OMMs appear to be valid alternatives to EBP and MBS for mild-to-moderate obesity, but surgery more effective for patients with a higher degree of obesity

For patients living with mild-to-moderate obesity, newer obesity management medications (OMM) such as tirzepatide and semaglutide, appear to be valid alternatives to endoscopic bariatric procedures (EBP) and metabolic bariatric surgery (MBS), according to researchers from the Italian Obesity Society (Società Italiana dell'Obesità (SIO)).

Their study reveals the results from a systematic review (SR) and a network meta-analysis (NMA) on randomised clinical trials (RCTs) comparing OMM, EBP and MBS versus either lifestyle interventions (LSI), placebo or no treatment, or other active comparators, in individuals with overweight or obesity.

However, the researchers noted that patients with a higher degree of obesity could be more effectively treated with MBS, the efficacy of which, with the notable exception of laparoscopic adjustable gastric banding (LAGB) and greater curvature plication (GCP), appears superior to other treatments, especially in the long term. Some types of MBS, such as BPD and SADI, although very effective, should be used with caution because of safety issues, whereas RYGB and LSG combine good efficacy with greater safety.

The principal endpoint was TBWL% (as change-from-baseline parameter); secondary endpoints were waist circumference, body composition, proportion of patients achieving at least 5%, 10%, 15%, 20%, and 25% body weight reduction and remission or improvement/resolution of OAMC, SAE, mortality, MACE, FPG, HbA1c, lipid profile, eGFR, creatinine, albuminuria, mental health parameters and QoL.

The SR identified 129 trials that fulfilled all the inclusion criteria: 52, 13, and 64 trials on MBS, EP, and OMMs were compared with either LSI/Pbo/NT or other active anti-obesity strategies. Therefore, the number of available comparisons was 140. The overall number of patients enrolled was 60,044, 2217, and 5991 in trials with OMM, EP, and MBS, respectively.

Only three trials reported separately the results on weight loss in different categories of BMI at study entry. One study with liraglutide27 provided data for patients with overweight (BMI 27–29.9 kg/m2) and with different degrees of obesity (class I, II, and III). All categories of patients reported a significantly higher placebo-subtracted TBWL%, ranging from 3.7% to 5.2%. The other two studies reported a significantly greater TBWL% at endpoint with semaglutide than with placebo in all BMI classes. The placebo-subtracted effect of semaglutide was 12.40 [7.13, 17.67], 15.60 [12.65, 18.55], 17.00 [13.64, 20.36], and 13.90 [10.78, 17.02]% for overweight, class I, II, and III of obesity, respectively (all p<0.001; test for subgroup differences: p=0.40). Similar figures were obtained for the other study, with a TBWL% ranging from 9.6% to 11.3%.

For patients with overweight (BMI 27–29.9 kg/m2), only one study with semaglutide reported data on incident MACE, showing that the interventional drug was associated with a significantly lower risk. For patients with BMI at study entry between 30 and 35 kg/m2, subgroup analyses were available for three trials with semaglutide, three with liraglutide and one with tirzepatide. A statistically significant reduction of incident MACE was observed only for semaglutide. For patients with BMI 35–39.9 kg/m2 and >39.9 kg/m2, only one study with semaglutide reported data on incident MACE, showing no between-group differences.

Only one study comparing RYGB with LSI and performed in an Asian population with type 2 diabetes reported a mean BMI at entry <30 kg/m2. The TBWL% at the end of the trial was significantly superior in the intervention arm at any assessed time points (i.e., WMD: 15.50 [12.53, 18.47], 12.50 [9.53, 15.47], 12.50 [9.53, 15.47], and 11.20 [8.23, 14.17] %, all p<0.001, at 52, 104, 156, and >156 weeks, respectively). The between-group difference of BMI at the endpoint was −5.20 [−7.12, −3.28] kg/m2 (p<0.001).

They reported on 22 trials with a mean baseline BMI between 30 and 34.9 kg/m2. Tirzepatide resulted in equal effectiveness to OAGB and RYGB, and it was significantly superior to all the other comparisons. Semaglutide was superior to liraglutide, orlistat, and IGB, and not inferior to the other comparisons, except tirzepatide. Results on weight loss at different time points were similar to those at the endpoint notably, results after 2 or more years were available only for semaglutide and RYGB.

A reduction of BMI at endpoint greater than 5 kg/m2 and a reduction of waist circumference greater than 10 cm were observed only for tirzepatide and RYGB. Fifty-seven trials with mean BMI at enrolment between 35 and 39.9 kg/m2 were available for analysis. Liraglutide was not superior to orlistat and NB, and equally effective as EBP (except for ESG). Semaglutide was associated with a higher TBWL% than the other OMMs (with the notable exception of tirzepatide) and was equally effective to EBP, GCP, and LAGB, but gastric bypass. Tirzepatide was significantly superior to all the other comparisons, except for GCP and ESG (not inferior), and it was associated with lower TBWL% than both OAGB and RYGB.

