CADTH Reimbursement Review

Liraglutide (Saxenda)

Sponsor: Novo Nordisk Canada Inc.

Therapeutic area: Chronic weight management in adults

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Introduction

The WHO (WHO) defines overweight and obesity as abnormal or excessive fat accumulation that presents a risk to health.¹ The 2 are distinguished from each other based on body mass index (BMI), where a BMI over 25 kg/m² is considered overweight, and over 30 kg/m² is obese.¹ Identified determinants of overweight and obesity include physical activity, diet, socioeconomic status (such as income and education), ethnicity, immigration, and environmental factors.² In 2019, overweight and obesity rates among adults aged 18 years to 79 years in Canada were 35.5% and 24.3%, respectively.³ The Canadian Task Force on Preventive Health Care (CTFPHC) has reported that over 2-thirds of Canadian men (67%) and more than half of Canadian women (54%) are overweight or living with obesity.⁴

Obesity is associated with an increased risk of a wide range of illnesses and long-term conditions, including type 2 diabetes, hypertension, gallstones, gastroesophageal reflux disease, and cancer, as well as psychological and psychiatric morbidities.⁵ WHO has reported that more than 4 million people die each year as a result of being overweight or living with obesity.¹ It is estimated that median survival is reduced by 2 years to 4 years for those with a BMI of 30 kg/m² to 35 kg/m² and by 8 years to 10 years for those with a BMI of 40 kg/m² to 50 kg/m,² whereas weight loss of 5 kg to 10 kg reduces the long-term risk of diseases associated with obesity.⁵ For example, it has been estimated that a loss of 10 kg may lead to a reduction in total cholesterol of 0.25 mmol/L and a reduction in diastolic blood pressure (DBP) of 4 mm Hg.⁶

The treatment of choice to manage overweight and obesity is a multi-component interventions approach with respect to provider discipline, as well as length and format of treatment.^4,7 The CTFPHC has stated that behavioural interventions focusing on diet, increasing exercise, making lifestyle changes, or any combination of these are the preferred options, as the benefit-to-harm ratio appears more favourable than for pharmacologic interventions.⁴ However, while initially effective for many individuals, diet or behaviour modification alone is often difficult to sustain and many individuals regain weight upon discontinuation.⁸ Therefore, drug therapy for chronic weight management as an adjunct to lifestyle intervention may help individuals to achieve and sustain clinically relevant weight loss.⁸ Pharmacotherapy for weight management is indicated only for those with a BMI of at least 30 kg/m², or those with a BMI of at least 27 kg/m² with at least 1 comorbidity who have failed a previous lifestyle attempt at weight loss,⁹ and may be considered only after dietary, exercise, and behavioural approaches have been started and evaluated, and for patients who have not reached their target weight loss or have reached a plateau on dietary, activity, and behavioural changes.⁷ Two other drugs approved for use in Canada with similar indications are orlistat and the naltrexone hydrochloride and bupropion hydrochloride (NB) combination tablet.

The drug under review is liraglutide 3 mg (Saxenda), which is a human glucagon-like peptide-1 (GLP-1) analogue with Health Canada approval to be used as an adjunct to a reduced-calorie diet and increased physical activity for chronic weight management as detailed in Table 1.¹⁰ The drug regulates appetite by increasing feelings of fullness and satiety, while lowering feelings of hunger and prospective food consumption.¹⁰ Liraglutide 3 mg is available as a solution for subcutaneous injection at a strength of 6 mg/mL in a pre-filled, multi-dose pen.¹⁰ Treatment is initiated at a dose of 0.6 mg daily for 1 week, after which the daily dose is escalated at weekly increments of 0.6 mg over 4 weeks to reach the recommended daily maintenance dose of 3 mg daily.¹⁰

The objective of this systematic review is to evaluate the beneficial and harmful effects of liraglutide 6 mg/mL for subcutaneous injection as an adjunct to a reduced-calorie diet and increased physical activity for chronic weight management in adults.

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient groups that responded to CADTH’s call for patient input and from clinical expert(s) consulted by CADTH for the purpose of this review.

Patient Input

Three Canadian patient organizations — Obesity Canada, Diabetes Canada, and the Gastrointestinal (GI) Society — provided input for this submission. Obesity Canada is Canada’s leading obesity registered charity association for professionals and patients, contributing research, education, and advocacy for Canadians living with obesity. Diabetes Canada is a national health charity that represents Canadians living with, or at risk of, diabetes and focuses on research and policy initiatives for diabetes prevention, care, and cure at the population level. The GI Society is a national charity committed to improving the lives of people with GI and liver conditions through research, advocating for increased access to health care, and promoting GI and liver health.

Obesity Canada conducted patient interviews and an online survey, where 5 of the individuals interviewed and 60% of the survey respondents had used liraglutide 3 mg for obesity management. Diabetes Canada submitted patient input using data from 2 online surveys conducted in July and August 2020 and in December 2020 and January 2021. The GI Society used data from patient and patient caregiver interviews, the results of published studies, and a survey conducted from October 6, 2020, to January 10, 2021, among individuals living with obesity. All 3 patient input groups submitted conflict of interest disclosures, which can be found on the CADTH website.

Patient groups reported that obesity not only increases the potential for the development of further disease(s), but it leads to inequities in access to employment, health care, and education due to the strong stigma associated with it. The GI Society reported that 72% of their survey respondents experienced social stigma as a result of living with obesity, with many reporting that they avoid getting medical care because they feel as though their physician shames them because of their weight. Individuals living with obesity also report frustration with the impact that the chronic and often misunderstood disease has on their overall quality of life.

Currently, most Canadians living with obesity reported using diet and exercise, medications, or bariatric surgery to combat the disease. Many who diet and exercise have difficulty sustaining their efforts or finding a program that suits their needs, which can lead to depression, hopelessness, and further weight gain. Approved medications include NB (Contrave), liraglutide (Saxenda), and orlistat (Xenical), all of which have undesirable side effects. Despite the availability of these drugs, the GI Society reported many concerns about obtaining and paying for the prescriptions of these drugs.

Patient groups reported that many patients would like a treatment that is effective in the long-term, is affordable, and has no or minimal side effects. Patient groups hope that liraglutide 3 mg may help people to better manage their weight, potentially delaying or preventing the development of comorbidities, such as the progression of prediabetes to type 2 diabetes. Furthermore, when asked about outcomes to consider, it was reported that patients focused less on improved weight than on improved health-related comorbidities (e.g., diabetes, hypertension, and sleep apnea) as well as outcomes related to everyday life such as productivity, energy levels, sleep, activity, and mental health.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinical expert consulted for this review noted that side effects limit the use of all 3 pharmacotherapies approved in Canada, with some patients needing to stop the medication completely due to side effects. The clinical expert acknowledged that the weight loss seen with medications for obesity will often fall within a range of 5% to 10%, which is considered adequate to ameliorate weight-related comorbidities such as type 2 diabetes and osteoarthritis. However, in an ideal world, we would have pharmacotherapy that promotes larger amounts of weight loss.

The expert indicated that in accordance with the Obesity Canada 2020 guidelines, pharmacotherapy can realistically be used anywhere in a patient’s weight management journey. Further, although the expert indicated that it is reasonable for a patient to try lifestyle intervention first before starting liraglutide 3 mg, that approach is not necessary, particularly for patients at higher BMIs. Since the mechanism of action is very different from NB and orlistat, some physicians are prescribing combinations of these agents for weight loss, despite minimal evidence to support this.

The clinical expert stated that currently, there is no way to predict which patients will lose the most weight with liraglutide, but that qualifying patients can be identified using the traditional definition of overweight and obesity based on BMI category. Patients most in need of intervention for weight loss are those who have the highest burden of weight-related comorbidities, those with Edmonton Obesity Staging System (EOSS) score ranging from EOSS 1 to EOSS 3, and patients with Class III obesity (i.e., BMI > 40 kg/m²).

The clinical expert consulted for this review stated that in general, weight-loss outcomes are assessed based on change in BMI and weight, whereas weight-related comorbidity outcomes are evaluated using change in parameters such as blood pressure, glycemic control, and lipid profile. Weight-loss response is assessed after 12 weeks to 16 weeks at the maximum dose (along with behavioural and lifestyle changes) to decide whether to continue the medication at that point. The expert indicated that most physicians would agree that a 5% to 10% total body weight loss is clinically meaningful and is typically felt to be associated with improved metabolic parameters. Based on experience, the expert stated that once the patient is successful on a stable dose of liraglutide 3 mg, assessments could be spaced at 3-month intervals, and eventually every 6 months.

According to the clinical expert consulted for this review, treatment discontinuation decisions are influenced by patients’ preferences, side effects (most commonly, GI side effects), and rare but serious adverse effects such as pancreatitis, worsening in mood, or increase in anxiety.

According to the clinical expert, primary care providers can diagnose obesity and overweight, safely prescribe liraglutide 3 mg, and monitor their patients over time.

Clinician Group Input

No clinician group input was received for this review.

Drug Program Input

The drug plans requested clarification regarding the potential target patient population and the anticipated treatment duration of liraglutide. The clinical expert considered that it would be reasonable to consider pharmacotherapy in patients with a BMI between 27 kg/m² and 30 kg/m² without comorbidity as second-line after lifestyle changes, although this would be beyond the Health Canada–approved indication. The expert felt that patients would regain the weight they had lost if pharmacologic treatment for weight management was discontinued; therefore, such treatments would need to be continued in the long-term, even in patients whose BMI dropped below 30 kg/m² (or 27 kg/m² in patients with weight-related comorbidities). The public drug plans also requested clarification regarding re-treatment in patients who regain weight, or if liraglutide 3 mg becomes ineffective over time after an initial desired response. The clinical expert noted that if there is no benefit after a patient tries weight-loss medication for the first time, it is unlikely they will respond better to it in the future. Therefore, it is unlikely that the drug will be prescribed for the same indication again in that patient.

Clinical Evidence

Pivotal Studies and Protocol Selected Studies

Description of Studies

A total of 4 phase III RCTs (i.e., Study 1839,¹¹ Study 1922,¹² Study 1923,¹³ and Study 3970¹⁴) met the inclusion criteria of this CADTH systematic review. They were all parallel-group, multi-centre, double-blind, placebo-controlled trials, conducted in multiple sites and in 2 countries or more, including Canada. A total of 5,358 adult patients were randomized across the 4 studies. Study 1839 consisted of a 56-week main phase and a 104-week extension phase. The primary objective of the main phase of Study 1839¹¹ was to establish the efficacy of liraglutide 3 mg compared with placebo in inducing and maintaining weight loss in patients who are overweight or living with obesity without diabetes over 56 weeks, whereas the objective of the extension phase¹⁵ was to investigate the long-term efficacy of liraglutide 3 mg compared to placebo in delaying the onset of type 2 diabetes mellitus (T2DM) in patients who are overweight or living with obesity, and who had prediabetes at screening, after a total of 160 weeks of treatment in the main and extension phases. Of the total 3,731 patients randomized to take part in the main phase of the trial, 2,200 patients with prediabetes at screening continued into the 104-week extension phase.¹⁵ The primary objectives of the other trials were as follows:

• to investigate the efficacy of liraglutide 3 mg compared to placebo in inducing and maintaining weight loss in patients who are overweight or living with obesity with T2DM after 56 weeks (Study 1922)¹²

• to compare the efficacy of liraglutide 3 mg versus placebo in maintaining run-in weight loss (≥ 5% achieved in a 4-week to-12-week low-calorie run-in period) over 56 weeks in patients living with obesity or patients who are overweight with comorbidities (Study 1923)¹³

• to compare the efficacy of liraglutide 3 mg versus placebo in inducing weight loss beyond that achieved in run-in (≥ 5% achieved in a 4-week to 12-week low-calorie run-in period) over 56 weeks in patients living with obesity or patients who are overweight with comorbidities (Study 1923)¹³

• to investigate if treatment with liraglutide 3 mg reduces the severity of obstructive sleep apnea (OSA) (assessed by the apnea-hypopnea index [AHI]) compared to placebo, both in combination with lifestyle intervention in patients living with obesity and moderate or severe OSA who were unable or unwilling to use continuous positive airway pressure (CPAP) treatment (Study 3970).¹⁴

In all the studies, patients were randomly assigned to receive once-daily subcutaneous injections of liraglutide 3 mg or placebo that matched the active drug in appearance, quantity, and route of administration. Treatments were started in accordance with the product monograph recommendation of 0.6 mg daily for 1 week, titrated at weekly increments of 0.6 mg over 4 weeks to reach the recommended daily maintenance dose of 3 mg to mitigate GI side effects. Patients in both the liraglutide 3 mg and placebo groups received counselling on lifestyle modification from randomization and throughout the entire trials. Counselling was provided by a qualified dietitian according to local standards and the patients were put on a reduced-calorie diet containing a maximum of 30% of energy from fat, approximately 20% of energy from protein, and approximately 50% of energy from carbohydrates, and with an energy deficit of approximately 500 kcal/day compared with the patients’ estimated total energy expenditure.

Three studies^11-13 had 3 co-primary end points assessed after 56 weeks of treatment. These comprised the percentage change in fasting body weight from baseline,^11-13 the proportion of patients losing 5% or more of baseline fasting body weight (5% responders),^11-13 the proportion of patients losing more than 10% of baseline fasting body (10% responders),^11,12 and the percentage of patients maintaining run-in fasting weight loss from baseline.¹³ The primary end point for the extension phase of Study 1839 was the proportion of patients with onset of T2DM at week 160 among patients who had prediabetes at baseline,¹⁵ and the primary end point for Study 3970 was change from baseline in AHI after 32 weeks.¹⁴ Secondary end points reported by the trials that met the protocol’s listed outcomes for this review included changes in BMI, health-related quality of life (HRQoL), glycemic control, and weight-related comorbidity (i.e., blood pressure and lipid profile parameters outcomes), though no secondary end points were controlled for multiple comparisons.

Overall, the treatment groups in all the included studies appeared well balanced with respect to baseline demographics and other characteristics. The study populations in all the trials were predominantly White, with percentages ranging from 72% to 88%, and most of the patients (> 85%) had a BMI of 30 kg/m² or more. In 3 of the studies, the percentage of female participants was between 54% and 78%. Patients’ mean ages ranged between 45 years and 55 years across the studies. In Study 1923, only patients who achieved at least 5% loss in body weight on a low-calorie diet during the run-in period continued to the randomization phase for allocation to either liraglutide 3 mg or placebo for 56 weeks.

