- Email: [email protected]
Comparison of Rates of Resolution of Diabetes Mellitus after Gastric Banding, Gastric Bypass, and Biliopancreatic Diversion
ORIGINAL SCIENTIFIC ARTICLES
Comparison of Rates of Resolution of Diabetes Mellitus after Gastric Banding, Gastric Bypass, and Biliopancreatic Diversion Manish Parikh, MD, Patricia Ayoung-Chee, BA, Eleny Romanos, BS, Nichole Lewis, BA, H Leon Pachter, MD, FACS, George Fielding, MD, Christine Ren, MD, FACS Bariatric operation is the most effective treatment for diabetes mellitus in the morbidly obese. The purpose of this study is to compare the rate of resolution of diabetes mellitus after three common laparoscopic bariatric procedures: laparoscopic adjustable gastric banding (LAGB), Roux-en-Y gastric bypass (RYGB), and biliopancreatic diversion with or without duodenal switch (BPD/DS). STUDY DESIGN: All data were prospectively collected and entered into an electronic registry. Characteristics evaluated for this study included preoperative age, body mass index, duration of diabetes, race, gender, operative time, length of stay, percent excess weight loss, oral hypoglycemic requirements, and insulin requirements. RESULTS: A total of 282 bariatric patients with diabetes mellitus were analyzed (218 LAGB, 53 RYGB, and 11 BPD/DS). Preoperative age (46 to 50 years), body mass index (46 to 50; calculated as kg/m2), race and gender breakdown, and baseline oral hypoglycemic (82% to 87%) and insulin requirements (18% to 28%) were comparable among the three groups (p ⫽ NS). Percent excess weight loss at 1, 2, and 3 years was: 43%, 50%, and 45% for LAGB; 66%, 68%, and 66% for RYGB; and 68%, 77%, and 82% for BPD/DS (p ⬍ 0.01 LAGB versus RYGB and LAGB versus BPD/DS at all time intervals). At 1 and 2 years, the proportion of patients requiring oral hypoglycemics postoperatively was 39% and 34% for LAGB; 22% and 13% for RYGB; and 11% and 13% for BPD/DS (p ⫽ NS). At 1 and 2 years, the proportion of patients requiring insulin postoperatively was 14% and 18% for LAGB; 7% and 13% for RYGB; and 11% and 13% for BPD/DS (p ⫽ NS). CONCLUSIONS: Despite the disparity in percent excess weight loss between LAGB, RYGB, and BPD/DS, the rate of resolution of diabetes mellitus is equivalent. (J Am Coll Surg 2007;205:631–635. © 2007 by the American College of Surgeons) BACKGROUND:
Competing Interests Declared: Drs Ren and Fielding are on the Medical Advisory Board for Allergan. Dr Ren is on the Medical Advisory Board for Ethicon Endosurgery. Unrestricted research grant was received from US Surgical Corporation. Presented at the American College of Surgeons’ 92nd Annual Clinical Congress, October 9, 2006, Chicago, Illinois.
type 2 diabetics is tissue insensitivity to insulin. A 2004 metaanalysis by Buchwald and colleagues3 revealed that 85% of morbidly obese patients with diabetes experienced improvement in their diabetes after bariatric operation (including banding, bypass, gastroplasty, and biliopancreatic diversion). What is unclear is whether a difference truly exists in the rate of diabetes improvement or resolution after malabsorptive versus restrictive procedures. This study aims to compare rates of resolution of diabetes mellitus after the three most common laparoscopic bariatric procedures: laparoscopic adjustable gastric banding (LAGB), Roux-en-Y gastric bypass (RYGB), and biliopancreatic diversion with and without duodenal switch (BPD/DS).
Received February 24, 2007; Revised May 24, 2007; Accepted May 29, 2007. From the Department of Surgery, New York University Program for Surgical Weight Loss, New York University School of Medicine, New York, NY. Correspondence address: Christine J Ren, MD, Department of Surgery, New York University Program for Surgical Weight Loss, New York University School of Medicine, 530 First Ave, Ste 10S, New York, NY 10016. email: [email protected]
METHODS Data from 1,293 patients undergoing laparoscopic bariatric operations (952 LAGB, 273 RYGB, and 68 BPD/DS) at our institution from July 2001 through December 2004 were prospectively collected and entered into an electronic
According to the Center for Disease Control and Prevention, 15 million Americans have diabetes mellitus and an additional 6 million Americans have either undiagnosed diabetes or impaired glucose tolerance.1 Approximately 95% of diagnosed diabetics are type 2 diabetics, and the majority of these patients (at least 60% to 70% in the US) are obese.2 The common pathophysiologic pathway in
© 2007 by the American College of Surgeons Published by Elsevier Inc.
631
ISSN 1072-7515/07/$32.00 doi:10.1016/j.jamcollsurg.2007.05.033
632
Parikh et al
Rate of DM Resolution after Bariatric Operation
J Am Coll Surg
Table 2. Characteristics of Diabetes (Preoperative) LAGB RYGB BPD/DS p (n ⴝ 218) (n ⴝ 53) (n ⴝ 11) Value
Abbreviations and Acronyms
BMI ⫽ body mass index BPD/DS ⫽ biliopancreatic diversion with or without duodenal switch LAGB ⫽ laparoscopic adjustable gastric banding %EWL ⫽ percent excess weight loss RYGB ⫽ Roux-en-Y gastric bypass
registry (Exemplo Medical). Technical aspects of each procedure have been described previously.4 All operations were performed by two laparoscopic bariatric surgeons at a university center with a comprehensive bariatric program. All patients met the 1991 National Institutes of Health Consensus Development Conference criteria5 for bariatric operation: patients had either a body mass index (BMI; calculated as kg/m2) ⱖ 40 or BMI ⱖ 35 with comorbidities. Of these 1,293 patients, 282 diabetic patients were identified. For the purposes of this study, having diabetes was determined by the requirement of diabetic medications preoperatively or having been diagnosed with diabetes (or glucose intolerance) by the patient’s primary care physician. The following information was extracted for analysis: age, admission BMI, gender, race, procedure performed, operative time, length of stay, preoperative insulin or oral hypoglycemic requirements, duration of diabetes mellitus preoperatively, percent excess weight loss (%EWL) with followup, and postoperative medication requirements. Data were supplemented retrospectively, as needed. Resolution of diabetes was defined as discontinuation of oral hypoglycemic agents or insulin postoperatively. All data were updated through February 2006. ANOVA methods, along with Tukey multiple comparisons, were used to compare the groups at different time points for %EWL, age, and initial BMI. Chi-square testing procedures were used to compare groups on race, gender, and oral hypoglycemic and insulin use.
