Er zijn in Nederland vermoedelijk heel wat mensen die in aanmerking kunnen komen voor gewichtsreducerende medicatie, omdat leefstijlaanpassingen bij hen onvoldoende effect hebben. Velen zullen voor aanvullende behandelingen in aanmerking komen, zoals farmacotherapie of metabole chirurgie. Desondanks wordt deze medicatie maar zeer beperkt voorgeschreven. Er is onduidelijkheid over welke patiënten het meest gebaat zijn bij medicatie, wat het beste moment is om te starten, kosteneffectiviteit en lange termijneffecten.

Introduction

There are probably quite a few people in the Netherlands who would qualify for weight-reducing medication, because lifestyle changes have insufficient effect on them. Many patients are eligible for additional treatments, such as pharmacotherapy or bariatric surgery. Nevertheless, this medication is only prescribed to a very limited extent. There is uncertainty about which patients would benefit from weight-reducing medication, the best time to start medication use, cost-effectiveness, and long-term effects.

Description of studies

Orlistat, liraglutide, naltrexon-bupropion vs placebo

Khera (2016) compares weight loss among drug treatments for obesity and covers literature until March 2016. Khera (2016) includes 28 RCTs comparing orlistat (16 trials), lorcaserin (3 trials),  naltrexon-bupropion (4 trials), phentermine-topiramate (2 trials), liraglutide (2 trial) against placebo. One three-group trial was included comparing liraglutide and orlistat against placebo. The study size of the included trials varied from 220 to 3.731 participants. The outcomes were the proportion of participants achieving at least 5% weight loss from baseline, and the proportion of participants with at least 10% weight loss and change in weight from baseline. The median of the average age of study participants was 45,9 years (range of average age, 40.0–59.8 years) and 74% of participants were women (range, 45%–92%). The median of average body mass index (BMI) of patients was 36.1 kg/m² (range, 32.6–42.0) and the median of average baseline weight was 100.5kg (range, 95.3–115.8kg). The risk of bias of individual studies was assessed using the Cochrane Risk of Bias assessment tool. Overall, studies were considered to be at high risk of bias, with attrition rates of 30% to 45% in all trials (see Figure 2; Khera, 2016). This literature analysis only focuses on the effect of orlistat,  naltrexon- bupropion, and liraglutide as drug treatment for obesity, since use of phentermine-topiramate and lorcaserin is not approved in the Netherlands. Specific baseline characteristics for these trials are included in the evidence table.

Singh (2020) compares weight loss among drug treatments for obesity using a and covers literature until September 2019. Singh (2020) includes 31 RCTs comparing orlistat (17 trials), lorcaserin (4 trials),  naltrexon-bupropion (4 trials), phentermine-topiramate (2 trials), liraglutide (4 trial) against placebo. The study size of the included trials varied from 45 to 12,000 participants. The outcome was a reduction in body weight (in kilograms) compared to a placebo at the end of 1-year follow-up. The risk of bias in individual studies was assessed in the context of the outcome using the Cochrane Risk of Bias assessment tool on seven domains that include – random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. Overall, studies were considered to be at high risk of bias, due to incomplete data as a result of attrition bias (see Table S1; Singh, 2020). Specific baseline characteristics for orlistat,  naltrexon-bupropion, and liraglutide are included in the evidence table.

Khera (2018) compares the effect of weight loss medication on cardiovascular risk profiles in adults with obesity using a systematic review and network meta-analysis and covers literature until February 2017. Khera (2018) includes 28 RCTs comparing orlistat (16 trials), lorcaserin (3 trials),  naltrexon-bupropion (4 trials), phentermine-topiramate (2 trials), liraglutide (2 trials) against placebo. One three-group trial was included comparing liraglutide and orlistat against placebo. The study size of the included trials varies from 220 to 3.731 participants. Outcomes were assessed at one year of follow-up. The outcomes were defined a priori and included: glucose profile – fasting blood glucose (mg/dL) and/or HbA1c; markers of lipoprotein metabolism – low-density lipoprotein (LDL)-cholesterol (mg/dL), high-density lipoprotein(HDL)-cholesterol (mg/dL); systolic and diastolic blood pressure (mmHg); and central adiposity, assessed using waist circumference (cm). The median age of study participants across studies was 46 years (range 41–60) and 77% (range 45–90) were women. The median baseline BMI was 36.1 kg/m² (range 32.7–42.0). The baseline median fasting blood glucose was 105,6 mg/dL (interquartile range [IQR, 94.6–161.5) and hemoglobin A1c was 7,9% (IQR, 5.7–8.1). The median baseline LDL-cholesterol was 122.7mg/dL (IQR, 112.9–137.2) and HDL-cholesterol was 46.4mg/dL (IQR, 45.5–51.3); median 34.5% (IQR, 10.3–52.5) participants had dyslipidemia. Median average systolic blood pressure was 127.5mmHg (IQR, 121.9–135.8) and diastolic blood pressure was 79.1mmHg (IQR, 77.5–84.2) at baseline; median 23% (IQR, 17–36) participants had hypertension at baseline. Waist circumference was 110cm (IQR, 109–115) at baseline. The risk of bias of individual studies was assessed using the Cochrane Risk of Bias assessment tool. Overall, studies were considered to be at high risk of bias, with attrition rates of 30% to 45% in all trials (See table 8; Khera, 2018). Specific baseline characteristics for orlistat,  naltrexon-bupropion and liraglutide are included in the evidence table.

