Controversies in Obesity Management A Technology Assessment Final Report August 10, 2015 Completed by: Institute for Clinical and Economic Review ©Institute for Clinical and Economic Review, 2015 AUTHORS: Daniel A. Ollendorf, PhD Chief Review Officer, Institute for Clinical and Economic Review Karen K. Shore, PhD Program Director, Institute for Clinical and Economic Review Chris Cameron, PhD Senior Decision Scientist, Institute for Clinical and Economic Review Jed Weissberg, MD, FACP Senior Fellow, Institute for Clinical and Economic Review Anne M. Loos, MA Research Associate, Institute for Clinical and Economic Review Patricia Synnott, MALD, MS Research Associate, Institute for Clinical and Economic Review Steven D. Pearson, MD, MSc President, Institute for Clinical and Economic Review DATE OF PUBLICATION: August 10, 2015 We would also like to thank Erin Lawler and Matt Seidner of ICER for their contributions to this report. ©Institute for Clinical and Economic Review, 2015 ii [Return to Table of Contents] About ICER The Institute for Clinical and Economic Review (ICER) is an independent non-profit research organization that evaluates medical evidence and convenes public deliberative bodies to help stakeholders interpret and apply evidence to improve patient outcomes and control costs. ICER receives funding from government grants, non-profit foundations, health plans, provider groups, and health industry manufacturers. Through all its work, ICER seeks to help create a future in which collaborative efforts to move evidence into action provide the foundation for a more effective, efficient, and just health care system. More information about ICER is available at www.icer-review.org About CTAF The California Technology Assessment Forum (CTAF) – a core program of ICER – reviews evidence reports and provides a public venue in which the evidence on the effectiveness and value of health care services can be discussed with the input of all stakeholders. CTAF seeks to help patients, clinicians, insurers, and policymakers interpret and use evidence to improve the quality and value of health care. CTAF is supported by grants from the Blue Shield of California Foundation and the California HealthCare Foundation. The CTAF Panel is an independent committee of medical evidence experts from across California, with a mix of practicing clinicians, methodologists, and leaders in patient engagement and advocacy, all of whom meet strict conflict of interest guidelines, who are convened to evaluate evidence and vote on the comparative clinical effectiveness and value of medical interventions. More information about CTAF is available at www.ctaf.org ©Institute for Clinical and Economic Review, 2015 iii [Return to Table of Contents] Table of Contents List of Abbreviations Used in this Report............................................................................................... v Executive Summary ............................................................................................................................ ES1 Introduction ........................................................................................................................................... 1 1. Background ........................................................................................................................................ 2 2. Clinical Guidelines ............................................................................................................................ 12 3. Previous Technology Assessments................................................................................................... 17 4. Evidence Review (Methods & Results) ............................................................................................ 19 5. Model of Clinical and Economic Outcomes ..................................................................................... 68 6. Questions and Discussion ................................................................................................................ 88 References ......................................................................................................................................... 100 ©Institute for Clinical and Economic Review, 2015 iv [Return to Table of Contents] List of Abbreviations Used in this Report AHRQ Agency for Healthcare Research and Quality ASMBS American Society for Metabolic and Bariatric Surgery BMI Body mass index BOLD Bariatric Outcomes Longitudinal Database BPD Biliopancreatic diversion BPD±DS Biliopancreatic diversion with or without duodenal switch CADTH Canadian Agency for Drugs and Technologies in Health CI Confidence interval CMS Centers for Medicare & Medicaid Services DEA Drug Enforcement Agency DJBL Duodenal-jejunal bypass liner DS Duodenal switch EWL Excess weight loss FDA Food and Drug Administration GERD Gastroesophageal reflux disease HbA1c Hemoglobin A1c HRQoL Health-Related Quality of Life HR Hazard ratio IGB Intragastric balloon IWQOL Impact of Weight on Quality of Life LAGB Laparoscopic adjustable gastric banding LCD Local coverage decision MAX Medicaid Analytic eXtract MBSAQIP Metabolic Bariatric Surgery Accreditation and Quality Improvement Program N/B Naltrexone/bupropion NCD National coverage determination NHANES National Health and Nutrition Examination Survey NICE National Institute for Health and Care Excellence NS Not significant OR Odds ratio P/T Phentermine/topiramate PHQ-9 Patient Health Questionnaire PMPM Per-member per-month QALY Quality-adjusted life year QoL Quality of life RCT Randomized controlled trial RR Risk ratio RYGB Roux-en-Y gastric bypass ©Institute for Clinical and Economic Review, 2015 v [Return to Table of Contents] SF-36 Standard form – 36 SOS Swedish Obese Subjects T2DM Type 2 diabetes mellitus USPSTF United States Preventive Services Task Force vBloc Vagus nerve block VSG Vertical sleeve gastrectomy ©Institute for Clinical and Economic Review, 2015 vi [Return to Table of Contents] Executive Summary As the obesity epidemic has spread around the world, a deeper understanding of the drivers of weight gain has emerged. Urbanization; change in diet to more refined foods with high added sugars; differing gut microbiomes, influenced by factors as diverse as breast feeding practices, antibiotic exposure, and dietary fiber; genetics; and our sedentary lifestyle all play a role and have contributed to the American Medical Association’s redefinition of obesity as a chronic disease, implying a need for both prevention and long-term management.1 A recent estimate found that two-thirds of Americans are overweight or obese.2 The health effects of excess weight are myriad and include the development of type 2 diabetes mellitus (T2DM), hypertension, cardiovascular disease, cancer, high blood pressure, and sleep apnea. Obesity and its sequelae generate an estimated $147 billion annually in health care costs in the US alone.3 The complexity involved in managing excess weight may affect both patient candidacy for certain treatment options as well as adherence to lifestyle changes necessary to sustain weight loss and improve health outcomes.4 As a result, there is great clinical interest in treatments that may be used for patients at various levels of obesity, as well as in interventions that promote better adherence to lifestyle change. However, success rates from lifestyle modifications alone have been modest at best. A variety of options await the patient seeking or needing to lose weight beyond conventional weight-loss approaches such as general lifestyle counseling, personal dieting, exercise regimens, and commercial weight loss plans. Bariatric surgery remains a mainstay for those with severe or morbid obesity, and it is also being explored in patients with lower levels of obesity (body mass index [BMI] <35 kg/m2); new devices have emerged or are being tested to suppress appetite and/or reduce food intake; and several medications have recently been approved by the FDA specifically for weight loss. The availability of treatments of differing intensity and invasiveness raises a number of questions, however, including 1) their effects on patients at different thresholds of BMI, 2) their proper place in the treatment continuum for any given group of patients, and 3) their performance and durability of treatment effects over the long term. The purpose of this report for CTAF, therefore, is to examine the comparative clinical effectiveness and comparative value of surgical-, device-, and medication-based treatments in relation to conventional weight-loss management as well as across intervention types. Special attention is also paid to studies conducted in individuals at lower levels of BMI (i.e., 25-35 kg/m2), a key area of uncertainty and controversy. ©Institute for Clinical and Economic Review, 2015 ES1 [Return to Table of Contents] Scope of the Assessment Our primary goal for this assessment was to compare the evidence on clinical benefits and harms for surgical procedures, devices, and medications of interest in this review to that of conventional weight-loss management (i.e., combinations of diet, exercise, and/or behavioral and lifestyle interventions). However, we also included studies that actively compared different types of interventions, both across (e.g., device vs. surgery) and within (e.g., gastric bypass vs. gastric banding) categories where available. Evidence on clinical benefit and harms was primarily derived from good- and fair-quality randomized controlled trials (RCTs) and prospective comparative cohort studies. Retrospective comparative cohort studies were summarized where feasible – particularly for harms data, given the large sample sizes available in these types of cohorts. Finally, case series of >50 patients were considered only for evidence on harms and long-term benefit. Surgical series were limited to those with two or more years of follow-up given the expected maturity of this evidence base. However, given that follow-up for emerging devices and drugs is not yet adequate to impose such a threshold, and/or is limited by the intervention approach itself (e.g., temporary balloon insertion), we did not impose a strict follow-up limit on these case series. Analytic Framework The analytic framework for this evaluation is depicted in Figure ES1 on the next page. Because there were expected limitations on the impact of obesity interventions on long-term measures of morbidity, mortality, and health-related quality of life (HRQoL), a series of conceptual links was required to tie shorter-term impact on measures of body weight and resolution of or improvement in key comorbidities to longer-term outcomes. ©Institute for Clinical and Economic Review, 2015 ES2 [Return to Table of Contents] Figure ES1. Analytic Framework for Evaluation of Obesity-related Treatments Vertical Sleeve Roux-en-y Gastric Bypass Gastrectomy Benefits Adjustable Gastric Biliopancreatic Diversion Devices & Procedures Banding (± Duodenal Switch) Weight/BMI Comorbidities Intragastric Balloon Vagus Nerve Block Morbidity Patients who are overweight Duodenal-Jejunal Bypass Mortality or obese Liner (BMI ≥25) Harms Quality of life Naltrexone/bupropion Phentermine/topiramate Complications Medications extended-release extended-release Revisions Lorcaserin Liraglutide Removal Other adverse effects Other Conventional Treatment (Other medications, diet & exercise, psychotherapy, etc.) Interventions For assessment of surgery, studies were limited to those that involve the four bariatric procedures shown in Figure ES1 and commonly used in the US: Roux-en-Y gastric bypass (RYGB), laparoscopic adjustable gastric banding (LAGB), vertical sleeve gastrectomy (VSG), and biliopancreatic diversion with or without duodenal switch (BPD±DS). We also considered the evidence for three newer types of devices and four medications listed below: Devices: temporary intragastric balloon systems (IGB) (e.g., Silimed®, ReShape®), duodenal- jejunal bypass liner (DJBL) (EndoBarrier®), and vagus nerve block devices (e.g., Maestro® system) Medications: liraglutide (Saxenda®), lorcaserin (BELVIQ®), naltrexone/bupropion extended- release (N/B) (Contrave®), and phentermine/topiramate extended-release (P/T) (Qsymia®) ©Institute for Clinical and Economic Review, 2015 ES3 [Return to Table of Contents] We focused on devices and medications that are approved by the Food and Drug Administration (FDA) for weight loss or are imminently expecting FDA approval as of July 14, 2015. a Note that interventions differ in terms of their approved indications; these differences are noted in the full report. Comparative Effectiveness and Safety Evidence on comparative clinical effectiveness and safety is described separately for bariatric surgery, devices, and medications in this report, and because evidence in patients with a BMI <35 was deemed to be of particular interest, findings for this subpopulation are highlighted separately for each intervention. Of note, although each of the included medications is intended for patients with a BMI ≥30 (or >27 with comorbidities), the mean baseline BMI of patients across medication studies was approximately 36. Availability of subgroup analysis by BMI class is limited. Table ES1 on the next page summarizes the level of certainty in the evidence according to BMI class, as well as by the presence of T2DM, the comorbidity most commonly studied for BMI levels <30 and 30-34.9. Because uncertainties remain with long-term outcomes and durability of clinical benefit, we did not consider the evidence to provide high levels of certainty for any intervention. Across interventions, the evidence is strongest for surgery, as well as for all interventions in patients with a BMI ≥35. A detailed summary of the evidence for each intervention type is described in Section 4 of the full report. a Between July 14, 2015 (the date of the CTAF public meeting) and August 10, 2015 (the date of this report), two additional intragastric balloons were approved by the FDA. ©Institute for Clinical and Economic Review, 2015 ES4 [Return to Table of Contents] Table ES1: Strength of Evidence by BMI Class BMI <30 30-34.9 35-39.99 ≥40 T2DM Yes No Yes No --- --- Bariatric Surgery BPD LAGB RYGB VSG Devices Key IGB No evidence DJBL Low certainty vBloc Moderate certainty Drugs High certainty Liraglutide Lorcaserin N/B P/T Clinical Effectiveness Bariatric Surgery Across a range of procedures, study designs, and duration of follow-up, bariatric surgery results in greater sustained weight loss (on average, 7-8 kg/m2, or 30-40% of total body weight) and resolution of comorbidities (primarily T2DM) than nonsurgical management. These results are limited by a lack of good-quality long-term data on durability of benefit. There is also a lack of both short- and long-term data demonstrating effectiveness of any bariatric surgery procedure for pediatric populations. There is little evidence to distinguish performance of individual bariatric procedures in patients with a BMI ≥35, as LAGB use has declined substantially, BPD±DS is technically complex and performed only at certain centers, and the performance of RYGB and VSG is similar. Evidence on the differential effectiveness and safety of bariatric surgery procedures according to patient characteristics such as age, sex, race/ethnicity, BMI, or psychosocial factors is largely inconclusive. Nevertheless, there is evidence, albeit limited, that surgeon experience, high procedure volume, multidisciplinary care, adherence to pre- and post-operative follow-up, and post-procedure support may positively influence outcomes. ©Institute for Clinical and Economic Review, 2015 ES5 [Return to Table of Contents] For patients with a BMI of 35 and above (with or without clinical comorbidities), we judge there to be moderate certainty of a substantial net health benefit of bariatric surgery compared to nonsurgical management; certainty remains moderate because of a lack of long-term data on durability of benefit. Across the different types of bariatric procedures, we judge there to be moderate certainty of a small net health benefit with BPD±DS, RYGB, and VSG over LAGB, and moderate certainty of a comparable net benefit between RYGB and VSG. For individuals with a BMI 30-34.9, we judge there to be moderate certainty of a small-to-moderate net benefit of bariatric surgery compared to nonsurgical management among patients with T2DM who receive RYGB or LAGB procedures. The evidence base provides low certainty for all other procedures and for any surgery in this BMI range in patients with comorbidities other than T2DM. Our judgment of the net benefit of bariatric surgery relative to nonsurgical management in pediatric populations is ”insufficient”, as we found only one good-quality RCT of 50 patients that evaluated outcomes in this population. Finally, we found no evidence of the effectiveness of bariatric surgery in patients with a BMI <30. Intragastric Balloon (IGB) Temporary IGB insertion was associated with statistically significant improvements in measures of weight change at a median of one year of follow-up relative to conventional approaches (e.g., diet, exercise, behavioral modification). Benefits were modest, however – higher-quality studies showed incremental BMI changes and percentage weight loss of 1-3 points and 4-8%, respectively – and also tended to worsen after balloon removal (typically after six months). There was great variability in study design, duration of follow-up (particularly after balloon removal), and treatment approach (i.e., single vs. multiple balloon insertions). Evidence comparing IGB to bariatric surgery was mixed; several studies showed greater weight loss with IGB vs. LAGB, but differences diminished with greater duration of follow-up. Data are limited on all other outcomes, including resolution of comorbidities and quality of life (QoL), and in populations with a BMI less than 35. Although there is insufficient evidence on the long-term durability of treatment effect or safety of IGB therapy following balloon removal, weight regain appears to be common, and major treatment-related complications are rare. We judge there to be low certainty of a comparable net benefit for temporary IGB insertion relative to either lifestyle intervention or bariatric surgery in patients with a BMI ≥35. In comparison to lifestyle interventions, IGB therapy appeared to provide modest reductions in weight across studies, but data on weight change were mixed after balloon removal, there was variability in balloon duration and number of placements, and there is the potential for harm. Evidence of benefit was truly mixed in available comparisons of IGB to bariatric surgery. Evidence on the benefit of the balloon in patients with a BMI <35 or its impact on any other outcome is insufficient to draw any firm conclusions. ©Institute for Clinical and Economic Review, 2015 ES6 [Return to Table of Contents] Duodenal-Jejunal Bypass Liner (EndoBarrier®) We identified one good- and one fair-quality RCT that compared the DJBL to conventional weight- loss treatment. Neither study found statistically significant differences in reduction of mean BMI between control and DJBL patients. Patients in both studies reduced or discontinued anti-diabetic medications; however, data on comorbidities were not reported for control patients in one study, and statistical testing for differences was not conducted in either study. Complications of the DJBL included device or anchor migration, epigastric pain, and sleeve obstruction; 2.9-20.5% of patients had the device removed prematurely across studies. Finally, enrollment in a large clinical trial of DJBL is currently stopped pending investigation of a higher-than-expected rate of bacterial liver infection. Key questions remain as to the efficacy and safety of the DJBL. Preliminary evidence suggests that there is low certainty that the DJBL has a comparable net health benefit or even a negative net health benefit for weight loss and comorbidity resolution relative to lifestyle interventions in patients with a BMI ≥30. Given this level of certainty, and both published and manufacturer- reported data on potentially serious harms, we judge the current evidence to be insufficient to draw firm conclusions on the DJBL’s effects.b Vagus Nerve Block (Maestro®) We identified two RCTs that examined the vagus nerve block (vBloc), one of which assessed an earlier generation of the Maestro device. After 12 months follow-up in each study, vBloc patients lost 17-24.4% of excess body weight versus 15.9-16.0% excess weight loss (EWL) in control patients. The included studies did not report outcomes related to resolution or improvement in comorbidities. In both studies, 5.4% of patients had the device removed before study end; 3.0-3.7% of patients experienced serious complications related to the device, implantation, or revision. We judge there to be low certainty of a small or comparable net benefit for the vBloc device compared to a sham device in patients with a BMI ≥35. There is low certainty because only two studies examined the vBloc, one of which assessed an earlier generation device that was not FDA- approved. While the ReCharge trial results indicated greater weight loss with vBloc vs. sham, many questions remain, including the impact of the vBloc on comorbidities, whether the device can be safely used for durations longer than 18 months, and whether the benefits demonstrated in the ReCharge trial can be repeated in future studies of the device. We found no evidence of the vBloc’s effects in patients with BMI levels <35. b Subsequent to the July 14, 2015 CTAF public meeting, the manufacturer of the DJBL and the FDA agreed to terminate the US trial due to safety concerns. ©Institute for Clinical and Economic Review, 2015 ES7 [Return to Table of Contents] Medications A number of medications have recently been approved for the treatment of obesity or overweight, and these are described below. Reflecting the populations included in the key clinical studies for these medications (i.e. patients with a BMI between 27 and 45 kg/m2), all are indicated for use in adults with a BMI ≥30 or in adults with a BMI ≥27 who have at least one obesity-related comorbidity. The majority of studies in our sample set compared the medications of interest to placebo plus lifestyle intervention. Liraglutide (Saxenda®) The three studies with labeled dosing for weight loss (3 mg) found that patients lost 6-8% of total body weight after one year of follow-up and were more likely to have normal glucose tolerance relative to those in the placebo group. The most commonly reported adverse effects (AEs) were gastrointestinal disorders including nausea, vomiting, diarrhea, and constipation. Overall AE rates ranged from 80.3-95.7% across studies, and discontinuation due to AEs ranged from 7.5-9.9%. We judge there to be moderate certainty that liraglutide (3 mg) plus lifestyle modification versus lifestyle modification alone provides a small net benefit in patients with a BMI ≥27, due to modest levels of incremental weight loss and comorbidity resolution. Certainty was judged to be moderate because while three good quality studies reported consistent results, uncertainty remains around the durability of liraglutide’s benefits and safety over long-term follow-up (i.e., >2 years). Lorcaserin (BELVIQ®) In three good-quality RCTs, reductions in total body weight were modest, ranging from 4.5-5.8% among lorcaserin recipients, compared to a 1.5-2.8% mean decrease among those taking the placebo (p<0.001 for lorcaserin vs. placebo in all studies). A single study reported outcomes related to comorbidity status and found 50.4% of lorcaserin patients versus 26.3% of placebo patients achieved a hemoglobin A1c (HbA1c) <7% (p>0.001).5 Discontinuation of lorcaserin from drug- related AEs occurred in 4.3-8.6% of patients across studies, and approximately 80% of study participants experienced any AE. We judge there to be moderate certainty of a small net benefit of lorcaserin (10 mg, administered once or twice daily) plus lifestyle modification versus lifestyle modification alone for weight loss in patients with a BMI between 27-45. There is moderate certainty because while three good quality studies reported consistent weight-loss results, only one study reported outcomes related to improvement of obesity-related comorbidities. Furthermore, we found a single study that stratified outcomes according to various BMI classes. Finally, there is uncertainty around the durability of lorcaserin’s benefits as well as safety over longer-term follow-up (i.e., >2 years). ©Institute for Clinical and Economic Review, 2015 ES8 [Return to Table of Contents] Naltrexone/Bupropion (Contrave®) Across four available RCTs, participants who received standard doses of naltrexone (extended- release, 32 mg daily) combined with bupropion (extended-release, 360 mg daily) lost 5.0-9.3% of total body weight after 56 weeks follow-up; patients who received a placebo lost 1.2-5.1% of total body weight. A single RCT reported outcomes related to improvement of comorbidities and found that 44.1% of patients taking N/B achieved a target HbA1c <7% compared to 26.3% of placebo patients.6 Discontinuation of N/B from adverse effects occurred in 19.5-29.3% of patients receiving standard-dose N/B across studies, and 83.1-90.4% of patients experienced any AE. We judge there to be moderate certainty of a small net benefit associated with N/B plus lifestyle intervention versus lifestyle intervention alone in patients with a BMI between 27 and 45. There is moderate certainty of benefit because four good-quality studies showed consistent weight loss relative to comparator treatment; however, benefits were modest. In addition, there is uncertainty around the durability of benefit from N/B as well as safety over longer-term follow-up, since we found no studies with follow-up greater than one year. We also judge the evidence to be of low certainty for a small net benefit in patients with BMI 30-34.9 and T2DM based on the results of a single RCT in this population. Phentermine/Topiramate (Qsymia®) We identified a total of eight good- or fair-quality reports from five RCTs that compared phentermine/topiramate extended release combination therapy (P/T) to placebo or to phentermine or topiramate monotherapy. In these trials, patients receiving the recommended dose combination (7.5/46 mg) lost 7.8-8.5% of total body weight (vs. 1-2% for placebo), while the range for those receiving a higher dose (15/92 mg) was 9.2-10.9%. Patients who received any dose of P/T experienced greater improvement in obesity-related comorbidities such as T2DM, hypertension, and sleep apnea. We found no studies that had a patient population with mean BMI <35. Overall, 91-95% of patients experienced one or more AEs, and 1.3-16.0% discontinued P/T due to AEs. We judge there to be moderate certainty of a small net benefit of P/T plus lifestyle intervention versus lifestyle intervention alone in improving weight loss in patients with BMI levels 27-45. As with the other medications, certainty was moderate because of the modest levels of weight loss and comorbidity resolution observed. In addition, there is uncertainty around the durability of benefit from P/T as well as safety over longer-term follow-up (i.e., >2 years). Clinical Effectiveness Summary Our review of the evidence suggests that obesity treatment is a dynamic clinical area, and that efforts to innovate have grown along with the prevalence of the clinical problem itself. That said, ©Institute for Clinical and Economic Review, 2015 ES9 [Return to Table of Contents] there are distinct differences and challenges with the level and types of evidence available for surgical procedures, devices, and medications. The evidence base for surgery is the most mature, and comparative data on procedures has accelerated in recent years (two-thirds of the publications in our sample were published after 2011). For patients with class III obesity (BMI ≥40) or class II obesity (BMI 35-39.9) with comorbidities, bariatric surgery consistently outperforms nonsurgical treatment approaches in terms of sustained weight loss (7-8 kg/m2, or 30-40% of total body weight) and resolution of comorbidities. There are clear tradeoffs to consider, as each individual procedure confers its own risks of complication and reoperation. However, the choice of procedure has essentially narrowed to gastric bypass or sleeve gastrectomy; current trends suggest that gastric banding has fallen out of favor due to lower effectiveness, and biliopancreatic diversion is a technically demanding and much less-frequently performed operation. Nevertheless, our conclusions on the benefits of any bariatric procedure are challenged by a general lack of high-quality long-term data on effectiveness, safety, weight regain, and return of comorbidities. In patients at lower levels of BMI (30-34.9), evidence for bariatric surgery is accumulating among patients with T2DM and consistently shows substantially greater levels of diabetes improvement vs. intensive lifestyle and medical management. Lack of long-term data is also a challenge with these studies, and comparative data for bariatric surgery in patients with lower BMI levels and types of comorbidities other than T2DM are lacking. Regardless of surgical approach or outcome of interest, there are data to suggest that there are core bariatric program components associated with success. Unfortunately, the best data are on elements (e.g., surgeon volume, learning curve) that are important for any surgical procedure, and evidence on other components (e.g., care team members, accreditation, psychological testing, pre- and/or post-operative support) are mixed at best. By contrast, information for devices is truly emerging. For one, the first FDA-approved device among the three types of interest in our review is the Maestro vBloc system; even with this approved device, benefits are modest (mean difference of 8.5% of body weight lost vs. sham device) in one RCT, and no benefit was observed in another (with a prior generation device). More comparative data are available on temporary intragastric balloon insertion, but benefits are again modest (4-8% total weight loss), there is a high rate of early removal due primarily to adverse events, and there are very little reliable data on trends in body weight after balloon removal or comparisons among the different balloon designs. Limited available data on the EndoBarrier DJBL ©Institute for Clinical and Economic Review, 2015 ES10 [Return to Table of Contents] show no material benefit over control therapy, and the current registrational clinical trial has been suspended due to safety concerns. c Finally, all four of the newer weight-loss medications have been FDA-approved, but benefits are again relatively modest in comparison to conventional weight-loss management. Across all four medications, total weight loss was 3-7% higher than achieved with placebo or active comparator therapy. In addition, there are limited data on resolution or improvement in comorbidities, lack of information on long-term weight trends, high rates of discontinuation in many studies, and not- inconsequential concerns about potential to harm (two of the four are scheduled substances). Economic Outcomes Published evidence accumulated to date on care value (see Section 5 of the report for details) suggests that bariatric surgery meets commonly-accepted thresholds for cost-effectiveness in comparison to standard care across multiple BMI classes, time horizons, and procedure types. By contrast, there are no currently published estimates of cost-effectiveness for newly-approved devices and drugs such as vBloc and Contrave®. To better inform the comparisons of interest in our review, we developed a simulation model to compare the care value of four bariatric surgery procedures, the Maestro® vBloc system, and N/B extended-release (Contrave®) to conventional weight-loss management for all individuals with obesity (BMI ≥30), as well as for classes defined by BMI range (i.e., 30-34.9, 35-39.9, and ≥40). Over the course of one year, across all levels of BMI and procedure type, we found that all interventions improved BMI levels but were subject to varying levels of complications and AEs, as well as increased overall costs. Over a 10-year timeframe, each intervention also resulted in improved quality-adjusted survival due primarily to the beneficial effects of lower weight on QoL. The more prominent weight loss from surgery also lowered obesity-related costs, offsetting the initial costs of surgery by as much as 30% over 10 years. Cost-effectiveness estimates for bariatric surgery over 10 years ranged from approximately $24,000 to $63,000 per quality-adjusted life year (QALY) gained vs. conventional treatment, which is within commonly-accepted thresholds for cost-effectiveness (i.e., $50,000-$100,000 per QALY gained). These findings were robust to a range of sensitivity analyses, including elimination of mortality benefit for bariatric surgery and complete weight regain five years after surgery. Importantly, while the most favorable results were seen in patients with BMI ≥40 due to greater weight loss (and c As noted previously, subsequent to the July 14, 2015 CTAF public meeting, the manufacturer of the DJBL and the FDA agreed to terminate the US trial due to safety concerns. ©Institute for Clinical and Economic Review, 2015 ES11 [Return to Table of Contents] corresponding gains in QoL), surgery produces cost-effectiveness ratios within the commonly- accepted range among those with a BMI level of current policy interest (30-34.9), with findings ranging from $43,000 - $63,000 per QALY gained vs. conventional treatment. In contrast, the much more modest weight loss achieved with the vBloc device and N/B pharmacotherapy, coupled with their high implantation and ongoing therapy costs, respectively, resulted in much higher cost-effectiveness ratios (>$100,000 per QALY vs. conventional treatment). Results were more favorable when these treatment options were considered “in sequence” with bariatric surgery for those failing initial treatment, in particular for a “drug first” regimen in which those with successful weight loss at one year continued to receive medication while remaining patients were able to receive surgery after some initial weight loss. Given the emerging nature of the evidence on devices and the modest benefits afforded by newer medications, we opted to focus our health-system value analysis on the use of bariatric surgery in patients with a BMI of 30-34.9. The potential budgetary impact of adopting vertical sleeve gastrectomy in 10-25% of all patients in this BMI class is substantial – between $27 and $66 PMPM, a 5-12% increase over the current total Medi-Cal expenditure rate of $552. When availability is restricted to the subset of these patients who have diabetes, however, the budgetary impact declines to approximately $3-$7 PMPM, a 0.5-1.3% increase over current total expenditures. It is not known how many people in this population would opt for surgery, even if offered. While findings from both our evidence review and assessment of comparative value are intriguing, gaps in available evidence continue to pose challenges in understanding the appropriate sequence of treatments and relevant candidates for each option along the treatment continuum. There is a need for rigorous long-term and comparative study to better understand the durability of weight loss and comorbidity resolution, as well as patterns of harm over longer durations of follow-up. CTAF Votes on Comparative Clinical Effectiveness and Value During public meetings, the CTAF Panel deliberates and votes on key questions related to the systematic review of the clinical evidence, an economic analysis of the applications of the medical technologies or treatments under examination, and the supplementary information presented. Because any judgment of evidence is strengthened by real-life clinical and patient perspectives, subject matter experts are recruited for each meeting and provide input to the CTAF Panel, serve as a resource to the Panel during their deliberation, and help form recommendations with the Panel on ways the evidence can be applied to policy and practice. At each meeting, after the CTAF Panel vote, a policy roundtable discussion is held with the Panel, clinical experts, and representatives from provider groups, payers, and patient groups. ©Institute for Clinical and Economic Review, 2015 ES12 [Return to Table of Contents] At a July 14, 2015 public meeting, the CTAF Panel discussed issues regarding the application of the available evidence to help patients, providers, and payers address important questions related to multiple surgical procedures, devices, and medications available to treat patients who are overweight or obese. Following the evidence presentation and public comments, the CTAF Panel voted on key questions concerning the comparative clinical effectiveness and comparative value of various treatments. The CTAF Panel used a value assessment framework (described in more detail in Section 6 of this report) with four different components of care value, which is a judgment comparing the clinical outcomes, average per-patient costs, and broader health effects of two alternative interventions or approaches to care. Once they made an overall assessment of care value as low, intermediate, or high, the CTAF Panel then explicitly considered the affordability of several treatments for obesity in assessing health system value. Health system value is a judgment of the affordability of the short- term budget impact that would occur with a change to a new care option for all eligible patients, assuming the current price and payment structure. The components of care value and health system value are depicted in Figures ES2 and ES3 below. Figure ES2. Care Value Framework Figure ES3. Health System Value Framework ©Institute for Clinical and Economic Review, 2015 ES13 [Return to Table of Contents] In assessing health system value, the CTAF Panel reviews estimates of the potential budget impact for a change in care and votes on whether interventions represent a “high,” “intermediate,” or “low” health system value from the generalized perspective of a state Medicaid program. CTAF Panel Voting Results Clinical Effectiveness Surgical procedures and devices 1. For adult patients with a BMI of 30.0-34.9 and Type 2 diabetes, is the evidence adequate to demonstrate that the net health benefit of bariatric surgery is greater than that of conventional weight-loss management? CTAF Panel Vote: 11 Yes (100%) 0 No (0%) 2. For adult patients with a BMI 30.0-34.9 who do not have Type 2 diabetes, is the evidence adequate to demonstrate that the net health benefit of bariatric surgery is greater than that of conventional weight-loss management? CTAF Panel Vote: 1 Yes (9%) 10 No (91%) 3. For adult patients with a BMI ≥35, is the evidence adequate to demonstrate that RYGB and VSG have a comparable net health benefit? CTAF Panel Vote: 11 Yes (100%) 0 No (0%) 4. For adult patients with a BMI ≥35, is the evidence adequate to demonstrate that the net health benefit of the Maestro vBloc system is greater than that of conventional weight-loss management? CTAF Panel Vote: 6 Yes (55%) 5 No (45%) Drugs 5. For adult patients with a BMI ≥30 or BMI ≥27 with an obesity-related comorbidity, is the evidence adequate to distinguish the net health benefit among d the four drugs reviewed? CTAF Panel Vote: 0 Yes (0%) 11 No (100%) d The CTAF Panel understood this question to mean distinguishing among or between these four medications. ©Institute for Clinical and Economic Review, 2015 ES14 [Return to Table of Contents] Value Surgical procedures and devices 6. [If yes to question 1 – adult patients with a BMI of 30.0-34.9 and Type 2 diabetes] Given the available evidence, what is the care value of bariatric surgery vs. conventional weight-loss management? CTAF Panel Vote: 2 Low (18%) 6 Intermediate (55%) 3 High (27%) 7. [If yes to question 1] Given the available evidence, what is the overall health system value of bariatric surgery for a state Medicaid program? CTAF Panel Vote: 3 Low (27%) 5 Intermediate (45%) 3 High (27%) 8. [If yes to question 2 – adult patients with a BMI 30.0-34.9 who do not have Type 2 diabetes] Given the available evidence, what is the care value of bariatric surgery vs. conventional weight-loss management? Vote Not Taken 9. [If yes to question 2] Given the available evidence, what is the overall health system value of bariatric surgery for a state Medicaid program? Vote Not Taken 10. [If yes to question 4 – adult patients with a BMI ≥35] Given the available evidence, what is the care value of the Maestro vBloc system vs. conventional weight-loss management? CTAF Panel Vote: 8 Low (73%) 3 Intermediate (27%) 11. [If yes to question 4] Given the available evidence, what is the overall health system value of the Maestro vBloc system for a state Medicaid program? CTAF Panel Vote: 10 Low (91%) 1 Intermediate (9%) Policy Roundtable Discussion and Key Policy Recommendations Coverage policy 1. The CTAF panel confirmed by unanimous vote that there is accumulating evidence of positive longer-term clinical outcomes for bariatric surgery in patients who have a body mass index (BMI) of 30.0-34.9 and type 2 diabetes. Payers should consider expanding coverage for bariatric surgery to this group of patients based on its comparative clinical effectiveness. ©Institute for Clinical and Economic Review, 2015 ES15 [Return to Table of Contents] 2. Given the rapid development of therapies for obesity, more frequent reviews of payer coverage policies are required and should be based on systematic reviews and rigorous technology assessments. 