การผ่าตัด metabolic surgery ในผู้ป่วยเบาหวานชนิดที่ 2

Diabetes Care Volume 39, June 2016 861Metabolic Surgery in the Francesco Rubino,1 David M. Nathan,2 METABOLIC SURGERYTreatment Algorithm for Type 2 Robert H. Eckel,3 Philip R. Schauer,4Diabetes: A Joint Statement by K. George M.M. Alberti,5 Paul Z. Zimmet,6International Diabetes Stefano Del Prato,7 Linong Ji,8Organizations Shaukat M. Sadikot,9 William H. Herman,10Diabetes Care 2016;39:861–877 | DOI: 10.2337/dc16-0236 Stephanie A. Amiel,1 Lee M. Kaplan,2 Gaspar Taroncher-Oldenburg,11 and David E. Cummings,12 on behalf of the Delegates of the 2nd Diabetes Surgery Summit*BACKGROUNDDespite growing evidence that bariatric/metabolic surgery powerfully improvestype 2 diabetes (T2D), existing diabetes treatment algorithms do not includesurgical options.AIMThe 2nd Diabetes Surgery Summit (DSS-II), an international consensus conference,was convened in collaboration with leading diabetes organizations to developglobal guidelines to inform clinicians and policymakers about benefits andlimitations of metabolic surgery for T2D.METHODS 1King’s College London, London, U.K. 2Harvard Medical School, Boston, MAA multidisciplinary group of 48 international clinicians/scholars (75% nonsur- 3University of Colorado Anschutz Medicalgeons), including representatives of leading diabetes organizations, participated Campus, Aurora, COin DSS-II. After evidence appraisal (MEDLINE [1 January 2005–30 September 4Cleveland Clinic, Cleveland, OH2015]), three rounds of Delphi-like questionnaires were used to measure consen- 5Imperial College London, London, U.K.sus for 32 data-based conclusions. These drafts were presented at the combined 6Monash University, Melbourne, Victoria,DSS-II and 3rd World Congress on Interventional Therapies for Type 2 Diabetes Australia(London, U.K., 28–30 September 2015), where they were open to public comment 7University of Pisa, Pisa, Italyby other professionals and amended face-to-face by the Expert Committee. 8Peking University, Beijing, China 9Diabetes India, Mumbai, IndiaRESULTS 10University of Michigan, Ann Arbor, MI 11Philadelphia, PAGiven its role in metabolic regulation, the gastrointestinal tract constitutes a 12University of Washington, Seattle, WAmeaningful target to manage T2D. Numerous randomized clinical trials, albeitmostly short/midterm, demonstrate that metabolic surgery achieves excellent Corresponding authors: Francesco Rubino,glycemic control and reduces cardiovascular risk factors. On the basis of such [email protected], and David E.evidence, metabolic surgery should be recommended to treat T2D in patients with Cummings, [email protected] III obesity (BMI ‡40 kg/m2) and in those with class II obesity (BMI 35.0–39.9kg/m2) when hyperglycemia is inadequately controlled by lifestyle and optimal This article contains Supplementary Data onlinemedical therapy. Surgery should also be considered for patients with T2D and BMI at http://care.diabetesjournals.org/lookup/30.0–34.9 kg/m2 if hyperglycemia is inadequately controlled despite optimal suppl/doi:10.2337/dc16-0236/-/DC1.treatment with either oral or injectable medications. These BMI thresholds shouldbe reduced by 2.5 kg/m2 for Asian patients. F.R. and D.E.C. chaired the writing committee for this report.CONCLUSIONS *The 2nd Diabetes Surgery Summit voting dele-Although additional studies are needed to further demonstrate long-term benefits, gates are listed in Table 2.there is sufficient clinical and mechanistic evidence to support inclusion of metabolicsurgery among antidiabetes interventions for people with T2D and obesity. To date, the © 2016 by the American Diabetes Association.DSS-II guidelines have been formally endorsed by 45 worldwide medical and scientific Readers may use this article as long as the worksocieties. Health care regulators should introduce appropriate reimbursement policies. is properly cited, the use is educational and not for profit, and the work is not altered. See accompanying articles, pp. 857, 878, 884, 893, 902, 912, 924, 934, 941, 949, and 954.

862 Metabolic Surgery in the Treatment Algorithm for T2D Diabetes Care Volume 39, June 2016Several gastrointestinal (GI) operations, In 2007, the delegates from the 1st changed their names to include the wordincluding partial gastrectomies (1,2) and Diabetes Surgery Summit (DSS-I), an inter- “metabolic” (9).bariatric procedures (Fig. 1) (3–5), pro- national consensus conference, reviewedmote dramatic, durable improvement of available clinical and mechanistic evi- Since DSS-I, a substantial body of addi-type 2 diabetes (T2D). Given the magni- dence and recommended expanding tional evidence has accumulated, includingtude and rapidity of the effect of GI the use and study of GI surgery to treat from numerous randomized clinical trialssurgery on hyperglycemia, along with diabetes, including for individuals with (RCTs), demonstrating that bariatric/experimental evidence that rearrange- only mild obesity (5,8). In the ensuing metabolic surgery achieves superior glyce-ments of GI anatomy similar to those years, the concept of “metabolic sur- mic control and reduction of cardiovascu-in some bariatric procedures directly gery” or “diabetes surgery” has become lar risk factors in obese patients with T2Daffect glucose homeostasis (6), GI inter- widely recognized in academic circles, compared with various medical/lifestyleventions have been suggested as a treat- and, accordingly, most major world- interventions (10–25). Further researchment for T2D (7). wide bariatric surgery societies have on mechanisms of action of these proce- dures (5,6,26–34) has corroborated evi-Figure 1—Diagrams of the four bariatric/metabolic operations currently in common clinical use. dence in animal studies demonstratingBPD can be performed as the classic type (shown) or with the duodenal switch variant. Reprinted an important role for the GI tract in glucosewith permission from the Cleveland Clinic Foundation (CCF). homeostasis (35), providing a biological rationale for the use of GI-based interven- tions to treat T2D. Available data, based predominantly on modeling studies, sug- gest that bariatric/metabolic surgery is also cost-effective, especially in patients with diabetes (36,37). On the basis of this mounting evidence, several international professional organiza- tions and government agencies have re- cently suggested expanding the indications for bariatric/metabolic surgery to include patients with inadequately controlled T2D and BMI as low as 30 kg/m2, and down to 27.5 kg/m2 for Asians (8,9,38,39). However, whereas obesity guidelines by national and international societies and government agencies recommend the use of bariatric surgery in individuals with T2D (9,40), clinical guidelines for diabetes care paradoxically provide little or no men- tion of a role for surgical interventions for T2D, even in patients with severe obesity (41). Despite the growing popularity of this topic in scientific communities (9) and the media (42), most diabetes care providers and patients are still inadequately in- formed about the indications, benefits, and potential risks of surgical treatments for diabetes. Moreover, insurance reim- bursement policies for bariatric/metabolic surgery continue to reflect only body weight–centric criteria and do not in- clude diabetes-related metrics or cost- effectiveness. Consequently, access to surgery for patients with diabetes is not adequately prioritized. In fact, no existing treatment algorithm for T2D includes a role for surgical intervention. Using surgery as a diabetes interven- tion, however, implies conceptual and practical differences from the traditional practice of bariatric surgery for obesity. For instance, the criteria currently used to select candidates for bariatric/metabolic

care.diabetesjournals.org Rubino and Associates 863surgery do not include metrics of meta- Executive Summarybolic disease severity, predictors of success T2D is associated with complex metabolic dysfunctions, leading to increased morbid-of treatment, or an evaluation of risks and ity, mortality, and cost. Although population-based efforts through lifestyle inter-benefits of surgery as contrasted to those ventions are essential to prevent obesity and diabetes, people who developof alternative diabetes treatment options. this disease should have access to all effective treatment options.In addition, preoperative diagnostics,perioperative management, and postop- Given its role in metabolic regulation, the GI tract constitutes a clinically and bi-erative follow-up of traditional bariatric ologically meaningful target for the management of T2D.surgery are not consistent with the needto identify and monitor diabetes-related A substantial body of evidence has accumulated, including numerous, albeit mostly short/parameters and complications. Further- midterm RCTs, demonstrating that metabolic surgeryddefined here as the use of GImore, there are no strategies yet for in- surgery with the intent to treat T2D and obesitydcan achieve excellent control oftegrating complementary pharmaceutical hyperglycemia and reduce cardiovascular risk factors.and surgical therapies to optimize out-comes of diabetes management. Although additional studies are needed to further demonstrate long-term benefits, there is now sufficient clinical and mechanistic evidence to support inclusion of met- Whereas selection criteria for bariatric abolic surgery among antidiabetes interventions for people with T2D and obesity.surgery have been standardized worldwidefor many years through an influential Na- Complementary criteria to the sole use of BMI, the traditional criterion used to selecttional Institutes of Health (NIH) consensus candidates for bariatric surgery, need to be developed to achieve a better patientstatement (43), that document is now con- selection algorithm for metabolic surgery.spicuously outdated, and there is no refer-ence for surgical treatment of diabetes to Metabolic surgery should be a recommended option to treat T2D in appropriate sur-globally raise the standards of such practice. gical candidates with class III obesity (BMI $40 kg/m2), regardless of the level of glycemic control or complexity of glucose-lowering regimens, as well as in patients Recognizing the need to inform diabetes with class II obesity (BMI 35.0–39.9 kg/m2) with inadequately controlled hyperglycemiacare providers about the benefits and lim- despite lifestyle and optimal medical therapy.itations of metabolic surgery, the 2ndDiabetes Surgery Summit (DSS-II) was con- Metabolic surgery should also be considered to be an option to treat T2Dvened in collaboration with six leading in- in patients with class I obesity (BMI 30.0–34.9 kg/m2) and inadequately con-ternational diabetes organizations: the trolled hyperglycemia despite optimal medical treatment by either oral or in-American Diabetes Association, Interna- jectable medications (including insulin).tional Diabetes Federation, Chinese Dia-betes Society, Diabetes India, European All BMI thresholds should be reconsidered depending on the ancestry of the patient.Association for the Study of Diabetes, and For example, for patients of Asian descent, the BMI values above should be reducedDiabetes UK. The overarching aim of this by 2.5 kg/m2.consensus conference was to review avail-able evidence and to develop global rec- Metabolic surgery should be performed in high-volume centers with multidisciplinaryommendations that integrate medical and teams that understand and are experienced in the management of diabetes and GIsurgical therapies in a rational treatment surgery.algorithm for T2D. Specific goals includedproviding guidance for selection of surgical Ongoing and long-term monitoring of micronutrient status, nutritional supplementation,candidates and use of diabetes-specific and support must be provided to patients after surgery, according to guidelines formeasures in the preoperative workup and postoperative management of bariatric/metabolic surgery by national andpostoperative follow-up of patients. international professional societies. At the time this article went to press, Metabolic surgery is a potentially cost-effective treatment option in obese patients withthe DSS-II consensus statements and T2D. The clinical community should work together with health care regulators to recognizeguidelines had been officially endorsed by metabolic surgery as an appropriate intervention for T2D in people with obesity and to45 leading professional societies across introduce appropriate reimbursement policies.the globe, of which 30 are primarily med-ical (diabetes, endocrinology, and gastro- METHODS surgery, including official representa-enterology) and 15 are primarily surgical tives of partner diabetes organizationsorganizations (Table 1). Additional medical DSS-II Partners and Selection of (Table 2). To ensure maximum scholarship,and scientific societies are currently con- Voting Delegates voting delegates were chosen entirelysidering endorsing these results as well. The DSS-II organizing committee and the from academicians, with no representa- partner diabetes organizations tasked a tives from industry. To further minimize These recommendations reflect cur- multidisciplinary group of 48 interna- potential conflicts of interest, nonsur-rently available data and will need to be tional authorities to develop a set of geons were purposefully overrepresentedupdated as new evidence is developed in evidence-based recommendations. This (75%) and were complemented by aca-the future. DSS-II Expert Committee included scholars demic surgeons with relevant publication representing diabetology, endocrinology, records. Two independent, nonvoting Here we report the methods for DSS-II, internal medicine, cardiology, gastroen- moderators/adjudicators developedthe resulting recommendations, and their terology, primary care, nutrition, andsupporting evidence.

