CADTH Reimbursement Review

Semaglutide (Rybelsus)

Sponsor: Novo Nordisk Canada Inc.

Therapeutic area: Diabetes mellitus, type 2

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

Diabetes mellitus is a metabolic disease that is characterized by persistent elevations in blood glucose, or hyperglycemia. Type 2 diabetes mellitus (T2DM) accounts for approximately 90% of cases of diabetes mellitus.¹ Onset of T2DM typically occurs around 40 years of age or older,² though this is changing with the increase in obesity and sedentary behaviours leading to more frequent diagnosis of T2DM in children and younger people.³ Diabetes is a significant problem in Canada, and is 1 of the most common chronic diseases in the country. Diabetes Canada estimated that 3.8 million people in Canada (10% of the population) were living with diabetes in 2020, and that this number will increase to 4.9 million people (12%) by 2030.⁴

Treatment regimens and therapeutic targets should be individualized in patients with T2DM due to the heterogeneous nature of the disease. Initial treatment often consists of lifestyle modifications through diet and exercise, and pharmacological treatment becomes necessary when blood glucose levels are not adequately controlled by these means.⁵ There are many classes of antihyperglycemic agents used to treat T2DM, which include both insulin and noninsulin therapies.⁵ Metformin (MET) is considered first-line therapy and is indicated for most patients. If treatment through lifestyle modifications and MET monotherapy fail to achieve adequate glycemic control, a second or third agent may be added in addition to MET. There are certain disadvantages to consider with some of the options, such as weight gain and/or hypoglycemia associated with the use of thiazolidinediones (TZDs), sulfonylureas (SUs), and insulin.^5,6 In contrast, some agents, such as SGLT2 inhibitors and GLP-1 receptor agonists, may be advantageous with respect to improved renal outcomes with SGLT2 inhibitors as well as improved cardiovascular (CV) outcomes with both of these classes of medications, which is a particular concern as CV effects are common and a leading cause of death among those with diabetes.^7-9 Additional considerations include patient’s renal function, other comorbidities, planning pregnancy, cost and coverage, ease of administration, and patient preference.⁵

The drug under review is semaglutide (Rybelsus), available as an oral tablet at 3 dosage strengths: 3 mg, 7 mg, and 14 mg.¹⁰ Semaglutide is a selective GLP-1 receptor agonist that acts on the same receptor as native GLP-1, an endogenous incretin hormone.¹⁰ Semaglutide tablets received Health Canada Notice of Compliance (NOC) on March 30, 2020. Semaglutide tablets are indicated as an adjunct to diet and exercise to improve glycemic control in adults with T2DM: as monotherapy when metformin is considered inappropriate due to intolerance or contraindications; and in combination with other medicinal products for the treatment of diabetes.¹⁰ The recommended dose and dosage adjustment for semaglutide tablets is to begin with a starting dose of 3 mg once daily. After 30 days, the dose should be increased to a maintenance dose of 7 mg once daily. If additional glycemic control is needed after at least 30 days on the 7 mg dose, the dose can be increased to a maintenance dose of 14 mg once daily.¹⁰

The sponsor has requested that semaglutide tablets be reimbursed for the treatment of adult patients with T2DM in combination with MET, and in combination with MET and SU.

The objective of this review was to perform a systematic review of the beneficial and harmful effects of semaglutide oral tablets (3 mg, 7 mg, and 14 mg) as an adjunct to diet and exercise to improve glycemic control in adult patients with T2DM:

• as monotherapy when metformin is considered inappropriate due to intolerance or contraindications; or

• in combination with other medicinal products for the treatment of diabetes

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient groups who responded to CADTH’s call for patient input and from 1 clinical expert consulted by CADTH for the purpose of this review.

Patient Input

Two patient group input submissions from Diabetes Canada and 1 from the type 2 Diabetes Experience Exchange (T2DXX), were provided for this review. Diabetes Canada used a series of online surveys with 1770 Canadian patients and caregivers that responded. T2DXX obtained data for their input from personal interviews and facilitated group discussions in their Experience Exchange forums, and through social media conversation threads. It is unclear how many patients contributed to the submission from T2DXX.

Patients reported that common symptoms of T2DM included extreme fatigue, unusual thirst, frequent urination and weight change (gain or loss). Hyperglycemia and hypoglycemia are often experienced by people with diabetes; high blood pressure and high cholesterol are common comorbid conditions. Patient groups reported that many healthy behaviours are required to manage diabetes including diet, physical activity, maintenance of a healthy body weight, taking medications (oral and/or injectable) as prescribed, monitoring blood glucose levels, and managing stress. Other health complications or comorbidities and financial barriers can make management of T2DM challenging. The management of blood glucose levels and the frequent visits to health care providers were highlighted as being constant and burdensome. One of the patient groups also described feelings of shame, guilt, and stigma in people with diabetes. Further, the stress of the disease and its potential complications was stated to be emotionally taxing for respondents, negatively influencing social interactions, mental health, and, ultimately, overall quality of life of patients.

The majority of respondents reported that keeping blood glucose at satisfactory levels during the day or after meals, avoiding weight change and avoiding gastrointestinal side effects (i.e., nausea, vomiting, diarrhea, abdominal pain), avoiding low blood sugar and reducing risk of heart problems were the most important considerations for medications for diabetes management. Other considerations reported by approximately 75% of respondents were avoiding urinary tract and/or yeast infections, avoiding fluid retention and reducing high blood pressure. A total of 6 patients reported having experience with semaglutide tablets, but their feedback indicated mixed results in terms added benefit for glycemic control and reduction of side effects when compared to other treatments.

Patients hoped new treatments would be safe, minimize side effects and damage to organs, and improve overall health outcomes. Respondents reported a strong desire to reduce the pill burden associated with treatment, or to be off medication entirely, for treatments to help resume ‘normal living’, such as the ability to eat without restrictions, for treatments with fewer unpleasant side effects (i.e., weight gain, hypoglycemia, gastrointestinal side effects) and which are less physically invasive (i.e., do not require an injection) and for treatments which can normalize/stabilize blood glucose levels, and improve A1C.

Clinician input

Input from clinical experts consulted by CADTH

The clinical expert indicated factors of current treatment that need to be improved upon include: better glycemic control, modification and ideally slowing down the progression of disease, prevention of complications (both microvascular and macrovascular), better side effect and safety profiles, and treatments that are more user-friendly to patients. The clinical expert stated that expected use of semaglutide tablets is aligned with the indication and that semaglutide tablets would be used as an add-on treatment in patients with T2DM when metformin is no longer effective as monotherapy (second-line treatment), as a first-line treatment when metformin is not tolerated, and as a third-line treatment on occasion.

Outcomes identified by the clinical expert that are important when assessing whether a patient is responding to treatment in clinical practice include: improvement in glycemic parameters, improvement in body weight with the attendant improvement in blood pressure and lipids, long-range improvement in microvascular and macrovascular events, and HRQoL. Patients that are most likely to exhibit a response to treatment with semaglutide are those individuals with T2DM who are not controlled with MET alone, who are overweight, who can tolerate minor GI discomforts and who are compliant with taking medication. Treatment should be discontinued if the glycemic response is inadequate, side-effects are intolerable, other treatments in development prove more effective, and according to patient preference. Lastly, the clinical expert felt that this semaglutide tablets can be prescribed by specialty and community-based clinics.

Clinician group input

CADTH did not receive any input from clinician groups for this review.

Drug program input

Representatives from the drug plans acknowledged that there was a lack of evidence comparing semaglutide tablets to semaglutide injections. They were also interested in the clinical expert’s opinion regarding whether there was sufficient evidence to support whether semaglutide tablets offered CV benefits. The clinical expert consulted by CADTH anticipated that, based on the evidence available at this time, physicians would continue to use semaglutide injections in patients when CV benefit was a priority.

Clinical Evidence

Pivotal Studies and Protocol Selected Studies

Description of studies

A total of 10 RCTs met the inclusion criteria for the systematic review. PIONEER 1 to 6 and 8 to 10 have been summarized in detail for this report (details for PIONEER 7 are provided in Appendix 3 as the intervention, semaglutide with flexible dosing, is not aligned with the criteria specified in the CADTH review protocol or the dosing recommended by Health Canada). All of the included studies were randomized, parallel-group, multi-centre, double-blind trials, except PIONEER 2 and 10, which were open-label. PIONEER 9 was a double-blind study with an and open-label treatment arm for liraglutide. A total of 9039 adult patients with T2DM were randomized in PIONEER 1 to 6 and 8 to 10. The trials evaluated the efficacy and safety of semaglutide tablets (3 mg, 7 mg, and 14 mg once daily) over 26 to 78 weeks of therapy. The trials were designed to assess semaglutide in comparison to a SGLT2 inhibitor (empagliflozin, PIONEER 2), a DPP-4 inhibitor (sitagliptin, PIONEER 3), and subcutaneous GLP-1 RAs (liraglutide, PIONEER 4 and 9, and dulaglutide, PIONEER 10), as well as placebo (PIONEER 1, 4 to 6, 8, and 9). Of note, PIONEER 4 and 9 were both active- and placebo-controlled trials. Semaglutide was evaluated as monotherapy (PIONEER 1, 6 and 9), as an add-on to metformin (PIONEER 2), as an add-on to 1 to 2 oral antidiabetics (OADs) (PIONEER 3, 4, 10) or insulin with or without metformin (PIONEER 8). The primary and key secondary outcomes in PIONEER 1 to 5, 8 and 9 was change from baseline to week 26 in A1C (%) and change from baseline to week 26 in body weight (kg), respectively. PIONEER 6 was an event-driven cardiovascular outcome trial (CVOT) that used time from randomization to first occurrence of a major cardiovascular event (MACE) as the primary outcome. Additionally, PIONEER 6 was the only trial to report diabetes-related morbidity and mortality outcomes. The number of treatment-emergent adverse events (TEAEs) during exposure to treatment was the primary outcome in the Japanese safety study, PIONEER 10. Other outcomes reported include HRQoL outcomes, blood pressure, and lipid profiles.

At baseline, patients had lived with T2DM for 3 to 16 years, had A1C levels that ranged from 7.9% to 8.4%, and were receiving treatment that ranged from diet and exercise alone to stable treatment with at least 1 antidiabetic medication. Patients included in PIONEER 5 and PIONEER 6 were living with moderate renal impairment (eGFR 30 to 59 mL/min/1.73 m²), and cardiovascular disease, respectively. In general, the baseline characteristics were similar between treatment groups within each of the included studies; however, there are a few differences to note. There were also differences across trials. The mean age of patients ranged from 54 to 61 years of age across all studies except PIONEER 5 and 6, where the mean age was 70 to 71 years and 66 years, respectively. The proportion of male patients per treatment group ranged from 47% to 57% in PIONEER 1 to 5 and 8, but was greater in PIONEER 6, representing 68% to 69% of enrolled patients, as well as in PIONEER 9 and 10, where 68% to 83% of patients were male. The trials also differed in terms of the race/ethnicity of participating patients. PIONEER 9 and 10 were conducted in Japanese patients only, and 94% to 97% of patients included in PIONEER 5 were White. The proportion of patients who were White ranged from 48% to 86%, Black ranged from 3% to 8%, Asian ranged from 7% to 36%, and Hispanic or Latino ranged from 4% to 30% in the rest of the PIONEER trials. The background medications used differed between the patient populations of included studies; however, this was due to the trial designs. The duration of diabetes ranged from 3 to 4 years in PIONEER 1, 14 to 16 years in PIONEER 5, 6, and 8, and ranged from 7 to 10 years in PIONEER 2 to 4, 9 and 10. Body weight was notably lower in PIONEER 9 and 10, which ranged from 68.0 kg to 74.7 kg, compared to the other PIONEER trials where the mean body weight was between 84.6 kg and 95.5 kg. Lastly, the mean eGFR was notably lower in PIONEER 5 and 6, which specifically included patients with impaired renal function.