Chao (2019) conducted a single-site, open-label, parallel-group design RCT on the effect of liraglutide combined with intensive behavioural therapy (IBT) on weight loss and quality of life compared to a multicomponent group and a IBT alone group (IBT combined with placebo). The RCT by Chao (2019) is of moderate quality (see table of quality assessment) and was commercially sponsored by Novo Nordisk. Chao (2019) included adults aged 21-70 years; body mass index (BMI) of 30-55 kg/m²; and prior lifetime weight-loss effort with diet and exercise (n=150). All participants received 21 sessions of IBT and were instructed to engage in low-to-moderate intensity physical activity 5 days per week, gradually. Participants who weighed <113.6 kg were prescribed a diet of 1200- 1499 kcal/day, while those ≥113.6 kg received 1500-1800 kcal/day. Outcomes include weight loss, health-related quality of life, and weight-related quality of life. The 150 participants included 67 (44.7%) individuals who identified as black and 119 (79.3%) women. The total sample had a mean age of 47.6 (SD=11.8) years, a weight of 108.4 (SD=17.5) kg, and a BMI of 38.4 kg/m² (SD=4.9). Baseline scores did not differ significantly among treatment groups. At baseline, 28% of the sample had class I obesity, 36.0% had class II and 36% had class III.

Wadden (2020) conducted a 56-week, randomized, double-blind, placebo-controlled, two-armed, multicenter phase 3b trial. The primary objective of the trial was to compare the effect of liraglutide 3.0 mg versus placebo, as an adjunct to standard-based IBT, on weight loss in individuals with obesity. The RCT by Wadden (2020) is of moderate quality (see table of quality assessment) and was commercially sponsored by Novo Nordisk. Wadden (2020) included adults aged ≥ 18 years, with stable body weight and BMI ≥30 kg/m² (n=282). Participants were randomized to treatment with liraglutide combined with IBT (liraglutide-IBT; n=142) or a placebo combined with IBT (placebo-IBT; n=140). The primary outcomes were change in body weight (percent) from baseline to week 56 and the proportion of participants who lost ≥ 5% of baseline body weight. Secondary outcomes included the proportion of participants who lost > 10% or > 15% of baseline body weight at week 56. Of the 142 participants included in the liraglutide-IBT arm 78.9% of individuals were identified as white and 16.2% were male. The liraglutide-IBT arm had a mean age of 45.4 (SD=11.6) years, the weight of 108.5 kg (SD=22.1), and a BMI of 39.3 kg/m² (SD=6.8). Of the 140 participants included in the placebo-IBT arm 82.1% of individuals were identified as white and 17.1% were male. The placebo-IBT arm had a mean age of 49.0 (SD=11.2) years, the weight of 106.7 kg (SD=22.0), and a BMI of 38.7 kg/m² (SD=7.2). Baseline scores did not differ significantly among treatment groups.

Semaglutide vs placebo

Zhong (2022) conducted a systematic review and meta-analysis assessing the efficacy and safety of once-weekly semaglutide among adults with overweight or obesity. Zhong (2022) covers the literature until July 2021. Literature searches were conducted in MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials. Zhong (2022) included randomized controlled trials which enrolled adults (≥18 years old) with a body-mass index of 27 kg/m² or higher and without diabetes and compared once-weekly semaglutide and placebo. In total, 4 studies were included. All included studies were randomized double-blind trials that compared the treatment effect of once-weekly subcutaneous semaglutide 2.4 mg with placebo in patients with overweight or obesity. Article quality of individual studies was assessed using the Cochrane Collaboration’s Risk of Bias tool. The primary outcomes were the percentage change and absolute change in body weight from baseline. Secondary outcomes included achievement of categorical weight loss targets (at least 5, 10, 15, or 20%) and change from baseline in waist circumference, body-mass index, glycated hemoglobin (HbA1c), fasting plasma glucose (FPG), systolic and diastolic blood pressure, physical functioning score on the 36-item Short-form Health Survey (SF-36), the SF-36 physical and mental component summary scores, and lipid profiles (total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), and triglycerides).