Efficacy Results

Table 2 presents a summary of key end point results from the included studies.

Percentage Change in Body Weight From Baseline

Primary analysis results from the main phase of Study 1839 showed that liraglutide 3 mg was superior to placebo regarding the percentage weight loss from baseline after 56 weeks of treatment, with a treatment difference of −5.39 (95% CI, −5.82 to −4.95; P < 0.0001). The other studies reported consistent findings with the main phase of Study 1839, as shown by the following treatment estimate differences:

• Study 1839 extension: Difference = –4.32 (95% CI, –4.94 to −3.70; not controlled for multiplicity)

• Study 1922: Difference = –3.97 (95% CI, –4.84 to –3.11; P < 0.0001)

• Study 1923: Difference = –6.06 (95% CI, –7.50 to –4.62; P < 0.0001)

• Study 3970: Difference = –4.15 (95% CI, –5.21 to –3.09; not controlled for multiplicity)

5% Responders

Primary analysis results from the main phase of Study 1839 showed that liraglutide 3 mg was superior to placebo for the odds of achieving at least 5% reduction from baseline body weight after 56 weeks of treatment. The odds ratio (OR) was 4.80 (95% CI, 4.12 to 5.60; P < 0.0001). The other studies reported consistent findings, as shown by the following ORs:

• Study 1839 extension: OR = 3.22; (95% CI, 2.63 to 3.94; not controlled for multiplicity)

• Study 1922: OR = 6.81 (95% CI, 4.34 to 10.68; P < 0.0001)

• Study 1923: OR = 3.86 (95% CI, 2.44 to 6.09; P < 0.0001)

• Study 3970: OR = 3.92 (95% CI, 2.41 to 6.38; not controlled for multiplicity)

10% Responders

Primary analysis results from the main phase of Study 1839 showed that liraglutide 3 mg was superior to placebo for the odds of achieving greater than 10% reduction from baseline body weight after 56 weeks of treatment. The OR was 4.34 (95% CI, 3.54 to 5.32; P < 0.0001). The other studies reported consistent findings with the pivotal study, as shown by the following ORs:

• Study 1839 extension: OR = 3.09 (95% CI, 2.35 to 4.05; not controlled for multiplicity)

• Study 1922: OR = 7.10 (95% CI, 3.48 to 14.48; P < 0.0001)

• Study 1923: OR = 5.30 (95% CI, 2.79 to 10.08; not controlled for multiplicity)

• Study 3970: OR = 18.96 (95% CI, 5.69 to 63.14; not controlled for multiplicity)

Maintaining Run-In Weight Loss

Primary analysis results from Study 1923 showed that liraglutide 3 mg was superior to placebo for the odds of maintaining run-in weight loss after 56 weeks of treatment. The OR was 4.82 (95% CI, 3.01 to 7.71; P < 0.0001). No other study measured this outcome.

Time to Onset of Type 2 Diabetes Mellitus

Primary analysis results from the extension phase of Study 1839 showed that liraglutide 3 mg was superior to placebo in delaying the progression to T2DM among patients with prediabetes after 160 weeks of treatment in the trial’s main and extension phases. A Weibull analysis showed an annualized T2DM incidence rate of 0.8 for liraglutide 3 mg versus 3.2 for placebo, with a treatment estimate of 2.681 (95% CI, 1.856 to 3.872; P < 0.0001). Thus, after 160 weeks of treatment, the estimated time to onset of T2DM for prediabetes patients treated with liraglutide 3 mg was close to 3 times longer than prediabetes patients treated with placebo. No other study measured this outcome.

Secondary outcomes identified as relevant to the CADTH review included change in BMI, HRQoL, and outcomes associated with weight-related comorbidities such as the development of T2DM, glycemic control, and change in other medications. However, none of these outcomes was controlled for multiplicity; therefore, results must be considered with regard to type I error.

Harms Results

Key harms results have been summarized in Table 3. Treatment-emergent adverse events (TEAEs) were more common with liraglutide 3 mg than placebo in all the trials. The overall adverse event (AE) rates associated with liraglutide 3 mg were between 80.1% and 94.7%, whereas the incidence of AEs was between 69.3% and 89.4% with placebo. The most common AEs (i.e., occurring in ≥ 5% of patients) with liraglutide 3 mg across all the included studies were nausea (26.7% to 47.6%), diarrhea (16.5% to 25.6%), and constipation (11.9% to 26.9%).

Serious adverse event (SAE) rates were between 3.4% and 15.1% with liraglutide 3 mg compared with 2.4% to 12.9% with placebo. The most frequent SAEs with liraglutide 3 mg (i.e., occurring in 1% or more of patients) were hepatobiliary disorders (2.5% or less, only in Study 1839), and infections and infestations (2.3%, in Study 1839 only). Rates of neoplasms (i.e., benign, malignant, and unspecified) were 1.7%, 1.9%, and 2.1% in Study 1922, Study 1923, and the Study 1839 extension, respectively.

The percentage of patients who discontinued treatment prematurely due to AEs was higher with liraglutide 3 mg than with placebo in all the studies. The rate of discontinuation due to AEs ranged from 8.6% to 13.3% in the liraglutide 3 mg group compared with 3.3% to 11.1% in the placebo group.

Briefly, in Study 1839, 1 death occurred in the liraglutide 3 mg group and 2 deaths occurred in the placebo group during the main phase of the study. By the end of the study’s extension phase, each group had a total of 2 deaths, corresponding to mortality rates of 0.1% and 0.3% for the liraglutide 3 mg and placebo groups, respectively. No deaths were reported during Study 1922 or Study 3970, and 1 death occurred in the placebo group in Study 1923.

GI symptoms were the most frequent AEs overall, and they were also the most common notable harms. The clinical expert consulted for the review noted that GI AEs are common with all drugs approved in Canada for chronic weight management, adding that they can be managed by introducing the drug gradually over a period of time to get to the maximum effective dose. The product monograph of liraglutide 3 mg provides a dose escalation schedule intended to help mitigate GI AEs.

Critical Appraisal

Although all the trials were double-blind randomized studies with a placebo control group, the proportion of patients who discontinued prematurely was high in all of the studies, resulting in significant amounts of missing data, which were imputed using last observation carried forward (LOCF). However, the higher rate of discontinuation due to AEs and the significant difference in weight loss with liraglutide 3 mg compared with placebo, as well as the larger proportion of patients who discontinued the study due to ineffective therapy in the placebo group, may have resulted in unblinding for some patients. In all the trials, statistical testing procedures were based on the primary end point such that the determination of study power and sample sizes did not consider secondary outcomes. Thus, there is a risk of type I error inflation in key outcomes such as HRQoL, glycemic control, and weight-related CV comorbidities. Moreover, depending on a patient’s response to treatment after 28 weeks, the study permitted the recalculation of dietary portions. However, there was no data to independently verify how often the diet adjustments happened and if they occurred in a balanced manner across treatment groups. It is important to note that considering the critical importance of calorie intake to weight management, an imbalance in this component of the co-intervention has the potential to tilt the outcomes in favour of 1 group over the other.

In all the trials, patients within the BMI bracket of 30 kg/m² to 40 kg/m² or more accounted for more than 85% of the study population. This indicates that a group of patients specified in the indication (i.e., patients who are overweight with a BMI of 27 kg/m² to less than 30 kg/m² with comorbidities) was not adequately represented in any of the studies. Further, the trials enrolled predominantly White patients (72% to 88%) and in 3 of the studies, the patients were mostly women (54% to 78%). According to the clinical expert consulted in this review, the study population does not reflect the ethnicity mix of patients who are overweight or living with obesity in Canada. Also, the high proportion of women in the study populations differs from the 67% rate of obesity in adult males in Canada, as reported by CTFPHC.⁴ The exclusion criteria denied entry to some patients, such as those on medication that causes weight gain and those regaining weight after a previous bariatric surgery, who would be considered clinically relevant patients and who may require pharmacotherapy for chronic weight management. The extent to which these issues affect the generalizability of the reported findings is unknown.

Indirect Treatment Comparisons

Description of Studies

One systematic review with network meta-analysis (NMA)¹⁶ published in 2016 was included in the review. It compared 5 FDA-approved weight-loss drugs for efficacy and adverse effects in patients living with obesity (BMI ≥ 30 kg/m²) or overweight (BMI ≥ 27 kg/m²) with at least 1 weight-related comorbidity. The drugs tested were liraglutide 3 mg, orlistat, naltrexone hydrochloride 8 mg and bupropion hydrochloride 90 mg combination in extended-release tablets, lorcaserin, and phentermine-topiramate [PT]). Eligible studies had to be RCTs using the most effective FDA-approved dosage of the drug for at least 1 year and reporting outcomes on differences in mean weight loss between treatment groups or the proportion of patients achieving at least 5% weight loss. Primary studies for the systematic review of the NMA were identified through a systematic literature search of multiple databases from inception until March 2016. Study screening and data extraction were performed independently by 2 reviewers, resolving conflicts by consensus involving a third reviewer. The quality of the primary RCTs was assessed using the Cochrane Risk of Bias tool, and the GRADE method¹⁷ was used to evaluate the quality of evidence in the NMA.

The primary efficacy outcome of the NMA was the proportion of patients with 5% or more weight loss from baseline at 1 year. Other efficacy outcomes assessed were the proportion of patients with at least 10% weight loss and the incremental change in weight in kilograms from baseline over placebo after 1 year of follow-up. The NMA did not assess efficacy outcomes concerning weight-related comorbidities or HRQoL. The only harms outcome assessed was the proportion of patients discontinuing treatment due to AEs. Overall AEs and SAEs were not evaluated.

Random-effects Bayesian NMAs with Markov chain Monte Carlo methods. Publication bias was assessed by examining the funnel plot and using the Egger regression test. However, clinical heterogeneity was not assessed among RCTs for the different direct or pairwise comparisons.

Efficacy Results

A total of 28 relevant RCTs were included. There were 16 RCTs of orlistat versus placebo, 2 RCTs of liraglutide 3 mg versus placebo, 4 RCTs of NB versus placebo, 3 RCTs of lorcaserin versus placebo, 2 RCTs of PT versus placebo, and 1 3-armed RCT comparing orlistat and liraglutide 3 mg with placebo. The mean weight of the study populations across the RCTs was between 95.3 kg and 115.8 kg but was similar for treatment groups in each study. Liraglutide 3 mg and the relevant comparators (orlistat) were administered at the Health Canada–approved dosages, each as an adjunct to low-calorie diet and physical activity co-intervention. Outcome from comparisons between liraglutide 3 mg versus orlistat are reported as follows.

Achieving Weight Loss of 5% or More

The results of a direct meta-analysis of data from 3 studies (N = 4,563) showed that patients treated with liraglutide 3 mg had greater odds of achieving at least 5% weight loss compared with orlistat, as indicated by an OR of 3.66 (95% credible interval [CrI], 1.79 to 7.46). The NMA also found that treatment with liraglutide 3 mg increases the odds of a patient losing at least 5% of body weight compared with orlistat (OR = 2.06; 95% CrI, 1.51 to 2.96).

Achieving Weight Loss of 10% or More

The results of a direct meta-analysis of data from 3 studies (N = 4,563) showed that patients treated with liraglutide 3 mg had greater odds of achieving at least 10% weight loss compared with orlistat, as indicated by an OR of 3.87 (95% CrI, 1.65 to 9.04). The results of the NMA comparison between liraglutide 3 mg and orlistat were consistent with this finding, with an OR of 2.07 (95% CrI, 1.48 to 3.20) in favour of liraglutide 3 mg.

Mean Weight Loss in Excess of Placebo

Treatment with liraglutide 3 mg resulted in greater incremental weight loss (in kilograms) over placebo compared with orlistat, as indicated by results of both the direct meta-analysis (weighted mean difference =  − 3.99; 95% CrI, −5.18 to −2.62) and the NMA (weighted mean difference = –2.68; 95% CrI, –3.35 to –1.83).

Discontinuation of Therapy Due To Adverse Events

The direct meta-analysis found no difference between liraglutide 3 mg and orlistat regarding discontinuation of therapy as due to AEs (weighted mean difference = 3.50; 95% CrI, 0.70 to 17.49). However, findings from the NMA indicate that treatment discontinuation due to adverse AEs occurs more with liraglutide 3 mg than with orlistat (weighted mean difference = 1.6; 95% CrI, 1.10 to 2.40).

Harms Results

The NMA reported on therapy discontinuation due to AEs. The direct meta-analysis found no difference between liraglutide 3 mg and orlistat regarding discontinuation of therapy due to AEs (weighted mean difference = 3.50; 95% CrI, 0.70 to 17.49). However, findings from the NMA indicate that treatment discontinuation due to adverse AEs occurred more with liraglutide 3 mg than with orlistat (weighted mean difference = 1.6; 95% CrI, 1.10 to 2.40).

Critical Appraisal

The following limitations have potential to impact the findings of the NMA:

• The primary RCTs were low in quality and the high attrition in all the RCTs undermined confidence in the outcomes due to a high risk of attrition bias.

• There were variations in the dietary component of the co-intervention between orlistat and liraglutide that may have contributed to the statistical heterogeneity and distorted the outcomes.

• A closed loop could be formed by orlistat, liraglutide, and placebo. Therefore, consistency throughout the network could not be assessed. Also, the portion of the loop connecting orlistat and liraglutide 3 mg was contributed by a single phase II RCT¹⁸ with 4 different liraglutide doses, including liraglutide 3 mg once daily, in which the placebo-controlled portion was double-blind whereas the orlistat comparator was open label. Thus, there is a low-quality issue that limits the evidence.

• A run-in placebo treatment phase was more common in the orlistat versus placebo RCTs than in the RCTs of other comparators versus placebo. As a run-in phase may enrich the trial population with patients more likely to adhere to treatment, there may have been a bias in the results for any of the outcomes in favour of orlistat relative to the other comparators.

• The encouragement of patients to continue with study assessment or return for the end-of-study assessment following treatment discontinuation was more common in the non-orlistat RCTs than in the orlistat RCTs. The potential direction of bias from this source of heterogeneity is unclear.

• The GRADE quality of evidence assessment for the NMA’s comparisons was low for NB versus liraglutide 3 mg.