864 Metabolic Surgery in the Treatment Algorithm for T2D Diabetes Care Volume 39, June 2016Table 1—International societies that have ratified and/or endorsed the DSS-II and administered questionnaires forconsensus statements and guidelines the Delphi process and chaired the face- to-face meeting of voting delegates (videPartner diabetes organizations that helped develop and have ratified Country infra). the DSS-II consensus statements and guidelines: USA Methods for Collection and Evaluation American Diabetes Association (ADA) International of Evidence International Diabetes Federation (IDF) UK Criteria used for evidence searching Diabetes UK (DUK) China were based on methods used in previous Chinese Diabetes Society (CDS) India consensus development conferences and Diabetes India (DI) systematic reviews of evidence (44,45), adapted to serve the DSS-II objectives.Other organizations that formally endorse the DSS-II consensus USA We used a highly selective, diabetes- statements and guidelines (to date): USA focused approach (only level-1 evidence USA from RCTs) to assess comparative effec- American Association of Clinical Endocrinologists (AACE) USA tiveness of surgery versus nonsurgical American College of Surgeons (ACS) Argentina therapies for T2D and to compare the American Gastroenterological Association (AGA) Argentina glycemic effects of different operations. American Society for Metabolic and Bariatric Surgery (ASMBS) International A broader evidence base was used (RCTs Argentinian Society of Diabetes (SAD) UK plus high-quality observational studies) Argentinian Society for Bariatric and Metabolic Surgery (SACO) Australia for matters such as durability of glycemic Asia-Pacific Bariatric and Metabolic Surgery Society (APBMSS) Belgium control, surgical safety, and cardiovascu- Association of British Clinical Diabetologists (ABCD) Brazil lar disease (CVD) risk reduction. Economic Australian Diabetes Society (ADS) Brazil implications of bariatric/metabolic sur- Belgian Diabetes Association (ABD) UK gery were assessed using available stud- Brazilian Society of Diabetes (SBD) Czech Republic ies of cost-effectiveness and systematic Brazilian Society of Bariatric and Metabolic Surgery (SBCBM) Chile reviews with specific reference to patients British Obesity and Metabolic Surgery Society (BOMSS) Chile with T2D. Czech Society for the Study of Obesity (CSSO) USA Chilean Society of Endocrinology and Diabetes (SCED) International Questions for evidence assessment in- Chilean Society for Bariatric and Metabolic Surgery (SCCBM) France cluded the following: 1) long-term effects Endocrine Society France of surgery on glycemic control in patients European Association for the Study of Obesity (EASO) Germany with T2D; 2) effectiveness of surgery com- French Society of Diabetes (SFD) Germany pared with medical/lifestyle interventions French Society of Bariatric and Metabolic Surgery (SOFFCO) Greece on glycemic control; 3) comparative German Diabetes Society (DDG) effectiveness of different procedures on German Society for Obesity Surgery (CA-ADIP) International T2D; 4) effects of surgery on microvas- Hellenic Diabetes Association (HDA) Israel cular complications of diabetes, CVD risk, International Federation for the Surgery of Obesity & Metabolic Italy CVD events, and mortality; 5) short- and Italy long-term surgical safety; and 6) compara- Disorders (IFSO) Japan tive safety profile of different operations. Israel Diabetes Association (IDA) International Italian Society of Bariatric & Metabolic Surgery (SICOB) Mexico MEDLINE from 1 January 2005 through Italian Society of Diabetology (SID) Mexico 15 June 2015 was searched to generate Japan Diabetes Society (JDS) Qatar the first draft of the consensus document. Latin American Association of Diabetes (ALAD) Saudi Arabia New evidence published by 30 September Mexican College of Bariatric and Metabolic Surgery (CMCOEM) USA 2015 was available for discussion in face- Mexican Society of Nutrition and Endocrinology (SMNE) UK to-face DSS-II meetings and is incorpo- Qatar Diabetes Association (QDA) USA rated into this document, using the same Saudi Diabetes and Endocrine Association (SDEA) South Africa inclusion/exclusion criteria for evidence Society of American Gastrointestinal and Endoscopic Surgeons (SAGES) Spain evaluation as in the initial draft. Society for Endocrinology (SfE) Spain Society for Surgery of the Alimentary Tract (SSAT) USA Studies considered to appraise the evi- South African Society for Surgery Obesity and Metabolism (SASSO) dence included RCTs and observational Spanish Society for Bariatric and Metabolic Surgery (SECO) studies (case-control and case-series), as Spanish Society of Diabetes (SED) appropriate for specific questions (vide The Obesity Society (TOS) infra). For both RCTs and observational studies, only reports documenting at leastThis table indicates the societies that, at the time this article went to press, had officially ratified 1-year follow-up and with 80% retentionand/or endorsed the DSS-II consensus statements and guidelines. Additional international at 2 years and 70% beyond 2 years weremedical and scientific societies are currently considering endorsing these results as well. included. These criteria are adapted from the methods of recent systematic reviews of bariatric surgery (46).

care.diabetesjournals.org Rubino and Associates 865Table 2—The DSS-II voting delegates Affiliation Nationality SpecialtyDSS-II delegate U.K. DiabetologyK. George M.M. Alberti* Imperial College London Syria EndocrinologyNizar Albache Aleppo University U.K. DiabetologyStephanie A. Amiel* King’s College London U.K. EndocrinologyRachel L. Batterham University College London U.S. CardiologyDeepak L. Bhatt Harvard Medical School Chile SurgeryCamilo Boza Cl´ınica Las Condes U.S. DiabetologyWilliam T. Cefalu Pennington Biomedical Research Center, Brazil SurgeryRicardo V. Cohen* Louisiana State University U.S. SurgeryAnita P. Courcoulas Oswaldo Cruz Hospital U.S. EndocrinologyDavid E. Cummings*† University of Pittsburgh Italy DiabetologyStefano Del Prato University of Washington Ireland EndocrinologySean F. Dinneen University of Pisa Australia General MedicineJohn B. Dixon* Galway University Hospitals U.S. EndocrinologyRobert H. Eckel Baker IDI Heart and Diabetes Institute Italy DiabetologyEle Ferrannini University of Colorado Anschutz Medical Campus Italy EndocrinologyPaola Fioretto University of Pisa Spain EndocrinologyGema Fru¨hbeck University of Padova U.S. and Canada SurgeryMichel Gagner University of Navarra, CIBERobn Florida International University and Hoˆpital du U.S. Internal MedicineRichard W. Grant U.S. EndocrinologyWilliam H. Herman Sacre´-Coeur de Montre´al U.S. SurgerySayeed Ikramuddin Kaiser Permanente Division of Research China DiabetologyLinong Ji* University of Michigan U.K. DiabetologyDesmond G. Johnston University of Minnesota U.S. GastroenterologyLee M. Kaplan*† Peking University U.S. EndocrinologySangeeta R. Kashyap Imperial College London U.K. NutritionTracy Kelly Harvard Medical School Poland DiabetologyTomasz Klupa Cleveland Clinic U.S. EndocrinologyJudith Korner Diabetes UK U.S. EndocrinologyBlandine Laferre` re Jagiellonian University U.S. DiabetologyHarold E. Lebovitz Columbia University Taiwan SurgeryWei-Jei Lee Columbia University Ireland Metabolic MedicineCarel W. le Roux* State University of New York U.S. EndocrinologyJeffrey I. Mechanick Min-Sheng General Hospital Italy Internal MedicineGeltrude Mingrone* University College Dublin U.S. SurgeryJohn M. Morton Icahn School of Medicine at Mount Sinai U.S. DiabetologyDavid M. Nathan Catholic University of Rome U.S. SurgeryWalter J. Pories Stanford University U.S. DiabetologyRobert E. Ratner* Harvard Medical School East Carolina University U.K. DiabetologyGerry Rayman American Diabetes Association, U.K. SurgeryFrancesco Rubino*† India DiabetologyShaukat M. Sadikot* Chief Scientific and Medical Officer U.S. SurgeryPhilip R. Schauer*† Ipswich Hospital NHS Trust U.S. SurgeryHarvey J. Sugerman King’s College London and King’s College Hospital Belgium EndocrinologyLuc Van Gaal Diabetes India Spain EndocrinologyJosep Vidal Cleveland Clinic China DiabetologyJianping Weng Virginia Commonwealth University U.S. Surgery/NutritionBruce M. Wolfe* University of Antwerp Australia DiabetologyPaul Z. Zimmet* Hospital Clinic Sun Yat-sen University Oregon Health & Science University Monash University†DSS-II conference codirectors. *DSS-II conference organizing committee.

