Quantity of Research Available
A total of 438 publications were indentified in the initial literature search. After review of the titles, 47 articles were selected for full text review. A total of 21 relevant articles were identified in the grey literature search; as a result, 68 articles were assessed for inclusion in this report.
A total of 21 studies met the inclusion criteria for this review. There was one HTA,11 ten MAs,12–21 three SRs,22–24 and five RCTs25–29 included that addressed question 1. In terms of question 2, there were three studies included (the one HTA also included in question 1, and two additional economic analyses).11, 30, 31 There were no guidelines identified that recommended a particular bariatric surgery technique over another. As a result, this review was unable to address question 3.
Appendix 1 provides the PRISMA flowchart for study selection in this report.
Summary of Study Characteristics
Details on clinical and safety study characteristics, economic study characteristics, critical appraisal, and study findings are located in Appendices 2 through 5, respectively.
Study Design
Among the studies included, there was one HTA,11 ten MAs,12–16, 18–21, 32 three SRs,22–24 five RCTs,25–29 and two economic analyses.30, 31 The number of included studies in the HTAs, MAs, and SRs ranged from 5 to 164, and the publication dates of the included studies was 1986 to 2013.11–24 It must be noted that many of the studies reviewed in the HTA, MA, and SR overlapped, particularly some RCTs.11–24 Of note, while the HTA included a SR assessing effectiveness of bariatric procedures, a SR evaluating the economic literature, and an economic analysis, the economic analysis was not applicable to this Rapid Report because it did not evaluate the costs associated with the different types of bariatric procedures, and instead combined costs.11
Country of Origin
The countries of origin included Canada,11, 22 China,14–16, 18, 32 Finland,25 France,26 Israel,28 Italy,23, 27 New Zealand,19 South Korea,12, 13 Switzerland,29 United Kingdom,20, 24 and the United States,12, 13, 21 The economic analyses were from Portugal and the United States.30, 31
Patient Population
The patient populations were adults with obesity, based on BMI, for all studies.11–31 A number of studies used the definition of a BMI > 40 kg/m2 or a BMI of > 35 kg/m2 plus the presence of comorbidities (for example, cardiovascular disease, type 2 diabetes, or obstructive sleep apnea).15, 24, 25, 27–29 Two studies focused on people with obesity and type 2 diabetes,21, 28 and one study evaluated individuals 55 years of age and older.20 Most studies reported a greater proportion of women than men, and baseline BMI ranged from 30.3 to 79.9 kg/m2.11–31
Interventions and Comparators
Three types of bariatric surgeries were evaluated in this analysis: RYGB, LAGB, and SG. In terms of the number of studies evaluating each type of bariatric procedure, 17 evaluated RYGB,11, 13–16, 18–21, 23–29, 33 10 evaluated LAGB,11, 12, 20, 21, 23, 24, 27, 28, 32, 33 and 16 evaluated SG.11–16, 18, 19, 23–26, 28, 29, 32, 33 A total of five studies compared all three procedures,11, 12, 22–24 ten studies compared RYGB and SG,13–16, 18, 19, 25, 26, 28, 29 five studies compared RYGB and LAGB,20, 21, 27, 30, 31 and one study compared SG to LAGB.32
Clinical Outcomes
In terms of clinical outcomes, all studies evaluated weight loss in some form, either by percent excess weight loss, reduction in BMI, reduction in weight, or percent excess BMI reduction, except one study that only looked at risk for anemia, iron deficiency, and vitamin B12 deficiency.13 Percent excess weight loss is calculated by determining excess body weight (subtracting ideal body weight from total body weight at the time of surgery) and total weight loss after surgery, and dividing total weight loss by excess body weight. Other clinical outcomes evaluated included length of operation, length of hospital stay, complications associated with the procedure, need for reoperation, improvement and/or resolution in obesity-related complications, quality of life, and changes in nutrient levels.