3. Bariatric surgery is rarely the initial treatment option for patients with obesity; however, it should be considered when weight-loss management approaches and medical treatments have failed or have had insufficient benefit. Given the risks of surgery, many patients will wish to use conventional or intensive weight-loss management approaches, medical treatment options, and/or less invasive surgical procedures. Payers should structure coverage policies to help guide clinicians and patients through the available lifestyle interventions, drugs, devices, and surgical procedures as they select treatments most appropriate for the patient. Clinical care 4. Obesity is not a surgical disorder; rather, it is a chronic condition that benefits from care that is coordinated across multiple specialty areas. As part of their care, patients with obesity should be considered for treatment by primary care physicians, bariatricians, e nutritionists, mental health and/or addiction experts, and other medical specialists who can help ensure successful treatment outcomes. If bariatric surgery is the treatment selected, surgical teams should work closely with primary care providers and bariatricians for pre-surgery consultation and post-surgery follow-up. Clinicians should give additional consideration to the physical, emotional, social, and environmental factors that contribute to a patient’s health, particularly for underserved and vulnerable populations, so that the obesity treatment chosen is most likely to be successful. 5. Health systems and provider groups should use more sophisticated obesity classification systems beyond simple BMI. Electronic documentation will allow for stratification of the patient population by obesity class; health systems and provider groups can use this information in planning for clinical capacity and to more effectively target treatment resources to the patient population. 6. Because of the high prevalence of obesity and the discomfort both providers and patients experience in discussing it, clinicians should develop new materials or use existing materials to engage with patients in discussions about the health effects of and treatment options for obesity. Payment 7. Fee-for-service payment does not adequately support treatment of obesity, which is a chronic condition requiring multidisciplinary, longitudinal care. Provider groups, payers, and e A bariatrician is a licensed physician who has received special training in bariatric medicine, which is the medical treatment of overweight and obesity and its associated conditions. ©Institute for Clinical and Economic Review, 2015 ES16 [Return to Table of Contents] other stakeholders should consider developing and implementing alternative payment structures for obesity care, and then evaluate whether or not they improve outcomes and value. Examples of alternative approaches are bundled payments for an episode of care, a “warranty” on surgical procedures, and risk sharing between device manufacturers and health plans. Measures that assess quality of care for obesity (including patient-reported outcomes) should be coupled with payment to provide incentives for high quality, efficient care. Prevention 8. Organizations engaged in efforts to prevent and reduce obesity in the overall population should continue this important work. Such programs are often part of broader social policy changes that address violence, promote access to healthy foods and safe places to exercise, and seek healthier environments in which people can live, work, and play. Health care providers and systems should serve as an example and lead by creating healthier environments for their employees, patients, and visitors. Complications of weight gain threaten to undo the health and mortality improvements attained through improved treatments for heart disease and cancer. The magnitude of the problem requires evidence-based public health approaches to prevention as well as medical care to aid people seeking to reduce their weight, risk, and disease burden. The many millions of people currently affected are seeking approaches that meet their current health needs, desire to change, and risk tolerance. As this report demonstrates, pharmaceutical, medical device, and surgical innovators, as well as researchers, are actively engaged in addressing this pressing need. The clinical evidence and the value recommendations summarized in this report should guide policymakers, patients, payers, and providers in their efforts to address obesity. ©Institute for Clinical and Economic Review, 2015 ES17 [Return to Table of Contents] Introduction A recent estimate found that two-thirds of Americans are overweight or obese.2 The health effects of obesity are myriad and include the development of type 2 diabetes mellitus (T2DM), hypertension, cardiovascular disease, cancer, high blood pressure, and sleep apnea. Obesity and its sequelae generate an estimated $147 billion annually in health care costs in the US alone.3 The complexity involved in managing excess weight may affect both patient candidacy for certain treatment options as well as adherence to lifestyle changes necessary to sustain weight loss and improve health outcomes.4 As a result, there is great clinical interest in treatments that may be used for patients at various levels of obesity, as well as in interventions that promote better adherence to lifestyle change. However, success rates from lifestyle modifications alone have been modest at best. A variety of options await the patient seeking or needing to lose weight beyond conventional weight-loss approaches such as general lifestyle counseling, personal dieting, exercise regimens, and commercial weight loss plans. Bariatric surgery remains a mainstay for those with severe or morbid obesity, and it is also being explored in patients with lower levels of obesity (body mass index [BMI] <35 kg/m2); new devices have emerged or are being tested to suppress appetite and/or reduce food intake; and several medications have recently been approved by the FDA specifically for weight loss. The availability of treatments of differing intensity and invasiveness raises a number of questions, however, including 1) their effects on patients at various levels of obesity, 2) their proper place in the treatment continuum for any given group of patients, and 3) their performance and durability of treatment effects over the long term. The purpose of this report for CTAF, therefore, is to examine the comparative clinical effectiveness and comparative value of surgical-, device-, and medication-based treatments in relation to conventional weight-loss management as well as across intervention types. Special attention is also paid to studies conducted in individuals at lower levels of BMI (i.e., 25-35 kg/m2), a key area of uncertainty and controversy. ©Institute for Clinical and Economic Review, 2015 1 [Return to Table of Contents] 1. Background 1.1 Background Historically, options for treating obesity have been limited to lifestyle modifications such as dietary changes and exercise, as well as the use of weight-loss medications and dietary supplements, all of which have struggled to show evidence of persistent and long-term weight reduction. In addition, earlier-generation medications were often found to pose significant health risks of their own.7 More recently, additional treatment options have become available for managing excess weight. Most prominently, use of surgical interventions has become more widespread. The term “bariatric surgery” refers to a collective group of procedures that involve modifications to the digestive system that promote weight loss; procedures currently performed in US settings include gastric bypass, gastric banding, sleeve gastrectomy, and biliopancreatic diversion (with or without duodenal switch).8 Most patients are able to undergo these procedures via laparoscopy. The choice of procedure depends primarily on the severity of obesity, the presence of comorbid conditions, the experience of the surgeon, and the patient’s individual preferences or other contraindications.9 In certain settings and populations, bariatric surgical procedures have resulted in substantial reductions in body weight and remission of some obesity-related comorbidities (e.g., hypertension, diabetes). Long-term observational studies also suggest that bariatric surgery may reduce the risk of newly developing these comorbidities, an important consideration in adolescents or adults without longstanding obesity.10,11 Early use of the procedures focused on individuals with class III obesity (BMI ≥40.0 kg/m2) or class II obesity (BMI 35.0-39.9 kg/m2) with at least one obesity-linked comorbid condition. Subsequent studies have been conducted in individuals at lower levels of BMI, which has led to regulatory approval specific to this population: in 2011, the Food and Drug Administration (FDA) approved the use of laparoscopic adjustable gastric banding (LAGB) in patients with class I obesity (BMI 30.0-34.9) and at least one obesity-linked condition.12 As a result, clinical interest remains high in expanding the use of bariatric surgery to a broader set of individuals. Questions remain, however, regarding the performance of these procedures in patients with higher levels of obesity versus those with BMI ≤35 as well as the health-system impact given the higher prevalence of moderate obesity versus severe/morbid obesity. An additional and considerable challenge to the potential expansion of bariatric surgery is a lack of long-term data on the safety and effectiveness of these procedures. A recent systematic review attempted to quantify the number of studies with sufficient long-term follow-up and found that only 29 (2.6%) of 1,136 long-term studies maintained at least 80% of the original sample after two or more years.13 In addition, even those studies with sufficient sample retention were often missing data on weight ©Institute for Clinical and Economic Review, 2015 2 [Return to Table of Contents] changes and comorbidity remission. Long-term follow-up is perhaps even more critical with bariatric surgery than in other clinical areas, as weight regain is not an uncommon phenomenon.4 There are also specific risks associated with bariatric procedures, which may include bowel obstruction, development of gallstones or hernias, stomach perforation and ulcer, “dumping syndrome” (diarrhea and other related symptoms caused by rapid movement of undigested food to the small bowel), and in some cases, death.14 Additional surgeries may be required as part of a multi-phase procedure (as with biliopancreatic diversion), to implement an entirely new treatment modality, remedy a complication, or reverse the procedure altogether if complications are life- threatening.15 Surgical revisions comprise about 6% of all weight loss surgeries performed annually in the US.16 Because of the complexities, risks, and uncertainties that remain with even well-accepted bariatric surgical procedures, there is continued interest in the development of less-invasive approaches to weight loss. New medications have recently been approved specifically for weight loss, including serotonin receptor agonists for appetite suppression, anti-diabetic medications with known weight- loss effects, and fixed-dose combinations such as phentermine/topiramate (P/T) and naltrexone/ bupropion (N/B) that have been studied and/or approved individually for weight loss.17 Several devices are also under active study, including a recently-approved vagus nerve-blocking system, a plastic barrier to reduce food absorption, and temporary intragastric balloons (IGBs) to reduce food intake and promote satiety.18 Because the role of each type of intervention across multiple levels of obesity remains uncertain, this review focuses on the evidence for comparative clinical effectiveness and comparative value for each intervention of interest in relation to conventional weight-loss management as well as available head-to-head studies across intervention types. Interventions of interest for this review are described in further detail in the following section. ©Institute for Clinical and Economic Review, 2015 3 [Return to Table of Contents] 1.2 Treatment Strategies of Interest Bariatric Surgery Roux-en-Y Gastric Bypass Roux-en-Y gastric bypass (RYGB) is the most commonly performed bariatric procedure worldwide.19 RYGB can be performed laparoscopically, robotically, or through a traditional open approach and typically lasts one-and-a-half to four hours. Figure 1: Roux-en-Y Gastric During the procedure, a surgeon separates the upper and lower Bypass portions of the stomach by creating a small pouch in the top of the stomach. The pouch is approximately two tablespoons in volume, and it is intended to restrict food intake and promote satiety after Source: http://asmbs.org/patients/bariatric-surgery-procedures small amounts of food are consumed.20 The remaining portion of the stomach is bypassed by dividing the small intestine into two limbs: the Roux limb and the biliopancreatic limb. The Roux limb, which is also referred to as the jejunum, is the middle section of the small intestine. This limb is connected to the gastric pouch so that food bypasses both the lower portion of the stomach and the beginning portion of the small intestine. The biliopancreatic limb, which contains the beginning part of the small intestine, is reconnected below the Roux limb so that digestive juices from the remnant stomach may flow to the remaining intestine. The intersection of the biliopancreatic and Roux limbs forms the shape of a “Y,” giving this procedure its name. The bypass causes malabsorption, in which patients absorb fewer calories and nutrients from food. After RYGB, patients remain in the hospital for one or two nights and recover within approximately one month. Possible complications include bleeding, pouch ulcers, dehydration, leakages, internal hernias, blockages, blood clots, and infection. “Dumping syndrome” can occur when food and digestive juices move to the small intestine at an abnormally fast pace. Patients are also required to take nutritional supplements for the remainder of their lives and monitor their intake of carbohydrates to avoid gastric discomfort, vomiting, and diarrhea. ©Institute for Clinical and Economic Review, 2015 4 [Return to Table of Contents] Biliopancreatic Diversion with or without a Duodenal Switch (BPD±DS) Biliopancreatic diversion is commonly performed on individuals with a BMI of 50 kg/m2 or greater.21 BPD first involves the removal of about 70% of the stomach in order to reduce acid production. The remaining portion of the stomach is larger than the pouch formed by RYGB, which allows the patient to ingest more food before feeling satiated.22 The small intestine is then divided and one end is attached to the new stomach pouch, creating an "alimentary limb" through which food travels with limited calorie and nutrient absorption. Digestive enzymes travel through a biliopancreatic limb that is connected near the end of the small intestine, meeting up with ingested food and forming a common limb. While the resulting anatomy of this procedure is similar to that of RYGB, the intestine length from stomach to colon is much shorter in BPD.23 Figure 2: Biliopancreatic The duodenal switch (DS) is a modification of the Diversion/Duodenal Switch biliopancreatic diversion. Instead of removing the lower half of the stomach (as with the BPD), the DS cuts the Source: http://asmbs.org/patients/bariatric-surgery-procedures stomach vertically and leaves a tube of stomach that empties into a very short (2-4 cm) segment of duodenum.23 Whereas the BPD involves forming a connection between the stomach and the intestine, the DS preserves the duodenum, attaching this upper portion of the small intestine to the lower portion of the small intestine. Patients typically remain in the hospital for four to seven nights after BPD and take three to four weeks to recover. As with RYGB, BPD±DS is a malabsorptive procedure that requires patients to remain on vitamin and mineral supplements for the remainder of their lives. Possible complications may include kidney stones, ulcers, internal bleeding, infection, blood clots, hernias, dumping syndrome, and death. Additionally, patients are prone to diarrhea and foul smelling gas, with an average of three to four loose bowel movements a day. ©Institute for Clinical and Economic Review, 2015 5 [Return to Table of Contents] Laparoscopic Adjustable Gastric Banding (LAGB) Figure 3: Laparoscopic Adjustable Gastric Banding Adjustable gastric banding is a purely restrictive Source: http://asmbs.org/patients/bariatric-surgery-procedures procedure that induces weight loss by restricting food intake. During the procedure, a band containing an inflatable balloon is fixed around the upper part of the stomach. This creates a small stomach pouch above the band with a narrow opening into to the rest of the stomach.21 The band can be adjusted by injecting or removing fluid from the balloon through a port under the skin of the abdomen. After surgery, some patients spend a night at the hospital while others recover at home. After one week, patients can return to work, provided it is not too physically taxing, and are usually fully recovered within one to two weeks. Unlike other bariatric procedures, LAGB is a reversible procedure with a lower risk of nutritional deficiencies and lower mortality. However, optimal results require frequent follow-up visits for band adjustments. Complications can include hemorrhage, port infection, band infection, obstruction, nausea, vomiting, band erosion into the stomach, and esophageal dilation. Figure 4: Vertical Sleeve Vertical Sleeve Gastrectomy (VSG) Gastrectomy Vertical Sleeve Gastrectomy (VSG) can be performed as part of a two-staged approach to surgical weight loss or as a Source: http://asmbs.org/patients/bariatric-surgery-procedures stand-alone procedure. Patients who have a very high BMI, are at high risk for surgical complications from longer procedures, have an excessively large liver, or have extensive scar tissue are considered possible candidates for sleeve gastrectomy.24 Patients sometimes return to the hospital to undergo RYGB as a second stage procedure after VSG. Similar to BPD±DS, 60-75% of the stomach is removed during VSG, leaving a narrow gastric “tube” or “sleeve.”24 This small remaining “tube” cannot hold as much food and produces less of the appetite-regulating hormone ghrelin, lessening a patient’s desire to eat. ©Institute for Clinical and Economic Review, 2015 6 [Return to Table of Contents] If conducted laparoscopically, sleeve gastrectomy requires an overnight hospital stay, and recovery time is approximately one to two weeks. VSG is not a purely malabsorptive procedure, so there is no requirement for lifetime nutritional supplementation. Potential complications include bleeding, infection, injury to other organs, and leakage from the staple line that divides the stomach.24 Devices EnteroMedics vBloc®/Maestro System® The newly FDA-approved Maestro system is a subcutaneous implant that generates an intermittent electrical pulse that blocks the vagus nerve, the primary nerve regulating the digestive system, in order to reduce feelings of hunger and promote earlier feelings of satiety. Pulse frequency and intensity can be modified externally by the physician.25 The device has been approved for adults with a BMI of 40 to 45 kg/m2 or a BMI of 35 and greater accompanied by at least one other obesity- related comorbidity. Adverse events reported during a recent clinical study of the Maestro system included complications related to intra-abdominal surgery, nausea and vomiting, implant site pain, heartburn, difficulty swallowing, and intestinal gas.26 The hospital list price for the implant kit is $17,500.27 EndoBarrier® The EndoBarrier is intended to function similarly to RYGB without involving fully-invasive surgery. The procedure is typically performed in less than an hour and consists of passing a thin plastic sleeve via the mouth to the small intestine where it is fixed in place by a metal anchor. The sleeve lines the first 60 cm of the small intestine, causing food to be absorbed further down in the intestine. Once implanted, the EndoBarrier influences certain gastrointestinal hormones that play a role in insulin sensitivity, glucose metabolism, and satiety.28 The device is being investigated for its effects on both weight loss and T2DM. Complications reported in a clinical study included esophageal and pharyngeal tears, sleeve migration, anchor dislocation, sleeve obstruction, and abdominal pain. Importantly, while the EndoBarrier has been approved for use in Europe, the regulatory approval process in the US has been delayed due to a higher-than-expected rate of bacterial liver infection among patients enrolled in the Phase III trial of the device.29 As a result, enrollment in the trial has been suspended pending a safety investigation. ©Institute for Clinical and Economic Review, 2015 7 [Return to Table of Contents] Intragastric Balloon (IGB) There are several intragastric balloons undergoing clinical development, although none were FDA- approved as of July 14, 2015. f These include saline balloons such as the BioEnterics® Intragastric Balloon (BIB® or “Orbera”), g the Silimed® Intragastric Balloon (“BIS”), the Spatz3 Adjustable Balloon System, and the ReShape Medical® Intragastric balloon, h as well as the air-based Helioscopie Heliosphere® Bag System. Unlike earlier-generation balloon systems, which were intended to be permanent, these systems involve temporary insertion, usually for no longer than six months. The purpose of the balloon is to provide patients with an opportunity to adjust eating habits and make other lifestyle or behavioral changes, or to promote weight loss prior to bariatric surgery. The balloons are inserted in the stomach via endoscopy and then filled with saline or air; removal is also performed endoscopically. The space filled by the balloon is intended to promote earlier feelings of satiety and thereby reduce portion sizes. Reported complications include balloon migration, infection, and bowel obstruction, as well as nausea and vomiting.30 Medications A number of medications have recently been approved for the treatment of obesity or overweight, and these are described below. Reflecting the populations included in the key clinical studies for these medications, all are indicated for use in adults with a BMI ≥30 or in adults with a BMI ≥27 who have at least one obesity-related comorbidity. Liraglutide (Saxenda®) Liraglutide is an injectable glucagon-like peptide-1 (GLP-1) agonist that was previously approved for glycemic control in T2DM under the brand name Victoza®.17 The version indicated for weight loss involves a higher daily dose (3 mg/day vs. a maximum of 1.8 mg/day for glycemic control). The drug is self-injected daily.31 Side effects of liraglutide include nausea, vomiting, constipation, and diarrhea, and the medication has also been found to slow gastric emptying and decrease the rate and/or extent of absorption of other medications.17 Finally, thyroid C-cell hyperplasia with liraglutide has been reported in animal studies, and the medication carries a “black box” warning because of this.31 The annual cost of liraglutide is estimated to be about $12,000.32 f Between July 14, 2015 (the date of the CTAF public meeting) and August 10, 2015 (the date of this report), two additional intragastric balloons were approved by the FDA. g Received FDA approval on August 5, 2015. h Received FDA approval on July 28, 2015. ©Institute for Clinical and Economic Review, 2015 8 [Return to Table of Contents] Lorcaserin (BELVIQ®) Lorcaserin is a selective serotonin 2C agonist with appetite-suppressant effects. Because of its potential to affect cognition and attention levels, lorcaserin is listed as a Schedule IV controlled substance.33 The recommended dose is 10 mg orally twice daily.34 Adverse effects of lorcaserin have included headache, nausea, and dizziness, as well as episodes of euphoria and cognitive/attention disturbance. The label also carries warnings concerning serotonin syndrome or neuroleptic malignant syndrome if lorcaserin is co-administered with other serotonergic agents as well as a general warning regarding possible development of valvular disorders.34 The annual cost of lorcaserin is about $2,400.32 Naltrexone/Bupropion (Contrave®) (N/B) Contrave is a fixed-dose combination of the antidepressant and smoking-cessation agent bupropion as well as the opioid receptor antagonist naltrexone. Although Contrave’s precise neurochemical effects are not yet fully understood, studies suggest that naltrexone and bupropion regulate food intake through the hypothalamus (appetite regulation center) and mesolimbic dopamine circuit (reward system).35 Each tablet contains 8 mg naltrexone and 90 mg bupropion; dosage escalates over the first month, culminating in a daily dose of four tablets from week 4 onward.35 Side effects include nausea, vomiting, headache, constipation, dizziness, and dry mouth. The package insert also contains a “black box” warning regarding suicidal thoughts and neuropsychiatric reactions, both of which have been reported with bupropion use for smoking cessation.35 The annual cost of N/B is about $2,400.32 Phentermine/Topiramate (Qsymia®) (P/T) Qsymia is a fixed-dose combination of the appetite suppressant phentermine and the anti-seizure medication topiramate.17 It is a Schedule IV controlled substance because of the abuse potential of phentermine.36 The medication is available in four dosage strengths, and a dose-escalation schedule is recommended; the maximum daily dose includes 15 mg of phentermine and 92 mg of topiramate in a single tablet.37 Side effects may include dry mouth, constipation, disordered taste, paresthesia, and insomnia at higher doses; cognition and attention disturbances have also been reported.17 While there are no black-box warnings, the use of Qsymia is not recommended in patients with underlying cardiovascular disease.17 Depending on dosage strength, the annual cost of Qsymia is estimated to range from $2,000 - $2,400.32 ©Institute for Clinical and Economic Review, 2015 9 [Return to Table of Contents] 1.3 Public and Representative Private Insurer Coverage Policies To understand the insurance landscape for bariatric surgery, weight loss devices, and weight loss drugs, we reviewed the publicly available coverage policies and formularies of the Centers for Medicare & Medicaid Services (CMS), California Department of Health Care Services (DHCS), Aetna, Anthem, CIGNA, Humana, UnitedHealthcare (UHC), Health Net, and Blue Shield of California (BSCA). Overall, public and private payers generally cover RYGB, BPD±DS, LAGB, and VSG for patients with a BMI ≥40, or ≥35 with an obesity-related comorbidity. There is some variation among payers as to which comorbidities are qualifying illnesses for patients with a BMI ≥35, but T2DM, cardiovascular disease, cardiopulmonary disease, and hypertension are broadly included. With three exceptions, payers require that patients seeking bariatric surgery have failed prior conservative management, which typically includes documented participation in a physician-supervised diet and exercise program for three to six consecutive months out of the past 18 to 24 months. Humana and Medicare require that patients have failed prior medial management, and Health Net requires patient failure of a reasonable, documented weight-loss attempt. In addition, all payers except for CMS and Health Net require that candidates for bariatric surgery undergo a pre-operative psychological screening to ensure that the patient is capable of providing informed consent and is able to comply with pre- and post-operative regimens. Of the payers included in our review, only Aetna covers liraglutide, lorcaserin, N/B, and P/T at a non-specialty tier. Aetna covers these drugs for patients with a BMI >30 kg/m2 or >27 kg/m2 who meet the following criteria: 1) serious risk factors such as coronary heart disease, dyslipidemia, hypertension, obstructive sleep apnea, or T2DM; and 2) have failed to lose at least one pound per week after six months of a weight-loss regimen that includes diet, increased physical activity, and behavioral therapy. None of the public or private payers surveyed cover vBloc (IGB and DJBL are not yet FDA-approved). Table 1 on the following page describes general insurer criteria for the bariatric surgeries, weight- loss devices, and weight-loss drugs included in this review. Appendix I contains a more detailed description of public and private insurance coverage policies, along with links to the policies. ©Institute for Clinical and Economic Review, 2015 10 [Return to Table of Contents] Table 1. Coverage Matrix of Public and Representative Private Insurance Coverage Policies CMS Medi-Cal Aetna Anthem CIGNA Humana UHC Health Net BSCA Bariatric Surgery Adult BMI ≥35 with >40, >35 with >40, >35 with >40, >35 with ≥40, ≥35 with ≥40, ≥35 with >40, >35 with >40, >35 with >40, >35 with Criteria comorbidity comorbidity comorbidity comorbidity comorbidity comorbidity comorbidity comorbidity comorbidity Open Covered Open Covered Open Open Covered Open Open RYGB Laparoscopic Laparoscopic Laparoscopic Laparoscopic Laparoscopic Laparoscopic Not covered Not covered Open Not covered Not covered Open Covered Not Covered Not covered BPD Laparoscopic Laparoscopic Open Covered Open Covered BMI >50 Open Covered BMI >50 BMI >50 BPD+DS Laparoscopic Laparoscopic Laparoscopic Laparoscopic Covered Laparoscopic Laparoscopic Open Laparoscopic Laparoscopic Open Laparoscopic LAGB Laparoscopic Laparoscopic At MAC discretion, Covered Open Open Open Open Covered Laparoscopic Covered VSG Yes in California Laparoscopic Laparoscopic Laparoscopic Laparoscopic Medications >30, >27 w/ BMI Criteria N/A N/A N/A N/A N/A -- -- N/A risk factors Liraglutide N/A Not listed Covered Not listed Not listed Not listed Tier 3 Not listed Not listed Lorcaserin N/A Not listed Covered Not listed Not listed Not covered Tier 3 Specialty Not covered N/B N/A Not listed Covered Not listed Not listed Not covered Tier 3 Not listed Not covered P/T N/A Not listed Covered Not listed Not listed Not covered Tier 3 Specialty Not covered Devices IGB Not covered -- Not covered -- Not covered Not covered Not covered Not covered -- vBloc Not covered -- Not covered Not Covered Not covered Not covered Not covered Not covered Not covered DJBL Not covered -- Not covered -- Not covered Not covered Not covered -- -- -- = Not mentioned in coverage policy MAC = Medicare Administrative Contractor ©Institute for Clinical and Economic Review, 2015 11 [Return to Table of Contents] 2. Clinical Guidelines Recently, many clinicians have recommended that bariatric surgery guidelines, which established the use of the traditional BMI thresholds to restrict surgery to those with more severe disease, need to be updated to include patients with lower classes of obesity.38 Still others have pointed out concerns with using BMI as an obesity indicator, suggesting that it does not accurately measure fat content, reflect the proportions of muscle and fat, or account for sex and racial differences in fat content and distribution of intra-abdominal and subcutaneous fat.39 For example, the World Health Organization has noted that BMI cut-off points for obesity are lower in Asian populations, whose body fat distributions differ relative to persons of European descent.40 The guidelines below reflect the most current policies from national and international clinical societies concerning the management of excess weight. 2.1 Guidelines for Adult Care Endocrine Society (2015) http://press.endocrine.org The Endocrine Society recommends that pharmacotherapy be used as an adjunct to diet, exercise, and behavioral modification in patients with a BMI ≥27 kg/m2 with a comorbidity or BMI >30 kg/m2. Pharmacotherapy is also appropriate to maintain long-term weight loss in patients meeting the above BMI and comorbidity criteria. Drug therapy should be continued if a patient loses >5% of their body weight by the third month of treatment and discontinued if a patient loses <5% of their body weight or if there are safety or tolerability issues at any time. Patients should be evaluated for the efficacy and safety of their regimens every three months. Phentermine is not recommended for patients with uncontrolled hypertension or a history of heart disease; sympathomimetics are not recommended for patients with cardiovascular disease, and lorcaserin and orlistat are listed as alternatives; and patients with T2DM should be prescribed antidiabetic medications that have been shown to promote weight loss such as glucagon-like peptide-1 (GLP-1) analogs or sodium-glucose-linked transporter-2 (SGLT-2) inhibitors in addition to metformin. US Department of Veteran’s Affairs / Department of Defense (VA/DoD) (2014) http://www.healthquality.va.gov/guidelines/CD/obesity/ The VA/DoD guidelines recommend bariatric surgery for patients with a BMI >40 kg/m2 or a BMI between 35 and 39.9 kg/m2 and one or more obesity-related comorbidities but note that there is ©Institute for Clinical and Economic Review, 2015 Page 12 [Return to Table of Contents] insufficient evidence to support surgery in individuals over age 65 or with a BMI <35 kg/m2. Surgery may be considered to improve obesity-related comorbid conditions in individuals with a BMI >35 kg/m2. The guidelines recommend that pharmacological management with P/T, orlistat, or lorcaserin is appropriate as an adjunct to lifestyle change in patients for whom lifestyle change alone has failed to produce adequate weight loss and with a BMI > 30 kg/m2 or a BMI > 27 kg/m2 and at least one obesity-related comorbid condition. American Heart Association / American College of Cardiology / The Obesity Society (AHA/ACC/TOS) (2013) http://content.onlinejacc.org/article.aspx?articleid=1770219 The AHA/ACC/TOS joint guidelines recommend bariatric surgery for patients who have not responded to previous behavioral modification, with or without pharmacotherapy, with a BMI >40 kg/m2 or a BMI >35 kg/m2 and one or more obesity-related comorbidities. No recommendation was made for patients with a BMI <35 kg/m2 due to a lack of evidence. The societies do not favor one type of surgery over another; instead, they recommend that the decision be made with regard to the individual circumstances of the patient. The guidelines did not specifically examine the evidence on pharmaceutical approaches to weight loss, but note that, based on expert opinion, medication management may be appropriate as an adjunct to lifestyle intervention for individuals with a BMI >30 kg/m2 or a BMI >27 kg/m2 and at least one obesity-related comorbidity. American Association of Clinical Endocrinologists / The Obesity Society / American Society for Metabolic and Bariatric Surgery (AACE/TOS/ASMBS) (2013) http://asmbs.org/resource-categories/guidelines-recommendations The AACE/TOS/ASMBS guidelines state that bariatric surgery is appropriate for patients with a BMI ≥40 kg/m2 who do not have existing medical complications or for whom surgery would not pose excessive risk. Patients with a BMI ≥35 kg/m2 may also be eligible in the presence of at least one severe obesity-related comorbidity. The guidelines note that limited evidence suggests that surgery may be appropriate in patients with a BMI from 30 to 35 kg/m2 who have diabetes or metabolic syndrome. Bariatric surgery to address glycemic control, lipid lowering, or cardiovascular disease risk reduction is not recommended in the absence of the BMI criteria listed above. ©Institute for Clinical and Economic Review, 2015 Page 13 [Return to Table of Contents] The societies do not recommend one bariatric procedure over another due to insufficient evidence in favor of one approach, instead suggesting that the decision be made based on individualized goals, patient preferences, and a personalized risk stratification. Laparoscopic procedures are generally preferred due to lower early postoperative morbidity and mortality. LAGB, VSG, RYGB, and BPD with or without (±) DS are considered to be the primary procedures of interest, though the guidelines recommend that caution should be taken with BPD±DS due to the greater nutritional risks associated with a large length of bypassed small intestine. 2.2 Position Statements American Society for Gastrointestinal Endoscopy / American Society for Metabolic and Bariatric Surgery Task Force on Endoscopic Bariatric Therapy (ASGE/ASMBS) (2011) https://asmbs.org/wp/uploads/2011/11/PathwayToEndoscopicBarTherapies-Nov2011.pdf The ASGE/ASMBS Task Force on Endoscopic Bariatric Therapy (EBT) issued a white paper in 2011 to propose a framework for the development, investigation, and adoption of EBT including intragastric balloon, duodenal-jejunal bypass sleeves, and gastric electrical stimulation devices. An EBT intended for use as a primary therapy should produce a minimum of 25% excess weight loss (EWL) when measured at 12 months. Additionally, studies should demonstrate that the mean percent EWL difference between primary EBT and control groups is at least 15% and statistically significant. These figures may vary for an EBT intended for a non-primary use such as early intervention, bridge therapy to induce weight loss prior to another bariatric procedure, or to improve metabolic illness. When used as a non-primary therapy, an EBT should produce a minimum 5% total weight loss and exceed the weight loss attainable through medical and lifestyle intervention. ASMBS (2011) https://asmbs.org/resources/preoperative-supervised-weight-loss-requirements In a 2011 position statement, the ASMBS argues that pre-operative diet requirements prior to bariatric surgery are unsupported by evidence and that they are “inappropriate, capricious, and counter-productive.” Pre-operative weight loss requirements increase patient health care costs and are an inconvenience to patients due to mandated physician visits that may not be covered by health insurance. The ASMBS notes that it may be appropriate for treatment teams to recommend pre-operative weight loss to improve surgical risk and evaluate adherence based on a patient’s individual needs, but there is only low-level evidence in support of this practice. ©Institute for Clinical and Economic Review, 2015 Page 14 [Return to Table of Contents] ASMBS (2012) https://asmbs.org/resources/bariatric-surgery-in-class-i-obesity This 2012 position statement from the ASMBS states that bariatric surgery should be an option for patients with a BMI from 30 – 35 kg/m2 whose weight and comorbidities are not substantially and durably improved by non-surgical interventions. The society further states that the current BMI cut- off point for surgery (>35 kg/m2 with comorbidities, >40 kg/m2 regardless of comorbidity) “was established arbitrarily nearly 20 years ago,” and that “there is no current justification on grounds of evidence of clinical effectiveness, cost effectiveness, ethics, or equity that this group should be excluded from life-saving treatment.” 