Davies (2021) conducted an double-blind, double-dummy, phase 3, superiority study evaluating the efficacy and safety of once a week subcutaneous semaglutide 2.4 mg versus semaglutide 1.0 mg (the dose approved for diabetes treatment) and placebo for bodyweight management in adults with overweight or obesity. Included were adults (≥18 years old) with a body-mass index of 27 kg/m2 or higher that reported at least one unsuccessful dietary effort to lose weight and who were diagnosed with diabetes at least 180 days before screening. Patients were randomly assigned (1:1:1) to semaglutide 2.4 mg, semaglutide 1.0 mg, or placebo. In total, 404 patients were allocated to the semaglutide 2.4 mg group (mean age 55 ± 11 years; mean BMI 35.9 ± 6.4 kg/m²; 55.2% female), 403 were allocated to the semaglutide 1.0 mg group (mean age 56 ± 10 years; mean BMI 35.3 ± 5.9 kg/m²; 50.4% female) and 403 were allocated to the placebo group (mean age 55 ± 11.0 years; mean BMI 35.3 ± 6.4 kg/m²; 47.1% female). Patients received semaglutide 2.4 mg or semaglutide 1.0 mg or placebo once a week for 68 weeks, plus a lifestyle intervention, followed by 7 weeks without treatment. Semaglutide was started at 0,25 mg per week and escalated in a fixed-dose regimen every 4 weeks until the target dose was reached. The lifestyle intervention involved counselling on diet (500 kcal per day reduction relative to the estimated total daily energy expenditure calculated at time of random allocation) and physical activity (150 min per week). Outcomes (semaglutide 2.4 mg vs placebo), were percentage change in bodyweight from baseline to week 68 and loss of at least 5% of baseline weight at week 68 (semaglutide 2,4 mg vs placebo), proportions of patients achieving bodyweight reductions of at least 10% or 15% at week 68, change from baseline to week 68 in waist circumference, change from baseline to week 68 in HbA1c, systolic and diastolic blood pressure, SF-36v2 physical functioning score, and IWQOL-Lite-CT physical function score.

Results

Weight loss

Orlistat, liraglutide,  naltrexon-bupropion vs placebo

In the direct pairwise meta-analysis by Khera (2016), weight loss compared with placebo at one year (52+4weeks) resulted in a mean difference (MD) of MD=2.6 kg (95% CI 2.3 to 2.9) with orlistat, MD=5.0 kg (95% CI 4.4 to 5.5 kg) with  naltrexon-bupropion, and MD=5.2kg (95% CI 4.9 to 5.6) with liraglutide (see Figure 8.1).

Figure 8.1 Forrest plot showing the effect of orlistat,  naltrexon-bupropion, and liraglutide on weight in comparison to placebo. Mean differences, random-effects model (Khera, 2016).

In a direct meta-analysis by Khera (2016), all agents were associated with higher odds of achieving clinically relevant weight loss. Odds of achieving 5% weight loss from baseline compared with placebo at one year (52+4weeks) were OR=2.69 (95% CI 2.31to 3.12) for orlistat, OR=3.90 (95% CI 2.43 to 6.26) for  naltrexon-bupropion, and OR=5.19 (95% CI 2.95 to 9.15) for liraglutide (See Figure 8.2). All agents were also associated with higher odds of at least 10% weight loss from baseline compared with placebo at one year (52+4weeks) with OR=2.41 (95% CI 2.05 to 2.83) for orlistat, OR=4.11 (95% CI 2.22 to 7.63) for  naltrexon-bupropion, and OR=4.57 (95% CI 2.43 to 8.58) for liraglutide (See Figure 8.3).

Figure 8.2 Forrest plot showing the odds of patients achieving 5% weight loss using orlistat,  naltrexon and liraglutide in comparison to placebo. Odds ratio, random effects model (Khera, 2016).

Figure 8.3. Forrest plot showing the odds of patients achieving 10% weight loss using  naltrexon-bupropion in comparison to placebo. Odds ratio, random effects model (Khera, 2016).

Network meta-analysis demonstrated that significantly more patients treated with liraglutide (OR=2.06; 95% CI 1.51 to 2.96) and  naltrexon-bupropion (OR=1.47; 95% CI 1.09 to 1.96) had achieved at least 5% weight loss compared. More patients treated with liraglutide also achieved 5% weight loss compared to naltrexon-bupropion (OR=1.4; 95% CI 0.96 to 2.18), but this difference was not statistically significant (Figure 8.4)

Fig 8.4. Comparing the proportion achieving 5% weight loss in network meta-analysis. Odds ratios for comparisons are in the cell in common between the column-defining and row-defining treatment. (Khera, 2016)

In the meta-analysis by Singh (2020) significant reduction in body weight compared to baseline was observed for orlistat (mean difference, MD=−3.07Kg; 95% CI −3.76 to −2.37),  naltrexon-bupropion (MD=−4.39Kg; 95% CI −5.05 to −3.72) and liraglutide (MD=−5.25Kg; 95% CI −6.17 to −4.32), compared to placebo (See Figure 8.5).