Economic Outcomes
The economic analyses each used both cost-effectiveness analysis and cost-utility analysis to for comparing RYGB to LAGB.30, 31 The perspective of one of the analyses was the health care system, and the other was the societal perspective.30, 31 The economic analyses did not specify the assumptions for their analyses.30, 31 Both manuscripts conducted sensitivity analyses to test the robustness of their findings. Wang and colleagues used three different BMI trajectories, including having the BMI approach the same trajectory as a non-surgical patient five years after the surgery was completed, having BMI remain stable five years after the surgery was completed, and regaining 100% of the weight lost in the first 5 years of surgery, up to 15 years post-surgery.30 They also varied BMI at baseline, age at baseline, sex, early complication rate, discount rate, and early mortality rate to conduct further sensitivity analyses.30 Faria and colleagues conducted sensitivity analyses by varying BMI, age, and presence or absence of comorbidities at the time of surgery.31 Lastly, one HTA included a systematic review on available studies assessing cost-effectiveness of RYGB, SG, and LAGB.11
Summary of Critical Appraisal
There was one HTA that included an effectiveness SR and an economic SR, 11 MAs, and 3 SRs that evaluated clinical and safety endpoints. In general, these manuscripts were found to be of moderate quality. A few studies had inadequate search strategies (for example, only searching two databases for relevant literature) to identify all relevant literature.13, 15, 19–21, 32 In addition, only two studies provided a list of excluded studies and the reason for exclusion.11, 24 Many studies reported using duplicate study selection.12–14, 16, 21, 22, 24 Fewer studies reported duplicate data extraction.12, 18, 21, 22, 24 Most studies reported assessing quality of the included studies, and all included MAs reported the statistical methods for combining study results to produce an overall effect size. However, three of the studies did not report the amount of heterogeneity associated with combining studies in their MAs.12, 19, 20 Lastly, possibility of publication bias was assessed in seven of the 15 manuscripts, and authors concluded that it was unlikely that publication bias was present in their reviews.14, 15, 18–21, 32
There were five RCTs that compared one form of bariatric surgery to another included in this report.25–29 In general, the studies were of moderate quality. None of the studies were blinded, which would be expected for individuals undergoing the procedure and surgeons conducting the surgery, however, none of the studies had blinded outcome assessors or analysts.25–29 Losses to follow up were reported in all of the studies, and some losses were substantial, particularly in the study conducted by Peterli and colleagues at three years follow-up.25–29 The losses to follow up did not appear to differ between surgical groups, however.25–29 Four of the studies did not use intention-to-treat analysis, thereby excluding losses to follow up from the study analyses.25–27, 29 The process of randomization was not documented in three of the studies.25–27
Two economic analyses compared cost-effectiveness and cost-utility of RYGB and LAGB.30, 31 The quality of each of these analyses was poor. While both studies had a clear objective and stated that costs were discounted at 3% per year, neither study reported assumptions of the analyses, and the only evaluated direct costs associated with each procedure.30, 31 In addition, the studies did not clearly state what items or costs associated with these items were included in the direct costs, therefore it was not possible to assess generalizability of costs included in each model.30, 31 Faria and colleagues stated that they used a societal perspective “of universal coverage for health care” to conduct their economic analyses, however, they do not specify what costs were included, and did not include indirect costs associated with bariatric procedures.31
Summary of Findings
What is the comparative clinical effectiveness and safety of specific bariatric surgical interventions (roux-en-y gastric bypass, sleeve gastrectomy, and laparoscopic adjustable gastric band) for obese and morbidly obese patients?