2.3 Selected International Guidelines Guidelines published by the United Kingdom’s National Institute for Health and Care Excellence (NICE) are summarized below. National Institute for Health and Care Excellence, UK (NICE) (2014) https://www.nice.org.uk/guidance Bariatric surgery is appropriate for patients with a BMI ≥40 kg/m2 or a BMI ≥35 kg/m2 and a significant comorbidity (e.g., T2DM or high blood pressure) that is expected to improve through weight loss. Bariatric surgery should only be pursued in patients for whom all non-surgical approaches have failed, who have been receiving intensive management, are suitable candidates for anesthesia and surgery, and are able to commit to long-term follow-up. An expedited assessment should be provided for patients with recent-onset T2DM and a BMI ≥35 kg/m2. Recent- onset T2DM in patients with a BMI of 30 to 34.9 kg/m2 or in individuals of Asian ethnicity at lower BMI thresholds for obesity can also be assessed on an individual basis. Surgery is generally not recommended for pediatric populations, though exceptional circumstances may supersede this rule for adolescents who have reached or are near physical maturity. A medical assessment that includes genetic screening is recommended to identify rare but treatable causes of obesity. NICE recommends that duodenal-jejunal bypass sleeves be used only in the context of research due to a lack of evidence on safety and efficacy. ©Institute for Clinical and Economic Review, 2015 Page 15 [Return to Table of Contents] Pharmacotherapy is appropriate for adults with a BMI ≥ 28 kg/m2 and at least one comorbidity or a BMI ≥ 30 kg/m2 who have been unable to attain their weight target or have reached a plateau with lifestyle modification, and should be used as an adjunct to diet, exercise, and behavioral change. Adolescents should only receive drug treatment in the presence of physical or severe psychological comorbidities, and treatment should be administered by a multidisciplinary team that can provide drug monitoring; psychological support; and diet, exercise, and behavioral interventions. Treatment with orlistat should continue for longer than three months only if the patient has lost > 5% of their body weight and continuation of medical therapy is appropriate as a strategy to maintain weight loss. NICE has suspended their review of lorcaserin pending submission of a licensing application by the manufacturer and also suspended their review of P/T following a negative opinion of the drugs issued by the Committee for Medicinal Products for Human Use (CHMP). A review of N/B is currently under development. ©Institute for Clinical and Economic Review, 2015 Page 16 [Return to Table of Contents] 3. Previous Technology Assessments 3.1 Health Technology Assessments We identified four health technology assessments evaluating at least one of the interventions of interest for this review, as summarized below. Agency for Healthcare Research and Quality (AHRQ, 2013): http://effectivehealthcare.ahrq.gov/index.cfm/search-for-guides-reviews-and- reports/?productid=595&pageaction=displayproduct The AHRQ review evaluated the comparative effectiveness of bariatric surgery procedures (LAGB, RYGB, VSG, or BPD) for adult patients with a BMI of 30.0 to 34.9 kg/m² and a metabolic condition compared to nonsurgical interventions. For LAGB, RYGB, and VSG, there was moderate strength of evidence to show that bariatric surgery procedures were effective in treating diabetes, hypertension, and hyperlipidemia in the short-term. The strength of evidence is low for BPD due to fewer studies and smaller sample sizes, and the evidence is insufficient for comparing the outcomes of multiple procedures directly. There is a low strength of evidence for adverse events (AEs) associated with all four surgical procedures, as well as insufficient evidence for determining long- term safety of these procedures in a patient population with moderate obesity. Blue Cross BlueShield Association Technology Evaluation Center (BCBS TEC, 2012): http://www.bcbs.com/blueresources/tec/vols/27/27_02.pdf http://www.bcbs.com/blueresources/tec/vols/27/27_03.pdf There are two technology assessments from BCBS TEC that were published around the same time: one evaluates the effectiveness of bariatric surgery procedures in diabetic patients with BMI of 30.0 to 34.9 kg/m², and the other evaluates all patients with moderate obesity undergoing LAGB. With the exception of RYGB, there is limited evidence demonstrating the effectiveness of bariatric surgery to treat diabetes in patients with moderate obesity. For those undergoing RYGB, the data are variable but promising to show that remission is achieved in the majority of patients. For those undergoing LAGB, the evidence is lacking in both quality and quantity to determine comparative effectiveness against other bariatric surgery procedures with regard to both weight outcomes and AEs, specifically in the long-term. ©Institute for Clinical and Economic Review, 2015 Page 17 [Return to Table of Contents] Canadian Agency for Drugs and Technologies in Health (CADTH, 2010): http://www.cadth.ca/media/pdf/H0485_Bariatric_Surgery_for_Severe_Obesity_tr_e.pdf In a technology assessment focused on the use of bariatric surgery procedures for the treatment of severe obesity compared with standard care (e.g., lifestyle modification) and/or pharmacological therapy, the available evidence suggests (although data from good-quality, long-term randomized controlled trials (RCTs) are lacking) that bariatric surgery appears to be more effective than nonsurgical interventions for treating severe obesity. While RYGB and LAGB have certain tradeoffs with regard to risk of complications and reoperations, diversionary procedures such as BPD result in the greatest weight loss relative to other procedures. There was a lack of evidence to determine the effectiveness of VSG. This review also assessed the economic impact of treating patients with severe obesity by means of surgery or standard care and found that surgical intervention is more effective and less costly in patients with a BMI greater than 35 kg/m² and an obesity-related comorbidity or a BMI greater than 40 kg/m². The results also suggest that both a high procedure volume and extensive surgical experience are associated with better clinical outcomes. California Technology Assessment Forum (CTAF, 2009, 2010, 2012) CTAF previously evaluated the evidence on LAGB (2009), VSG (2010), and all four surgical procedures in patients with diabetes (2012). Findings are summarized below. http://www.ctaf.org/reports/laparoscopic-adjustable-silicone-gastric-banding-obesity http://www.ctaf.org/reports/sleeve-gastrectomy-stand-alone-bariatric-procedure-obesity http://www.ctaf.org/reports/bariatric-surgery-treatment-type-2-diabetes-mellitus CTAF found that the effectiveness of LAGB was inferior to that of RYGB. There is some trade-off, however, since RYGB is a more technically-demanding procedure and is associated with longer operating times and hospital stays, as well as higher early complication and reoperation rates. LAGB should remain an option for those who are given appropriate informed consent about the benefits and harms of LAGB relative to RYGB. Evidence on VSG was found to be promising, as it is less invasive than RYGB and BPD, but longer-term outcomes are uncertain. Finally, several bariatric procedures were found to significantly increase resolution of diabetes compared to intensive lifestyle and medical therapy. However, diabetes alone is not sufficient to justify surgical intervention as it remains unclear whether the harms of surgery outweigh the benefits of disease remission. ©Institute for Clinical and Economic Review, 2015 Page 18 [Return to Table of Contents] 4. Evidence Review (Methods & Results) 4.1. Methods Scope of the Assessment The scope for this assessment is described in the sections that follow using the PICOTS (Population, Intervention, Comparators, Outcomes, Timing, Settings) framework. Our primary goal was to compare the evidence on clinical benefits and harms for surgical procedures, devices, and medications of interest in this review to that of conventional weight-loss management (i.e., combinations of diet, exercise, and/or behavioral and lifestyle interventions). However, we also included studies that actively compared different types of interventions, both across (e.g., device vs. surgery) and within (e.g., gastric bypass vs. gastric banding) categories where available. Evidence on clinical benefit and harms was primarily derived from good- and fair-quality RCTs and prospective comparative cohort studies. Retrospective comparative cohort studies were summarized where feasible – particularly for harms data, given the large sample sizes available in these types of cohorts. Finally, case series of >50 patients were considered only for evidence on harms and long-term benefit. Surgical series were limited to those with two or more years of follow-up given the expected maturity of this evidence base. However, given that follow-up for emerging devices and drugs is not yet adequate to impose such a threshold, and/or is limited by the intervention approach itself (e.g., temporary balloon insertion), we did not impose a strict follow-up limit on these case series. Analytic Framework The analytic framework for this evaluation is depicted in Figure 5 on the following page. Because there were expected limitations on the impact of obesity interventions on long-term measures of morbidity, mortality, and health-related quality of life (HRQoL), a series of conceptual links was required to tie shorter-term impact on measures of body weight and resolution of or improvement in key comorbidities to longer-term outcomes. Population The population of interest included adolescent (i.e., age 12-17) and adult (age 18+) individuals who are overweight or obese according to established thresholds of BMI in US-based treatment guidelines (i.e., BMI ≥25) who received an intervention of interest for this assessment (see “Interventions” below). Studies that focused exclusively on populations with specific conditions (e.g., Prader Willi syndrome, psoriasis, polycystic ovary syndrome) were excluded unless the ©Institute for Clinical and Economic Review, 2015 Page 19 [Return to Table of Contents] condition of interest was a common obesity-related comorbidity such as hypertension, T2DM, sleep apnea, or dyslipidemia. Interventions For assessment of surgery, studies were limited to those that involve the four bariatric procedures commonly used in the US: Roux-en-Y gastric bypass, laparoscopic adjustable gastric banding, vertical sleeve gastrectomy, and biliopancreatic diversion (with or without duodenal switch). We also considered the evidence for three newer types of devices and four medications, as listed below. Figure 5. Analytic Framework for CTAF Evaluation of Obesity-related Treatments Vertical Sleeve Roux-en-Y Gastric Bypass Gastrectomy Benefits Adjustable Gastric Biliopancreatic Diversion Devices & Procedures Banding (± Duodenal Switch) Weight/BMI Comorbidities Intragastric Balloon Vagus Nerve Block Morbidity Patients who are overweight Duodenal-Jejunal Bypass Mortality or obese Liner (BMI ≥25) Harms Quality of life Naltrexone/bupropion Phentermine/topiramate Complications Medications extended-release extended-release Revisions Lorcaserin Liraglutide Removal Other adverse effects Other Conventional Treatment (Other medications, diet & exercise, psychotherapy, etc.) Devices • Temporary intragastric balloon systems (e.g., Silimed®, ReShape®) • Duodenal-jejunal bypass liner (EndoBarrier®) • Vagus nerve block devices (Maestro® system) Medications • Liraglutide (Saxenda®) ©Institute for Clinical and Economic Review, 2015 Page 20 [Return to Table of Contents] • Lorcaserin (BELVIQ®) • Naltrexone/bupropion extended-release (Contrave®) • Phentermine/topiramate extended-release (Qsymia®) We aimed to focus attention on drugs and devices that have received or are imminently expecting FDA approval. However, while all four medications of interest are FDA approved, only the Maestro device has received such approval. The FDA asked the manufacturer to stop enrollment in the EndoBarrier trial due to a higher-than-expected rate of bacterial liver infection, and the balloons are still undergoing or preparing for FDA review.29 We nevertheless summarize all available information on these devices to more fully describe the expected landscape for obesity treatment in the near future. Comparators Our primary comparator of interest was conventional weight-loss management programs, which may have included combinations of diet, exercise, nutritional and/or behavioral counseling, and medications other than those of focus in our review. We note, however, that many studies compared an intervention of interest to placebo and/or sham devices or to a generalized “usual- care” condition; we abstracted data from these studies as well. Finally, head-to-head studies comparing the interventions of interest were also included for completeness. Outcomes Outcomes of interest included the impact of obesity-related interventions on: • Mortality • Weight loss related outcomes (e.g., reduction in BMI, % EWL) • Improvement/resolution of comorbidities • Measures of pain, function, HRQoL, and/or patient satisfaction • Short- and long-term complications of surgical procedures and devices • Reported overall drug-related adverse events (AEs) and AEs leading to discontinuation • Economic outcomes, including payer costs, patient productivity, and cost-effectiveness We assessed the evidence on an overall basis as well as stratified by important baseline (e.g., pre- intervention BMI, age, presence of selected comorbidities) and program (e.g., pre-operative weight loss, multidisciplinary care, pre-surgery waiting period) characteristics. While studies evaluating devices and medications frequently use percent of weight loss or the proportion of patients achieving certain thresholds (e.g., 5-10% reduction in body weight), studies of bariatric surgery instead often report % EWL which represents weight loss that has been achieved relative to a defined goal (i.e., the patient’s individualized ideal body weight); these data ©Institute for Clinical and Economic Review, 2015 Page 21 [Return to Table of Contents] are reported according to the study’s criteria for determining successful outcomes. Because the range and severity of complications varied between the surgeries and devices, we focused on treatment-related morbidity, mortality, and complications leading to revision, repair, or removal. For medications, the focus of attention was on overall rates of all reported AEs, defined by the FDA as any undesirable experience associated with the use of a medicinal product,41 as well as the proportion of AEs leading to drug discontinuation during follow-up. Timing Evidence on intervention effectiveness was limited to comparative studies with at least six months of follow-up as a gauge of the durability of weight-loss and related outcomes. Evidence on harms was derived from comparative studies of any duration, as surgical and device-based interventions report complications at the greatest level of detail during the peri-procedure period (i.e., within 30 days after surgery or minimally-invasive interventions). Evidence on effectiveness and harms from case series focused on studies meeting sample-size and quality criteria (described on page 23). Settings All relevant settings were considered, including inpatient, clinic, and outpatient settings. Literature Search and Retrieval The timeframe spanned the period from January 2000 to the most recently published data available. The electronic databases searched as part of the systematic review included MEDLINE, EMBASE, the Cochrane Register of Controlled Trials, PsycINFO, and CINAHL. Electronic searches were supplemented by manual review of retrieved references, previously published technology assessments, and systematic reviews. Further details on the literature search strategy can be found in Appendix A. The combined search results identified 18,941 potentially relevant studies for this assessment (see Figure 6 on the next page). After elimination of duplicate and non-relevant references, we identified 87 RCTs, 156 comparative cohort studies, and 126 case series, for a total of 369 included studies. The total number of studies for each intervention is as follows: 275 for bariatric surgery, 46 for intragastric balloon, two for vagus nerve block, six for DJBL, 22 for liraglutide (Saxenda®), five for lorcaserin (BELVIQ®), five for N/B extended-release (Contrave®), and eight for P/T extended-release (Qsymia®). ©Institute for Clinical and Economic Review, 2015 Page 22 [Return to Table of Contents] Figure 6. PRISMA flow chart showing results of literature search 18,942 potentially relevant references screened 15,497 citations excluded Population: 138 Intervention: 1653 Comparator: 339 Outcomes: 305 3,445 references for full Study Type: 7147 text review Duplicates: 5915 3,075 citations excluded (no outcomes of interest, reviews, case series with <50 patients, not an active comparator, 370 TOTAL conference abstracts) 88 RCTs 156 comparative cohorts 126 case series Data Synthesis Data on relevant outcomes were synthesized quantitatively where feasible. Quantitative evaluation was limited to good- and fair-quality RCTs and prospective cohort studies. Random-effects models were specified and focused on odds ratios (ORs) for binary measures such as comorbidity resolution. Weighted mean differences in continuous variables such as body weight/BMI and QoL were also assessed. Note that because of expected heterogeneity in patient populations (e.g., BMI levels, comorbidity), study design, duration of follow-up, and other important study characteristics, we did not attempt to conduct network meta-analysis or other forms of indirect comparison across different intervention categories; however, meta-analyses as described above were conducted within intervention categories where feasible. Qualitative evidence tables for all RCTs and comparative cohort studies selected for review can be found in Appendix B. ©Institute for Clinical and Economic Review, 2015 Page 23 [Return to Table of Contents] Study Quality We used criteria published by the US Preventive Services Task Force42 to assess the quality of RCTs and comparative cohort studies, using the categories “good,” “fair,” or “poor.” Guidance for quality ratings using these criteria is presented below, as is a description of any modifications we made to these ratings specific to the purposes of this review. Good: Meets all criteria: Comparable groups are assembled initially and maintained throughout the study; reliable and valid measurement instruments are used and applied equally to the groups; interventions are spelled out clearly; all important outcomes are considered; and appropriate attention paid to confounders in analysis. In addition, intention to treat analysis is used for RCTs. Specifically for this review, attrition did not appreciably differ between study groups (≤20% difference). Fair: Studies will be graded "fair" if any or all of the following problems occur, without the fatal flaws noted in the "poor" category below: Generally comparable groups are assembled initially but some question remains whether some (although not major) differences occurred with follow-up; measurement instruments are acceptable (although not the best) and generally applied equally; some but not all important outcomes are considered; and some but not all potential confounders are addressed. Intention to treat analysis is done for RCTs. Specifically for this review, differences in baseline characteristics and/or duration of follow-up were allowed only if appropriate statistical methods were used to control for these differences (e.g., multiple regression, survival analysis). Poor: Studies will be graded "poor" if any of the following fatal flaws exists: Groups assembled initially are not close to being comparable or maintained throughout the study; unreliable or invalid measurement instruments are used or not applied equally among groups (including not masking outcome assessment); and key confounders are given little or no attention. For RCTs, intention to treat analysis is lacking. Note that case series are not considered under this rating system – because of the lack of comparator, these are generally considered to be of poor quality. Nevertheless, we restricted our use of case series to those that met specific criteria, including a minimum of six months follow-up, sample size >50 patients, clearly defined entry criteria, and use of consecutive samples of patients. Evidence Ratings We used the ICER Evidence Rating Matrix (see Figure 7 on the next page) to evaluate the evidence for a variety of outcomes. The evidence rating reflects a joint judgment of two critical components: ©Institute for Clinical and Economic Review, 2015 Page 24 [Return to Table of Contents] a) The magnitude of the difference between a therapeutic agent and its comparator in “net health benefit” – the balance between clinical benefits and risks and/or adverse effects AND b) The level of certainty in the best point estimate of net health benefit. Figure 7. ICER Evidence Rating Matrix Comparative Clinical Effectiveness High Certainty D C B A Moderate B+ Certainty C+ P/I Low I Certainty I Negative Comparable Small Substantial Net Benefit Net Benefit Net Benefit Net Benefit A = “Superior” - High certainty of a substantial (moderate-large) net health benefit B = “Incremental” - High certainty of a small net health benefit C = “Comparable”- High certainty of a comparable net health benefit D=”Negative”- High certainty of an inferior net health benefit B+=”Incremental or Better” – Moderate certainty of a small net health benefit, with high certainty of at least incremental net health benefit C+=”Comparable or Better” - Moderate certainty of a comparable net health benefit, with high certainty of at least comparable net health benefit P/I = “Promising but Inconclusive” - Moderate certainty of a small or substantial net health benefit, small (but nonzero) likelihood of a negative net health benefit I = “Insufficient” – Either moderate certainty that the best point estimate of comparative net health benefit is comparable or inferior; or any situation in which the level of certainty in the evidence is low ©Institute for Clinical and Economic Review, 2015 Page 25 [Return to Table of Contents] 4.2 Overall Evidence Quality Of the 243 comparative studies identified for this evaluation, we rated only 65 (27%) to be of good quality. An additional 87 studies (36%) were rated fair quality; issues with comparability, duration of follow-up, and/or attrition were identified in these studies, but attempts were generally made to control for confounding in the analytic methods (e.g., survival analysis techniques, multivariate regression). We considered another 91 studies (37%) to be of poor quality because at least one key quality issue was present and not adequately addressed in either study design or analysis, as described further below. For discussions of evidence, we prioritized all good- and fair-quality RCTs and prospective comparative cohort studies, although good- and fair-quality retrospective cohort studies were also summarized; while not discussed in detail, study characteristics from poor-quality studies can be found in Appendix Table B3. The total number of good- and fair-quality RCTs and comparative cohort studies for each intervention of interest is as follows: 100 for bariatric surgery, 11 for intragastric balloon, two for vagus nerve block, two for DJBL, 19 for liraglutide, five for lorcaserin, five for N/B extended-release, and eight for P/T extended-release. Importantly, however, many of the medication studies were performed using doses and regimens that are not part of the FDA approval package. For example, only two of the 19 liraglutide studies used the dose that is approved for weight loss. Across all interventions, follow-up was generally limited to 1-2 years in available studies. Only bariatric surgery had a significant number of studies with two or more years of follow-up, but many of these saw high rates of patient attrition over time due to poor post- surgery program adherence. As such, it is problematic to draw firm conclusions from these studies due to the potential for “informed” censoring (i.e., patients with negative outcomes are less likely to be available for long-term follow-up). We did not identify any studies evaluating patient/program characteristics associated with treatment success for any intervention other than bariatric surgery and intragastric balloon. Also, the intragastric balloon studies were highly variable in terms of comparator (different forms of conventional management or bariatric surgery) as well as treatment regimen (e.g., single vs. multiple temporary insertions). There were specific quality issues with the evidence for surgery and devices. First, treatment groups were often imbalanced with respect to baseline characteristics with the potential to influence outcomes. This was not only frequently encountered but also seldom controlled for in statistical analyses of outcome, even if differences were in baseline body weight or BMI. Another important concern was with follow-up, manifested in both systematic differences between groups in duration of follow-up as well as high rates of loss to follow-up in many long-term studies (due to informed censoring as mentioned above). Large-scale patient attrition is certainly understandable in these patients given the clinical complexity involved in obesity-related illness; however, very few studies ©Institute for Clinical and Economic Review, 2015 Page 26 [Return to Table of Contents] accounted for patient attrition using well-accepted methods such as survival analysis and/or actuarial reporting. Finally, most studies were lacking standardized definitions for important outcomes. For example, relatively few surgical or device studies used an accepted classification system (e.g., Clavien) for categorizing the severity of treatment-related complications; we were therefore limited to tracking overall complication rates alone across studies. In addition, definitions of comorbidity resolution varied across studies. For example, resolution of T2DM was determined based on reductions of hemoglobin A1c (HbA1c) below a clinically-important threshold in some studies (the thresholds themselves also varied), and in others, reduction or elimination of diabetes medications was also required. 4.3 Comparative Effectiveness and Safety Evidence on comparative clinical effectiveness and safety is described separately for bariatric surgery, devices, and medications in the sections that follow. Because evidence in patients with BMI <35 was deemed to be of particular interest, findings for this subpopulation are highlighted separately for each intervention. Of note, although each of the included medications is intended for patients with a BMI ≥30 (or >27 with comorbidities), the mean baseline BMI of patients across medication studies was approximately 36. Availability of subgroup analysis by BMI class is limited. Table 2 on the following page summarizes the level of certainty in the evidence according to BMI class, as well as by the presence of T2DM, the comorbidity most commonly studied for BMI levels <30 and 30-34.9. Because uncertainties remain with long-term outcomes and durability of clinical benefit across all interventions, we did not consider the evidence to provide high levels of certainty for any intervention. Across interventions, the evidence is strongest for surgery, as well as for all interventions in patients with a BMI ≥35. A detailed summary of the evidence is then described according to intervention type in the sections that follow. ©Institute for Clinical and Economic Review, 2015 Page 27 [Return to Table of Contents] Table 2. Strength of Evidence by BMI Class BMI <30 30-34.9 35-39.99 ≥40 T2DM Yes No Yes No --- --- Bariatric Surgery BPD LAGB RYGB VSG Devices Key IGB No evidence DJBL Low certainty vBloc Moderate certainty Drugs High certainty Liraglutide Lorcaserin N/B P/T 4.3.1 Bariatric Surgery The evidence comparing bariatric surgical procedures to conventional weight-loss is summarized below by key outcome of interest. Because the majority of studies presented outcomes after a minimum of one year, the primary focus of discussion is on good- or fair-quality RCTs and prospective cohort studies with at least 12 months of follow-up, although higher-quality retrospective studies are also discussed in some detail (as these tend to involve larger sample sizes). Impact of Bariatric Surgery on Overall and/or Cause-Specific Mortality Importantly, none of the studies in our comparative set directly addressed the impact of bariatric surgery on all-cause or obesity-related mortality; this is not surprising given the significant patient attrition in long-term follow-up for the comparative studies in our sample. A recently-published meta-analysis of long-term data from older trials and cohort studies (published 1986-1997) showed a significantly reduced risk of all-cause mortality from RYGB or LAGB relative to nonsurgical controls (Odds Ratio [OR] 0.55; 95% CI: 0.49, 0.63) and a similarly reduced risk of cardiovascular mortality, but major limitations were noted in the available data, including sample attrition, lack of statistical control for other mortality risk factors, differential ascertainment of causes of death for surgical and ©Institute for Clinical and Economic Review, 2015 Page 28 [Return to Table of Contents] control patients, and a trend toward overstating mortality benefits in smaller studies.43 We did not include the Swedish Obese Subjects (SOS) study in our analytic set because the primary surgical intervention was gastroplasty, which is no longer performed in the US. Long-term follow-up from this study in a matched set of surgical and control patients also suggests that bariatric surgery reduces the risk of all-cause mortality (Hazard Ratio [HR] 0.71; 95% CI: 0.54, 0.92).44 However, the authors note that the recorded death rate was more modest than expected (5% and 6.3% over 15 years for surgical and control patients, respectively), and there was not sufficient discriminatory power in the analysis to ascribe mortality benefit to surgery-induced weight loss. Other mortality studies were not included in our set because they did not include a comparison to a control group that featured an active comparator; these studies have produced somewhat conflicting results. Adams and colleagues assessed overall and cause-specific mortality over a mean of 7.1 years in nearly 10,000 surgical patients matched to a control group of patients with a self- reported BMI of 35 or more who had applied for driver’s licenses in Utah.45 They found significantly reduced rates of mortality from cardiovascular-, diabetes-, and cancer-related causes; however, a key limitation of this study was a lack of information on the baseline health status of control patients. Another large (n=42,094) comparison of bariatric surgery patients and nonsurgical controls treated at 12 Department of Veterans Affairs (VA) centers found a borderline significant reduction in all-cause mortality (HR 0.80; 95% CI: 0.63, 0.995) over a mean of 6.7 years of follow- up;46 however, additional analyses in a subset of patients matched on the propensity score for bariatric surgery failed to yield a statistically-significant result. By contrast, a more recent VA-based evaluation examined all-cause mortality at multiple time points during and up to 14 years of follow- up in 2,500 surgical patients matched on a 1:3 basis to nonsurgical controls (demographics for matched cohorts: mean age 53, 74% male, mean BMI 46).47 No significant differences between groups in all-cause mortality were observed at one year of follow-up. At one to five years, however, surgical patients experienced significantly lower rates of mortality (HR: 0.45; 95% CI: 0.36, 0.56); findings were similar at five to 14 years of follow-up. Bariatric Surgery vs. Nonsurgical Management: Other Outcomes We identified a total of 21 reports of good- or fair-quality RCTs (14) and prospective cohort studies (seven) comparing one or multiple forms of bariatric surgery to nonsurgical management. Characteristics of included studies can be found in Appendix Tables B1-B2. Mean age ranged between 41.4 and 57.7 years (average across studies: 46.4); however, most studies had relatively strict age criteria for entry (e.g., 20-50 years), and elderly patients were examined in only two.48,49 Across all studies, 70-80% or more of subjects were female. Consistent with the selection criteria for this evaluation, nonsurgical comparators involved some form of active diet, lifestyle, and/or medical intervention. In some studies, the intervention was labeled “intensive”; this was variably defined, ranging from dietary and exercise therapy in a ©Institute for Clinical and Economic Review, 2015 Page 29 [Return to Table of Contents] supervised rehabilitation setting50 to outpatient programs involving behavior modification, medication, and dietary counseling,51 to fully-integrated multidisciplinary programs involving physicians, dietitians, psychologists, and occupational/physical therapists.52 Surgical interventions also varied in these studies. RYGB was assessed in 13 studies, followed by LAGB (6), VSG (4), and BPD±DS (3) (note: some studies involved multiple procedures). In most studies, lifestyle interventions were compared to surgical intervention alone or with limited lifestyle support; in a few, however, the intensive lifestyle intervention was provided to all patients, and surgery was added.53–55 Studies were typically performed in all potential candidates for bariatric surgery, but some focused solely on patients with specific comorbidities, typically T2DM.48,49,54–61 Impact on Measures of Body Weight In comparison to nonsurgical management approaches, bariatric surgical procedures were associated with substantial and statistically significant improvements in measures of weight change at a median of two years of follow-up, irrespective of the type of procedure performed or the measure of weight change (e.g., change in BMI, percentage of excess and/or total body weight lost, changes in fat mass or waist circumference). Data for weight loss beyond a median of two years of follow-up is lacking. Figure 8 on the next page presents the results of our meta-analysis of mean BMI at study end for the good- and fair-quality studies that produced these measures along with an appropriate measure of variance (e.g., standard deviation, standard error, 95% confidence interval). The pooled mean difference in BMI at study end was 7.4 points (95% CI: 6.2, 8.6). There was a relatively high degree of heterogeneity in these estimates (I²=84%), but in this case the variability is in the degree of treatment effect across studies; the direction of the effect of surgery in reducing BMI is quite consistent across all studies in the analysis. ©Institute for Clinical and Economic Review, 2015 Page 30 [Return to Table of Contents] Figure 8. Meta-analysis of Mean BMI at Study End: Bariatric Surgery vs. Nonsurgical Management i Heterogeneity: Tau2= 2.81; Q=55.8; df=9; I2=84% Test for overall effect: Z=-12.1 (p<0.001) Impact on Resolution of Comorbidities Improvement and/or resolution of comorbidities was reported in 16 of 21 studies (76%); however, in some of these studies, improvement was measured only in terms of mean changes in laboratory parameters. The most frequently-reported comorbidity was T2DM. Figure 9 on the following page shows the results of our meta-analysis of resolution of T2DM in studies conducted solely in patients with this condition; bariatric surgery was associated with a substantial increase in the likelihood of full resolution (Mantel-Haenzel log OR 3.62; 95% CI 2.49, 4.74). Two studies examined the impact of bariatric surgery on comorbidity resolution using composite measures. Ikramuddin and colleagues randomized 120 patients (mean age 49, 76% female, mean BMI 35) to receive RYGB or lifestyle medical management (nutritional and exercise counseling, weight-control medications, medication optimization for cardiovascular risk factors) over 12 months of follow-up.60 The primary treatment goal was a composite of HbA1c <7%, LDL cholesterol <100 mg/dl, and systolic blood pressure <130 mm Hg, and this was reached by 49% of those receiving surgery and 19% in the lifestyle intervention group (OR 4.8; 95% CI: 1.9, 11.7). A two-year RCT assessed the impact of LAGB versus intensive medical therapy (very low-calorie diet, weight-loss medication, and intensive physician and dietary counseling) in 80 patients (mean age 41, 76% female, mean BMI 34),51 and found that LAGB resolved “metabolic syndrome” as defined using i Follow-up in these studies ranged from 1-2 years. ©Institute for Clinical and Economic Review, 2015 Page 31 [Return to Table of Contents] Adult Treatment Panel III (ATP III) criteria (i.e., obesity plus at least two of: hypertriglyceridemia, reduced HDL cholesterol, hypertension, raised plasma glucose) in 14 of 15 patients diagnosed at baseline (93.3%) versus resolution in seven of 15 (46.7%) for the comparison (p<0.002). Similar patterns were observed in a ten-year follow-up from this study, although nearly half of those originally randomized to nonsurgical management crossed over to LAGB surgery.62 Figure 9. Meta-analysis of Resolution of T2DM: Bariatric Surgery vs. Nonsurgical Managementj Heterogeneity: Tau2= 1.58; Q=20.5; df=8; I2=61% Test for overall effect: Z=6.32 (p<0.001) Other individual comorbidities commonly evaluated in these comparative studies included hypertension and hyperlipidemia. In studies evaluating resolution of these conditions and/or discontinuation of relevant medications as a binary variable, bariatric surgery was associated with two- to three-fold reductions in the prevalence of these comorbidities at the end of follow-up, while nonsurgical management resulted in no appreciable change from baseline.48,49,57–59,61 Detailed findings are presented in Appendix B. We identified three good- or fair-quality studies of the effects of bariatric surgery on sleep apnea. One was a good-quality RCT of 60 patients (mean age 49, 82% female, mean BMI 45) who were randomized to receive LAGB or conventional weight-loss treatment (individualized dietary, exercise, and behavior-modification services) and were followed for two years.63 Sleep apnea, as measured by reductions in the number of events per hour on the Apnea-Hypopnea Index, improved in both groups and did not statistically differ between them. The prevalence of sleep apnea was reduced significantly in 30 patients with T2DM who received VSG and were followed for 18 months in a prospective cohort (from 15% at baseline to 3% at end of follow-up, p=0.03);61 unfortunately, this j Follow-up in these studies ranged from 1-2 years. ©Institute for Clinical and Economic Review, 2015 Page 32 [Return to Table of Contents] measure was not reported for the control group receiving intensive medical therapy. Resolution of sleep apnea also did not statistically differ between groups in a prospective cohort of 179 patients receiving RYGB or one of three nonsurgical options: a residential program, a commercial weight-loss camp, and a hospital outpatient program.64 The Martins cohort study was also the only comparative study that evaluated the impact of bariatric surgery on asthma or arthritis relative to nonsurgical management.64 Unfortunately, the methods for defining resolution of these comorbidities were not reported; in any event, the rate of resolution of asthma and arthritis did not statistically differ between the RYGB group and any of the three nonsurgical intervention groups. Retrospective Cohort Studies We identified a single retrospective cohort study comparing the effects of bariatric surgery to an active form of nonsurgical management, a matched study of 58 patients with T2DM (mean age 52, 59% female, mean BMI 41) undergoing RYGB or receiving medical management (usual care attendance at an endocrinology clinic) over 12 months of follow-up.65 RYGB was associated with statistically-significantly greater reductions in BMI, HbA1c, and use of lipid-lowering medications relative to medical management, as well as significantly greater resolution of T2DM. Outcomes in BMI <35 Among our set of good- and fair-quality RCTs and prospective cohort studies, a total of nine enrolled patients with BMI levels <35.48,49,51,53,54,56,57,60,66 Importantly, seven of the 10 studies included presence of T2DM as an entry criterion, one recruited individuals based on the presence of metabolic syndrome, and two had no specific comorbidity-based entry criteria. All studies involved comparisons of surgery to medical/lifestyle management; procedures evaluated included RYGB (6 studies), LAGB (4), VSG (2), and BPD±DS (1). Outcomes for studies with a mean preoperative BMI of 30-34.9 are summarized in Appendix L; patterns of weight loss across procedures were similar to those in studies of patients at higher BMIs. More broadly, however, all of the seven studies that measured complete T2DM resolution as a binary variable at 12-24 months of follow-up reported substantially and statistically-significantly greater resolution with surgery (range: 26-73%; median 42%) than with nonsurgical management (range: 0-16%; median 9%). Studies that also reported improvement in or partial remission of T2DM (e.g., reduced HbA1c, reduced insulin use) showed between-group differences of even greater magnitude. An additional RCT evaluated the effects of LAGB versus intensive medical therapy on metabolic syndrome in 80 patients with mild or moderate obesity51 and observed resolution in 93% and ©Institute for Clinical and Economic Review, 2015 Page 33 [Return to Table of Contents] 47% for surgery and medical management, respectively (p<0.002). Finally, another study compared RYGB to lifestyle management in 120 patients60 and found that 49% of surgical patients achieved a composite goal of reductions in HbA1c, LDL cholesterol, and systolic blood pressure below common clinical thresholds, versus 19% in the nonsurgical group (p<0.05). Most of these studies also reported improvements in measures of cholesterol and blood pressure, but these were most commonly reported as mean changes in laboratory parameters rather than as binary measures of resolution. Improvements were also noted in other laboratory measures such as plasma insulin, HOMA-IR (a measure of insulin resistance), and C-reactive protein. However, neither laboratory measurement nor binary assessment of resolution was reported for other obesity-related comorbidities of interest for this assessment such as sleep apnea, arthritis pain and function, and asthma in studies of lower BMI levels. Impact on Other Outcomes Two studies reported the impact of bariatric surgery on HRQoL relative to nonsurgical management. One was a prospective