RESULTS A total of 282 diabetic bariatric patients were identified during the 4-year period: 218 LAGB, 53 RYGB, and 11
Oral hypoglycemics (%) Insulin (%) Length of DM (y)
83 18 6.3
87 28 6.4
82 18 7.4
0.8 0.2 0.83
BPD/DS, biliopancreatic diversion with or without duodenal switch; DM, diabetes mellitus; LAGB, laparoscopic adjustable gastric banding; RYGB, Roux-en-Y gastric bypass.
BPD/DS. Patient characteristics are shown in Table 1. Mean age, preoperative BMI, and race and gender breakdown were similar among all three surgical groups. Preoperative duration of diabetes and baseline oral hypoglycemics and insulin requirements were also comparable among the 3 groups (Table 2). At each time interval, there was a significant disparity in %EWL between LAGB, RYGB, and BPD/DS (Table 3). At 1 and 2 years, the proportion of patients requiring oral hypoglycemics and insulin postoperatively was similar among the 3 surgical groups (Tables 4, 5). When comparing purely restrictive (LAGB) to more malabsorptive procedures (RYGB and BPD/DS), there was still no significant difference in rate of resolution at 2 years (p ⫽ 0.12 for oral hypoglycemics and p ⫽ 0.72 for insulin). Subgroup analysis of LAGB patients who still required either oral hypoglycemics or insulin, or both, 2 years postoperatively revealed an overall increased duration of diabetes preoperatively (8 years) and lower %EWL (40%) compared with the overall patient cohort. The vast majority of our patients experienced resolution of diabetes irrespective of bariatric operation performed.
DISCUSSION There is no doubt that bariatric operations have a powerful impact on obesity-related comorbidities, especially diabetes.3 There are ample data showing that bariatric operation in the obese patient with diabetes has a much greater benefit in terms of resolution of diabetes and overall survival compared with medical therapies alone.6,7 There is a relative scarcity of data directly comparing different bariatric
Table 1. Characteristics of Diabetic Patients Undergoing Laparoscopic Bariatric Operation Age (y) Preoperative BMI Caucasian (%) Women (%) Median operative time (min) Median LOS (h)
LAGB (n ⴝ 218)
RYGB (n ⴝ 53)
BPD/DS (n ⴝ 11)
p Value
49.8 ⫾ 11 46.1 ⫾ 7.2 89 56 60 23
46.1 ⫾ 9.6 47.2 ⫾ 6.5 79 55 135 72
46 ⫾ 10.6 50.2 ⫾ 7.7 91 73 280 96
0.06 0.13 0.08 0.58 ⬍ 0.05 ⬍ 0.05
BMI, body mass index (calculated as kg/m2); BPD/DS, biliopancreatic diversion with or without duodenal switch; LAGB, laparoscopic adjustable gastric banding; LOS, length of stay; RYGB, Roux-en-Y gastric bypass.
Vol. 205, No. 5, November 2007
Rate of DM Resolution after Bariatric Operation
Parikh et al
633
Table 3. Weight Loss in Diabetic Population Year of followup
1 2 3
%EWL
LAGB % Followup
43 50 45
%EWL
87 87 65
RYGB % Followup
66 68 66
72 56 65
%EWL
BPD/DS % Followup
68 77 82
p Value
⬍ 0.01* ⬍ 0.01* ⬍ 0.01*
55 64 56
*LAGB versus RYGB and LAGB versus BPD/DS. BPD/DS, biliopancreatic diversion with or without duodenal switch; LAGB, laparoscopic adjustable gastric banding; %EWL, percent excess weight loss; RYGB, Roux-en-Y gastric bypass.
operations in terms of comorbidity resolution, especially diabetes. The purpose of this study was to compare the rates of resolution of diabetes mellitus among three commonly performed laparoscopic bariatric operations. We found that, despite the disparity in weight loss between LAGB, RYGB, and BPD/DS, rates of resolution of diabetes were similar among these patient groups. Specifically, at 1 and 2 years postoperatively, the percent of patients requiring oral hypoglycemics or insulin was comparable in each surgical group. Consistent with current literature, the majority of our patients experienced resolution of diabetes within 2 years of bariatric operation.8-10 Most current data about diabetes resolution after bariatric operation involve RYGB. There is evidence that hormonal effects secondary to surgical bypass of the proximal small intestine leads to immediate resolution of diabetes, even before substantial weight loss has occurred.11,12 This theory is based on studies in the nonobese diabetic rat model that show that surgically bypassing the proximal intestine directly ameliorates type 2 diabetes, independent of effects of food intake and body weight.13,14 These immediate effects are not as clear in the clinical realm. In perhaps the most comprehensive evaluation of the effect of RYGB on diabetes (n ⫽ 240), Schauer and colleagues9 noted an 83% rate of diabetes resolution; the rate of “immediate medication independence,” ie, discontinuation of medications immediately after hospital discharge and before substantial weight loss had taken place, was only 30%. The authors concluded that the benefit of RYGB in terms of resolution of diabetes was primarily “weight-loss driven.”9 They acknowledged that the reduced caloric intake in the immediate postoperative period (often ⬍ 800 calories/ day) might have contributed to the immediate improve-
ment in diabetes. Although several other studies have shown that the RYGB is highly effective in delivering resolution of diabetes (with rates ranging from 74% to 83%), none looked specifically at immediate medication independence.11,15,16 Purely restrictive operations, such as the LAGB, have also been shown to be effective in delivering diabetes resolution, with resolution rates ranging from 64% to 80%.8,17-19 The concept of immediate medication independence after a restrictive operation has been reported before. In 1991, Deitel and colleagues20 reported that their diabetic patients who had undergone verticalbanded gastroplasty had achieved adequate glycemic control before discharge. Seventy-five percent of their insulin-requiring diabetics were off insulin by the time of discharge. This improved glycemic control before any substantial weight loss was accompanied by decreased caloric intake: patients were consuming a mean 575 cal/kg daily in the immediate postoperative period. Also in 1991, Jensen and colleagues,21 in a series of 83 diabetic patients undergoing vertical-banded gastroplasty, reported that 46% of their insulin users were off all medications by the time of discharge. They also attributed this immediate glycemic improvement to decreased caloric intake. Neve and colleagues,22 in a small series of diabetic patients undergoing vertical-banded gastroplasty, also reported immediate medication independence before discharge. Caloric restriction clearly plays a major role in early glycemic control in patients undergoing a restrictive operation. The clinical impact of excluding the foregut is unclear. It is difficult to determine whether immediate postoperative diabetes resolution after a bypass procedure is a result of en-
Table 4. Diabetes Resolution: Percent of Patients Requiring Oral Hypoglycemics Postoperatively
Table 5. Diabetes Resolution: Percent of Patients Requiring Postoperative Insulin
Postoperative year
Postoperative year
1 2
LAGB (%)
RYGB (%)
39 34
22 13
BPD/DS (%) p Value
11 13
0.10 0.33
BPD/DS, biliopancreatic diversion with or without duodenal switch; LAGB, laparoscopic adjustable gastric banding; RYGB, Roux-en-Y gastric bypass.