Data on longer-term (>2 years) weight loss were available only for RYGB and LABG. Heterogeneity (τ2 values) was assessed for all the available comparisons, showing some concerns NB, liraglutide, orlistat, and POSE versus the reference category.

The procedure with the highest estimated weight loss was BPD (for which no trial on patients with mean BMI <40 kg/m2 was available). All the other surgical procedures produced a weight loss greater than 15%, with the only exception of LABG. Semaglutide was statistically less effective than SG and gastric bypass, but not inferior, from a statistical point of view, to LVBG, GCP and LAGB. Among different types of MBS, BPD was associated with a higher TBWL% than all the other interventions. RYGB and OAGB (equally effective with each other) were superior to SG. LAGB and GCP were associated on average with a lower TBWL% (<20%).

Results on %TBWL at different time points were similar to those at endpoint. However, the efficacy of LAGB appeared to decrease over time, whereas this phenomenon was not observed with other surgical procedures. Results on BMI and (when available) on waist circumference were consistent with those on TBWL. No heterogeneity (τ2 values) was detected for any of the available comparisons.

No trial enrolling patients with a mean BMI below 30 kg/m2 was available for this endpoint. In the 30–34.9 kg/m2 BMI category, the SELECT trial reported a significant reduction of events with semaglutide, compared to placebo. Twenty-one trials enrolling patients with a mean BMI between 35 and 39.9 kg/m2 performed with liraglutide, semaglutide, tirzepatide, NB, and orlistat, which reported information on adjudicated MACE, failed to show any significant effect of any treatment on this endpoint. Only two RCTs with a mean BMI at enrolment >39.9 kg/m2, provided information on this endpoint, with no events reported.

Only one trial enrolling patients with mean BMI 30–34.9 kg/m2 reported information on this endpoint, showing a non-significant reduction of HHF for semaglutide. In 4 trials with mean BMI at enrolment between 35 and 39.9 kg/m2, both semaglutide (N = 3 studies) and tirzepatide (N = 1 study) were associated with a significant reduction of HHF. No information on HHF was available for trials with mean BMI at enrolment below 30 or over 40 kg/m2.

Two trials with semaglutide with mean BMI at enrolment between 30 and 34.9 kg/m2 reported a significant reduction of the incidence of diabetes with the active treatment. In trials enrolling patients with a mean BMI between 35 and 39.9 kg/m2 (N = 419, 68, 83, 91), a lower risk of incident diabetes was observed with liraglutide and orlistat, but not semaglutide. Only 2 studies with mean BMI at enrolment >39.9 kg/m2, one with semaglutide and one comparing RYGB and OAGB reported data on incident diabetes, with no significant between-group differences.

There was no significant increase in the risk of serious adverse events (SAE) observed for any therapy in trials enrolling patients with a mean BMI between 30 and 34.9 kg/m2 and comparing an active treatment with LSI/Pbo/NT. In a NMA of trials with a mean BMI at enrolment between 35 and 39.9 kg/m2, MBS - with the exception of SG and GCP - was associated with the highest risk; ESG was also associated with an increased risk of SAE, unlike other types of EBP, whereas among OMMs only NB was associated with an increased risk of SAE. In trials enrolling patients with a mean BMI >40 kg/m2, BPD and SADI were the only treatments associated with an increase in the risk of overall SAE versus LSI/Pbo/NT.

In a NMA of trials with mean BMI at enrolment 30–34.9 kg/m2, only semaglutide was associated with a significant reduction of all-cause mortality versus LSI/Pbo/NT. No significant effect on all-cause mortality was detectable in trials with mean BMI at enrolment 35–39.9 kg/m2, or >40 kg/m2.

They reported that there was no data on the quality of life is available for trials enrolling patients with a mean BMI <30 kg/m2. In higher BMI categories, only a minority of trials reported quality of life results, using a variety of instruments, and therefore preventing a formal meta-analysis. The most effective treatments (OMM or MBS) on weight loss were usually associated with improvements of QoL versus LSI/Pbo/NT in all BMI categories, whereas most direct comparisons between active treatments failed to detect significant differences.

“These results are of interest to clinicians involved in the management of obesity. For the first time, performing a thorough evaluation and synthesis of RCTs and adopting GRADE methodology, different anti-obesity approaches have been meta-analysed in different categories of patients (overweight, and obesity class I, II, and III), providing a clearer picture of their effectiveness,” the authors concluded. “A systematic disclosure of results in different classes of BMI would enhance our knowledge of the profile of action of different treatments, allowing for a more rational choice of therapy in individual patients.”

The findings were reported in the paper, 'Efficacy and safety of European Medicines Agency (EMA)-approved pharmacological, endoscopic, and surgical treatments in different classes of obesity: A network meta-analysis of randomised controlled trials for the development of the SIO (Società Italiana Obesità) Italian guidelines for the diagnosis and treatment of overweight and obesity', published in Diabetes, Obesity and Metabolism. To access this paper, please click here