cohort study of 139 patients (mean age 45, 70% female, mean BMI 45) who received RYGB or intensive lifestyle intervention (four inpatient rehabilitation admissions totaling seven weeks) and were followed for 12 months.50 HRQoL was measured by the SF-36 as well as two disease-specific scales, the Obesity and Weight-Loss Quality of Life (OWLQOL) and Weight-Related Symptom Measure (WRSM) scales. RYGB was associated with statistically-significantly greater improvement than lifestyle intervention on all summary measures from each of these three scales. The other study was a 10-year follow-up of an RCT62 comparing LAGB to intensive medical therapy. Results from the O’Brien study are previously discussed on page 32. Head-to-Head Comparisons of Surgical Procedures We identified a total of 24 good- and fair-quality RCTs and prospective cohort studies of bariatric surgery procedures relative to each other with at least 12 months of follow-up. Because the primary comparators of interest in this evaluation were conventional weight-loss management or an alternative treatment approach (e.g., device, new medication), these studies are not discussed in detail here. However, detailed descriptions of findings by procedure type are presented in Appendix H. The most frequent comparator was RYGB; 12 studies made comparisons to sleeve gastrectomy, seven to gastric banding, and five to biliopancreatic diversion (with or without duodenal switch). Among these procedures, BPD±DS appears to result in the greatest reduction in BMI (median: 24.2; range: 17.9-25.9), followed by RYGB (median: 15.9; range: 10.2-26.8), VSG (median: 13.1; range: 8.9-24.2), and LAGB (median: 9.8; range: 4.8-16.2). Data were sufficient for meta-analysis only for the comparison of RYGB to VSG. As shown in Figure 10 on the next page, of the four RCTs of RYGB ©Institute for Clinical and Economic Review, 2015 Page 34 [Return to Table of Contents] vs. VSG reporting a difference in mean BMI at study end, no statistically-significant difference in BMI change was observed in any individual study or on a pooled basis. While data were insufficient to conduct meta-analyses on resolution of comorbidities or other outcomes of interest, no differences in outcomes between RYGB and VSG were apparent. Figure 10. Meta-analysis of mean BMI at study end: RYGB vs. VSG Heterogeneity: Tau2= 0.28; Q=3.7; df=3; I2=20% Test for overall effect: Z=0.52 (p=0.605) Findings for all available comparisons of bariatric surgical procedures are described in further detail in Appendix H. Bariatric Surgery in Pediatric Populations We found only two studies of sufficient quality demonstrating effectiveness for any bariatric surgery procedure in pediatric populations: one RCT67 compared LAGB to conventional weight-loss treatment, and one retrospective cohort68 compared LAGB to RYGB. The RCT67 involved adolescents with obesity undergoing any bariatric surgery procedure of interest for this review. A total of 50 patients between 14 and 18 years old (mean age 16.6, 69% female, mean BMI 41.4) with comorbidities who were unable to lose weight through conventional methods received either LAGB or lifestyle intervention. The nonsurgical group received an individualized reduced-calorie diet and exercise program, and compliance was monitored via a food diary and step counts on a pedometer. The mean BMI at baseline was higher in the LAGB group, though the difference was not statistically significant (42.3 vs. 40.4 kg/m² for conventional treatment). After two years, the mean BMI was 29.6 kg/m² in the surgical cohort and 39.2kg/m² in the lifestyle intervention group, representing a significantly greater percentage of EWL among those undergoing LAGB (78.8% vs. 13.2%, p<0.001). For those presenting with metabolic syndrome at study entry, the condition was completely resolved in all nine patients in the surgical cohort compared to six out of 10 patients in the non- surgical group (100% vs. 60%, p=0.025). Mortality was not reported. ©Institute for Clinical and Economic Review, 2015 Page 35 [Return to Table of Contents] Of the five comparative cohort studies we identified in our search, only one study was found to be of fair quality.68 The authors retrospectively evaluated 890 adolescents with obesity from the Bariatric Outcomes Longitudinal Database (BOLD) between the ages of 11 and 19 (mean age 18.5, 75% female, mean BMI 51.4) who received either LAGB or RYGB. Outcomes were assessed every three months up to one year of follow-up. At every time point, patients in both groups had significant weight loss and significant improvement of comorbidities, including T2DM, hypertension, asthma, and obstructive sleep apnea compared to baseline. After one year, patients in the RYGB group lost more than twice as much weight (48.6 vs. 20.0 kg, p<0.001) and had a significantly greater improvement in hyperlipidemia (58.8% vs. 23.3%, p<0.05) compared to those in the LAGB cohort. However, after controlling for selection bias and differences in clinical characteristics between groups at baseline, the mixed model analysis did not yield any significant differences between groups for weight outcomes. There was only one death due to cardiac failure during the study period that occurred in the RYGB group. Long-term Durability of Treatment Effect While there are many long-term studies of bariatric procedures, available studies consistently suffer from quality issues that include significant loss to follow-up and a lack of appropriate statistical techniques to account for patient attrition. Many studies examine outcomes of weight reduction, weight regain, comorbidity relapse, and long-term harms only in the patients still enrolled in follow- up programs – populations that likely differ significantly from those who dropped out of the program. Further, analyses conducted at discrete time points discard the information provided by patients who were lost to follow-up between the time points of interest. We identified 28 good- and fair-quality reports of 23 RCTs, prospective cohort, and retrospective cohort studies; thirteen of the 28 reports were retrospective comparative cohorts and are not discussed in this section. Details of these studies are available in Appendix Tables B1-B2. Long Term Impact on Measures of Body Weight We identified 28 reports of 23 good- and fair- quality RCTs and comparative cohort studies that provided data on weight change outcomes beyond two years. Only three RCTs compared bariatric surgery to non-surgical management.51,54,55,62 The first RCT was a three-year study of 137 patients with T2DM (n=137; 66% female; mean age 48.7; mean BMI 36.5) who were randomized to receive intensive medical therapy alone (lifestyle counseling, weight management, home glucose monitoring, and optimized use of anti-diabetic medications), medical therapy + RYGB, or medical therapy + VSG.55 After three years, patients lost 24.5%, 21.1%, and 4.2% of total body weight for the RYGB, VSG, and medical therapy groups, respectively (p<0.001 for RYGB and VSG vs. medical therapy). ©Institute for Clinical and Economic Review, 2015 Page 36 [Return to Table of Contents] A second RCT was also a three-year study that compared 61 patients with T2DM (82% female; mean age 47.3; mean BMI 35.7) who underwent a lifestyle intervention (modeled after the Diabetes Prevention Program and Look AHEAD) with or without RYGB or LAGB.69 At year 3, patients lost 24.6 kg (25% total body weight), 14.9 kg (15.0% total body weight), and 5.03 kg (5.7% total body weight) in the RYGB, LAGB, and lifestyle groups, respectively (p<0.001 for all comparisons). The final study to compare surgery to conventional obesity management was a 10-year follow-up study of 80 patients (mean age 41, 76% female, mean BMI 34) who received LAGB or intensive medical therapy (very low-calorie diet, weight-loss medication, and intensive physician and dietary counseling).51 Although nearly half of those originally randomized to nonsurgical management crossed over to LAGB surgery after the initial two-year trial, patients who did not go on to have the LAGB procedure gained 2.63% excess weight by 10 years, while those who were originally randomized to LAGB lost 63.04% of excess weight (p<0.05).70 The remaining studies evaluated the long-term comparative effectiveness of various bariatric procedures. Angrisani and colleagues randomized 51 patients (mean age 34, 82% female, mean BMI 44) to receive RYGB or LAGB in a single-center evaluation in which patients were followed for five years;71 one of the 27 LAGB patients was lost to follow-up during this period. At five years, mean BMI was statistically-significantly lower for RYGB relative to LAGB (29.8 vs. 34.9, p<0.001), while the percentage of EWL was significantly greater for RYGB (67% vs. 48%, p<0.001). At 10 years, a total of 5/27 LAGB (19%) and 3/24 (13%) RYGB patients were lost to follow-up.70 Among remaining patients at 10 years, BMI was essentially unchanged in the RYGB group (30.0 vs. 29.8 at five years), while BMI increased somewhat in the LAGB group (36.0 vs. 34.9 at five years). EWL remained in favor of RYGB (69% vs. 46% for LAGB, p=0.03). Another fair-quality RCT evaluated 111 RYGB and 86 LAGB patients (mean age 43, 77% female, mean BMI 47) who were followed for a mean of 4.2 years at a single bariatric surgical clinic.72 Treatment groups were imbalanced because a greater number of LAGB patients could not obtain insurance approval for surgery. EWL was significantly higher in the RYGB group (68.4% vs. 45.4%, p<0.05). In addition, treatment failure, defined as conversion to another procedure because of failure to lose weight or <20% EWL, occurred in 17% of LAGB patients and zero RYGB patients (not statistically tested). The durability of procedure performance was also examined in the three reports of the Søvik RCT. In the 2010 Søvik study, 60 patients with a BMI of 50-60 kg/m2 (mean age 35, 70% female, mean BMI 55) were randomized to RYGB or BPD±DS and followed for two years.73 Mean BMI at 12 months was statistically-significantly lower in the BPD±DS group (32.5 vs. 38.5 for RYGB, p<0.001). At 24 months of follow-up, BMI continued to decline in both groups but the magnitude of differences was similar (30.1 vs. 37.5, p<0.001).74 After five years of follow-up, with a 92% retention rate, the mean BMI for the BPD±DS group remained significantly lower than for the RYGB group ©Institute for Clinical and Economic Review, 2015 Page 37 [Return to Table of Contents] (33.1 vs. 41.2 respectively, p<0.001), but weight regain (9-10 kg) was comparable for the two groups.75 Noticeably missing from weight-change data is any analysis of long-term weight regain following surgery. Although there were limited data available in the studies that met our inclusion criteria, two cohort studies that were excluded from our primary study sample provide information on weight recidivism over longer durations of follow-up. Adams and colleagues used propensity score matching to compare 418 RYGB patients (n=1,156; 82% female; mean age 45.0; mean BMI 45.9) to two non-intervened control groups.76 The two control groups consisted of participants who sought but did not have bariatric surgery at the same surgical center as the RYGB group (n=417; control group 1) and participants who were recruited at random from a population-based sample (n=321; control group 2). Participants were matched according to sex, age, BMI, income, education level, and marital status and were examined at two and six years. During the first two years after surgery, RYGB patients lost a mean 34.9% of total body weight; by year six, however, total weight loss declined to 27.7%, which meant that patients gained back approximately 21.2% (9.9 kg) of their 2- year weight loss. The SOS study, which followed patients for over 15 years, reported that weight increases did occur 1-2 years after surgery but eventually leveled off. After 10 years, weight loss remained 25% and 14% below baseline weight for the subgroups of patients who underwent RYGB and LAGB, respectively (note that the SOS study was not part of our primary set because a majority of patients underwent gastroplasty, a procedure no longer performed in the US). These results were included in a 2013 systematic review of 16 studies, primarily consisting of case series and cross-sectional surveys.77 Weight regain was defined variably in these studies, ranging from gains in absolute weight from a nadir value, to gains above a certain kilogram threshold, to reductions in the percentage of excess body weight lost. In most of these studies, weight regain was common, occurring in 70-80% of subjects, but was moderate for most patients (5-10% of original weight loss regained). However, 10-20% of patients also reported weight regain that exceeded predetermined clinically-important thresholds over 1-11 years of follow-up. Long Term Impact on Resolution of Comorbidities We found three studies that compared bariatric surgery to non-surgical management with follow- up longer than two years. Looking at three years after surgery, Courcoulas and colleagues found that 65% of RYGB patients and 33% of LAGB patients no longer required insulin or oral diabetes; no patients who received a lifestyle intervention alone discontinued diabetes medication (study described on page 37).69 As described previously, the O’Brien RCT62 assessed the impact of LAGB versus intensive medical therapy. After an initial two-year study, the authors found that LAGB resolved “metabolic syndrome” as defined using ATP III criteria in 14 of 15 patients diagnosed at baseline (93.3%) vs. resolution in 7 of 15 (46.7%) for the comparison (p<0.002). Similar patterns ©Institute for Clinical and Economic Review, 2015 Page 38 [Return to Table of Contents] were observed in a 10-year follow-up from this study, although nearly half of those originally randomized to nonsurgical management crossed over to LAGB surgery.62 As with weight changes, degradation in performance of bariatric surgery with respect to comorbidity resolution was rarely evaluated in available RCTs. One RCT (n=150, mean age 48.6, 66% female, mean BMI 36.5) reported achievement of HBA1c levels <6% in 42% and 37% of the RYGB and VSG groups, respectively, versus 12% in those receiving medical therapy alone (p<0.01 for both comparisons) after 12 months.54 Achievement of HbA1c <6% was lower over the next two years but remained substantially higher in the surgical groups (38%, 24%, and 5% for RYGB, VSG, and medical therapy, respectively, p≤0.01 for both surgeries vs. medical therapy). However, relapse, defined as meeting the HbA1c target and discontinuing anti-diabetic medications at 12 months but not at three years, was also common, occurring in 38% and 46% of RYGB and VSG patients respectively (note: relapse could not be calculated in the medical therapy group because no patients achieved the HbA1c target and discontinued anti-diabetic medications). Although the results of the SOS study were not included in our meta-analysis, long-term data on T2DM remission are available from this study. While 72% of surgery patients with T2DM experienced remission at two years of follow-up, the rate of relapse among patients with initial remission and 10 years of follow-up was 50%. Bariatric surgery was associated with reductions in the risk of new-onset T2DM, however (96%, 84%, and 78% after two, 10, and 15 years, respectively).11 Similarly, six-year results from the Adams study (described on page 38) indicate that 75% of surgical patients experienced remission of T2DM at year two, which decreased to 62% by year six. Concomitantly, however, T2DM incidence was significantly lower in the surgical group than it was in the two non-intervened control groups: new diagnoses of T2DM occurred in 2% of the surgical group in the six years following surgery, compared to 15-17% of the control groups. There were also fewer diagnoses of hypertension among surgical patients than control patients (16% vs. 31-33%; p- value not reported). Long Term Impact on Other Outcomes Limited data were available from RCTs and prospective cohort studies on the long-term durability of other outcomes from bariatric surgery. In the O’Brien study previously described on page 38, no statistically-significant differences were found between groups in the physical or mental summary component measures of the SF-36 after 10 years follow-up.62 HRQoL was also reported in the Søvik RCT.75 Although there were statistically-significant improvements from baseline in domain-specific scores of the SF-36 as well as in the Obesity-related ©Institute for Clinical and Economic Review, 2015 Page 39 [Return to Table of Contents] Problems Scale, there were no statistical differences between RYGB and BPD±DS groups after five years follow-up. Long Term Impact in Pediatric Populations To assess long-term outcomes of bariatric surgery in pediatric populations, we attempted to identify any case series with at least 25 patients and a mean follow-up of least two years with 80% participation at the end of the study. We found only one study78 that met these criteria. The authors evaluated the clinical effectiveness of LAGB in 50 pediatric patients between nine and 19 years old (mean age 17.1, mean BMI 45.2) over a mean follow-up of slightly more than seven years. At five years, with only 10% lost to follow-up, the mean BMI was 27.3 kg/m², representing a mean EWL of 92.6%, and the difference between time points was significant up to three years (p<0.01). All patients with a functional band had 100% resolution of all comorbidities, and QoL after surgery continued to improve over time with significant differences between all points of follow-up up to five years (p=0.01). Harms Associated with Bariatric Surgery We identified a total of 32 reports of 28 RCTs and prospective cohort studies that met our criteria for good- or fair-quality and reported on harms of the four bariatric surgery procedures of interest for this review. There were seven comparisons involving BPD, 14 of LAGB, 26 of RYGB, and 12 of VSG, with the most frequent comparison between RYGB and VSG. Eight of these studies compared a single bariatric surgery procedure to conventional treatment; although not discussed in detail here, any reported complications, reoperations, or deaths reported in these studies are represented in the overall calculations of harms in Appendix Table E1. The overall complication rate is comparable between RYGB and LAGB (19.4% vs. 17.9% for LAGB), but the reoperation rate is higher for LAGB (14.8% vs. 6.0%), which also has the highest rate of reoperations across all procedures. VSG is associated with the fewest overall complications (9.5%) and reoperations (2.0%), and BPD±DS has the highest complication rate (31.6%). Most studies were small and too underpowered to detect any statistical differences between procedures for AEs, however. Deaths were rarely or not reported; we identified <100 reported deaths in studies of over 30,000 patients. An additional 29 good- or fair-quality retrospective comparative cohort studies were also identified and had outcomes similar to those of the RCTs and prospective cohorts. There is a lack of both short- and long-term data evaluating safety for any bariatric surgery procedure in pediatric populations. Table 3 on the next page presents the median overall complication and reoperation rate by procedure across all good and fair quality RCTs and prospective cohort studies regardless of duration. Deaths are reported as absolute values, as they were rarely reported. The detailed findings for each surgical comparison can be found in Appendix Table E1. ©Institute for Clinical and Economic Review, 2015 Page 40 [Return to Table of Contents] Table 3. Median Complication and Reoperation Rates for all Good- and Fair-Quality RCTs and Prospective Comparative Cohort Studies, by Procedure Complication Rate; # of # of Follow-Up; Range, Reoperation Rate; # of Procedure Range, Median Studies Patients Median (Months) Range, Median (%) Deaths (%)* BPD 7 189 12-60, 18 17-79, 31.6 3-45, 13.0 0 LAGB 14 13,005 12-120, 24 3-61, 17.9 1-33, 14.8 11 RYGB 26 15,830 1-120, 16 0-78, 19.4 0-33, 6.0 62 VSG 12 2,613 12-36, 12 1-80, 9.5 0-17, 2.0 2 *Complication rate may include reoperations in some studies. Retrospective Cohort Studies We identified a total of 59 retrospective cohort studies that reported on harms of surgery. Table 4 below represents the median complication, reoperation, and mortality rate across all retrospective comparative studies of good or fair quality (note: some studies involved multiple procedures). Median rates tended to be lower than in RCTs and prospective cohort studies, which is not surprising given the information biases attendant in many retrospective evaluations. Nevertheless, the relative effects between procedures are similar to the prospectively-reported data (see Table 3) with VSG representing the lowest complication rate (8.8%), LAGB with the highest reoperation rate (7.4%), and BPD±DS with the highest overall complication rate (26.9%). Table 4. Median Complication and Reoperation Rates for all Good- and Fair-Quality Retrospective Comparative Cohort Studies, by Procedure Follow-Up; Complication Reoperation Mortality Rate; # of # of Procedure Range, Median Rate; Range, Rate; Range, Range, Median Studies Patients (Months) Median (%)* Median (%) (%) BPD 9 2,659 3-63 (24) 8-83, 26.9 0-30, 3.6 0-2.9, 1.40 LAGB 17 16,335 3-72 (29) 0-53, 10.1 0-44, 7.4 0-2.0, 0.15 RYGB 23 840,895 2-72 (29) 0-78, 9.2 0-22, 5.8 0-4.3, 1.94 VSG 11 16,574 2-63 (23) 0-80, 8.8 0-17, 3.9 0-3.9, 0.07 *Complication rate may include reoperations in some studies. Case Series We attempted to identify any case series with at least 100 patients and a mean follow-up of least two years with 70% participation at the end of the study that also reported on harms related to surgery; only 12 studies (two for BPD, seven for LAGB, three for RYGB, and none for VSG) met these criteria for inclusion due to inconsistent reporting of complications and substantial sample attrition. Data abstracted from these studies can be found in Appendix Table E2. ©Institute for Clinical and Economic Review, 2015 Page 41 [Return to Table of Contents] Pediatric Populations Only two studies that met our quality standards reported on harms of bariatric surgery in a pediatric population.67,68 The single RCT67 compared 50 patients (mean age 16.6, 69% female, mean BMI 41.4) receiving either LAGB or lifestyle intervention. In the non-surgical group, 11 patients experienced 18 AEs, of which eight were hospital admissions due to depression or hypertension. Twelve patients experienced 13 AEs in the surgical cohort, including nine reoperations (eight revision procedures and one cholecystectomy) and one readmission due to depression. Of the seven patients who withdrew in the lifestyle intervention group, six had gained weight. Only one patient in the LAGB group was lost to follow-up, though the reason was not reported. Mortality was also not reported. Another comparative cohort study retrospectively evaluated 890 adolescent patients with obesity (mean age 18.5, 75% female, mean BMI 51.4) undergoing LAGB or RYGB.68 The RYGB cohort had 45 readmissions and 29 reoperations, compared to 10 readmissions and eight reoperations in the LAGB cohort. The overall complication rate was 21.6% and 5.0% in the RYGB and LAGB groups, respectively; the majority of complications in both groups were the result of gastrointestinal issues. One death due to cardiac failure during the study period occurred in the RYGB group. Patient/Program Characteristics Associated with Treatment Success There are few good-quality comparative studies that stratify outcomes according to various patient characteristics and procedure type. As such, evidence about the differential effectiveness and safety of bariatric surgery procedures according to patient characteristics such as age, sex, race/ethnicity, BMI, or psychosocial factors is inconclusive. Moreover, evidence of the effectiveness of certain programmatic elements, such as participation in mandatory pre-operative weight loss programs and accreditation of bariatric surgery centers, has been largely inconsistent. Nevertheless, there is evidence, albeit limited, that surgeon experience, high procedure volume, multidisciplinary care, adherence to pre- and post-operative follow-up, and post-procedure support may positively influence outcomes. The key factors of each of the studies reviewed for this section are summarized in Appendix Table C1. Surgeon Experience The majority of studies that assessed surgeon experience with various bariatric procedures examined the learning curve of individual surgeons or surgical groups. These studies stratified patients into consecutive groups and compared outcomes between the first patients to receive a particular procedure at a single institution with later groups receiving the same procedure. The primary outcomes reported included operative time, complication rate, and length of hospital stay. A large proportion of these studies monitored the RYGB learning curve (n=13), although we did ©Institute for Clinical and Economic Review, 2015 Page 42 [Return to Table of Contents] identify four VSG and two LAGB studies; studies related to surgeon experience with BPD±DS are still lacking. The range in operative time, length of hospital stay, and complication rate varied widely, and data appeared to be institution-specific in many instances. Because these studies typically reported outcomes from a single bariatric facility and/or a limited number of individual surgeons and had observational study designs, they have limited external validity. Procedure Volume The majority of studies assessing outcomes according to surgeon and hospital volume were based on data derived from administrative databases. Several studies aggregated bariatric procedures in their analyses or focused only on RYGB. There is likely bias from unobserved confounding factors in the results of the studies described within this section. The majority of studies report an inverse relationship between surgeon or hospital volume and AEs. Nguyen et al. (2004) found that in-hospital mortality was lower in academic medical centers with more than 100 RYGB cases per year (0.3%) compared to centers with fewer than 50 cases per year (1.2%, p<0.01).79 This relationship was more pronounced among patients 55 years of age or above, with whom the observed in-hospital mortality was 0.9% at high-volume hospitals and 3.1% at low- volume hospitals (p<0.01). Likewise, the overall complication rate was significantly lower at high- volume hospitals (10.2% versus 14.5%, respectively; p<0.01), and the mean length of hospital stay was shorter (3.8 versus 5.1 days; p<0.01). Moreover, Nguyen and colleagues found that the mean cost for an RYGB operation was significantly higher at low volume hospitals ($13,908 versus $10,292 for high-volume, p<0.01). Despite the evidence that higher procedure volumes produce better results, Livingston et al. caution that many of these studies rely on statistical methods that amplify the effects.80 Specifically, the authors used Monte-Carlo simulated data to demonstrate that as sample size decreases as a result of low-volume, the uncertainty of the true mortality rate as estimated from the observed mortality rate increases. Relatively few extra deaths in low-volume facilities can result in significant volume effects when analyzed with chi-square tests or logistic regression analysis. Furthermore, the logistic regression models employed in volume studies tend to rely on patient data with incomplete clinical information. Models that incorporate “high-fidelity disease-specific clinical information” allow for high quality risk adjustment, after which volume-outcome relationships may disappear.80 Certification On February 21, 2006, the Centers for Medicare and Medicaid Services (CMS) released the National Coverage Determination (NCD) for Bariatric Surgery for Treatment of Morbid Obesity. This ©Institute for Clinical and Economic Review, 2015 Page 43 [Return to Table of Contents] restricted coverage of bariatric surgery to procedures performed at facilities that were accredited by either the American College of Surgeons (ACS) or the American Society for Metabolic and Bariatric Surgery (ASMBS). k CMS removed this requirement in September 2013, citing evidence that outcomes were comparable at accredited and non-accredited facilities; however, accreditation continues to be touted by some as an important pathway to optimal care.81 The impact of efforts to certify bariatric facilities has proven challenging to measure, with many studies showing no or marginal differences in mortality82–84 and complication rates85 before and after the publication of CMS’s 2006 NCD. With respect to mortality, we identified six studies that compared accredited hospitals with non-accredited hospitals: three studies did not find statistical differences in long-term or inpatient mortality between accredited and non-accredited hospitals,86– 88 two studies found in-hospital mortality was lower in accredited facilities,89,90 and one study reported that accreditation was associated with lower in-hospital mortality (0.06% vs. 0.21%, p=0.003) but was unable to determine conclusively whether the findings were due solely to accreditation, procedure volume, or a combination of both. Six studies evaluated complication rates in accredited versus non-accredited hospitals. Similar to mortality outcomes, findings were inconsistent across studies: four studies did not find a statistically significant association between complications and accreditation,85,86,88,91 while two studies reported that complication rates were lower at accredited facilities. An additional two studies found that complication rates fell after the 2006 NCD.83,84 Detailed outcomes of the studies discussed can be found in Appendix Table C2. Multidisciplinary Care A multidisciplinary care approach has become a common element of bariatric surgery, both before and after the procedure, though very few studies have examined the differential effectiveness of multidisciplinary care across the various bariatric procedures. We found a single study that compared outcomes between patients who received care through a multidisciplinary team approach with those that were treated and followed by the surgical team alone.92 In this study, 200 patients (mean age 31, 62% female, mean BMI 43 kg/m²) were followed for up to 12 months. At 12 months, the percentage of overall weight loss was statistically significantly greater among patients treated by a multidisciplinary team as compared to two cohorts treated by a single surgical group (mean % weight loss 74.3% vs. 59.8-65.0%, p=0.008). Operative time, hospital length of stay, and overall complications were also statistically significantly lower in the multidisciplinary care group. The researchers credited these improved outcomes to a specialized dietician who met with patients k In April 2012, the ASMBS and the ACS formed the Metabolic Bariatric Surgery Accreditation and Quality Improvement Program (MBSAQIP). This unified national accreditation program maintained many of the standards of the previous programs but adjusted volume requirements to 50 or more cases per year. ©Institute for Clinical and Economic Review, 2015 Page 44 [Return to Table of Contents] preoperatively and at consistent post-operative follow-up appointments to evaluate and educate patients on their eating patterns and lifestyles. Additionally, the authors suggested that by sharing perioperative care tasks, surgeons were given more time to focus on improving their technique and gaining experience. In an additional study of 1,236 consecutive LAGB patients in France (49% age 15-39 years, 29% age 40-49 years, 65% BMI 40-49 kg/m2, % female not reported), authors found that patients who did not change their eating habits after surgery were 2.2 times less likely to have weight loss success (defined as EWL >50% at 2 years post-surgery; p=0.009), and patients who did not recover or increase their physical activity were 2.3 times less likely to have success (p<0.001).93 Although they did not directly measure the effects of a multidisciplinary care team, the authors emphasized that these findings were indicative of the need to employ a multidisciplinary care team before and after the operation.93 Type of Pre-procedure Preparation/Post-procedure Support Patient adherence to pre- and post-operative programs and follow-up has been shown to be an important predictor of % EWL. In a subgroup analysis of 177 LAGB patients, those who missed more than 25% of their pre-procedure appointments lost 23% EWL at 12 months compared with 32% for those who missed fewer appointments (p=0.01).94 Not surprisingly, program adherence after surgery has also been shown to be one of the most important predictors of treatment success. In a study comparing 32 RYGB patients who completed 12 months of follow-up to 28 patients who did not (mean age 46.8, 72% female, mean BMI 52 kg/m2), Compher and colleagues (2012)95 calculated that the odds of >50% EWL increased 3.3-fold with each unit increase in the number of follow-up visits (95% CI 1.6, 6.8) and 2.8-fold at 24 months (95% CI 1.4, 5.7). Gould and colleagues (2007)96 had similar findings after following gastric bypass patients 3-4 years post-operatively. The authors found that patients who attended all scheduled post-operative appointments achieved greater EWL (mean of 70% vs. 60% for those followed for only one year, and 56% among those lost to follow-up before one year; p<0.05). Correspondingly, adherence to scheduled follow-up visits and compliance with recommended post-operative care predict a greater decrease in BMI during the first four years after LAGB l.97 Additionally, there is some evidence that post-operative support groups help patients to make positive lifestyle changes, improve psychological comorbidities, and achieve greater weight loss. Post-operatively, support groups have been associated with greater weight loss success and a reduction in patients’ depressive mood.98–100 In a RCT by Nijamkin and colleagues, 144 Hispanic- l Study only reported p-values and f-values; both were significant. ©Institute for Clinical and Economic Review, 2015 Page 45 [Return to Table of Contents] American RYGB patients (mean age 44.5, 83% female, mean BMI 49 kg/m2) were randomized to receive either comprehensive nutrition and lifestyle support or brief, printed healthy lifestyle guidelines six months after surgery.100 At 12 months post-surgery, patients in the comprehensive support group experienced greater EWL (80% versus 64%; p<0.001) and BMI reduction (6.48 kg/m2 vs. 3.63 kg/m2, p<0.001).100 Summary: Bariatric Surgery For patients with a BMI of 35 and above with or without clinical comorbidities, we judge there to be moderate certainty of a substantial net health benefit of bariatric surgery compared to nonsurgical management; certainty remains moderate because of a lack of long-term data on durability of benefit. Across the different types of bariatric procedures, we judge there to be moderate certainty of a small net health benefit with BPD±DS, RYGB, and VSG over LAGB, and moderate certainty of a comparable net benefit between RYGB and VSG. For individuals with a BMI 30-34.9, we judge there to be moderate certainty of a small-to-moderate net benefit of bariatric surgery compared to nonsurgical management among patients with T2DM who receive RYGB or LAGB procedures. The evidence base provides low certainty for all other procedures and for any surgery in this BMI range in patients with comorbidities other than T2DM. Our judgement of the net benefit of bariatric surgery relative to nonsurgical management in pediatric populations is ”insufficient”, as we found only one good-quality RCT of 50 patients that evaluated outcomes in this population. Finally, we found no evidence of the effectiveness of bariatric surgery in patients with a BMI <30. Across a range of procedures, study designs, and duration of follow-up, bariatric surgery results in greater sustained weight loss (on average, 7-8 kg/m2, or 30-40% of total body weight) and resolution of comorbidities (primarily T2DM) than nonsurgical management. The level of certainty in these results is limited by a lack of good-quality long-term data on durability of benefit. There is also a lack of both short- and long-term data demonstrating effectiveness for any bariatric surgery procedure in pediatric populations. There is little to distinguish performance of individual bariatric procedures in patients with a BMI ≥35, as LAGB use has declined substantially, BPD±DS is technically complex and performed only at certain centers, and the performance of RYGB and VSG is similar. Evidence on the differential effectiveness and safety of bariatric surgery procedures according to patient characteristics such as age, sex, race/ethnicity, BMI, or psychosocial factors is largely inconclusive. Nevertheless, there is evidence, albeit limited, that surgeon experience, high procedure volume, multidisciplinary care, adherence to pre- and post-operative follow-up, and post-procedure support may positively influence outcomes. ©Institute for Clinical and Economic Review, 2015 Page 46 [Return to Table of Contents] 4.3.2 Devices The evidence for each device of interest is summarized below by key outcome. The primary focus of discussion is on good- or fair-quality RCTs and prospective cohort studies with at least 6 months of follow-up, although higher-quality retrospective studies are also discussed in some detail. Of note, none of the available studies assessed device impact on mortality or other long-term clinical outcomes (e.g., development of cardiovascular disease). Intragastric Balloon (IGB) Intragastric Balloon vs. Conventional Management We identified a total of seven reports of good- or fair-quality RCTs (five) and comparative cohort studies (two) comparing intragastric balloon to conventional management of obesity. Mean age ranged between 32 and 46 years (average across studies: 39.2), and patients were primarily female (range: 56-87%). Mean BMI across all studies was 41.6, and only three studies allowed participants with a BMI <35.101–103 Consistent with studies for bariatric surgery, conventional approaches involved some form of active diet, lifestyle, and/or behavioral modification intervention. We did not find any studies directly comparing the intragastric balloon to any form of medical management, including those interventions of interest to this review, although drug therapy for weight loss was permitted as part of the standard treatment group. Only one study104 had more than six months of follow-up after IGB removal, and two studies included a cohort that received two consecutive temporary balloons.104,105 Impact on Measures of Body Weight Of the available RCTs, four and one were good- and fair-quality respectively and measured weight outcomes for IGB compared to conventional approaches.101,102,104,106 A good-quality RCT101 assigned 66 patients (mean age 45.9, 68% female, mean BMI 36.4) with a diagnosis of metabolic syndrome to receive 12 months of behavioral modification based on a T2DM lifestyle intervention program; the balloon was implanted in 31 of these participants. At time of removal, those subjects undergoing IGB therapy had a statistically-significantly greater reduction in BMI at six months (5.1 kg/m² vs. 1.7 kg/m² for behavioral modification group, p<0.0001), although reductions were modest in comparison to bariatric surgery. Weight regain was observed in the IGB group at 12 months, but BMI changes remained in favor of the balloon (-3.4 kg/m² vs. -1.9 kg/m², p=0.01). Two additional good-quality RCTs102,106 with a total of 66 patients found statistically-significantly greater percentages of weight loss three months after balloon removal (5.4-11.9%) compared to diet and exercise alone (2.1-4.6%). ©Institute for Clinical and Economic Review, 2015 Page 47 [Return to Table of Contents] Two additional RCTs104,105 evaluated the use of two consecutive balloons compared to diet alone in patients with a BMI of 40-49.9 kg/m². In both studies, all patients initially received IGB therapy for six months, followed by a second balloon for an additional six months with a one-month interval between placements. The first RCT104 included 100 patients (mean age 31.8, 80% female, mean BMI 42.8) and found no significant difference in BMI after the first balloon placement but a significantly lower mean BMI than those in the diet group after the second placement (33.1 kg/m² vs. 37.1 kg/m², p<0.05). After two years of follow-up, weight regain occurred in both groups but differences remained in favor of IGB. The second RCT105 included 50 patients (mean age 32.0, 77% female, mean BMI 42.0) and also found no differences between groups at six months, but at 13 months, the patients receiving a second balloon had a significantly lower BMI (30.9 kg/m² vs. 35.1 kg/m², p<0.005). A small, fair-quality prospective cohort study107 of 16 patients (mean age 44.2, 56% female, mean BMI 46.9) was the only study in our set that did not find any significant differences between groups for weight outcomes after six months of therapy with IGB or intensive lifestyle modification. However, the study may have been underpowered to detect body-weight changes given the small