1 2
LAGB (%)
RYGB (%)
14 18
7 13
BPD/DS (%) p Value
11 13
0.80 0.99
BPD/DS, biliopancreatic diversion with or without duodenal switch; LAGB, laparoscopic adjustable gastric banding; RYGB, Roux-en-Y gastric bypass.
634
Parikh et al
Rate of DM Resolution after Bariatric Operation
teroinsular axis nullification or postoperative caloric restriction. It has already been well established that very low-calorie diets play a substantial role in glycemic control and insulin sensitivity, even independent of weight loss.23-25 A recent comprehensive review about changes in insulin resistance and glycemic control after bariatric operation proposed that caloric restriction in the short-term and weight loss in the longterm are the main mechanisms responsible for improved glucose metabolism and insulin resistance.26 Prospective studies are needed to determine the hormonal effect of bypassing the enteroinsular axis after controlling for reduced caloric intake. Although we did not specifically evaluate immediate medication independence in our patients, we have noticed this phenomenon in all three groups and believe that the postoperative low-calorie diet plays the most important role. There are very few direct comparisons of the resolution of comorbidities among various bariatric procedures. The metaanalysis by Buchwald and colleagues3 reported a lower rate of diabetes remission with restrictive procedures (48% for LAGB) versus malabsorptive procedures (84% and 99% for RYGB and BPD, respectively). The number of RYGB patients analyzed for diabetes resolution far exceeded the number of LAGB patients. In a singleinstitution comparison of outcomes of RYGB and LAGB, Kim and colleagues27 found that, despite a greater %EWL with RYGB up to 18 months postoperatively, there were no differences in improvement or resolution of comorbidities between the two groups (72% versus 77%, p ⫽ NS). Likewise, Dolan and colleagues compared BPD to a matched cohort of superobese LAGB and found a similar rate of resolution of comorbidities, despite a substantially greater %EWL with the BPD.28 These findings raise an intriguing question: What is the minimum %EWL required to deliver resolution of obesityrelated comorbidities, such as diabetes mellitus? Modest weight loss (5% to 7%) from medical therapies (diet, exercise, and medications) demonstrates improvement in diabetes but not resolution and an unacceptably high rate of weight regain.29,30 Randomized controlled data show that surgical therapy is more effective than medical therapy alone in reducing weight and resolving the metabolic syndrome.31 Clearly, a bariatric procedure is the most effective means of obtaining sustained weight loss and resolution of diabetes. In an analysis of 88 diabetic patients undergoing LAGB, Dolan and colleagues8 determined that %EWL ⬎ 31% at 6 months was the only substantial predictor of remission of diabetes. Our data substantiate this finding that ⬍ 50% EWL can still be successful in resolving obesity-related comorbidities. Subgroup analysis of our LAGB patients who still required medications 2 years postoperatively revealed a
J Am Coll Surg
longer duration of diabetes preoperatively (8 years) and a lower %EWL (40%). This is in accordance with other published reports, that patients who do not resolve their diabetes after bariatric operation typically have a longer preoperative duration of diabetes and lose less weight than those who do resolve their diabetes.15,17,18 This reinforces the concept that early surgical intervention is key. We did not specifically compare the weight loss of diabetics with nondiabetics. There are already several reports in the literature showing that diabetics lose less weight than nondiabetics (after both band and bypass).9,15,17,18 In addition, the main purpose of this study was to compare the rate of resolution of diabetes in these patients and not compare %EWL between diabetics and nondiabetics. Although not the focus of our study, this issue about the impact of diabetes on weight loss (after both restrictive and malabsorptive procedures) deserves additional study and is a topic for another article. This study has obvious limitations because of its small sample size. It is possible that our data for the RYGB and BPD/DS diabetes resolution are underpowered, especially at 2 years, because of the disproportionately lower number of patients undergoing RYGB and BPD/DS versus LAGB. When we combined the RYGB and BPD/DS into 1 group (malabsorptive) and compared them with the LAGB (restrictive), there was still no significant difference in rate of resolution at 2 years (p ⫽ 0.12 for oral hypoglycemics and p ⫽ 0.72 for insulin). Our finding of equivalent diabetes resolution substantiates the findings of single-institution comparisons done previously.27,28 Patient referral patterns accounted for the disparity between the numbers of patients in each surgical group. Because of the reputation of New York University as an experienced center for LAGB, many patients came to our program specifically seeking this operation. Patients are typically referred to us with a particular operation in mind, based on earlier education that consists of an information session, a comprehensive Web site, and standardized information booklets about all types of laparoscopic bariatric procedures. Patient preference played a major role in determining the type of operation performed and a selection bias is inevitable. We did not find a difference in the duration of diabetes preoperatively among the patient groups. We did not have adequate hemoglobin A1c data preoperatively or postoperatively to perform a meaningful statistical analysis. In conclusion, our data confirm that laparoscopic bariatric operation is highly effective in delivering diabetes resolution in the morbidly obese patient. Our comparison between three commonly performed bariatric procedures (LAGB, RYGB, and BPD/DS) reveals that, despite the disparity in %EWL between LAGB, RYGB, and BPD/DS,
Vol. 205, No. 5, November 2007
Parikh et al
the rates of resolution of diabetes are equivalent. Our data suggest that caloric restriction and weight loss alone can deliver diabetes resolution, irrespective of the hormonal effects from a bypass procedure. Longterm followup continues to be problematic in this patient population. Author Contributions Study conception and design: Parikh, Fielding, Ren Acquisition of data: Parikh, Ayoung-Chee, Romanos, Lewis Analysis and interpretation of data: Parikh, Ayoung-Chee, Romanos, Lewis Drafting of manuscript: Parikh, Fielding, Ren Critical revision: Parikh, Pachter, Fielding, Ren Acknowledgment: We thank Michel Secic, MS, for her statistical analysis.