Efficacy Results

Detailed results for the primary outcome in PIONEER 1 to 5, 8 and 9, change from baseline in A1C at week 26, are presented in Table 2. In active-controlled trials where semaglutide was evaluated as an add-on to 1 to 2 OADs:

• SEM 14 mg was superior to empagliflozin with a between-group difference in A1C reduction of –0.4% (95 CI, –0.6 to –0.3, P < 0.0001) (PIONEER 2)

• SEM 14 mg and 7 mg were superior to sitagliptin with a between-group difference in A1C reduction of –0.3% (95% CI, –0.4 to –0.1, P < 0.0001) and –0.5% (95% CI, –0.6 to –0.4, P < 0.0001), respectively (PIONEER 3)

• SEM 3 mg failed to demonstrate non-inferiority to sitagliptin with a difference of 0.2% (95% CI, 0.1 to 0.3, P = 0.0856) in favour of sitagliptin (PIONEER 3)

• SEM 14 mg was non-inferior to liraglutide 1.8 mg, the between-group difference in A1C reduction was –0.1% (95% CI, –0.3 to 0.0, P < 0.0001) (PIONEER 4)

When compared to placebo, semaglutide 3 mg, 7 mg, and 14 mg (unless otherwise noted) demonstrated superiority based on:

• A between-group difference in A1C reduction of –0.6% (95% CI, –0.8 to –0.4, P < 0.0001) to –1.1% (95% CI, –1.3 to –0.9, P < 0.0001) when used as monotherapy in treatment-naïve patients (PIONEER 1)

• A between-group difference in A1C reduction of –0.8% (95% CI, –1.0 to –0.6, P < 0.0001) (SEM 14 mg only) as an add-on to MET alone, SU with or without MET, and basal insulin with or without MET, in patients with moderate renal impairment (PIONEER 5)

• A between-group difference in A1C reduction of–0.5% (95% CI, –0.7 to –0.3, P < 0.0001) to –1.2% (95% CI, –1.4 to –1.0, P < 0.0001) as an add-on to insulin with or without MET (PIONEER 8)

• A between-group difference in A1C reduction of –1.1% (95% CI, –1.2 to –0.9, P < 0.0001) (SEM 14 mg only) as an add-on to MET with or without a SGLT2 inhibitor (PIONEER 4)

Additionally, a between-groups difference of –0.8% to –1.4% was reported for semaglutide 3 mg, 7 mg, and 14 mg compared to placebo, and 0.2% to –0.4% compared to liraglutide 0.9 mg in PIONEER 9. In PIONEER 10, a between-groups difference of 0.4% to –0.4% was reported for semaglutide 3 mg, 7 mg, and 14 mg compared to dulaglutide. The clinical expert indicated a reduction of 0.5% in A1C or achievement of A1C between 8 and 8.5% or lower was clinically meaningful, 1 of which was achieved by all treatment groups in the PIONEER studies.

In terms of a reduction in body weight from baseline to week 26, semaglutide as an add-on to 1 to 2 OADs in active-controlled trials (Table 3):

• demonstrated superiority to sitagliptin with a between-groups difference of –1.6 kg (95% CI, –2.0 to –1.1, P < 0.0001) and –2.5 kg (95% CI, –3.0 to –2.0, P < 0.0001) for SEM 14 mg and 7 mg, respectively (PIONEER 3)

• demonstrated superiority to liraglutide with a between-groups difference of –1.2 kg (95% CI, –1.9 to –0.6, P = 0.0003) for SEM 14 mg (PIONEER 4)

• reported a between-groups difference of –0.6 kg (95% CI, –1.1 to –0.1, P = 0.0185) for SEM 3 mg compared to sitagliptin (PIONEER 3); however, the analysis was conducted following a failure in the statistical testing hierarchy and therefore must be interpreted nominally

• reported a between-groups difference of –0.1 kg (95% CI, –0.7 to 0.5, P = 0.7593) for SEM 14 mg compared to empagliflozin, which corresponded to no difference in treatment effect (PIONEER 2)

In placebo-controlled trials, the change in body weight at week 26 was evaluated compared to placebo, where:

• SEM 14 mg demonstrated superiority as monotherapy in treatment-naïve patients with a between-groups difference of –2.3 kg (95% CI, –3.1 to –1.5, P < 0.0001) compared to placebo; however, (PIONEER 1)

• SEM 14 mg demonstrated superiority in patients with renal impairment with a between-groups difference of –2.5 kg (95% CI, –3.2 to –1.8, P < 0.0001) compared to placebo (PIONEER 5)

• SEM 3 mg, 7 mg, and 14 mg demonstrated superiority as an add-on to insulin with or without MET in patients with a between-groups difference of –0.9 kg (95% CI, –1.8 to –0.0, P = 0.0392), –2.0 kg (95% CI, –3.0 to –1.0, P < 0.0001), and –3.3 kg (95% CI, –4.2 to –2.3, P < 0.0001) (PIONEER 8)

• A between-groups difference of –0.1 kg (95% CI, –0.9 to 0.8, P = 0.8692) for SEM 3 mg and 7 mg –0.9 (95% CI, –1.9 to 0.1, P = 0.0866) did not demonstrate a difference in treatment effect

The clinical expert consulted for this review suggested a change in weight of at least 2 kg over 26 weeks would be a meaningful change in clinical practice. This was achieved by patients treated with semaglutide 7 mg and 14 mg in PIONEER 1 to 8, and patients treated with semaglutide 14 mg in PIONEER10. Of note, patients in PIONEER 9 and 10 weighed less at baseline compared to patients in PIONEER 1 to 6 and 8.

Mortality (as an efficacy outcome) and diabetes-related morbidity were only reported in PIONEER 6, and time from randomization to first event adjudication committee (EAC)-confirmed MACE was the primary outcome in the trial. Non-inferiority based on a margin of 1.8 for the hazard ratio (HR) required confirmation before assessing for superiority. The HR for semaglutide 14 mg compared to placebo was 0.79 (95% CI, 0.57 to 1.11), therefore demonstrating non-inferiority by the pre-specified non-inferiority margin; however, the analysis for superiority was not confirmed (P = 0.1749). EAC-confirmed all-cause deaths were reported for 23 patients (1.4%) in the semaglutide 14 mg treatment group and 45 patients (2.8%) in the placebo treatment group of the CVOT. Ten of the 23 deaths in the semaglutide 14 mg treatment group, and 23 of the 45 deaths in the placebo treatment group were caused by CV events.

Health-related quality of life was evaluated in PIONEER 1 to 5, and 8 to 10 using the Short-Form Survey version 2 (SF-36v2), Diabetes Treatment Satisfaction Questionnaire (DTSQ), Diabetes Treatment-Related Quality of Life Questionnaire (DTR-QOL), Control of Eating Questionnaire (CoEQ) and the Impact of Weight on Quality of Life Questionnaire-Lite version (IWQOL). These outcomes were exploratory and measured as a change from baseline. Overall, semaglutide did not show benefit in terms of HRQoL when evaluated against active and placebo comparators.

Change in blood pressure and lipid profile were also evaluated as exploratory outcomes in the included PIONEER studies, the results for which were not notable. Health care resource utilization was included as an outcome in the systematic review protocol, but was not assessed in any of the included studies.

Pre-specified subgroup analyses on the primary analysis in PIONEER 6 were conducted by: sex, age (less than 65 years, 65 years or greater), region, race, BMI, A1C (8.5% or less, greater than 8.5%), renal function (less than 60 mL/min/1.73m², 60 mL/min/1.73m² or greater), and evidence of CV disease at screening. The treatment effect may be greater for patients that weight less (BMI of 30 or less), without a history of myocardial infarction (MI) or stroke before randomization, and for patients exhibiting CV risk factors; however, the latter is limited by a wide confidence interval. Subgroup analyses by A1C, renal function or for patients with a BMI greater than 30, prior MI or stroke, and presence of CV disease do not appear to have a differential treatment effect. Subgroup analyses by background therapy on the change in A1C and body weight in PIONEER 3 and PIONEER 4 were also reported, and were consistent with the primary analysis.

Harms Results

A summary of key harms results is provided in Table 4. The overall frequency of AEs was similar between treatment groups. In the active-controlled trials, AEs were reported by 71% to 80% of patients treated with semaglutide, 70% to 83% of patients treated with active comparators (all: empagliflozin, sitagliptin, liraglutide, and dulaglutide), and 67% of patients in the placebo group of PIONEER 4. In placebo-controlled trials, between 53 and 58% of patients in the semaglutide groups and 56% of patients in the placebo group of PIONEER 1 reported AEs. In PIONEER 5 and 8, between 74% and 83% of patients in semaglutide treatment groups and 65% to 76% of patients in the placebo treatment groups reported AEs. In PIONEER 9 and 10, between 71% and 85% of patients in semaglutide treatment groups, 67% to 82% of patients in the active comparator groups (liraglutide and dulaglutide), and 80% of patients in the placebo treatment group reported AEs. Overall AEs were not reported in PIONEER 6.

In PIONEER 1 to 5, and 8 to 10, SAEs were reported by 0% to 14% of patients across all treatment groups and the frequency of SAEs was similar between treatment groups in all trials. Serious AEs were a key focus of PIONEER 6; 18.9% and 22.5% of patients in the semaglutide 14 mg and placebo treatment groups, respectively, reported a SAE. Individual SAEs were infrequently reported. In PIONEER 1 to 5, and 8 to 10, WDAEs ranged from 2% to 15% in semaglutide treatment groups, 0% to 9% of active comparator groups (empagliflozin, sitagliptin, and liraglutide), and 0% to 5% of placebo groups. Gastrointestinal disorders were the most commonly reported reasons for WDAEs in all studies. In PIONEER 6, 27% of patients in the semaglutide 14 mg treatment group and 17% of patients in the placebo treatment group WDAE, with the most common reasons for WDAE attributed to gastrointestinal disorders. Few deaths were reported in the PIONEER trials. A total of 16 deaths were reported in semaglutide treatment groups across PIONEER 1 to 5 and 8 to 10, 8 deaths were reported in active treatment groups (all), and 3 deaths were reported in placebo groups. No deaths were reported in PIONEER 1, 9, or 10. Deaths for PIONEER 6 were reported in the efficacy section under mortality outcomes.