sample size. Impact on Resolution of Comorbidities Improvement and/or resolution of comorbidities with IGB therapy were not reported in any good- or fair-quality RCTs or prospective comparative cohort studies. Retrospective Cohort Studies We identified only one large, fair-quality retrospective cohort103 with 260 patients (mean age 37.9, 77% female, mean BMI 42.0) who received either IGB or a structured diet and behavioral modification program; it found that the mean BMI at balloon removal was significantly, albeit modestly, lower in the IGB group vs. controls (35.4 kg/m² vs. 38.9 kg/m², p<0.01). Although data for weight outcomes were not reported after the active treatment period, the authors noted that patients in the IGB cohort began to regain weight after 24 months, whereas those in the diet group had already regained any weight they lost while on active treatment. More patients in the IGB group also had resolution of T2DM (39% vs. 28%), hypertension (46% vs. 36%), and joint disease (37% vs. 25%) compared to those in the diet group after 24 months, but these results were not tested statistically. Outcomes in BMI <35 We found no comparative studies involving subgroups of patients with BMI <35 that met our inclusion criteria. ©Institute for Clinical and Economic Review, 2015 Page 48 [Return to Table of Contents] Impact on Other Outcomes Two good-quality RCTs were available to assess the comparative impact of IGB relative to conventional management for QoL measures. The first101 assessed QoL in 66 subjects (mean age 45.9, 68% female, mean BMI 36.4) based on the Impact of Weight on Quality of Life (IWQOL)-Lite form and found a significant difference in favor of the balloon group at six months, which was maintained through the next six months following removal. Another RCT (mean age 40.8, 87% female, mean BMI 35.0) evaluated QoL using the SF-36 and reported a substantial improvement for IGB in all measures of physical functioning but did not test these results statistically.102 Intragastric Balloon vs. Bariatric Surgery We found a total of four good- or fair-quality comparative cohort studies evaluating balloon versus bariatric surgery on measures of body weight and comorbidities, two of which were prospective and compared IGB to LAGB.108,109 One of these studies109 assessed the impact of these interventions on QoL outcomes. Two fair-quality retrospective cohort studies also compared IGB to LAGB and RYGB110 and to VSG.111 Impact on Measures of Body Weight We identified two prospective studies that met our quality criteria.108,109 Both were comparisons of IGB to LAGB. One of these was a good-quality prospective cohort study108 evaluating 32 patients (mean age 35.0, 75% female, mean BMI 38.3) who received two consecutive balloons or underwent LAGB surgery, with outcomes reported every six months up to 18 months of follow-up. Although results were comparable for weight outcomes at six months, differences in mean BMI (27.5 kg/m² vs. 36.6 kg/m² for LAGB) and % EWL (70% vs. 57% for LAGB) were statistically significant in favor of IGB after one year (p<0.05), but differences were no longer significant by the end of the study. A fair-quality prospective study109 included 47 subjects (mean age 40.3, 75% female, mean BMI 54.7) who also received the balloon or LAGB surgery. In contrast to Peker et al., mean BMI was significantly lower (39.7 kg/m² vs. 52.1 kg/m², p<0.001) and % EWL was significantly greater (44.0% vs. 26.2% for IGB, p=0.004) for LAGB at 14 months. Impact on Resolution of Comorbidities Both prospective studies also evaluated resolution and/or improvement in comorbidities for IGB versus LAGB.108,109 Due to the small sample size, Peker et al. presented data on comorbidity resolution and improvement as absolute numbers rather than as a proportion of patients achieving these outcomes and did not test for statistical differences. Tayyem and colleagues did statistically test for differences in comorbidity resolution but did not find any significant differences for resolution or improvement in T2DM, hypertension, hyperlipidemia, or sleep apnea. ©Institute for Clinical and Economic Review, 2015 Page 49 [Return to Table of Contents] Retrospective Cohort Studies We also identified two retrospective cohort studies, both of fair quality. The largest of these studies110 evaluated 983 patients (mean age 48.1, 79% female, mean BMI 47.5) and found that RYGB was significantly more effective compared to either LAGB or IGB for % EWL after one year (71%, 27%, and 9% for RYGB, LAGB, and IGB, respectively, p<0.001), and continued to be more effective than the balloon up to two years of follow-up (81% vs. 12% EWL for balloon, p=0.05). In addition, although weight loss was similar for LAGB and IGB at one year, % EWL for LAGB was significantly greater than for IGB after two years (44% vs. 12%, p=0.05). Genco et al.111 assessed weight outcomes for 120 patients (mean age 41, 70% female, mean BMI 54.3) and did not find any statistically-significant differences between VSG and IGB patients after one year. The authors noted, however, that after removal of the balloon, patients in the IGB cohort started to regain weight while those in the VSG group continued to lose weight; results were not reported beyond 12 months of follow-up. Outcomes in BMI <35 We found no comparative studies involving subgroups of patients with BMI <35 that met our inclusion criteria. Impact on Other Outcomes We found only one study109 that evaluated QoL for patients receiving IGB or bariatric surgery (LAGB in this instance). Although participants in both groups reported similar improvements of QoL for all domains from baseline, those in the LAGB cohort reported significantly greater improvements in SF- 36 scores for physical functioning (89 vs. 75, p=0.025), general health (80 vs. 73, p=0.011), and pain (82 vs. 70, p=0.024) relative to the IGB group. Long-term Durability of Treatment Effect No RCTs or comparative cohort studies evaluated outcomes beyond one year of follow-up after IGB removal, so we were unable to assess the long-term impact of the balloon relative to other treatments for obesity. Nevertheless, shorter-term data from these studies suggest that there is a tendency to regain weight when the balloon is removed after six months. Most long-term studies included at least some proportion of patients who had two or more consecutive balloons, had a single balloon longer than six months, or underwent a bariatric surgery procedure within the selected timeframe; these were excluded from our analysis as they are more likely to favor IGB. Table 5 on the next page provides details on the only two long-term case series that met our inclusion criteria on the durability of treatment effect. ©Institute for Clinical and Economic Review, 2015 Page 50 [Return to Table of Contents] Table 5. Long-term Data on Weight Outcomes for Intragastric Balloon Mean # of patients at Last point of Mean BMI Mean BMI @ # patients Study baseline @ last point of follow-up @ balloon last point of @ baseline age/BMI follow-up (months) removal follow-up Benamouzig 2013 67 38.9/36.6 29 53 32.2 33.4 Kotzampassi 2012 474 39.4/43.7 195 60 31.9 39.3 Harms Associated with the Intragastric Balloon As with the other interventions in this assessment, reporting of harms is highly variable for IGB therapy. The most frequently reported complications are nausea, vomiting, and abdominal pain; balloon migration or deflation is infrequent, and the balloon usually passes through the intestinal tract without injury. Early removal of the balloon can occur for any number of reasons, including patient preference (i.e., not safety-related), but is often due to balloon intolerance (e.g., persistent vomiting despite treatment). More serious complications, including gastric perforation or ulcers, bowel obstruction, and esophagitis, are rare. Table 6 below summarizes the incidence rate for the most frequently-reported complications associated with balloon therapy across all comparative cohort studies of any quality. We did not identify any long-term case series that reported complications after removal of the balloon. Some patients may have experienced more than one symptom or complication; data are reported on a per-patient basis wherever possible. Table 6. Harms Associated with Intragastric Balloon Nausea/vomiting/ Device Complication Early removal abdominal pain migration/deflation # of studies reporting outcome 7 1 5 Incidence rate (range, median) 19-98%, 37.8% 25.8% 2-20%, 8.7% To further our understanding of the harms associated with the balloon, we identified a systematic review and meta-analysis of RCTs involving a total of 395 patients from the Cochrane Collaboration that evaluated the relative risk of developing complications from temporary IGB therapy relative to diet or no treatment.112 Complications found to be significantly associated with IGB included gastric erosions (RR 9.78, 95% CI: 3.87, 24.69) and abdominal pain (RR 14.00, 95% CI: 3.45, 56.74). Other complications under consideration were vomiting and migration/deflation of the balloon; data from these studies did not indicate a statistically-significant difference between interventions for these outcomes, however. The authors also evaluated early removal of the balloon, but a meta-analysis was not conducted. ©Institute for Clinical and Economic Review, 2015 Page 51 [Return to Table of Contents] Patient/Program Characteristics Associated with Treatment Success No RCTs or comparative cohort studies evaluated patient or program characteristics associated with treatment success in individuals undergoing IGB treatment for obesity. We did, however, find 20 case series of IGB patients that met our inclusion criteria; results from selected case series are summarized as follows. One series113 of 815 patients (mean age 36.5, 84% female, mean BMI 38.1) evaluated the influence of comorbidities and found that the presence of T2DM was associated with a lower rate of successful weight loss (defined as losing ≥10% of initial weight) (1.8% vs. 5.9% without T2DM, p=0.015); findings were similar for patients with osteoarthritis (7.2% vs. 16.8%, p=0.002). Another study114 assessing 140 patients (mean age 38, 76% female, mean BMI 42.3) found that the duration of balloon therapy was associated with greater reductions in weight. Those with successful outcomes (defined as ≥25% EWL) had the balloon implanted a median of 206 days compared to those who were considered “failures” (median 187 days, p<0.001). Two other studies115,116 found similar results for duration of balloon treatment, but statistical testing was not performed. Data on the influence of age, sex, BMI, other comorbidities, smoking status, psychosocial health, and presence of an eating disorder (e.g., binge-eating) were limited or inconsistent in these studies. Summary: Intragastric Balloon We judge there to be low certainty of a comparable net benefit for temporary IGB insertion relative to either lifestyle intervention or bariatric surgery in patients with a BMI ≥35. In comparison to lifestyle interventions, IGB therapy appeared to provide modest reductions in weight across studies, but data on weight change were mixed after balloon removal, there was variability in balloon duration and number of placements, and there is the potential for harm. Evidence of benefit was truly mixed in available comparisons of IGB to bariatric surgery. Evidence on the benefit of the balloon in patients with a BMI <35 or its impact on any other outcome is insufficient to draw any firm conclusions. Temporary IGB insertion was associated with statistically significant improvements in measures of weight change at a median of one year of follow-up relative to conventional approaches (e.g., diet, exercise, behavioral modification). Benefits were modest, however – higher-quality studies showed incremental BMI changes and percentage weight loss of 1-3 points and 4-8%, respectively – and also tended to worsen after balloon removal (typically after six months). Certainty was low because of great variability in study design, duration of follow-up (particularly after balloon removal), and treatment approach (i.e., single vs. multiple balloon insertions). Evidence comparing IGB to bariatric surgery was mixed; several studies showed greater weight loss with IGB vs. LAGB, but differences diminished with greater duration of follow-up. Data are limited on all other outcomes, including resolution of comorbidities and QoL, and in populations with a BMI ©Institute for Clinical and Economic Review, 2015 Page 52 [Return to Table of Contents] less than 35. Although there is insufficient evidence on the long-term durability of treatment effect or safety of IGB therapy following balloon removal, weight regain appears to be common, and major treatment-related complications are rare. Duodenal-Jejunal Bypass Liner (EndoBarrier®) We identified two good- or fair-quality RCTs that examined the duodenal-jejunal bypass liner (DJBL). Detailed characteristics of these studies can be found in Appendix Tables B1 and B2. There were no treatment-attributed deaths in either study. Impact on Measures of Body Weight We found one good-quality RCT that evaluated the DJBL and met our inclusion criteria. In this study, 73 diabetic patients (35.7% female; mean age 37; mean BMI 35.7) followed a calorie-restricted diet alone or in combination with an implanted DJBL.117 After six months, the DJBL was removed and all patients were followed for a subsequent six months while maintaining diet and exercise. At month 12, the percent total weight loss was 5.8% in the DJBL group versus 3.5% in the control group (p<0.05). Differences in BMI were not statistically significant between groups. A second fair-quality RCT randomized 41 patients (75.6% female; mean age 41; mean BMI 49) to be implanted with the DJBL and/or to follow a low calorie diet.118 After 12 weeks, differences in BMI were not statistically significant between groups. A notable limitation of this study, however, was that only three patients kept the device for a full 24 weeks. As discussed in further detail below, safety concerns and complications related to the device led to early explantation of the DJBL in the majority of patients. Impact on Resolution of Comorbidities Both RCTs of interest evaluated the impact of the DJBL on improvement or remission of T2DM. Schouten and colleagues observed that after 12 weeks, five (62.5%) of eight diabetic DJBL patients decreased insulin dosages and/or anti-diabetic medication, while one patient discontinued medication altogether; outcomes in the two control patients with T2DM were not reported.118 Koehestanie and colleagues reported that after one year, 3.3% of DJBL patients discontinued use of metformin, 13.3% discontinued sulfonylurea derivatives, and 13.3% discontinued insulin; similar percentages of control patients discontinued metformin (2.6%) and sulfonylureas (13.9%), but no control patients discontinued insulin.117 ©Institute for Clinical and Economic Review, 2015 Page 53 [Return to Table of Contents] Outcomes in BMI <35 We found no comparative studies involving subgroups of patients with BMI <35 that met our inclusion criteria. Impact on Other Outcomes We found no studies that reported other outcomes of interest and met our inclusion criteria. Harms Associated with the Duodenal-Jejunal Bypass Liner We found two good- or fair-quality RCTs that met our inclusion criteria and reported data on DJBL- related AEs. In the first study, described above, 76.3% of DJBL patients had at least one AE, which consisted primarily of minor gastrointestinal complaints, abdominal pain, or discomfort.117 The majority of these events occurred within the two weeks after implantation. Five (14.7%) events that were determined to be device-related required hospitalization. One patient had the device removed prematurely after it became blocked with food. In a second fair quality RCT, described above, no serious AEs were reported.118 However, 26 patients (100%) in the device group had at least one mild or moderate AE during follow-up. These events were most commonly nausea (76.9%) and upper abdominal pain (50%), primarily occurring during the first week after the procedure. Four (15.4%) patients had the device removed prior to the 12 or 24 week study period because of sleeve obstruction, dislocation or migration of the device, or epigastric pain. After migration occurred in one patient at four months post-implantation, safety concerns prompted researchers to remove the devices in four remaining patients who still had the device in situ. As such, only three (11.5%) patients kept the device for the full period of 24 weeks. We found one additional case series that met our inclusion criteria and reported on harms. This study of 78 patients (84.4% female; mean age 50.8; mean BMI 43.8) reported that early removal occurred in 16 (20.5%) patients; 14 (18%) of these removals were due to complications such as device or anchor migration.119 Summary: Duodenal-Jejunal Bypass Liner Key questions remain as to the efficacy and safety of the DJBL. Preliminary evidence suggests that there is low certainty that the DJBL has a comparable net health benefit or even a negative net health benefit for weight loss and comorbidity resolution relative to lifestyle interventions in patients with a BMI ≥30. Given this level of certainty, and both published and manufacturer- reported data on potentially serious harms, we judge the current evidence to be insufficient to draw firm conclusions on the DJBL’s effects. ©Institute for Clinical and Economic Review, 2015 Page 54 [Return to Table of Contents] We identified one good- and one fair-quality RCT that compared the DJBL to conventional weight- loss treatment. Neither study found statistically significant differences in BMI change between control and DJBL patients. Patients in both studies reduced or discontinued anti-diabetic medications; however, data on comorbidities were not reported for control patients in one study, and statistical testing for differences was not conducted in either study. Complications of the DJBL included device or anchor migration, epigastric pain, and sleeve obstruction; 2.9-20.5% of patients had the device removed prematurely across studies. Finally, enrollment in a large clinical trial of DJBL is currently stopped pending investigation of a higher-than-expected rate of bacterial liver infection. Vagus Nerve Block (Maestro® vBloc) We identified two good quality RCTs26,120 that examined the vagus nerve block device (vBloc). Detailed characteristics of these studies can be found in Appendix Table B1. There were no treatment-attributed deaths in either study. Impact on Measures of Body Weight Both RCTs of interest reported the percentage of EWL achieved by patients. In the first study, 239 patients (84% female; mean age 47; mean BMI 41) were randomized to receive the vBloc or sham device.26 After 12 months, vBloc recipients lost an average of 24.4% (95% CI: 20.8, 28.1) excess weight compared to 15.9% (95% CI: 11.9, 19.9) in the sham group (p=0.002 in post hoc testing). Thirty-eight percent of the vBloc group achieved at least 25% EWL, compared to 23% in the sham group (OR 2.0; 95% CI: 1.1, 3.8; p-value not reported). At 18 months when study participants were no longer blinded to treatment assignment, the treatment difference between groups increased; the estimated mean EWL was 23.5% in the vBloc group and 10.2% in the sham group (p<0.001).121 In the second RCT, 294 patients (88% female; mean age 46; mean BMI 41) were implanted with an earlier generation of the Maestro vBloc device and received either a complete block of vagal neural impulses or a very low, clinically unimportant level of charge.120 In contrast to the first RCT, there were no statistical differences in the amount of EWL achieved between groups after 12 months (17% vs. 16% for vBloc and low-charge, respectively). Correspondingly, the proportion of patients achieving 25% or more EWL did not statistically differ between groups (22-25%; p=NS). Within both treatment groups, EWL was greater among patients who had a greater daily mean duration of device use. Given this observation, as well as the lack of difference in EWL between groups, authors of the EMPOWER study suggested that the seemingly unimportant level of charge in the control group may have had an effect on vagal function. However, the mechanism behind these results remains uncertain. ©Institute for Clinical and Economic Review, 2015 Page 55 [Return to Table of Contents] Impact on Resolution of Comorbidities We found no studies that reported resolution of comorbidities and met our inclusion criteria. Outcomes in BMI <35 We found no comparative studies involving subgroups of patients with BMI <35 that met our inclusion criteria. Impact on Other Outcomes Sarr et al. (2012) measured QoL outcomes using both the IWQOL-Lite and SF-36 instruments. Scores on both improved for both groups between baseline and 12 months of follow-up; however, these improvements were not statistically different between groups.120 Harms Associated with the Vagus Nerve Block We found two good-quality RCTs that reported data on AEs. The first RCT, described on page 55 reported that 3.7% (n=6; 95% CI: 1.4%, 7.9%; p<0.001) of recipients experienced a serious AE directly related to the device, implantation or revision, or therapy; an additional nine serious AEs were related to general intra-abdominal surgery.26 Serious AEs included neuroregulator malfunction requiring replacement (2), pain at the neuroregulator site (1), atelectasis (1), emesis (1), and gallbladder disease (1). Five (3.1%) participants in the vBloc group and 8 (10.4%) in the sham group required removal of the device before 12 months. The most commonly reported AEs included pain at the neuroregulator site, heartburn and dyspepsia, abdominal pain, eructation, and dysphagia. The second RCT of interest, described on page 55 reported 35 serious AEs, 13 of which were related to the operative procedure (4), implantation or revision (5), or the device itself (4). Sixteen subjects (5.4%) required removal of the device before the trial’s 12-month endpoint (eight for an AE, eight for subject decision), and 14 subjects (4.8%) required a revisionary procedure.120 Summary: vBloc Device We judge there to be low certainty of a small or comparable net benefit for the vBloc device compared to a sham device in patients with a BMI ≥35. There is low certainty because only two studies examined the vBloc, one of which assessed an earlier generation device that was not FDA- approved. While the ReCharge trial results indicated greater weight loss with vBloc vs. sham, many questions remain, including the impact of the vBloc on comorbidities, whether the device can be safely used for durations longer than 18 months, and whether the benefits demonstrated in the ReCharge trial can be repeated in future studies of the device. We found no evidence of the vBloc’s effects in patients with BMI levels <35. ©Institute for Clinical and Economic Review, 2015 Page 56 [Return to Table of Contents] We identified two RCTs that examined the vagus nerve block (vBloc), one of which assessed an earlier generation of the Maestro device. After 12 months follow-up in each study, vBloc patients lost 17-24.4% of excess body weight versus 15.9-16.0% excess weight loss (EWL) in control patients. The included studies did not report outcomes related to resolution or improvement in comorbidities. In both studies, 5.4% of patients had the device removed before study end; 3.0-3.7% of patients experienced serious complications related to the device, implantation, or revision. 4.3.3 Medications The key studies and outcomes for each medication are summarized in detail in the sections that follow. We also conducted Bayesian network meta-analyses using a random-effects approach to facilitate comparisons of each drug on a) the proportion of patients achieving 10% or more total weight loss, and b) the proportion of patients discontinuing due to an adverse event. All four medications were significantly more likely to achieve 10% weight loss vs. lifestyle intervention alone, but only one comparison between agents was significant (phentermine/topiramate vs. lorcaserin: odds ratio [OR] 2.42, 95% Credible Interval [CrI] 1.03, 5.64). The likelihood of discontinuation due to adverse events did not differ across medications, but both phentermine/topiramate and naltrexone/bupropion were significantly more likely to be discontinued for adverse events than lifestyle intervention alone. Detailed methods and results of these analyses are available in Appendices D and E. Of note, none of the studies included in this section addressed the impact of obesity medication on all-cause or obesity-related mortality; this is not surprising given the relatively short duration of the trials in our sample. In addition, incidence of cardiovascular events was not assessed in the included studies, although a large trial designed to study the cardiovascular safety of naltrexone plus bupropion (Contrave®) was recently halted after interim data were released to internal and external stakeholders without authorization.122,123 The FDA has requested that a second study be conducted because of this release.124,125 The executive steering committee expects to report the original study’s data in a scientific forum after all of the cardiovascular events have been collected and adjudicated.123 Liraglutide (Saxenda®) We identified a total of five reports of three good-quality RCTs that compared one or multiple doses of liraglutide to a placebo and/or orlistat (120 mg administered three times daily). Although liraglutide is prescribed at lower doses for management of T2DM, we focused exclusively on studies in which at least one treatment arm received the recommended dose for obesity management (3 mg daily). Characteristics of included studies can be found in Appendix Table B1. All trials provided lifestyle counseling and/or prescribed a low calorie diet to all participants.126,127 ©Institute for Clinical and Economic Review, 2015 Page 57 [Return to Table of Contents] Impact on Measures of Body Weight Three RCTs reported weight-related outcomes for liraglutide at the recommended dose for obesity management (3 mg).126–128 In a two-year crossover study, 564 participants (76% female; mean age 46; mean BMI 34.7) were randomly assigned to once-daily subcutaneous liraglutide (1.2, 1.8, 2.4 or 3.0 mg), placebo, or open-label orlistat (120 mg, three times daily). At the end of year one, all low- dose liraglutide and placebo recipients switched to liraglutide 2.4 mg, and eventually to 3.0 mg (between weeks 70-96).128 At 12 months follow-up, patients in the liraglutide 3 mg group lost 8% of total body weight, compared to 2.9% and 2.4% weight loss in the placebo and orlistat groups, respectively (p<0.0001 for both comparisons). In addition, 73% and 37% of patients taking liraglutide 3 mg lost more than 5% and 10% of total body weight, respectively, compared to 28% and 10% in the placebo group and 44% and 14% in the orlistat group (p<0.001 for all comparisons of liraglutide to placebo or orlistat). Results from the end of year two, which incorporated those patients who shifted from the placebo and lower-dose liraglutide groups into liraglutide 3 mg, maintain significantly better weight loss outcomes among liraglutide recipients relative to orlistat recipients (52% vs. 29% achieving more than 5% weight loss; p<0.001). The second study was the SCALE Maintenance trial.127 In this study, 422 patients (81.5% female; mean age 46.2; mean BMI 35.6) who lost at least 5% of total weight during a low-calorie diet run-in were randomly assigned to liraglutide 3 mg or placebo for 56 weeks; both groups received in- person diet and exercise counseling during the duration of the trial. After 56 weeks, the liraglutide group had lost an average of 6% total body weight, and were 3.9 times more likely to achieve more than 5% weight loss (95% CI: 2.4, 6.1; p<0.001) relative to patients in the placebo group. Twenty-six percent of liraglutide patients achieved 10% or more weight loss versus 6.3% of placebo patients (o<0.0001).127 The third study to evaluate liraglutide at labeled dosing for weight loss was the SCALE Obesity and Prediabetes trial.129 In this trial, 3,731 patients (78.5% female; mean age 45.1; mean BMI 38.3) were randomly assigned to receive liraglutide 3 mg or placebo. After 56 weeks, patients in the liraglutide group lost a mean 8.0% of their body weight versus patients in the placebo group who lost 2.6% of their body weight (p<0.001). A greater proportion of patients lost at least 5% or 10% of their body weight with liraglutide (63.2% and 33.1%, respectively) than with the placebo (27.1% and 10.6%, respectively; p<0.001 for both comparisons to liraglutide). Impact on Resolution of Comorbidities One RCT of interest evaluated the impact of liraglutide on improvement or remission of obesity- related comorbidities. Astrup and colleagues found a beneficial effect of liraglutide 3 mg on T2DM indicators, finding 95.1% with normal glucose tolerance (defined as fasting plasma glucose <5.6 mmol/L or <7.8 mmol/L during oral glucose tolerance test) versus 65% in the placebo group (OR: ©Institute for Clinical and Economic Review, 2015 Page 58 [Return to Table of Contents] 12.5; 95% CI: 2.9, 55.0; p<0.01) at 20 weeks follow-up.126 Similarly, metabolic syndrome improved more in patients taking standard-dose liraglutide than placebo; the proportion of patients with metabolic syndrome improved from baseline prevalence of 28% and 34%, in liraglutide and placebo groups, respectively to 11% and 21% at follow-up (significance not reported). A second RCT reported that T2DM developed in more patients in the placebo group (14 of 813 at- risk patients) than in the liraglutide group (4 of 2,130 at-risk patients; p<0.001) at week 56. Similarly, the prevalence of prediabetes was significantly lower in the liraglutide group (30.8%) than in the placebo group at week 56 (67.3%; p<0.001).129 Outcomes in BMI <35 We found a single comparative study that stratified changes in body weight by baseline BMI category. The SCALE Obesity and Prediabetes trial (described on page 58) found that weight loss outcomes did not differ appreciably between patients in any BMI class. Weight loss for patients with a baseline BMI ≥35 ranged from 7.4% to 9.3% in the liraglutide group and from 2.8% to 3.4% in the placebo group; 60.7-63.5% achieved at least 5% weight loss with liraglutide compared to 25.9-28.5% with a placebo. Comparably, patients with a baseline BMI <35 experienced 7.5-8.3% and 0.9-2.6% weight loss in the liraglutide and placebo groups, respectively; 66.0-66.9% achieved 5% weight loss with liraglutide versus 24.9-30.0% with the placebo.129 Impact on Other Outcomes Two studies reported QoL-related outcomes.130 In the Pi-Sunyer trial, liraglutide was associated with significantly higher scores on the SF-36 for overall physical and mental health, as well as a higher total score on the IWQOL-Lite compared with the placebo. In a secondary analysis of the Astrup trial, Lean and colleagues stratified subscale outcomes of the IWQOL-Lite questionnaire (described previously) according to those who did and did not experience one or more episodes of nausea or vomiting.130 Scores improved across all subscales between baseline and follow-up with only marginal differences between those who experienced symptoms of nausea or vomiting. Harms Associated with Liraglutide The three RCTs of interest reported on AEs. The most commonly reported AEs were gastrointestinal disorders, including nausea, vomiting, diarrhea, and constipation. These events occurred most frequently during the first four weeks of treatment. The rate of overall AEs ranged from 80.3-95.7% between studies, and discontinuation from AEs occurred in 7.5-9.9% of patients. These results are summarized in Appendix Table E3. ©Institute for Clinical and Economic Review, 2015 Page 59 [Return to Table of Contents] Summary: Liraglutide We judge there to be moderate certainty that liraglutide (3 mg) plus lifestyle modification versus lifestyle modification alone provides a small net benefit in patients with a BMI ≥27, due to modest levels of incremental weight loss and comorbidity resolution. Certainty was judged to be moderate because while three good quality studies reported consistent results, uncertainty remains around the durability of liraglutide’s benefits as well as safety over longer-term follow-up (i.e., >2 years). The three studies with labeled dosing for weight loss (3 mg) found that patients lost 6-8% of total body weight after one year of follow-up and were more likely to have normal glucose tolerance relative to those in the placebo group. Changes in QoL were reported in a single study showing that patients who experienced nausea or vomiting had marginally lower improvements in quality of life. The most commonly reported AEs were gastrointestinal disorders including nausea, vomiting, diarrhea, and constipation. Overall AE rates ranged from 80.3-95.7% across studies, and discontinuation due to AEs ranged from 7.5-9.9%. Lorcaserin (BELVIQ®) We identified a total of three good quality RCTs comparing one or multiple doses of lorcaserin to placebo. Characteristics of included studies can be found in Appendix B. Mean age ranged between 43.8 and 52.7 (average across studies: 44.6); no trial exclusively examined elderly or pediatric patients. Across studies, 55-84% of patients were female (average: 79.4%), and mean BMI was 36.2. All studies compared lorcaserin (10 mg, administered once or twice daily) to placebo and provided standardized nutritional and exercise counseling for all participants. All three studies reported results at 52 weeks of follow-up, although the BLOOM trial randomly reassigned patients in the treatment arm during year 2 to continue taking lorcaserin or to switch to placebo.131 Aggregated results reported in this section are from 52 weeks of follow-up. No deaths occurred in any study. Impact on Measures of Body Weight Across all three RCTs of interest,5,131,132 patients who took twice daily lorcaserin achieved greater weight loss than those receiving a placebo. At 52 weeks of follow-up, reductions in total body weight ranged from 4.5-5.8% among lorcaserin recipients, compared to a 1.5-2.8% mean decrease among those taking the placebo (p<0.001 for lorcaserin vs. placebo in all studies). A significantly larger proportion of participants achieved 5% and 10% or more total weight loss with lorcaserin than with the placebo. Across studies, 44.1% (range 37.5%-47.5%) and 20.5% (range 16.3%-22.6%) of patients lost 5% and 10% or more total weight, respectively with lorcaserin, compared to 20.5% (range 16.1%-25%) and 7.3% (range 4.4%-9.7%) with a placebo. In the BLOOM trial, a greater proportion of patients who continued to receive lorcaserin maintained 5% or more ©Institute for Clinical and Economic Review, 2015 Page 60 [Return to Table of Contents] weight loss in year two than those who were reassigned to receive placebo (67.9% vs. 50.3%, p<0.001).131 Impact on Resolution of Comorbidities Outcomes related to the resolution of comorbidities were reported in a single study of 603 T2DM patients (55% female; mean age 52.7; mean BMI 36) who took metformin and/or a sulfonylurea for the duration of the trial.5 Authors of the BLOOM-DM trial reported that at 52 weeks follow-up, 50.4% of the lorcaserin group and 26.3% of the placebo group achieved an HbA1c <7% (p<0.001).5 Outcomes in BMI <35 We found a single comparative study involving subgroups of patients with BMI <35 that met our inclusion criteria. The BLOSSOM trial (n=4,004; 79.8% female; mean age 43.8; mean BMI 36.2) found that weight loss outcomes did not differ for patients in any BMI class for either the lorcaserin or placebo groups.132 Impact on Other Outcomes Three studies reported HRQoL outcomes using the IWQOL-Lite.5,131,132 IWQOL-Lite scores improved in two studies in both the lorcaserin (+11.8 - +12.4) and placebo groups (+10.0 - +10.7; p<0.001 for both study comparisons).131,132 Harms Associated with Lorcaserin We identified five good-quality RCTs that reported on harms. Across studies, headache, upper respiratory infection, nausea, and nasopharyngitis were the most commonly reported AEs. No study reported drug-related adverse effects of lorcaserin on laboratory measures, vital signs, or cardiac valvulopathy. No deaths occurred in any study. Discontinuation of lorcaserin from drug-related adverse effects occurred in 4.3-8.6% of patients across studies. A single study reported that 82.6% of study participants experienced an AE.132 These results are summarized in Appendix Table E4. Summary: Lorcaserin We judge there to be moderate certainty of a small net benefit of lorcaserin (10 mg, administered once or twice daily) plus lifestyle modification versus lifestyle modification alone for weight loss in patients with a BMI between 27-45. There is moderate certainty because while three good quality studies reported consistent weight-loss results, only one study reported outcomes related to improvement of obesity-related comorbidities. Furthermore, we found a single study that stratified outcomes according to various BMI categories. Finally, there is uncertainty around the durability of lorcaserin’s benefits as well as safety over longer-term follow-up (i.e., >2 years). ©Institute for Clinical and Economic Review, 2015 Page 61 [Return to Table of Contents] In the three good-quality RCTs, reductions in total body weight were modest, ranging from 4.5-5.8% among lorcaserin recipients, compared to a 1.5-2.8% mean decrease among those taking the placebo (p<0.001 for lorcaserin vs. placebo in all studies). A single study reported outcomes related to comorbidity status and found 50.4% of lorcaserin patients versus 26.3% of placebo patients achieved an HbA1c <7% (p>0.001).5 Discontinuation of lorcaserin from drug-related adverse effects occurred in 4.3-8.6% of patients across studies, and approximately 80% of study participants experienced any AE. Naltrexone/Bupropion (Contrave®) We identified a total of four good-quality RCTs comparing various doses of N/B combination therapy to placebo. Characteristics of included studies are reported in Appendix B. Mean age ranged between 44.2 and 53.9 (average across studies: 45.4); no study focused exclusively on elderly or pediatric patients. Three of the four trials were composed of over 84% females (average across five studies: 82.9%). The mean BMI ranged from 36.2-36.7 across studies (average: 36.3). All four studies had a duration of follow-up of 56 weeks and focused solely on patients with specific comorbidities, including dyslipidemia, hypertension, and T2DM. All participants in all studies also received a lifestyle intervention, which included guidance on following a calorie-deficit diet, increasing physical activity, and modifying behavior. Impact on Measures of Body Weight We identified four good-quality RCTs that reported weight outcomes for N/B.6,133–136 Across these four studies, participants received standard doses of naltrexone (extended-release, 32 mg daily) combined with bupropion (extended-release, 360 mg daily) and lost 5.0-9.3% of total body weight after 56 weeks of follow-up; patients who received a placebo lost 1.2-5.1% of total body weight (p<0.05 for all comparisons to N/B). N/B was also associated with a significantly larger proportion of participants achieving 5% and 10% weight loss versus placebo. Across studies, 52.4% (range 44.5%-66.4%) and 28.4% (range 18.5%- 41.5%) of patients who received N/B achieved 5% and 10% or more total weight loss, respectively, compared to 23.6% (range 16.0%-42.5%) and 9.7% (range 5.7%-20.2%) in the placebo groups (p<0.05 for all study comparisons). Patients who participated in the COR-BMOD trial (n=793; 90.5% female; mean age 45.8; mean BMI 36.7) experienced better weight loss outcomes in both treatment and control groups than patients in other trials.136 This difference may be the result of the COR-BMOD trial’s more intensive group behavior modification program. Whereas the lifestyle intervention in the other trials typically consisted of advising patients to follow a calorie-deficit diet, increase physical activity, and modify certain behaviors, the BMOD program was an intensive intervention involving 28 group sessions led ©Institute for Clinical and Economic Review, 2015 Page 62 [Return to Table of Contents] by registered dietitians, behavioral psychologists, or exercise specialists. During each session, group leaders led discussions about new topics in weight control, including meal planning, stimulus control, slower eating, problem solving, social support, coping with high risk situations, and maintaining lost weight. In addition, patients were taught to increase weekly exercise, measure portion sizes, count calories, and record both food intake and physical activity. Relative to patients in other trials, who lost on average 5.8% and 1.4% of total body weight in the N/B and placebo groups, respectively, N/B and placebo patients in the COR-BMOD trial lost 9.3% and 5.1% of total body weight, respectively (p<0.001).136 Furthermore, a considerably greater proportion of N/B and control group patients achieved 5% and 10% weight loss through the BMOD program than patients who participated