• An assessment of clinical heterogeneity or the appropriateness of pooling trial results was not done. Also, an evaluation of heterogeneity was not feasible for NB versus orlistat or liraglutide 3 mg in the absence of any direct comparisons.

Other Relevant Evidence

Description of Studies

Two additional studies of comparative RCTs evaluating liraglutide 3 mg with intensive lifestyle intervention were identified as relevant to this review. Both trials provide additional evidence of liraglutide 3 mg compared with intensive lifestyle modification, which was identified as a comparator of interest in the CADTH review protocol.

Study 4274 was a prospective, multi-centre, double-blind, placebo-controlled, phase IIIb randomized trial to evaluate the health benefits of combining intensive behavioural therapy (IBT) with liraglutide 3 mg in adult patients living with obesity without diabetes, and Study NCT02911818 was a single-site, open-label, parallel-group randomized trial to assess whether the addition of liraglutide 3 mg to an IBT intervention would increase weight loss compared to IBT alone in adult patients living with obesity.

Across the 2 studies, a total of 432 patients were randomly assigned to be treated with liraglutide 3 mg or placebo. The participants were predominantly female (79% to 84%) and the mean age of patients was between 45 years and 49 years. The patients in Study 4274 were mostly White (79% and 82% for the liraglutide 3 mg and placebo groups, respectively), whereas for Study NCT02911818, 54.0% self-identified as non-Hispanic White, 44.7% as Black, and 6.7% as Hispanic. Overall, baseline characteristics appeared similar for the treatment groups of each study. However, in Study 4274, the liraglutide 3 mg group had a greater proportion of patients with a BMI of 40 kg/m² or more than the placebo group (40.8% versus 30.7%).

In both trials, liraglutide 3 mg was used as an adjunct to the US Centers for Medicare and Medicaid Services (CMS)-IBT (CMS-IBT) at the approved dose and following recommended titration strategies previously described. The CMS-IBT consisted of weekly, 15-minute, in-person lifestyle counselling visits the first month, followed by visits every other week the next 5 months, approximating 14 contacts to 15 contacts over 6 months together with increased physical activity and specific daily energy intake based on patients’ weight. In Study 4274, the comparator was placebo plus CMS-IBT, whereas in Study NCT02911818, the comparator was CMS-IBT alone.

Study 4274 had 2 co-primary end points: change in body weight (%) from baseline to week 56, and the proportion of patients losing at least 5% of baseline body weight at week 56. The primary end point of Study NCT02911818 was the mean percentage reduction in baseline body weight at week 52.

Efficacy Results

Primary analysis results from Study 4274 showed that liraglutide 3 mg was superior to placebo with respect to relative mean change (%) in body weight and proportion of patients reaching a clinically relevant weight loss of 5% at week 56, shown as follows:

• The mean change (%) baseline body weight at week 56 was −7.46% versus −4.01% for liraglutide 3 mg versus placebo, respectively, with an estimated treatment difference of −3.45% (95% CI, −5.31 to −1.59) in favour of the liraglutide 3 mg group (P = 0.0003).

• The probability for achieving 5% or more loss of baseline body weight at week 56 was 60.6% versus 32.9% for liraglutide 3 mg versus placebo, respectively, with an OR of 2.51 (95% CI, 1.53 to 4.14; P = 0.0003).

The findings were consistent for the secondary outcome of the proportion of patients losing at least 10% of baseline body weight at week 56.

In Study NCT02911818, liraglutide 3 mg resulted in a significant reduction from baseline in body weight (%) at week 52, where the mean reduction in the IBT-liraglutide 3 mg group was 11.5% (standard error of the mean [SEM]) =  ± 1.3), 6.1% (SEM =  ± 1.3) in the IBT-alone group, and 11.8% (SEM =  ± 1.3) in the multi-component group in favour of the IBT-liraglutide 3 mg group (P = 0.005) and the multi-component group (P = 0.003). Further, the mean reduction from baseline in BMI at week 52 in the IBT-liraglutide 3 mg group was 4.3 (SEM =  ± 0.5), 2.3 (SEM =  ± 0.5) in the IBT-alone group, and 4.6 (SEM =  ± 0.5) in the multi-component group. Compared to the IBT-alone group, the results favoured the IBT-liraglutide 3 mg group (P = 0.003) and the multi-component group (P = 0.001).

Overall, after a year of treatment, results from both studies suggest that the patients in the liraglutide-based treatment groups had better outcomes than those in the placebo plus IBT group or IBT-alone group regarding percent mean weight loss from baseline and the proportion of patients losing 5% or more of baseline body weight. Results from Study 4274 also showed that significantly more individuals on liraglutide 3 mg than placebo achieved more than 10% weight loss and more than 15% weight loss relative to baseline body weight. The clinical expert consulted for this review noted that there are not many intensive lifestyle intervention programs available in Canada and that the number of patients accessing such programs is likely limited.

Harms Results

The overall AE rate for liraglutide 3 mg plus IBT was 96% in Study 4274 and 90% in NCT02911818. The corresponding rates for placebo plus IBT and IBT alone was 89% and 60%, respectively. SAE rates for the liraglutide 3 mg group and the placebo group were 4.2% and 1.4%, respectively, in Study 4274. In Study NCT02911818, the SAE rate was 4.0% in the IBT-alone group, whereas there were no SAEs reported in the IBT plus liraglutide group. No deaths were reported in either study. In Study 4274, premature discontinuation of treatment due to AEs was 9% and 4% in the liraglutide 3 mg and placebo groups, respectively. Study NCT02911818 did not report treatment discontinuations due to AEs. Similar to the AEs reported in the studies in the systematic review, GI symptoms were the most frequent AEs overall, and they were also the most common notable harms.

Critical Appraisal

Both Study 4274 and Study NCT02911818 used appropriate randomization methods — a web-based randomization system (Interactive Voice/Web Response System [IV/WRS]) to allocate patients to their treatment group. Study 4274 was a double-blind trial with blinding maintained over the entire treatment period, whereas Study NCT02911818 was an open-label study with an attempt at concealment or blinding. However, even for Study 4274, which implemented blinding, the high and unbalanced attrition rate driven by AEs in the liraglutide 3 mg group and less effective therapy within the placebo group may have resulted in unblinding for some patients. Thus, there was a risk of altered response and assessor bias caused by the study design of Study NCT02911818 and easily distinguishable effects of the intervention and control groups. Further, Study NCT02911818 was a single-site, open-label study, with a relatively small sample size (50 patients in each group). Therefore, evidence provided by this study is limited and unlikely to have a confirmatory value.

The patients enrolled in both Study 4274 and Study NCT02911818 were predominantly women, and for Study 4274, also mostly White. Furthermore, both studies excluded patients with a history of bariatric surgery and those with recent use of medications that cause weight loss, who make up a clinically relevant proportion of individuals seeking pharmacotherapy for management of obesity in real life. As previously discussed, these issues indicate that the study populations in these 2 trials may not be representative of Canadian patients living with obesity who have clinical needs for drug therapy to manage a chronic problem. Thus, there is uncertainty about the generalizability of the conclusion of the studies regarding the effectiveness and safety of liraglutide 3 mg-based interventions in the diverse population of patients in Canada who are overweight or living with obesity.

Conclusions

Overall, the results of 4 RCTs demonstrated that once-daily treatment with liraglutide 3 mg in addition to a background regimen of diet and exercise resulted in statistically significant and clinically meaningful reductions in body weight compared with placebo (in addition to diet and exercise). Further, results demonstrated that liraglutide 3 mg increased the likelihood of achieving 5% or less or more than 10% reduction in body weight. These results were consistently observed in a variety of patient populations, including those without diabetes, those with prediabetes, those with T2DM, and those with OSA. Results of subgroup analyses based on baseline BMI and prediabetes status at screening were consistent with these results. In patients with prediabetes, liraglutide 3 mg also showed superiority over placebo in delaying progression to T2DM. HRQoL was a secondary outcome in each of the trials, but results were inconsistent across measures and studies. Other outcomes of interest to the CADTH review included change in BMI, HRQoL, and glycemic control, and change in other medications. However, none of these outcomes was controlled for multiplicity; therefore, results must be considered with regard to type I error.

TEAEs occurred more frequently with liraglutide 3 mg than placebo. GI disorders were the most common AEs with liraglutide 3 mg and are generally manageable with a dose escalation strategy, as used in the 4 studies and as recommended in the product monograph.

Key limitations associated with the evidence reviewed are that patients with comorbidities and a BMI between 27 kg/m² and less than 30 kg/m² appeared to be underrepresented, and that there was a lack of comparative evidence.

Although indirect evidence from 1 NMA may suggest that patients treated with liraglutide 3 mg had greater odds of achieving clinically relevant weight loss (5% to 10%) compared with orlistat, confidence in these results is limited by significant heterogeneity and high attrition rates across all the included primary studies and by significant limitations involving the quality of the primary studies and methodological rigour.

Introduction

Disease Background

WHO defines overweight and obesity as abnormal or excessive fat accumulation that presents a risk to health.¹ The 2 are distinguished from each other based on BMI, where a BMI over 25 kg/m² is considered overweight, and over 30 kg/m² is obese.¹ Determinants associated with obesity have been identified as physical activity, diet, socioeconomic status (such as income and education), ethnicity, immigration, and environmental factors.² WHO has reported that the problem of overweight and obesity has grown to epidemic proportions,¹ and the Canadian Health Measures Survey found that in 2019, overweight and obesity rates among adults aged 18 years to 79 years in Canada were 35.5% and 24.3%, respectively.³ The CTFPHC has reported that more than 2-thirds of Canadian men (67%) and more than half of Canadian women (54%) are overweight or living with obesity.⁴

Obesity is associated with an increased risk of a wide range of illnesses and long-term conditions, including type 2 diabetes, hypertension, gallstones, gastroesophageal reflux disease, and cancer, as well as psychological and psychiatric morbidities.⁵ For instance, a positive association has been established between BMI and some psychiatric disorders, including depression and anxiety, and it has been reported that individuals living with obesity were 1.5 times more likely than individuals of normal weight to report such disorders.¹⁹ Also, a relationship has been established between obesity and sleep disturbances such as insomnia and OSA. Individuals living with obesity are also more likely to experience insomnia.²⁰ According to the clinical expert consulted for this review, other weight-related comorbidities are extensive and include metabolic complications such as prediabetes, T2DM, and dyslipidemia, as well as osteoarthritis, certain types of cancers (e.g., endometrial, colon, renal, esophageal, and breast cancer in women), infertility, gallbladder disease, nonalcoholic fatty liver disease, and gout. Therefore, the goal of therapy in managing excess weight goes beyond weight loss to also target a reduction or improvement in weight-related comorbidities, and improve longevity and the patient’s quality of life.

WHO has reported that more than 4 million people die each year as a result of being overweight or living with obesity, according to the global burden of disease.¹ It is estimated that median survival is reduced by 2 years to 4 years for those with a BMI of 30 kg/m² to 35 kg/m² and by 8 years to 10 years for those with a BMI of 40 kg/m² to 50 kg/m².⁵ Also, weight loss of 10 kg may lead to a reduction in total cholesterol of 0.25 mmol/L and DBP of 4 mm Hg.⁶ Thus, the treatment of overweight and obesity has health benefits, with weight loss of 5 kg to 10 kg reported to reduce the long-term risk of diseases associated with obesity.⁵

Standards of Therapy

The clinical expert consulted for this review stated that diagnosis of all weight-related comorbidities can be performed within the scope of any primary care provider without any foreseeable challenges with diagnosing overweight, obesity, or comorbidities. Thus, community care facilities and hospital outpatient clinics are appropriate settings for diagnosis and treatment. This is in consonance with recommendations of CTFPHC that interventions could be offered in primary care settings or settings where primary care practitioners may refer patients, such as credible commercial or community programs.⁴

The treatment of choice to manage overweight and obesity is a multi-component interventions approach with respect to provider discipline, length, and format.^4,7 A critical component of this strategy is to support patients to achieve sustainable weight loss through the modification of diet, physical activity, and behaviour.⁷ The CTFPHC has stated that behavioural interventions focusing on diet, increasing exercise, making lifestyle changes, or any combination of these are the preferred option, as the benefit-to-harm ratio appears more favourable than for pharmacologic interventions.⁴ It has been estimated that a diet that provides a deficit of 600 kcal/day may be expected to produce weight loss of 5 kg over 1 year, whereas exercise and behavioural therapy may provide weight loss of approximately 2 kg and 8 kg, respectively, when added to a calorie-restricted diet.⁶

However, while initially effective for many individuals, diet or behaviour modification alone is often difficult to sustain and many individuals regain weight upon discontinuation.⁸ Therefore, drug therapy for chronic weight management as an adjunct to lifestyle intervention may help individuals to achieve and sustain clinically relevant weight loss.⁸ According to the Obesity Canada 2020 guidelines, drug therapy for weight management is indicated only for those with a BMI of 30 kg/m² or more, or those with a BMI of 27 kg/m² or more with at least 1 comorbidity who have failed a previous lifestyle attempt at weight loss.⁹ The UK’s National Institute for Health and Care Excellence advises that pharmacologic treatment be considered only after dietary, exercise, and behavioural approaches have been started and evaluated, and drug treatment should be considered for people who have not reached their target weight loss or have reached a plateau on dietary, activity, and behavioural changes.⁷

Currently, 3 drugs (liraglutide 3 mg, NB fixed-dose combination, and orlistat) have been approved in Canada for chronic weight management in adult patients with an initial BMI of 30 kg/m² or greater or 27 kg/m² or greater in the presence of at least 1 weight-related comorbidity (e.g., hypertension, type 2 diabetes, or dyslipidemia) and who have failed a previous weight management intervention. The clinical expert consulted for this review stated that it is likely that the most used off-label medication for weight loss currently is weekly injectable semaglutide, a GLP-1 receptor agonist that is indicated in Canada for the treatment of T2DM. The expert noted that there is some evidence for the use of metformin to prevent weight gain associated with the use of antipsychotic medications, and topiramate may be chosen for patients with migraines and seizures who need drug therapy for weight loss since the drug is indicated for those conditions.