In all studies, AEs were largely driven by GI disorders; nausea, vomiting, and diarrhea in particular. In general, GI-related AEs were higher in patients treated with semaglutide compared to placebo, as well as active comparators with the exception of other GLP-1 RAs. Health Canada’s review of the safety data concluded that the safety profile of semaglutide tablets, including the frequency of GI AEs, was comparable to the other previously authorized GLP-1 RAs, including Ozempic (semaglutide injection).¹¹

Critical Appraisal

PIONEER 2 to 4 were required to demonstrate non-inferiority for comparisons to active treatments before testing for superiority. Pre-defined non-inferiority margins of 0.4% were used in PIONEER 2 and PIONEER 4 and 0.3% was used in PIONEER 3; however, justification for the use of a 0.4% non-inferiority margin was weak. Further, the primary analysis of semaglutide 14 mg compared to placebo in PIONEER 6 used a non-inferiority margin corresponding to a HR of 1.8, which was considered inappropriate by Health Canada¹¹ as 1.3 is recommended.²¹ Another limitation of the trials was that the statistical testing procedures were only used to account for multiple comparisons among the primary and key secondary end points in PIONEER 1 to 5, and 8 and was thus limited to change from baseline to week 26 in A1C and body weight in these studies. Consequently, many of the outcomes (HRQoL, lipid profile outcomes, BMI) reported in the included studies were subject to type I error. Also of note, the open-label study design of PIONEER 2 and 10, as well as discontinuation from treatment due to adverse events and inferred treatment received may have introduced bias to patient reported outcomes and safety analyses, creating uncertainty for these results.

Included studies provided evidence for a heterogenous population of patients with T2DM in terms of disease background, treatment experience, background therapies, and comorbid conditions (renal impairment and CV disease). The comparators used in the trials (empagliflozin, sitagliptin, liraglutide, and dulaglutide) were representative of treatment options for T2DM that are currently used in Canadian clinical practice. The clinical expert consulted for this review supported that the trials overall were fairly generalizable to Canadian patients living with T2DM; however, there are some issues to note with regards to the demographic characteristics of patients, which do not reflect the racial and ethnic diversity of Canadian patients (for example, the majority of patients were white, and Asian patients were typically underrepresented). None of the trials included patients that were specifically contraindicated or intolerant of metformin. In PIONEER 1 and 9 where semaglutide was used as monotherapy, patients were previously treated with diet and exercise, or an OAD (in PIONEER 9) that required a wash-out period. The clinical expert suggested that there is uncertainty regarding whether the from PIONEER 1 are applicable to patients who are intolerant to metformin. Most of the outcomes assessed in the included studies were relevant to clinical practice and based on clinical outcome such as change in A1C, body weight, lipid profile, blood pressure, mortality, and diabetes-related morbidity.

Indirect Comparisons

Description of studies

The ITC consisted of 2 components analyzing change in A1C and change in body weight. The first explored semaglutide as a second-line treatment added to metformin and the second investigated semaglutide as third-line treatment added to metformin and a SU. The methods and analysis used in both were similar and drew from the same systematic review. Forty-three studies were included in NMA for second-line therapies, with 10 of the trials compared to placebo. All trials included a total of 22,721 patients with an average of 220 patients per treatment group, ranging from 14 to 780 patients. The baseline A1C between the different studies ranged from 7.2% to 8.8%. All studies reported age and gender. The baseline weight was an average of 89.2 kg ranging from 79.7 kg to 101.9 kg. The studies were drawn from 2 decades with all studies from 2004 to 2018.

The NMA for third-line therapies included 9 studies in the network with 1 of the trials compared to placebo. All trials included a total of 3,867 patients with an average of 184 patients per treatment group, ranging from 40 to 378. The baseline A1C between the different studies ranged from 8% to 9%. All studies reported age and gender. The baseline weight was an average of 85 kg ranging from 76 kg to 91 kg; but half the studies did not report a baseline weight. The studies were published from 2014 to 2018.

Efficacy Results

Second-line ITC: Overall, semaglutide tablets was found to be more efficacious for reducing A1C versus the majority of other second-line treatments. It was found to be more efficacious than placebo [–1.25 (–1.41, –1.09] and as efficacious to other drugs within the class. The network was found to be largely consistent with only 3 comparisons (1 loop) found to be inconsistent (saxagliptan- > placebo- > dapagliflozin), additional analysis did not change the results. The treatment difference in weight analysis was limited to 35 studies and overall, results suggested that semaglutide tablets were more efficacious for treatment differences in weight versus the majority of other second-line treatments. It was found to be more efficacious than placebo [–3.09 (–3.72, –2.54)] and as efficacious to other drugs within the class. The network was found to be consistent throughout.

Third-line ITC: Overall, semaglutide tablets were found to be the most efficacious for reducing A1C versus all other third-line treatments. It was found to be more efficacious than placebo [–1.33(–1.55, –1.12)]. The network was found to be consistent throughout. The treatment differences in weight analysis was limited to 8 studies and overall, results suggested that semaglutide tablets were to be more efficacious for weight loss versus all other third-line treatments. It was not found to be more efficacious than placebo [–2.20 (–6.88, 2.50] and with only other drugs within the class being as efficacious. The network was found to be consistent throughout.

Harms Results

An analysis of safety was not conducted in the ITC reviewed.

Critical Appraisal

The applicability of the sponsor’s ITC is impacted of the limited scope of the analysis and minimalistic analysis conducted. As described above, the sponsor-submitted ITC did include an extensive systematic review but was limited by the research question, especially limiting to only 2 outcomes. This restriction significantly limited the utility and the robustness of the results. Importantly, no exploration of baseline differences between studies was included. Overall, the results of the submitted ITC indicate semaglutide is likely better than placebo both as second- and third-line therapy. Further, the results may suggest superiority to other treatment classes, specifically SGLT-2 inhibitors, DPP-4 inhibitors, TZD, and SUs; however, all of the results should be interpreted with consideration for the previously described limitations. No conclusions can be made for efficacy or safety outcomes beyond glycemic reduction and weight loss since these outcomes were not evaluated.

Conclusions

The safety and efficacy of semaglutide tablets was evaluated in a total of 9 studies in patients on a variety of background therapies. In terms of glycemic control, once daily treatment with semaglutide tablets demonstrated superiority compared to placebo as monotherapy and as add-on therapy, and as add-on therapy in patients with moderate renal impairment (semaglutide 14 mg). When compared to active treatments as an add-on therapy, semaglutide 14 mg demonstrated superiority to empagliflozin and sitagliptin, and was non-inferior to liraglutide. Semaglutide 7 mg was superior to sitagliptin as well; semaglutide 3 mg failed to demonstrate non-inferiority. In terms of a reduction in body weight, semaglutide demonstrated mixed results. In general, superiority was demonstrated with semaglutide 14 mg with all comparators, but semaglutide 7 mg and 3 mg did not consistently show benefit. Of note, semaglutide 7 mg and 3 mg as monotherapy did not demonstrate superiority in terms of a reduction in body weight when compared to placebo. Regarding CV safety, semaglutide 14 mg was non-inferior to placebo based on time from randomization to first EAC-confirmed MACE indicating no increase in risk in the occurrence of MACE with semaglutide compared to placebo. Based on currently available evidence, CV benefit with semaglutide tablets cannot be claimed. Other outcomes such as HRQoL, blood pressure, and lipid profile were also included in the PIONEER studies as supportive outcomes; however, none of these outcomes were controlled for multiplicity.

The safety profile of semaglutide tablets is comparable to other GLP-1 RAs, with GI disorders such as nausea frequently reported. A clear benefit in HRQoL was not demonstrated based on the included studies, and with a lack of additional evidence regarding outcomes such as diabetes-related morbidity beyond the CVOT, or a direct comparison to semaglutide injection.

Introduction

Disease Background

Diabetes mellitus is a metabolic disease that is characterized by persistent elevations in blood glucose, or hyperglycemia. There are 2 main subtypes of diabetes mellitus: type 1 diabetes mellitus, which is caused by inadequate secretion of insulin from pancreatic beta cells, and type 2 diabetes mellitus (T2DM), which results from target cells for insulin that are unresponsive to the insulin that is produced as well as inadequate production of insulin from the beta cells of the pancreas. Type 2 diabetes mellitus is more common than type 1, accounting for approximately 90% of cases of diabetes mellitus.¹

The etiology of diabetes mellitus is associated with genetic factors and environmental triggers are believed to play a role in the development of disease.²² Onset of T2DM typically occurs around 40 years of age or older,² though this is changing with the increase in obesity and sedentary behaviours leading to more frequent diagnosis of T2DM in children and younger people.³ Poor diet and minimal exercise, and associated weight gain, are considered major risk factors for T2DM.²³ In the earlier stages of disease, patients with T2DM are able to secrete insulin and may even be hyperinsulinemic; however, the disease may progress to a stage where insulin secretion is reduced, similar to type 1 diabetes mellitus. As described by the patient input received for this review, common symptoms of diabetes include extreme fatigue, unusual thirst, frequent urination, and weight change. More serious complications may present for patients with poor glucose control. For example, low glucose may cause confusion, coma, or seizures. High glucose levels may lead to more long-term issues such as damage to the nerves and blood vessels, which increases the risk of blindness, heart disease, stroke, peripheral vascular disease, kidney disease, neuropathy, and damage to the extremities. Patients also reported that diabetes has a great impact on the patients’ emotional, social, and economic status.

The prevalence of diabetes is increasing at a dramatic rate around the world. In a report produced by the WHO, there was an estimated 422 million adults living with diabetes globally in 2014, up from 108 million in 1980.³ Further, this number is projected to increase to 693 million by 2045 if the current trends continue.²² Diabetes is a significant problem in Canada, and is 1 of the most common chronic diseases in the country. Diabetes Canada estimated that 3.8 million people in Canada (10% of the population) were living with diabetes in 2020, and that this number will increase to 4.9 million people (12%) by 2030.⁴ People with diabetes are more likely to be hospitalized and to experience complications requiring care by a specialist. It is estimated that by 2030, the direct costs of diabetes for the Canadian health care system will increase to C$4.9 billion per year.⁴

Standards of Therapy

Treatment regimens and therapeutic targets should be individualized in patients with T2DM due to the heterogeneous nature of the disease. Initial treatment often consists of lifestyle modifications through diet and exercise, in addition to nutrition counselling, smoking cessation and avoidance of excess intake of alcohol as noted by the clinical expert. When blood glucose levels are not adequately controlled by lifestyle modifications (such as diet and exercise) alone, pharmacological treatment becomes necessary.⁵ There are many classes of antihyperglycemic agents used to treat T2DM, which include both insulin and noninsulin therapies.⁵ Metformin (MET) is considered first-line therapy and is indicated for most patients. If treatment through lifestyle modifications and MET monotherapy fail to achieve adequate glycemic control, a second or third agent may be added in addition to MET.

There are several oral antidiabetic (OAD) agents that may be used with MET, such as sulfonylureas (SU), meglitinides, thiazolidinediones (TZD), alpha-glucosidase inhibitors, dipeptidyl peptidase-4 (DPP-4) inhibitors, and sodium-glucose cotransporter-2 (SGLT2) inhibitors. Injectable agents, such as glucagon-like peptide-1 receptor agonists (GLP-1 RAs), and insulin and insulin analogues (rapid-acting, intermediate, or long-acting forms), may also be considered as an add-on to MET, or patients can be switched to insulin.⁹ However, according to the 2018 Clinical Practice Guidelines from Diabetes Canada, it is recommended that DPP-4 inhibitors, GLP-1 receptor agonists, or SGLT2 inhibitors be considered first as hypoglycemia and weight gain are less of an issue with these agents, provided contraindications, accessibility, and affordability are considered.⁵ As noted by the clinical expert consulted with on this review, the most important goals of an ideal treatment would be to improve acute symptoms related to elevated glucose levels, to prevent macrovascular and microvascular disease, to improve quality of life, to minimize drug side effects, and societal goals of ensuring ongoing employment and cost-effectiveness of the treatment.