866 Metabolic Surgery in the Treatment Algorithm for T2D Diabetes Care Volume 39, June 2016Question-Specific Inclusion Criteria medical consensus conferences (49). DSS-II SUMMARY OF EVIDENCE:c For evidence related to the effective- delegates who did not agree with pro- CLINICAL AND BIOLOGICAL posed statements were asked to state RATIONALE FOR SURGICAL ness of surgery versus medical/lifestyle their reasons and propose amendments. TREATMENT OF T2D interventions to control T2D, only RCTs Three rounds of questionnaires were ad- were considered. A comprehensive al- ministered to test various amendments to Evidence Supporting Surgical gorithm was used to identify all RCTs the original statements that could increase published by 30 September 2015 re- consensus levels from the group. Draft Treatment of T2D porting the effects of bariatric/metabolic conclusions generated through this itera- The GI tract is an important contributor surgery in patients with diabetes. Evi- tive process were presented at the com- to normal glucose homeostasis (35), and dence listed includes studies of patients bined DSS-II and 3rd World Congress on mounting evidence, especially over the with BMI $35 kg/m2 and ,35 kg/m2. A Interventional Therapies for Type 2 Dia- past decade, has demonstrated benefits simple meta-analysis was performed betes (WCITD 2015, London, U.K., 28–30 of bariatric/metabolic surgery to treat to present an integrated picture of ex- September 2015). Proceedings were open and prevent T2D (3,5,10–25,51–53). isting evidence. to public comment by other experts in the Beyond inducing weight loss–relatedc For evidence regarding comparative ef- field (members of the Faculty of WCITD) metabolic improvements, some opera- fectiveness of different surgical proce- and by the entire audience through opin- tions engage mechanisms that improve dures on T2D, data were obtained only ion polls, using real-time electronic voting. from RCTs in which different procedures Approximately 630 professionals and stake- glucose homeostasis independent of were used expressly to treat diabetes. holders from 50 nations on five conti- weight loss (6), such as changes in gutc For evidence regarding the effect of GI nents contributed to those discussions. hormones, bile acid metabolism, micro- surgery on CVD events and CVD risk re- duction, data were obtained from RCTs Finally, on 30 September 2015, voting biota, intestinal glucose metabolism, and when available, as well as from long- DSS-II delegates met face-to-face to de- nutrient sensing (5,6,26–34). Bariatric/ term case-control studies, and from fine a final consensus document. Several metabolic surgery confers sustained favor- the most recent relevant meta-analyses. relevant professional organizations andc For evidence regarding the durability of stakeholders were invited to observe the able effects on glycemiadup to 20 years in postoperative glycemic control, surgical proceedings by sending official representa- one observational study (52)dalthough safety in general, and comparative tives to WCITD 2015/DSS-II (Supplementary benefits can decrease over time, with or safety profiles of different operations, Table 1). The document with conclusions without weight regain (3,51,52,54–56). data were obtained from RCTs when reached by the experts underwent a final available, from longitudinal case-series review by DSS voting delegates and was Data from a growing number of re- and case-control studies, and from the then submitted to the appropriate commit- cent RCTs in patients with T2D (10–25), most recent relevant meta-analyses. tees and executive boards of partner orga- including mainly individuals with BMI nizations for formal approval (Table 1). $35 kg/m2 (the most commonly usedDescriptors of Level of Evidence threshold for traditional bariatric sur-Herein we use standard level of evidence Grade of Consensus gery) as well as some patients with BMI(LoE) descriptors, defined as follows: We used a supermajority rule to define ,35 kg/m2 (range 25–35 kg/m2), consis-IA, evidence from meta-analysis of RCTs; consensus. Consistent with other stud- tently demonstrate superior efficacyIB, evidence from at least one RCT; IIA, ies (8,50), consensus was considered of bariatric/metabolic surgery in re-evidence from at least one controlled to have been reached when $67% ofstudy without randomization; IIB, evi- the experts agreed on a given topic. ducing weight and lowering glycemiadence from at least one other type of However, language was iteratively mod- compared with a variety of medical/quasi-experimental study; III, evidence ified to maximize agreement, and thefrom nonexperimental descriptive degree of consensus for each statement lifestyle interventions (LoE IA) (Fig. 2A).studies (e.g., comparative, correlation, was graded according to the following Although the antidiabetes benefits ofor case-control); IV, evidence from expert scale: grade U 5 100% agreement surgery often wane over time, the rela-committee reports, opinions or clinical (unanimous); grade A 5 89–99% agree-experience of respected authorities, or ment; grade B 5 78–88% agreement; tive superiority of surgery over medical/both. grade C 5 67–77% agreement (Table 3). lifestyle interventions in RCTs is similar This grading scale is meant to indicate throughout a range of 1–5 years (Fig. 2B).Consensus Development Process statements that reflect unanimous or Our analysis of these trials shows a medianAfter review and appraisal of evidence, near-unanimous opinions (grade U HbA1c reduction of 2.0% for surgery versustwo independent moderators developed and grade A), strong agreement with 0.5% for conventional therapies (P ,online Delphi-like questionnaires (47,48) little variance (grade B), or a consensus 0.001) (Figs. 2C and 3A). Each of the 11to measure the degree of consensus for a statement that reflects an averaging of existing surgery-versus-medicine/lifestyleset of statements and recommendations more and possibly extremely diverse RCTs reported greater HbA1c reductionthat were believed to summarize and re- opinions (grade C). We report here both following surgery (Figs. 2C and 3B). In allflect available evidence. For each of these, the grade of consensus and the exact per- of these trials, final HbA1c in the surgicalwe sought to achieve consensus, defined centage of agreement for each statement. groups was near 6.0%, regardless of the levelas agreement by a supermajority ($67%)of voting delegates, consistent with other of baseline HbA1c (Fig. 3C). However, the majority of these RCTs have only examined 1- to 2-year results, and only a handful of them have examined results for 3–5 years. Several classic “bariatric” operations cause T2D remissionddefined as achiev- ing nondiabetic HbA1c levels off all diabe- tes medicationsdin a majority of cases

care.diabetesjournals.org Rubino and Associates 867Figure 2—A: Forest plot of Peto odds ratios (ORs) of main glycemic end points, as defined in each trial, from published RCTs of bariatric/metabolicsurgery compared with medical/lifestyle treatments for diabetes. Data are arranged in order of ascending mean baseline BMI; the dotted lineseparates trials performed with cohorts exhibiting an average baseline BMI above or below 35 kg/m2. Study duration and HbA1c end point thresholdsare shown in brackets in column 1, where “off meds” indicates a threshold achieved off all diabetes medications; otherwise, end points representHbA1c thresholds achieved with or with such medications. ORs .1 indicate a positive effect of surgery compared with medical/lifestyle treatment.For each study, the OR is shown with its 95% CI. The pooled Peto ORs (95% CI) for all data were calculated under the assumption of a fixed-effectsmodel. Weights represent inverse variance of ORs (or mean differences [MDs]) and provide an indirect measure of the relevance of each study withinthe meta-analysis, as a function of individual study size and variance. B: Forest plot of the trials depicted in Fig. 2A, with data arranged in order ofincreasing length of follow-up. C: Forest plot of MDs of HbA1c serum levels after bariatric/metabolic surgery compared with medical/lifestyletreatments in published RCTs related to diabetes. Data are arranged in order of increasing follow-up time. Negative MDs denote lower HbA1c levelsfollowing surgery than medical/lifestyle treatment. Data for each study are shown as the MD with its 95% CI. A random-effects model was used tocalculate the pooled standardized MD. Glyc. Endp., glycemic end point; mo, month; SG, sleeve gastrectomy.

868 Metabolic Surgery in the Treatment Algorithm for T2D Diabetes Care Volume 39, June 2016Figure 3—A: Box plot comparing the average changes in HbA1c between surgery and medical/lifestyle (LoE IA) (Fig. 2A). Numerous RCTs withtreatments in the first reports of the 11 RCTs published to date. The plot shows 15 sample points because postoperative follow-up ranging betweensome RCTs reported results from two different surgical arms separately. Center lines show medians; box 1 and 5 years have consistently docu- mented sustained diabetes remission inlimits indicate the 25th and 75th percentiles, as determined by R software; whiskers extend 1.5 times the 30–63% of patients (LoE IB) (10–25). Avail- able data suggest an erosion of diabetesinterquartile range from the 25th and 75th percentiles. Data points are plotted as open circles. B: Change remission over time: 35–50% or more of patients who initially achieve remission offrom baseline HbA1c in each of the 11 RCTs displayed in Fig. 3A. In trials where more than one type of surgery diabetes eventually experience recur-was studied, each operation is displayed separately, compared with the medical/lifestyle group. C: Dot plot rence. However, the median disease-freecomparing baseline with final HbA1c levels following surgery in each of the 11 RCTs displayed in Fig. 3A. period among such individuals with Roux- en-Y gastric bypass (RYGB) is 8.3 years (52,56). With or without diabetes relapse, the large majority of patients who undergo surgery maintain substantial improve- ment of glycemic control from baseline for at least 5 (LoE IB) (20) to 15 (LoE IIA) (52,55–59) years. Baseline duration of diabetes (e.g., .8 years) (LoE IB) (19), use of insulin, and poorer glycemic control (LoE IIA) are consistently associated with lower rates of diabetes remission and/or higher risk of recidivism (19,52,58). Baseline visceral fat area may also help to predict postoperative outcomes, es- pecially among Asian patients with T2D, who typically have more visceral fat compared with Caucasians with di- abetes of the same BMI (60). Beyond improving glycemia, bariatric/ metabolic surgery has been shown to confer additional health benefits in RCTs, including greater reductions com- pared with medical/lifestyle interven- tions in other CVD risk factors (10–25), and enhancements in quality-of-life measures (LoE IB) (15,19,20). Improve- ments in other critical outcomes, such as micro- and macrovascular complications of diabetes, CVD, cancer, and death, have been observed only in nonrandomized studies (LoE IIA) (3,52,57,61–65). Small retrospective analyses and a recent prospective multicenter non- randomized study (LoE IIA) (66) sug- gest that bariatric/metabolic surgery may induce similar benefits in obese ad- olescents with T2D. Teenagers appear to experience similar degrees of weight loss, diabetes remission, and improve- ment of cardiometabolic risk factors for at least 3 years after surgery (66). No randomized trials, however, have yet compared the effectiveness and safety of surgery to those of conven- tional treatment options in adolescents. Available data from economic analyses, albeit predominantly based on modeling