Obesity Canada guidelines recommend that bariatric surgery may be considered for people with a BMI of 40 kg/m² or more or a BMI of 35 kg/m² or more with at least 1 obesity-related disease. The decision regarding the type of surgery should be made in collaboration with a multidisciplinary team, balancing the patient’s expectations, medical condition, and expected benefits and risks of the surgery.⁹

Concerning goals of therapy, the clinical expert consulted for this review stated that most physicians would agree that a 5% to 10% total body weight loss is clinically meaningful because it can ameliorate weight-related comorbidities such as type 2 diabetes and osteoarthritis. The clinical expert noted that improved weight or BMI alone is the least clinically meaningful outcome; therefore, treatments should target a reduction or improvement in weight-related comorbidities and improve longevity. However, weight stability for some patients may be a reasonable goal, especially in those with a rapid upwards weight trajectory (e.g., for some patients with weight regain after bariatric surgery, stabilizing weight is a success).

Drug

Liraglutide is a human GLP-1 analogue, which acts as a GLP-1 receptor agonist to regulate appetite by increasing feelings of fullness and satiety, while lowering feelings of hunger and prospective food consumption.¹⁰

Liraglutide 3 mg (Saxenda) is available as a solution for subcutaneous injection at a strength of 6 mg/mL in a pre-filled, multi-dose pen that delivers doses of 0.6 mg, 1.2 mg, 1.8 mg, 2.4 mg, or 3 mg.¹⁰ The initial dose of liraglutide 3 mg is 0.6 mg daily for 1 week, after which the daily dose is escalated at weekly increments of 0.6 mg to reach the recommended daily maintenance dose of 3 mg per day over 4 weeks.¹⁰

Liraglutide 3 mg received a Health Canada Notice of Compliance on February 26, 2015, with an initial indication to be used as an adjunct to a reduced-calorie diet and increased physical activity for chronic weight management in adult patients with an initial BMI of 30 kg/m² or greater (obese), or 27 kg/m² or greater (overweight) in the presence of at least 1 weight-related comorbidity (e.g., hypertension, type 2 diabetes, or dyslipidemia) and who have failed a previous weight management intervention.¹⁰ The sponsor has requested that liraglutide 3 mg be listed as an adjunct to a reduced-calorie diet and increased physical activity for chronic weight management in adult patients who have been diagnosed with obesity (BMI ≥ 30 kg/m²) and prediabetes, or overweight (BMI ≥ 27 kg/m² and < 30 kg/m²) with 1 or more weight-related comorbidity and prediabetes.²¹

On February 25, 2021, Health Canada–approved the indication for liraglutide 3 mg for chronic weight management in adolescent patients. Thus, in addition to the previously stated indication, liraglutide 3 mg is indicated as an adjunct to a reduced-calorie diet and increased physical activity for chronic weight management in pediatric patients aged 12 years to less than 18 years with the following¹⁰:

• an inadequate response to a reduced-calorie diet and increased physical activity alone

• a body weight above 60 kg (132 lbs.)

• an initial BMI corresponding to 30 kg/m² or more for adults (obesity).

However, the sponsor’s requested reimbursement criteria for this review remains the same as previously stated with no changes to the submission. Thus, the requested reimbursement is focused only on a subset of patients with prediabetes and not the entire population for which the drug has received Health Canada approval.

Two other drugs approved for use in Canada with similar indications are orlistat and the NB combination tablet. Table 4 summarizes the characteristics of key treatments currently available for chronic weight management.

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups.

About the Patient Groups and Information Gathered

Three Canadian patient organizations — Obesity Canada, Diabetes Canada, and the GI Society — provided input for this submission. Obesity Canada is Canada’s leading obesity registered charity association for professionals and patients, contributing research, education, and advocacy for Canadians living with obesity. Diabetes Canada is a national health charity that represents Canadians living with, or at risk of, diabetes and focuses on research and policy initiatives for diabetes prevention, care, and cure at the population level. The GI Society is a national charity committed to improving the lives of people with GI and liver conditions through research, advocating for increased access to health care, and promoting GI and liver health.

The groups submitted input gathered through a variety of sources, including surveys, interviews, and published studies. Obesity Canada conducted patient interviews and an online survey, where 5 of the individuals interviewed and 60% of the survey respondents had used liraglutide 3 mg for obesity management. Diabetes Canada submitted patient input using data from 2 online surveys conducted in July and August 2020 and in December 2020 and January 2021. They declared that the survey in July and August 2020 was jointly created by themselves, a research and advocacy organization for type 1 diabetes (JDRF), and an advocacy organization for individuals living with type 1 diabetes (Type 1 Together). They also declared that they consulted with Obesity Canada regarding the creation of the December 2020 and January 2021 patient input survey but drafted it independently. The GI Society used data from patient and patient caregiver interviews, the results of published studies, and a survey conducted from October 6, 2020, to January 10, 2021, among individuals living with obesity. All 3 patient input groups submitted conflict of interest disclosures, which can be found on the CADTH website.

Disease Experience

Patient group input identified obesity as a chronic, multifactorial, relapsing disease characterized by excessive or abnormal body fat that can impair health. Due to the influence that adipose tissue has on the central regulation of energy homeostasis, excessive adiposity can lead to the development of a variety of health complications such as diabetes, high blood pressure, heart disease, sleep apnea, mental health problems, and osteoarthritis.

Patient group input reported that an estimated 80% to 90% of people with type 2 diabetes live with overweight or obesity. Prediabetes is a precursor to type 2 diabetes, and for those living with both prediabetes and overweight or obesity, various weight management approaches can help reduce the likelihood of progression to diabetes.

Patient groups reported that obesity not only increases the potential for the development of further disease(s), but it leads to inequities in access to employment, health care, and education due to the strong stigma associated with it. Many individuals have an incorrect perception that obesity is a self-inflicted disease, demonstrating a societal misunderstanding of obesity. The GI Society reported that 72% of their survey respondents experienced social stigma as a result of living with obesity, with many reporting that they avoid getting medical care because they feel as though their physician shames them because of their weight:

“I don’t go to the doctor as often as I should because I feel like a failure and that all my medical issues are caused by my obesity.”

“I’ve received the most shame about my weight from doctors to the point I’m scared to go. They should help, not shame.”

Individuals living with obesity also report frustration with the impact that the chronic and often misunderstood disease has on their overall quality of life:

“It’s frustrating to know that no matter what you do, weight management will always be an issue. Eating right is not always good enough – and the older I get the more difficult it is to keep the weight off.”

“It’s like I’m in my very own prison that I have a hard time fitting into.”

Not only does obesity affect the individual, but it also affects families and society as well. Despite its widespread impact, patient groups that provided input expressed that few provincial or territorial governments in Canada have implemented health promotion efforts or treatment programs for children and adults living with obesity, forcing individuals to self-manage their disease.

Experience With Treatment

Although there are treatments currently available in Canada, patient groups reported that there is a lack of access to these treatments:

“It is so frustrating and demoralizing that the things that work for me are unattainable, I cannot afford the medications or to see a therapist regularly and the wait time for surgery is several years. I am left to try and manage on my own and it is just not possible.”

Currently, most Canadians living with obesity report using diet and exercise, medications, or bariatric surgery to combat the disease. Patient groups report that diet and exercise do not address biologic, psychological, or environmental factors contributing to the disease and instead place the blame on the individual. Additionally, due to stigma and discrimination, many individuals do not turn to licensed health care professionals for obesity management and instead rely on commercial weight-loss programs that are unregulated and untested. Many who diet and exercise have difficulty sustaining their efforts or finding a program that suits their needs, which can lead to depression, hopelessness, and further weight gain:

“I have tried countless diets and participate in a number of activities that support physical health. They are not effective at lowering weight and keeping weight at a healthy level for a long period of time. I have ended up gaining back the lost weight and even gaining more. It was very frustrating and took away from my quality of life.”

In the survey conducted by Diabetes Canada, only 1 of 12 respondents with type 2 diabetes or prediabetes reported taking liraglutide 3 mg (Saxenda) for overweight or obesity, or any medication, for that matter. The individual did report that weight loss was much improved on the medication. The lack of reported use of medications for overweight or obesity may be because there are few medication options available in Canada, and the options do not have full public or private coverage. Approved medications include NB (Contrave), liraglutide (Saxenda), and orlistat (Xenical), all of which have undesirable side effects. The side effects of the medications include nausea, diarrhea, constipation, oily stools, bowel urgency, low blood sugar, headaches, and dizziness. Despite the availability of these drugs, the GI Society reported many concerns obtaining and paying for the prescriptions of these drugs:

“Most of us who could benefit from the medication do not have coverage to use the medication that could actually be beneficial.”

“I have tried going on weight loss medication but unfortunately it has never gone past the discussion point. I have been eagerly looking forward to trying any sort of medication for my weight loss.”

“My doctor refused to try any weight loss drugs for me.”

The patient groups reported that individuals who have experience with liraglutide have reported weight loss while on the drug. Specifically, Obesity Canada reported that the respondents in their survey who had experience with Saxenda reported an average of 11% weight loss. Among reports of weight loss, patient groups also noted that although the participants did find the side effects undesirable, they were manageable and not significant enough to deter them from taking the drug.

“I have been on Saxenda for 8 months and lost 40 lbs, sustained loss for the 9th month, and will continue to use it.”

Although bariatric surgery is currently considered the gold standard for treating obesity, it is often used as a last resort due to the potential serious side effects. According to the GI Society, 33% of respondents would not consider the surgery. For individuals in Canada who would consider the surgery, wait lists and the cost of the service are significant barriers. The surgery can lead to complications, hospital readmission, severe nutritional deficiencies, and GI symptoms.

Improved Outcomes

Currently, no provincial drug formularies include anti-obesity medication. As well, treatments that are currently available are not effective in the long-term. Due to this, patient groups report that many patients would like a treatment that is effective in the long-term, is affordable, and has no or minimal side effects:

“Reduce the costs if possible to make them more affordable.”

“No side effects that can add to my stress.”

Patient groups hope that liraglutide 3 mg may help people to better manage their weight, potentially delaying or preventing the development of comorbidities, such as the progression of prediabetes to type 2 diabetes. Furthermore, when asked about outcomes to consider, it was reported that patients focused less on improved weight than improved related comorbidities (e.g., diabetes, hypertension, and sleep apnea) as well as outcomes related to everyday life such as productivity, energy levels, sleep, activity, and mental health:

“I need to lose weight so I can have the energy and mobility to play with my kids/grandkids”

“I am so preoccupied with worrying about my weight that my productivity and mental health suffer, if I can lose some weight, everything else will get better.”

“Steady weight loss would help reduce joint pain so I can return to work, steady blood glucose levels would reduce the stress of always starving myself to keep sugars under control without insulin.”

Patient organizations noted that even when individuals lose a significant amount of weight, many of them gain it back within 5 years. This, in combination with a lack of effective treatments, has led to a 455% increase in severe obesity within Canada over the past 3 decades. Given the impact of obesity on multiple comorbidities and an individuals’ quality of life, patient groups hope for an effective, affordable, and accessible treatment for individuals living with obesity that will improve lives and avoid further direct health care costs.

Clinician Input

All CADTH review teams include at least 1 clinical specialist with expertise regarding the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 1 clinical specialist with expertise in chronic weight management.

Input From Clinical Experts Consulted by CADTH

Unmet Needs

The clinical expert noted that drug therapies are most effective when combined with lifestyle and behavioural changes. However, there is a lack of funded lifestyle programs in Canada. Although 3 medications have been approved for the treatment of obesity in Canada, the minimal coverage for these drugs outside private insurance plans limits access for many patients. Therefore, it is up to the patient and health care provider to decide which lifestyle changes will be implemented along with pharmacotherapy. Access to bariatric surgery for those meeting the criteria is limited in some Canadian provinces.

The clinical expert noted that side effects limit the use of all 3 pharmacotherapies, with some patients needing to stop the medication completely due to side effects. More typically, however, the dose would be limited by side effects. For example, a patient may be able to tolerate a half dose of NB or take only 1.2 mg or 1.8 mg daily of liraglutide, and still see some weight-lowering efficacy while avoiding side effects.

The clinical expert acknowledged that the weight loss seen with medications for obesity will often fall within a range of 5% to 10%, which is considered adequate to ameliorate weight-related comorbidities such as type 2 diabetes and osteoarthritis. However, in an ideal world, we would have pharmacotherapy that promotes larger amounts of weight loss.

Place in Therapy

The clinical expert consulted for this review stated that liraglutide is used both before and following bariatric surgery, and it is a common treatment strategy to prescribe liraglutide 3 mg in patients experiencing weight regain after bariatric surgery. The expert indicated that in accordance with the Obesity Canada 2020 guidelines, pharmacotherapy can realistically be used anywhere in a patient’s weight management journey. It was explained further that though it is reasonable for a patient to try lifestyle intervention first before starting liraglutide 3 mg, that approach is not necessary, particularly for patients at higher BMIs. The expert added that the long-term data for behavioural and lifestyle interventions alone is very discouraging and suggests that weight is typically regained within a relatively short period because of compensatory mechanisms in the brain that promote positive caloric intake by increasing hunger and ultimately causing weight gain. However, the clinical expert stated that liraglutide 3 mg for chronic weight management should always be combined with at least some behavioural changes or lifestyle interventions. Since the mechanism of action is very different from NB and orlistat, some physicians are prescribing combinations of these agents for weight loss. There is minimal evidence to support this; however, combination therapy is a well-established treatment paradigm for treatment of type 2 diabetes and hypertension.

On the question of whether liraglutide addresses the underlying process or manages symptoms of the condition, the clinical expert stated that liraglutide slows postprandial emptying of the stomach, acts centrally on the brain to increase satiety, reduces hunger, and (when blood glucose is elevated) stimulates insulin secretion and reduces glucagon secretion from the pancreas. Patients typically report being less interested in food, being able to reduce portion sizes more easily at meals, and having an overall lower appetite. Therefore, liraglutide targets 1 of the many underlying hormonal mechanisms for obesity and is not just managing symptoms. The clinical expert thought that the approval of liraglutide 3 mg (along with the NB combo) in Canada has started to shift the treatment paradigm for obesity management, in that the availability of pharmacotherapy options has appropriately given more legitimacy to the idea that obesity should be managed as a chronic disease.