Figure 8.5 Forrest plot showing the effect of orlistat,  naltrexon-bupropion, and liraglutide on weight in comparison to placebo. Mean differences, random-effects model (Singh, 2020).

Chao (2019) reported that weight loss in the IBT liraglutide (11.5%; SD=1.3%; P=0.005) and multicomponent (11.8%; SD=1.3%; P=0.003) groups were significantly greater compared to IBT alone (6.1; SD=1.3%). At week 52, 44% of participants in IBT-alone, 70% in IBT-liraglutide and 74% in multicomponent achieved clinically relevant weight loss of ≥5% of initial weight; 26%, 46%, and 52% of these participants, respectively, lost ≥10% of initial weight.

Wadden (2020) estimated mean weight change at week 56 was −7.5% for liraglutide-IBT and −4.0% for placebo-IBT. The estimated mean difference (MD) between groups was MD=−3.4%; (95% CI −5.3% to −1.6%). A higher proportion of 61.5% of participants achieved clinically relevant weight loss of ≥ 5% at 56 weeks with liraglutide-IBT compared to 38.8% of participants with placebo-IBT (OR 2.5%; 95% CI 1.5% to 4.1%). The proportions who lost > 10% were 30.5% (liraglutide-IBT) and 19.8% (placebo-IBT), and the proportions who lost > 15% were 18.1% (liraglutide-IBT) and 8.9% (placebo-IBT).

Semaglutide vs placebo

Two studies reported on the effect of 2.4 mg semaglutide on the percentage change from baseline in bodyweight compared to placebo (Zhong, 2022; Davies, 2021). One study was published after the systematic review by Zhong (2022) and was used to update the meta-analysis.

Zhong (2022) included 3 RCTs (3375 patients) comparing 2.4 mg semaglutide with placebo that reported on the percentage change from baseline in bodyweight. The pooled mean difference (MD) was -12.57% (95% CI -14.80 to -10.35; Figure 6) in favour of semaglutide. In addition, Zhong included 4 RCTs (3345 patients) comparing 2.4 mg semaglutide with placebo that reported on the absolute change from baseline in bodyweight. The pooled mean difference (MD) was -11.90% (95% CI -13.24 to -10.56; Figure 6) in favour of semaglutide

Davies (2021) found an mean weight loss of -9.64% + 8.04% in the semaglutide group compared to -3.42% + 8.03% in the placebo group. The MD was -6.22% (95% -7.33 to -5.11; Figure 8.6).

In summary, there were 4 RCTs (4252 patients) comparing 2.4 mg semaglutide with placebo that reported the percentage change from baseline in bodyweight. The pooled MD was -10.94% (95% CI -14.67 to -7.21; Figure 6) in favour of semaglutide. This is a clinically relevant difference. 4 RCTs (3345 patients) reported absolute change from baseline in bodyweight.

Figure 8.6 Forrest plot showing the effect of semaglutide on absolute weight change from baseline in comparison to placebo. Mean differences, random-effects model.

In addition, two studies reported on the effect of 2.4 mg semaglutide on achieving clinically relevant weight loss of 5% and 10% from baseline compared to placebo (Zhong, 2022; Davies, 2021). One study was published after the systematic review by Zhong (2022) and was used to update the meta-analysis.

Zhong (2022) included 3 RCTs (3121 patients) and found that semaglutide was associated with achieving clinically relevant weight loss of 5% (RR=2.11; 95% CI 1.59 to 2.78) and 10% (RR=3.98; 95% CI 2.58 to 6.13) from baseline compared with placebo (See Figure 7).

Davies (2021) found that semaglutide was associated with achieving clinically relevant weight loss of 5% (RR=8.35; 95% CI 5.92 to 11.77) and 10% (RR=5.53; 95% CI 3.88 to 7.89; figure 7) from baseline compared to placebo.

In summary, there were 4 RCTs (3885 patients) that reported achieving 5% and 10% weight loss compared to baseline. They found 2098/2493 events in the semaglutide group compared to 422/1392 in the placebo group of patients achieving 5% weight loss. The RR was 2.88 (95% CI 1.85 to 4.46; Figure 7) in favour of semaglutide. Moreover, they found 1707/2493 events in the semaglutide group and 202/1392 in the placebo group of patients achieving 10% weigh loss. The RR was 4.29 (95% CI 3.00 to 6.12; Figure 8.7) in favour of semaglutide.

Figure 8.7 Forrest plot showing the effect semaglutide on achieving 5% and 10% weight loss from baseline in comparison to placebo. Risk Ratio, random-effects model.