care.diabetesjournals.org Rubino and Associates 869studies, support cost-effectiveness of important factor determining mortal- has been observed in up to 44% ofbariatric/metabolic surgery, especially in ity, complications, reoperations, and adults prior to bariatric surgery (85).patients with T2D (37). Cost per quality- readmissions (70). Hence, differences in baseline iron sta-adjusted life-year (QALY) of bariatric/ tus may explain the large variability inmetabolic surgery in general is approxi- Safety of bariatric/metabolic surgery reported rates of postoperative ironmately $3,200–$6,300, well below the in general has improved significantly deficiency.range of $50,000/QALY deemed appro- over the last two decades, with contin-priate for coverage (36,67). In a U.S. ued refinement of minimally invasive Nutritional complications, as well asstudy of obese patients, RYGB had incre- approaches (laparoscopic surgery), en- bone demineralization, are more likelymental cost-effectiveness ratios (ICERs) hanced training and credentialing, and with intestinal bypass operations, par-of $7,000/QALY for newly diagnosed di- involvement of multidisciplinary teams. ticularly BPD (20), and less common/abetes and $12,000/QALY for estab- Mortality rates with bariatric/metabolic severe with standard RYGB, LAGB, andlished diabetes (68). As a comparison, operations are typically 0.1–0.5%, simi- VSG. Risk of bone fractures after surgeryother treatments for diabetes, such as lar to cholecystectomy or hysterectomy is unclear. One retrospective cohortintensive glycemic and lipid control, have (71–75). Morbidity has also dramatically study showed no increased fractureICERs of $41,384/QALY and $51,889/ declined with laparoscopic approaches. risk, whereas another reported a 1.2-QALY, respectively (69). Although some Major complications rates are 2–6%, fold increase in the surgery versus controlmodels have suggested that bariatric with minor complications in up to 15% groups (86,87). Postprandial hypogly-surgery may even be cost-saving, direct (71–79), comparing favorably with other cemia can also occur, especially withmeasurements of health care costs from commonly performed elective opera- RYGB (83,88). The exact prevalence ofclinical studies have not demonstrated tions (75). symptomatic hypoglycemia is unknown.that surgery decreases overall health care In one study, it affected 11% of 450expenditures. There are, however, still complica- patients who had undergone RYGB or tions of surgery that may require re- VSG (88). Severe hypoglycemia resistant A long-term assessment of health operations and rehospitalizations. A to conservative therapy, however, iscare costs in subjects enrolled in the recent multicenter study showed early rare (89).Swedish Obese Subjects (SOS) study reoperation and readmission rates afterwas performed according to diabetes laparoscopic operations of 2.5% and Novel Device-Based Interventions forstatus at baseline, providing a compari- 5.1% for RYGB, versus 0.6% and 2.0% Diabetesson of drug-related and total health care for laparoscopic adjustable gastric There has recently been increased inter-expenditure for patients who undergo banding (LAGB), versus 0.6% and 5.5% est in device-based GI interventionsbariatric surgery versus matched con- for vertical sleeve gastrectomy (VSG), designed to reproduce some of the ben-trol participants over 15 years (37). Drug after a median 3-year follow-up (76). efits of metabolic surgery. Small humancosts were lower for the surgery patients Long-term studies (.5 years) demon- studies have examined numerous ap-who started with prediabetes ($3,329 strate low rates of reoperation after proaches, including space-occupyingless per patient) or diabetes ($5,487 most bariatric/metabolic procedures endoluminal devices (90), gastric elec-less per patient). Although total health except LAGB, which is associated with trical stimulation (91), duodenal andcare costs for the surgery group were removal or revision rates of .20% over gastroduodenal endoluminal barriershigher for patients with euglycemia or 5–10 years (72,77–79). Biliopancreatic (92,93), and duodenal mucosal resurfac-prediabetes, there was no difference be- diversion (BPD), classic type or duode- ing (clinical trial reg. no. NCT01927562,tween the surgery and conventional nal switch (BPD-DS), is the most com- clinicaltrials.gov). Preliminary short-termtreatment groups for patients with dia- plex procedure, requires longer operative results show variable degrees of efficacy,betes at baseline. These findings further time, and is associated with the highest depending on the device, in improving gly-support the economic value of bariatric/ perioperative mortality and morbidity cemic and metabolic control in patientsmetabolic surgery, specifically in patients rates (80). Compared with RYGB, BPD re- with obesity and T2D. Because of limita-with obesity and T2D. There are, how- sults in more surgical complications and tions in sample size and/or relativelyever, several limitations of economic greater incidence of GI side effects (81), short-term follow-up of existing studies,studies in this field, warranting further as well as nutritional deficiencies (20) however, the current LoE for these de-research (vide infra). (LoE IB). vices was not yet deemed sufficient for formal recommendation.Safety of Bariatric/Metabolic Surgery Long-term nutritional and micronutrientProcedures used in bariatric/metabolic deficiencies with related complications, Knowledge Gapssurgery are characterized by distinct such as anemia, bone demineralization, Available RCTs do not allow an assess-anatomic rearrangements (Fig. 1). This and hypoproteinemia, may occur with ment of the relative role of surgery ver-implies differences in technical com- variable frequency depending on the sus conventional therapies in manyplexity, mechanisms of action, clinical type of procedure, requiring lifelong clinical scenarios, including the long-outcomes, and safety profiles. Safety vitamin/nutritional supplementation term effects of the most commonly per-of bariatric/metabolic surgery also (82,83). Iron deficiency after bariatric formed current procedure (VSG), or ofvaries across hospitals and surgeons. surgery, with or without clinical ane- the effectiveness of surgery in differentEmpirical data suggest that profi- mia, has been observed in 5–64% of stages of disease severity. Factors pre-ciency of the operating surgeon is an adults (84). One study reported iron dicting glycemic control after surgery deficiency in up to 50% of operated ad- olescents (66). Of note, iron deficiency

870 Metabolic Surgery in the Treatment Algorithm for T2D Diabetes Care Volume 39, June 2016are incompletely characterized, and strategies for assessing the expertise of Modeling studies are prone to risks ofthere is insufficient evidence from teams/centers providing metabolic surgery overestimating cost-savings becauseRCTs to clearly define cutoffs in diabetes to increase standardization of outcomes they make assumptions about the dura-duration and/or laboratory markers across hospitals and geographic areas. bility of clinical benefits from metabolicthat could quantitatively predict the surgery. For instance, weight regain andsuccess of treatment over time. Fur- There is also limited evidence regard- diabetes relapse have not been prop-thermore, the number of patients with ing the appropriate frequency of moni- erly accounted for in many economicBMI ,35 kg/m2 studied in RCTs is still toring of nutritional status and the analyses. Variations in nonsurgical treat-modest, and there are even fewer pa- effectiveness of different types and dos- ments of obesity and diabetes, plustients with BMI ,30 kg/m2. Few RCTs age of nutritional and vitamin supple- costs across different types of payershave compared surgical procedures mentations. The exact prevalence and (private versus public) and across coun-head-to-head, specifically to treat T2D. causes of severe hypoglycemia after tries, also are likely to determine differ-Further studies are needed to under- bariatric/metabolic surgery remain un- ent levels of return on investment.stand the roles of different operations known (89); hence, studies investigating Uncertainty also exists about the cost-in specific clinical scenarios, especially the best means of preventing and treat- effectiveness or savings of bariatric/in adolescents and patients with BMI ing this condition are warranted. metabolic surgery for patients with,35 kg/m2, and to determine what lower BMIs. On the other hand, mostexactly constitutes failure of medical/ There is a paucity of studies investi- studies so far examined patients re-lifestyle management before surgery is gating the role of multimodality therapy ceiving bariatric surgery primarily for se-considered. The current LoE is not suf- with integration of pharmaceutical and vere obesity and with a relatively lowficient to determine the role of surgery surgical treatment to optimize out- prevalence of diabetes; these studiesas a first-line treatment in most clinical comes of diabetes management. In par- might underestimate economic value ofscenarios, especially in mildly obese or ticular, little is known about the role of surgery because cost-effectiveness ap-merely overweight patients. complementary postoperative lifestyle pears to be greater in obese patients and pharmaceutical interventions to in- with diabetes at baseline compared Although it is likely that major glyce- crease and maintain diabetes remission with those without diabetes (37). Addi-mic improvements and/or prolonged or enhance glycemic control and lower tional cost-effectiveness studies of spe-diabetes remission after bariatric/ the risk of diabetes complications. cific metabolic surgery procedures inmetabolic surgery lead to reductions in different clinical scenarios, and baseddiabetes-related complications, data Although available data suggest that on RCT data, would greatly facilitateregarding micro- and macrovascular metabolic surgery may be as effective in the decision-making process of policy-events, cancer, and mortality can be ex- adolescents as in adults (66), there is makers determining insurance coveragetrapolated only from nonrandomized presently no level-1 evidence to assess for surgical treatment of T2D.trials (3,52,57,61–65,94). There are no the effectiveness of surgery comparedavailable long-term RCTs directly com- with conservative treatment in this pop- Finally, although numerous physio-paring surgery versus modern phar- ulation. In particular, there are minimal logical consequences of GI operationsmacological therapies with diabetes long-term data regarding the safety of appear to contribute to the antidiabetescomplications or CVD events as primary metabolic surgery and the potential and weight-reducing benefits of bariatric/end points, or with sufficient size, dura- negative impact of nutritional deficits metabolic surgery (5,6,26,28–34), thetion, and completeness of follow-up on growth. exact mechanisms mediating diabetesto conclusively determine the effects remission after various procedures areof surgery on these hard outcomes. Although preliminary clinical evi- not fully known. Studies designed toSuch trials, which are clearly war- dence for some device-based GI inter- further elucidate these mechanismsranted, should ideally be randomized, ventions is promising, appropriate represent an important research prior-with adequate power and follow-up to RCTs with adequate end points, sample ity. Such knowledge holds promise toexamine microvascular and CVD out- size, and follow-up are necessary for for- inform decisions regarding the choicecomes as primary end points. mal consideration of such approaches in of procedures for individual patients, the treatment algorithm for T2D. Studies to optimize surgical design, and hope- Although long-term safety and efficacy should investigate the role of these ap- fully to provide targets for novel device-of metabolic surgery have been demon- proaches in specific clinical scenarios, based and/or pharmaceutical approachesstrated in several studies (20,52,55–59), alone or in combination with medica- to T2D.investigations with follow-up beyond tions and/or lifestyle interventions, and5 years are limited. This is particularly rel- their potential value to predict surgical STATEMENTS ANDevant for some procedures such as VSG outcomes (e.g., screening of surgical RECOMMENDATIONSbecause of their relatively recent introduc- candidates).tion into clinical practice. Further evalua- (See Table 3.)tion of long-term outcomes of bariatric/ Although most studies suggest a sig-metabolic surgery, particularly in compari- nificant positive economic impact of Metabolic Surgery Versus Traditionalson with available alternative treatments of bariatric/metabolic surgery, especially Bariatric Surgerydiabetes, should be pursued. The surgeon’s in patients with T2D, current evidence Although obesity and T2D are oftenexperience appears to influence outcomes has limitations. Most assessments of the associated with one another, T2D is a dis-(70), and there is a need to identify effective economic impact of bariatric/metabolic ease entity with significant heterogeneity surgery derive from modeling studies rather than from direct measurements of economic costs from clinical trials.