Patient Population

The clinical expert consulted for this review stated that currently, there in no way to predict which patients will lose the most weight with liraglutide. The clinical expert consulted for this review stated that qualifying patients can be identified using the traditional definition of overweight and obesity based on BMI category. Patients most in need of intervention for weight loss are those who have the highest burden of weight-related comorbidities, which can be diagnosed using condition-dependent applicable diagnosis techniques, including laboratory tests, ultrasound, X-rays, and other clinical examination strategies. The expert indicated that patients with an EOSS score ranging from EOSS 1 to EOSS 3 could most typically benefit from a weight-loss intervention, noting that it is reasonable to intervene with liraglutide 3 mg at the EOSS 1 category to prevent future progression to the later stages of obesity. The clinical expert noted that patients with Class III obesity (i.e., BMI > 40) could be considered an important category for liraglutide 3 mg since they have the most excess weight. However, in such patients, a more intensive intervention such as bariatric surgery should also be considered or offered.

The clinical expert indicated that some patients have BMI below the traditional cut-off for overweight (i.e., < 27 kg/m²) but have metabolic complications from their weight, especially in ethnicities shown to have metabolic risk at lower BMIs. In such patients, it is reasonable to consider pharmacotherapy as second-line after lifestyle changes; however, that would be off-label use of the medication.

The clinical expert stated that at this time, the only way to know how a patient will respond to a medication is to try it. However, it was expected that patients who complain of a large amount of physical hunger limiting their weight-loss efforts tend to respond very well to the drug, since it promotes satiety. However, this was based on clinical observation and not on evidence.

Patients who would be unsuitable for treatment with liraglutide would be those with an absolute contraindication for the drug (e.g., planning pregnancy, pregnant, breastfeeding, personal or family history of medullary thyroid cancer or multiple endocrine neoplasia type 1). Patients with a history of pancreatitis should also usually not be prescribed the drug. Patients who are also unwilling to take an injectable therapy or adhere to any sort of behavioural intervention or changes are also not good candidates.

Assessing Response to Treatment

The expert indicated that most physicians would agree that a 5% to 10% total body weight loss is clinically meaningful and is typically felt to be associated with improved metabolic parameters. However, improved weight or BMI alone are the least clinically meaningful outcomes. Other meaningful responses to treatment include improvement in weight-related comorbidities (e.g., type 2 diabetes, prediabetes, hypertension, dyslipidemia, OSA), the prevention of progression of preclinical conditions (e.g., reduced progression from prediabetes to diabetes, from prehypertension to hypertension), improved quality of life, improved survival, reduced cardiovascular (CV) and renal events, and reduced osteoarthritis symptoms.

The clinical expert consulted for this review stated that, in general, weight-loss outcomes are assessed based on change in BMI and weight, whereas weight-related comorbidity outcomes are evaluated using change in parameters such as blood pressure, glycemic control, and lipid profile. Initial assessment takes place between 4 weeks to 6 weeks. Weight-loss response is assessed after 12 weeks to 16 weeks at the maximum dose (along with behavioural and lifestyle changes) to decide whether to continue the medication at that point. Based on experience, the expert stated that once the patient is successful on a stable dose of liraglutide 3 mg, assessments could be spaced at 3-month intervals, and eventually every 6 months.

Discontinuing Treatment

According to the clinical expert consulted for this review, treatment discontinuation decisions are influenced by patients’ preference not to rely on long-term medication for weight management, experiences of side effects (most commonly, GI side effects) with even the lowest dose of liraglutide, and rare but serious adverse effects such as pancreatitis, worsening in mood, or increase in anxiety. Where a patient decides not to use the drug for long-term weight management, a gradual tapering-off with close follow-up and of weight monitoring is a reasonable approach. Liraglutide would be stopped during the post-operative period for a patient who goes for a bariatric surgery. However, it could be reintroduced 18 months to 24 months after surgery if the patient experienced weight regain.

Prescribing Conditions

The clinical expert consulted for this review stated that community care facilities (either specialist or primary care) and hospital outpatient clinics (either bariatric clinics or specialty clinics) are appropriate settings for treatment with liraglutide. The expert pointed out that the inability for long-term patients to follow-up at some hospital-based specialty clinics (e.g., tertiary care centres) may be a barrier to liraglutide 3 mg treatment. Therefore, discharge back to primary care for ongoing treatment with liraglutide 3 mg and management of obesity may be necessary.

According to the clinical expert, primary care providers can diagnose obesity and overweight, safely prescribe liraglutide 3 mg, and monitor their patients over time. It was stated that some family physicians and a variety of specialists, including endocrinologists, general internists, and nephrologists, offer weight management services.

Clinician Group Input

No clinician group input was received for this review.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may impact their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 5.

Clinical Evidence

The clinical evidence included in the review of liraglutide 3 mg is presented in 3 sections. The first section, the systematic review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those studies that were selected according to an a priori protocol. The second section includes indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the systematic review.

Systematic Review (Pivotal and Protocol Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of liraglutide 6 mg/mL for subcutaneous injection as an adjunct to a reduced-calorie diet and increased physical activity for chronic weight management in adults

Methods

Studies selected for inclusion in the systematic review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria in Table 6. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

The literature search was performed by an information specialist using a peer-reviewed search strategy. The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS Peer Review of Electronic Search Strategies checklist.²⁴

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946–) via Ovid and Embase (1974–) via Ovid. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concepts were Saxenda (liraglutide) and weight management. Clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

Search filters were applied to limit retrieval to randomized controlled trials (RCTs) or controlled clinical trials. Retrieval was not limited by publication date or by language. Where possible, retrieval was limited to the human population. Conference abstracts were excluded from the search results. See Appendix 1 for the detailed search strategies.

The initial search was completed on January 27, 2021. Regular alerts updated the search until the meeting of the CADTH Canadian Drug Expert Committee on May 19, 2021.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist²⁵ Included in this search were the websites of regulatory agencies (the FDA and European Medicines Agency). Google was used to search for additional internet-based materials. See Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented through contacts with appropriate experts. In addition, the sponsor of the drug was contacted for information regarding unpublished studies.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

A focused literature search for NMAs dealing with weight management was run-in MEDLINE All (1946–) on January 27, 2021. The search was limited to selected subject headings and keywords appearing in the title only. No other limits were applied.

Findings From the Literature

A total of 4 studies were identified from the literature for inclusion in the systematic review (Figure 1). The included studies are summarized in Table 7 through Table 10. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Description of Studies

A total of 4 phase III RCTs (i.e., Study 1839, Study 1922, Study 1923, and Study 3970) met the inclusion criteria of this CADTH systematic review. The primary objectives of the sponsor-submitted pivotal study (i.e., Study 1839) were to establish the efficacy of liraglutide 3 mg compared with placebo in inducing and maintaining weight loss over 56 weeks in adults without diabetes, and to investigate the long-term efficacy of liraglutide 3 mg in delaying the onset of type 2 diabetes in patients living with obesity with prediabetes and in patients who are overweight with prediabetes and treated or untreated comorbidities such as dyslipidemia and/or hypertension. Thus, the trial had 2 phases: a main phase with primary analysis based on results after 56 weeks of treatment (N = 3,731) and a 104-week extension phase (N = 2,254) that included only patients who had prediabetes at screening. Study 1839 was a global, randomized, parallel-group, multi-centre, double-blind trial conducted at 191 sites in 27 countries in Africa, Asia, Australia, Europe, North America, and South America. Patients were randomized to receive treatment with liraglutide 3 mg or matched placebo. After a 2-week screening period, patients were randomly assigned in a 2:1 ratio to either the liraglutide or placebo treatment groups using a telephone or web-based system. The randomization was stratified according to prediabetes status and baseline BMI (BMI ≥ 30 kg/m² or < 30 kg/m²) at screening in the main phase and by BMI alone in the extension phase. Patients classified as not having prediabetes at screening were treated with either liraglutide 3 mg or placebo for 56 weeks whereas patients classified as having prediabetes at screening received either liraglutide 3 mg or placebo for 160 weeks (56 weeks during the main trial plus 104 weeks for the extension study). Following 56 weeks of treatment, patients without prediabetes at screening were re-randomized in a 1:1 manner to either continue treatment with liraglutide 3 mg (liraglutide/liraglutide group), switch to placebo (liraglutide/placebo group) or continue with placebo (placebo group) for 12 weeks. However, this 12-week period was not considered part of the main study phase; rather, it was meant to assess the effects of drug cessation on weight control, possible withdrawal, and rebound effects. The non-drug follow-up period was 2 weeks for the main phase and 12 weeks for the extension phase. Further details about the main and extension phases of Study 1839 have been summarized in Table 7.

Study 1922 (N = 846) was a phase IIIa, double-blind, placebo-controlled, 3-armed, parallel-group, multi-centre, multinational trial to investigate the efficacy of liraglutide compared to placebo after 56 weeks of treatment in inducing and maintaining weight loss in patients with type 2 diabetes who were overweight or living with obesity. The trial was conducted at 126 sites in 9 countries, including 6 European countries, Israel, South Africa, and the US. In addition to the liraglutide 3 mg and placebo arms, there was a third treatment arm for liraglutide 1.8 mg, which was included upon request from regulatory agencies. The patients were randomized in a 2:1:1 manner using a centralized IV/WRS to receive 3 mg of liraglutide, 1.8 mg of liraglutide, or placebo as an add-on to their background diabetes treatment. Randomization was stratified according to background treatment. The results of the liraglutide 1.8 mg dose will not be presented in the current review as this is not an approved dose for the indication under review. Other aspects of the design of Study 1922 were similar to the main phase of Study 1839 as previously described. Further details about Study 1922 have been summarized in Table 8.

Study 1923 (N = 422) was a 56-week, phase IIIa, randomized, double-blind, placebo-controlled, parallel-group, multi-centre trial in patients who were overweight with comorbidities or living with obesity. The primary objectives of the study were to compare the efficacy of liraglutide 3 mg versus placebo in maintaining run-in weight loss over 56 weeks, and to compare the efficacy of liraglutide 3 mg versus placebo in inducing weight loss beyond what was achieved during run-in over the same treatment period. The trial was conducted at 36 sites in Canada and the US. Prior to randomization, patients were treated with a low-calorie diet (total energy intake of 1,200 kcal/day to 1,400 kcal/day) in the run-in period lasting up to 12 weeks under the instruction and supervision of a qualified nutritionist. Those who achieved at least 5% loss in body weight on the low-calorie diet during the run-in period were randomized in a 1:1 manner to receive either liraglutide 3 mg or placebo for 56 weeks. Randomization was stratified according to comorbidity status (i.e., the presence or absence of treated or untreated hypertension or dyslipidemia). Further details about Study 1923 have been summarized in Table 9.

Study 3970 (N = 359) was a 32-week phase IIIa, randomized, double-blind, placebo-controlled, parallel-group, multi-centre, multinational trial conducted at 5 sites in Canada and 35 sites in the US. The primary objective of the trial was to investigate if treatment with liraglutide 3 mg reduces the severity of OSA, as assessed by the AHI, compared to placebo both in combination with lifestyle intervention in patients living with obesity and moderate or severe OSA unable or unwilling to use CPAP treatment. After a 2-week screening period, patients were randomized using an IV/WRS in a 1:1 manner to receive either once-daily liraglutide 3 mg or placebo using IV/WRS, without stratification. Randomization was not stratified. Further details about Study 3970 have been summarized in Table 10.

Populations

Inclusion and Exclusion Criteria

Study 1839 enrolled adult patients without diabetes who were living with obesity (i.e., having a stable BMI of ≥ 30 kg/m²) or were overweight (i.e., BMI ≥ 27 kg/m²) with comorbidities such as dyslipidemia or hypertension. The major inclusion and exclusion criteria have been summarized in Table 7. Of note, patients were eligible for inclusion if they were adults without diabetes diagnosed with obesity or overweight and had failed previous dietary interventions to manage weight. Patients were not eligible to be included in the study if they had type 1 or type 2 diabetes, a family or personal history of multiple endocrine neoplasia type 2 or familial medullary thyroid carcinoma, obesity induced by drug treatment, previous surgical treatment of obesity, or uncontrolled hypertension, among others. Of the total 3,731 patients randomized to take part in the main phase of Study 1839, 2,200 patients who were identified as having prediabetes at screening continued into the 104-week extension phase that investigated the long-term efficacy of liraglutide 3 mg in delaying the onset of type 2 diabetes in patients who are overweight or living with obesity with prediabetes, for a total treatment period of 160 weeks.

Study 1922 enrolled adult patients with type 2 diabetes who were on diet and exercise treatment alone, or with 1 to 3 oral antidiabetic drugs (OADs) (i.e., metformin, SU, or glitazone, either as monotherapy or in combination with any of the other 2 compounds). To be eligible for inclusion in the study, patients had to have hemoglobin A1C between 7.0% and 10.0% (both inclusive) and BMI of 27.0 kg/m² or more and failed a previous dietary effort for weight management. Patients were excluded from the study if they experienced recurrent major hypoglycemia or hypoglycemic unawareness as judged by the investigator, received treatment with any hypoglycemic agent(s) other than metformin, SU, and glitazone in the 3 months before screening, or used any drug (except for metformin, SU, or glitazone) that, in the investigator’s opinion, could interfere with their glucose level. Other exclusion criteria were like those previously described for the main phase of Study 1839. The major inclusion and exclusion criteria of Study 1922 have been summarized in Table 8.

Study 1923 enrolled adult patients with BMI of 30 kg/m² or more or with BMI of 27 kg/m² or more with the presence of comorbidities of treated or untreated dyslipidemia and/or hypertension. Untreated dyslipidemia was defined as low-density lipoprotein (LDL) of 160 mg/dL or more (i.e., ≥ 8.9 mmol/L) or triglycerides of 150 mg/dL or more (i.e., ≥ 8.3 mmol/L), or high-density lipoprotein (HDL) of less than 40 mg/dL (i.e., ≥ 2.2 mmol/L) for males and less than 50 mg/dL (i.e., ≥ 2.8 mmol/L) for females. Untreated hypertension was defined as systolic blood pressure (SBP) of 140 mm Hg or more and/or DBP of 90 mm Hg or more. To be eligible for consideration for inclusion in the trial, patients had to have a stable body weight during the previous 3 months (< 5 kg self-reported weight change) and must have undergone a previous weight loss with dietary adjustments without being able to maintain reduced weight. Exclusion criteria included patients with type 1 or type 2 diabetes, a history of cancer, a positive screening for hepatitis B surface antigen, hepatitis C antibodies, and HIV antibodies, and other criteria as previously described for the main phase of Study 1839. A summary of the key inclusion and exclusion criteria of Study 1923 have been provided in Table 9.

