Initiation Criteria

1. Adult patients diagnosed with T2DM with inadequate glycemic control.

Administration Criteria

1. In combination with metformin (MET) alone, when diet and exercise plus maximal tolerated dose of MET do not achieve adequate glycemic control.
2. Semaglutide should not be reimbursed for use as add-on therapy to MET and another antihyperglycemic drug.

Pricing Conditions

1. Drug plan costs for semaglutide should not exceed the drug plan costs of the least costly currently reimbursed drug used when MET alone is insufficient to achieve glycemic control in the treatment of patients with T2DM.

Summary of Patient Input

Two patient groups, Diabetes Canada and Patient Commando, provided patient input for this submission. Information for the patient input submissions was obtained from online surveys, personal interviews and group discussions, social media platforms, a website story collection, and community responses to a program (carried out by one of the patient groups). The following is a summary of key input from the perspective of the patient groups:

• T2DM is a chronic progressive disease without a cure that brings a common set of symptoms and risk of complications. Beyond the physical impact, T2DM has a burden on patients’ emotional, social, and financial well-being.
• Poor glucose control is serious and problematic, and can lead to a range of serious comorbidities, such as heart disease, blindness, kidney disease, and lower limb amputations. Both patient groups indicated the complexity of treatment modalities for diabetes and the need for constant monitoring and self-management.
• While one patient group indicated that the majority of its respondents reported improvements in meeting glycemic control targets after initiation on their current treatment and were “satisfied” or “very satisfied” with these treatments for diabetes, the other group reported that an increasing number of patients escalated to insulin therapy. In addition, this group noted geographical access inequalities, institutional protocols, and access to specialized diabetes teams as components of treatment that further complicate the access to treatment.
• Both groups expressed a strong desire for medications that can normalize and stabilize blood glucose levels and improve hemoglobin A1C without causing weight gain or hypoglycemia, be safe and affordable, be easily administered with minimal disruption to lifestyle, minimize the risk of diabetes-related complications, avoid polypharmacy, and reduce or eliminate the need for insulin. The need to attend to the broader needs of patients, taking a holistic approach, was also noted.

Clinical Trials

The CDR systematic review included eight RCTs (SUSTAIN-1 to -7 and the Seino study) of patients with T2DM (N = 308 to 3,297).

These trials evaluated the efficacy and safety of semaglutide 0.5 mg or 1 mg once weekly, alone or in combination with an OAD such as MET or MET plus a SU, or basal insulin, compared with placebo or active comparators in adults with T2DM with inadequate glycemic control with background therapy. All three placebo-controlled trials (SUSTAIN-1, 5 and 6) and one active-controlled trial (SUSTAIN-2) included a randomized, double-blind treatment period, and all other active-controlled trials had an open-label design (SUSTAIN-3, -4, -7, and the Seino study). The primary objective of the included trials was to compare the effect of semaglutide once weekly with its respective comparators on change in hemoglobin A1C from baseline, except for SUSTAIN-6 and the Seino study. “Time from randomization to first occurrence of MACE” was the primary outcome in SUSTAIN-6. The occurrence of treatment-emergent adverse event (AE) was the primary outcome in the Seino study. The occurrence of diabetes-related comorbidities (macrovascular and microvascular) was also measured in the only CV outcomes trial (SUSTAIN-6). Change in body weight, body mass index, blood pressure, and blood lipid profile were evaluated in all trials. Health-related quality of life (HRQoL) was evaluated in all trials but SUSTAIN-1 and the Seino study. Noninferiority of treatment with semaglutide versus active comparators on glycemic control was assessed in four SUSTAIN trials (SUSTAIN-2, -3, -4, and -7). Noninferiority of treatment with semaglutide compared with placebo on increase in CV events was assessed in patients who had prior or concomitant CV conditions in SUSTAIN-6. In SUSTAIN-2, -3, -4, -6, and -7, superiority of semaglutide compared with placebo or active treatment for either change in hemoglobin A1C or change in body weight was tested if the noninferiority test criterion for the primary end point was met. Treatment duration of the included trials ranged from 30 weeks to 104 weeks.