RYGB versus LAGB
Among the studies that compared RYGB with LAGB, it was consistently noted that RYGB was associated with a statistically significantly greater reduction in weight (measured by percent excess weight loss, percent excess BMI loss, reduction in BMI, or reduction in weight in kilograms) and improvement or resolution of obesity-related comorbidities including type 2 diabetes, dyslipidemia, obstructive sleep apnea, and hypertension.11, 12, 20–24, 27 In addition, reoperation was necessary in more people who received LAGB compared to those who received RYGB.12, 27 However, duration of surgery, risk for complications related to the procedure, and length of hospital stay were consistently higher in people who underwent RYGB compared to those who underwent LAGB.11, 12, 23, 24, 27
RYGB versus SG
A total of fourteen studies compared RYGB with SG.11, 13–16, 18, 19, 22–26, 28, 29 Among these studies, most found that SG was less effective than RYGB for weight outcomes and improvement in obesity-related comorbidities, but were also less likely to have procedural complications relative to RYGB.12, 14–16, 18, 22–24, 29 The only available studies assessing risk of nutrient deficiency were done in studies comparing RYGB to SG. In a study conducted by Kwon and colleagues, RYGB was associated with vitamin B12 deficiency compared to SG (odds ratio [OR]: 3.55; 95% confidence interval [CI]: 1.26 to 10.01).13 Also, Vix and colleagues evaluated vitamin D and parathyroid hormones in people randomized to RYGB or SG, and found that RYGB was associated with significantly lower vitamin D and parathyroid hormone levels up to 12 months after surgery compared with SG.26 These studies suggest a greater risk for nutrient deficiencies in patients who undergo RYGB compared to those who undergo SG. The need for reoperation was lower for SG compared with RYGB in four studies,.12, 14, 18, 23 but was not significantly different in two studies.16, 29
LAGB versus SG
There were six studies that compared LAGB to SG.11, 12, 22–24, 32 The HTA from the Institute of Health Economics noted that clinical evidence was limited for SG, and did not make any direct comparisons with LAGB.11 SG was associated with a greater reduction in BMI at 1 and 5 years relative to LAGB in a MA of RCTs and observational studies.12 SG was also associated with a greater likelihood for remission of diabetes, hypertension, dyslipidemia, and sleep apnea in this MA.12 Complications rates were similar between SG and LAGB, whereas reoperation was less likely in SG relative to LAGB.12 Similar results were seen in terms of weight reduction and resolution of type 2 diabetes in a MA conducted by Wang and colleagues, where SG was associated with a greater reduction in weight at 6 and 12 months, but these comparisons were limited by high heterogeneity.32 In a SR conducted by Sarkosh and colleagues, it was found that LAGB was associated with a greater mean percent weight loss (66.8%) compared with SG (46.1%).22
Lastly, it must be noted that some studies concluded that, based on the differing benefit and risk profiles associated with each bariatric procedure, that assessment of the patient, preferences of the patient, and experience of the surgeon will influence the choice of type of procedure.11, 14, 24
What is the comparative cost-effectiveness of specific bariatric surgical interventions (roux-en-y gastric bypass, sleeve gastrectomy, and laparoscopic adjustable gastric band ) for obese and morbidly obese patients?
Wang and colleagues conducted cost-effectiveness and cost-utility analyses to compare lifetime direct medical costs, quality-adjusted life years (QALYs), and the incremental cost effectiveness ratio (ICER) of RYGB compared to LABG.30 They used a healthcare system perspective, and a reference case of a 53-year old female with a BMI of 44 kg/m2 to conduct their analyses. They found the direct medical costs for RYGB ($169,074) to be comparable to LAGB ($164,313) with a standard BMI trajectory (BMI decreasing after surgery), and the QALYs for RYGB (13.4) to be higher than for LAGB (12.8), producing an ICER of $7,935 per QALY gained for RYGB compared to LAGB.30 Sensitivity analyses varying the BMI trajectory (weight stable and maximum weight regain) produced similar results.30
Faria and colleagues also conducted cost-effectiveness and cost-utility analyses to compare lifetime costs and QALYs of RYGB compared to LAGB.31 In contrast to the study by Wang et al, the lifetime costs of LAGB (41,056 euro) was consistently more than the lifetime costs of RYGB (29,254 euro), although it was unclear as to what costs were driving the difference.30, 31 However, similarly to the previous economic analyses, RYGB was consistently associated with greater QALYs (16.36) relative to LAGB (15.09).31 These analyses were robust over a number of sensitivity analyses, including those with type 2 diabetes, and varying levels of BMI.31 The study authors concluded that RYGB was dominant as it was consistently associated with greater QALYs and lower costs relative to LABG across sensitivity analyses.31
Lastly, in the one HTA included in this review there were no direct economic comparisons identified in the cost-effectiveness literature review.11
Limitations
There are a number of limitations that must be noted when considering the information reported in this review. A number of pooled comparisons within the MAs were associated with significant heterogeneity, and therefore results must be interpreted with caution. In addition, some of the MAs/SRs did not assess the possibility of publication bias, and as a result, studies may be missing from these reviews, which could impact the conclusions drawn in each review. Some studies within this review had conflicting conclusions. In addition, many of the studies included in the HTA, MAs, and SRs overlapped, which may overemphasize the conclusions drawn from the overlapping studies. Lastly, the included economic analyses were of poor quality and lacked important information regarding costs, limiting the generalizability and applicability of the results. Also, there was no economic information available for SG.