Although there are currently numerous therapeutic options and combination therapy strategies available, none of the available therapies are curative and many patients still have difficulty achieving adequate glycemic control.²⁴ Further, there are certain disadvantages to consider with some of the options, such as weight gain and/or hypoglycemia associated with the use of TZDs, SUs, and insulin.^5,6 In contrast, some agents, such as SGLT2 inhibitors and GLP-1 RAs, may be advantageous in terms of cardiovascular (CV) effects, which is a particular concern as CV effects are common and a leading cause of death among those with diabetes.^7-9 In addition, SGLT2 inhibitors may also have some benefits on renal outcomes, as noted by the clinical expert.

It is recommended that the selection of a second agent is patient specific, and based on the efficacy and safety profile of available agents.⁵ This includes various factors, such as the effectiveness of an agent at lowering blood glucose and glycated hemoglobin (A1C), concerns regarding hypoglycemia, effects on body weight, and the ability to reduce the risk of diabetic microvascular and/or CV complications.⁵ Additional considerations include patient’s renal function, other comorbidities, planning pregnancy, cost and coverage, ease of administration, and patient preference.⁵

Drug

Semaglutide is a selective GLP-1 RA that acts on the same receptor as native GLP-1, an endogenous incretin hormone.¹⁰ In doing so, semaglutide simultaneously increases insulin secretion and decreases glucagon secretion, both in a glucose-dependent manner. The mechanism of blood glucose lowering also involves a delay in gastric emptying.^10,25

The drug under review is semaglutide (Rybelsus), available as an oral tablet at 3 dosage strengths: 3 mg, 7 mg, and 14 mg.¹⁰ Semaglutide tablets received Health Canada Notice of Compliance (NOC) on March 30, 2020. Semaglutide is indicated as an adjunct to diet and exercise to improve glycemic control in adults with T2DM: as monotherapy when metformin is considered inappropriate due to intolerance or contraindications; and in combination with other medicinal products for the treatment of diabetes.¹⁰ The sponsor has requested that semaglutide is reimbursed for the treatment of adult patients with T2DM in combination with metformin, and in combination with metformin and sulfonylurea.

The Health Canada-recommended starting dose of semaglutide tablets is 3 mg once daily.¹⁰ After 30 days, it is recommended that the dose is increased to 7 mg once daily as a maintenance dose. Following 30 days on the 7 mg dose, the maintenance dose may be increased to 14 mg once daily if additional glycemic control is needed. Semaglutide tablets should be taken on an empty stomach with no more than 120mL of water (a half of a glass) at least 30 minutes before food, beverage, or other oral medications to avoid a decrease in absorption.¹⁰

Other GLP-1 RAs currently approved in Canada are semaglutide injection (Ozempic), dulaglutide, exenatide, liraglutide, and lixisenatide. Semaglutide injection was previously reviewed by CADTH and a recommendation to reimburse with conditions was issued in May 2019.²⁶

Key characteristics of currently available antihyperglycemia treatments are presented in Table 5.

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

Two patient group input submissions from Diabetes Canada and 1 from the type 2 Diabetes Experience Exchange (T2DXX), were provided for this review. Diabetes Canada is a national health charity with a focus on research and policy initiatives to help deliver impact (i.e., prevention and treatment strategies) at a population level. T2DXX provides an open, safe, and non-judgmental space for sharing of personal experiences with T2DM to improve the outcomes and quality of life for patients with T2DM. Due to the timing of this submission, Diabetes Canada provided 2 patient group submissions for the review of semaglutide tablets, which were received on December 11, 2019 and December 17, 2020. The input received from T2DXX was provided in December 11, 2019.

Diabetes Canada used online surveys conducted in July/August 2020, November/December 2020, November 2019, November 2018 and October 2016. Data about the number of patients for each survey are included in Table 6. Respondents were between 25 and ≥ 70 years and reported living with diabetes between 1 and ≥ 20 years.

T2DXX obtained data for their input from personal interviews and facilitated group discussions in their Experience Exchange forums, and through social media conversation threads. It is unclear how many patients contributed to the submission from T2DXX.

Disease experience

T2DM was stated to be a chronic and progressive disease. Common symptoms of T2DM included extreme fatigue, unusual thirst, frequent urination and weight change (gain or loss). Hyperglycemia and hypoglycemia are often experienced by people with diabetes; high blood pressure and high cholesterol are common comorbid conditions. Other problems reported included skin infections, gastrointestinal disturbances (nausea, diarrhea), metabolic changes, lymphedema and other autoimmune disorders. Respondents of the November/December survey from Diabetes Canada also reported comorbidities alongside their diabetes, including weight management issues (79%), high blood pressure (64%), mental health concerns (43%), abnormal cholesterol levels (29%) and eye problems (29%). Other problems included fibromyalgia, chronic fatigue, epilepsy, and celiac disease.

Diabetes Canada reported that many healthy behaviours are required to manage diabetes including diet, physical activity, maintenance of a healthy body weight, taking medications (oral and/or injectable) as prescribed, and monitoring of blood glucose and managing stress. Respondents from T2DXX and Diabetes Canada highlighted the difficulty some respondents face with exercise to help manage variations in blood sugar, especially when faced with other health complications or comorbidities and financial barriers. The goal of managing diabetes through healthy behaviour interventions is meant to keep glucose levels within a target range to minimize side effects of the disease and prevent or delay potentially irreversible complications (i.e., blindness, heart disease, kidney problems and lower limb amputations). The management of blood glucose levels and the frequent visits to health care providers were highlighted as being constant and burdensome.

T2DXX highlighted feelings of shame, guilt, and stigma in people with diabetes, illustrating perceptions of T2DM being considered as the ‘bad’ diabetes as it may seem to be a condition brought on by patients, versus type 1 diabetes which is considered the ‘good’ diabetes. The stigma surrounding diabetes was stated by T2DXX to impact patients socially; 1 respondent described missing their insulin injections when at social functions to avoid misconceptions around their condition from their peers, risking further health complications. The stress of the disease and its potential complications was stated to be emotionally taxing for respondents, negatively influencing social interactions, mental health, and, ultimately, overall quality of life of patients.

Experiences with currently available treatments

T2DXX highlighted the complexity and frustration related to diabetes treatments as patients offered conflicting information. Depending on the awareness and access of optimal treatments for diabetes, the choice of interventions can vary at different stages of treatment for patients. Treatment choice may also be influenced by geography (urban versus rural), institutional protocols, and access to diabetes teams (i.e., nutritionists, social workers, ophthalmologists, vascular specialists). The lack of access to certain treatments and resources for some diabetes patients was indicated as a source of inequity within the health care system and may result in patients feeling powerless.

Most respondents (75%) of the November/December 2020 survey form Diabetes Canada reported that they did not have difficulty in accessing their medications. Although, other comments from respondents expressed concern about running out of or losing benefits to pay for medications, and the affordability of medications. Concerns over treatment cost were also highlighted by T2DXX, as choice of treatment may be made based on affordability for the patient in addition to what is most effective. Patients reported having to make trade-offs of therapy versus basic needs, resulting in suboptimal dosing of insulin and setting sensors on pumps to double or triple times the length recommended by manufacturers. Some patients reported there were able to self-manage their disease with the support of a health care team, including nurse educators and dietitians. However, management of diabetes was stated to eventually require insulin therapy. One patient reported that management of diabetes with medications could be addressed more appropriately, as “doctors tell patients ‘if you don’t follow my orders I’ll put you straight onto insulin’. Need to stop using insulin as a threat.”

In the November/December 2020 survey from Diabetes Canada, 13 respondents reported having experience with antihyperglycemic agents; most commonly, patients reported taking metformin (91%) and insulin glargine or an insulin glargine biosimilar (50%). In the November 2019 survey from Diabetes Canada, 11 respondents reported having experience with antihyperglycemic agents; further, 7 respondents reported taking insulin. In the October 2016 survey, 667 patients reporting receiving antihyperglycemic agents; most commonly respondents reported taking metformin (56%). Between 40% and 60% of survey respondents reported being “much better” or “better” able to meet target blood sugars upon fasting, waking or after eating; in addition, between 50% and 60% of respondents reported being “much better” or “better” able to meet target hemoglobin A1c levels, and between 46% and 50% reported being “much better” or “better” able to avoid hypoglycemia. Current treatments were reported to better help maintain or lose weight by 39% of respondents. Some of the respondents of these surveys also indicated having experience with, but were no longer taking, the following medications: sulfonylureas, GLP-1 receptor agonists, DPP-4 inhibitor, DPP-4 plus metformin, meglitinides, SGLT2 inhibitors, short-acting insulin, premixed insulin, U300/other long-acting insulin, orlistat and metformin.

The November/December 2020 survey indicated that patients liked that their current treatments helped with weight management and that “it isn’t insulin injections”. Comments from the survey indicated that respondents disliked the following about their current treatments: medications cause gastrointestinal upset, are difficult for someone with a disability to adjust independently, are expensive and not covered by the provincial drug plan, and are not effective at regulating post-prandial blood sugar levels. The following side effects of treatments were reported: gastrointestinal issues (including stomach pain, indigestion, nausea, vomiting, diarrhea, painful gas and flatulence), polyuria, weight gain, hypoglycemia, genital infections, mood swings, muscle aches and fatigue. Respondents of Diabetes Canada’s November/December 2020 survey were asked to indicate considerations when choosing medications to manage diabetes. Almost all (≥ 80%) respondents reported that keeping blood glucose at satisfactory levels during the day or after meals, avoiding weight change and avoiding gastrointestinal side effects (i.e., nausea, vomiting, diarrhea, abdominal pain), avoiding low blood sugar and reducing risk of heart problems were the most important considerations for medications for diabetes management. Other considerations reported by approximately 75% of respondents were avoiding urinary tract and/or yeast infections, avoiding fluid retention and reducing high blood pressure.

Improved outcomes

Respondents from the surveys conducted by Diabetes Canada were asked to report their expectations for new treatments. Patients hoped new treatments would be safe, minimize side effects and damage to organs, and improve overall health outcomes. Respondents reported a strong desire to reduce the pill burden associated with treatment, or to be off medication entirely, for treatments to help resume ‘normal living’, such as the ability to eat without restrictions, for treatments with fewer unpleasant side effects (i.e., weight gain, hypoglycemia, gastrointestinal side effects) and which are less physically invasive (i.e., do not require an injection) and for treatments which can normalize/stabilize blood glucose levels, and improve hemoglobin A1c. T2DXX also indicated a preference by patients to receive oral medications compared to injections; “once a day oral would be preferable – taking one pill a day would be attractive vs a once a week injectable.”

Respondents from Diabetes Canada also reported expectations for more affordable treatment options, with a desire to have both medications and devices covered by both public and private plans. One respondent stated, “I wish it was more affordable for the masses and covered by FNIHB [First Nations and Inuit Health Branch, Health Canada] for First Nation patients.” Survey respondents from Diabetes Canada also indicated a desire for methods of self-monitoring blood sugar which eliminate the need for finger pricks, and for investments into non-pharmacological interventions (e.g., affordable exercise programs and nutritional education). T2DXX also reported an expectation for improved perceptions about needles and the administration of insulin. Respondents reported that “there really is a stigma about injecting”; 1 patient reported that she had been injecting insulin “for over 30 years and people still make me feel dirty.”

Diabetes Canada and T2DXX also highlighted a hope for improved relationships with health care professionals. Respondents reported disjointed communication and coordination between specialists (i.e., endocrinologists) and general practitioners leading to distrust between patients and their health care providers. Varying levels of knowledge in managing issues related to diabetes among health care practitioners was stated to be a barrier to access, decrease clinical outcomes and lower quality of life indicators for patients.