Key limitations of the trials include uncertainty in data interpretation related to the outcome measures outside of the hierarchical testing procedure, which was used for controlling for possible inflated type I error; lack of subgroup-specific noninferiority margins; lack of adjustment for multiplicity with subgroup analyses; missing data in most of the trials; and the appropriateness of mixing patients who were controlled on two prior OADs with those controlled on one OAD in one analysis. In SUSTAIN-2, -3, and -4, which comprised patients treated with semaglutide as second- or third-line therapy, subgroup analyses based on prior antidiabetic therapy were performed. However, these were post hoc analyses and were exploratory in nature; therefore, interpreting the results of the subgroup analyses is difficult because of the potential imbalance of patients’ baseline characteristics across the subgroups, complexity of testing interaction effect, and the inconsistency between statistical significance and clinical importance. None of the reviewed studies were designed specifically to assess the effects of semaglutide added onto two antihyperglycemic drugs. Sustain-7 was the only study that enrolled a patient population receiving one OAD as background therapy (i.e., semaglutide as second-line treatment).

Outcomes

Outcomes were defined a priori in the CDR systematic review protocol. Of these, the committee discussed the following:

• glycemic control (e.g., change in A1C)
• body weight (e.g., change in body weight)
• risk of CV events in patients with existing CV disease
• HRQoL was assessed by generic and diabetes-specific questionnaires
• AEs, serious adverse events (SAEs), AEs leading to premature treatment discontinuation, and notable harms.

Change in A1C from baseline was the primary efficacy outcome in all included trials, except for SUSTAIN-6 and the Seino study. Time from randomization to first occurrence of MACE was the primary outcome in SUSTAIN-6, and the occurrence of AEs was the primary outcome in the Seino study.

Efficacy

Harms (Safety)

The AE profile of semaglutide appears to be similar to other drugs in the GLP-1 class and no new safety signals were identified based on the included trials. The overall frequency of AEs was similar between treatment groups within trials. In the placebo-controlled trials (except for SUSTAIN-6), AEs were reported by 56% to 69% of patients treated with semaglutide, and by 54% to 58% of patients treated with placebo. In the active-controlled trials, AEs were reported by 68% to 75% of patients with semaglutide, and 62% to 76% of patients with other active treatments.

SAEs were reported by 4% to 7% of patients who received placebo, 5% to 9% of patients who received semaglutide (2% to 3% of these SAEs were gastrointestinal disorders), and 2% to 8% of patients who received other active treatments. The frequency of AEs leading to treatment discontinuation were 4% to 7% in the placebo group, 3% to 10% in the semaglutide group, and 1% to 7% in other active treatment groups. The frequency of gastrointestinal disorders was higher in the semaglutide group compared with placebo or active comparator that was not a GLP-1 receptor agonist, such as sitagliptin or insulin. The frequency of hypoglycemia was highest in the insulin glargine group in SUSTAIN-4 (39.4%). The risk of hypoglycemia for semaglutide was similar to the other GLP-1 receptor agonists (such as exenatide and dulaglutide), and was lower than sitagliptin and insulin glargine. Severe hypoglycemia was reported infrequently in the included trials.

Patients in SUSTAIN-6 reported higher AE rates compared with the other SUSTAIN trials. The incidence of AEs was 88% to 89%. The rates of SAEs were 32.1%, 29.3%, and 34.9% in patients who received semaglutide 0.5 mg, semaglutide 1 mg, and placebo, respectively. The rates of AEs leading to premature treatment discontinuation were 11.5% to 14.5% in the semaglutide groups and 6.7% in the placebo groups. Furthermore, proportions of new or worsening nephropathy were numerically lower in the semaglutide groups than in the placebo groups, while proportions of patients experiencing diabetic retinopathy complications were numerically higher in the semaglutide groups than in the placebo groups: ▬ with semaglutide 0.5 mg, ▬ with semaglutide 1 mg, ▬ with placebo 0.5 mg, and ▬ with placebo 1 mg.

Indirect Treatment Comparisons

The manufacturer submitted three separate ITCs for patients inadequately control on one OAD that compared semaglutide with other GLP-1 agonists, SGLT-2 inhibitors, and SU. Results of these ITCs suggested a reduction in A1C and body weight with semaglutide versus the comparators; however, the data should be interpreted with caution due to the use of separate networks for each drug class rather than one connected network that included all drugs relevant to the decision-making comparator set, and the relatively poor quality of these ITCs.