Quality of life

Health-related quality of life

Orlistat, liraglutide,  naltrexon-bupropion vs placebo

Chao (2019) used the SF-36 short form health survey to assess health-related quality of life in participants treated with IBT combined with liraglutide (IBT-liraglutide), multicomponent IBT combined with liraglutide (multicomponent) and IBT combined with placebo (IBT-alone). The SF-36 scale creates two summary measures: (1) a physical component summary and (2) a mental component summary, as well as eight subscales. Scores are transformed using a 0 to 100 scale, with higher scores indicating better health related quality of life. Changes from baseline ≥3.8 points on the SF-36 physical component summary score or ≥4.6 points on the SF-36 mental component summary score are considered clinically relevant. From baseline to week 50, compared to IBT-alone, IBT-liraglutide had significantly greater improvements in bodily pain and mental health, while multicomponent had significantly greater improvements in mental component summary, bodily pain, social functioning, role emotion and mental health (see table 8.1). At week 52, 50% of participants in IBT-alone, 34% in IBT-liraglutide and 48% in multicomponent had clinically relevant improvements in the PCS score; 32%, 42% and 44% of these participants, respectively, reported clinically relevant improvements on the mental components summary score. There were no statistically significant differences between groups in the percent of participants with clinically relevant improvements in physical- or mental component scores.

Table 8.1 Estimated mean (±SE) changes in health-related quality of life at 52 weeks, as measured from randomization, in the intention-to-treat population (N = 150)

Values shown are means (±SE). For each variable, the three treatment groups were compared using pair-wise comparisons. Abbreviations: IBT, intensive behavioural therapy.

Weight-related quality of life

Orlistat, liraglutide,  naltrexon-bupropion vs placebo

Chao (2019) used the 31-item impact of weight quality of life-Lite scale to assess weight-related quality of life among participants in the IBT-liraglutide, multicomponent, and IBT-alone arms. Scores range from 0 to 100, with 0 representing the worst quality of life and 100 representing the best. A change of 7.7 to 12 points on the total score represents a clinically relevant change. From baseline to week 52, compared to IBT-alone, IBT-liraglutide significantly had significantly greater improvements in public distress, while multicomponent had significantly greater improvements in total score, and public distress (see Table 2). At week 52, 44.4% of IBT-alone, 64.4% of IBT-liraglutide, and 65.2% of multicomponent participants achieved clinically relevant improvements in weight-related quality of life (as measured by the total score). After covariate adjustment, compared to participants in IBT-alone, the odds of achieving clinically relevant improvements in total weight-related quality of life were significantly greater for individuals in IBT-liraglutide (adjusted odds ratio (OR) = 2.4; 95% CI 1.02 to 5.8; P = 0.046) and in multicomponent (adjusted OR = 2.4; 95% CI 1.01 to 5.6; P = 0.047).

Table 8.9 Estimated mean (±SE) changes in weight-related quality of life at 52 weeks, as measured from randomization, in the intention-to-treat population (N = 150)

Values shown are means (±SE). For each variable, the three treatment groups were compared using pair-wise comparisons. Abbreviations: IBT, intensive behavioural therapy.

Semaglutide vs placebo

Davies (2021) used the 31-item impact of weight quality of life-Lite scale to assess weight-related quality of life among participants in the semaglutide 2.4 mg and placebo group. Scores range from 0 to 100, with 0 representing the worst quality of life and 100 representing the best quality of life. They found that the semaglutide group showed a change from baseline to week 68 of 10.1 + 19.7 compared to 5.3 + 21.3 for the placebo group. The MD was 4.80 (95% CI 1.89 to 7.71) in favour of the semaglutide group, but this difference was not clinically relevant.

Comorbidities

Effects on blood glucose profile

Orlistat, liraglutide, naltrexon-bupropion vs placebo

Using meta-analysis to compare different agents to placebo, Khera (2018) found that liraglutide was associated with a moderate reduction in fasting blood glucose compared to placebo (mean difference, MD=-15.6mg/dL; 95% CI −23.7 to −7.6; SMD=−0.72) followed by orlistat (MD=−8.0mg/dl; 95% CI −12.2 to −3.7; SMD=−0.23). None of these differences were, however, considered clinically relevant. Naltrexon/bupropion was not associated with a significant or clinically relevant decline in fasting blood glucose compared to placebo (MD=-2.8 mg/dL; 95% CI -10.1 to 4.4; SMD=-0.13).

Khera (2018) observed moderate reduction for HbA1c with liraglutide compared to placebo (MD=−0.5%; 95% CI −0.9 to −0.2; SMD, −0.54) followed by orlistat (MD=−0.4%; 95% CI −0.6 to −0.2; SMD, −0.38). Only liraglutide resulted in a clinically relevant decline. Naltrexon/bupropion was not associated with a significant or clinically relevant decline in HbA1c compared to placebo (MD=-0.5%; 95% CI -1.1, 0.1; SMD=-0.33).

Wadden (2020) found a significant but not a clinically relevant difference in reduction of fasting blood glucose between liraglutide-ITD and placebo-ITD at 56 weeks (MD=−0.23 mmol/L; 95% CI −0.36 to −0.11). Wadden (2020) also found a significant and clinically relevant difference in reduction of HbA1c (MD−0.10%; 95% CI −0.2 to −0.04).

