Metabolic and bariatric surgery in adults with obesity and diabetes mellitus: a network meta-analysis

RATIONALE: Obesity and its associated comorbidities, particularly type 2 diabetes (T2D), have reached epidemic dimensions globally. Both conditions are related to an increased risk of serious health outcomes and impose a substantial burden on those affected. Metabolic and bariatric surgery (MBS) may offer a therapeutic option for achieving sustained weight loss, improving comorbidities and thereby enhancing long-term quality of life. OBJECTIVES: To investigate the medium- and long-term benefits and harms of different MBS procedures compared with each other and to non-surgical treatment on outcomes relevant to adults with obesity and T2D, and to obtain a clinically meaningful ranking of these interventions by considering both randomised controlled trials (RCTs) and cohort studies. SEARCH METHODS: We searched MEDLINE, Cochrane CENTRAL, CINAHL, LILACS, BASE as well as the WHO ICTRP Search Portal and ClinicalTrials.gov. The date of the last search for all databases was 25 June 2024. We did not apply any language restrictions. ELIGIBILITY CRITERIA: We included RCTs and both prospective and retrospective cohort studies with a minimum follow-up of three years in adults (≥ 18 years) with obesity (body mass index ≥ 30 kg/m² or ethnic-specific cut-off values) and T2D. Eligible interventions include MBS procedures (i.e. laparoscopic adjustable gastric banding [LAGB], sleeve gastrectomy [SG], Roux-en-Y gastric bypass [RYGB], one-anastomosis gastric bypass [OAGB], gastric plication [GCP], biliopancreatic diversion with duodenal switch, and their variations) or non-surgical treatment (i.e. lifestyle intervention, medical treatment [LI/MT]). OUTCOMES: Critical outcomes were medium- (≥ 3 to < 5 years) and long-term (≥ 5 years) weight loss (%), waist circumference (cm), T2D remission (complete, partial), and serious adverse events (SAE). Important outcomes included all-cause mortality, T2D complications, glycaemic control (glycated haemoglobin (HbA1c) in %) and health-related quality of life (HRQoL). RISK OF BIAS: We used the Risk of Bias 2 (RoB 2) tool for RCTs and Risk Of Bias In Non-randomized Studies - of Interventions (ROBINS-I) for cohort studies. SYNTHESIS METHODS: We performed random-effects network meta-analyses based on the frequentist approach and ranked treatments by P scores. We rated the certainty of evidence according to the GRADE approach for long-term outcomes (≥ 5 years). INCLUDED STUDIES: We included 18 studies (with 70 reports) with 15,282 participants in this review. Of these, 13 studies were RCTs (n = 911) and five were cohort studies (n = 14,371). Two RCTs compared MBS with LI/MT, ten compared two or three different MBS procedures and one with a 3-arm design investigated two different MBS procedures as well as LI/MT. All cohort studies compared different MBS procedures. The networks were based on the following MBS procedures: RYGB, SG, OAGB, LAGB, GCP, and LI/MT. SYNTHESIS OF RESULTS: NMA results (8 RCTs; n = 491) suggest greater long-term weight loss after RYGB compared to LI/MT (MD -16.95%; 95% CI -24.19 to -9.71). OAGB (MD -12.10%; 95% CI -22.85 to -1.34) and SG (MD -9.40 %; 95% CI -17.45 to -1.34) may also result in greater weight loss compared to LI/MT, but the evidence is very uncertain. A reduction in waist circumference at long-term follow-up (6 RCTs; n = 348) is likely higher after RYGB compared to LI/MT (MD -12.34 cm; 95% CI -17.88 to -6.81). OAGB (MD: -13.28 cm; 95% CI -22.04 to -4.53) and SG (MD-9.08 cm; 95% CI -15.74 to -2.43) may also lead to a higher reduction in waist circumference compared to LI/MT, but the evidence is very uncertain. The evidence is very uncertain about the effects of MBS procedures compared to LI/MT on long-term complete T2D remission (9 RCTs; n = 536). All MBS procedures (8 RCTs; n = 588) may result more often in partial T2D remissions than LI/MT, but the evidence is very uncertain. The evidence is very uncertain about the effect of all MBS procedures compared to LI/MT on SAE (9 RCTs; n = 598) and all-cause mortality in the long term (7 RCTs; n = 520). RYGB may reduce total long-term T2D complications compared to LI/MT (Rate ratio (RaR) 0.17; 95% CI 0.03 to 0.89; 5 RCTs; n = 356), but the evidence is very uncertain. Based on NMA results (10 RCTs; n = 612), RYGB may reduce HbA1c compared to LI/MT (MD -1.44%; 95% CI -2.26 to -0.63). OAGB (MD -2.01%; 95%CI -3.42 to -0.60) and SG (MD -1.01%; 95%CI -1.94 to -0.08) may also reduce HbA1c compared to LI/MT, but the evidence is every uncertain. Both RYGB and SG may increase mental and physical components of HRQoL compared to LI/MT (both 2 RCTs; n = 118), but evidence is very uncertain. The inclusion of cohort studies in the analyses of percentage weight loss, partial type 2 diabetes remission, SAE, and all-cause mortality confirmed the main findings. For most outcomes, the main reasons for downgrading certainty of evidence were related to risk of bias and imprecision. AUTHORS' CONCLUSIONS: In people with obesity and T2D, RYGB, OAGB and, to a lesser extent, SG may lead to long-term improvements in body weight, waist circumference, partial T2D remission, and HbA1c compared with LI/MT. However, due to imprecise results, the magnitude of the effects often remains uncertain or very uncertain. For other outcomes, especially complete T2D remission, SAE, all-cause mortality and T2D complications, the evidence is mainly very uncertain, due to a low numbers of participants and events. When interpreting these findings, it should be noted that cohort studies were not, as originally planned, included in the certainty of evidence assessment, as these studies were judged to have a high or critical risk of bias. Larger and well-designed studies may contribute to diminishing the existing uncertainties and enhancing our confidence in the findings. FUNDING: This Cochrane review was funded by the German Federal Ministry of Education and Research (BMBF) (grant number: 01KG2201). REGISTRATION: Protocol (2024) DOI: https://doi.org/10.1002/14651858.CD015622 PROSPERO ID: CRD42023457363.