REFERENCES 1. Center for Disease Control and Prevention. National diabetes fact sheet. 2005. Available at: http://www.cdc.gov/diabetes/ pubs/factsheet05.htm. Accessed: September 1, 2006. 2. National Task Force on the Prevention and Treatment of Obesity. Overweight, obesity, and health risk. Arch Intern Med 2000;160:898–904. 3. Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery: a systematic review and meta-analysis. JAMA 2004;292:1724–1737. 4. Parikh M, Laker S, Weiner M, et al. Objective comparison of complications resulting from laparoscopic bariatric procedures. J Am Coll Surg 2006;202:252–261. 5. Gastrointestinal Surgery for Severe Obesity. Gastrointestinal Surgery for Severe Obesity. NIH Consensus Statement, March 25⫺27, 1991;9:1⫺20. 6. Sjostrom L, Lindroos A, Peltonen M, et al. Lifestyle, diabetes and cardiovascular risk factors 10 years after bariatric surgery. N Engl J Med 2004;351:2683–2693. 7. MacDonald K, Long S, Swanson M, et al. The gastric bypass operation reduces the progression and mortality of non-insulin dependent diabetes mellitus. J Gastrointest Surg 1997;1:213–220. 8. Dolan K, Bryant R, Fielding G. Treating diabetes in the morbidly obese by laparoscopic gastric banding. Obes Surg 2003; 13:439–443. 9. Schauer P, Burguera B, Ikramuddin S, et al. Effect of laparoscopic Roux-en-Y gastric bypass on type 2 diabetes mellitus. Ann Surg 2003;238:467–485. 10. Scopinaro N, Adami G, Marinari G, et al. Biliopancreatic diversion. World J Surg 1998;22:936–946. 11. Pories W, Swanson M, MacDonald K, et al. Who would have thought it? An operation proves to be the most effective therapy for adult-onset diabetes mellitus. Ann Surg 1995; 222:339–352. 12. Rubino F, Gagner M, Gentileschi P, et al. The early effect of the Roux-en-Y gastric bypass on hormones involved in body weight regulation and glucose metabolism. Ann Surg 2004;240:236–242. 13. Rubino F, Marescaux J. Effect of duodenal-jejunal exclusion in a non-obese animal model of type 2 diabetes: a new perspective for an old disease. Ann Surg 2004;239:1–11.
Rate of DM Resolution after Bariatric Operation
635
14. Rubino F, Forgione A, Cummings D, et al. The mechanism of diabetes control after gastrointestinal bypass surgery reveals a role of the proximal small intestine in the pathophysiology of type 2 diabetes. Ann Surg 2006;244:741–749. 15. Sugerman H, Wolfe L, Sica D, et al. Diabetes and hypertension in severe obesity and effects of gastric bypass-induced weight loss. Ann Surg 2003;237:751–768. 16. Torquati A, Lutfi R, Abumrad N, et al. Is Roux-en-Y gastric bypass surgery the most effective treatment for type 2 diabetes mellitus in morbidly obese patients? J Gastrointest Surg 2005; 9:1112–1118. 17. Dixon J, O’Brien P. Health outcomes of severely obese type 2 diabetic subjects 1 year after laparoscopic adjustable gastric banding. Diabetes Care 2002;25:358–363. 18. Ponce J, Haynes B, Paynter S, et al. Effect of lap-band-induced weight loss on type 2 diabetes mellitus and hypertension. Obes Surg 2004;14:1335–1342. 19. Spivak H, Hewitt M, Onn A, et al. Weight loss and improvement of obesity-related illness in 500 U.S. patients following laparoscopic adjustable gastric banding procedure. Am J Surg 2005;189:27–32. 20. Deitel M, Sidhu P, Stone E. Effect of vertical banded gastroplasty on diabetes in the morbidly obese. Obes Surg 1991;1:113–114. 21. Jensen K, Mason E, Scott D. Changes in the postoperative hypoglycemic and antihypertensive medication requirements in morbidly obese diabetic patients after vertical banded gastroplasty. Obes Surg 1991;1:113–114. 22. Neve H, Soulsby C, Whiteley G, et al. Resolution of diabetes following vertical gastroplasty in morbidly obese patients. Obes Surg 1993;3:75–78. 23. Wing R, Marcus M, Salata R, et al. Effects of a very-low-calorie diet on long-term glycemic control in obese type 2 diabetic subjects. Arch Intern Med 1991;151:1334–1340. 24. Wing R, Blair E, Bononi P, et al. Caloric restriction per se is a significant factor in improvements in glycemic control and insulin sensitivity during weight loss in obese NIDDM patients. Diabetes Care 1994;17:30–36. 25. Jazet I, De Craen A, Van Schie E, et al. Sustained beneficial metabolic effects 18 months after a 30-day very low calorie diet in severely obese, insulin-treated patients with type 2 diabetes. Diabetes Res Clin Pract 2007;77:70–76. 26. Gumbs A, Modlin I, Ballantyne G. Changes in insulin resistance following bariatric surgery: role of caloric restriction and weight loss. Obes Surg 2005;15:462–473. 27. Kim T, Daud A, Ude A, et al. Early US outcomes of laparoscopic gastric bypass versus laparoscopic adjustable silicone gastric banding for morbid obesity. Surg Endosc 2006;20:202–209. 28. Dolan K, Hatzifotis M, Newbury L. A comparison of laparoscopic adjustable gastric banding and biliopancreatic diversion in superobesity. Obes Surg 2004;14:165–169. 29. Stern L, Iqbal N, Seshadri P, et al. The effects of lowcarbohydrate versus conventional weight loss diets in severely obese adults: one year follow-up of a randomized trial. Ann Intern Med 2004;140:778–785. 30. Foster G, Wyatt H, Hill J, et al. A randomized trial of a lowcarbohydrate diet for obesity. N Engl J Med 2003;348:2082–2090. 31. O’Brien P, Dixon J, Laurie C, et al. Treatment of mild to moderate obesity with laparoscopic adjustable gastric banding or an intensive medical program: a randomized trial. Ann Int Med 2006;144:625–633.