Semaglutide vs placebo

Zhong (2022) included 3 RCTs (3375 patients) and found that compared with placebo, once-weekly semaglutide significantly reduced HbA1c (MD –0.25%, 95% CI –0.31 to –0.19), but this was not a clinically relevant difference. Moreover, the reduction of fasting plasma glucose was greater in patients who received once-weekly semaglutide than in patients who received placebo (MD –7.40 mg/dL, 95% CI –8.40 to –6.41), but this was also not a clinically relevant difference.

Davies (2021) found that the semaglutide group showed a change in HbA1c from baseline to week 68 of -1.6% + 1.95% compared to -0.4% + 1.93% for the placebo group. The MD was -1.20 % (95% CI -1.48 to -0.92) in favour of the semaglutide group. This is a clinically relevant difference.

Effects on cholesterol profile

Orlistat, liraglutide, naltrexon-bupropion vs placebo

Using direct meta-analysis to compare different agents to placebo, Khera (2018) found that orlistat was associated with a significant and clinically relevant reduction in LDL-cholesterol compared to placebo (MD=−8.7mg/dl; 95% CI −10.7 to −6.7; SMD=−0.27). Liraglutide (MD=-3.6; 95% CI -7.8 to 0.6; SMD=-0.11) and naltrexon/bupropion (MD= -1.86; 95% CI -4.87 to 1.16; SMD=-0.09) was not associated with a significant or clinically relevant reduction in LDL-cholesterol compared to placebo. For HDL-cholesterol compared to placebo, naltrexon-bupropion (MD=2.5mg/dL; 95% CI 1.2 to 3.8; SMD 0.40) was associated with increase in HDL-cholesterol at 1 year follow-up. In contrast, orlistat was associated with decline in HDL-cholesterol (MD=−1.1mg/dL; 95% CI −1.9 to −0.4; SMD, −0.11) compared to placebo, but this decline was not clinically relevant.

Wadden (2020) found no significant difference in LDL-cholesterol between liraglutide-ITD and placebo-ITD at 56 weeks (MD=−0.07; 95% CI −0.21 to 0.06) and in HDL-cholesterol (MD=0.02; 95% CI −0.02 to 0.07).

Semaglutide vs placebo

Zhong (2022) included 2 RCTs (1414 patients) and found that semaglutide reduced the MD in percentage of total cholesterol by -5.93% (95% CI -4.34 to -7.51) when compared with placebo. In addition, compared with placebo, the reduction in the MD in percentage of LDL was -6.55% (95% CI -4.19 to -8.92), but semaglutide showed a neutral effect on the percentage of HDL with a MD of 0.54% (95% CI -1.46 to 2.53)

Davies (2021) found no significant or clinically relevant difference between the semaglutide and placebo group at 68 weeks to baseline ratio of total cholesterol (MD=0.99; 95% CI 0.96 to 1.02), HDL-cholesterol (MD=1.03; 95% CI 1.00 to 1.05) and in LDL-cholesterol (MD=1.00; 95% CI 0.96 to 1.05).

Effects on blood pressure

Orlistat, liraglutide, naltrexon-bupropion vs placebo

Using direct meta-analysis to compare different agents to placebo, Khera (2018) found that liraglutide (MD= -2.8mmHg; 95% CI −4.3 to −1.2; SMD, −0.24) and orlistat (MD= -1.7mm Hg; 95% CI −2.4 to −0.9; SMD, −0.19) were associated with modest reductions in systolic blood pressure, but these reductions were not considered clinically relevant. No reduction was observed for naltrexon/bupropion (MD=1.1; 95% CI -0.7 to 2.8, SMD=0.05). Similar changes were seen in diastolic blood pressure with modest reductions observed for liraglutide (MD=-0.83; 95% CI -1.7 to 0.0, SMD=-0.08) and orlistat (MD=-1.6; 95% CI -2.0 to -1.1, SMD=-0.71), but these reductions were also not considered clinically relevant. No reduction was observed for naltrexon/bupropion (MD=0.1; 95% CI -0.6 to 0.7; SMD=0.01).

Wadden (2020) found no significant or clinically relevant difference in systolic blood pressure between liraglutide-ITD and placebo-ITD at 56 weeks (MD−2.2; 95% CI −4.9 to 0.5) and in diastolic blood pressure (MD=−0.2; 95% CI −2.12 to 1.8).

Semaglutide vs placebo

Zhong (2022) included 3 RCTs (3375 patients) and found that once-weekly semaglutide resulted in a significant and clinically relevant reduced systolic blood pressure with a MD of -4.62 mm Hg (95% CI -5.57 to -3.66) when compared with placebo, and the decrease of diastolic blood pressure was also more significant but not clinically relevant in the once-weekly semaglutide group with a MD of -1.82 mm Hg (95% CI -2.94 to -0.70) when compared with placebo.