In Study 3970, patients living with obesity who had moderate to severe OSA were enrolled. Key inclusion criteria were BMI of 30 kg/m² or more, a diagnosis of moderate or severe OSA, unwillingness or inability to use CPAP, and a stable body weight (defined as < 5% self-reported change during the previous 3 months). Exclusion criteria included patients with type 1 or type 2 diabetes, significant craniofacial abnormalities that may be causing OSA, respiratory and neuromuscular diseases that could interfere with the results of the trial in the opinion of the investigator, and use of central stimulants, hypnotics, mirtazapine, opioids, or trazodone within the 3 months before screening. Other exclusion criteria were similar to those described for the main phase of Study 1839. The major inclusion and exclusion criteria of Study 3970 have been summarized in Table 10.

Baseline Characteristics

Overall, the treatment groups in all the included studies appeared well balanced with respect to baseline demographics and other characteristics.

For Study 1839, the mean age of included patients at baseline was 45 years in both treatment groups, and the study populations were predominantly White (85%) and female (≥ 78%). Most patients enrolled had a BMI of 30 or more, with less than 4% of patients having a BMI between 27.0 kg/m² and less than 30 kg/m². Also, most of the patients were designated as having prediabetes (61%) at screening and having no history of CV disease (> 57%) or gallbladder disease (> 85%). Cardiometabolic disease markers were reported in the form of dyslipidemia (≥ 29%), hypertension (≥ 34%), or both dyslipidemia and hypertension (17%) in both treatment groups. In general, the baseline characteristics were similar for patients in both the 56-week main study phase and those who continued into the 104-week extension phase, except for the fact that all patients in the extension phase had prediabetes, had a history of CV disease, and were taking antihypertensive drugs. Further details about the baseline characteristics of the main and extension phases of Study 1839 have been summarized in Table 11.

The mean age of patients in Study 1922 was 55 years at baseline and the study population was predominantly White (83%). The proportion of female patients and male patients was similar across treatment groups. Most enrolled patients had a BMI of 30 kg/m² or more. As at the time of randomization, the duration of diabetes was 7.4 years in the liraglutide 3 mg group and 6.7 years in the placebo group. More than 70% of patients in each group had a history of CV disease, and cardiometabolic disease markers were reported in a large proportion of patients in both treatment groups in the form of dyslipidemia (> 59%), hypertension (> 68%), or both dyslipidemia and hypertension (> 43%). The uses of concomitant medication and background diabetes treatment were similar across the study groups. Further details about the baseline characteristics of Study 1922 have been summarized in Table 12.

For Study 1923, the mean age of enrolled patients at baseline was about 46 years in both treatment groups and most patients were female (> 78%). Most patients enrolled had a BMI of 30 kg/m² or more, with fewer than 3% of patients having a BMI between 27.0 kg/m² and less than 30 kg/m². The proportion of patients with a history of CV disease or gallbladder disease and concomitant medication use were low (< 2.5%) at baseline. However, 33% and 28% of patients in the liraglutide group and placebo group, respectively, had hypertension at baseline. Further details about the baseline characteristics of Study 1923 have been summarized in Table 13.

The mean age of patients in Study 3970 was 48.5 years. The study enrolled mostly male patients (72% in each group). Most patients enrolled had a BMI of 30 kg/m² or more. One (0.6%) patient in the placebo group had a BMI between 27.0 kg/m² and less than 30 kg/m² but there was no patient in that BMI category in the liraglutide 3 mg group. Like other characteristics, the baseline OSA-related parameters were similar across study groups. More than 60% of patients in each treatment group had prediabetes. Further details about the baseline characteristics of Study 1922 have been summarized in Table 14.

Interventions

Study 1839 randomly assigned patients in a 2:1 ratio to receive once-daily subcutaneous injections of liraglutide at a dose of 3 mg (n = 2,487) or placebo (n = 1,244). To mitigate any potential GI side effects, treatment with liraglutide was initiated using a 4-week dose titration schedule, as follows: 0.6 mg for week 1; 1.2 mg for week 2; 1.8 mg for week 3; 2.4 mg for week 4 and 3 mg for week 5, and thereafter. Liraglutide was available in a concentration of 6 mg/mL in a 3 mL FlexPen. The placebo volume was adjusted to match the corresponding dose of liraglutide. Liraglutide 3 mg or placebo was administered once daily by subcutaneous injections either in the abdomen, thigh, or upper arm at any time of day, regardless of meals. However, it was preferable that injections were performed at approximately the same time on a day-to-day basis. Although not stated, it is reasonable to assume that the injections were self-administered by patients, given their outpatient status and the daily dosing requirement.

Patients in both the liraglutide group and placebo group received counselling on lifestyle modification from randomization and throughout the entire trial. Counselling was provided by a qualified dietitian according to local standards and the patients were put on a reduced-calorie diet containing a maximum of 30% of energy from fat, approximately 20% of energy from protein, and approximately 50% of energy from carbohydrates, and with an energy deficit of approximately 500 kcal/day compared with the patients’ estimated total energy expenditure. All patients were instructed by a qualified dietitian to keep a 3-day food diary every second month for the purpose of diet counselling. Compliance with the prescribed diet was at the discretion of the dietitian based on review of the food diary.

After 28 days of treatment (liraglutide 3 mg plus diet and exercise), it was acceptable for the recommended energy intake to be recalculated with no kcal deficit if patients were unable to lose additional weight despite having a BMI of 25 kg/m² or more. Also, if a BMI 22 kg/m² or less was reached, the recommended energy intake was recalculated with no kcal deficit (maintenance diet) for the remainder of the trial. An increase in physical activity (recommended for a minimum 150 minutes per week) was encouraged and re-enforced by use of pedometers.

Concomitant medications were to be taken as usual during the conduct of the trial, except on pre-specified clinic visits where they were withheld until after blood sampling had been done. The proportion of patients with change in concomitant medication from baseline to week 56 in antihypertensive drugs, lipid-lowering agents, and OADs were analyzed and the results reported in the Efficacy section of this report.

For Study 1922, 423 patients were randomized to receive treatment with liraglutide 3 mg, while 210 patients and 212 patients were allocated to receive liraglutide 1.8 mg and placebo, respectively. The placebo treatment group was further subdivided into 2 treatment groups, with different injection volumes corresponding to the different dose levels of liraglutide. Treatment initiation and dose titration was given to the recommended maintenance 3 mg daily dose, as well as counselling on diet and physical activity, and provision for recalculating energy intake followed a similar manner as previously described for Study 1839. Concomitant medication use was limited to metformin, SU, and glitazone, either as monotherapy or a combination of any of the 3.

In Study 1923, a total of 422 patients who lost 5% body weight or more on a reduced-calorie diet during the 4-week to 12-week run-in period were randomly assigned to treatment with liraglutide 3 mg (n = 212) or placebo (n = 210). Treatment initiation and dose titration were given to the recommended maintenance level of a 3 mg daily dose, as well as counselling on diet and physical activity, and provision for recalculating energy intake followed a similar manner as previously described for Study 1839.

Study 3970 randomly assigned a total of 359 eligible patients in a 1:1 ratio to receive either once-daily liraglutide 3 mg (n = 180) or placebo (n = 179). Initiation of treatment followed the dose escalation procedure as previously described. Patients in both groups were counselled throughout the trial by a dietitian, were maintained on a 500 kcal/day deficit diet and encouraged to exercise for a minimum of 150 minutes per week. The counselling on diet and physical activity, with provision for recalculating energy intake after 28 weeks of BMI review, followed in a similar manner as previously described for Study 1839.

Outcomes

A list of efficacy end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 15. These end points are further summarized as follows. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

Mortality

Mortality was not measured as an efficacy outcome in any of the included trials. However, where it occurred, the number and percentage of deaths were reported under the safety evaluation in the manner described as follows.

Body Weight

Weight was measured at all study visits except for screening visit 2. The same calibrated scales were preferably to be used throughout the trial. Measurements were made in the fasting state with an empty bladder, without shoes, only wearing light clothing, and recorded to the nearest 0.1 kg. Patients’ body weight measured at baseline was used in the assessment of change in body weight. The percentage change in fasting body weight from baseline, the proportion of patients losing 5% or more of body weight, and the proportion of patients losing more than 10% of baseline body weight after 56 weeks of treatment were co-primary efficacy end points in the main phase of Study 1839 and Study 1922. In Study 1923, the percentage change in weight, the percentage of patients maintaining run-in fasting weight loss, and the proportion of patients losing 5% or more baseline weight after 56 weeks of treatment were the co-primary end points, whereas in Study 3970, body weight outcomes were secondary end points.

Body Mass Index

BMI was calculated as follows: BMI is equal to weight (kg) divided by height² (m²), where change in BMI was calculated using weight and height as measured and recorded at screening visit 1. The change in BMI from baseline to the end of study was a secondary outcome in all the included studies.

Health-Related Quality of Life

Patient-reported HRQoL was assessed with tools such as the Impact of Weight on Quality of Life-Lite (IWQOL-Lite) questionnaire, the Short Form (36) Health Survey (SF-36) questionnaire, the Treatment Related Impact Measure of Weight (TRIM-Weight), and the Diabetes Treatment Satisfaction Questionnaire (DTSQ).

IWQOL-Lite

The IWQOL-Lite was assessed as a secondary outcome in Study 1839 and Study 1922. It is a disease-specific questionnaire designed to assess the effect of obesity on quality of life. The IQWOL-Lite has 31 self-administered items with 5 domains: self-esteem (7 items), sexual life (4 items), physical function (11 items), public distress (5 items), and work (4 items). Total scores and subscale scores on the IWQOL-Lite are transformed to a range from 0 to 100, with 100 being the best and a 0 being the poorest quality of life. The reliability of the IWQOL-Lite has been validated in patients living with obesity seeking treatment. The minimal important difference (MID) for an improvement in the IWQOL-Lite total score ranges from 7.7 to 12, depending on the baseline score. The MIDs for improvement were 7.7 to 7.8 for patients with no impairment at baseline (depending on exact baseline score), 7.9 to 8.1 for patients with mild impairment, 8.1 to 8.4 for patients with moderate impairment, and 12.0 for patients with severe impairment.³⁰ The MIDs for deterioration that were determined using the distribution-based method ranged from –7.8 to –4.4, depending on baseline severity of impairment.³⁰ Further information on the IWQOL-Lite questionnaire is provided in Appendix 4.

Short-Form (36) Health Survey

The SF-36 (version 2.0) tool was used to assess HRQoL as a secondary outcome in Study 1839 and Study 3970. It is a 36-item generic instrument used to measure general health, which has been used extensively in clinical trials. It has 8 health domains for scoring: physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional, and mental health. Scores from these 8 domains are summarized using a scoring algorithm into the 2 component summaries of the SF-36: the physical component summary (PCS) and the mental component summary (MCS). Scores on the PCS and MCS range from 0 to 100, with higher scores indicating better health status. The SF-36 has been validated in a variety of disease conditions. There is evidence of SF-36 MCS and PCS validity in the community-based population living with obesity. In the general population, clinically meaningful improvement is indicated by a change of 2 points in the SF-36 PCS and 3 points in the SF-36 MCS.³¹ Based on anchor data, the following minimal mean group differences, in terms of t score points, are described for SF-36 version 2.0 individual dimension scores: physical functioning, 3; role-physical, 3; bodily pain, 3; general health, 2; vitality, 2; social functioning, 3; role-emotional, 4; and mental health, 3.³¹ These MID values were determined as appropriate for groups with mean t score ranges of 30 to 40.³¹ For higher t score ranges, MID values may be higher.³¹ No information about the MID of the SF-36 in the population living with obesity was located. Further information on the SF-36 is provided in Appendix 4.

Treatment Related Impact Measure of Weight

The TRIM-Weight scale was assessed as a secondary outcome in Study 1839. It is a treatment-specific outcomes measure used in patients living with obesity to assess the impacts of anti-obesity medications. The TRIM-Weight scale is a validated 22-item instrument with 5 domains: daily life (6 items), weight management (3 items), treatment burden (4 items), experience of side effects (5 items), and psychological health (4 items). A higher TRIM-Weight score indicates greater improvement. No information was identified about the MID of the TRIM-Weight scale for adult patients who are overweight or living with obesity. However, it was estimated that the MID threshold met ½ standard deviation (SD) criteria for the total score, and all the domain scores.³² No significant relationships were found between the TRIM-Weight total score and BMI category, gender, age, or educational level.³²

Diabetes Treatment Satisfaction Questionnaire

Study 1922 reported mean change from baseline in total DTSQ score as a secondary outcome. The DTSQ is an 8-item questionnaire used to evaluate patients’ satisfaction with treatment regarding convenience, flexibility, and general feelings. The first 36 items measure treatment satisfaction and are summed up to derive a total response score between 0 and 36, with higher DTSQ scores indicating greater satisfaction with treatment.^33,34 The remaining 2 items assess patients’ perceived frequency of hyperglycemia or hypoglycemia, with responses scored on a 7-point scale from 0 (none of the time) to 6 (most of the time). Lower scores on these 2 items indicate greater perceived blood glucose control.^33,34 While the limited number of questionnaire items makes DTSQ convenient to use, it also limits the range that can be assessed concerning the impact and satisfaction that treatment had on patients’ quality of life. No evidence assessing the validity and reliability of the DTSQ in patients living with obesity was identified in the literature. No MID was identified for the change in DTSQ scores.

Development of T2DM and Time of Onset of T2DM

Development of T2DM was reported as a secondary end point in the main study and a primary end point in the extension phase of Study 1839. The presence of T2DM was determined through assessment of glycemic control parameters. In the event of hemoglobin A1C of 6.5% or more at any time during the trial, a repeated measurement was to be taken within 4 weeks to confirm or exclude a diagnosis of diabetes. The time of onset of T2DM was assessed as the annualized incidence rate, defined as the number of new cases of T2DM per 100 patient-years of exposure.

Glycemic Control

Glycemic control was assessed as a secondary outcome using a variety of measures in each of the included studies. However, glycemic control parameters listed in this review’s protocol were hemoglobin A1C and fasting plasma glucose (FPG). According to information provided by the sponsor, the assay method used to assess hemoglobin A1C was certified under the National Glycohemoglobin Standardization Program. Hemoglobin A1C and FPG assessments were performed on blood samples drawn at pre-specified clinic visits.