Comparison of Rates of Resolution of Diabetes Mellitus after Gastric Banding, Gastric Bypass, and Biliopancreatic Diversion Manish Parikh, MD, Patricia Ayoung-Chee, BA, Eleny Romanos, BS, Nichole Lewis, BA, H Leon Pachter, MD, FACS, George Fielding, MD, Christine Ren, MD, FACS Bariatric operation is the most effective treatment for diabetes mellitus in the morbidly obese. The purpose of this study is to compare the rate of resolution of diabetes mellitus after three common laparoscopic bariatric procedures: laparoscopic adjustable gastric banding (LAGB), Roux-en-Y gastric bypass (RYGB), and biliopancreatic diversion with or without duodenal switch (BPD/DS). STUDY DESIGN: All data were prospectively collected and entered into an electronic registry. Characteristics evaluated for this study included preoperative age, body mass index, duration of diabetes, race, gender, operative time, length of stay, percent excess weight loss, oral hypoglycemic requirements, and insulin requirements. RESULTS: A total of 282 bariatric patients with diabetes mellitus were analyzed (218 LAGB, 53 RYGB, and 11 BPD/DS). Preoperative age (46 to 50 years), body mass index (46 to 50; calculated as kg/m2), race and gender breakdown, and baseline oral hypoglycemic (82% to 87%) and insulin requirements (18% to 28%) were comparable among the three groups (p ⫽ NS). Percent excess weight loss at 1, 2, and 3 years was: 43%, 50%, and 45% for LAGB; 66%, 68%, and 66% for RYGB; and 68%, 77%, and 82% for BPD/DS (p ⬍ 0.01 LAGB versus RYGB and LAGB versus BPD/DS at all time intervals). At 1 and 2 years, the proportion of patients requiring oral hypoglycemics postoperatively was 39% and 34% for LAGB; 22% and 13% for RYGB; and 11% and 13% for BPD/DS (p ⫽ NS). At 1 and 2 years, the proportion of patients requiring insulin postoperatively was 14% and 18% for LAGB; 7% and 13% for RYGB; and 11% and 13% for BPD/DS (p ⫽ NS). CONCLUSIONS: Despite the disparity in percent excess weight loss between LAGB, RYGB, and BPD/DS, the rate of resolution of diabetes mellitus is equivalent. (J Am Coll Surg 2007;205:631–635. © 2007 by the American College of Surgeons) BACKGROUND:
Competing Interests Declared: Drs Ren and Fielding are on the Medical Advisory Board for Allergan. Dr Ren is on the Medical Advisory Board for Ethicon Endosurgery. Unrestricted research grant was received from US Surgical Corporation. Presented at the American College of Surgeons’ 92nd Annual Clinical Congress, October 9, 2006, Chicago, Illinois.
type 2 diabetics is tissue insensitivity to insulin. A 2004 metaanalysis by Buchwald and colleagues3 revealed that 85% of morbidly obese patients with diabetes experienced improvement in their diabetes after bariatric operation (including banding, bypass, gastroplasty, and biliopancreatic diversion). What is unclear is whether a difference truly exists in the rate of diabetes improvement or resolution after malabsorptive versus restrictive procedures. This study aims to compare rates of resolution of diabetes mellitus after the three most common laparoscopic bariatric procedures: laparoscopic adjustable gastric banding (LAGB), Roux-en-Y gastric bypass (RYGB), and biliopancreatic diversion with and without duodenal switch (BPD/DS).
Received February 24, 2007; Revised May 24, 2007; Accepted May 29, 2007. From the Department of Surgery, New York University Program for Surgical Weight Loss, New York University School of Medicine, New York, NY. Correspondence address: Christine J Ren, MD, Department of Surgery, New York University Program for Surgical Weight Loss, New York University School of Medicine, 530 First Ave, Ste 10S, New York, NY 10016. email: [email protected]
METHODS Data from 1,293 patients undergoing laparoscopic bariatric operations (952 LAGB, 273 RYGB, and 68 BPD/DS) at our institution from July 2001 through December 2004 were prospectively collected and entered into an electronic
According to the Center for Disease Control and Prevention, 15 million Americans have diabetes mellitus and an additional 6 million Americans have either undiagnosed diabetes or impaired glucose tolerance.1 Approximately 95% of diagnosed diabetics are type 2 diabetics, and the majority of these patients (at least 60% to 70% in the US) are obese.2 The common pathophysiologic pathway in
© 2007 by the American College of Surgeons Published by Elsevier Inc.
631
ISSN 1072-7515/07/$32.00 doi:10.1016/j.jamcollsurg.2007.05.033
632
Parikh et al
Rate of DM Resolution after Bariatric Operation
J Am Coll Surg
Table 2. Characteristics of Diabetes (Preoperative) LAGB RYGB BPD/DS p (n ⴝ 218) (n ⴝ 53) (n ⴝ 11) Value
Abbreviations and Acronyms
BMI ⫽ body mass index BPD/DS ⫽ biliopancreatic diversion with or without duodenal switch LAGB ⫽ laparoscopic adjustable gastric banding %EWL ⫽ percent excess weight loss RYGB ⫽ Roux-en-Y gastric bypass
registry (Exemplo Medical). Technical aspects of each procedure have been described previously.4 All operations were performed by two laparoscopic bariatric surgeons at a university center with a comprehensive bariatric program. All patients met the 1991 National Institutes of Health Consensus Development Conference criteria5 for bariatric operation: patients had either a body mass index (BMI; calculated as kg/m2) ⱖ 40 or BMI ⱖ 35 with comorbidities. Of these 1,293 patients, 282 diabetic patients were identified. For the purposes of this study, having diabetes was determined by the requirement of diabetic medications preoperatively or having been diagnosed with diabetes (or glucose intolerance) by the patient’s primary care physician. The following information was extracted for analysis: age, admission BMI, gender, race, procedure performed, operative time, length of stay, preoperative insulin or oral hypoglycemic requirements, duration of diabetes mellitus preoperatively, percent excess weight loss (%EWL) with followup, and postoperative medication requirements. Data were supplemented retrospectively, as needed. Resolution of diabetes was defined as discontinuation of oral hypoglycemic agents or insulin postoperatively. All data were updated through February 2006. ANOVA methods, along with Tukey multiple comparisons, were used to compare the groups at different time points for %EWL, age, and initial BMI. Chi-square testing procedures were used to compare groups on race, gender, and oral hypoglycemic and insulin use.