Experience with drug under review

A total of 3 patients from the November/December 2020 survey and 3 patients from the November 2019 survey from Diabetes Canada reported having experience with semaglutide tablets. Each respondent of the November 2019 survey reported switching to semaglutide tablets from another medication, and that their treatment was covered through private insurance. There were no respondents from T2DXX with experience with semaglutide.

Respondents of the November/December 2020 survey were also asked to report the effectiveness of semaglutide compared to other medications; of the 3 respondents, semaglutide was reported to have “about the same” effectiveness as previously received medications in terms of meeting target blood sugar levels or hemoglobin A1c levels, managing gastrointestinal side effects (i.e., diarrhea, nausea, vomiting, abdominal pain), and incidence/severity of yeast or urinary tract infections. Two out of 3 patients reported semaglutide was “better” or “much better” at reducing the incidence of extreme thirst and/or dehydration. Respondents currently receiving semaglutide reported that treatment was helping them lose weight, or that it had the potential to help them lose weight, and indicated that its oral administration was preferable to an injection; 1 respondent stated oral medications are “easier to take” while another stated “injections don’t bother me.” However, 1 respondent reported that semaglutide resulted in loss of appetite and fear of eating brought on by side effects; this respondent stated, “if I had known the pill was going to make me this sick (vomiting and diarrhea for two months) I never would have started it…I don’t leave the house. I don’t eat. I don’t enjoy food anymore. I am angry and irritable. My [spouse] is worried and tired…I have four other disabilities besides diabetes. Diabetes has now taken over my life and made me unable to leave the house....” Another respondent who just began treatment with semaglutide reported that they were trying to get used to it while also dealing with some gastrointestinal side effects.

One respondent of the November 2019 survey reported that semaglutide was better able than their previous treatments at achieving target hemoglobin A1C levels; 1 respondent was unsure of whether semaglutide was helping them achieve target hemoglobin A1c levels. Two respondents reported semaglutide was much better at helping them meet target fasting blood glucose levels. One respondent reported semaglutide was better at helping them avoid hypoglycemia and gastrointestinal side effects, while another indicated it was worse. Two of the 3 respondents reported they were very satisfied with semaglutide.

Additional information

According to Diabetes Canada and the 2018 Clinical Practice Guidelines for the Prevention and Management of Diabetes in Canada, to achieve optimal blood glucose levels, individualization of diabetes therapy is essential. This includes careful consideration of medication selection, route of administration (oral, injection, infusion), frequency with which someone monitors blood glucose and adjusts dosage, benefits and risks that the patient experiences and/or tolerates, and lifestyle changes the patient is willing or able to make. One of the respondents from the T2DXX input described the future of management of diabetes that they hope will be the gold standard, which included teamwork, cross-training and seeking partnerships (i.e., working with emergency departments to move patients with T2DM out of the emergency room and into the Diabetes Management Centre).

T2DXX highlighted that current health care systems, at the individual and government level, are in need of improvement to address the needs of diabetes patients and improve overall health. Relationships between patients and health care providers was stated to be complex with some patients struggling to navigate through the inconsistencies in health care. Further, T2DXX stated that governments may be unaware of costs for patients with diabetes, especially for those without insurance coverage, and that health outcomes for Canadian patients with diabetes are poor when ranked against countries across the globe. Overall, T2DM was stated to be a heterogenous disease with a lack of coherence in treatment and social and emotional barriers.

Clinician Input

Input from clinical experts consulted by CADTH

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 the diagnosis and management of T2DM.

Unmet Needs

The clinical expert consulted for this review stated that the most important goals of an ideal treatment would be to improve acute symptoms related to elevated glucose levels, to prevent macrovascular and microvascular disease, to improve quality of life, to minimize drug side effects, and societal goals of ensuring ongoing employment and cost-effectiveness of the treatment. The clinical expert indicated that the following elements of current treatment that need to be improved upon: better glycemic control, modification and ideally slowing down the progression of disease, prevention of long-term complications, both microvascular and macrovascular, improved cost-effectiveness, better side effect and safety profiles, and treatments that are more user-friendly to patients.

Place in therapy

The clinical expert stated semaglutide tablets would be used as an add-on treatment in patients withT2DM when metformin is no longer effective as monotherapy (second-line treatment), as a first-line treatment when metformin is not tolerated, and as a third-line treatment on occasion. Further, they felt it is still premature to speculate on the ability for semaglutide tablets to modify the disease process in T2DM as well as whether semaglutide tablets would shift the current treatment paradigm or not. The clinical expert also indicated that semaglutide tablets would likely be used where a GLP-1 RA is needed, but an injectable treatment is rejected by the patient.

With regards to whether or not the clinical expert thought it would be appropriate to recommend that patients try other treatments before initiating semaglutide tablets, the clinical expert stated that when a GLP-1 RA is considered appropriate, they would certainly recommend an injectable form of this class such as semaglutide for injection because, in addition to glycemic control and weight loss, this treatment has been shown to have cardiovascular benefits.

Patient population

According to the clinical expert, patients who are most likely to do well with semaglutide tablets are those whose glucose levels are not well controlled by metformin, who are overweight, who can tolerate the GI side-effects of semaglutide tablets, and who reject an injectable GLP-1 RA. They noted that this would include both patients with new onset T2DM and patients with established T2DM.

The clinical expert felt that patients best suited for treatment with semaglutide tablets would be individuals with T2DM not controlled with metformin or intolerant of metformin who are overweight. The clinical expert relayed that the diagnosis of T2DM is straightforward. They also stated that given that patients with T2DM have great variability in their symptoms, pre-symptomatic patients should not be precluded from treatment with semaglutide tablets if the treatment is indicated as noted above.

Patients least suited to treatment with semaglutide tablets would be individuals well controlled with Metformin, individuals not well controlled with Metformin who are overweight but who are willing to take an injectable GLP-1 RA, and patients who are intolerant of the GI side-effects of semaglutide tablets, according to the clinical expert.

Lastly, the clinical expert felt that patients most likely to exhibit a response to treatment are those individuals with T2DM who are not controlled with Metformin alone, who are overweight, who can tolerate minor GI discomforts and who are compliant with taking medication.

Assessing response to treatment

The clinical expert highlighted the following outcomes that are important when assessing whether a patient is responding to treatment in clinical practice: improvement in glycemic parameters, improvement in body weight with the attendant improvement in blood pressure and lipids, and long-range improvement in microvascular and macrovascular events, in addition to improvement in quality of life.

The clinical expert stated that they believe most physicians would consider a meaningful response to treatment to be a significant improvement in glycemic parameters with improvement in acute symptoms, significant weight loss, a decrease in microvascular and macrovascular disease, and improved quality of life with its attendant improvement of performance of activities of daily life.

Initially, treatment response should be assessed every 3 months according to the clinical expert although this is dependent on patient factors such as drug tolerance.

Discontinuing treatment

Treatment should be discontinued if the glycemic response is inadequate, side-effects are intolerable, other treatments in development prove more effective, and according to patient preference, as indicated by the clinical expert.

Prescribing conditions

The clinical expert felt that this semaglutide tablets can be prescribed by specialty and community-based clinics. In their opinion, this is not a medication that should be restricted to specialists and is 1 that can be prescribed by endocrinologists, internists, diabetes nurse practitioners and family doctors.

Clinician group input

CADTH did not receive any input from clinician groups 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 7.

Clinical Evidence

The clinical evidence included in the review of semaglutide 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 from the sponsor and indirect evidence selected from the literature that met the selection criteria specified in the review.

Systematic Review (Pivotal and Protocol Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of semaglutide oral tablets (3 mg, 7 mg, and 14 mg) as an adjunct to diet and exercise to improve glycemic control in adult patients with T2DM:

• as monotherapy when metformin is considered inappropriate due to intolerance or contraindications; or

• in combination with other medicinal products for the treatment of diabetes

Methods

Studies selected for inclusion in the systematic review will include pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 8. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

The literature search will be 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 (https://www.cadth.ca/resources/finding-evidence/press).³⁷

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 was semaglutide (Rybelsus). Clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit the retrieval by study type. 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 December 23, 2020. Regular alerts updated the search until the meeting of the CADTH Canadian Drug Expert Committee (CDEC) on April 21, 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 checklist (https://www.cadth.ca/grey-matters).³⁸ Included in this search were the websites of regulatory agencies (US 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.

In addition, the sponsor of the drug was contacted for information regarding unpublished studies.

Two CADTH clinical reviewers will independently select 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 will be acquired. Reviewers will independently make the final selection of studies to be included in the review, and differences will be resolved through discussion.

Findings from the Literature

A total of 10 studies were identified from the literature for inclusion in the systematic review (Figure 1). The included studies are summarized in Table 9, Table 10, Table 11, Table 12, and Appendix 3. 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 10 phase IIIa RCTs met the inclusion criteria for the CADTH systematic review: PIONEER 1 through 10.^12-20,52 Note that the intervention in the pivotal PIONEER 7 study, semaglutide with flexible dosing, is not aligned with the criteria specified in the CADTH review protocol or the dosing recommended by Health Canada. Therefore, all data for PIONEER 7 are presented in Appendix 3.

Details of the included studies are summarized in Table 9, Table 10, Table 11, and Table 12 and an overview of the trial designs are presented in Appendix 3. PIONEER 1 to 6 and 8 to 10 evaluated the efficacy and safety of semaglutide tablets 3 mg, 7 mg, or 14 mg once daily, alone or in combination with other anti-diabetic medication, compared to placebo (PIONEER 1, 5, 6, and 8) or active comparators (PIONEER 2, 3, 4, 9, and 10) in adults with T2DM and inadequate glycemic control with diet and exercise alone, or with background therapy. Patients included in PIONEER 5 and PIONEER 6 were living with moderate renal impairment, and cardiovascular disease or risk factors, respectively.

The primary objective of PIONEER 1 to 5 and 8 was to evaluate the effect of semaglutide tablets once daily on glycemic control (change from baseline in A1C). The primary objective of PIONEER 6 was to confirm cardiovascular safety by showing that treatment with semaglutide tablets did not result in an unacceptable increase in cardiovascular risk compared with placebo (rule out 80% excess risk) in subjects with T2DM at high risk of cardiovascular events. The primary objective of PIONEER 9 was to assess the dose-response relationship semaglutide tablets compared to placebo on glycemic control. The primary objective of PIONEER 10 was to evaluate the safety and tolerability of semaglutide tablets 3 mg, 7 mg, or 14 mg once daily. Secondary objectives for the PIONEER trials were to evaluate the effect of semaglutide tablets on body weight, and compare safety and tolerability.

All of the included studies were randomized, parallel-group, multi-centre, double-blind trials, except PIONEER 2 and 10, which were open-label. Also, PIONEER 9 was a combination of double-blind for semaglutide tablets and placebo, and open-label liraglutide. PIONEER 2 to 4 and 9 to 10 were active-controlled trials (PIONEER 4 and 9 also compared to placebo), and PIONEER 1, 5, and 8 were placebo-controlled trials. All of the studies assessed the once daily dosing of semaglutide tablets 3 mg, 7 mg, or 14 mg, except PIONEER 2, 4, and 5, which only assessed the 14 mg dose. Background therapy was permitted in each of the trials except PIONEER 1, 6, and 9, as described in Table 13. Briefly, semaglutide was evaluated as monotherapy (PIONEER 1, 6 and 9), as an add-on to metformin (PIONEER 2), as an add-on to 1 to 2 oral antidiabetics (OADs) (PIONEER 3, 4, 10) or insulin with or without metformin (PIONEER 8). The active-controlled trials included comparisons to a SGLT2 inhibitor (empagliflozin, PIONEER 2), DPP-4 inhibitor (sitagliptin, PIONEER 3), and subcutaneous GLP-1 RAs (liraglutide, PIONEER 4 and 9, and dulaglutide, PIONEER 10).