The ITC that examined all-cause mortality among patients with T2DM treated with GLP-1 agonists, dipeptidyl-peptidase-4 inhibitors, or SGLT-2 inhibitors suggested that the risk of death among those who received semaglutide was similar to other therapies, with the exception of empagliflozin. These data should be interpreted with caution due to potential heterogeneity in trial and patient characteristics of the included studies that was not adequately explored in the ITC.

Cost and Cost-Effectiveness

Semaglutide is available as a solution in a pre-filled multi-dose disposable pen in 1.5 mL or 3 mL cartridges, equivalent to 2 mg or 4 mg semaglutide. The starting dose of semaglutide is 0.25 mg weekly; after four weeks the dose should be increased to 0.5 mg weekly. After an additional four weeks, the dose can be increased to 1 mg once weekly. At the manufacturer’s submitted price of $195.06 per pre-filled pen, the annual treatment cost per patient is $2,544.

The manufacturer submitted a cost-utility analysis over a 40-year time horizon from the perspective of a Canadian public health care payer. Analysis was conducted for two populations: those who do not achieve adequate glycemic control with metformin (second-line treatment) and those who do not achieve adequate glycemic control with metformin and sulfonylurea (third-line treatment). The second-line treatment analysis compared both semaglutide 0.5 mg and 1 mg weekly with weekly regimens (dulaglutide 1.5 mg weekly and exenatide extended-release 2 mg weekly) and daily regimens (liraglutide 1.2 mg , liraglutide 1.8 mg , lixisenatide 20 mcg, sitagliptin 100 mg, insulin glargine 0.53 IU/kg, canagliflozin 300 mg, dapagliflozin 10 mg, empagliflozin 25 mg, and glyburide 15 mg). For third-line treatment, the same comparators were considered, with the exception of glyburide. The model incorporated a variety of health states relating to the important micro and macrovascular complications associated with diabetes, the incidence of hypoglycemic events, and the associated impact of complications and events on mortality.

In the manufacturer’s analysis for second-line treatment, glyburide had the lowest costs and fewest quality-adjusted life-years (QALYs). Relative to glyburide, canagliflozin was the next most cost-effective treatment, followed by semaglutide 1 mg. The manufacturer reported the incremental cost per QALY gained (incremental cost-utility ratio [ICUR]) for canagliflozin versus glyburide was $10,827, and the ICUR for semaglutide 1 mg versus canagliflozin was $714,488. For third-line treatment, canagliflozin had the lowest costs. Relative to canagliflozin, semaglutide 1mg was the next most cost-effective treatment. The manufacturer reported that the ICUR for semaglutide 1 mg versus canagliflozin was $136,653.

CADTH identified the following key limitations:

• A major concern is the clinical data used to inform the economic model. A comprehensive network meta-analysis (NMA) incorporating all treatment options and all clinical outcomes is required to fully assess the relative effectiveness of each treatment option considered. This has led to the evidence for treatments being inconsistent. For example, instead of providing a single NMA comparing all second-line treatments, three separate NMAs are provided covering different groups of second-line treatment. However, for certain treatments, data from single clinical trials are used, whereas for other treatments, data from the NMA were used. Thus, a consistent base comparator is not available across these analyses and comparisons are de facto based on naive indirect comparisons, which are inappropriate.
• The CADTH clinical review team noted that the NMAs used to inform the economic analysis for third-line treatment were not representative of the patient population as the study population was not specific to patients who were not adequately controlled on metformin and SU, and were therefore inappropriate for this analysis.
• The submitted model was not transparent and there were conceptual problems in how diabetes progression was modelled. Verification of the model code was not possible and there were concerns given the inconsistency of results provided by the manufacturer.
• The manufacturer included a disutility associated with body mass index. CADTH noted that it is unclear whether utility decrements for weight gain derived from larger weight differences (13 kg to 30 kg) can be applied in a proportional manner to the smaller weight differences observed in the NMA of second-line therapies.

CADTH noted significant limitations with the submitted economic analysis that could not be corrected. As such, CADTH concluded that the manufacturer’s economic submission did not provide an appropriate basis to assess the cost-effectiveness of semaglutide. The manufacturer suggested that if semaglutide was lowered in price by 28%, its incremental cost-effectiveness ratio versus canagliflozin would be less than $50,000.