RESULTS A total of 282 diabetic bariatric patients were identified during the 4-year period: 218 LAGB, 53 RYGB, and 11
Oral hypoglycemics (%) Insulin (%) Length of DM (y)
83 18 6.3
87 28 6.4
82 18 7.4
0.8 0.2 0.83
BPD/DS, biliopancreatic diversion with or without duodenal switch; DM, diabetes mellitus; LAGB, laparoscopic adjustable gastric banding; RYGB, Roux-en-Y gastric bypass.
BPD/DS. Patient characteristics are shown in Table 1. Mean age, preoperative BMI, and race and gender breakdown were similar among all three surgical groups. Preoperative duration of diabetes and baseline oral hypoglycemics and insulin requirements were also comparable among the 3 groups (Table 2). At each time interval, there was a significant disparity in %EWL between LAGB, RYGB, and BPD/DS (Table 3). At 1 and 2 years, the proportion of patients requiring oral hypoglycemics and insulin postoperatively was similar among the 3 surgical groups (Tables 4, 5). When comparing purely restrictive (LAGB) to more malabsorptive procedures (RYGB and BPD/DS), there was still no significant difference in rate of resolution at 2 years (p ⫽ 0.12 for oral hypoglycemics and p ⫽ 0.72 for insulin). Subgroup analysis of LAGB patients who still required either oral hypoglycemics or insulin, or both, 2 years postoperatively revealed an overall increased duration of diabetes preoperatively (8 years) and lower %EWL (40%) compared with the overall patient cohort. The vast majority of our patients experienced resolution of diabetes irrespective of bariatric operation performed.
DISCUSSION There is no doubt that bariatric operations have a powerful impact on obesity-related comorbidities, especially diabetes.3 There are ample data showing that bariatric operation in the obese patient with diabetes has a much greater benefit in terms of resolution of diabetes and overall survival compared with medical therapies alone.6,7 There is a relative scarcity of data directly comparing different bariatric
Table 1. Characteristics of Diabetic Patients Undergoing Laparoscopic Bariatric Operation Age (y) Preoperative BMI Caucasian (%) Women (%) Median operative time (min) Median LOS (h)
LAGB (n ⴝ 218)
RYGB (n ⴝ 53)
BPD/DS (n ⴝ 11)
p Value
49.8 ⫾ 11 46.1 ⫾ 7.2 89 56 60 23
46.1 ⫾ 9.6 47.2 ⫾ 6.5 79 55 135 72
46 ⫾ 10.6 50.2 ⫾ 7.7 91 73 280 96
0.06 0.13 0.08 0.58 ⬍ 0.05 ⬍ 0.05
BMI, body mass index (calculated as kg/m2); BPD/DS, biliopancreatic diversion with or without duodenal switch; LAGB, laparoscopic adjustable gastric banding; LOS, length of stay; RYGB, Roux-en-Y gastric bypass.
Vol. 205, No. 5, November 2007
Rate of DM Resolution after Bariatric Operation
Parikh et al
633
Table 3. Weight Loss in Diabetic Population Year of followup
1 2 3
%EWL
LAGB % Followup
43 50 45
%EWL
87 87 65
RYGB % Followup
66 68 66
72 56 65
%EWL
BPD/DS % Followup
68 77 82
p Value
⬍ 0.01* ⬍ 0.01* ⬍ 0.01*
55 64 56
*LAGB versus RYGB and LAGB versus BPD/DS. BPD/DS, biliopancreatic diversion with or without duodenal switch; LAGB, laparoscopic adjustable gastric banding; %EWL, percent excess weight loss; RYGB, Roux-en-Y gastric bypass.
operations in terms of comorbidity resolution, especially diabetes. The purpose of this study was to compare the rates of resolution of diabetes mellitus among three commonly performed laparoscopic bariatric operations. We found that, despite the disparity in weight loss between LAGB, RYGB, and BPD/DS, rates of resolution of diabetes were similar among these patient groups. Specifically, at 1 and 2 years postoperatively, the percent of patients requiring oral hypoglycemics or insulin was comparable in each surgical group. Consistent with current literature, the majority of our patients experienced resolution of diabetes within 2 years of bariatric operation.8-10 Most current data about diabetes resolution after bariatric operation involve RYGB. There is evidence that hormonal effects secondary to surgical bypass of the proximal small intestine leads to immediate resolution of diabetes, even before substantial weight loss has occurred.11,12 This theory is based on studies in the nonobese diabetic rat model that show that surgically bypassing the proximal intestine directly ameliorates type 2 diabetes, independent of effects of food intake and body weight.13,14 These immediate effects are not as clear in the clinical realm. In perhaps the most comprehensive evaluation of the effect of RYGB on diabetes (n ⫽ 240), Schauer and colleagues9 noted an 83% rate of diabetes resolution; the rate of “immediate medication independence,” ie, discontinuation of medications immediately after hospital discharge and before substantial weight loss had taken place, was only 30%. The authors concluded that the benefit of RYGB in terms of resolution of diabetes was primarily “weight-loss driven.”9 They acknowledged that the reduced caloric intake in the immediate postoperative period (often ⬍ 800 calories/ day) might have contributed to the immediate improve-
ment in diabetes. Although several other studies have shown that the RYGB is highly effective in delivering resolution of diabetes (with rates ranging from 74% to 83%), none looked specifically at immediate medication independence.11,15,16 Purely restrictive operations, such as the LAGB, have also been shown to be effective in delivering diabetes resolution, with resolution rates ranging from 64% to 80%.8,17-19 The concept of immediate medication independence after a restrictive operation has been reported before. In 1991, Deitel and colleagues20 reported that their diabetic patients who had undergone verticalbanded gastroplasty had achieved adequate glycemic control before discharge. Seventy-five percent of their insulin-requiring diabetics were off insulin by the time of discharge. This improved glycemic control before any substantial weight loss was accompanied by decreased caloric intake: patients were consuming a mean 575 cal/kg daily in the immediate postoperative period. Also in 1991, Jensen and colleagues,21 in a series of 83 diabetic patients undergoing vertical-banded gastroplasty, reported that 46% of their insulin users were off all medications by the time of discharge. They also attributed this immediate glycemic improvement to decreased caloric intake. Neve and colleagues,22 in a small series of diabetic patients undergoing vertical-banded gastroplasty, also reported immediate medication independence before discharge. Caloric restriction clearly plays a major role in early glycemic control in patients undergoing a restrictive operation. The clinical impact of excluding the foregut is unclear. It is difficult to determine whether immediate postoperative diabetes resolution after a bypass procedure is a result of en-
Table 4. Diabetes Resolution: Percent of Patients Requiring Oral Hypoglycemics Postoperatively
Table 5. Diabetes Resolution: Percent of Patients Requiring Postoperative Insulin
Postoperative year
Postoperative year
1 2
LAGB (%)
RYGB (%)
39 34
22 13
BPD/DS (%) p Value
11 13
0.10 0.33
BPD/DS, biliopancreatic diversion with or without duodenal switch; LAGB, laparoscopic adjustable gastric banding; RYGB, Roux-en-Y gastric bypass.