PIONEER 1 to 6 and 8 were conducted globally and PIONEER 6 and 8 included patients from 7 sites (each) in Canada. Between 243 and 1863 patients were included in the full analysis set of each of PIONEER 1 to 5, and 8 to 10, and 3183 patients were included in PIONEER 6. An interactive web/voice response system was used to randomize patients in all studies. A simple randomization scheme (for example, 1:1) was followed for all of the trials except PIONEER 4 and PIONEER 9, where patients were randomized to the active treatment groups and placebo treatment group at a 2:1 ratio. Randomization was stratified by: background therapy in all of the PIONEER trials except 1, 2, and 6; by renal impairment and CV disease classification in PIONEER 5 and 6, respectively; and by country (Japanese and non-Japanese) in PIONEER 1, 3, 4, and 8 (Table 13).

Each trial began with a screening period to assess eligibility and review screening data of patients, which was a duration of 2 weeks in the all of the PIONEER trials except PIONEER 6 that had a 3-week screening period. Additionally, PIONEER 9 included an 8-week screening period that only applied to patients that were receiving an OAD before enrolment to allow for discontinuation and wash-out of the OAD. The treatment period was 52 weeks in all of the included studies except PIONEER 1 and 5 (26 weeks), PIONEER 3 (78 weeks), and the PIONEER 6, which was an event-driven study that continued until at least 122 first EAC-confirmed MACE occurred. The treatment period of all of the trials began with an 8-week dose escalation period for semaglutide tablets (further described under “Interventions”) and were followed by a 5-week follow-up period. Active comparators were escalated according to the treatment’s label (note: liraglutide 0.9 mg follows the Japanese label). Subsequent to the dose escalation period, patients received a maintenance dose of semaglutide tablets for the remainder of the treatment period. A notable difference of PIONEER 8 was that the 52-week treatment period consisted of a 26-week fixed insulin treatment period, where an increase in total daily insulin dose was avoided, followed by a 26-week adjustable insulin treatment period, where the total daily insulin doses could be adjusted by the investigator (further described under “Interventions”).

Populations

Inclusion and exclusion criteria

PIONEER 1 to 5, and 8 to 10 included adult patients, defined as at least 18 years of age (except in Japan, which was at least 20 years of age) with T2DM that was inadequately controlled with current therapy, which ranged from diet and exercise alone to stable treatment with an antidiabetic medication. All patients in PIONEER 2 to 4 were required to have been previously treated with metformin; PIONEER 3 also included patients that were receiving metformin in combination with SU, and PIONEER 4 included patients that were receiving metformin in combination with a SGLT2 inhibitor. Patients included in PIONEER 5 were required to have been receiving metformin and/or SU, basal insulin alone, or metformin in combination with basal insulin. PIONEER 8 included patients that were receiving insulin therapy alone or in combination with metformin. PIONEER 1 and 9 included patients that were treated with diet and exercise alone, or with an OAD that was washed-out before randomization in PIONEER 9. Patients included in PIONEER 10 were receiving an OAD (SU, glinide, TZD, alpha-glucosidase inhibitor, or SGLT2 inhibitor). PIONEER 6 did not have any inclusion criteria for background therapy.

Stable antidiabetic treatment was defined as a stable treatment for at least 90 days before screening in PIONEER 1 to 5, and 8, which varied by treatment as follows:

• Metformin: at least 1500 mg or maximum tolerated

• SU: at least half of the maximum approved dose according to local labelling or maximum tolerated

• Insulin therapies (basal insulin, basal and bolus in any combination, premixed insulin): maximum 20% change in total daily dose

PIONEER 9 and PIONEER 10 required patients to have been on a stable dose of OAD as monotherapy for at least 30 days and 60 days before screening, respectively.

The required A1C level at baseline ranged from 7.0% in most trials (6.5% in PIONEER 9) to between 9.5% and 10.5%, as outlined in Table 9, Table 10, Table 11, and Table 12. Patients were required to have been diagnosed with T2DM for at least 30 days before screening, with the exception of in PIONEER 6 that did not specify a requirement for the duration of diagnosis with T2DM.

PIONEER 6 had a unique set of inclusion criteria aimed at including patients with T2DM at risk of CV outcomes. This included patients that were at least 50 years of age with established cardiovascular disease and/or chronic kidney disease, or patients at least 60 years of age with certain CV risk factors, based on their medical records (Table 11).

Exclusion criteria were similar across the included studies. Patients were excluded if they had a family or personal history of multiple MEN 2 or MTC, history of pancreatitis or major surgical procedures involving the stomach that could affect drug absorption, a recent major cardiovascular event (MACE) or heart failure [New York Heart Association (NYHA) Class IV], or treatment with any medication for the indication of diabetes or obesity other than stated in the inclusion criteria in a period of 90 days before the day of screening with the exception of short-term insulin treatment for acute illness (for a total maximum of 14 days). Patients with moderate renal impairment (eGFR < 60 mL/min/1.73 m²) were excluded from PIONEER 1 to 4 and 8; and PIONEER 6, 9, and 10 excluded patients with severe renal impairment (eGFR < 30 mL/min/1.73 m²). PIONEER 6 also excluded patients on chronic or intermittent hemodialysis or peritoneal dialysis, and PIONEER 5 excluded patients with rapidly progressing renal disease or known nephrotic albuminuria.

Additional details of the inclusion and exclusion criteria for the studies included in this review are summarized in Table 9, Table 10, Table 11, and Table 12.

Baseline characteristics

The baseline characteristics of the included studies are presented in Table 14, Table 15, Table 16, and Table 17. In general, the baseline characteristics were similar between treatment groups within each of the included studies; however, there are a few differences to note. There was an imbalance between treatment groups in terms of sex (% male) in PIONEER 5, 8, 9, and 10, race/ethnicity in PIONEER 1, 5, and 8, particularly due to the proportion of patients that identified as Hispanic or Latino, and by background medication in PIONEER 9. The demographic and disease characteristics of patients in PIONEER 6 were similar between treatment groups.

There were also differences across trials. The mean age of patients ranged from 54 to 61 years of age across all studies except PIONEER 5 and 6, where the mean age was 70 to 71 years and 66 years, respectively. Of note, PIONEER 5 included patients with impaired renal function and PIONEER 6 included patients with or at risk of CV disease. The proportion of male patients per treatment group ranged from 47% to 57% in PIONEER 1 to 5 and 8, but was greater in PIONEER 6, representing 68% to 69% of enrolled patients, as well as in PIONEER 9 and 10, where 68% to 83% of patients were male. The trials also differed in terms of the race/ethnicity of participating patients. PIONEER 9 and 10 were conducted in Japanese patients only, and 94% to 97% of patients included in PIONEER 5 were White. The proportion of patients who were White ranged from 48% to 86%, Black ranged from 3% to 8%, Asian ranged from 7% to 36%, and Hispanic or Latino ranged from 4% to 30% in the rest of the PIONEER trials. The background medications used differed between the patient populations of included studies; however, this was due to the trial designs. The duration of diabetes ranged from 3 to 4 years in PIONEER 1, 14 to 16 years in PIONEER 5, 6, and 8, and ranged from 7 to 10 years in PIONEER 2 to 4, 9 and 10. Body weight was notably lower in PIONEER 9 and 10, which ranged from 68.0 kg to 74.7 kg, compared to the other PIONEER trials where the mean body weight was between 84.6 kg and 95.5 kg. Lastly, the mean eGFR was notably lower in PIONEER 5 and 6, which specifically included patients with impaired renal function.

Interventions

Three doses of semaglutide tablets (3, 7 and 14 mg once daily) were investigated in 5 of the included studies (PIONEER 1, 3 and 8 − 10). Four of the included studies investigated semaglutide tablets 14 mg once daily only (PIONEER 2 and 4 to 6). In all included studies (PIONEER 1 to 6 and 8 to 10) semaglutide tablets were titrated upwards using a fixed schedule to achieve a maintenance dose greater than 3 mg once daily. Titration started with the 3 mg dose once daily for 4 weeks, followed by a dose escalation at 4-week intervals to 7 mg and then to 14 mg, depending on randomized dose. The 7 and 14 mg maintenance doses were achieved after 4 and 8 weeks, respectively. In all studies, patients continued on the maintenance dose of semaglutide tablets for the remainder of the treatment period. Semaglutide tablets were administered once daily in the morning while in a fasting state and up to 30 minutes before the first meal of the day, with up to half a glass or 120 mL of water, swallowed whole. Other oral medications could be taken 30 minutes following administration.

All of the trials that included placebo as the sole comparator (PIONEER 1, 5, 6, and 8) were double-blind. Sitagliptin 100 mg once-daily was the comparator used in PIONEER 3, and both SC liraglutide 1.8 mg and placebo were used as comparators in PIONEER 4, and both studies employed a double-blind, double-dummy study design. In PIONEER 4, liraglutide was administered once daily, and titrated upwards on a weekly basis from 0.6 mg to 1.2 mg, and finally to 1.8 mg over 2 weeks total. PIONEER 9 was both double-blind and open-label; once-daily semaglutide tablets and placebo were double-blind, and liraglutide 0.9 mg was open-label. Liraglutide was administered daily and titrated upwards on a weekly basis from 0.3 mg to 0.6 mg, and finally to 0.9 mg. PIONEER 2 and 10 used once-daily oral empagliflozin 25 mg and once-weekly SC dulaglutide 0.75 mg, respectively, as open-label comparators.

PIONEER 1, 9, and the PIONEER 6 evaluated semaglutide tablets as a monotherapy. Each of the other trials included patients receiving 1 or more of the following treatments in addition to semaglutide tablets: metformin, SU, SGLT2 inhibitor, and insulin (basal, basal-bolus, premix). PIONEER 10 included patients receiving glinide, TZD, or an alpha-glucosidase inhibitor as background therapy. Further details of background therapy used is described in the description of studies tables (Table 9, Table 10, Table 11, and Table 12) and Table 13.

In all of the included trials, rescue medication was permitted for patients with persistent and unacceptable hyperglycemia as judged by the investigator. Criteria for rescue medication are presented in Table 18. In PIONEER 1 to 4 and 9 to 10, rescue criteria based on fasting plasma glucose (FPG) applied from week 8 and onwards, and criteria based on A1C applied from week 26 and onwards in trials of that were greater than 26 weeks duration (all except PIONEER 1 and 5). In PIONEER 5 and 8, rescue criteria applied from week 12 and week 16, respectively, to allow the basal insulin dose to adjust. If an initial FPG value as well as follow-up re-test exceed the limits described below, rescue medication was offered.

There were no rescue criteria in PIONEER 6, but antidiabetic medication (excluding GLP-1 RAs, DPP-4 inhibitors and pramlintide) could be adjusted or added, at the investigator’s discretion and in accordance with standard of care and the current local label. This was aligned with standard-of-care treatment for glycemic control, which was permitted for patients during the trial in addition to standard-of-care treatment for management of complications, comorbidities, and CV risk factors. Concomitant medication was used at the investigator’s discretion and aligned with local practice and regulations. Briefly, the use of CV medication initiated after baseline until the end of treatment visit was reported by patients in the semaglutide and placebo treatment groups, respectively, as follows: anti-hypertensive medication by 27.7% and 29.9%; lipid lowering drugs by 16.0% and 16.1%; anti-thrombotic medication by 12.5% and 11.3%; and diuretics by 10.7% and 13.1%. Additional details regarding concomitant CV medication are presented in Appendix 3.