in other trials. Indeed, 66.4% and 41.5% of N/B patients achieved 5% and 10% weight loss, respectively, through BMOD versus 48.5% and 24.2% across the other three studies; 42.5% and 20.2% of BMOD control patients achieved 5% and 10% weight loss, respectively, compared to 17.3% and 6.1% in the other studies. Impact on Resolution of Comorbidities A single RCT of 424 patients (53.6% female; mean age 53.9; mean BMI 36.5) reported outcomes related to improvement of comorbidities.6 At week 56, 44.1% of N/B subjects achieved a target HbA1c <7% compared to 26.3% of placebo patients (p<0.001). Outcomes in BMI <35 We found no comparative studies involving subgroups of patients with BMI <35 that met our inclusion criteria. Impact on Other Outcomes Three studies reported HRQoL outcomes using the IWQOL-Lite. Across studies, overall weight- related QoL improved significantly more with N/B than with placebo, though both groups experienced an improvement. Between baseline and follow-up, patients taking standard dose N/B shifted from “moderate” to “mild” impairment with score improvements ranging from 10.9-13.4 points; by contrast, patients in the placebo groups improved their IWQOL-Lite scores by 6.4-10.3 points (p<0.001 for all comparisons with N/B).133,135,136 Harms Associated with Naltrexone/Bupropion We identified four good-quality studies that reported N/B-related AEs. Across studies, nausea, headache, and constipation were the most commonly reported treatment-emergent AEs. Nausea and other gastrointestinal symptoms occurred most frequently during dose escalation. There were no apparent drug-related negative effects of N/B on laboratory measures or electrocardiogram findings. Combination treatment was not associated with increased incidence of depression, ©Institute for Clinical and Economic Review, 2015 Page 63 [Return to Table of Contents] suicidality, or other mood-related AEs. Discontinuation of N/B from adverse effects occurred in 19.5-29.3% of patients receiving standard-dose N/B across studies and 83.1-90.4% of patients experienced any AE. These results are summarized by study in Appendix Table E5. Summary: Naltrexone/Bupropion We judge there to be moderate certainty of a small net benefit associated with N/B plus lifestyle intervention versus lifestyle intervention alone in patients with a BMI between 27 and 45. There is moderate certainty of benefit because four good-quality studies showed consistent weight loss relative to comparator treatment; however, benefits were modest. In addition, there is uncertainty around the durability of benefit from N/B as well as safety over longer-term follow-up, since we found no studies with follow-up greater than one year. We also judge the evidence to be of low certainty for a small net benefit in patients with BMI 30-34.9 and T2DM based on the results of a single RCT in this population. Across four available RCTs, participants who received standard doses of naltrexone (extended- release, 32 mg daily) combined with bupropion (extended-release, 360 mg daily) lost 5.0-9.3% of total body weight after 56 weeks follow-up; patients who received a placebo lost 1.2-5.1% of total body weight. A single RCT reported outcomes related to improvement of comorbidities and found that 44.1% of patients taking N/B achieved a target HbA1c <7% compared to 26.3% of placebo patients.6 Discontinuation of N/B from adverse effects occurred in 19.5-29.3% of patients receiving standard-dose N/B across studies, and 83.1-90.4% of patients experienced any AE. Phentermine/Topiramate (Qsymia®) We identified a total of eight good- or fair-quality reports of five RCTs; two reports were trial extensions, one of which maintained treatment and assignment, while the other changed the active treatment but maintained group assignment. Two reports were secondary analyses of previous RCTs. All studies compared one or multiple doses of P/T extended release combination therapy to a placebo and/or phentermine or topiramate monotherapy. Characteristics of included studies can be found in Appendix B. Across studies, mean age ranged between 42.6 and 52.4 (average across studies: 47.8), 47-83% of patients were female (average: 74.7%), and mean BMI was 38.0 (range 35.4-42.2). All studies provided a standardized lifestyle intervention for all participants; four of the five trials implemented the LEARN program (lifestyle, exercise, attitude, relationship, nutrition), which consisted of a calorie-deficit diet, increased water consumption, and increased physical activity. ©Institute for Clinical and Economic Review, 2015 Page 64 [Return to Table of Contents] Three studies focused solely on patients with obesity-related comorbidities.137–139 The median duration of follow-up was 28 weeks (range: 28-56). Impact on Measures of Body Weight Weight loss was assessed at various dose combinations of P/T in a total of five of the included studies. Across studies, P/T had a significant and dose-dependent effect on weight reduction. Two of the five trials prescribed phentermine (7.5 mg) plus topiramate (46 mg) at the recommended dose.137,140 In the CONQUER trial, 2,487 patients (70% female; mean age 51.1; mean BMI 36.5) with two or more obesity-related comorbidities were randomized to receive P/T (7.5/46 mg), P/T (15/92 mg), or placebo.137 At 56 weeks follow-up, patients who received the recommended dose lost 7.8% of total body weight, compared to 9.8% and 1.2% in the 15/92 mg and placebo groups, respectively (P/T vs. placebo: p<0.0001). In a 56-week extension to the CONQUER trial, recipients of P/T (7.5/46 mg) lost 9.3% of total weight while placebo patients lost 1.8% (p<0.001).141 By study end, 75.2% and 50.3% of P/T (7.5/46 mg) patients had lost at least 5% and 10%, respectively, of their initial body weight compared to 30% and 11.5% of patients in the placebo group. Correspondingly, in a 28 week trial of 756 patients (79% female; mean age 44.7; mean BMI: 36.2), P/T (7.5/46 mg) recipients lost 8.46% of weight (versus 1.7% weight loss in the placebo group; p<0.05), with 62.1% achieving 5% or more total weight loss (versus 15.5% in the placebo group; p<0.0001).140 Aronne and colleagues also examined outcomes in patients who received higher doses of P/T (15/92 mg, once daily), as did three additional RCTs.138,139,142 Despite a duration of follow-up that ranged from 28-56 weeks, these patients lost a similar amount of weight across studies: 9.2-10.9% of total body weight. Impact on Resolution of Comorbidities Five reports of three RCTs provided information on improvement in comorbidities. T2DM-related outcomes were described in three reports.137,138,143 In a subgroup analysis of the CONQUER trial, Garvey and colleagues found a lower annualized incidence rate of T2DM among pre-diabetic patients who were assigned to either dose of P/T versus placebo, but this was significant only for the higher dose (1.8, 0.4, and 3.5 for the 7.5/46, 15/92, and placebo groups, respectively; 15/92 vs. placebo: p=0.0125).143 Another RCT of 130 patients (69% female; mean age 49.6; mean BMI 35.4) reported that a significantly greater percentage of P/T subjects achieved ADA-targeted HbA1c levels of <7.0% compared with placebo subjects (53% vs. 40%; p<0.05). In addition, 18.7% of P/T patients ©Institute for Clinical and Economic Review, 2015 Page 65 [Return to Table of Contents] decreased the number of anti-diabetic medications taken during the 28-week study period versus 5.5% in the placebo group (significance not reported).138 A final RCT randomized 45 patients (47% female; mean age 52.4; mean BMI 35.7) with obstructive sleep apnea (OSA) to receive P/T (15/92 mg) or placebo.139 By Week 28, the authors found significant improvements in OSA and related symptoms with P/T versus placebo: AHI (apnea- hypopnea index) decreased 31.5 points in P/T patients compared to a 16.6 point reduction in placebo patients (p=0.0084). Outcomes in BMI <35 We found no comparative studies involving subgroups of patients with BMI <35 that met our inclusion criteria. Impact on Other Outcomes We identified two studies that assessed changes in QoL. The CONQUER trial found greater improvements in most measurements of the IWQOL-Lite and SF-36 scales in both doses of P/T than with placebo (itemized data not reported).137 Alternatively, Winslow and colleagues did not find significant differences between treatment groups (study characteristics described previously) in SF- 36 QoL parameters, with the exception of the general health perceptions score, which favored the P/T 15/92 mg group (p=0.0103).139 An additional three studies measured changes in depression symptoms using the PHQ-9.137,140,142 Two studies found improvements in PHQ-9 scores among both treatment and placebo groups, but no significant differences emerged between groups.140,142 A third study found that scores worsened slightly for all groups between baseline and follow-up but found no differences between groups.137 Harms Associated with Phentermine/Topiramate We identified five good quality RCTs that reported P/T-related AEs. Across studies, dry mouth, dysgeusia, paraesthesia, insomnia, and constipation were among the most commonly reported treatment-emergent AEs. In a 56-week extension of the CONQUER trial, authors found that the incidence of AEs attributed to P/T declined over time, generally resolved on drug discontinuation, and occurred at a higher frequency with higher doses of treatment.137,141 Discontinuation of the study drug due to AEs was most commonly attributed to insomnia, irritability, anxiety, headache, disturbance in attention, depression, dry mouth, and nephrolithiasis. Across studies, suicidality did not increase significantly, and psychiatric AEs such as depression and anxiety occurred primarily during the early phase of treatment. AEs occurred in 90.9-94.7% of ©Institute for Clinical and Economic Review, 2015 Page 66 [Return to Table of Contents] patients, and 1.3-16.0% discontinued P/T due to treatment-emergent AEs. These results are summarized in Appendix Table E6. Summary: Phentermine/Topiramate We judge there to be moderate certainty of a small net benefit of P/T plus lifestyle intervention versus lifestyle intervention alone in improving weight loss in patients with BMI levels 27-45. As with the other medications, certainty was moderate because of the modest levels of weight loss and comorbidity resolution observed. In addition, there is uncertainty around the durability of benefit from P/T as well as safety over longer-term follow-up (i.e., >2 years). We identified a total of eight good- or fair-quality reports from five RCTs that compared phentermine/topiramate extended release (P/T) combination therapy to a placebo or to phentermine or topiramate monotherapy, all but one of which were conducted in patients with BMI levels between 35 and 39.9. In these trials, patients receiving the initial recommended dose combination (7.5/46 mg) lost 7.8-8.5% of total body weight (vs. 1-2% for placebo), while the range for those receiving the higher dose (15/92 mg) was 9.2-10.9%. Patients who received any dose of P/T experienced greater improvement in obesity-related comorbidities such as T2DM, hypertension, and sleep apnea. We found no studies that had a patient population with mean BMI <35. Overall, 91-95% of patients experienced one or more AEs and 1.3-16.0% discontinued P/T due to AEs. ©Institute for Clinical and Economic Review, 2015 Page 67 [Return to Table of Contents] 5. Model of Clinical and Economic Outcomes ICER has adopted the following framework for assessing the comparative value of health care interventions, with value assessed according to two distinct constructs: • Care Value: 1. Comparative clinical effectiveness of each intervention vs. alternatives (considering both clinical benefits and harm) 2. Any additional “non-clinical” benefits (e.g., reduced caregiver burden) 3. Contextual considerations (no other acceptable treatment, vulnerable populations) 4. Cost-effectiveness (incremental cost to achieve important patient outcomes vs. alternatives) • Health System Value: 1. Care value of the intervention of interest (as above) AND 2. Potential effects of short-term budgetary impact from the intervention on other patients in the health care system 5.1 Prior Published Evidence on Costs and Cost-Effectiveness of Bariatric Surgery As clinical evidence has accumulated on bariatric surgery over more than two decades, so too have data on the costs and potential cost-effectiveness of bariatric procedures in multiple populations. Below we summarize the findings of a comprehensive systematic review on the economic impact of bariatric surgery as well as those of several key studies made available after the publication of this systematic review. Padwal et al., 2011144 Padwal and colleagues conducted a CADTH-sponsored systematic review of clinical evidence as well as information on costs and cost-effectiveness, based on available studies published through mid- January 2011 (Padwal, 2011). Economic studies were limited to those conducted for adult populations as well as to studies that adjusted estimates of survival for QoL (i.e., cost-utility studies). A total of 13 studies were evaluated, six of which were industry-sponsored. All evaluations involved comparisons of open or laparoscopic RYGB and/or LAGB, as well as usual or standard care. The primary focus of attention was on BMI levels of 35 or greater in all evaluations; in many of these, multiple BMI classes were tested. Studies varied in terms of primary data source (e.g., RCT vs. observational study), time horizon, geographic setting, and target population (e.g., all patients with obesity vs. patients with obesity and T2DM). ©Institute for Clinical and Economic Review, 2015 Page 68 [Return to Table of Contents] Despite these differences, bariatric procedures were shown to be cost-effective across all studies at willingness-to-pay thresholds <$50,000 per QALY gained over time horizons ranging from two years to lifetime. In eight of 13 studies, cost-effectiveness estimates were below $15,000 per QALY gained. Higher cost-effectiveness ratios tended to be produced over shorter time horizons (i.e., 2-5 years). One study (Picot, 2009) showed an increase in two-year cost-effectiveness ratios with declining BMI (i.e., $35,904 per QALY gained at pre-operative BMI of 37, $115,230 per QALY gained for BMI of 34), but 20-year cost-effectiveness estimates were substantially lower ($3,000-$24,000 per QALY gained). Results were generally robust in sensitivity analyses, with reported probabilities of values <$50,000 per QALY gained ranging from 84-100%. One evaluation reported that LAGB was less costly and more effective than standard care on a lifetime basis, but only if diabetes remission lasted longer than 10 years;145 LAGB was no longer considered cost-effective when remission was less than two years in duration. More recent economic evaluations focused on relevant US populations are summarized below. Weiner et al., 2013146 This was not a simulation model but a matched retrospective review of nearly 60,000 individuals enrolled at seven Blue Cross/Blue Shield health plans nationwide.146 Patients were matched on an obesity-related propensity score that included BMI and obesity-related comorbidity data, as well as on age, sex, availability of prescription drug coverage, and plan location. An evaluation of regression-adjusted costs for each of the six years following surgery showed that mean annual costs increased significantly in the second and third years after surgery (by $500-$1,000) but then declined to pre-operative levels thereafter. By contrast, costs remained relatively stable in the nonsurgical group throughout follow-up. As with the clinical studies highlighted in our evidence review, this claims-based analysis also saw significant patient attrition, as <10% of the original study sample was available at year 6 of follow- up. However, mean annual costs of care were higher in surgical vs. nonsurgical patients when patients available in each of the six years of the evaluation were assessed separately, particularly for inpatient services; the authors suggest that future studies should focus on the effects of bariatric surgery on overall health and well-being rather than its potential to produce a medical cost-offset. Maciejewski et al., 2012147 Another propensity-matched analysis compared 847 military veterans undergoing gastric bypass to 847 nonsurgical controls at 12 VA medical centers nationwide. Among surgical patients, 6-month expenditures steadily increased over the 3-year period prior to surgery, peaking at nearly $30,000 higher than nonsurgical controls during the 6-month period that included surgery. By 12 months ©Institute for Clinical and Economic Review, 2015 Page 69 [Return to Table of Contents] after surgery, costs had converged to levels similar to that of nonsurgical patients and remained similar through three years of postsurgical follow-up. Findings were similar to that of the Weiner analysis in that expenditures in the surgical group were never lower than those for nonsurgical patients at any point during follow-up. Wang et al., 2014148 Wang and colleagues developed a two-part simulation model to estimate the effects of bariatric procedures: a decision-analytic model focused on 1) the shorter-term (five-year) cost impact of surgery vs. standard care, and 2) a lifetime natural history model examining the possible trajectory of BMI change and its related effects beyond five years.148 Analyses were conducted for a 53 year- old female with a BMI of 44. On a lifetime basis, the cost-effectiveness of laparoscopic RYGB, open RYGB, and LAGB vs. standard care were estimated to be $6,600, $17,200, and $6,200 per QALY gained, respectively, based on available epidemiologic data on BMI change. Findings were similar when postsurgical BMI was assumed to remain stable. When patients were assumed to regain all weight by 15 years after surgery, cost-effectiveness estimates eroded somewhat but remained well below $50,000 per QALY gained for laparoscopic RYGB and LAGB, and only slightly above for open RYGB ($59,500 per QALY gained). Several commentaries have explored possible reasons why bariatric procedures appear to be cost- effective but do not produce overall cost savings. These include a duration of follow-up too short to capture cost offsets associated with reduced rates of cardiovascular disease, cancer, and premature death;149 a paucity of economic study on sleeve gastrectomy, which may strike an appropriate balance between effectiveness and long-term safety;150 and restriction of bariatric surgery as a “last resort” treatment in patients whose health care costs are too high to fully offset.149 Prior Published Evidence on Costs and Cost-Effectiveness of Other Obesity-Related Treatments We did not find any prior publications addressing the costs and/or cost-effectiveness of the Maestro vBloc device or N/B therapy, the other treatment options of interest in this review. ICER Simulation Model To augment the available evidence on the economic impact of bariatric surgery and newer types of devices and medications, and to compare all procedures of interest in this evaluation, we developed our own decision-analytic model. The primary focus of attention in our model was on all four bariatric procedures of interest (RYGB, LAGB, VSG, and BPD±DS). However, in order to better understand the impact of device- and medication-based treatment either alone or in sequence with surgery, we included the one currently FDA-approved device (vBloc) and the medication with the ©Institute for Clinical and Economic Review, 2015 Page 70 [Return to Table of Contents] largest evidence base at an approved dose for obesity management (naltrexone/bupropion extended-release) from our review scope. All intervention strategies were assessed in comparison to standard nonsurgical management for all individuals with obesity (BMI ≥30) and in classes defined by BMI range (i.e., 30-34.9, 35-39.9, and ≥40). 5.2 Methods: Care Value Type of Economic Evaluation As in Wang et al. above,148 we developed a two-part model for this evaluation. We first conducted a cost-consequence analysis over a one-year time horizon to assess the immediate clinical and economic effects of bariatric surgery, a newer medication, and a newer type of device. For the bariatric surgery procedures, this analysis compared change in BMI, and proportions of patients with perioperative mortality, reoperation, and medical complications. For N/B extended-release, we considered discontinuation due to AEs, while we included revision rates for the vBloc device. Costs of interest included those of treatment, reoperation, management of complications, and total costs. In addition, to explore the potential impact of obesity and its treatment on quantity and quality of life (QoL), a cost-utility analysis was also conducted over a ten-year time horizon based on assumed trajectories of BMI change after the various surgical, pharmacological, and device interventions. All analyses were conducted using Microsoft Excel 2010 (Microsoft Corporation, Seattle, Washington). Target Population The target population of the decision model included adults with obesity (BMI ≥30). We did not include pediatric populations in our modeling because of the paucity of comparative clinical evidence for each of the procedures of interest. We conducted an analysis for all patients with obesity (BMI ≥30) as well as for various classifications of obesity: moderate obesity (BMI 30.0 – 34.9), severe obesity (BMI 35.0 – 39.9), and morbid obesity (BMI ≥40). Study Perspective We adopted a public payer perspective for the reference case (i.e., primary analysis). In other words, costs were assumed to be those borne by the payer for services rendered. Patient out-of- pocket costs (e.g., copays, deductibles) were not considered. Indirect costs (e.g., lost work time, caregiver burden) were not included in the model. ©Institute for Clinical and Economic Review, 2015 Page 71 [Return to Table of Contents] Interventions We evaluated the cost-effectiveness of the four types of bariatric surgery of focus in this evidence review, as well as the newer medication and type of device: RYGB, LAGB, VSG, and BPD±DS, N/B pharmacotherapy, and the vBloc device. These latter interventions were assessed as stand-alone treatment and in sequence with surgery reserved for treatment failures. Comparator The reference case analysis compared the various forms of bariatric surgery and newer types of devices and medications with conventional weight-loss treatments, including other prescription medication, dietary supplements, diet-control programs, exercise, psychotherapy, and nutritional counseling. Conventional treatments may have been delivered individually or in combination. We also conducted analyses comparing LAGB, VSG, and BPD±DS to RYGB as the most widespread form of bariatric surgery in the US. Decision Modeling The model was structured to incorporate the findings of RCTs that were included in the clinical review. The RCT outcomes were limited because of the short period of follow-up and use of surrogate outcomes such as BMI changes. Two main models were constructed: 1) a short-term model using RCT data related to percent change in BMI, complications and comorbidities at one year; and 2) a longer-term ten-year model that includes short-term outcomes from RCTs and incorporates estimates of QoL, mortality, and comorbid illness using observational study data over 10 years. For the long-term model, a Markov process was used to estimate costs and clinical outcomes in one-year cycles (see Figure 11 on page 74). The costs and effectiveness of each Markov cycle were assigned based on the characteristics of survivors and equations relating these characteristics to costs and QoL (e.g., BMI, age of survivors). The model outputs included QALYs, life-years gained, reduction in BMI, total health care costs, and incremental cost per QALY gained. Summary estimates of health care costs and QALYs were derived based on equations exploring their relationship with levels of BMI (as opposed to costs/QALYs being derived from individual obesity-related complications). BMI The initial BMI is based on the classification of obesity considered: all patients with obesity (BMI ≥30), moderate obesity (BMI 30.0 – 34.9), severe obesity (BMI 35.0 – 39.9), and morbid obesity (BMI ≥ 40). Patients with moderate obesity, severe obesity, and morbid obesity were assumed to have mean baseline BMIs of 32.5, 37.5 and 45, respectively, while all patients with BMI of 40 or greater were assumed to have a BMI of 40. The percent change in BMI at one year between ©Institute for Clinical and Economic Review, 2015 Page 72 [Return to Table of Contents] bariatric surgery strategies was based on the data derived from the evidence review. We used percent change in BMI versus absolute change in BMI because the former translates better across the various obesity sub-populations considered in the model.151 The percent change for each intervention was then applied to the mean baseline BMI for each class to calculate the absolute expected reduction in BMI. For example, a 10% change in BMI for a given intervention would yield a reduction of 3.3 BMI points for a starting mean BMI of 32.5. After one year, subsequent changes in BMI were based on the results of observational studies. We assumed a 20% worsening in BMI change over 20 years for primarily restrictive procedures (RYGB, LAGB, VSG). For BPD, we assumed that the weight change is constant throughout, as the evidence suggests that primarily malabsorptive procedures may be better at sustaining weight loss.152 In the absence of data, we assumed that values remained constant for N/B, vBloc, and standard care. The trajectory of change in BMI is depicted graphically in Figure 12 on page 75. ©Institute for Clinical and Economic Review, 2015 Page 73 [Return to Table of Contents] Figure 11. Decision Model for Short- and Long-term Economic Outcomes of Weight-Loss Treatment ©Institute for Clinical and Economic Review, 2015 Page 74 [Return to Table of Contents] Figure 12. Percentage Decrease in BMI by Duration of Follow-up for Treatment Strategies in Decision Model Mortality The risk of perioperative mortality among patients undergoing any of the bariatric procedures was 1.4%. The relative risk of perioperative mortality in patients receiving each type of bariatric surgery was based on mortality among participants who were identified in the clinical review as undergoing each of the procedures. The differences in short-term mortality by surgical approach were based on the calculated relative risk of mortality in the first year. RYGB, VSG, LAGB, and BPD±DS were assumed to be associated with relative risks of perioperative mortality of 0.47, 0.52, 0.58, and 1.09 respectively. We assumed no treatment-related mortality risk for N/B, vBloc, or standard care. The risk of mortality among patients in subsequent years is based on age and BMI (see Appendix F). For standard care, N/B, and vBloc, we multiplied mortality rates in US Life Tables by BMI-specific mortality relative risks derived from the published literature.153,154 We assumed that bariatric surgery was associated with a reduction in the risk of mortality (RR 0.71; 95% CI 0.54, 0.92) in years 2+ for all bariatric surgeries versus standard care based on long-term data from the SOS study;11 given the controversy over this topic, we also conducted alternative analyses assuming no mortality benefit. ©Institute for Clinical and Economic Review, 2015 Page 75 [Return to Table of Contents] Quality of Life Improvements in QoL are thought to be a key benefit of weight loss. We derived the estimates of BMI-specific utilities from a regression analysis of EuroQOL five dimensions questionnaire (EQ-5D) data from 2013.155 In this study, the factors associated with change in HRQoL as assessed by the EQ- 5D between baseline and six-month follow-up were baseline EQ-5D score, baseline BMI, baseline number of comorbidities, and change in BMI. The relationship among the various factors and HRQoL as assessed by EQ-5D were as follows: Δ EQ-5D=0.71995-0.68279*EQ-5DBaseline-0.00285*BMIBaseline -0.00957*NoComorb+0.0073*ΔBMI Based on this, differential gains in HRQoL will be observed among the various BMI sub-populations considered. For example, patients with more severe obesity (i.e., BMI ≥40) will achieve greater gains in HRQoL than those with less severe obesity (i.e., BMI ≥30-34.9) given that patients with severe obesity (i.e., BMI ≥40) are more likely to have lower HRQoL at baseline, have more comorbidities, and achieve a greater reduction in weight. A detailed table of our estimates is in Appendix F. Time Horizon Various time horizons were considered. A one-year time frame focusing largely on clinical benefits and short-term complications was considered, as well a longer-term horizon of 10 years. We also considered five- and 25-year time horizons via sensitivity analysis. Complications of Bariatric Surgery, N/B Pharmacotherapy, and the vBloc Device The risk and relative risk of short-term complications was derived from the clinical review for each procedure of interest. The overall rate of reoperation was estimated to be 9.9%, to which relative risk estimates of 0.63, 0.32, 1.18, and 0.71 were applied for RYGB, VSG, LAGB, and BPD±DS, respectively. The rate of medically-managed complications was 11.8%; corresponding relative risk estimates by procedure were 0.93, 1.10, 0.14, and 1.76, respectively. We assumed 5.6% of patients who received the vBloc system would have the device removed by the end of year one due to complications. We assumed that 9.8% of patients using N/B therapy would discontinue due to AEs. For participants undergoing bariatric surgery or subsequent surgery, it was assumed that the QoL was reduced by 0.21153 for six weeks to account for surgery and recovery for all procedures except LAGB (a 4-week recovery was assumed). Medical complications were associated with a utility decrement of 0.11153 over two weeks, while reoperations were associated with a decrement of 0.32153 over four weeks. We made a simplifying assumption that all surgical complications would occur within the first year of the index surgery. We conservatively assumed that N/B discontinuation ©Institute for Clinical and Economic Review, 2015 Page 76 [Return to Table of Contents] and vBloc removal were not associated with a reduction in QoL, as treatment was assumed to stop immediately when AEs were experienced. Estimating Resources and Costs Direct costs for bariatric procedures were considered from the payer perspective; reimbursement rates from a public payer (the Washington State Health Care Authority [HCA]) were used based on data from a prior evidence review (see Table 7 below).156 Estimates of direct costs included professional and technical fees as well as facility charges for the bariatric surgery procedures. The cost of the vBloc system was assumed to be $17,500 based on the hospital list price supplied by the manufacturer.27 Finally, the cost of N/B pharmacotherapy was assumed to be approximately $1,800 annually, based on the published wholesale acquisition cost32 less the mandated Medicaid price rebate (23.1%) for innovative medications.157 Table 7. Costing Data for Health Economic Analysis Cost Element Total Costs Source Gastric bypass $24,277 Washington HCA156 Gastric banding $17,483 Washington HCA156 Biliopancreatic diversion (with or without $36,160 Washington HCA156 duodenal switch) Sleeve gastrectomy $18,788 Survey of surgeons from state of Washington Maestro® system $17,500 Manufacturer 27 Naltrexone/bupropion extended-release $1,807* REDBOOK, 201532 Medically managed complications $5,625 Washington HCA156 Surgically managed complications $12,673 Washington HCA156 Standard nonsurgical care $3,746 Østbye 2014158 Mortality $41,503 Wang 2014148 Annual costs – Obesity BMI (30-34.9) $3,246 Østbye 2014158 Annual costs – Obesity BMI (35-39.9) $3,783 “158 Annual costs – Obesity BMI (40+) $4,028 “158 % change in costs per BMI, Males 3.93% % change in costs per BMI, Females 2.18% “158 % change in costs per BMI, All 2.97% * 1,414 pills per year at $1.66 wholesale cost per pill (net of 23.1% Medi-Cal discount); ©Institute for Clinical and Economic Review, 2015 Page 77 [Return to Table of Contents] Health care costs for short-term surgical complications were also derived from the Washington HCA, while other costs at varying levels of BMI were derived from a recently published US study that reported costing data by BMI level for our population of interest.158 We assumed that each unit of BMI decrease was associated with an approximate 3% decrease in health care expenditures. Currency, Price Date, and Conversion All costs are provided in 2014 US dollars. Analytical Method In addition to stratifying results by BMI, several univariate sensitivity and variability analyses were conducted to explore the effect of varying parameter values and assumptions within the model. These included the following factors of interest: time horizon, cost of bariatric procedure, mortality benefit for bariatric surgery, variation in BMI trajectory, and relationships between BMI and costs/QALYs. We also conducted an analysis investigating the effect of treatment sequencing, in which we assumed patients using N/B or vBloc therapy who were unsuccessful with initial treatment would subsequently require RYGB surgery. In this analysis, 18% of patients using N/B are assumed to achieve “success” (10% or more total weight loss) and continue therapy, while 82% would switch to RYGB in year 2. For vBloc, 24% are assumed to achieve success (10% or greater total weight loss) and remain on vBloc, while 76% would switch to RYGB in year two. We also effect separate sensitivity analyses varying the discontinuation/removal and success rates for N/B and vBloc between 50% and 200% of the base case estimate. 5.3 Results: Care Value Reference Case Analysis – Costs and health consequences of bariatric surgery, N/B pharmacotherapy, vBloc, and nonsurgical standard care over one-year time frame When compared with standard care in all patients with obesity (BMI≥30), the use of N/B, vBloc, RYGB, VSG, LAGB, and BPD±DS was associated with approximate decrements in BMI of 3.0, 3.8, 10.4, 9.8, 7.8, and 12.5 respectively (vs. 1.4 for standard care). Corresponding one-year costs were $5,355, $22,574, $30,099, $24,357, $22,035, and $42,979 respectively vs. $3,710 for standard care (see Table 8 on the next page). Mortality rates were similar among bariatric procedures but reoperation rates were lowest for VSG and highest for LAGB, while medical complication rates were highest for VSG and BPD±DS. The rates of co-morbidity resolution were also similar among bariatric procedures but lowest for LAGB. Nearly 6% of vBloc patients had the device removed. ©Institute for Clinical and Economic Review, 2015 Page 78 [Return to Table of Contents] We also stratified results by BMI sub-categories; note that data on deaths, reoperations, and medical complications are not repeated in these stratified analyses since rates are identical to those in the overall population. Overall, the findings for BMI are more favorable for patients with a BMI ≥40 compared with those with lower BMI. For example, patients undergoing RYGB with a pre- operative BMI ≥40 achieved larger absolute and % reductions in BMI (11.7, 29%) compared with those who had BMI 30-34.9 (8.45, 26%). The same trend occurred for other treatments, although the differences were less pronounced for N/B and vBloc given their more modest treatment effects. Total costs were similar across BMI classes for patients undergoing the four procedures but did increase in the standard care group as BMI increased, owing to the greater complexity of managing patients at higher levels of BMI. Table 8. Costs and Consequences of Bariatric Surgery, N/B Pharmacotherapy, vBloc System, and Nonsurgical Standard Care over 1 year of Follow-up, among all Patients with BMI >30 Standard Outcome/Cost N/B vBloc RYGB VSG LAGB BPD±DS Care BMI ≥30 Clinical Outcome BMI loss (mean) 1.4 3.0 3.8 10.4 9.8 7.8 12.5 Death (%) 1% 1% 1% 2% 2% 2% 2% Reoperation (%) 0% 0% 6%** 6% 3% 12% 7% Medical 0% 0% 4% 11% 13% 2% 21% complication (%) Costs ($) Procedure $0 $1,645 $17,500 $24,277 $18,788 $15,987 $36,160 Reoperation $0 $0 $710 $787 $402 $1,478 $893 Other $3,710 $3,710 $4,364 $5,035 $5,167 $4,570 $5,925 Complications* TOTAL $3,710 $5,355 $22,574 $30,099 $24,357 $22,035 $42,979 BMI 30-34.9 BMI loss (mean) 1.2 2.4 3.1 8.5 7.9 6.3 10.1 Total Cost ($)* $3,042 $4,687 $21,813 $29,338 $23,596 $21,274 $42,218 BMI 35-39.9 BMI loss (mean) 1.4 2.8 3.6 9.8 9.2 7.3 11.7 Total Cost ($)* $3,500 $5,145 $22,384 $29,909 $24,167 $21,845 $42,789 BMI ≥40 BMI loss (mean) 1.6 3.4 4.3 11.7 11.0 8.7 14.0 Total Cost ($)* $4,269 $5,914 $23,359 $30,884 $25,142 $22,820 $43,764 NOTE: Because of rounding, performing calculations may not produce the exact results shown. * Includes age-related background health care costs for obesity derived from Østbye 2014. ** 5.6% of vBloc patients assumed to have the device removed ©Institute for Clinical and Economic Review, 2015 Page 79 [Return to Table of Contents] Reference Case Analysis – Cost-effectiveness of bariatric surgery, N/B extended-release, Maestro® vBloc system and nonsurgical standard care over 10 year of follow-up, among all patients with BMI >30 In the 10-year time horizon analysis, all active interventions of interest in this analysis, including surgery, device, and drug-based treatment, resulted in additional QALYs and increased costs compared with standard care (see Table 9 on the next page). The use of RYGB was associated with a gain of approximately 0.5 QALYs and incremental costs of nearly $20,000 ($54,110 vs. $34,923 for the standard care strategy). This led to an incremental cost per QALY of $37,423 for RYGB. VSG and LAGB are less costly but less effective than RYGB, while BPD±DS is more expensive and more effective. The cost per QALY gained for BPD±DS was $77,574 in comparison to RYGB across all levels of BMI. However, in comparison to standard care, cost-effectiveness estimates are similar for all surgery types (ranging from $29,000 - $47,000 per QALY gained). In contrast, the cost per QALY estimates for N/B therapy and the vBloc system relative to standard care were $131,250 and $109,543, respectively, owing to the more modest effects of these treatments on BMI. In keeping with clinical results at one year of follow-up, cost-effectiveness values were most favorable in patients with a BMI ≥40. For example, RYGB produced 0.57 QALYs vs. standard care in these patients (vs. 0.41 in those with BMI 30-34.9) and was associated with incremental costs of approximately $18,000 (vs. $22,000 in patients with lower levels of obesity). As a result, the cost- effectiveness of RYGB in individuals with morbid obesity was approximately $31,000 per QALY gained (vs. $53,000 in patients with BMI 30-34.9). Differences were similar for the more effective but more expensive BPD±DS procedure (cost-effectiveness ratios of ~$39,000 and ~$63,000 for BMI ≥40 and 30-34.9, respectively); differences were more pronounced for the less effective but less expensive N/B (~$123,000 and ~$173,000 respectively and vBloc therapies (~$102,000 and ~$143,000 respectively). ©Institute for Clinical and Economic Review, 2015 Page 80 [Return to Table of Contents] Table 9. Cost-effectiveness of Bariatric Procedures Over A 10-year Time Horizon, by Procedure and Preoperative BMI Level BMI Effectiveness Cost-effectiveness ($/QALY gained) Cost ($) Level/Procedure (QALYs) Vs. SC Vs. RYGB BMI≥30 Standard care $34,923 7.5680 N/A Less expensive & less effective N/B $47,732 7.6656 $131,250 Less expensive & less effective vBloc $51,471 7.7191 $109,543 Less expensive & less effective RYGB $54,110 8.0807 $37,423 N/A VSG $48,702 8.0417 $29,087 Less expensive & less effective LAGB $47,668 7.9252 $35,680 Less expensive & less effective BPD±DS $65,741 8.2307 $46,508 $77,574 BMI 30-34.9 Standard care $27,943 7.9418 N/A Less expensive & less effective N/B $41,722 8.0213 $173,469 Less expensive & less effective vBloc $45,487 8.0645 $143,023 Less expensive & less effective RYGB $49,735 8.3529 $53,021 N/A VSG $44,298 8.3211 $43,122 Less expensive & less effective LAGB $42,738 8.2273 $51,826 Less expensive & less effective BPD±DS $61,410 8.4730 $63,011 $97,194 BMI 35-39.9 Standard care $32,538 7.6567 N/A Less expensive & less effective N/B $46,184 7.7482 $149,212 Less expensive & less effective vBloc $49,994 7.7982 $123,392 Less expensive & less effective RYGB $52,886 8.1351 $42,534 N/A VSG $47,468 8.0986 $33,789 Less expensive & less effective LAGB $46,217 7.9898 $41,073 Less expensive & less effective BPD±DS $64,533 8.2751 $51,743 $83,224 BMI ≥40 Standard care $40,329 7.2846 N/A Less expensive & less effective N/B $53,746 7.3939 $122,737 Less expensive & less effective vBloc $57,650 7.4540 $102,249 Less expensive & less effective RYGB $58,257 7.8630 $30,995 N/A VSG $53,047 7.8194 $23,784 Less expensive & less effective LAGB $52,255 7.6882 $29,552 Less expensive & less effective BPD±DS $69,329 8.0322 $38,790 $65,431 NOTE: Because of rounding, performing calculations may not produce the exact results shown. ©Institute for Clinical and Economic Review, 2015 Page 81 [Return to Table of Contents] Sensitivity Analyses – Bariatric Surgery vs. Standard Care We performed a series of one-way sensitivity analyses on key model variables. A tornado diagram comparing RYGB with standard care using a 10-year time horizon for patients with BMI ≥30 is shown in Figure 13 on the following page. The incremental cost effectiveness rations (ICERs) range from $5,444 – $84,971/QALY. The model input having the greatest impact on incremental cost- effectiveness was time horizon. As the time horizon of the analysis is extended, the incremental cost per QALY gains for bariatric surgery estimates decrease. Similarly, as the time horizon of the analysis is reduced to as low as five years, the incremental cost per QALY gains for bariatric surgery increase (see Appendix F). The model was also sensitive to the cost of bariatric surgery. In the base case, we assumed that the RYGB procedure costs were $24,277. However, in analyses varying this cost from 50% to 200% of the base case level, cost-effectiveness ratios ranged from $10,009 to $72,968 per QALY gained vs. standard care. We also investigated a best case scenario in which we assumed that the BMI reduction observed in the first year is sustained over the time horizon as opposed to diminishing by 20%. Also considered was a worst case scenario of RYGB effectiveness, where we assumed that patients returned to preoperative BMI at 5 years post-surgery. The cost per QALY estimates for the 10-year time horizons did not materially differ from the base case analysis, ranging between $35,546 and $67,381 for RYGB compared with standard care under these scenarios. We considered the impact of a varying all-cause mortality risk associated with bariatric surgery. If the base-case hazard ratio was reduced to 0.50 (versus 0.71) or increased to 1.0 (no mortality benefit), the ICER changed only slightly, to $36,651 and $39,756, respectively. We also conducted a probabilistic sensitivity analysis on all relevant parameters. Results were similar to those of deterministic analyses and are summarized in Appendix F. ©Institute for Clinical and Economic Review, 2015 Page 82 [Return to Table of Contents] Figure 13. Tornado Diagram of Bariatric Surgery (RYGB) vs. Standard Care using 10 Year Time Horizon Sensitivity analyses – Treatment Sequencing Findings of our sensitivity analysis of treatment sequencing suggested somewhat improved cost- effectiveness ratios relative to the base case analysis. For example, continuation of N/B pharmacotherapy for patients achieving success but switching to RYGB surgery for those not successful would result in a cost per QALY gained of $44,196 vs. standard care, in comparison to a level of $131,250 in the base case. This was due to a number of factors, including a low incremental rate of success (18%), cessation of costs of drug therapy after failure, and ability to perform RYGB at a lower pre-operative BMI (due to the average effects of pharmacotherapy). In addition, because a 5% weight loss threshold is frequently used as a decision point for whether to continue pharmacotherapy, we used the aggregate data from the N/B trials in an alternative analysis. The incremental rate of success was somewhat higher in this analysis (28% vs. 18% for 10% weight loss), and the corresponding cost-effectiveness ratio somewhat lower ($41,211 per QALY gained vs. $44,196 for 10% weight loss). ©Institute for Clinical and Economic Review, 2015 Page 83 [Return to Table of Contents] In contrast, the cost-effectiveness of vBloc did not change appreciably in the sequencing scenario ($104,240 per QALY gained vs. $109,543 in the base case), as all patients would incur the cost of the device, eliminating most of the benefit of performing surgery at a lower BMI level. We also conducted more broad-based sensitivity analyses that varied the success rates of N/B and vBloc therapy from 50% to 200% of the base case levels. Cost-effectiveness of N/B therapy ranged from $40,856 to $52,834 per QALY gained, while vBloc estimates ranged from $95,233 to $128,595 per QALY compared with standard care. Varying the drug discontinuation and device removal rates had similar effects. 