Weight-Related Comorbidity

Change in weight-related comorbidity was assessed in Study 3970, which investigated the efficacy of liraglutide 3 mg in reducing the severity of OSA, as assessed by the AHI. The AHI is the number of apneas or hypopneas recorded during the study per hour of sleep, expressed as the number of events per hour.³⁵ Polysomnography recordings show the episodes of apneas or hypopneas and the severity of OSA is classified into 4 categories according to AHI scores, as follows³⁵:

• AHI score of less than 5 per hour — None/minimal

• AHI score of 5 or higher per hour but less than 15 per hour — Mild

• AHI score of 15 per hour or higher but less than 30 per hour — Moderate

• AHI score of 30 or higher per hour — Severe

Non-Fatal Cardiovascular Events

This outcome was not measured in any of the included trials as a measure of efficacy. However, Study 1839 reported tachycardia and atrioventricular block as AEs in the manner described as follows in the Safety section.

Renal Outcomes

This outcome was not measured in any of the included trials as a measure of efficacy. However, Study 1839 and Study 1922 reported on renal and urinary outcomes as AEs in the manner as described as follows in the Safety section.

Severity of Depression

This was a secondary outcome reported in Study 1923 using the 9-item Patient Health Questionnaire (PHQ-9) score. The PHQ-9 is a general measure of HRQoL and is a 9-item self-administered Likert scale. It is a reliable, validated scale comprising 9 items that directly correspond to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria for major depression. The total score ranges from 0 to 27, with cut-off scores recommended at 5 for mild, 10 for moderate, 15 for moderately severe, and 20 for severe depressive symptoms. The standard cut-off score of 10 is typically recommended for depression screening. The MID of PHQ-9 has been estimated at 5 points on a scale of 0 to 27 points.³⁶ The reliability and validity of PHQ-9 have been established in patients with persistent major depression, partial remission, and full remission.³⁶ Validity was not assessed in patients living with obesity.³⁷ However, there is some evidence of reliability for adult patients living with obesity awaiting bariatric surgery, although the applicable cut-off score is unclear.³⁷

Physical Function

This outcome was not measured in any of the included trials as a stand-alone efficacy measure. However, the IWQOL-Lite instrument used in Study 1839 and Study 1922 and SF-36 tools used in Study 1839 and Study 3970 have domain items for physical functioning in assessing patients’ HRQoL.

Impact on Work and Daily Activities

This outcome was not measured in any of the included trials as a stand-alone efficacy measure. However, the IWQOL-Lite tool used in Study 1839 and Study 1922 has a work score as 1 of the domain items in assessing patients’ HRQoL.

Health care resource utilization, dose reduction or complete withdrawal of concomitant medications for weight-related comorbidities, the elimination of non-drug interventions for weight-related comorbidities, fatigue, and pain intensity were identified as outcomes of interest in the CADTH review protocol but were not measured in any of the included trials.

Safety

TEAEs, SAEs, and AEs leading to premature withdrawal (withdrawal due to adverse events, or WDAEs) or discontinuation of treatment, as well as deaths, were reported as safety outcomes in the included studies. TEAEs were defined as an event that had onset on or after the first day of randomized treatment and no later than 14 days after the last day of randomized treatment. According to information provided by the sponsor, all AEs either observed by the investigator or reported spontaneously by the patients were to be recorded by the investigator on the AE form in the electronic data capture (the electronic case report form) and evaluated. Also, at each contact with the trial site, the patients were asked whether they had had any AEs (including any changes in concomitant illness or new illnesses) since their last visit. The safety items were summarized descriptively and presented in terms of the number of patients with at least 1 event (N), the percentage of patients with at least 1 event, and the event rate per 100 years (R).

Statistical Analysis

The sponsor-submitted pivotal study (Study 1839) had 3 co-primary end points for the 56-week main phase: 5% or more of baseline body weight, and the proportion of patients losing more than 10% of baseline body weight after 56 weeks of treatment. The primary end point of the study’s 104-week extension phase (i.e., the fourth co-primary end point of Study 1839) was new onset of T2DM among patients with prediabetes after 160 weeks of treatment. Each comparison of liraglutide 3 mg and placebo with respect to the primary end points were tested in a hierarchical manner, based on a predefined fixed sequence of the 4 primary end points (Table 16 and Table 17). The order of testing was as follows:

• relative change from baseline fasting body weight at 56 weeks

• the proportion of patients losing 5% or more of baseline fasting body weight at 56 weeks

• the proportion of patients losing more than 10% of baseline fasting body weight at 56 weeks

• time to onset of T2DM at 160 weeks (extension phase among patients with prediabetes).

The test for each primary end point was carried out based on a 2-sided test with a significance level of 0.05 (provided that superiority of liraglutide is demonstrated for all previous end points).

Relative change from baseline body weight was analyzed using an analysis of covariance (ANCOVA) model with adjustment for country, a BMI stratification factor, the prediabetes status at screening, an interaction between BMI strata and prediabetes status at screening and gender, and baseline fasting body weight. The expected differences between the treatment groups were estimated together with the corresponding 95% confidence interval (CI) and P value. The 2 end points of proportion of patients losing 5% or more and more than 10% of baseline body weight were both analyzed using a logistic regression model, adjusting for the same factors as those in the ANCOVA model. The proportion of patients with onset of T2DM at week 160 among patients with prediabetes at baseline was evaluated at specific visits during the trial and was modelled as a time-to-onset-of-T2DM ratio using Weibull analysis to account for interval censoring. The model adjusted for sex, the BMI stratification factor, and baseline FPG. The findings were presented as a time-to-event ratio along with a corresponding 95% CI and P value, as well as an equivalent estimated treatment hazard ratio.

Study 1922 had the same 3 co-primary end points as described earlier for the 56-week main phase of Study 1839. The hierarchical testing structure and analyses of the 3 primary end points were also similar to that of Study 1839. The factors adjusted for in each comparison for Study 1922 include country, hemoglobin A1C stratification factor, background treatment stratification factor, interaction between stratification factors, sex, and baseline body weight.

Study 1923 had 3 co-primary end points: the percentage of body weight loss, the proportion of patients who maintain run-in fasting weight loss after 56 weeks of treatment, and the proportion of patients losing at least 5% of baseline weight. Patients were deemed to have maintained run-in weight loss if they regained no more than 0.5% of weight from baseline. Each comparison of liraglutide 3 mg and placebo was tested in a hierarchical manner based on a predefined fixed sequence in the order in which the end points are mentioned. Each comparison was made based on a 2-sided test at the significance level of 0.05. The continuous and categorical end points of Study 1923 were similarly analyzed using ANCOVA and logistic regression, respectively. Each model adjusted for sex, country, comorbidities stratification, and baseline weight.

For Study 3970, the primary end point was the change from baseline in AHI (events per hour) after 32 weeks. The comparison between liraglutide 3 mg and placebo was made using ANCOVA, with adjustment for sex, country, baseline AHI, baseline BMI, and baseline age.

Each analysis for the primary outcomes of all studies was carried out using the full analysis set (FAS) and missing data were imputed using LOCF. In all the studies, secondary end points were tested using the same models as primary analyses (i.e., ANCOVA for continuous data and regression analysis for categorical data), but without control for multiplicity. Where continuous data were analyzed using a log-transformed value, an analogous model was applied, but with the corresponding baseline value also log-transformed. Secondary end points were summarized descriptively by visit using observed data based on the FAS.

Subgroup Analyses

All the included studies performed pre-specified subgroup analyses to compare the efficacy of liraglutide in various subgroups. The subgroup analyses pre-specified in the protocol for this systematic review are diabetes status (e.g., prediabetes, T2DM), baseline BMI (e.g., BMI 30 kg/m² or greater versus 27 kg/m² to less than 30 kg/m²), the number and/or type of weight-related comorbidities, patients with or without previous bariatric surgery, and ethnicity.

In Study 1839, subgroup analyses were conducted according to prediabetes status and baseline BMI (BMI ≥ 30 kg/m² or 27 kg/m² to < 30 kg/m²) at screening in the main phase and for BMI alone in the extension phase. Study 1922 performed subgroup analyses based on background treatment (i.e., metformin, SU, and glitazone) whereas Study 1923 conducted subgroup analysis according to comorbidity status (i.e., the presence or absence of treated or untreated hypertension or dyslipidemia. Three subgroup analyses were performed in Study 3970 to investigate the treatment effect for the genders (i.e., female and male), the initial AHI condition (moderate or severe sleep apnea), and initial BMI (30.0 kg/m² to 34.9 kg/m², 35.0 kg/m² to 39.9 kg/m², and ≥ 40.0 kg/m²).

For Study 1839, Study 1922, and Study 1923, it was unclear if statistical tests were performed to evaluate for differences in effects between subgroups. The model for assessing subgroups in Study 3970 tested the interaction of subgroups with fixed factors and covariates of the statistical analysis.

Sensitivity Analyses

All the included studies performed pre-specified sensitivity analyses to support the robustness of the evidence from the primary analyses. In this regard, the same methodologies used in analyzing the primary end points were applied using different analysis sets in all the studies. The key analyses sets that were used in sensitivity testing in the main phase of Study 1839 were:

• a set of completers

• all randomized patients for patients without post-baseline measurements, baseline values were carried forward

• the FAS, including fasting and non-fasting weight measurements and using follow-up measurements of fasting body

• the FAS, including fasting and non-fasting weight measurements, off-drug weight measurements, and follow-up weight measurements

• the FAS, but by imputing missing post-baseline observations of the primary end point using a multiple imputation procedure.

Similar analyses sets were used in sensitivity testing in Study 1922, Study 1923, and Study 3970, each according to their primary end points and end of treatment. In addition to the analysis sets described earlier, the 3 studies also applied the same methodologies used in analyzing their primary end points to the FAS but imputing missing observations with the regression method. Study 1923 also performed a sensitivity analysis using the FAS without imputation, by treating patients without a valid assessment of weight at the end of treatment as nonresponders.

For the extension phase of Study 1839, 7 sensitivity analyses of the primary end point, the onset of T2DM, were performed. This included applying the Weibull model and Cox regression to a set of completers. Other sensitivity analyses sets were as follows:

• patients without prediabetes stratified to join the extension phase

• patients with prediabetes who were re-randomized instead of continuing in the 160-week treatment period, and excluding patients who had normal glycemic parameters at baseline but were stratified to 160 weeks of treatment

• patients with possible T2DM, but without confirmation as having T2DM

• patients who were potentially unblinded during the trial.

Analysis Populations

All the included studies used the FAS in analyzing efficacy outcomes and the safety analysis set (SAS) for analyzing harms outcomes. The FAS was defined as all randomized patients exposed to at least 1 dose of the trial product and with at least 1 post-baseline assessment of any end point. Patients in the FAS were analyzed according to the treatment group to which they were originally randomized.

The SAS included all randomized patients exposed to at least 1 dose of trial product. If a patient received a treatment different than that to which they were randomized, data for the patient was analyzed, tabulated, and/or listed according to the actual treatment they had received.

Results

Patient Disposition

Patient disposition data for the included studies are summarized in Table 18, Table 19, Table 20, Table 21, and Table 22. For Study 1839 (both the main phase and extension phase) as well as Study 1922 and Study 1923, the study discontinuation rate was lower in the liraglutide 3 mg group (%) than in the placebo group (%), indicating a greater proportion of completers with liraglutide than with placebo. AEs were the most common reason for discontinuation in the liraglutide 3 mg group (8.5% to 13.5%), whereas the most common reason for discontinuation in the placebo group was withdrawal of consent (11.2% to 31.1%). The overall rates of discontinuation from the study were lower with liraglutide than with placebo in Study 1839 (liraglutide versus placebo: 28.1% versus 35.6% in the main phase, and 31.1% versus 55.0% in the extension phase, respectively), Study 1922 (liraglutide: 23.4%; placebo: 34.0%), and Study 1923 (liraglutide: 25.0%; placebo: 30.5%), but were lower in the placebo group in Study 3970 (liraglutide: 25.6%; placebo: 20.7%). Although the overall rates of discontinuation due to ineffective therapy were low across the study groups, they were lower with liraglutide than with placebo in Study 1839 (liraglutide versus placebo: 0.9% versus 2.9% in the main phase, and 1.9% versus 4.8% in the extension phase, respectively), Study 1922 (liraglutide: 0.0%; placebo: 1.4%), and Study 1923 (liraglutide: 0.0%; placebo: 1.0%), but were lower in the placebo group in Study 3970 (liraglutide: 1.1%; placebo: 0.6%). There was no consistency across the trials regarding discontinuation due to non-compliance with the study protocols. Table 19 summarizes the disposition of patients without prediabetes who were re-randomized after the 56-week main phase of the study for an additional 12 weeks to evaluate the effects of drug cessation on weight control, possible withdrawal, and rebound effects.

Exposure to Study Treatments

For Study 1839, Study 1922, and Study 3970, the total exposure was expressed as patient-years of exposure to treatments and reflected the randomization ratio used for each study. The computation was based on the SAS.

In the 56-week main phase of Study 1839, the total exposure was 2,234.7 years and 1,067.4 years with liraglutide 3 mg and placebo, respectively. Thus, the mean duration of exposure was the same for the 2 treatments (0.9 years) based on 2,481 patients and 1,242 patients in the liraglutide and placebo groups, respectively. Most of the patients were exposed for 53 weeks to 56 weeks, with 62.0% and 56.6% of patients in the liraglutide 3 mg and placebo groups, respectively, achieving that level of exposure. The difference reflects the higher withdrawal rate with placebo. In the 12-week re-randomization period, the total exposure was 80.2 years, 78.6 years, and 68.2 years for the liraglutide/liraglutide group, the liraglutide/placebo group, and the placebo group, respectively. Most of the patients were exposed 9 weeks to 12 weeks, with 84.6%, 87.1%, and 80.9% of patients in the liraglutide/liraglutide group, the liraglutide/placebo group, and the placebo group, respectively, achieving that level of exposure.

In the extension phase, the total exposure was 3,161 years with liraglutide 3 mg and 1,442 years with placebo, with respective mean exposures of 2.1 years and 1.9 years. Approximately half of all treated patients in the extension phase were exposed to the trial product for 36 months, although the proportion of patients exposed for 36 months or more was higher with liraglutide 3 mg (52.8%) than with placebo (45.2%), reflecting the higher withdrawal rate in the placebo group.