1 2
LAGB (%)
RYGB (%)
14 18
7 13
BPD/DS (%) p Value
11 13
0.80 0.99
BPD/DS, biliopancreatic diversion with or without duodenal switch; LAGB, laparoscopic adjustable gastric banding; RYGB, Roux-en-Y gastric bypass.
634
Parikh et al
Rate of DM Resolution after Bariatric Operation
teroinsular axis nullification or postoperative caloric restriction. It has already been well established that very low-calorie diets play a substantial role in glycemic control and insulin sensitivity, even independent of weight loss.23-25 A recent comprehensive review about changes in insulin resistance and glycemic control after bariatric operation proposed that caloric restriction in the short-term and weight loss in the longterm are the main mechanisms responsible for improved glucose metabolism and insulin resistance.26 Prospective studies are needed to determine the hormonal effect of bypassing the enteroinsular axis after controlling for reduced caloric intake. Although we did not specifically evaluate immediate medication independence in our patients, we have noticed this phenomenon in all three groups and believe that the postoperative low-calorie diet plays the most important role. There are very few direct comparisons of the resolution of comorbidities among various bariatric procedures. The metaanalysis by Buchwald and colleagues3 reported a lower rate of diabetes remission with restrictive procedures (48% for LAGB) versus malabsorptive procedures (84% and 99% for RYGB and BPD, respectively). The number of RYGB patients analyzed for diabetes resolution far exceeded the number of LAGB patients. In a singleinstitution comparison of outcomes of RYGB and LAGB, Kim and colleagues27 found that, despite a greater %EWL with RYGB up to 18 months postoperatively, there were no differences in improvement or resolution of comorbidities between the two groups (72% versus 77%, p ⫽ NS). Likewise, Dolan and colleagues compared BPD to a matched cohort of superobese LAGB and found a similar rate of resolution of comorbidities, despite a substantially greater %EWL with the BPD.28 These findings raise an intriguing question: What is the minimum %EWL required to deliver resolution of obesityrelated comorbidities, such as diabetes mellitus? Modest weight loss (5% to 7%) from medical therapies (diet, exercise, and medications) demonstrates improvement in diabetes but not resolution and an unacceptably high rate of weight regain.29,30 Randomized controlled data show that surgical therapy is more effective than medical therapy alone in reducing weight and resolving the metabolic syndrome.31 Clearly, a bariatric procedure is the most effective means of obtaining sustained weight loss and resolution of diabetes. In an analysis of 88 diabetic patients undergoing LAGB, Dolan and colleagues8 determined that %EWL ⬎ 31% at 6 months was the only substantial predictor of remission of diabetes. Our data substantiate this finding that ⬍ 50% EWL can still be successful in resolving obesity-related comorbidities. Subgroup analysis of our LAGB patients who still required medications 2 years postoperatively revealed a
J Am Coll Surg
longer duration of diabetes preoperatively (8 years) and a lower %EWL (40%). This is in accordance with other published reports, that patients who do not resolve their diabetes after bariatric operation typically have a longer preoperative duration of diabetes and lose less weight than those who do resolve their diabetes.15,17,18 This reinforces the concept that early surgical intervention is key. We did not specifically compare the weight loss of diabetics with nondiabetics. There are already several reports in the literature showing that diabetics lose less weight than nondiabetics (after both band and bypass).9,15,17,18 In addition, the main purpose of this study was to compare the rate of resolution of diabetes in these patients and not compare %EWL between diabetics and nondiabetics. Although not the focus of our study, this issue about the impact of diabetes on weight loss (after both restrictive and malabsorptive procedures) deserves additional study and is a topic for another article. This study has obvious limitations because of its small sample size. It is possible that our data for the RYGB and BPD/DS diabetes resolution are underpowered, especially at 2 years, because of the disproportionately lower number of patients undergoing RYGB and BPD/DS versus LAGB. When we combined the RYGB and BPD/DS into 1 group (malabsorptive) and compared them with the LAGB (restrictive), there was still no significant difference in rate of resolution at 2 years (p ⫽ 0.12 for oral hypoglycemics and p ⫽ 0.72 for insulin). Our finding of equivalent diabetes resolution substantiates the findings of single-institution comparisons done previously.27,28 Patient referral patterns accounted for the disparity between the numbers of patients in each surgical group. Because of the reputation of New York University as an experienced center for LAGB, many patients came to our program specifically seeking this operation. Patients are typically referred to us with a particular operation in mind, based on earlier education that consists of an information session, a comprehensive Web site, and standardized information booklets about all types of laparoscopic bariatric procedures. Patient preference played a major role in determining the type of operation performed and a selection bias is inevitable. We did not find a difference in the duration of diabetes preoperatively among the patient groups. We did not have adequate hemoglobin A1c data preoperatively or postoperatively to perform a meaningful statistical analysis. In conclusion, our data confirm that laparoscopic bariatric operation is highly effective in delivering diabetes resolution in the morbidly obese patient. Our comparison between three commonly performed bariatric procedures (LAGB, RYGB, and BPD/DS) reveals that, despite the disparity in %EWL between LAGB, RYGB, and BPD/DS,
Vol. 205, No. 5, November 2007
Parikh et al
the rates of resolution of diabetes are equivalent. Our data suggest that caloric restriction and weight loss alone can deliver diabetes resolution, irrespective of the hormonal effects from a bypass procedure. Longterm followup continues to be problematic in this patient population. Author Contributions Study conception and design: Parikh, Fielding, Ren Acquisition of data: Parikh, Ayoung-Chee, Romanos, Lewis Analysis and interpretation of data: Parikh, Ayoung-Chee, Romanos, Lewis Drafting of manuscript: Parikh, Fielding, Ren Critical revision: Parikh, Pachter, Fielding, Ren Acknowledgment: We thank Michel Secic, MS, for her statistical analysis.