5.4 Methods: Health-System Value Given the emerging nature and/or pre-approval status of the devices and medications of interest in this review, we opted not to assess the potential health-system impact of their introduction to one or more populations of interest. Similarly, because the use of bariatric surgery is generally well- accepted in patients with BMI levels ≥35, we did not focus on the impact of surgery in these populations. Instead, we assessed the effect of allowing coverage of bariatric surgery in patients with a BMI between 30 and 34.9. Results were applied to all adults estimated to be currently enrolled in Medi-Cal, as well as the subset of these individuals estimated to have T2DM. An analysis of merged National Health and Nutrition Examination Survey (NHANES) and Medicaid Analytic eXtract (MAX) data indicate that the prevalence of BMI 30-34.9 in a Medicaid population is 27.7%;159 a separate analysis of NHANES trend data indicates that the prevalence of diagnosed and undiagnosed diabetes in this BMI range is 11%.160 Findings were applied to a current estimate of Medi-Cal adult enrollees. Total enrollment (post- expansion) is currently estimated to be 11.7 million, of whom 56% are adults (~6.6 million).161 We assumed that 25% of patients with BMI 30-34.9 (as well as among the subset with diabetes) would opt for surgery over standard treatment in any given year. We calculated budget impact by taking the difference in year 1 costs from the care value analysis between VSG surgery and standard care; we chose VSG as the lowest-cost bariatric procedure still in widespread use (as noted previously, LAGB use has declined precipitously in the state). We also explored the effects if 10-year cost offsets were included, again based on data from the care value model. The overall PMPM Medi-Cal premium was estimated to be $552, based on data from Kaiser State Health Facts updated to 2014 levels.162,163 ©Institute for Clinical and Economic Review, 2015 Page 84 [Return to Table of Contents] 5.5 Results: Health System Value Findings from our health-system value analysis are presented in Table 10 below. One year costs for standard obesity treatment are estimated to total $3,042 per patient, including costs of treatment and management of obesity-related complications. Offering VSG surgery would add $20,554 in costs over that of standard treatment, including the costs of surgery, complications, and ongoing patient management. Of the 6.6 million adults currently enrolled in Medi-Cal, approximately 1.8 million are estimated to have a BMI in the range of 30-34.9. Replacing standard treatment with VSG surgery in 25% of these patients (~450,000) would increase one-year expenditures by approximately $9 billion, or approximately $66 PMPM, representing a 12% increase in the base PMPM of $552. Restricting availability of VSG to the subset of these patients with diabetes (~50,000) would increase expenditures by approximately $1 billion, or about $7 on a PMPM basis (1.3%). We also adjusted cost differences for any offsets that would be realized from BMI improvements with surgery at 10 years. Data were again obtained from the care value analysis, and the difference in costs between VSG and standard care declines from $20,554 to $16,355 (a 20% cost offset). Under this scenario, use of VSG in 25% of Medi-Cal adults with BMI of 30-34.9 would result in a PMPM impact of approximately $53 (9.6%), whereas use of VSG in 25% of those with a BMI of 30- 34.9 and diabetes would increase expenditures by ~$6 PMPM (1.1%). Table 10. Estimated Budgetary Impact of Use of Vertical Sleeve Gastrectomy among Medi-Cal Enrollees with BMI 30-34.9 Adults with BMI 30-34.9 All Enrollees Adults with BMI 30-34.9 Measure and diabetes (25% receive (11.7 million) (25% Receive Surgery) surgery) One-Year Costs* Total Expenditures $77,500,800,000 $9,325,884,204 $1,025,847,262 PMPM $552 $66.42 $7.31 % Increase --- 12.03% 1.32% One-Year Costs with Offset** Total Expenditures $77,500,800,000 $7,420,688,730 $816,275,760 PMPM $552 $52.85 $5.81 % Increase --- 9.57% 1.05% *Assumed one-year difference in cost between VSG and standard care of $20,554 **Adjusted one-year difference for downstream cost offsets of $16,355 We recognize that surgical capacity may not be sufficient to accommodate the volume assumed in this analysis. Under an alternative assumption that 10% of patients with a BMI of 30-34.9 ©Institute for Clinical and Economic Review, 2015 Page 85 [Return to Table of Contents] (~180,000) would opt for surgery, costs would rise by approximately $3.7 billion, or $27 PMPM (a 4.8% increase). Restricting availability to the approximately 20,000 patients with a BMI of 30-34.9 and Type 2 diabetes would increase costs by $410 million, or ~$3 PMPM (a 0.5% increase). When cost offsets are considered, increases would be 3.8% among all patients with BMI 30-34.9 and 0.4% for the subset of these patients with T2DM. 5.6 Summary Published evidence accumulated to date on care value suggests that bariatric surgery meets commonly-accepted thresholds for cost-effectiveness in comparison to standard care across multiple BMI classes, time horizons, and procedure types. By contrast, there are no currently published estimates of cost-effectiveness for newly-approved devices and drugs such as vBloc and Contrave®. To better inform the comparisons of interest in our review, we developed a simulation model to compare the care value of four bariatric surgery procedures, the Maestro® vBloc system, and N/B extended-release (Contrave®) to conventional weight-loss management for all individuals with obesity (BMI ≥30), as well as for classes defined by BMI range (i.e., 30-34.9, 35-39.9, and ≥40). Over the course of one year, across all levels of BMI and procedure type, we found that all interventions improved BMI levels but were subject to varying levels of complications and AEs, as well as increased overall costs. Over a 10-year timeframe, each intervention also resulted in improved quality-adjusted survival due primarily to the beneficial effects of lower weight on QoL. The more prominent weight loss from surgery also lowered obesity-related costs, offsetting the initial costs of surgery by as much as 30% over 10 years. Cost-effectiveness estimates for bariatric surgery over 10 years ranged from approximately $24,000 to $63,000 per quality-adjusted life year (QALY) gained vs. conventional treatment, which is within commonly-accepted thresholds for cost-effectiveness (i.e., $50,000-$100,000 per QALY gained). These findings were robust to a range of sensitivity analyses, including elimination of mortality benefit for bariatric surgery and complete weight regain five years after surgery. Importantly, while the most favorable results were seen in patients with BMI ≥40 due to greater weight loss (and corresponding gains in QoL), surgery produces cost-effectiveness ratios within the commonly- accepted range among those with a BMI level of current policy interest (30-34.9), with findings ranging from $43,000 - $63,000 per QALY gained vs. conventional treatment. In contrast, the much more modest weight loss achieved with the vBloc device and N/B pharmacotherapy, coupled with their high implantation and ongoing therapy costs, respectively, resulted in much higher cost-effectiveness ratios (>$100,000 per QALY vs. conventional treatment). Results were more favorable when these treatment options were considered “in sequence” with ©Institute for Clinical and Economic Review, 2015 Page 86 [Return to Table of Contents] bariatric surgery for those failing initial treatment, in particular for a “drug first” regimen in which those with successful weight loss at one year continued to receive medication while patients requiring surgery were able to receive surgery after an initial weight loss. Given the emerging nature of the evidence on devices and the modest benefits afforded by newer medications, we opted to focus our health-system value analysis on the use of bariatric surgery in patients with a BMI of 30-34.9. The potential budgetary impact of adopting vertical sleeve gastrectomy in 10-25% of all patients in this BMI class is substantial – between $27 and $66 PMPM, a 5-12% increase over the current total Medi-Cal expenditure rate of $552. When availability is restricted to the subset of these patients who have diabetes, however, the budgetary impact declines to approximately $3-$7 PMPM, a 0.5-1.3% increase over current total expenditures. It is not known how many people in this population would opt for surgery, even if offered. We note certain limitations of our analysis. First, the cost-effectiveness and budgetary impact of all of the interventions of interest are highly dependent on assumptions related to initial BMI loss and forecasted change in BMI over any given time horizon. To address this limitation, we employed several BMI trajectory scenarios and found that results were robust under several assumptions. In addition, our budgetary impact analysis assumed that between 10% and 25% of eligible patients with a BMI of 30-34.9 would opt for surgery in any given year. Our review has documented multiple issues with access to accredited and comprehensive bariatric programs in this country, and studies conducted in select populations suggest rates of utilization of <5%. 164–167 Finally, there was considerable variation in patient populations, study design, and other features across studies, which limits the comparability of clinical evidence among interventions. As a result, we needed to make assumptions about the comparative clinical effects in our models. Rigorous, long-term studies are needed to better characterize the clinical benefits of drugs, devices, and surgical procedures for obesity, particularly in relation to durability of weight loss and comorbidity remission over the long term. ©Institute for Clinical and Economic Review, 2015 Page 87 [Return to Table of Contents] 6. Questions and Discussion 6.1 About the CTAF Process During CTAF public meetings, the CTAF Panel deliberates and votes on key questions related to the systematic review of the clinical evidence, an economic analysis of the applications of the medical technologies or treatments under examination, and the supplementary information presented. Panel members typically serve for two or more years and are intentionally selected to represent a range of expertise and diversity in perspective. To maintain the objectivity of the CTAF Panel and ground the conversation in the interpretation of the published evidence, they are not pre-selected based on the topic being addressed. Acknowledging that any judgment of evidence is strengthened by real-life clinical and patient perspectives, subject matter experts are recruited for each meeting topic and provide input to CTAF Panel members before the meeting to help clarify their understanding of the different interventions being analyzed in the evidence review. The same clinical experts serve as a resource to the CTAF Panel during their deliberation, and they help form recommendations with CTAF on ways the evidence can be applied to policy and practice. At each meeting, after the CTAF Panel vote, a policy roundtable discussion is held with the CTAF Panel, clinical experts, and representatives from provider groups, payers, and patient groups. This is intended to bring stakeholders into the discussion on how best to apply the evidence to guide patient education, clinical practice, and coverage policies. Participants on policy roundtables are selected for their expertise on the specific meeting topic, are different for each meeting, and do not vote on any questions. At the July 14, 2015 meeting, the CTAF Panel discussed issues regarding the application of the available evidence to help patients, providers, and payers address the important questions related to multiple surgical procedures, devices, and medications for the treatment of overweight and obesity. Following the evidence presentation and public comments, the CTAF Panel voted on key questions concerning the comparative clinical effectiveness and comparative value of various treatments. These questions are developed by the ICER research team for each assessment, with input from the CTAF Advisory Board to ensure that the questions are framed to address the issues that are most important in applying the evidence to support clinical practice and medical policy decisions. The voting results are presented below, along with comments reflecting considerations mentioned by CTAF Panel members during the voting process. In its deliberations and voting related to value, the CTAF Panel made use of a value assessment framework with four different components of care value, which is a judgment comparing the clinical outcomes, average per-patient costs, and broader health effects of two alternative interventions or approaches to care. The four components of care value are comparative clinical ©Institute for Clinical and Economic Review, 2015 Page 88 [Return to Table of Contents] effectiveness, incremental cost per outcomes achieved, additional benefits, and contextual considerations regarding the illness or therapy. Once they made an overall assessment of care value considering these four components, the CTAF Panel then explicitly considered the affordability of several treatments for overweight and obesity in assessing health system value as low, intermediate, or high (see Figure 14 below and Figure 15 on the next page, as well as the detailed explanation that follows). Figure 14. Care Value Framework There are four elements to consider when deliberating on care value: 1. Comparative clinical effectiveness is a judgment of the overall difference in clinical outcomes between two interventions (or between an intervention and placebo), tempered by the level of certainty possible given the strengths and weaknesses of the body of evidence. CTAF now uses the ICER Evidence Rating Matrix as its conceptual framework for considering comparative clinical effectiveness. 2. Incremental cost per outcomes achieved is the average per-patient incremental cost of one intervention compared to another to achieve a desired “health gain,” such as an additional stroke prevented, case of cancer diagnosed, or gain of a year of life. Alternative interventions are compared in terms of cost per unit of effectiveness, and the resulting comparison is presented as a ratio: a “cost per outcome achieved.” Relative certainty in the cost and outcome estimates continues to be a consideration. 3. Additional benefits refers to any significant benefits offered by the intervention to caregivers, the delivery system, or other patients in the health care system that would not have been captured in the available “clinical” evidence. Examples of additional benefits include mechanisms of treatment delivery that require many fewer visits to the clinician’s office, treatments that reduce disparities across various patient groups, and new potential mechanisms of action for treating clinical conditions (e.g., mental illness) that have demonstrated low rates of response to currently available therapies. For each intervention evaluated, it will be open to discussion whether additional benefits such as these are ©Institute for Clinical and Economic Review, 2015 Page 89 [Return to Table of Contents] important enough to factor into the overall judgment of care value. There is no quantitative measure for additional benefits. 4. Contextual considerations can include ethical, legal, or other issues (but not cost) that influence the relative priority of illnesses and interventions. Examples of contextual considerations include whether there are currently any existing treatments for the condition, whether the condition severely affects quality of life or not, and whether the condition affects priority populations. There is no quantitative measure for the role of contextual considerations in an overall judgment of care value. CTAF uses this conceptual description of the elements of care value when deliberating on the evidence and voting. In assessing health system value, the CTAF Panel was asked to vote whether interventions represent a “high,” “intermediate,” or “low” value from the generalized perspective of a state Medicaid program. Figure 15. Health System Value Framework Health system value is a judgment of the affordability of the short-term budget impact that would occur with a change to a new care option for all eligible patients, assuming the current price and payment structure. Usually, the care value and the health care system value of an intervention or approach to care will align, whether it is “high,” “intermediate,” or “low.” For example, a treatment that is judged to represent high care value from the perspective of per-patient costs and benefits will almost always represent a high health system value as well. But health system value also takes into consideration the short-term effects of the potential budget impact of a change in care across the entire population of patients. Rarely, when the additional per-patient costs for a new care option are multiplied by the number of potential patients treated, the short-term budget impact of a new intervention of intermediate or even high care value could be so substantial that the intervention would be “unaffordable” unless the health system severely restricts its use, delays or cancels other valuable care programs, or undermines access to affordable health insurance for all patients by sharply increasing health care premiums. Under these circumstances, unmanaged change to a new ©Institute for Clinical and Economic Review, 2015 Page 90 [Return to Table of Contents] care option could cause significant harm across the entire health system, in the short-term possibly even outweighing the good provided by use of the new care option itself. To consider this possibility, CTAF reviews estimates of the potential budget impact for a change in care as measured by the estimated increase in “per-member-per-month” health care premiums that would be needed to fund a new care option in its first year of use were all eligible patients to be treated. It should be noted that if, after considering potential budget impact, a health intervention judged to have high care value receives a judgment of “low” health system value from the CTAF Panel, this does not imply that the health system should not adopt the intervention; rather, the vote indicates that policy makers should consider implementing mechanisms related to patient selection, step therapy, pricing, and/or financing to ensure that the short-term budget impact of a high care value intervention does not lead to more harm than good. CTAF votes on health system value will therefore serve an important function by highlighting situations when policymakers need to take action and work together to align care value with health system value. 6.2 Summary of the Votes Clinical Effectiveness (based on the evidence presented) Surgical procedures and devices 1. For adult patients with a BMI of 30.0-34.9 and Type 2 diabetes, is the evidence adequate to demonstrate that the net health benefit of bariatric surgery is greater than that of conventional weight-loss management? CTAF Panel Vote: 11 Yes (100%) 0 No (0%) 2. For adult patients with a BMI 30.0-34.9 who do not have Type 2 diabetes, is the evidence adequate to demonstrate that the net health benefit of bariatric surgery is greater than that of conventional weight-loss management? CTAF Panel Vote: 1 Yes (9%) 10 No (91%) 3. For adult patients with a BMI ≥35, is the evidence adequate to demonstrate that RYGB and VSG have a comparable net health benefit? CTAF Panel Vote: 11 Yes (100%) 0 No (0%) ©Institute for Clinical and Economic Review, 2015 Page 91 [Return to Table of Contents] 4. For adult patients with a BMI ≥35, is the evidence adequate to demonstrate that the net health benefit of the Maestro vBloc system is greater than that of conventional weight-loss management? CTAF Panel Vote: 6 Yes (55%) 5 No (45%) Comment: Some CTAF Panel members expressed concerns about the large placebo effect and high dropout rate in the one published trial available for review at the meeting. Drugs 5. For adult patients with a BMI ≥30 or BMI ≥27 with an obesity-related comorbidity, is the evidence adequate to distinguish the net health benefit among m the four drugs reviewed? CTAF Panel Vote: 0 Yes (0%) 11 No (100%) Value Surgical procedures and devices 6. [If yes to question 1 – adult patients with a BMI of 30.0-34.9 and Type 2 diabetes] Given the available evidence, what is the care value of bariatric surgery vs. conventional weight-loss management? CTAF Panel Vote: 2 Low (18%) 6 Intermediate (55%) 3 High (27%) Comment: Panelists voting intermediate or high care value noted the likelihood that bariatric surgery would resolve type 2 diabetes as well as the favorable incremental cost- per-QALY presented in the ICER report. The rationale from a panelist voting low care value was the lack of data on a patient population more representative of Medicaid members – a more diverse group that is likely burdened with more severe obesity and that has less access to surgical and support services. In written notes, CTAF Panel members offered their thoughts on the components of care value. With regard to comparative clinical effectiveness, seven CTAF members judged bariatric surgery to provide superior benefit, while four determined the benefit to be incremental. One panelist highlighted the lack of evidence on long-term outcomes for patients undergoing surgery. In the additional benefits category, it was noted that patients could have fewer comorbidities after surgery and there would potentially be better long-term outcomes for patients with class I obesity since treatment would be offered earlier in the disease course. m The CTAF Panel understood this question to mean distinguishing among or between these four medications. ©Institute for Clinical and Economic Review, 2015 Page 92 [Return to Table of Contents] 7. [If yes to question 1] Given the available evidence, what is the overall health system value of bariatric surgery for a state Medicaid program? CTAF Panel Vote: 3 Low (27%) 5 Intermediate (45%) 3 High (27%) Comment: In the general care value description above, the budget impact is based on an assumption that all eligible patients would seek treatment in one year. For this report, however, the budget impact was modeled with two scenarios: one assuming 25% of eligible Medi-Cal patients would seek bariatric surgery in one year, and the other assuming 10% of eligible patients would do so. The estimated budget impact under the second scenario fell below the threshold of 0.5-1% increase in PMPM. Consistent with the statement by a policy roundtable clinical expert who said the likely number was closer to 1%, two CTAF panelists who voted intermediate or high health system value thought that the number of patients in this group likely to pursue bariatric surgery would be lower than 10% and the resulting increase in PMPM health insurance costs would therefore likely be less than 0.5-1%. 8. [If yes to question 2 – adult patients with a BMI 30.0-34.9 who do not have Type 2 diabetes] Given the available evidence, what is the care value of bariatric surgery vs. conventional weight-loss management? Vote Not Taken Comment: The CTAF Panel did not vote on this question because the majority voted there is insufficient evidence for this patient group to demonstrate bariatric surgery’s superiority to conventional weight management. 9. [If yes to question 2] Given the available evidence, what is the overall health system value of bariatric surgery for a state Medicaid program? Vote Not Taken Comment: The CTAF Panel did not vote on this question because the majority voted there is insufficient evidence for this patient group to demonstrate bariatric surgery’s superiority to conventional weight management. 10. [If yes to question 4 – adult patients with a BMI ≥35] Given the available evidence, what is the care value of the Maestro vBloc system vs. conventional weight-loss management? CTAF Panel Vote: 8 Low (73%) 3 Intermediate (27%) Comment: Panelists who chose the low care value category expressed concerns about the large placebo effect and high dropout rate in the one published trial available for review at the meeting. Some panelists noted that the device produced incremental costs per QALY over $100,000 for a modest benefit, but several felt that vBloc has the potential to provide ©Institute for Clinical and Economic Review, 2015 Page 93 [Return to Table of Contents] significant additional benefits to patients because it offers patients a less invasive treatment option than bariatric surgery and works through a new mechanism of action. In written notes, half the CTAF Panel thought that vBloc was “incremental” (high certainty of a small net health benefit), while the other half thought it fell into either the insufficient (as described in the report: low certainty of small or comparable net health benefit) or promising but inconclusive (moderate certainty of small net health benefit, with a small possibility of comparable or negative net benefit) categories in terms of comparative clinical effectiveness. Panelists pointed to the trial results of modest weight loss with a low rate of adverse events, but they expressed concern about the existence of only one clinical trial without data on comorbidity resolution at the time of the meeting. 11. [If yes to question 4] Given the available evidence, what is the overall health system value of the Maestro vBloc system for a state Medicaid program? CTAF Panel Vote: 10 Low (91%) 1 Intermediate (9%) 6.3 Roundtable Discussions and Key Policy Implications Following its deliberation on the evidence and subsequent voting, the CTAF Panel engaged in a moderated discussion with a Policy Roundtable that included clinical experts, a patient advocate, representatives of one private and one public payer, and a representative from a professional society. The policy roundtable discussion with the CTAF Panel reflected multiple perspectives and opinions, and therefore, none of the recommendations below should be taken as a consensus view held by all participants. The names of the Policy Roundtable participants are shown in Table 11 below. Table 11. Policy Roundtable Participants Joel Brill, MD, FACP, AGAF Chief Medical Officer, Predictive Health, LLC; Former chair, Practice Management and Economics Committee, American Gastroenterological Association Alexander Li, MD Deputy Director, Care Transitions, Los Angeles Department of Health Services John Morton, MD, MPH, Chief of Bariatric and Minimally Invasive Surgery, Stanford University School of FACS, FASMBS Medicine; President, American Society for Metabolic and Bariatric Surgery Sameer Murali, MD, MSHS Bariatric/Internal Medicine, Southern California Permanente Medical Group Philippe J. Quilici, MD, Chief of Minimally Invasive and Bariatric Surgery, Providence Saint Joseph Medical FACS, FASMBS Center, and Attending Surgeon, Cedars Sinai Medical Center Michelle Vicari Patient Advocate John Yao, MD, MBA, MPH, Staff Vice President of Medical Policy, Anthem MPA, FACP ©Institute for Clinical and Economic Review, 2015 Page 94 [Return to Table of Contents] The roundtable discussion was facilitated by Jed Weissberg, MD, Senior Fellow at ICER. The main themes and recommendations from the discussion are summarized below. Coverage policy 1. The CTAF panel confirmed by unanimous vote that there is accumulating evidence of positive longer-term clinical outcomes for bariatric surgery in patients who have a body mass index (BMI) of 30.0-34.9 and type 2 diabetes. Payers should consider expanding coverage for bariatric surgery to this group of patients based on its comparative clinical effectiveness. Many insurers base their current coverage policies on a 1991 statement from the National Institutes of Health (NIH) Consensus Development Conference on Gastrointestinal Surgery identifying patients with a BMI of 40 or above, or 35 and above with high-risk comorbid conditions, as appropriate candidates for bariatric surgery. These policies should be updated to reflect current evidence. The CTAF Panel voted unanimously that bariatric surgery is an effective treatment for patients with a BMI of 30.0-34.9 who also have type 2 diabetes. The majority of the CTAF Panel also voted that the care value of surgery in this group, as well as health system value, was intermediate to high. The budget impact analysis in this report was based on 10% of patients with a BMI of 30.0-34.9 who also have type 2 diabetes seeking bariatric surgery; since it is extremely unlikely that such a large number of patients would seek surgery in one year, the short-term budget impact (estimated to be a 0.5% increase in PMPM for this patient population) is likely to be manageable, particularly if patients are provided with information about the range of conventional weight-loss management approaches and medical treatments other than surgery. So that patients have health plans from which to choose and that no plan has a disproportionate number of patients with Class I obesity seeking bariatric surgery, all payers should make this change to their coverage policy. 2. Given the rapid development of therapies for obesity, more frequent reviews of payer coverage policies are required and should be based on systematic reviews and rigorous technology assessments. Both clinicians and patients have expressed a desire for a wider variety of medical treatments along the continuum of medications, devices, and bariatric surgery. Additional treatment modalities that will be appropriate for a broad variety of patient populations are anticipated. Given the rapidly evolving evidence base, payers should frequently review emerging therapies to ensure that coverage policies are up-to-date with current evidence. In the two weeks following the CTAF meeting in July 2015, for example, a new device (ReShape Integrated Dual Balloon System) was approved to treat class I and II obesity, and clinical trials for another device (Endobarrier) were suspended. 3. Bariatric surgery is rarely the initial treatment option for patients with obesity; however, it should be considered when weight-loss management approaches and medical treatments have ©Institute for Clinical and Economic Review, 2015 Page 95 [Return to Table of Contents] failed or have had insufficient benefit. Given the risks of surgery, many patients will wish to use conventional or intensive weight-loss management approaches, medical treatment options, and/or less invasive surgical procedures. Payers should structure coverage policies to help guide clinicians and patients through the available lifestyle interventions, drugs, devices, and surgical procedures as they select treatments most appropriate for the patient. A variety of treatment options are currently available for patients who are unable to lose excess weight through conventional weight-loss management approaches, with more likely to become available over time. Medicare, for example, covers weight-loss management that involves behavioral counseling and behavioral therapy to promote sustained weight loss through high intensity diet and exercise interventions (see accompanying Action Guide). Newer drugs and devices have recently been approved for patients with Class I or II obesity. An approach to coverage policy that identifies the most appropriate treatment for patients within specific obesity classes would help clinicians and patients select from the available weight-loss management techniques, drugs, devices, and surgical procedures. Some flexibility would be needed in the steps for such an approach, since, as noted in recommendation 5 below, the current classification based solely on BMI does not capture the full picture of a patient’s health. Clinical care 4. Obesity is not a surgical disorder; rather, it is a chronic condition that benefits from care that is coordinated across multiple specialty areas. As part of their care, patients with obesity should be considered for treatment by primary care physicians, bariatricians, n nutritionists, mental health and/or addiction experts, and other medical specialists who can help ensure successful treatment outcomes. If bariatric surgery is the treatment selected, surgical teams should work closely with primary care providers and bariatricians for pre-surgery consultation and post- surgery follow-up. Clinicians should give additional consideration to the physical, emotional, social, and environmental factors that contribute to a patient’s health, particularly for underserved and vulnerable populations, so that the obesity treatment chosen is most likely to be successful. The contributors to excess weight are multiple and varied, and obesity is now considered to be a chronic disease. Many patients with obesity also have other chronic health conditions requiring multiple medications and involving multiple body systems. Because of these factors, treatment approaches that involve clinicians from a variety of disciplines are needed to effectively treat patients with obesity. While some patients will lose excess weight through conventional weight-loss management approaches or use of medications, surgical implantation of a device or bariatric n A bariatrician is a licensed physician who has received special training in bariatric medicine, which is the medical treatment of overweight and obesity and its associated conditions. ©Institute for Clinical and Economic Review, 2015 Page 96 [Return to Table of Contents] surgery may be the best approach for others. For patients undergoing surgery, concomitant with the surgical procedure itself are diet, nutritional supplements, and other lifestyle modifications that may require lifetime maintenance to be successful; therefore, the patient’s care team should ensure that the patient is well-prepared prior to surgery and has access to post-surgical resources to address any complications and offer support. Care managers can play important roles in facilitating communication among the patient, primary care, and other members of the patient’s care team, and in ensuring that patients have access to long-term psychosocial support to enable continued progress toward or maintenance of weight loss. The current options for treatment vary in terms of effectiveness and safety. Clinicians and patients should work together to share clinical decisions based on matching clinical needs and risk tolerance to available options. As noted above, the variety and complexity of factors influencing obesity means that good care requires attention to factors beyond physical health such as employment (and health insurance); housing; and behavioral health conditions such as depression, anxiety, or alcohol/drug addictions. More comprehensive approaches will be needed that acknowledge and address the needs of individuals seeking treatment for obesity who are, for example, homeless, living in areas that are unsafe for exercising outdoors, or have limited access to transportation and grocery stores, as these all have the potential to reduce the effectiveness of chosen treatments. 5. Health systems and provider groups should use more sophisticated obesity classification systems beyond simple BMI. Electronic documentation will allow for stratification of the patient population by obesity class; health systems and provider groups can use this information in planning for clinical capacity and to more effectively target treatment resources to the patient population. The current classification of obesity is based solely on BMI. Since lowering a patient’s BMI alone does not necessarily reflect improvement in overall health or functioning, a classification system such as the Edmonton Obesity Staging System that measures the presence and severity of risk factors, comorbidities, and functional limitations should be used to supplement BMI. Based on the number of patients in each obesity class, health systems and provider groups can more effectively plan for the various types of providers needed to care for patients, and they can target resources to more effectively address the needs of the patient population. 