care.diabetesjournals.org Rubino and Associates 871that presents distinct challenges for clin- baseline, possibly reflecting more pre- outcomes, their applications in routineical care. Therefore, the traditional model served b-cell function (19,52,96). This clinical practice and research are prob-of bariatric surgery practice, which is suggests that unnecessarily delaying ac- lematic. The DSS delegates felt that re-shaped around the goal to induce weight cess to surgery might reduce health ben- mission as currently defined should notloss and treat severe obesity, is not con- efits and cost-effectiveness of surgery in be considered to be the sole clinical ben-sistent with the principles and standards patients with diabetes. Moreover, exist- efit justifying metabolic surgery usage,of modern diabetes care. A few examples ing criteria used for coverage of bariatric especially since remission requires re-below provide an idea of the conceptual surgery are of low relevance for meta- moval of all diabetes medications, andand practical ramifications of using a dis- bolic surgery. For example, because BMI metformin is often used in individualsease-specific model of care when surgery is not a standard diagnostic parameter without diabetes. Furthermore, comple-is used specifically to treat T2D. or a measure of severity of T2D, using mentary pharmaceutical therapies such BMI thresholds as stand-alone criteria as metformin should not be discontin- Offering GI surgery with the primary for metabolic surgery does not allow ued simply to meet the definition ofintent to treat T2D, instead of just as a health care providers to appropriately remission, and metformin as well asweight reduction therapy, can influence select candidates for such operations ACE inhibitors and statins should bethe demographic characteristics and or to define criteria for prioritization of maintained as needed to sustain ade-baseline disease states of patients who this type of approach. quate glycemic control and preventelect to undergo surgery. Patients diabetes complications. Additional stud-choosing bariatric surgery are typically Defining Goals and Success of ies are warranted to identify more reli-young, predominantly female, with re- Metabolic Surgery able biological and/or clinical markers forlatively low prevalence of T2D for The loss of 50% of excess body weight an exact definition of remission and/ortheir BMI (3,4,95). In contrast, a study (a somewhat arbitrary metric) is considered cure in diabetes.comparing patient populations in a to be a successful outcome of tradi-“metabolic surgery” program versus a tional bariatric surgery. T2D, however, Patient Selectiontraditional “bariatric surgery” program describes a continuum of hyperglyce- Patient selection for metabolic surgerywithin the same academic center mic states, is a heterogeneous disorder, should be based on balancing surgicalshowed that despite being similarly and is associated with complex meta- and other long-term risks with potentialobese, patients who sought metabolic bolic dysfunctions that increase CVD long-term benefits to individual pa-surgery were older, were more often risk, as well as morbidity and mortality. tients, as with any operation (Fig. 4).male, and had more severe T2D and Thus, it is necessary to define meaning- This trade-off needs to take into accountCVD at baseline (95). Although not ful definitions of goals and successful factors such as baseline CVD risk due tosurprising, these differences can signif- treatment when surgery is used with metabolic disease and hyperglycemiaicantly influence the outcomes of sur- the primary intent to treat T2D. Because that do not adequately respond to non-gery (e.g., rates of diabetes remission/ even temporary (months to years) nor- surgical treatments, as well as condi-control, cost-effectiveness, etc.), and malization of glycemic control or major tions that could contraindicate anythey have important ramifications for long-term improvement of glycemia with- elective operation, such as prior abdom-all aspects of patient care. These implica- out remission confers potential benefits inal surgery, risk of anastomotic dehis-tions, rather than the BMI of the target for patients with T2D, remission of diabe- cence, or risks of deep vein thrombosispopulation, represent the fundamen- tes, although desirable, should not be and pulmonary embolism.tal distinction between bariatric and regarded as the only goal of metabolicmetabolic surgery, necessitating the surgery or the only measure of success. In addition, preoperative indicatorsdevelopment of a new, disease-based The success of metabolic surgery needs other than BMI should be establishedmodel of practice. to be defined in the larger context of to make patient selection for metabolic comprehensive diabetes care plans. surgery diabetes relevant. There are Traditional bariatric surgery is primar- Metabolic surgery should be considered no data showing that baseline BMI pre-ily conceived of as an intervention that a means to achieve the glycemic control dicts metabolic surgery success. In-reduces the risk of future disease (i.e., to necessary to reduce risk of microvascu- stead, strong evidence indicates thatprevent metabolic or CVD complications lar complications and CVD. To date, no preoperative BMI, at least within theof severe obesity) rather than as an ap- high-quality (RCT) data have directly obese range, does not predict theproach to treat established disease. demonstrated reductions in microvas- benefits of GI surgery with regard to di-Such (mis)conception is reflected in cular complications or CVD events, com- abetes prevention (51,57), remissionthe fact that most guidelines and criteria pared with standard therapy. (11,20,52,53,56,98,99), relapse afterfor coverage of bariatric surgery today initial remission (20), or the magni-make no recommendation for early in- An ADA expert panel in 2009 defined tude of its effects on CVD eventstervention and often delay access to partial and complete remission of T2D (62,100), cancer (61), or death (LoE IIA)surgery. However, T2D is a progressive as achievement of HbA1c ,6.5% and (51,53,56,61–63,98). Of note, a recentdisease associated with increased risk ,6.0%, respectively, off all diabetes meta-analysis of all studies reportingfor CVD and microvascular complica- medications, and maintenance of these diabetes remission rates following bari-tions. Furthermore, evidence shows glycemic levels for at least 1 year (97). atric surgerydincluding 94,579 surgicalthat metabolic improvement following Although these definitions have helped patients with T2Ddshowed that the ratesurgery in patients with T2D correlates to improve standardization of reporting of remission was equivalent among thewith shorter diabetes duration at

872 Metabolic Surgery in the Treatment Algorithm for T2D Diabetes Care Volume 39, June 201660 studies in which mean preoperative surgery, based on extant data. However, unrecognized type 1 diabetes [T1D]),BMI was $35 kg/m2 compared with additional diabetes-specific parameters planning the frequency of postoperativethe 34 studies with mean preoperative should help to identify clinical scenarios monitoring of glycemic control andBMI ,35 kg/m2 (71% versus 72%, re- where surgical treatment of T2D should diabetes complications, and use ofspectively) (98). Overall, the surgical be prioritized. complementary postoperative medicalvalue seems to be more related to therapy.improved glucose homeostasis than Preoperative Workupweight loss per se (11,12,51,54,55, Indications for surgical treatment of T2D Choice of Procedure61–63). should be evaluated by a multidisciplinary The choice of surgical procedure should clinical team following a comprehensive be based on evaluation of the risk-to- Although baseline BMI per se does preoperative assessment of diabetes and benefit ratio in individual patients,not predict outcomes in metabolic sur- metabolic health. Exact diagnosis of the weighing long-term nutritional hazardsgery, available evidence, including all type of diabetes, screening for diabetes versus effectiveness on glycemic controlexisting RCTs, is based on studies that complications, and measurement of resid- and CVD risk.have included BMI ranges among their ual insulin secretory reserve have specialprimary criteria for eligibility. The number relevance for the practice of metabolic It is too early to establish a gold stan-of patients with BMI ,35 kg/m2 in such surgery. This knowledge can inform dard operation for metabolic surgery be-studies is also limited. Inevitably, and clinicians about patients’ counseling cause of the paucity of RCTs comparinguntil additional studies identify more (e.g., the likelihood of diabetes remis- surgical procedures head-to-head. How-robust predictors of outcomes, BMI sion after surgery), risks of postoperative ever, available RCTs and nonrandom-ranges remain necessary to select pa- diabetic ketoacidosis (for patients with ized studies specifically designed totients who might benefit from metabolic compare different procedures againstFigure 4—Algorithm for the treatment of T2D, as recommended by DSS-II voting delegates. The indications above are intended for patients who areappropriate candidates for elective surgery. meds, medications.

care.diabetesjournals.org Rubino and Associates 873Table 3—Statements and recommendations Grade; LoC Grade U; LoC 100%Generalities Grade A; LoC 97% 1. Given its role in metabolic regulation, the GI tract constitutes a clinically and biologically meaningful target for Grade A; LoC 92% the management of T2D. Grade U; LoC 100% 2. There is now sufficient clinical and mechanistic evidence to support inclusion of GI surgery among antidiabetes Grade U; LoC 100% interventions for people with T2D and obesity. Grade U; LoC 100% 3. Algorithms for treating T2D should include specific scenarios in which metabolic surgery is considered to be a Grade U; LoC 100% treatment option in addition to lifestyle, nutritional, and/or pharmacological approaches. Grade A; LoC 91% 4. The development of an integrated chronic disease care model of lifestyle, nutritional, pharmacological, and Grade U; LoC 100% surgical approaches is an important priority for modern diabetes care. Grade A; LoC 97% 5. The clinical community should work together with health care regulators to recognize metabolic surgery as a Grade A; LoC 97% valid intervention for T2D in people with obesity and to introduce appropriate reimbursement policies. Grade B; LoC 85%Metabolic surgery versus traditional bariatric surgery Grade A; LoC 93% 6. Metabolic surgeryddefined here as the use of GI surgery with the intent to treat T2D and obesitydrequires the development of a diabetes-based model of clinical practice consistent with international standards of Grade U; LoC 100% diabetes care. Grade A; LoC 97% 7. Complementary criteria to the sole use of BMI, the traditional criterion used to select candidates for bariatric Grade B; LoC 87% surgery, need to be developed to achieve a better patient selection algorithm for metabolic surgery. Grade B; LoC 86% 8. RYGB, VSG, LAGB, and BPD classic or duodenal switch variant (BPD-DS), are common metabolic operations, each Grade U; LoC 100% with its own risk-to-benefit ratio. All other metabolic operations are considered to be investigational at this time. 9. Metabolic surgery should be performed in high-volume centers with multidisciplinary teams that understand Grade U; LoC 100% and are experienced in the management of diabetes and GI surgery. Grade A; LoC 98%Defining goals and success of metabolic surgery Grade A; LoC 95% 10. Although more studies are needed to further demonstrate long-term benefits, evidence exists for GI surgery to Grade U; LoC 100% Continued on p. 874 be considered as an additional approach beyond lifestyle modifications and current medical therapies to reduce complications of T2D. 11. The aim of metabolic surgery in people with T2D and obesity is to improve their hyperglycemia and other metabolic derangements, while reducing their complications of diabetes, in order to improve their long-term health.Patient selection 12. Patients’ eligibility for metabolic surgery should be assessed by a multidisciplinary team including surgeon(s), internist(s) or diabetologist(s)/endocrinologist(s), and dietitian(s) with specific expertise in diabetes care. Also, depending on individual circumstances, other relevant specialists could be consulted to evaluate the patient. 13. Contraindications for metabolic surgery include diagnosis of T1D (unless surgery is indicated for other reasons, such as severe obesity); current drug or alcohol abuse; uncontrolled psychiatric illness; lack of comprehension of the risks/benefits, expected outcomes, or alternatives; and lack of commitment to nutritional supplementation and long-term follow-up required with surgery. 14. Metabolic surgery is recommended as an option to treat T2D in patients with the following conditions: c Class III obesity (BMI $40 kg/m2), regardless of the level of glycemic control or complexity of glucose- lowering regimens. c Class II obesity (BMI 35.0–39.9 kg/m2) with inadequately controlled hyperglycemia despite lifestyle and optimal medical therapy. 15. Metabolic surgery should also be considered to be an option to treat T2D in patients with class I obesity (BMI 30.0–34.9 kg/m2) and inadequately controlled hyperglycemia despite optimal medical treatment by either oral or injectable medications (including insulin). 16. All BMI thresholds used in these recommendations should be reconsidered depending on the ancestry of the patient. For example, for patients of Asian descent, the BMI values above should be reduced by 2.5 kg/m2. 17. Given the lack of level-1 evidence involving the effects of metabolic surgery on T2D in adolescent patients, the DSS-II committee feels a recommendation for use of GI surgery in this population is inappropriate at present. However, the committee does consider this a high priority for future research.Preoperative workup 18. Preoperative patient evaluation should include assessment of endocrine, metabolic, physical, nutritional, and psychological health. 19. Preoperative evaluation should include a combination of routine clinical tests and diabetes-specific metrics. The following tests are recommended by the DSS-II expert group: c Standard preoperative tests used for GI surgery at individual providers’ institutions. c Recent tests to characterize current diabetes statusdfor example, but not limited to, HbA1c, fasting glucose, lipid profile, and tests for retinopathy, nephropathy, and neuropathy. c Tests to distinguish T1D from T2D (fasting C-peptide; anti-GAD or other autoantibodies). 20. In order to reduce the risk for postoperative infection due to hyperglycemia, an attempt should be made to improve glycemic control before surgery.Choice of procedure 21. RYGB is a well-standardized surgical procedure, and among the four accepted operations for metabolic surgery, it appears to have a more favorable risk-benefit profile in most patients with T2D.