Patients prematurely discontinued from treatment for safety and tolerability concerns, intending to or becoming pregnant, simultaneously participating in another clinical trial, or if a patient’s calcitonin levels were 100 ng/L or greater.

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 19.

These end points are further summarized below. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

Glycemic control

Glycemic control was measured using a variety of measures in each of the included studies. For the purposes of this review, change from baseline in A1C (%) was reported. The change from baseline in A1C at week 26 was the primary end point in PIONEER 1 to 5, 8, and 9. A1C was measured at every in-person study visit in the PIONEER trials, which typically occurred at week 0 (randomization), week 4, week 8, week 14, and then every 6 or 7 weeks until end of treatment.

Mortality

Mortality as an efficacy outcome was only assessed in PIONEER 6. Mortality was reported as all-cause deaths and CV-related deaths, which included undetermined cause of death (i.e., undetermined cause of death was assumed to be CV-related) and required adjudication by an EAC. Further, mortality was reported via the time from randomization to first occurrence of CV-death, all-cause death, and time from randomization to all-cause death. Time from randomization to CV-related deaths were incorporated into the expanded composite MACE, which was the secondary outcome used in PIONEER 6.

Diabetes-related morbidity and mortality

Diabetes-related morbidity and mortality was only assessed in PIONEER 6. Measures of morbidity and mortality used the MACE composite end point defined as CV death, non-fatal MI, or non-fatal stroke. An expanded MACE composite outcome was also used, which included the same outcomes as the MACE in addition to unstable angina pectoris requiring hospitalization or heart failure requiring hospitalization. In this review, the time from randomization to first occurrence of a EAC-confirmed MACE and the time from randomization to first occurrence of EAC-confirmed all-cause death, non-fatal stroke, or non-fatal MI were reported. A breakdown of EAC-confirmed expanded MACE was also reported.

HRQoL

Patient’s HRQoL was evaluated using a generic measure of HRQoL, the Short Form-36 version 2 (SF-36v2), 2 diabetes-specific measures, the Diabetes Treatment Satisfaction Questionnaire (DTSQ) and the Diabetes Treatment Related-Quality of Life (DTR-QOL) outcomes, and 2 generic/weight-management outcomes, the control of eating questionnaire (CoEQ) and Impact of Weight on Quality of Life (IWQOL). The HRQoL outcomes were reported as a change from baseline to week 26 and end of study.

SF-36v2

The SF-36v2 is a 36-item, generic health status instrument that has been used extensively in clinical trials in many disease areas.⁵³ It consists of 8 health domains: physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health. The 8 domains are aggregated to create 2 component summaries: the physical component summary (PCS) and the mental component summary (MCS), with scores ranging from zero to 100 with higher scores indicating better health status. Previous research suggested a lack of improvement in SF-36 scores (deteriorated or remained stable) following interventions demonstrating modest improvement in A1C levels, blood lipid and blood pressure in patients with T2DM.⁵⁴ There is evidence of validity and reliability in the general population, as well as evidence supporting adequate validity among patients with T2DM; however, the validity and reliability in some dimensions among diabetes patients were not optimal and therefore revalidation of the questionnaire among this patient population has been suggested. The non-disease specific MID for the PCS was 2 points and 3 points for the MCS. A benchmark based on 1-point change was suggested for the MID for patients with T2DM, but the validity of this benchmark is unclear.⁵⁵ The SF-36v2 was measured in PIONEER 1 to 3, 5, and 8 to 10.

DTSQ

The DTSQ was used to assess patient satisfaction with treatment (6 items) and perception of change in hyperglycemia and hypoglycemia (2 items).⁵⁶ The DTSQ has 2 versions that have 8 items each: the DTSQ original status version (DTSQs) and the DTSQ change version (DTSQc). The DTSQs was used in the PIONEER trials. Six of the 8 items measure treatment satisfaction (satisfaction with current treatment, convenience, flexibility, satisfaction with own understanding of diabetes, and likelihood of continuing on or recommending current treatment). The item scores range from “very satisfied” ( = a score of 6) to “very unsatisfied” ( = a score of 0), and the sum of these items is taken to generate a DTSQs score, ranging from 0 to 36. Higher DTSQs scores indicate greater satisfaction with treatment. For the 2 items measuring perceived frequency of hyperglycemia and frequency of hypoglycemia, the items are scored on 7‐point response scales ranging from “most of the time” ( = a score of 6) to “none of the time” ( = a score of 0). Lower DTSQs scores indicate more ideal blood glucose levels in this case. No minimal clinically important difference (MID) was identified for the change in DTSQs scores. The psychometric properties of different language versions of the DTSQs were assessed in a study of type 1 and type 2 diabetes patients treated with insulin or poorly controlled on SUs who then started on insulin treatment. The DTSQs was shown to be consistently reliable in all languages studied and significantly sensitive to change in type 1 diabetes patients at weeks 8, 20, 24, and at last available visit. However, it has also been observed that because patients tend to report satisfaction with current treatment in the absence of experience with alternatives for comparison, the DTSQs often exhibits a ceiling effect.⁵⁶ Change in DTSQs from baseline to end of study was measured in PIONEER 4, 5, and 8.

DTR-QOL

The DTR-QOL was used in PIONEER 9 and 10 and is a Japanese questionnaire which assesses the influence of diabetes treatment on a patient’s HRQoL. Four domains are assessed in this questionnaire using 29 items, including “burden on social activities and daily activities”, “anxiety and dissatisfaction with treatment”, “hypoglycemia” and “satisfaction with treatment ”. The domains for assessment of treatment impact on quality of life in the DTR-QOL were daily activity, social activities, and somatic symptoms. Questionnaire items were adapted from the following questionnaires: Insulin Therapy Related Quality of Life, the Japanese version of the DTSQ, and the Japanese version of the Diabetes Medication Satisfaction Questionnaire. Responses to questionnaire items were captured using a 7-point Likert scale with a score of ‘1’ indicating “strongly agree” and ‘–7’ indicating “strongly disagree”. Item scores are reversed making a score of 7 representative of the highest quality of life. The total score was reported for this review, which is derived from a sum of item scores and converted to a scale that ranges from zero (indicating worse-case scenario) to 100 (indicating best-case scenario).⁵⁷ Validity and reliability were assessed and considered adequate in Japanese patients in with diabetes.⁵⁷ An MID was not identified for this outcome. Additional information about the psychometric properties of the DTR-QOL are summarized in Appendix 4.

CoEQ

The CoEQ was reported in PIONEER 2 and 3. The CoEQ questionnaire has its origins in the Food Craving Record. The questionnaire contains 21 items using 6 sections assessing the intensity and type of food cravings, and subjective sensations of appetite and mood, and the individual’s perceived level of control against a craved food item.⁵⁸ Sections 1 and 2 of the questionnaire pertain to questions of general levels of appetite and overall mood (independent of food craving). Sections 3 and 4 assess the frequency and intensity of food cravings in general. Section 5 assesses cravings for specific foods (e.g., dairy, starch, sweet or non-sweet foods). Section 6, which includes items 20 and 21, assesses the perceived level of control over resisting a nominated, craved food item. Twenty items in the questionnaire are assessed using a visual analogue scale, while 1 item (item 20) allows patients to enter their own nominated food.⁵⁹ A 19-item version of the CoEQ was used in the PIONEER trials.⁵⁸ Evidence of validity and reliability was demonstrated for the 21-item version of the CoEQ, but this was not specific to patients with T2DM. Evidence of validity and reliability in the 19-item version, or an associated MID was not identified.

IWQOL-Lite Clinical Version (IWQOL-Lite-CT)

The IWQOL-Lite-CT was designed to evaluate the impact of change in weight on HRQoL. This outcome measure is composed of 22 items that can be summarized by 5 domains including: psychosocial, physical, physical function, pain/discomfort, and IWQOL-Lite-CT Total. Each of the 22 items are answered based on a 5-point scale with the following options: “1=Never”, “Rarely”, “Sometimes”, Usually”, and “5=Always”.^12,14,17 Lower-level scores indicate higher levels of functioning. The IWQOL-Lite-CT was assessed in PIONEER 1, 3, and 8 and the total score and domain scores were reported for this review.

The IWQOL-Lite-CT was adapted from the original IWQOL-Lite to address inadequacies related to clinical trials as the original IWQOL-Lite was meant for patients enrolled in residential/day treatment programmes. The original IWQOL-Lite was developed before recommendations for medical product labelling based on patient reported outcomes were developed by the FDA; however, it has since been validated extensively for use in weight-loss trials.^60,61 Following a recommendation by the FDA, the use of the IWQOL-Lite-CT in patients with T2DM was evaluted to support broader use of the questionnaire. The evaluation of the psychometric properties of this outcome in patients with T2DM based on non-weight loss trials was limited in terms of responsiveness, but considered satisfactory in terms of validity and reliability.^60,61 An MID was not identified for this outcome.

Please note the IWQOL-Lite-CT will be referred to simply as the IWQOL throughout the remainder of the report.

Blood pressure

Blood pressure was reported as a safety end point in all included studies. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) measured in mmHg were reported as a change from baseline to end of study, and have been summarized for this review.

BMI and/or body weight

Change from baseline to end of study in body weight and BMI were secondary and supportive secondary efficacy end points, respectively, in all included trials.

Lipid profile

Fasting blood lipids (e.g., total cholesterol, HDL cholesterol and LDL cholesterol) was secondary efficacy end point in all included trials.

Health care resource utilization

This outcome was not measured in any of the included trials.

Safety

Treatment emergent AEs, SAEs, WDAEs (adverse events leading to premature treatment discontinuation), and deaths were reported as safety outcomes in this review. Of note, TEAE was the primary outcome measure in PIONEER 10.

According to the sponsor, a targeted approach for the collection of safety data was taken for PIONEER 6 as the trial was designed to evaluate CV outcomes. This included reporting: SAEs, AEs leading to discontinuation of treatment, diabetic retinopathy and related complications, episodes of severe hypoglycemia, hepatic events, pregnancies, and medication errors.

Statistical analysis

Estimands

PIONEER 1 to 5, and 8 to 10 implemented 2 estimands to address different aspects of the primary trial objectives, the treatment policy estimand and the hypothetical estimand, which are defined as follows:

• the treatment policy estimand evaluated the treatment difference at week 26 for all randomized patients regardless of adherence to randomized treatment and initiation of rescue medication; and

• the hypothetical estimand evaluated the treatment difference at week 26 for all randomized patients that adhered to treatment and did not initiate rescue medication.

The treatment policy and hypothetical estimands were the primary and secondary estimands, respectively, in PIONEER 1 to 5, 8, and 10. In PIONEER 9, the hypothetical estimand was the primary estimand and the treatment policy estimand was the secondary estimand. Analyses based on the treatment policy estimand were reported for this review as they align with an intention-to-treat analysis.

PIONEER 6 only used a treatment policy estimand, defined as a comparison of semaglutide tablets and placebo for all randomized patients according to the planned visit schedule regardless of treatment discontinuation.