According to information provided by the sponsor, treatment adherence was good, in general, for both the main phase and the extension phase of the study, with most patients reaching the target daily dose of 3 mg. However, protocol deviations regarding missing injections and nonadherence with the dose escalation scheme occurred (357 protocol deviations occurred during the main phase; the number of protocol deviations during the extension phase was not reported). None of these deviations was considered to have any effect on the trial results.

The total exposure in Study 1922 was 379.86 years and 179.71 years for liraglutide 3 mg and placebo, respectively, with corresponding mean duration of exposure of 0.9 years and 0.85 years reflecting the higher withdrawal rate with placebo. The level of compliance with the use of trial interventions was not adequately discussed.

In Study 1923, the mean (SD) duration of treatment was 335.0 (115.2) days and 321.1 (124.2) days, and the mean total exposure was 194.5 years and 184.6 years for the liraglutide 3 mg and placebo groups, respectively. The level of compliance with treatment was not adequately discussed in Study 1923. Although it was indicated for 457 important patient-level protocol deviations, 53 of those deviations (11.6%) were related to treatment compliance.

The total exposure in Study 3970 was 89.5 and 96.0 years for liraglutide 3 mg and placebo, respectively. The mean duration of treatment was slightly lower in the liraglutide 3 mg group (26.5 weeks) than the placebo group (28.0 weeks). The difference in the total exposure reflects the higher number of withdrawn patients in the liraglutide 3 mg group and the slightly greater number of patients exposed to placebo. The level of compliance with the use of trial treatment was not adequately discussed.

Efficacy

Only those efficacy outcomes identified in the review protocol are reported as follows. The investigators conducted sensitivity analyses for the primary end points using different analyses sets. Of the subgroups listed in the protocol for this systematic review, Study 1839 reported subgroup results for both the prediabetes and BMI categories, whereas Study 1922, Study 1923, and Study 3970 reported subgroup outcomes for BMI but not for prediabetes. None of the included studies analyzed efficacy data with respect to weight-related comorbidities, bariatric surgery history, or ethnicity subgroups. The results of the subgroup analyses and sensitivity analyses were consistent with the findings of the primary analyses for each study. A summary of results for subgroup analyses is available in Appendix 3.

Mortality

Mortality was not reported as an efficacy outcome in any of the included studies. A detailed presentation and description of mortality is presented in the Harms section of this report. Briefly, in Study 1839, 1 death occurred in the liraglutide 3 mg group and 2 deaths occurred in the placebo group during the main phase of the study. By the end of the study’s extension phase, each group had a total of 2 deaths, corresponding to mortality rates of 0.1% and 0.3% for the liraglutide 3 mg and placebo groups, respectively. No deaths were reported during Study 1922 or Study 3970, and 1 death occurred in the placebo group in Study 1923.

Change From Baseline in Body Weight

Body weight outcomes reported in the studies were the percentage change in body weight from baseline, the proportion of patients losing at least 5% of baseline body weight, the proportion of patients losing more than 10% of baseline body weight, and the change in kilogram body weight from baseline. In addition, Study 1923 reported on patients maintaining run-in weight loss as well as maintaining more than 50% or more than 75% of baseline body weight loss. The results of body weight outcomes from the included studies are summarized in Table 23, Table 24, Table 25, Table 26, and Table 27.

Percentage of Body Weight Change

The percentage change in body weight from baseline after 56 weeks of treatment was a co-primary end point in the main phase of Study 1839, as well as in Study 1922 and Study 1923. Study 3970 and the extension phase of Study 1839 reported this outcome as a secondary end point after 32 weeks and 160 weeks of treatment, respectively, and neither study controlled for multiplicity for percentage change in body weight from baseline. In Study 1839, at week 56, the mean percentage change in body weight from baseline was –7.98% (SD = 6.67%) in the liraglutide 3 mg group and –2.62% (SD = 5.74%) in the placebo group, with a between-groups difference of –5.39% (95% CI, –5.82 to –4.95; P < 0.0001) in favour of liraglutide 3 mg.

In the extension phase of Study 1839, the mean percentage body weight change from baseline at week 160 was –6.14% (SD = 7.34%) in the liraglutide 3 mg group and –1.89% (SD = 6.27%) in the placebo group, with a between-groups difference of –3.97% (95% CI, –4.84 to –3.11; P < 0.0001), although the analyses did not control for multiplicity.

In Study 1922, the mean percent body weight change from baseline at week 56 was –5.9% (SD = 5.5%) in the liraglutide 3 mg group and –2.0% (SD = 4.3%) in the placebo group, with a between-groups difference of –4.32% (95% CI, –4.94 to –3.70; P < 0.0001) in favour of liraglutide 3 mg.

In Study 1923, the mean percentage body weight change from baseline at week 56 was –6.2% (SD = 7.3%) in the liraglutide 3 mg group and –0.2 (SD = 7.0%) in the placebo group, with a between-groups difference of –6.06% (95% CI, –7.50 to –4.62; P < 0.0001) in favour of liraglutide 3 mg.

Consistent results in favour of liraglutide 3 mg were reported in Study 3970 (–4.15; 95% CI, –5.21 to –3.09; P < 0.0001), as well as in the extension phase of Study 1839 (–4.32; 95% CI, –4.94 to –0.70; P < 0.0001), although this outcome was not controlled for multiplicity.

In the re-randomization phase of Study 1839, the mean percentage change in body weight from baseline was –9.09% (SD = 6.91%) in the liraglutide/liraglutide group, –9.33% (SD = 7.58%) in the liraglutide/placebo group, and –3.47% (SD = 5.74%) in the placebo group, with a between-groups difference (liraglutide/liraglutide versus liraglutide/placebo) of –2.18% (95% CI, –2.60 to –1.75; P < 0.0001) in favour of the liraglutide/liraglutide group, although the analysis did not control for multiplicity.

Responders, 5%

Patients losing 5% or more in baseline body weight was a co-primary end point assessed after 56 weeks of treatment in the main phase of Study 1839, and in Study 1922 and Study 1923. Study 3970 and the extension phase of Study 1839 reported this as a secondary outcome after 32 weeks and 160 weeks of treatment, respectively, and neither study controlled for multiplicity for this end point.

In the main phase of Study 1839, the proportion of patients who lost 5% or more in baseline body weight after 56 weeks of treatment was higher with liraglutide 3 mg than with placebo (63.3% versus 27.1%, respectively) and the OR indicated greater odds of losing 5% or more of baseline body weight with liraglutide 3 mg than with placebo (OR = 4.80; 95% CI, 4.12 to 5.60; P < 0.0001).

In Study 1922, the proportion of patients who lost 5% or more in baseline body weight after 56 weeks of treatment was higher with liraglutide 3 mg than with placebo (49.9% versus 13.8%, respectively) and the OR indicated greater odds of losing 5% or more of baseline body weight with liraglutide 3 mg than with placebo (OR = 6.81; 95% CI, 4.34 to 10.68; P < 0.0001).

In Study 1923, the proportion of patients who lost 5% or more in baseline body weight after 56 weeks of treatment was higher with liraglutide 3 mg than with placebo (46.4% versus 20.9%, respectively) and the OR indicated greater odds of losing 5% or more of baseline body weight with liraglutide 3 mg than with placebo (OR = 3.86; 95% CI, 2.44 to 6.09; P < 0.0001)

Consistent results were reported in Study 3970 (OR = 3.92; 95% CI, 2.41 to 6.38; P < 0.0001), as well as in the extension phase of Study 1839 (OR = 3.22; 95% CI, 2.63 to 3.94; P < 0.0001).

Responders, 10%

Patients losing more than 10% in baseline body weight after 56 weeks of treatment was a co-primary outcome in the main phase of Study 1839 and in Study 1922. Study 1923 and Study 3970, as well as the extension phase of Study 1839, also reported more than 10% loss in baseline body weight after 56 weeks, 32 weeks, and 160 weeks of treatment, respectively, and none of the studies controlled for multiplicity for this end point.

In the main phase of Study 1839, the proportion of patients who lost more than 10% in baseline body weight after 56 weeks of treatment was higher with liraglutide 3 mg than with placebo (33.1% versus 10.6%, respectively) and the OR indicated a greater likelihood of losing more than 10% baseline body weight with liraglutide 3 mg than with placebo (OR = 4.34; 95% CI, 3.54 to 5.32; P < 0.0001).

In Study 1922, the proportion of patients who lost more than 10% in baseline body weight after 56 weeks of treatment was higher with liraglutide 3 mg than with placebo (26.1% versus 6.3%, respectively) and the OR indicated a greater likelihood of losing more than 10% baseline body weight with liraglutide 3 mg than with placebo (OR = 7.10; 95% CI, 3.48 to 14.48; P < 0.0001).

Consistent results in favour of liraglutide 3 mg were reported in Study 1923 (OR = 5.30; 95% CI, 2.79 to 10.08; P < 0.0001) and Study 3970 (OR = 18.96; 95% CI, 5.69 to 63.14; P < 0.0001), as well as in the extension phase of Study 1839 (OR = 3.086; 95% CI, 2.350 to 4.052; P < 0.0001).

Change In Kilogram Body Weight

Change from baseline in kg body weight was a secondary end point in all the included studies, but was not controlled for multiplicity. For this outcome, the least squares mean (LSM) difference in changes in kg body weight after 56 weeks of treatment was –5.56 (95% CI, –6.04 to –5.09; P < 0.0001) for the main phase of Study 1839, –5.56 (95% CI, –6.04 to –5.09; P < 0.0001) for Study 1922, and –5.86 (95% CI, –7.30 to –4.43; P < 0.0001) for Study 1923, all in favour of liraglutide 3 mg. In Study 3970 and the extension phase of Study 1839, the LSM difference from baseline in kg body weight was –4.92 (95% CI, –6.18 to –3.66; P < 0.0001) after 32 weeks of treatment and −4.57 (95% CI, –5.27 to –3.88; P < 0.0001) after 160 weeks, respectively, both in favour of liraglutide 3 mg.

Maintaining Run-In Weight Loss

The percentage of patients maintaining run-in weight loss after 56 weeks of treatment was a co-primary outcome in Study 1923, assessed after 56 weeks of treatment. In this study, only patients who lost 5% or more body weight during a 4-week to 12-week run-in period on a low-calorie diet were randomized to take part in the 56-week treatment period. To be considered as maintaining run-in weight loss, patients should have regained no more than 0.5% of weight lost during run-in at baseline. The proportion of patients who maintained run-in weight loss after 56 weeks of treatment was higher with liraglutide 3 mg than with placebo (82.1% versus 47.9%, respectively). The odds of maintaining run-in weight loss were significantly increased with liraglutide 3 mg than with placebo (OR = 4.82; 95% CI, 3.01 to 7.71; P < 0.0001).

Study 1923 also reported outcomes for patients who maintained more than 50% or more than 75% baseline weight loss after 56 weeks of treatment as secondary end points, though these end points did not account for multiplicity. For these outcomes, the LSM difference was 5.86 (95% CI, 3.12 to 10.98; P < 0.0001) in patients maintaining more than 50% baseline weight loss and 6.02 (95% CI, 3.65 to 9.92; P < 0.0001) in those maintaining more than 75% baseline weight loss, all in favour of liraglutide 3 mg (Table 26).

Change From Baseline in Body Mass Index

Change from baseline in BMI was a secondary end point in all the included studies (Table 28, Table 29, Table 30, and Table 31), and was not controlled for multiplicity. For this outcome, the LSM difference after 56 weeks of treatment was –2.04 (95% CI, –2.10 to –1.87; P < 0.0001) for the main phase of Study 1839, –1.50 (95% CI, –1.83 to –1.18; P < 0.0001) for Study 1922, and –2.05 (95% CI, –2.53 to –1.57; P < 0.0001) for Study 1923, all in favour of liraglutide 3 mg. In Study 3970 and the extension phase of Study 1839, the LSM difference from baseline in BMI was –1.59 (95% CI, –2.00 to –1.17; P < 0.0001) after 32 weeks of treatment and –1.69 (95% CI, –1.93 to –1.44; P < 0.0001) after 160 weeks of treatment, respectively, all in favour of liraglutide 3 mg.

Health-Related Quality of Life

All the included studies evaluated HRQoL outcomes as secondary end points. Scales used to assess HRQoL were the IWQOL-Lite, SF-36, TRIM-Weight, and DTSQ, and none of these outcomes was controlled for multiplicity.

Impact of Weight on Quality of Life-Lite

The IWQOL-Lite scale was 1 of the HRQoL assessment tools in Study 1839 and Study 1922 (Table 32 and Table 33). The LSM difference in the IWQOL-Lite total score after 56 weeks of treatment was 3.13 (95% CI, 2.24 to 4.01; P < 0.0001) for the main phase of Study 1839 and 2.75 (95% CI, 0.57 to 4.93; P = 0.0136) for Study 1922, suggesting positive treatment effects, all in favour of liraglutide 3 mg. Consistent results were reported for each of the 5 domains. Consistent results were observed in the extension phase of Study 1839, where the LSM difference in the IWQOL-Lite total score after 160 weeks of treatment was 3.35 (95% CI, 2.04 to 4.66; P < 0.0001) in favour of liraglutide 3 mg.

Short Form (36) Health Survey

The SF-36 questionnaire was used to assess HRQoL in Study 1839 and Study 3970 (Table 34 and Table 35). For overall physical health, the LSM difference in the SF-36 score was 1.73 (95% CI, 1.22 to 2.24; P < 0.0001) for the main phase of Study 1839 after 56 weeks of treatment and 0.86 (95% CI, –0.49 to 2.20; P = 0.2113) for Study 3970 after 32 weeks of treatment. The corresponding LSM difference in overall mental health scores were 0.90 (95% CI, 0.30 to 1.50; P = 0.0034) and 0.59 (95% CI, –0.86 to 2.04; P = 0.4252) for Study 1839 and Study 3970, respectively.

In the extension phase of Study 1839, the LSM difference suggests that liraglutide 3 mg maintained higher positive treatment effects than placebo in the overall physical health score (estimated difference = 0.87; 95% CI, 0.17 to 1.58; P = 0.0034) but not in the overall mental health score (estimated difference = 0.77; 95% CI, −0.09 to 1.68; P = 0.778) after 160 weeks of treatment (Table 34).

Higher mean changes were also reported with liraglutide 3 mg than with placebo for all 8 SF-36 domains in both studies.