874 Metabolic Surgery in the Treatment Algorithm for T2D Diabetes Care Volume 39, June 2016Table 3—Continued Grade; LoC 22. Although longer-term studies are needed, current data suggest that VSG is an effective procedure that results Grade B; LoC 80% in excellent weight loss and major improvement of T2D, at least in the short to medium term (1–3 years) in Grade B; LoC 85% which outcomes have been measured in RCTs. It could be a valuable option to treat diabetes, especially in Grade B; LoC 83% patients for whom concerns exist about the risk of operations that involve bowel diversion. Grade A; LoC 98% 23. LAGB is effective in improving glycemia in patients with obesity and T2D, to the degree that it causes weight Grade U; LoC 100% loss. The procedure, however, is associated with greater risk for reoperation/revision compared with RYGB Grade U; LoC 100% due to failure or band-related complications, e.g., slippage/migration, erosion, etc. Grade A; LoC 95% Grade B; LoC 85% 24. Although clinical evidence suggests that BPD/BPD-DS may be the most effective procedure in terms of glycemic control and weight loss, the operation is associated with significant risk of nutritional deficiencies, Grade B; LoC 82% making its risk-benefit profile less favorable than that of the other bariatric/metabolic procedures for most patients. BPD/BPD-DS should be considered only in patients with extreme levels of obesity (e.g., Grade A; LoC 98% BMI .60 kg/m2). Grade U; LoC 100%Postoperative follow-up 25. After surgery, patients should continue to be managed by multidisciplinary teams including diabetologists/ endocrinologists, surgeons, nutritionists, and nurses with specific diabetes expertise. 26. Postoperative follow-up should include surgical and nutritional evaluations at least every 6 months, and more often if necessary, during the first 2 postoperative years and at least annually thereafter. 27. Unless patients have a documented, stable condition of nondiabetic glycemia, glycemic control should be monitored with at least the same frequency as in standard diabetes care of nonoperated patients. 28. In patients who have reached stable normalization of hyperglycemia for at least 6 months, monitoring of glycemic control should be performed with the same frequency as recommended for patients with prediabetes because of the potential for relapse. 29. Patients with a stable condition of nondiabetic glycemia for less than 5 years should be monitored for complications of diabetes at the same frequency as before remission. Once remission reaches the 5-year mark, monitoring of complications can be done at a reduced frequency, depending on the status of each complication. Complete cessation of screening for a particular complication should be considered only if nondiabetic glycemia persists and there is no history of that complication. 30. Within the first 6 months after surgery, patients should be carefully evaluated for glycemic control and antidiabetes medication(s) tapered according to the professional opinion of the physician(s). Further medical treatment of T2D after this initial 6-month period should be dosed accordingly, but not discontinued until laboratory proof of stable glycemic normalization is obtained. Stable nondiabetic glycemia (i.e., HbA1c in the normal range) should be documented for at least two 3-month HbA1c cycles (6 months in total) before considering complete withdrawal of glucose-lowering drugs, although withdrawal of certain frontline medications (e.g., metformin) should be considered more carefully. 31. In the event of plasma glucose levels rapidly approaching the normal range early postoperatively, appropriate adjustments to medical therapy (medication types and dosage) should be implemented to prevent hypoglycemia. Metformin, thiazolidinediones, GLP-1 analogs, DPP-4 inhibitors, a-glucosidase inhibitors, and SGLT2 inhibitors are suitable drugs for early postoperative diabetes management due to their low risk of inducing hypoglycemia. 32. Ongoing and long-term monitoring of micronutrient status, nutritional supplementation, and support must be provided to patients after surgery, according to guidelines for postoperative management of metabolic/ bariatric surgery by national and international societies (for example, AACE/TOS/ASMBS, IFSO, BOMSS).Grade U 5 100% agreement (unanimous); grade A 5 89–99% agreement; grade B 5 78–88% agreement; grade C 5 67–77% agreement. AACE,American Association of Clinical Endocrinologists; ASMBS, American Society for Metabolic and Bariatric Surgery; BOMSS, British Obesity & MetabolicSurgery Society; DPP-4, dipeptidyl peptidase 4; GLP-1, glucagon-like peptide 1; IFSO, International Federation for the Surgery of Obesity andMetabolic Disorders; LoC, level of consensus; SGLT2, sodium–glucose cotransporter 2; TOS, The Obesity Society.medical/lifestyle interventions or metabolic complications compared and higher incidence of postoperativeother operations in patients with T2D with RYGB (LoE IB). complications (LoE IA).show a gradient of efficacy among the c RYGB versus LAGB: RYGB achievesfour accepted surgical approaches greater diabetes remission com- Postoperative Follow-upfor weight loss and diabetes remission, pared with LAGB (LoE IA). RYGB is Regardless of the level of diabetesas follows: BPD.RYGB.VSG.LAGB. associated with higher risk of early control and/or remission achieved byThe opposite gradient exists for com- postoperative complications but patients following surgery, diabetesparative safety of these operations lower risk of long-term reoperations management should includedin addition(10–25,72,76–79,101–104). Evidence (LoE IIA). to optimizing glycemic controldmonitoringfrom these studies can be summarized c RYGB versus VSG: Compared with VSG, and ameliorating CVD risk factors,as follows: RYGB promotes higher diabetes remis- such as hypertension and dyslipide- sion rates (LoE IA), better lipid control mia, because it is reasonable to assumec RYGB versus BPD: BPD promotes (LoE IA), similar risk of reoperation that these patients remain at higher risk greater T2D remission but more (LoE IA), better quality of life (LoE IB), of CVD complications and disease

care.diabetesjournals.org Rubino and Associates 875relapse than does the general popula- and postoperative lifestyle-based proves to be the most effective therapy fortion. Thus, until surgery-specific pre- treatments. adult-onset diabetes mellitus. Ann Surg 1995;dicting factors of diabetes relapse are 10. Identify optimal definitions of out- 222:339–350; discussion 350–352better developed, patients should con- come to be used across treatment 5. Rubino F, Schauer PR, Kaplan LM, Cummingstinue to be monitored by primary care modalities. DE. Metabolic surgery to treat type 2 diabetes:physicians, endocrinologists, and inter- 11. Increase understanding of surgical clinical outcomes and mechanisms of action.nal medicine specialists as appropriate mechanisms, so as to improve use of Annu Rev Med 2010;61:393–411and have regular postoperative screen- current treatment options and develop 6. Rubino F, Marescaux J. Effect of duodenal-ing for development and/or progression effective, new alternative therapies. jejunal exclusion in a non-obese animal modelof microvascular complications of T2D 12. Investigate the role of device-based of type 2 diabetes: a new perspective for an old(e.g., retinopathy, nephropathy, and GI interventions (“interventional disease. Ann Surg 2004;239:1–11neuropathy). Because sudden im- diabetology”) to treat T2D, in combi- 7. Rubino F, Gagner M. Potential of surgery forprovement of prolonged hyperglyce- nation with lifestyle and/or pharma- curing type 2 diabetes mellitus. Ann Surg 2002;mia can acutely worsen microvascular ceutical approaches. 236:554–559disease, particularly intensive early 13. Investigate cost-effectiveness of 8. Rubino F, Kaplan LM, Schauer PR, Cummingspostoperative monitoring is warranted specific procedures and of the use DE; Diabetes Surgery Summit Delegates. Thein patients known to be afflicted of surgery in distinct clinical scenar- Diabetes Surgery Summit consensus conference:(Table 3). ios to inform policymakers about recommendations for the evaluation and use of optimal strategies to prioritize sur- gastrointestinal surgery to treat type 2 diabetesFuture Research gical access. mellitus. Ann Surg 2010;251:399–405The DSS delegates identified the follow- 9. Cummings DE, Cohen RV. Beyond BMI: theing arenas for future research in meta- Funding and Duality of Interest. The DSS-II need for new guidelines governing the use ofbolic surgery: and WCITD 2015 were supported by the In- bariatric and metabolic surgery. Lancet Diabe- ternational Diabetes Surgery Task Force tes Endocrinol 2014;2:175–181 1. Develop and evaluate criteria for (a nonprofit organization), King’s College Lon- 10. Dixon JB, O’Brien PE, Playfair J, et al. Adjust- surgery that are more appropriate don, King’s College Hospital, Johnson & able gastric banding and conventional therapy than BMI alone in people with Johnson, Medtronic, Novo Nordisk, Fractyl, for type 2 diabetes: a randomized controlled T2D. DIAMOND MetaCure, Gore, MedImmune, and trial. JAMA 2008;299:316–323 NGM Biopharmaceuticals. These sponsors 11. Mingrone G, Panunzi S, De Gaetano A, et al. 2. Investigate the long-term effect of played no role in the selection of voting dele- Bariatric surgery versus conventional medical surgery on microvascular disease gates, the Delphi process, the DSS-II and WCITD therapy for type 2 diabetes. N Engl J Med 2012; and CVD in high-quality studies 2015 programs, or the writing of this article. 366:1577–1585 (RCTs especially and prospective, None of the DSS-II codirectors, members of 12. Schauer PR, Kashyap SR, Wolski K, et al. well-matched case-control studies). the organizing committee, or voting delegates Bariatric surgery versus intensive medical ther- received payment for their efforts. No other apy in obese patients with diabetes. N Engl J 3. Refine the structure of therapeutic potential conflicts of interest relevant to this Med 2012;366:1567–1576 algorithms to incorporate meta- article were reported. 13. Ikramuddin S, Korner J, Lee WJ, et al. Roux-en-Y bolic surgery. Author Contributions. F.R. and D.E.C. chaired gastric bypass vs intensive medical management the writing committee for this article and spear- for the control of type 2 diabetes, hypertension, 4. Establish appropriate national/ headed its development. D.M.N., R.H.E., P.R.S., and hyperlipidemia: the Diabetes Surgery Study international registries of metabolic K.G.M.M.A., P.Z.Z., S.D.P., L.J., S.M.S., W.H.H., randomized clinical trial. JAMA 2013;309:2240– surgery in patients with T2D, es- S.A.A., L.M.K., and G.T.-O. contributed to the 2249 pecially designed to facilitate preparation of this report. The 48 voting dele- 14. Courcoulas AP, Goodpaster BH, Eagleton JK, standardized collection of quality gates listed in Table 2 participated in a 4-month- et al. Surgical vs medical treatments for type 2 long-term data about CVD, mortal- long Delphi-like process to craft the 32 consensus diabetes mellitus: a randomized clinical trial. ity, and other relevant outcomes. statements, culminating in the DSS-II conference JAMA Surg 2014;149:707–715 in London, U.K. F.R., D.E.C., P.R.S., and L.M.K. 15. Halperin F, Ding SA, Simonson DC, et al. 5. Investigate the long-term effec- served as codirectors of DSS-II. Roux-en-Y gastric bypass surgery or lifestyle tiveness and safety of metabolic with intensive medical management in patients surgery in adolescents as com- References with type 2 diabetes: feasibility and 1-year re- pared with alternative treatment 1. Leyton O. Diabetes and operation: a note on sults of a randomized clinical trial. JAMA Surg options. the effect of gastrojejunostomy upon a case of 2014;149:716–726 mild diabetes mellitus with a low renal thresh- 16. Liang Z, Wu Q, Chen B, Yu P, Zhao H, Ouyang 6. Determine which operation is the old. Lancet 1925;206:1162–1163 X. Effect of laparoscopic Roux-en-Y gastric by- optimal choice for individual 2. Friedman MN, Sancetta AJ, Magovern GJ. pass surgery on type 2 diabetes mellitus with patients. The amelioration of diabetes mellitus following hypertension: a randomized controlled trial. subtotal gastrectomy. Surg Gynecol Obstet 1955; Diabetes Res Clin Pract 2013;101:50–56 7. Determine the optimal time to inter- 100:201–204 17. Wentworth JM, Playfair J, Laurie C, et al. vene in individual patients. 3. Sjo¨stro¨m L, Lindroos AK, Peltonen M, et al.; Multidisciplinary diabetes care with and with- Swedish Obese Subjects Study Scientific Group. out bariatric surgery in overweight people: a 8. Identify specific clinical scenarios in Lifestyle, diabetes, and cardiovascular risk fac- randomised controlled trial. Lancet Diabetes patients with diabetes that warrant tors 10 years after bariatric surgery. N Engl J Med Endocrinol 2014;2:545–552 escalation of treatment and earlier 2004;351:2683–2693 18. Parikh M, Chung M, Sheth S, et al. Random- consideration of surgery. 4. Pories WJ, Swanson MS, MacDonald KG, ized pilot trial of bariatric surgery versus inten- et al. Who would have thought it? An operation sive medical weight management on diabetes 9. Define the optimal use of therapies remission in type 2 diabetic patients who do that combine surgical, pharmacological, NOT meet NIH criteria for surgery and the role of soluble RAGE as a novel biomarker of suc- cess. Ann Surg 2014;260:617–622; discussion 622–624 19. Schauer PR, Bhatt DL, Kirwan JP, et al.; STAMPEDE Investigators. Bariatric surgery