Davies (2021) found that the semaglutide group showed a change in systolic blood pressure from baseline to week 68 of -3.9 mm Hg + 13.8 mm Hg compared to -0.5 mm Hg + 15.5 mm Hg for the placebo group. The MD was -3.40 mm Hg (95% CI -5.48 to -1.32) in favour of the semaglutide group. In addition, diastolic blood pressure showed showed a change in diastolic blood pressure from baseline to week 68 of -1.6 mm Hg + 7.9 mm Hg compared to -0.9 mm Hg + 9.7 mm Hg for the placebo group. The MD was -0.70 mm Hg (95% CI -1.95 to 0.55) in favour of the semaglutide group.

Effects on waist circumference

Orlistat, liraglutide, naltrexon-bupropion vs placebo

Using direct meta-analysis to compare different agents to placebo, Khera (2018) found that all agents were associated with a decrease in waist circumference liraglutide by 4cm (95% CI −5.0, −3.3; SMD=−0.63), naltrexon-bupropion 3.5cm (95% CI −4.4 to −2.6; SMD=−0.37 and orlistat by 2.3cm (95% CI −2.8 to −1.7; SMD=−0.26). Only liraglutide and naltrexon-bupropion were also associated with clinically relevant decrease of at least 3cm.

Wadden (2020) showed that waist circumference decreased significantly more in liraglutide-IBT than in the placebo-IBT participants (MD=−2.7 cm; 95% CI −4.7 to −0.8), but this difference was not considered clinically relevant.

Semaglutide vs placebo

Zhong (2022) included 3 RCTs (3375 patients) and fount that compared with placebo, once-weekly semaglutide was associated with a MD reduction in waist circumference of –9.34 cm (95% CI –10.00 to –8.68).

Davies (2020) found that the semaglutide group showed a change from baseline to week 68 of -9.4 cm + 7.9 cm compared to -4.5 cm + 7.8 cm for the placebo group. The MD was -4.90 cm (95% CI -6.01 to -2.79) in favour of the semaglutide group.

Adverse events

Orlistat, liraglutide, naltrexon-bupropion vs placebo

Wadden (2020) reported that Liraglutide 3.0 mg combined with IBT was generally well tolerated. The most frequent adverse events were gastrointestinal (71% with liraglutide-IBT vs. 49% with placebo-IBT), The incidence of nausea was greater with liraglutide-IBT (47.9%) than with placebo-IBT (17.9%). Three acute gallstone disease events occurred with liraglutide-IBT and two with placebo-IBT. No events of pancreatitis were observed in the study. The proportion of participants with reported serious adverse events was 4.2% (seven events in six participants) in liraglutide-IBT and 1.4% (two events in two participants) with placebo-IBT. There were no deaths.

Non of de included assessed the association between adverse events and other weight loss medication (orlistat and naltrexon-bupropion) in treating obesity or overweight in adults.

Semaglutide vs placebo.

Zhong (2022) found no significant differences in the risk of hypoglycemic events, acute pancreatitis, acute renal failure, malignant neoplasms, and all-cause death between once-weekly semaglutide and placebo. However, nausea (RR=2.60; 95% CI 2.25 to 2.99), vomiting RR(=3.35; 95% 2.65 to 4.24), diarrhoea (RR=1.92; 95% CI 1.61 to 2.29), and constipation (RR=1.91; 95% CI 1.34 to 2.72) were reported more frequently with once-weekly semaglutide than with placebo.

Davies (2021) found that 353/503 (87.6%) of patients reported adverse events in the semaglutide 2.4 mg group compared to 309/402 (76.9%) patients in the placebo group. The RR was 1.14 (95% CI 1.07 to 1.22) in favour of placebo.

Cost effectiveness

No studies assessed the cost effectiveness of weight loss medication on treating obesity or overweight in adults.

Level of evidence of the literature

Orlistat, liraglutide, naltrexon-bupropion vs placebo

The level of evidence of RCTs comparing orlistat, liraglutide and/or  naltrexon-bupropion vs placebo starts high. The level of evidence regarding the outcome measure weight loss was downgraded by one level because of study limitations (risk of bias due to high attrition rates for all trials).

The level of evidence regarding quality of life were downgraded to very low because of risk of bias and serious imprecision. There was high risk of bias since blinding of the trials was unclear or not possible and only one study reported on quality of life with large standard errors of effect size (imprecision).

The level of evidence regarding the outcome measure comorbidities was downgraded moderate because of study limitations (risk of bias due to high attrition rates for all trials). It was downgraded to low for HbA1c, LDL-cholesterol and blood pressure outcomes because of conflicting results (inconsistency).