6. Because of the high prevalence of obesity and the discomfort both providers and patients experience in discussing it, clinicians should develop new materials or use existing materials to engage with patients in discussions about the health effects of and treatment options for obesity. While acknowledging the importance of addressing excess weight and related health conditions, clinicians and patients both expressed a need for tools to support conversations about obesity; its ©Institute for Clinical and Economic Review, 2015 Page 97 [Return to Table of Contents] effects on health; and the risks, benefits, and outcomes of various treatment options. Experts in diet, nutrition, exercise, and behavior change, as well as patients who have successfully lost excess weight, may be helpful resources to primary care providers and patients. A variety of materials to facilitate patient-provider discussions about excess weight are available online (see accompanying Action Guide), and some clinicians may also find it useful to develop tools that are more tailored to their patient population. Payment 7. Fee-for-service payment does not adequately support treatment of obesity, which is a chronic condition requiring multidisciplinary, longitudinal care. Provider groups, payers, and other stakeholders should consider developing and implementing alternative payment structures for obesity care, and then evaluate whether or not they improve outcomes and value. Examples of alternative approaches are bundled payments for an episode of care, a “warranty” on surgical procedures, and risk sharing between device manufacturers and health plans. Measures that assess quality of care for obesity (including patient-reported outcomes) should be coupled with payment to provide incentives for high quality, efficient care. Acknowledging the changing payment landscape and the need to identify creative ways of funding obesity management as the industry evolves, some clinicians on the policy roundtable experienced in alternative payment approaches stated that approaches involving shared risk among provider groups, payers, and manufacturers have the potential to enhance both quality of care and value. With bundled payment, provider groups assume financial risk for the cost of all services associated with treatment such as bariatric surgery, as well as for costs associated with preventable complications. A “warranty” on surgical procedures may involve treating any and all complications without additional payment. These approaches provide incentives for care coordination, high- quality care, and efficient use of resources. For devices, shared risk between manufacturers and payers so that there is a reduced or refunded price if the therapy is not effective also provides incentives for efficient use of resources. Organizations seeking to implement alternative payment approaches will need to ensure compliance with state and federal regulations, and such approaches may require buy-in from state insurance commissioners, state regulators, and/or consumer advocacy organizations. Prevention 8. Organizations engaged in efforts to prevent and reduce obesity in the overall population should continue this important work. Such programs are often part of broader social policy changes that address violence, promote access to healthy foods and safe places to exercise, and seek healthier environments in which people can live, work, and play. Health care providers and systems should serve as an example and lead by creating healthier environments for their employees, patients, and visitors. ©Institute for Clinical and Economic Review, 2015 Page 98 [Return to Table of Contents] Many private and public sector entities have acknowledged the increase in rates of overweight and obesity, and the associated detriments to health, and have designed and implemented prevention programs with a goal of promoting healthy weight and helping individuals to lose excess weight through better eating and increased physical activity. Broader social policies to prevent obesity may include shared use of school playgrounds as safe places to exercise; limits on the number of liquor and convenience stores in a neighborhood; increased access to healthy food choices while eliminating unhealthy choices in work environments, cafeterias, and vending machines; and taxes on sodas or other sugary beverages. Because they can serve as important role models for other organizations whose primary focus is elsewhere, health care systems, provider groups, public and private payers, and other health care organizations should lead by creating healthier environments for their employees, patients, and visitors. **** The California Technology Assessment Forum has previously reviewed: • Bariatric Surgery for the Treatment of Type 2 Diabetes Mellitus, June 20, 2012 • Sleeve Gastrectomy as a Stand Alone Bariatric Procedure for Obesity, October 13, 2010 • Laparoscopic Adjustable Silicone Gastric Banding for Obesity, October 28, 2009 (update of 2007 review) • Laparoscopic Adjustable Gastric Banding for Morbid Obesity, February 28, 2007 This is the first CTAF review of all other interventions contained in this report. ©Institute for Clinical and Economic Review, 2015 Page 99 [Return to Table of Contents] References 1. Hurt RT, Edakkanambeth Varayil J, Mundi MS, Martindale RG, Ebbert JO. Designation of Obesity as a Disease: Lessons Learned From Alcohol and Tobacco. Curr Gastroenterol Rep. 2014;16:415. doi:10.1007/s11894-014-0415-z. 2. Centers for Disease Control and Prevention. Obesity and Overweight. 2015. http://www.cdc.gov/nchs/fastats/obesity-overweight.htm. Accessed June 18, 2015. 3. Finkelstein E a., Trogdon JG, Cohen JW, Dietz W. Annual medical spending attributable to obesity: Payer-and service-specific estimates. Health Aff. 2009;28(5):822-831. doi:10.1377/hlthaff.28.5.w822. 4. Magro DO, Geloneze B, Delfini R, Pareja BC, Callejas F, Pareja JC. Long-term weight regain after gastric bypass: A 5-year prospective study. Obes Surg. 2008;18:648-651. doi:10.1007/s11695-007-9265-1. 5. O’Neil PM, Smith SR, Weissman NJ, et al. Randomized Placebo-Controlled Clinical Trial of Lorcaserin for Weight Loss in Type 2 Diabetes Mellitus: The BLOOM-DM Study. Obesity. 2012;20(7):1426-1436. doi:10.1038/oby.2012.66. 6. Hollander P, Gupta AK, Plodkowski R, et al. Effects of naltrexone sustained- release/bupropion sustained-release combination therapy on body weight and glycemic parameters in overweight and obese patients with type2 diabetes. Diabetes Care. 2013;36(12):4022-4029. doi:10.2337/dc13-0234. 7. National Institutes of Health. Prescription medications for the treatment of obesity. Natl Inst Diabetes Dig Kidney Dis Weight Inf Network. 2013. http://win.niddk.nih.gov/publications/pdfs/prescriptionmedications.pdf. Accessed September 1, 2014. 8. National Institutes of Health. Bariatric surgery for severe obesity. Natl Inst Diabetes Dig Kidney Dis Weight Inf Netw. 2009. http://win.niddk.nih.gov/publications/pdfs/prescription_medications.pdf. Accessed September 1, 2014. 9. Colquitt JL, Pickett K, Loveman E, Frampton GK. Surgery for weight loss in adults. Cochrane database Syst Rev. 2014;8(8):CD003641. doi:10.1002/14651858.CD003641.pub4. ©Institute for Clinical and Economic Review, 2015 Page 100 [Return to Table of Contents] 10. Booth H, Khan O, Prevost T, et al. Incidence of type 2 diabetes after bariatric surgery : population-based matched cohort study. LANCET Diabetes Endocrinol. 2014;2:963-968. doi:10.1016/S2213-8587(14)70214-1. 11. Sjöström L. Review of the key results from the Swedish Obese Subjects (SOS) trial - a prospective controlled intervention study of bariatric surgery. J Intern Med. 2013;273:219-234. doi:10.1111/joim.12012. 12. U.S. Food and Drug Administration. FDA expands use of banding system for weight loss. http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm245617.htm. Accessed September 1, 2014. 13. Puzziferri N, Roshek TB, Mayo HG, Gallagher R, Belle SH, Livingston EH. Long-term Follow-up After Bariatric Surgery A Systematic Review. JAMA. 2014;312(9):934-942. doi:10.1001/jama.2014.10706. 14. Mayo Clinic. Gastric bypass surgeries: risks. http://www.mayoclinic.org/tests- procedures/bariatric-surgery/basics/risks/prc-20019138. Accessed September 1, 2014. 15. Brethauer S a., Kothari S, Sudan R, et al. Systematic review on reoperative bariatric surgery. American Society for Metabolic and Bariatric Surgery Revision Task Force. Surg Obes Relat Dis. 2014:1-21. doi:10.1016/j.soard.2014.02.014. 16. American Society for Metabolic and Bariatric Surgery. Estimate of bariatric surgery numbers. http://asmbs.org/resources/estimate-of-bariatric-surgery-numbers. Accessed September 1, 2014. 17. The Medical Letter on Drugs and Therapeutics. Diet, drugs, and surgery for weight loss. 2015. 18. U.S. Food and Drug Administration. FDA approves first-of-kind device to treat obesity. January . http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm430223.htm. Accessed April 20, 2005. 19. American Society for Metabolic and Bariatric Surgery. Bariatric Surgery Procedures. http://asmbs.org/patients/bariatric-surgery-procedures. Accessed January 1, 2015. ©Institute for Clinical and Economic Review, 2015 Page 101 [Return to Table of Contents] 20. University of Illinois Bariatric Surgery Program. Roux-en-Y Gastric Bypass (RYGB). 2015. http://hospital.uillinois.edu/Patient_Care_Services/Bariatric_Surgery/Our_Services/Roux -en-. Accessed January 1, 2015. 21. Mayo Clinic. Guide to types of weight loss surgery. http://www.mayoclinic.org/tests- procedures/bariatric-surgery/in-depth/weight-loss-surgery/art-20045334. Accessed January 1, 2015. 22. Kaleida Health. Bariatric Program: Weight Loss Surgery Options. http://www.kaleidahealth.org/Services/display.asp?pt=G&p=1301&s=690. Accessed January 1, 2015. 23. American Society for Metabolic and Bariatric Surgery. Story of Obesity Surgery. http://asmbs.org/resources/story-of-obesity-surgery. Accessed January 1, 2015. 24. Cleveland Clinic. Procedures: Sleeve Gastrectomy. https://weightloss.clevelandclinic.org/sleevegastrectomy.aspx. Accessed January 1, 2015. 25. EnteroMedics. EnteroMedics announces FDA approval of VBLOC® Vagal Blocking Therapy for the treatment of obesity. http://ir.enteromedics.com/releasedetail.cfm?ReleaseID=891316. Accessed May 1, 2015. 26. Ikramuddin S, Blackstone RP, Brancatisano A, et al. Effect of Reversible Intermittent Intra-abdominal Vagal Nerve Blockade on Morbid Obesity. JAMA. 2014;312(9):915. doi:10.1001/jama.2014.10540. 27. EnteroMedics. 2015 Maestro Device & Patient Kit Hospital Sale Pricing.manufacturer. Published 2015. 28. Rattue P. EndoBarrier Re-Implantation Feasible. Med News Today. http://www.medicalnewstoday.com/articles/245022.php. Accessed May 20, 2001. 29. Fierce Medical Devices. FDA puts hold on pivotal GI Dynamics trial for obesity device due to bacterial infection. March 2015. http://www.fiercemedicaldevices.com/story/fda- puts-hold-pivotal-gi-dynamics-trial-obesity-device-due-bacterial-infect/2015-03-06. Accessed May 20, 2005. ©Institute for Clinical and Economic Review, 2015 Page 102 [Return to Table of Contents] 30. UK Health Centre. Gastric Balloon London & UK. http://www.healthcentre.org.uk/weight-loss-surgery/weight-loss-surgery-gastric- balloon.html. Accessed May 1, 2015. 31. Novo Nordisk. Product Backgrounders: Victoza. https://www.novomedlink.com/diabetes-videos/product-backgrounders/victoza- injection-conversation.html. Accessed May 1, 2015. 32. Micromedex Healthcare Series. Truven Health Analytics. RED BOOK® Online. http://truvenhealth.com/. Published 2014. Accessed May 20, 2005. 33. U.S. Drug Enforcement Administration. 21 CFR Part 1308, Placement of Lorcaserin into Schedule IV. https://www.federalregister.gov/articles/2013/05/08/2013- 10895/schedules-of-controlled-substances-placement-of-lorcaserin-into-schedule-iv. Accessed May 1, 2015. 34. Arena Pharmaceuticals/Eisai Inc. BELVIQ for Chronic Weight Management. http://www.arenapharm.com/belviq. Accessed May 1, 2015. 35. Orexigen/Takeda Inc. Important safety information for Contrave. http://contrave.com/?gclid=Cj0KEQjwhPaqBRDG2uiHzpKLi6ABEiQAk_XXiS1Nm_74U1u_Z 44P6L9-LrK6yVZtLfZFRYOtPnNDXu8aAq1u8P8HAQ#inline-isi. Accessed May 1, 2015. 36. U.S. Drug Enforcement Administration. Drug Fact Sheets. http://www.dea.gov/druginfo/concerns.shtml. Accessed May 1, 2015. 37. Vivus I. Full prescribing information. http://www.vivus.com/docs/Qsymia_prescribing_information.pdf. Accessed March 1, 2015. 38. Cummings DE, Cohen R V. Beyond BMI: The need for new guidelines governing the use of bariatric and metabolic surgery. Lancet Diabetes Endocrinol. 2014;2:175-181. doi:10.1016/S2213-8587(13)70198-0. 39. Ahima RS, Lazar MA. The Health Risk of Obesity--Better Metrics Imperative. Science (80- ). 2013;341:856-858. 40. World Health Organization Expert Consultation. Public health Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363:157-163. ©Institute for Clinical and Economic Review, 2015 Page 103 [Return to Table of Contents] 41. U.S. Food and Drug Administration. What is a serious adverse event. http://www.fda.gov/Safety/MedWatch/ucm053087.htm. Accessed June 1, 2015. 42. U.S Preventive Services Task Force. Procedure manual - appendix VII. http://www.uspreventiveservicestaskforce.org/Page/Name/procedure-manual--- appendix-vii. Accessed June 1, 2015. 43. Pontiroli AE, Morabito A. Long-term prevention of mortality in morbid obesity through bariatric surgery. a systematic review and meta-analysis of trials performed with gastric banding and gastric bypass. Ann Surg. 2011;253:484-487. doi:10.1097/SLA.0b013e31820d98cb. 44. Sjöström L, Narbro K, Sjöström CD, et al. Effects of Bariatric Surgery on Mortality in Swedish Obese Subjects. N Engl J Med. 2007;357(8):741-752. doi:10.1056/NEJMoa1109071. 45. Adams TD, Gress RE, Smith SC, et al. Long-Term Mortality after Gastric Bypass Surgery. N Engl J Med. 2007;357(8):753-761. 46. Maciejewski ML, Livingston EH, Smith V a, et al. Survival among high-risk patients after bariatric surgery. JAMA. 2011;305(23):2419-2426. doi:10.1097/01.SA.0000407027.11707.2f. 47. Arterburn DE, Olsen MK, Smith V a, et al. Association Between Bariatric Surgery and Long-term Survival. JAMA - J Am Med Assoc. 2015;313(1):62-70. doi:10.1001/jama.2014.16968. 48. Halperin F, Ding S-A, Simonson DC, et al. Roux-en-Y Gastric Bypass Surgery or Lifestyle With Intensive Medical Management in Patients With Type 2 Diabetes: Feasibility and 1- Year Results of a Randomized Clinical Trial. JAMA Surg. 2014;02215. doi:10.1001/jamasurg.2014.514. 49. Scopinaro N, Adami GF, Papadia FS, et al. Effects of biliopanceratic diversion on type 2 diabetes in patients with BMI 25 to 35. Ann Surg. 2011;253(4):699-703. doi:10.1097/SLA.0b013e318203ae44. 50. Karlsen TI, Lund RS, Røislien J, et al. Health related quality of life after gastric bypass or intensive lifestyle intervention: a controlled clinical study. Health Qual Life Outcomes. 2013;11:17. doi:10.1186/1477-7525-11-17. ©Institute for Clinical and Economic Review, 2015 Page 104 [Return to Table of Contents] 51. O’Brien PE, Dixon JB, Laurie C, et al. Treatment of mild to moderate obesity with laparoscopic adjustable gastric banding or an intensive medical program: a randomized trial.[Summary for patients in Ann Intern Med. 2006 May 2;144(9):I12; PMID: 16670127]. Ann Intern Med. 2006;144:625-633. http://ezproxy.library.usyd.edu.au/login?url=http://ovidsp.ovid.com/ovidweb.cgi?T=JS& CSC=Y&NEWS=N&PAGE=fulltext&D=med4&AN=16670131\nhttp://dd8gh5yx7k.search.se rialssolutions.com/?sid=OVID:medline&id=pmid:16670131&id=doi:&issn=0003- 4819&isbn=&volume=144&is. 52. Padwal RS, Rueda-Clausen CF, Sharma a M, et al. Weight Loss and Outcomes in Wait- listed, Medically Managed, and Surgically Treated Patients Enrolled in a Population-based Bariatric Program: Prospective Cohort Study. Med Care. 2014;52(3):208-215. doi:10.1097/MLR.0000000000000070. 53. Kashyap SR, Bhatt DL, Wolski K, et al. Metabolic Effects of Bariatric Surgery. 2Diabetes care. 2013;36:2175-2182. doi:10.2337/dc12-1596. 54. Schauer PR, Kashyap SR, Wolski K, et al. Bariatric Surgery versus Intensive Medical Therapy in Obese Patients with Diabetes. N Engl J Med. 2012;366(17):1567-1576. doi:10.1056/NEJMoa1407764. 55. Schauer PR, Bhatt DL, Kirwan JP, et al. Bariatric Surgery versus Intensive Medical Therapy for Diabetes - 3-Year Outcomes. N Engl J Med. 2014:1-12. doi:10.1056/NEJMoa1401329. 56. Courcoulas AP, Goodpaster BH, Eagleton JK, et al. Surgical vs Medical Treatments for Type 2 Diabetes Mellitus: A Randomized Clinical Trial. JAMA Surg. 2014;15213(7):707- 715. doi:10.1001/jamasurg.2014.467. 57. Dixon JB, Brien PEO, Playfair J, et al. Adjustable Gastric Banding and Conventional Therapy for Type 2 Diabetes. JAMA. 2008;299(3):316-323. 58. Liang Z, Wu Q, Chen B, Yu P, Zhao H, Ouyang X. Effect of laparoscopic Roux-en-Y gastric bypass surgery on type 2 diabetes mellitus with hypertension: A randomized controlled trial. Diabetes Res Clin Pract. 2013;101(1):50-56. doi:10.1016/j.diabres.2013.04.005. 59. Mingrone G, Panunzi S, De Gaetano A, et al. Bariatric Surgery versus Conventional Medical Therapy for Type 2 Diabetes. N Engl J Med. 2012;366:1577-1585. doi:10.1056/NEJMoa1200111. ©Institute for Clinical and Economic Review, 2015 Page 105 [Return to Table of Contents] 60. Ikramuddin S, Korner J, Lee W-J, et al. Roux-en-Y Gastric Bypass vs Intensive Medical Management for the Control of Type 2 Diabetes, Hypertension, and Hyperlipidemia. JAMA. 2013;309(21):2240-2249. 61. Leonetti F, Capoccia D, Coccia F, et al. Obesity, Type 2 Diabetes Mellitus, and Other Comorbidities: A Prospective Cohort Study of Laparoscopic Sleeve Gastrectomy vs Medical Treatment. Arch Surg. 2012;147(8):694-700. doi:10.1001/archsurg.2012.222. 62. O’Brien PE, Brennan L, Laurie C, Brown W. Intensive medical weight loss or laparoscopic adjustable gastric banding in the treatment of mild to moderate obesity: Long-term follow-up of a prospective randomised trial. Obes Surg. 2013;23:1345-1353. doi:10.1007/s11695-013-0990-3. 63. Dixon J, Mbbs P, Schachter L, et al. Surgical vs Conventional Therapy for Weight Loss Treatment of Obstructive Sleep Apnea: A Randomized Controlled Trial. Jama. 2012;308:1142-1149. http://ovidsp.ovid.com/ovidweb.cgi?T=JS&PAGE=reference&D=ovftn&NEWS=N&AN=00 005407-201209190-00034. 64. Martins C, Strømmen M, Stavne O a, Nossum R, Mårvik R, Kulseng B. Bariatric surgery versus lifestyle interventions for morbid obesity--changes in body weight, risk factors and comorbidities at 1 year. Obes Surg. 2011;21:841-849. doi:10.1007/s11695-010-0131-1. 65. Dorman RB, Rasmus NF, Al-Haddad BJS, et al. Benefits and complications of the duodenal switch/biliopancreatic diversion compared to the Roux-en-Y gastric bypass. Surg (United States). 2012;152(4):758-767. doi:10.1016/j.surg.2012.07.023. 66. Dixon JB, Strauss BJG, Laurie C, O’Brien PE. Changes in Body Composition with Weight Loss: Obese Subjects Randomized to Surgical and Medical Programs*. Obesity. 2007;15(5):1187-1198. doi:10.1038/oby.2007.639. 67. O’Brien PE, Sawyer SM, Laurie C, et al. Laparoscopic adjustable gastric banding in Severely Obese Adolescents. JAMA - J Am Med Assoc. 2010;303(6):519-526. 68. Messiah SE, Lopez-Mitnik G, Winegar D, et al. Changes in weight and co-morbidities among adolescents undergoing bariatric surgery: 1-year results from the Bariatric Outcomes Longitudinal Database. Surg Obes Relat Dis. 2013;9:503-513. doi:10.1016/j.soard.2012.03.007. ©Institute for Clinical and Economic Review, 2015 Page 106 [Return to Table of Contents] 69. Courcoulas AP, Belle SH, Neiberg RH, et al. Three-Year Outcomes of Bariatric Surgery vs Lifestyle Intervention for Type 2 Diabetes Mellitus Treatment. JAMA Surg. 2015:E1-E9. doi:10.1001/jamasurg.2015.1534. 70. Angrisani L, Cutolo PP, Formisano G, Nosso G, Vitolo G. Laparoscopic adjustable gastric banding versus Roux-en-Y gastric bypass: 10-year results of a prospective, randomized trial. Surg Obes Relat Dis. 2013;9:405-416. doi:10.1016/j.soard.2006.12.005. 71. Angrisani L, Lorenzo M, Borrelli V. Laparoscopic adjustable gastric banding versus Roux- en-Y gastric bypass: 5-year results of a prospective randomized trial. Surg Obes Relat Dis. 2007;3(2):127-132. doi:10.1016/j.soard.2006.12.005. 72. Nguyen NT, Slone J a, Nguyen X-MT, Hartman JS, Hoyt DB. A prospective randomized trial of laparoscopic gastric bypass versus laparoscopic adjustable gastric banding for the treatment of morbid obesity: outcomes, quality of life, and costs. Ann Surg. 2009;250(4):631-641. doi:10.1097/SLA.0b013e3181b92480. 73. Søvik TT, Taha O, Aasheim ET, et al. Randomized clinical trial of laparoscopic gastric bypass versus laparoscopic duodenal switch for superobesity. Br J Surg. 2010;97:160-166. doi:10.1002/bjs.6802. 74. Søvik TT, Aasheim ET, Taha O, et al. Weight Loss, Cardiovascular Risk Factors, and Quality of Life After Gastric Bypass and Duodenal Switch. Ann Intern Med. 2011;155:281-291. 75. Risstad H, Søvik TT, Engström M, et al. Five-Year Outcomes After Laparoscopic Gastric Bypass and Laparoscopic Duodenal Switch in Patients With Body Mass Index of 50 to 60. JAMA Surg. 2015:E1-E10. doi:10.1001/jamasurg.2014.3579. 76. Adams TD, Davidson LE, Litwin SE, et al. Health Benefits of Gastric Bypass Surgery After 6 Years. JAMA. 2012;308(11):1122-1131. 77. Karmali S, Brar B, Shi X, Sharma AM, De Gara C, Birch DW. Weight recidivism post- bariatric surgery: A systematic review. Obes Surg. 2013;23:1922-1933. doi:10.1007/s11695-013-1070-4. 78. Silberhumer GR, Miller K, Pump A, et al. Long-term results after laparoscopic adjustable gastric banding in adolescent patients: follow-up of the Austrian experience. Surg Endosc. 2011;25:2993-2999. doi:10.1007/s00464-011-1658-5. ©Institute for Clinical and Economic Review, 2015 Page 107 [Return to Table of Contents] 79. Nguyen NT, Paya M, Stevens CM, Mavandadi S, Zainabadi K, Wilson SE. The relationship between hospital volume and outcome in bariatric surgery at academic medical centers. Ann Surg. 2004;240(4):586-593; discussion 593-594. doi:10.1097/01.sla.0000140752.74893.24. 80. Livingston EH, Elliott AC, Hynan LS, Engel E. When Policy Meets Statistics: The Very Real Effect that Questionable Statistical Analysis Has on Limiting Health Care Access for Bariatric Surgery. Arch Surg. 2007;142(10):979-987. 81. MBSAQIP. Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program Standards and Pathways Manual. 2014. http://www.mbsaqip.org/docs/Resources for Optimal Care of the MBS Patient.pdf. 82. Flum DR, Kwon S, MacLeod K, et al. The Use, Safety and Cost of Bariatric Surgery Before and After Medicare's National Coverage Decision. Ann Surg. 2011;254(6):860-865. doi:10.1097/SLA.0b013e31822f2101. 83. Kwon S, Wang B, Wong E, et al. The Impact of Accreditation on Safety and Cost of Bariatric Surgery. Surg Obes Relat Dis. 2013;9(5):617-622. doi:10.1016/j.biotechadv.2011.08.021.Secreted. 84. Nguyen NT, Hohmann S, Slone J, Varela E, Smith BR, Hoyt D. Improved bariatric surgery outcomes for Medicare beneficiaries after implementation of the medicare national coverage determination. Arch Surg. 2010;145(1):72-78. doi:10.1001/archsurg.2009.228. 85. Dimick JB, Nicholas LH, Ryan AM, Thumma JR, Birkmeyer JD. Bariatric surgery complications before vs after implementation of a national policy restricting coverage to centers of excellence. JAMA. 2013;309(8):792-799. doi:10.1016/j.jemermed.2013.05.040. 86. Livingston EH. Bariatric surgery outcomes at designated centers of excellence vs nondesignated programs. Arch Surg. 2009;144(4):319-325; discussion 325. doi:10.1016/S0739-5930(09)79326-2. 87. Telem DA, Talamini M, Altieri M, Yang J, Zhang Q, Pryor AD. The effect of national hospital accreditation in bariatric surgery on perioperative outcomes and long-term mortality. Surg Obes Relat Dis. 2015;11(4):749-757. doi:10.1016/j.soard.2014.05.012. 88. Kohn GP, Galanko J a., Overby DW, Farrell TM. High Case Volumes and Surgical Fellowships Are Associated with Improved Outcomes for Bariatric Surgery Patients: A ©Institute for Clinical and Economic Review, 2015 Page 108 [Return to Table of Contents] Justification of Current Credentialing Initiatives for Practice and Training. J Am Coll Surg. 2010;210(6):909-918. doi:10.1016/j.jamcollsurg.2010.03.005. 89. Morton JM, Garg T, Nguyen N. Does Hospital Accreditation Impact Bariatric Surgery Safety ? Ann Surg. 2014;260(3):504-508. doi:10.1097/SLA.0000000000000891. 90. Jafari MD, Jafari F, Young MT, Smith BR, Phalen MJ, Nguyen NT. Volume and outcome relationship in bariatric surgery in the laparoscopic era. Surg Endosc Other Interv Tech. 2013;27(12):4539-4546. doi:10.1007/s00464-013-3112-3. 91. Birkmeyer NJO, Dimick JB, Share D, et al. Hospital Complication Rates with Bariatric Surgery in Michigan. JAMA. 2010;304(4):435-442. 92. Chen T, Godebu E, Horgan S, Mirheydar HS, Sur RL. The Effect of Restrictive Bariatric Surgery on Urolithiasis. J Endourol. 2013;27(2):242-244. doi:10.1089/end.2012.0408. 93. Chevallier J-M, Paita M, Rodde-Dunet M-H, et al. Predictive factors of outcome after gastric banding: a nationwide survey on the role of center activity and patients’ behavior. Ann Surg. 2007;246(6):1034-1039. doi:10.1097/SLA.0b013e31813e8a56. 94. El Chaar M, McDeavitt K, Richardson S, Gersin KS, Kuwada TS, Stefanidis D. Does patient compliance with preoperative bariatric office visits affect postoperative excess weight loss? Surg Obes Relat Dis. 2011;7(6):743-748. doi:10.1016/j.soard.2010.10.020. 95. Compher CW, Hanlon A, Kang Y, Elkin L, Williams NN. Attendance at clinical visits predicts weight loss after gastric bypass surgery. Obes Surg. 2012;22:927-934. doi:10.1007/s11695-011-0577-9. 96. Gould JC, Beverstein G, Reinhardt S, Garren MJ. Impact of routine and long-term follow- up on weight loss after laparoscopic gastric bypass. Surg Obes Relat Dis. 2007;3:627-630. doi:10.1016/j.soard.2007.07.005. 97. Pontiroli AE, Fossati A, Vedani P, et al. Post-surgery adherence to scheduled visits and compliance, more than personality disorders, predict outcome of bariatric restrictive surgery in morbidly obese patients. Obes Surg. 2007;17:1492-1497. doi:10.1007/s11695- 008-9428-8. 98. Elakkary E, Elhorr A, Aziz F, Gazayerli MM, Silva YJ. Do support groups play a role in weight loss after laparoscopic adjustable gastric banding? Obes Surg. 2006;16(248):331- 334. doi:10.1381/096089206776116499. ©Institute for Clinical and Economic Review, 2015 Page 109 [Return to Table of Contents] 99. Nijamkin MP, Campa A, Samiri Nijamkin S, Sosa J. Comprehensive behavioral- motivational nutrition education improves depressive symptoms following bariatric surgery: A randomized, controlled trial of obese hispanic Americans. J Nutr Educ Behav. 2013;45(6):620-626. doi:10.1016/j.jneb.2013.04.264. 100. Nijamkin MP, Campa A, Sosa J, Baum M, Himburg S, Johnson P. Comprehensive Nutrition and Lifestyle Education Improves Weight Loss and Physical Activity in Hispanic Americans Following Gastric Bypass Surgery: A Randomized Controlled Trial. J Acad Nutr Diet. 2012;112(3):382-390. doi:10.1016/j.jada.2011.10.023. 101. Fuller NR, Pearson S, Lau NS, et al. An intragastric balloon in the treatment of obese individuals with metabolic syndrome: A randomized controlled study. Obesity. 2013;21(8):1561-1570. doi:10.1002/oby.20414. 102. Ponce J, Quebbemann BB, Patterson EJ. Prospective, randomized, multicenter study evaluating safety and efficacy of intragastric dual-balloon in obesity. Surg Obes Relat Dis. 2013;9:290-295. doi:10.1016/j.soard.2012.07.007. 103. Genco A, Balducci S, Bacci V, et al. Intragastric balloon or diet alone? A retrospective evaluation. Obes Surg. 2008;18(8):989-992. doi:10.1007/s11695-007-9383-9. 104. Genco A, Cipriano M, Bacci V, et al. Intragastric balloon followed by diet vs intragastric balloon followed by another balloon: A prospective study on 100 patients. Obes Surg. 2010;20(11):1496-1500. doi:10.1007/s11695-010-0231-y. 105. Genco A, Maselli R, Frangella F, et al. Effect of consecutive intragastric balloon (BIB®) plus diet versus single BIB® plus diet on eating disorders not otherwise specified (EDNOS) in obese patients. Obes Surg. 2013;23(12):2075-2079. doi:10.1007/s11695-013-1028-6. 106. Konopko-Zubrzycka M, Baniukiewicz a., Wróblewski E, et al. The effect of intragastric balloon on plasma ghrelin, leptin, and adiponectin levels in patients with morbid obesity. J Clin Endocrinol Metab. 2009;94(5):1644-1649. doi:10.1210/jc.2008-1083. 107. Takihata M, Nakamura A, Aoki K, et al. Comparison of intragastric balloon therapy and intensive lifestyle modification therapy with respect to weight reduction and abdominal fat distribution in super-obese Japanese patients. Obes Res Clin Pract. 2014;8(4):e331- e338. doi:10.1016/j.orcp.2013.07.002. 108. Peker Y, Coskun H, Bozkurt S, Cin N, Atak T, Genc H. Comparison of results of laparoscopic gastric banding and consecutive intragastric balloon application at 18 ©Institute for Clinical and Economic Review, 2015 Page 110 [Return to Table of Contents] months: a clinical prospective study. J Laparoendosc Adv Surg Tech A. 2011;21(6):471- 475. doi:10.1089/lap.2010.0439. 109. Tayyem R, Obondo C, Ali A. Short-term outcome and quality of life of endoscopically placed gastric balloon and laparoscopic adjustable gastric band. Saudi J Gastroenterol. 2011;17(6):400. doi:10.4103/1319-3767.87182. 110. Alfa Wali M, Ashrafian H, Schofield KL, et al. Is Social Deprivation Associated with Weight Loss Outcomes Following Bariatric Surgery? A 10-Year Single Institutional Experience. Obes Surg. 2014:2126-2132. doi:10.1007/s11695-014-1290-2. 111. Genco A, Cipriano M, Materia A, et al. Laparoscopic sleeve gastrectomy versus intragastric balloon: A case-control study. Surg Endosc Other Interv Tech. 2009;23(8):1849-1853. doi:10.1007/s00464-008-0285-2. 112. Fernandes M, Atallah a. N, Soares BGO, et al. Intragastric balloon for obesity. Cochrane Database Syst Rev. 2007;(1). doi:10.1002/14651858.CD004931.pub2. 113. Mathus-Vliegen EMH, Alders PRH, Chuttani R, Scherpenisse J, Mathus- EMH, Centre AM. Outcomes of intragastric balloon placements in a private practice setting. Endoscopy. 2015;47:302-307. 114. Melissas J, Mouzas J, Filis D, et al. The intragastric balloon - Smoothing the path to bariatric surgery. Obes Surg. 2006;16(7):897-902. doi:10.1381/096089206777822188. 115. Lecumberri E, Krekshi W, Matía P, et al. Effectiveness and safety of air-filled balloon heliosphere BAG® in 82 consecutive obese patients. Obes Surg. 2011;21(10):1508-1512. doi:10.1007/s11695-010-0314-9. 116. Dastis SN, François E, Deviere J, et al. Intragastric balloon for weight loss: Results in 100 individuals followed for at least 2.5 years. Endoscopy. 2009;41:575-580. doi:10.1055/s- 0029-1214826. 117. Koehestanie P, de Jonge C, Berends FJ, Janssen IM, Bouvy ND, Greve JWM. The Effect of the Endoscopic Duodenal-Jejunal Bypass Liner on Obesity and Type 2 Diabetes Mellitus, a Multicenter Randomized Controlled Trial. Ann Surg. 2014;260(6):984-992. doi:10.1097/SLA.0000000000000794. ©Institute for Clinical and Economic Review, 2015 Page 111 [Return to Table of Contents] 118. Schouten R, Rijs CS, Bouvy ND, et al. A multicenter, randomized efficacy study of the EndoBarrier Gastrointestinal Liner for presurgical weight loss prior to bariatric surgery. Ann Surg. 2010;251(2):236-243. doi:10.1097/SLA.0b013e3181bdfbff. 119. De Moura EGH, Orso IRB, Martins BDC, et al. Improvement of insulin resistance and reduction of cardiovascular risk among obese patients with type 2 diabetes with the duodenojejunal bypass liner. Obes Surg. 2011;21(7):941-947. doi:10.1007/s11695-011- 0387-0. 120. Sarr MG, Billington CJ, Brancatisano R, et al. The EMPOWER Study: Randomized, prospective, double-blind, multicenter trial of vagal blockade to induce weight loss in morbid obesity. Obes Surg. 2012;22(11):1771-1782. doi:10.1007/s11695-012-0751-8. 121. Shikora SA, Wolfe BM, Apovian CM, et al. Sustained Weight Loss with Vagal Nerve Blockade but Not with Sham: 18-Month Results of the ReCharge Trial. J Obes. 2015;2015:1-8. 122. Cleveland Clinic. Clinical Trial Testing Safety of Obesity Drug Contrave Halted; 50 Percent Interim Data Released by the Study’s Executive Committee.http://my.clevelandclinic.org/about-cleveland-clinic/newsroom/releases- videos-newsletters/2015-5-12-clinical-trial-testing-safety-of-obesity-drug-contrave- halted. Published May 15, 2015. 123. Takeda Pharmaceuticals USA. Inc. Takeda Pharmaceuticals and Orexigen Therapeutics Announce Termination of the Cardiovascular Outcomes Study (Light Study) of the Obesity Drug Contrave® (naltrexone HCl and bupropion HCl).http://www.takeda.us/newsroom/press_release_detail.aspx?id=334&year=2015. Published May 12, 2015. 124. Herper M. The FDA is Forcing Orexigen to Do a Second Safety Study Because of Contrave Disclosures. Forbes. http://www.forbes.com/sites/matthewherper/2015/03/03/the-fda- will-force-orexigen-to-do-a-second-safety-study-because-of-contrave-disclosures/. Accessed May 1, 2015. 125. US Food and Drug Administration. FDA approves weight-management drug Contrave. 2014. http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm413896.htm. Accessed June 23, 2015. ©Institute for Clinical and Economic Review, 2015 Page 112 [Return to Table of Contents] 126. Astrup A, Rössner S, Van Gaal L, et al. Effects of liraglutide in the treatment of obesity: a randomised, double-blind, placebo-controlled study. Lancet. 2009;374(9701):1606-1616. doi:10.1016/S0140-6736(09)61375-1. 127. Wadden T, Hollander P, Klein S, et al. Weight maintenance and additional weight loss with liraglutide after low-calorie-diet-induced weight loss: The SCALE Maintenance randomized study. Int J Obes. 2013;37:1443-1451. doi:10.1038/ijo.2013.162. 128. Astrup A, Carraro R, Finer N, et al. Safety, tolerability and sustained weight loss over 2 years with the once-daily human GLP-1 analogue, liraglutide. Int J Obes. 2012;36(6):890- 890. doi:10.1038/ijo.2011.203. 129. Pi-Sunyer X, Astrup A, Fujioka K, et al. A Randomized, Controlled Trial of 3.0 mg of Liraglutide in Weight Management. N Engl J Med. 2015;373(1):11-22. doi:10.1056/NEJMoa1411892. 130. Lean M, Carraro R, Finer N, et al. Tolerability of nausea and vomiting and associations with weight loss in a randomized trial of liraglutide in obese, non-diabetic adults. Int J Obes. 2014;38(5):689-697. doi:10.1038/ijo.2013.149. 131. Smith SR, Weissman NJ, Anderson CM, et al. Multicenter, Placebo-Controlled Trial of Lorcaserin for Weight Management. N Engl J Med. 2010;363(3):245-256. doi:10.1056/NEJMc1009975. 132. Fidler MC, Sanchez M, Raether B, et al. A one-year randomized trial of lorcaserin for weight loss in obese and overweight adults: The BLOSSOM Trial. J Clin Endocrinol Metab. 2011;96(10):3067-3077. doi:10.1210/jc.2011-1256. 133. Apovian CM, Aronne L, Rubino D, et al. A randomized, phase 3 trial of naltrexone SR/bupropion SR on weight and obesity-related risk factors (COR-II). Obesity. 2013;21(5):935-943. doi:10.1002/oby.20309. 134. Greenway FL, Dunayevich E, Tollefson G, et al. Comparison of combined bupropion and naltrexone therapy for obesity with monotherapy and placebo. J Clin Endocrinol Metab. 2009;94(12):4898-4906. doi:10.1210/jc.2009-1350. 135. Greenway FL, Fujioka K, Plodkowski R a., et al. Effect of naltrexone plus bupropion on weight loss in overweight and obese adults (COR-I): A multicentre, randomised, double- blind, placebo-controlled, phase 3 trial. Lancet. 2010;376(9741):595-605. doi:10.1016/S0140-6736(10)60888-4. ©Institute for Clinical and Economic Review, 2015 Page 113 [Return to Table of Contents] 136. Wadden T a, Foreyt JP, Foster GD, et al. Weight Loss With Naltrexone SR/Bupropion SR Combination Therapy as an Adjunct to Behavior Modification: The COR-BMOD Trial. Obesity. 2011;19(1):110-120. doi:10.1038/oby.2010.147. 137. Gadde KM, Allison DB, Ryan DH, et al. Effects of low-dose, controlled-release, phentermine plus topiramate combination on weight and associated comorbidities in overweight and obese adults (CONQUER): A randomised, placebo-controlled, phase 3 trial. Lancet. 2011;377(9774):1341-1352. doi:10.1016/S0140-6736(11)60205-5. 138. Garvey WT, Ryan DH, Bohannon NJ V., et al. Weight-Loss Therapy in Type 2 Diabetes: Effects of Phentermine and Topiramate Extended Release. Diabetes Care. 2014;37(12):3309-3316. doi:10.2337/dc14-0930. 139. Winslow DH, Bowden CH, DiDonato KP, McCullough P a. A Randomized, Double-Blind, Placebo-Controlled Study of an Oral, Extended-Release Formulation of Phentermine/Topiramate for the Treatment of Obstructive Sleep Apnea in Obese Adults. Sleep. 2012. doi:10.5665/sleep.2204. 140. Aronne LJ, Wadden T a., Peterson C, Winslow D, Odeh S, Gadde KM. Evaluation of phentermine and topiramate versus phentermine/topiramate extended-release in obese adults. Obesity. 2013;21(11):2163-2171. doi:10.1002/oby.20584. 141. Garvey W, Ryan D, Look M, et al. Two-year sustained weight loss and metabolic benefits with controlled-release phentermine. Am J Clin Nutr. 2012;95(2):297-308. doi:10.3945/ajcn.111.024927.INTRODUCTION. 142. Allison DB, Gadde KM, Garvey WT, et al. Controlled-Release Phentermine/Topiramate in Severely Obese Adults: A Randomized Controlled Trial (EQUIP). Obesity. 2012;20(2):330- 342. doi:10.1038/oby.2011.330. 143. Garvey WT, Ryan DH, Henry R, et al. Prevention of type 2 diabetes in subjects with prediabetes and metabolic syndrome treated with phentermine and topiramate extended release. Diabetes Care. 2014;37(4):912-921. doi:10.2337/dc13-1518. 144. Padwal R, Klarenbach S, Wiebe N, et al. Bariatric surgery: A systematic review of the clinical and economic evidence. J Gen Intern Med. 2011;26:1183-1194. doi:10.1007/s11606-011-1721-x. ©Institute for Clinical and Economic Review, 2015 Page 114 [Return to Table of Contents] 145. Keating CL, Dixon JB, Moodi ML, et al. Cost-Effectiveness of Surgically Induced Weight Loss for the Management of Type 2 Diabetes: Modeled Lifetime Analysis. Diabetes Care. 2009;32(4):567-574. doi:10.2337/dc08-1749. 146. Weiner JP, Goodwin SM, Chang H-Y, et al. Impact of bariatric surgery on health care costs of obese persons: a 6-year follow-up of surgical and comparison cohorts using health plan data. JAMA Surg. 2013;148:555-562. doi:10.1001/jamasurg.2013.1504. 147. Maciejewski ML, Livingston EH, Smith VA, Kahwati LC, Henderson WG, Arterburn DE. Health Expenditures Among High-Risk Patients After Gastric Bypass and Matched Controls. Arch Surg. 2012;147(7):633-640. doi:10.1001/archsurg.2012.818. 148. Wang BCM, Wong ES, Alfonso-Cristancho R, et al. Cost-effectiveness of bariatric surgical procedures for the treatment of severe obesity. Eur J Health Econ. 2014;15(3):253-263. doi:10.1007/s10198-013-0472-5. 149. Goldfine A, Vernon A, Zinner M. Where Are the Health Care Cost Savings With Bariatric Surgery in Obesity Management? JAMA Surg. 2014;149(1):5-6. doi:10.1001/jamasurg.2013.3060. 150. Maciejewski ML, Arterburn DE. Cost-effectiveness of Bariatric Surgery. JAMA - J Am Med Assoc. 2013;310(7):742-743. doi:10.10001/jama.2013.276131. 151. Benoit SC, Hunter TD, Francis DM, De La Cruz-Munoz N. Use of bariatric outcomes longitudinal database (BOLD) to study variability in patient success after bariatric surgery. Obes Surg. 2014;24(6):936-943. doi:10.1007/s11695-014-1197-y. 152. Dolan K, Hatzifotis M, Newbury L, Fielding G. A Comparison of Laparoscopic Adjustable Gastric Banding and Biliopancreatic Diversion in Superobesity. 2004:165-169. 153. Campbell J, McGarry L, Shikora S, Hale B, Lee J, Weinstein M. Cost-effectiveness of laparoscopic gastric banding and bypass for morbid obesity. Am J Manag Care. 2010;16(7):e174-e187. 154. Flegal K, Graubard B, Williamson D, Gail M. Excess deaths associated with underweight, overweight, and obesity. JAMA. 2005;293(15):1861-1867. 155. Rothberg AE, McEwen LN, Kraftson AT, et al. The impact of weight loss on health-related quality-of-life: implications for cost-effectiveness analyses. Qual Life Res. 2014;23(4):1371-1376. doi:10.1007/s11136-013-0557-8. ©Institute for Clinical and Economic Review, 2015 Page 115 [Return to Table of Contents] 156. Ollendorf DA, Cameron CG, Loos AM, Synnott PG, Pearson SD. Final Appraisal Document: Bariatric Surgery.; 2015. http://www.hca.wa.gov/hta/Documents/bariatric_final_rpt_040315.pdf. 157. Centers for Medicare and Medicaid Services. Medicaid Drug Rebate Program. 2015. http://www.medicaid.gov/Medicaid-CHIP-Program-Information/By- Topics/Benefits/Prescription-Drugs/Medicaid-Drug-Rebate-Program.html. Accessed June 10, 2015. 158. Østbye T, Stroo M, Eisenstein EL, Peterson B, Dement J. Is Overweight and Class I Obesity Associated with Increased Health Claims Costs ? 2014;22(4):1179-1186. doi:10.1002/oby.20669. 159. Dodd AH, Gleason PM, Mathematica Policy Research. Using the MAX-NHANES Merged Data to Evaluate the Association of Obesity and Medicaid Costs.; 2013. 160. Gregg EW, Cadwell BL, Cheng YJ, et al. Trends in the prevalence and ratio of diagnosed to undiagnosed diabetes according to obesity levels in the U.S. Diabetes Care. 2004;27(12):2806-2812. doi:10.2337/diacare.27.12.2806. 161. Gorn D. Medi-Cal Enrollment Jumps to 11.3 Million. California Healthline. http://www.californiahealthline.org/capitol-desk/2014/11/medical-jumps-to-11-3- million. Published November 12, 2014. Accessed May 1, 2015. 162. The Henry J. Kaiser Family Foundation. State Health Facts. 2013. http://kff.org/statedata/. Accessed May 1, 2015. 163. U.S. Bureau of Labor Statistics. Consumer Price Index. 2014. http://www.bls.gov/cpi/cpid1501.pdf. Accessed May 1, 2015. 164. Kim K, White V, Buffington CK. Utilization rate of bariatric surgery in an employee-based healthcare system following surgery coverage. Obes Surg. 2010;20(11):1575-1578. http://link.springer.com/article/10.1007%2Fs11695-010-0193-0. 165. Trus TL, Pope GD, Finlayson SRG. National trends in utilization and outcomes of bariatric surgery. Surg Endosc Other Interv Tech. 2005;19(5):616-620. http://link.springer.com/article/10.1007%2Fs00464-004-8827-8. ©Institute for Clinical and Economic Review, 2015 Page 116 [Return to Table of Contents] 166. Reames BN, Birkmeyer NJ, Dimick JB, et al. Variation in the Care of Surgical Conditions: Obesity; A Dartmouth Atlas of Health Care Series.; 2014. http://www.dartmouthatlas.org/downloads/reports/Obesity_report_09_16_14.pdf. 167. Funk LM, Jolles S, Fischer LE, Voils CI. Patient and Referring Practitioner Characteristics Associated With the Likelihood of Undergoing Bariatric Surgery. JAMA Surg. July 2015. doi:10.1001/jamasurg.2015.1250. ©Institute for Clinical and Economic Review, 2015 Page 117 [Return to Table of Contents]