876 Metabolic Surgery in the Treatment Algorithm for T2D Diabetes Care Volume 39, June 2016versus intensive medical therapy for diabetesd 34. Saeidi N, Meoli L, Nestoridi E, et al. Repro- diabetes in Swedish obese subjects. N Engl J3-year outcomes. N Engl J Med 2014;370:2002– gramming of intestinal glucose metabolism and Med 2012;367:695–7042013 glycemic control in rats after gastric bypass. Sci- 52. Sjo¨stro¨m L, Peltonen M, Jacobson P, et al.20. Mingrone G, Panunzi S, De Gaetano A, et al. ence 2013;341:406–410 Association of bariatric surgery with long-termBariatric-metabolic surgery versus conventional 35. Drucker DJ. The role of gut hormones in glu- remission of type 2 diabetes and with microvas-medical treatment in obese patients with type 2 cose homeostasis. J Clin Invest 2007;117:24–32 cular and macrovascular complications. JAMAdiabetes: 5 year follow-up of an open-label, 36. Picot J, Jones J, Colquitt JL, et al. The clinical 2014;311:2297–2304single-centre, randomised controlled trial. Lan- effectiveness and cost-effectiveness of bariatric 53. Sjo¨holm K, Pajunen P, Jacobson P, et al. In-cet 2015;386:964–973 (weight loss) surgery for obesity: a systematic cidence and remission of type 2 diabetes in re-21. Ikramuddin S, Billington CJ, Lee WJ, et al. review and economic evaluation. Health Tech- lation to degree of obesity at baseline and 2 yearRoux-en-Y gastric bypass for diabetes (the Di- nol Assess 2009;13:1–190, 215–357 weight change: the Swedish Obese Subjectsabetes Surgery Study): 2-year outcomes of a 37. Keating C, Neovius M, Sjo¨holm K, et al. (SOS) study. Diabetologia 2015;58:1448–14535-year, randomised, controlled trial. Lancet Di- Health-care costs over 15 years after bariatric 54. Sjo¨stro¨m L. Review of the key results fromabetes Endocrinol 2015;3:413–422 surgery for patients with different baseline glu- the Swedish Obese Subjects (SOS) trialda pro-22. Ding SA, Simonson DC, Wewalka M, et al. cose status: results from the Swedish Obese spective controlled intervention study of bari-Adjustable gastric band surgery or medical man- Subjects study. Lancet Diabetes Endocrinol atric surgery. J Intern Med 2013;273:219–234agement in patients with type 2 diabetes: a ran- 2015;3:855–865 55. Cohen RV, Pinheiro JC, Schiavon CA, Sallesdomized clinical trial. J Clin Endocrinol Metab 38. Zimmet P, Alberti KG, Rubino F, Dixon JB. JE, Wajchenberg BL, Cummings DE. Effects of2015;100:2546–2556 IDF’s view of bariatric surgery in type 2 diabetes. gastric bypass surgery in patients with type 223. Cummings DE, Arterburn DE, Westbrook Lancet 2011;378:108–110 diabetes and only mild obesity. Diabetes CareEO, et al. Gastric bypass surgery vs intensive 39. Kasama K, Mui W, Lee WJ, et al. IFSO-APC 2012;35:1420–1428lifestyle and medical intervention for type consensus statements 2011. Obes Surg 2012; 56. Arterburn DE, Bogart A, Sherwood NE, et al.2 diabetes: the CROSSROADS randomised 22:677–684 A multisite study of long-term remission andcontrolled trial. Diabetologia 2016;59:945– 40. National Institute for Health and Care Excel- relapse of type 2 diabetes mellitus following953 lence. Obesity: identification, assessment and gastric bypass. Obes Surg 2013;23:93–10224. Courcoulas AP, Belle SH, Neiberg RH, et al. management of overweight and obesity in chil- 57. Adams TD, Davidson LE, Litwin SE, et al.Three-year outcomes of bariatric surgery vs life- dren, young people and adults. London, National Health benefits of gastric bypass surgery afterstyle intervention for type 2 diabetes mellitus Institute for Health and Care Excellence, 2014 6 years. JAMA 2012;308:1122–1131treatment: a randomized clinical trial. JAMA 41. National Institute for Health and Care Ex- 58. Brethauer SA, Aminian A, Romero-Talama´sSurg 2015;150:931–940 cellence. Algorithm for blood glucose lowering H, et al. Can diabetes be surgically cured? Long-25. Gloy VL, Briel M, Bhatt DL, et al. Bariatric therapy in adults with type 2 diabetes. London, term metabolic effects of bariatric surgery insurgery versus non-surgical treatment for obe- National Institute for Health and Care Excel- obese patients with type 2 diabetes mellitus.sity: a systematic review and meta-analysis of lence, 2015 Ann Surg 2013;258:628–636; discussion 636–randomised controlled trials. BMJ 2013;347: 42. Grady D. Surgery for diabetes may be better 637f5934 than standard treatment. The New York Times, 59. Hsu CC, Almulaifi A, Chen JC, et al. Effect of26. Thaler JP, Cummings DE. Minireview: hor- 26 March 2012 bariatric surgery vs medical treatment on type 2monal and metabolic mechanisms of diabetes 43. Consensus Development Conference Panel. diabetes in patients with body mass index lowerremission after gastrointestinal surgery. Endo- NIH conference. Gastrointestinal surgery for severe than 35: five-year outcomes. JAMA Surg 2015;crinology 2009;150:2518–2525 obesity. Ann Intern Med 1991;115:956–961 150:1117–112427. Madsbad S, Dirksen C, Holst JJ. Mechanisms 44. Burns PB, Rohrich RJ, Chung KC. The levels 60. Yu H, Di J, Bao Y, et al. Visceral fat area as aof changes in glucose metabolism and body- of evidence and their role in evidence-based new predictor of short-term diabetes remissionweight after bariatric surgery. Lancet Diabetes medicine. Plast Reconstr Surg 2011;128:305–310 after Roux-en-Y gastric bypass surgery in Chi-Endocrinol 2014;2:152–164 45. Shekelle PG, Woolf SH, Eccles M, Grimshaw nese patients with a body mass index less than28. Salehi M, Woods SC, D’Alessio DA. Gastric J. Developing clinical guidelines. West J Med 35 kg/m2. Surg Obes Relat Dis 2015;11:6–11bypass alters both glucose-dependent and 1999;170:348–351 61. Sjo¨stro¨m L, Gummesson A, Sjo¨stro¨m CD,glucose-independent regulation of islet hor- 46. Jensen MD, Ryan DH, Apovian CM, et al.; et al.; Swedish Obese Subjects Study. Effectsmone secretion. Obesity (Silver Spring) 2015; American College of Cardiology/American Heart of bariatric surgery on cancer incidence in obese23:2046–2052 Association Task Force on Practice Guidelines; patients in Sweden (Swedish Obese Subjects29. Tremaroli V, Karlsson F, Werling M, et al. Obesity Society. 2013 AHA/ACC/TOS guideline Study): a prospective, controlled interventionRoux-en-Y gastric bypass and vertical banded for the management of overweight and obesity trial. Lancet Oncol 2009;10:653–662gastroplasty induce long-term changes on the in adults: a report of the American College of 62. Sjo¨stro¨m L, Peltonen M, Jacobson P, et al.human gut microbiome contributing to fat Cardiology/American Heart Association Task Bariatric surgery and long-term cardiovascularmass regulation. Cell Metab 2015;22:228–238 Force on Practice Guidelines and The Obesity events. JAMA 2012;307:56–6530. Dirksen C, Jørgensen NB, Bojsen-Møller KN, Society. J Am Coll Cardiol 2014;63(25 Pt B): 63. Sjo¨stro¨m L, Narbro K, Sjo¨stro¨m CD, et al.;et al. Mechanisms of improved glycaemic con- 2985–3023 Swedish Obese Subjects Study. Effects of bari-trol after Roux-en-Y gastric bypass. Diabetologia 47. Dalkey N, Helmer O. An experimental appli- atric surgery on mortality in Swedish obese sub-2012;55:1890–1901 cation of the DELPHI method to the use of ex- jects. N Engl J Med 2007;357:741–75231. Breen DM, Rasmussen BA, Kokorovic A, perts. Manage Sci 1963;9:458–467 64. Adams TD, Gress RE, Smith SC, et al. Long-Wang R, Cheung GW, Lam TK. Jejunal nutrient 48. Milholland AV, Wheeler SG, Heieck JJ. Med- term mortality after gastric bypass surgery. Nsensing is required for duodenal-jejunal bypass ical assessment by a Delphi group opinion tech- Engl J Med 2007;357:753–761surgery to rapidly lower glucose concentrations nic. N Engl J Med 1973;288:1272–1275 65. Arterburn DE, Olsen MK, Smith VA, et al.in uncontrolled diabetes. Nat Med 2012;18: 49. Gabel MJ, Shipan CR. A social choice ap- Association between bariatric surgery and long-950–955 proach to expert consensus panels. J Health term survival. JAMA 2015;313:62–7032. Ryan KK, Tremaroli V, Clemmensen C, et al. Econ 2004;23:543–564 66. Inge TH, Courcoulas AP, Jenkins TM, et al.;FXR is a molecular target for the effects of ver- 50. Kleynen M, Braun SM, Bleijlevens MH, et al. Teen-LABS Consortium. Weight loss and healthtical sleeve gastrectomy. Nature 2014;509:183–188 Using a Delphi technique to seek consensus re- status 3 years after bariatric surgery in adoles-33. Liou AP, Paziuk M, Luevano JM Jr, garding definitions, descriptions and classifica- cents. N Engl J Med 2016;374:113–123Machineni S, Turnbaugh PJ, Kaplan LM. Con- tion of terms related to implicit and explicit 67. Laiteerapong N, Huang ES. The publicserved shifts in the gut microbiota due to gastric forms of motor learning. PLoS One 2014;9: health implications of the cost-effectiveness ofbypass reduce host weight and adiposity. Sci e100227 bariatric surgery for diabetes. Diabetes CareTransl Med 2013;5:178ra41 51. Carlsson LM, Peltonen M, Ahlin S, et al. 2010;33:2126–2128 Bariatric surgery and prevention of type 2