REFERENCES 1. Center for Disease Control and Prevention. National diabetes fact sheet. 2005. Available at: http://www.cdc.gov/diabetes/ pubs/factsheet05.htm. Accessed: September 1, 2006. 2. National Task Force on the Prevention and Treatment of Obesity. Overweight, obesity, and health risk. Arch Intern Med 2000;160:898–904. 3. Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery: a systematic review and meta-analysis. JAMA 2004;292:1724–1737. 4. Parikh M, Laker S, Weiner M, et al. Objective comparison of complications resulting from laparoscopic bariatric procedures. J Am Coll Surg 2006;202:252–261. 5. Gastrointestinal Surgery for Severe Obesity. Gastrointestinal Surgery for Severe Obesity. NIH Consensus Statement, March 25⫺27, 1991;9:1⫺20. 6. Sjostrom L, Lindroos A, Peltonen M, et al. Lifestyle, diabetes and cardiovascular risk factors 10 years after bariatric surgery. N Engl J Med 2004;351:2683–2693. 7. MacDonald K, Long S, Swanson M, et al. The gastric bypass operation reduces the progression and mortality of non-insulin dependent diabetes mellitus. J Gastrointest Surg 1997;1:213–220. 8. Dolan K, Bryant R, Fielding G. Treating diabetes in the morbidly obese by laparoscopic gastric banding. Obes Surg 2003; 13:439–443. 9. Schauer P, Burguera B, Ikramuddin S, et al. Effect of laparoscopic Roux-en-Y gastric bypass on type 2 diabetes mellitus. Ann Surg 2003;238:467–485. 10. Scopinaro N, Adami G, Marinari G, et al. Biliopancreatic diversion. World J Surg 1998;22:936–946. 11. Pories W, Swanson M, MacDonald K, et al. Who would have thought it? An operation proves to be the most effective therapy for adult-onset diabetes mellitus. Ann Surg 1995; 222:339–352. 12. Rubino F, Gagner M, Gentileschi P, et al. The early effect of the Roux-en-Y gastric bypass on hormones involved in body weight regulation and glucose metabolism. Ann Surg 2004;240:236–242. 13. Rubino F, Marescaux J. Effect of duodenal-jejunal exclusion in a non-obese animal model of type 2 diabetes: a new perspective for an old disease. Ann Surg 2004;239:1–11.
Rate of DM Resolution after Bariatric Operation
635
14. Rubino F, Forgione A, Cummings D, et al. The mechanism of diabetes control after gastrointestinal bypass surgery reveals a role of the proximal small intestine in the pathophysiology of type 2 diabetes. Ann Surg 2006;244:741–749. 15. Sugerman H, Wolfe L, Sica D, et al. Diabetes and hypertension in severe obesity and effects of gastric bypass-induced weight loss. Ann Surg 2003;237:751–768. 16. Torquati A, Lutfi R, Abumrad N, et al. Is Roux-en-Y gastric bypass surgery the most effective treatment for type 2 diabetes mellitus in morbidly obese patients? J Gastrointest Surg 2005; 9:1112–1118. 17. Dixon J, O’Brien P. Health outcomes of severely obese type 2 diabetic subjects 1 year after laparoscopic adjustable gastric banding. Diabetes Care 2002;25:358–363. 18. Ponce J, Haynes B, Paynter S, et al. Effect of lap-band-induced weight loss on type 2 diabetes mellitus and hypertension. Obes Surg 2004;14:1335–1342. 19. Spivak H, Hewitt M, Onn A, et al. Weight loss and improvement of obesity-related illness in 500 U.S. patients following laparoscopic adjustable gastric banding procedure. Am J Surg 2005;189:27–32. 20. Deitel M, Sidhu P, Stone E. Effect of vertical banded gastroplasty on diabetes in the morbidly obese. Obes Surg 1991;1:113–114. 21. Jensen K, Mason E, Scott D. Changes in the postoperative hypoglycemic and antihypertensive medication requirements in morbidly obese diabetic patients after vertical banded gastroplasty. Obes Surg 1991;1:113–114. 22. Neve H, Soulsby C, Whiteley G, et al. Resolution of diabetes following vertical gastroplasty in morbidly obese patients. Obes Surg 1993;3:75–78. 23. Wing R, Marcus M, Salata R, et al. Effects of a very-low-calorie diet on long-term glycemic control in obese type 2 diabetic subjects. Arch Intern Med 1991;151:1334–1340. 24. Wing R, Blair E, Bononi P, et al. Caloric restriction per se is a significant factor in improvements in glycemic control and insulin sensitivity during weight loss in obese NIDDM patients. Diabetes Care 1994;17:30–36. 25. Jazet I, De Craen A, Van Schie E, et al. Sustained beneficial metabolic effects 18 months after a 30-day very low calorie diet in severely obese, insulin-treated patients with type 2 diabetes. Diabetes Res Clin Pract 2007;77:70–76. 26. Gumbs A, Modlin I, Ballantyne G. Changes in insulin resistance following bariatric surgery: role of caloric restriction and weight loss. Obes Surg 2005;15:462–473. 27. Kim T, Daud A, Ude A, et al. Early US outcomes of laparoscopic gastric bypass versus laparoscopic adjustable silicone gastric banding for morbid obesity. Surg Endosc 2006;20:202–209. 28. Dolan K, Hatzifotis M, Newbury L. A comparison of laparoscopic adjustable gastric banding and biliopancreatic diversion in superobesity. Obes Surg 2004;14:165–169. 29. Stern L, Iqbal N, Seshadri P, et al. The effects of lowcarbohydrate versus conventional weight loss diets in severely obese adults: one year follow-up of a randomized trial. Ann Intern Med 2004;140:778–785. 30. Foster G, Wyatt H, Hill J, et al. A randomized trial of a lowcarbohydrate diet for obesity. N Engl J Med 2003;348:2082–2090. 31. O’Brien P, Dixon J, Laurie C, et al. Treatment of mild to moderate obesity with laparoscopic adjustable gastric banding or an intensive medical program: a randomized trial. Ann Int Med 2006;144:625–633.