resolution in the first 12 months after surgery were higher for patients undergoing RYGB (60% EWL and 80% diabetes resolution) compared with patients undergoing LB (46% EWL and 57% diabetes resolution). Although incomplete follow-up data in more than 40% of patients in the studies included in the analysis were a problem, the results are similar to those of series with more complete follow-up40,41 and also to ours.
Many explanations have been proposed to account for why—beyond greater weight loss—RYGB may offer a better rate of resolution of T2DM than LB. For example, resolution may arise from the anatomical rearrangement of the gastrointestinal tract and resultant altered secretion of gut and pancreatic hormones that are known to affect glucose metabolism and also alter glucose kinetics.42 However, recent evidence suggests that the incretin effect alone cannot account for improvements in insulin sensitivity after RYGB, and the weight loss is indeed essential for normalization of glucose disposal.43- 45
To our knowledge, the only other study directly comparing RYGB with LB in terms of quality of life is the recently published randomized trial.24 Quality-of-life outcomes are not always strongly correlated with weight loss46,47 and thus provide independent information. Furthermore, quality of life is critical because the success of bariatric surgery requires a passionate commitment from the patient. If the patients see improvement in the quality of life, they are more likely to continue long-term weight loss efforts. In contrast to the recent randomized trial in which quality of life, measured using the 36-Item Short Form Health Survey, improved similarly at 1 year for RYGB and LB groups,24 our study used a measurement tool specific to bariatric surgery and showed better quality of life after RYGB.
The main limitations of our study are the small to moderate sample size, a patient population drawn from a tertiary academic referral medical center, and the relatively short follow-up. In addition, and because it is not a randomized trial, it is likely that selection bias and confounding by severity may have affected the results. Its strengths include prospective collection of data by a trained research coordinator, the pair-matched design and resulting comparability of the RYGB and LB groups, and the 90% follow-up rate at 1 year.
Our study shows that RYGB, when performed in high-volume centers by expert surgeons, has a similar rate of overall complications and lower rate of reoperations than LB. Because it achieves greater weight loss, increased resolution of diabetes, and better improvement in quality of life, we conclude that, in the setting we studied, RYGB has a better risk-benefit profile than LB. This information should be provided when discussing bariatric surgical options with patients.
Correspondence: Guilherme M. Campos, MD, Department of Surgery, University of Wisconsin School of Medicine and Public Health, 600 Highland Ave, Suite H4/744 CSC, Madison, WI 53792-7375 (email@example.com).
Accepted for Publication: November 20, 2009.
Author Contributions: Dr Campos had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Campos, Rabl, Roll, Peeva, Smith, and Vittinghoff. Acquisition of data: Campos, Rabl, Roll, Peeva, Prado, and Smith. Analysis and interpretation of data: Campos, Rabl, Roll, Peeva, and Vittinghoff. Drafting of the manuscript: Campos, Rabl, Roll, Peeva, Prado, Smith, and Vittinghoff. Critical revision of the manuscript for important intellectual content: Campos. Statistical analysis: Vittinghoff. Obtained funding: Campos. Administrative, technical, and material support: Campos, Rabl, Roll, Peeva, Prado, and Smith. Study supervision: Campos.
Financial Disclosure: None reported.
Funding/Support: This study was supported by grant KL2 RR024130 from the National Center for Research Resources (NCRR).
Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official view of the NCRR or the National Institutes of Health.
Additional Contributions: Pamela Derish, MA, provided editorial review of the manuscript.