care.diabetesjournals.org Rubino and Associates 87768. Hoerger TJ, Zhang P, Segel JE, Kahn HS, and laparoscopic duodenal switch in patients 92. Rohde U, Hedba¨ck N, Gluud LL, Vilsbøll T,Barker LE, Couper S. Cost-effectiveness of bari- with body mass index of 50 to 60: a randomized Knop FK. Effect of the EndoBarrier Gastrointes-atric surgery for severely obese adults with di- clinical trial. JAMA Surg 2015;150:352–361 tinal Liner on obesity and type 2 diabetes: aabetes. Diabetes Care 2010;33:1933–1939 82. Mechanick JI, Kushner RF, Sugerman HJ, systematic review and meta-analysis. Diabetes69. CDC Diabetes Cost-effectiveness Group. et al.; American Association of Clinical Endocri- Obes Metab 2016;18:300–305Cost-effectiveness of intensive glycemic con- nologists; Obesity Society; American Society for 93. Schouten R, Rijs CS, Bouvy ND, et al. Atrol, intensified hypertension control, and se- Metabolic & Bariatric Surgery. American Asso- multicenter, randomized efficacy study of therum cholesterol level reduction for type 2 ciation of Clinical Endocrinologists, The Obesity EndoBarrier Gastrointestinal Liner for presurgi-diabetes. JAMA 2002;287:2542–2551 Society, and American Society for Metabolic & cal weight loss prior to bariatric surgery. Ann70. Birkmeyer JD, Finks JF, O’Reilly A, et al.; Bariatric Surgery medical guidelines for clinical Surg 2010;251:236–243Michigan Bariatric Surgery Collaborative. Surgi- practice for the perioperative nutritional, met- 94. Iaconelli A, Panunzi S, De Gaetano A, et al.cal skill and complication rates after bariatric abolic, and nonsurgical support of the bariatric Effects of bilio-pancreatic diversion on diabeticsurgery. N Engl J Med 2013;369:1434–1442 surgery patient. Obesity (Silver Spring) 2009;17 complications: a 10-year follow-up. Diabetes71. Flum DR, Belle SH, King WC, et al.; Longitu- (Suppl. 1):S1–S70 Care 2011;34:561–567dinal Assessment of Bariatric Surgery (LABS) 83. Mechanick JI, Youdim A, Jones DB, et al.; 95. Rubino F, Shukla A, Pomp A, Moreira M,Consortium. Perioperative safety in the longitu- American Association of Clinical Endocrinolo- Ahn SM, Dakin G. Bariatric, metabolic, and di-dinal assessment of bariatric surgery. N Engl J gists; Obesity Society; American Society for abetes surgery: what’s in a name? Ann SurgMed 2009;361:445–454 Metabolic & Bariatric Surgery. Clinical practice 2014;259:117–12272. Courcoulas AP, Christian NJ, Belle SH, et al.; guidelines for the perioperative nutritional, 96. Nguyen KT, Billington CJ, Vella A, et al. PreservedLongitudinal Assessment of Bariatric Surgery metabolic, and nonsurgical support of the insulin secretory capacity and weight loss are the(LABS) Consortium. Weight change and health bariatric surgery patientd2013 update: co- predominant predictors of glycemic control in pa-outcomes at 3 years after bariatric surgery sponsored by American Association of Clinical tients with type 2 diabetes randomized to Roux-among individuals with severe obesity. JAMA Endocrinologists, The Obesity Society, and en-Y gastric bypass. Diabetes 2015;64:3104–31102013;310:2416–2425 American Society for Metabolic & Bariatric Sur- 97. Buse JB, Caprio S, Cefalu WT, et al. How do73. Arterburn DE, Courcoulas AP. Bariatric sur- gery. Obesity (Silver Spring) 2013;21(Suppl. 1): we define cure of diabetes? Diabetes Care 2009;gery for obesity and metabolic conditions in S1–S27 32:2133–2135adults. BMJ 2014;349:g3961 84. Del Villar Madrigal E, Neme-Yunes Y, 98. Panunzi S, De Gaetano A, Carnicelli A,74. Young MT, Gebhart A, Phelan MJ, Nguyen Clavellina-Gaytan D, Sanchez HA, Mosti M, Mingrone G. Predictors of remission of diabetesNT. Use and outcomes of laparoscopic sleeve Herrera MF. Anemia after Roux-en-Y gastric by- mellitus in severely obese individuals under-gastrectomy vs laparoscopic gastric bypass: pass. How feasible to eliminate the risk by going bariatric surgery: do BMI or procedureanalysis of the American college of surgeons proper supplementation? Obes Surg 2015;25: choice matter? A meta-analysis. Ann Surg 2015;NSQIP. J Am Coll Surg 2015;220:880–885 80–84 261:459–46775. Aminian A, Brethauer SA, Kirwan JP, 85. Flancbaum L, Belsley S, Drake V, Colarusso T, 99. Panunzi S, Carlsson L, De Gaetano A, et al.Kashyap SR, Burguera B, Schauer PR. How safe Tayler E. Preoperative nutritional status of patients Determinants of diabetes remission and glyce-is metabolic/diabetes surgery? Diabetes Obes undergoing Roux-en-Y gastric bypass for morbid mic control after bariatric surgery. DiabetesMetab 2015;17:198–201 obesity. J Gastrointest Surg 2006;10:1033–1037 Care 2016;39:166–17476. Birkmeyer NJ, Dimick JB, Share D, et al.; 86. Lalmohamed A, de Vries F, Bazelier MT, 100. Eliasson B, Liakopoulos V, Franze´ n S, et al.Michigan Bariatric Surgery Collaborative. Hospi- et al. Risk of fracture after bariatric surgery in Cardiovascular disease and mortality in patientstal complication rates with bariatric surgery in the United Kingdom: population based, retro- with type 2 diabetes after bariatric surgery inMichigan. JAMA 2010;304:435–442 spective cohort study. BMJ 2012;345:e5085 Sweden: a nationwide, matched, observational co-77. Altieri MS, Yang J, Telem DA, et al. Lap band 87. Lu CW, Chang YK, Chang HH, et al. Fracture hort study. Lancet Diabetes Endocrinol 2015;3:outcomes from 19,221 patients across centers risk after bariatric surgery: a 12-year nationwide 847–854and over a decade within the state of New York. cohort study. Medicine (Baltimore) 2015;94: 101. Mu¨ller-Stich BP, Senft JD, Warschkow R,Surg Endosc. 23 July 2015 [Epub ahead of print]. e2087 et al. Surgical versus medical treatment ofDOI: 10.1007/s00464-015-4402-8 88. Lee CJ, Clark JM, Schweitzer M, et al. Preva- type 2 diabetes mellitus in nonseverely obese78. Hutter MM, Schirmer BD, Jones DB, et al. lence of and risk factors for hypoglycemic symp- patients: a systematic review and meta-analysis.First report from the American College of Sur- toms after gastric bypass and sleeve gastrectomy. Ann Surg 2015;261:421–429geons Bariatric Surgery Center Network: laparo- Obesity (Silver Spring) 2015;23:1079–1084 102. Lee WJ, Chong K, Ser KH, et al. Gastricscopic sleeve gastrectomy has morbidity and 89. Service GJ, Thompson GB, Service FJ, bypass vs sleeve gastrectomy for type 2 diabe-effectiveness positioned between the band Andrews JC, Collazo-Clavell ML, Lloyd RV. Hy- tes mellitus: a randomized controlled trial. Archand the bypass. Ann Surg 2011;254:410–420; perinsulinemic hypoglycemia with nesidioblas- Surg 2011;146:143–148discussion 420–422 tosis after gastric-bypass surgery. N Engl J Med 103. Li JF, Lai DD, Ni B, Sun KX. Comparison of79. Nguyen NT, Slone JA, Nguyen XM, Hartman 2005;353:249–254 laparoscopic Roux-en-Y gastric bypass with lap-JS, Hoyt DB. A prospective randomized trial of 90. Ponce J, Woodman G, Swain J, et al.; aroscopic sleeve gastrectomy for morbid obe-laparoscopic gastric bypass versus laparoscopic REDUCE Pivotal Trial Investigators. The REDUCE sity or type 2 diabetes mellitus: a meta-analysisadjustable gastric banding for the treatment of pivotal trial: a prospective, randomized con- of randomized controlled trials. Can J Surg 2013;morbid obesity: outcomes, quality of life, and trolled pivotal trial of a dual intragastric balloon 56:E158–E164costs. Ann Surg 2009;250:631–641 for the treatment of obesity. Surg Obes Relat Dis 104. Lee WJ, Chong K, Lin YH, Wei JH, Chen80. Morino M, Toppino M, Forestieri P, 2015;11:874–881 SC. Laparoscopic sleeve gastrectomy versusAngrisani L, Allaix ME, Scopinaro N. Mortality 91. Lebovitz HE, Ludvik B, Yaniv I, Schwartz T, single anastomosis (mini-) gastric bypass forafter bariatric surgery: analysis of 13,871 mor- Zelewski M, Gutterman DD; Metacure Investigators. the treatment of type 2 diabetes mellitus:bidly obese patients from a national registry. Treatment of patients with obese type 2 diabetes with 5-year results of a randomized trial and studyAnn Surg 2007;246:1002–1007; discussion Tantalus-DIAMOND gastric electrical stimulation: of incretin effect. Obes Surg 2014;24:1552–1007–1009 normal triglycerides predict durable effects for at 156281. Risstad H, Søvik TT, Engstro¨m M, et al. Five- least 3 years. Horm Metab Res 2015;47:456–462year outcomes after laparoscopic gastric bypass