Observation periods

Three observation periods were defined for the evaluation of efficacy and safety in PIONEER 1 to 5 and 8 to 10:

• The in-trial observation period – starts at randomization, includes the time period from when a patient was randomized (includes any period after initiation of rescue medication or premature discontinuation of treatment) until the final scheduled visit

• The on-treatment observation period – starts at the date of the first dose of treatment, includes the time period when a patient was on treatment with treatment (includes any period after initiation of rescue medication) until the final scheduled visit (last date on treatment + 38 days or end-date for in-trial observation period)

• The on-treatment without rescue medication observation period – starts at the date of the first dose of treatment, includes the time period when a patient was on treatment with treatment, excluding any period after initiation of rescue medication, that is, ending following the last dose of treatment + 3 days or when rescue medication is initiated

Only the in-trial and on-treatment observation periods were used in PIONEER 6.

The analyses of the treatment policy estimand were estimated using measurements from the in-trial observation period in all studies (PIONEER 1 to 6, and 8 to 10). Safety assessments reported in this review were evaluated based on the on-treatment observation period. Of note, PIONEER 6 and 9 reported safety assessments in both the in-trial and on-treatment observation periods; only the latter has been presented in this review. Analyses pertaining to the hypothetical estimand were estimated using measurements from the on-treatment with rescue medication observation period and have not been presented in this review.

Primary Outcomes of the Studies

Power calculation

An overview of the parameters used for power calculations in PIONEER 1 to 6, and 8 to 10 is available in Table 20. The primary outcome in PIONEER 1 to 5, 8, and 9 was the change from baseline in A1C (%) at week 26. Each of these studies was powered at 90% to detect a difference based on this outcome, assuming a withdrawal rate of 10%, and at a 5% significance level. Studies that included change in body weight as a key secondary outcome were also powered to detect a change in this measure.

The primary outcome in PIONEER 6 was time from randomization to first occurrence of a MACE. The study was designed to have at least 90% power to test the primary analysis for non-inferiority. PIONEER 6 was an event-driven trial that planned to collect information until at least 122 first MACEs accumulated within the planned trial duration of 19 months. The estimated sample size was based on first MACEs occurring at a rate of 3 per 100 patient years of observation time in both treatment groups and a lost-to-follow-up rate of 1% per year throughout the trial.

The primary outcome in PIONEER 10 was number of TEAEs during exposure to treatment. Details regarding a power calculation were not reported.

Non-inferiority analyses

PIONEER 2 to 4 and 6 were required to demonstrate non-inferiority for the primary analysis before conducting a test for superiority. PIONEER 2 and 4 used a non-inferiority margin of 0.4%, and PIONEER 3 use a margin of 0.3%. PIONEER 2 and 4 reported that the 0.4% margin was selected based on the effect of comparators on glycemic effect in similar trials. Further, the sponsor reported that for PIONEER 2, 0.4% was selected instead of 0.3% (the accepted standard) because of an anticipated advantage in terms of body weight for comparisons to empagliflozin in PIONEER 2. Further rationale for the use of the broader non-inferiority margin in PIONEER 4 was not provided. Details regarding the methodology for the selection of the non-inferiority margin used in PIONEER 3 were unclear. In PIONEER 6, the non-inferiority margin used was a HR of 1.8; no rationale for the use of this margin was provided.

Statistical test or model

A summary of the statistical testing used for the studies included in this review is provided in Table 21. In PIONEER 1 to 5 and 8, the primary analysis was estimated based on the FAS using week 26 measurements from the in-trial observation period. The primary analysis for the treatment policy estimand was based on a pattern mixture model where baseline A1C was included as a covariate and region and/or stratification factor were included as fixed effects. The model employed a multiple imputation approach to account for missing data which assumed that data was missing at random (MAR). Analysis of the key secondary outcomes (change from baseline in body weight at week 26) and continuous supportive secondary outcomes were conducted using similar methods as the primary analysis. The multiple imputation approach imputed missing data using analysis of covariance (ANCOVA) in PIONEER 1 to 5 and 8.

The primary analysis for the treatment policy estimand in PIONEER 9 evaluated the dose-response in change in A1C (%) using a mixed model for repeated measurements (MMRM), which was estimated based on the FAS using post-baseline measurements up to and including week 26 from the on-treatment without rescue medication period. The model incorporated baseline A1C as a covariate and region and stratification factor as fixed effects. The MMRM method assumes that missing data are MAR. Missing data was not imputed except for patients with no post-baseline assessments for whom the baseline value was carried forward to ensure that all randomized subjects contributed to the statistical analysis. The primary analysis for the treatment policy estimand (the secondary estimand for this study) in PIONEER 9 used similar methods to the analysis of the hypothetical estimand in PIONEER 1 to 5 and 8.

The primary analysis in PIONEER 10 was based on the on-treatment observation period and evaluated using the SAS.

In PIONEER 6 the primary analysis was a time to event analysis for MACE. Patients that did not have a MACE within the observation period were censored at the end of the observation period in the analysis of the primary end point and therefore considered to be still at risk at that end point. Time to event and time to censoring were calculated from randomization for the in-trial observation period. Time to first MACE was measured from randomization to the first occurrence of an event defined as a MACE, regardless of any MACE that follow. If events had the same date of onset, the events are prioritized as follows: CV death > non-fatal MI > non-fatal stroke. Of note, deaths of unknown cause were presumed CV deaths in the statistical analyses. The number of serious adverse events (SAEs) was a secondary end point in the trial. Additional secondary efficacy end points such as A1C, body weight, and lipids were evaluated using descriptive statistics.

A hierarchical testing strategy was used in PIONEER 5 and 6. PIONEER 1 to 4 and 8 used a pre-specified multi-branched gatekeeping procedure with a weighted Bonferroni-based adjustment to control for inflated risk of type I error. Each of the statistical testing strategies were based on the following principles:

• For a specific dose of semaglutide tablets (e.g., SEM 14 mg), demonstration of superiority on A1C was required before testing for superiority based on other outcomes, such as body weight

• Establishment of superiority for A1C was required at all higher dosages before continuing testing at lower dosages

• Non-inferiority must be demonstrated for comparisons to active treatments (or placebo in the CVOT) before testing for superiority

PIONEER 9 did not report the use of a testing strategy to control for type 1 error and PIONEER 10 did not adjust for multiplicity.

Additional details of the statistical testing procedures are provided in Table 22 and graphical representations of the closed testing procedures used in PIONEER 1 to 4 and 8 are provided in Appendix 3 (Figure 15, Figure 16, Figure 17, Figure 18, and Figure 19).

Briefly, a closed testing procedure that operated based on an overall significance level of alpha = 0.05 was allocated to the initial test, which was either a test of non-inferiority for SEM 14 mg compared to active comparators (empagliflozin in PIONEER 2 and sitagliptin in PIONEER 3) or a test of superiority for SEM 14 mg compared to placebo (PIONEER 1, 4, and 8). The alpha level was reallocated to subsequent tests following a confirmed hypothesis. The details of the reallocation of alpha are described in the figures presented in Appendix 3.

Subgroup analyses

PIONEER 6 was the only study that included pre-specified subgroup analyses on the primary end point. Subgroup analyses were conducted based on: sex, age (less than 65 years, 65 years or greater), region, race, BMI, A1C (8.5% or less, greater than 8.5%), renal function (less than 60 mL/min/1.73m², 60 mL/min/1.73m² or greater), and evidence of CV disease at screening. Subgroup analyses were based on the FAS using the in-trial observation period. A stratified Cox proportional hazards model was used for each subgroup analysis and a forest plot with P values for the interaction effect were presented.

The sponsor submitted subgroup analyses by background therapy for the change from baseline in A1C (%) and body weight (kg) at week 26 in PIONEER 3 (metformin with or without SU) and 4 (metformin with or without a SGLT2 inhibitor). Whether or not the methodology for the subgroup analyses were the same as the primary analysis was not specified.

An integrated summary of efficacy that was conducted according to guidance from regulatory agencies included subgroup analyses for the primary end point by baseline A1C and body weight for PIONEER 1 to 5 and 8. The subgroup analyses were conducted using a similar methodology to the primary analysis in the overall population.

Sensitivity analyses

Three types of pre-specified sensitivity analyses were included in PIONEER 1 to 5 and 8 to evaluate the robustness of the primary analysis results regarding missing data. This included a pattern mixture model that used comparator-based multiple imputation and AE-determined comparator based multiple imputation, as well as a tipping-point analysis (Table 21).

PIONEER 6 included 3 pre-specified analyses on the primary analysis to investigate the robustness of the primary end point. They were analyzed using the same methodology as the primary analysis. The first sensitivity analyses included additional covariates; the second was conducted using the on-treatment observation period (with onset during treatment and until 38 days after the last day on treatment); and the third evaluated the effect while patients were considered to be exposed to treatment but used a shorter observation period, i.e., up to 7 days after the last day on treatment (as opposed to 38).

Sensitivity analyses were not conducted in PIONEER 9 or 10.

Analysis populations

PIONEER 1, 5, and 8 to 10 used 2 analysis populations, the full analysis set (FAS) and safety analysis set (SAS).

The FAS included all randomized patients. Patients contributed to a treatment group based on the treatment they were randomized to receive.

The SAS included all randomized patients who received at least 1 dose of treatment. Patients contributed to a treatment group based on the treatment they actually received for the majority of the on-treatment observation period.

PIONEER 2 to 4 used the same FAS and SAS described above, in addition to a per protocol (PP) analysis set.

The PP analysis set comprised all patients in the FAS who have not violated any inclusion criteria, have not fulfilled any exclusion criteria, have a valid baseline A1C measurement and were exposed to treatment and have at least 1 valid A1C measurement while on treatment without rescue medication at or after week 14. Patients contributed to a treatment group based on the treatment they actually received for the majority of the on-treatment observation period.

PIONEER 6 only used a FAS, which included all randomized patients. Each patient belongs to a treatment group based on the treatment which the subject was randomized to receive.

Results

Patient Disposition

Patient disposition for active-controlled RCTs (PIONEER 2 to 4), placebo-controlled RCTs (PIONEER 1, 5, 8), the CVOT (PIONEER 6), and the population-specific safety studies (PIONEER 9, 10) are presented in Table 23, Table 24, Table 25, and Table 26, respectively.

The proportion of screening failures ranged from 7% to 30% in PIONEER 1 to 4, 6, 8 to 10; in PIONEER 5, 55.1% of patients failed screening. All studies reported 92% or more of patients as trial completers, defined as patients who attended the final scheduled visit. In other words, 8% or less of patients discontinued from study across the PIONEER trials. The most common reasons for discontinuation from study were withdrawal by patient and lost to follow-up. There were no major imbalances in discontinuations between treatment groups within each study; however small differences in patients who discontinued from study due to death were reported. PIONEER 1 to 5 and 8 also reported patients that discontinued from study due to death at a frequency of less than 2% in any treatment group. Further, discontinuation due to death was only reported within the SEM 14 mg treatment group of PIONEER 1 and 8 (none for placebo or SEM 3 mg, 7 mg). In PIONEER 5, 1 and 2 patients in SEM 14 mg and placebo treatment groups, respectively, discontinued due to death.

The proportion of patients that discontinued from treatment varied across studies, ranging from 0% to 20%, and was predominantly due to adverse events. Discontinuation from treatment was greater among the semaglutide tablets treatment groups, particularly with the 7 mg and 14 mg dosage strengths, than comparator treatment groups in the active-controlled RCTs, PIONEER 5, 8, 10, and the CVOT.