Semaglutide vs Placebo

The level of evidence of RCTs comparing semaglutide vs placebo starts high. The level of evidence regarding the outcome measure weight loss was downgraded by one level to moderate because of inconsistency (due to considerable heterogeneity among the included RCTs, possibly attributable to the variation in sample sizes, baseline characteristics of patient populations, and trial durations).

The level of evidence regarding quality of life was downgraded with two levels to low because of serious imprecision (95% confidence intervals crossing the border for clinical relevance).

The level of evidence regarding the outcome measure HbA1c, blood pressure, and waist circumference were downgraded to moderate. The level of evidence for the outcome measure HbA1 was downgraded because of conflicting results (inconsistency). The level of evidence for blood pressure and waist circumference was downgraded because the 95% confidence interval crossed the border for clinical relevance (imprecision).

The level of evidence regarding the outcome measure glucose and cholesterol was downgraded to low because of conflicting results (inconsistency) and the 95% confidence interval crossing the border for clinical relevance, and only one study reporting on the outcome with a low number of events (imprecision).

A systematic review of the literature was performed to answer the following question:

What is the effect of pharmaceutical treatment combined with behavioural or lifestyle intervention compared to placebo in treating overweight with comorbidities or obesity in adults?

P: Adults with obesity (BMI ≥30 kg/m²) and adults with overweight (BMI ≥25 kg/m²) with comorbidities.

I: Weight loss medication (orlistat, naltrexon/bupropion, liraglitude 3 mg, semaglutide).

C: Placebo.

O: Weight loss, maintaining the quality of life, comorbidities (glucose/diabetes), adverse events, cost-effectiveness.

Relevant outcome measures

The working group considered weight as a crucial outcome measure for decision making; and comorbidities and quality of life as important outcome measures for decision making.

The working group defined the outcome measure comorbidities as follows: diabetes-related parameters (HbA1c, glucose) and cardiovascular risk factors (HDL cholesterol, LDL cholesterol, systolic blood pressure, diastolic blood pressure)

The working group defined the following outcomes as a minimal clinically (patient) important difference.

• Weight loss: 5% (Khera, 2016).
• Quality of Life: changes from baseline ≥ 3.8 points on the SF-36 physical component summary score or ≥ 4.6 points on the SF-36 mental component summary score (Chao 2019).
• Weight related quality of life: A change of 7.7 to 12 points on the total score (Chao 2019).
• Comorbidities
  • Cardiovascular risk factors: SBP: 3 mmHg; DBP: 2 mmHg; LDL: 0.13 mmol/L (5 mg/dL) 5% reduction; and HDL: 0.05 mmol/L (2 mg/dL) or a 5% increase
  • Glucose: 1 mmol/L (18 mg/dL).
  • HbA1c: 0.5%
  • Waist circumference: 3 cm

Search and select (Methods)

Three different literature searches were performed. At first we searched for systematic reviews and after this we performed an additional search to supplement the selected review(s) with RCT’s that were published after the search date of the selected review(s). Finally, during development of this module we learned that semaglutide was approved as weight loss medication for the treatment of obesity and updated our search to include it.

Search 1: Systematic reviews (SR)

• The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms from the 1^st of January 2005 until the 27^th of June 2020. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 382 hits. Studies were selected based on the following criteria: systematic reviews (SR) investigating the effect of pharmacological treatments for obesity on weight loss and cardiovascular risk in adults with obesity and overweight compared to adults with obesity and overweight receiving placebo. Fifteen studies were initially selected based on title and abstract screening. After reading the full text, 12 studies were excluded (see the table with reasons for exclusion under the tab Methods), and three studies were included (Khera, 2016; Khera, 2018; Singh; 2020).

Search 2: Randomized controlled trials (RCTs)

• The search strategy of the systematic review (Singh, 2020) was complete up on 30 September 2019. Therefore, we performed an additional search on RCTs in the databases Medline (via OVID) and Embase (via Embase.com) with relevant search terms between 1^st of January 2019 until and the 22^nd of July 2020. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 184 hits. Studies were selected based on the following criteria: RCTs investigating the effect of pharmacological treatments for obesity on weight loss and cardiovascular risk in adults with obesity and overweight compared to adults with obesity and overweight receiving a placebo. Four studies were initially selected based on title and abstract screening. After reading the full text, two studies were excluded (see the table with reasons for exclusion under the tab Methods), and two studies were included (Chao, 2019; Wadden, 2020).

Search 3: Semaglutide

• The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms from the 1^st of January 2005 until the 7^th of February 2022. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 59 hits. Studies were selected based on the following criteria: SRs and RCTs investigating the effect of semaglutide as treatment for obesity on weight loss and cardiovascular risk in adults with obesity and overweight compared to adults with obesity and overweight receiving placebo. 12 studies were initially selected based on title and abstract screening. After reading the full text, 10 studies were excluded (see the table with reasons for exclusion under the tab Methods), and 2 studies were included Zhong, 2022; Davies, 2021).

Results

7 studies were included in the analysis of the literature. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables.