CADTH Reimbursement Review

Insulin icodec (Awiqli)

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.

Table 1: Background Information of Application Submitted for Review

NOC = Notice of Compliance; SC = subcutaneous.

Source: Insulin icodec product monograph. Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Introduction

Type 2 diabetes mellitus (T2DM) is a chronic health condition that develops when the body is no longer able to use insulin efficiently or produce enough insulin to manage blood glucose levels within a normal range.² This persistent hyperglycemia results in a constellation of symptoms and downstream impacts on the body.³ Diabetes Canada estimates that more than 4 million people living in Canada, representing approximately 10% of the population, live with diabetes mellitus in 2023, and that this will increase to more than 5 million (12%) by 2033.⁴ Approximately 90% of patients with diabetes specifically have T2DM.⁵ The prevalence of T2DM may be higher in racialized and minority groups^4,5 such as Indigenous Peoples in Canada⁶ and South Asian or Black populations,⁷ compared to white populations. Indigenous Peoples in Canada are also at higher risk for diabetes-related complications.^6,8

While T2DM does not directly result in death, it is associated with increased mortality from downstream ischemic heart disease, stroke, chronic liver disease, infections, and cancer of the liver, pancreas, breast, and female reproductive system.⁹ In Canada, diabetes mellitus is associated with a 20-fold increased risk of hospitalization for nontraumatic lower limb amputations, with an estimated 65% of lower limb amputations due to diabetes.¹⁰ Patients with T2DM may have lowered health-related quality of life (HRQoL) and social stigma associated with their condition.^11-13

The main goals of treatment for patients with T2DM are to reduce the risk of long-term complications through control of glycemia and blood pressure, and cardiovascular (CV) risk reduction through control of lipids and hypertension.¹⁴ Management of T2DM is individualized and ideally combines lifestyle modifications (e.g., dietary modification, exercise, quitting smoking) with pharmacological interventions.¹⁴ As T2DM is a progressive condition, most patients will require pharmacological intervention eventually, and most commonly are first prescribed metformin. If a patient is unable to lower or maintain their glycated hemoglobin (hemoglobin A1C) or blood glucose levels with metformin treatment alone, additional therapies may be combined with continued metformin therapy, such as sulfonylureas, dipeptidyl peptidase-4 inhibitors (DPP-4is), sodium-glucose cotransporter-2 inhibitors (SGLT2is), glucagon-like peptide-1 receptor agonists (GLP-1 RAs), and insulin.¹⁴

According to Diabetes Canada treatment guidelines, insulin (in combination with metformin) should be initiated in a patient in whom fasting blood glucose levels and/or hemoglobin A1C are not at target on current antihyperglycemic drugs, or with symptomatic hyperglycemia and/or metabolic decompensation.¹⁴ Basal insulin should be initiated and titrated to achieve fasting blood glucose targets, and metformin should be continued unless contraindicated.¹⁴ Other antihyperglycemic drugs may also be used in combination with these therapies as needed, and therapy should be advanced if the patient’s hemoglobin A1C is not at target within 3 months to 6 months despite adequate titration of basal insulin and supports for lifestyle and other pharmacotherapeutic interventions.¹⁴ Basal insulins for treatment of T2DM can include long-acting or intermediate-acting insulins. Currently available long-acting insulins include insulin degludec (U100 or U200), insulin glargine (U100 or U300), and insulin detemir, while neutral protamine Hagedorn insulin is an intermediate-acting insulin. Insulin and its analogues work to lower blood glucose by stimulating peripheral glucose uptake and by inhibiting hepatic glucose production.¹⁵

Insulin icodec is a long-acting insulin which is administered subcutaneously on a once-weekly basis, in contrast to the currently available once-daily long-acting basal insulins (insulin degludec and insulin glargine).¹⁶ The Health Canada–approved indication is for the once-weekly treatment of adults with diabetes mellitus to improve glycemic control, and the reimbursement request is for adult patients with T2DM.

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of insulin icodec once-weekly subcutaneous (SC) injection in the treatment of T2DM in adults.

Stakeholder Perspectives

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

Patient Input

Patient input was submitted for this CADTH review by Diabetes Canada, which fielded a self-directed questionnaire to people living with T2DM and their caregivers across Canada between October 3, 2023, and October 23, 2023, inquiring about respondents’ lived experiences with diabetes and with several questions pertaining to insulin icodec. Of the 21 respondents, 13 identified as living with T2DM and 1 identified as a caregiver, 93% (of 14 respondents for the question) were aged older than 55 years, 35% were aged 75 years to 84 years, and 71% reported living with T2DM for more than 10 years (of which 29% reported living with T2DM for more than 20 years).

Most respondents indicated that living with T2DM was preoccupying, inconvenient, and burdensome, with constant management requiring foresight and planning. A total of 24% (out of 20 respondents) reported experiencing hyperglycemia more than once per day and 10% reported experiencing it more than once per week. A total of 43% (out of 20 respondents) indicated they did not experience hypoglycemia or experienced it in the past but not currently, while 14% experienced it more than once per week; none reported experiencing hypoglycemia daily.

All respondents who provided data on diabetes-related drug use (n = 19) reported taking antihyperglycemic medication including long-, short-, and rapid-acting insulin, insulin icodec, and other noninsulin antihyperglycemic drugs, either as single-drug products or combined with metformin. A total of 5 of 18 (28%) respondents reported current insulin icodec use. A total of 61% (out of 18) respondents said they were very satisfied or satisfied with their medication; no respondents indicated dissatisfaction. Respondents indicated that ease of use, lack of side effects, and helping to lower hemoglobin A1C were aspects they liked about their medications.

When choosing a medication for diabetes management, several considerations were important to respondents including avoiding hypoglycemia and hyperglycemia, reducing the risk of heart problems, reducing high blood pressure, maintaining satisfactory blood glucose levels throughout the day, and avoiding yeast infections, urinary tract infections, fluid retention, or weight gain. Affordability was also highlighted as an important consideration. Improvements that respondents wished to see in a new treatment that are not currently being achieved with available therapies included fewer side effects, blood flow improvement to extremities, weight control, and better hemoglobin A1C results.

Clinician Input

Input From Clinical Expert Consulted by CADTH

The clinical expert consulted by CADTH indicated that treatment of T2DM must be individualized, provided in a culturally appropriate manner, and equitably and affordably accessible across Canada. Treatment goals for patients with T2DM include reducing symptoms of hyperglycemia, reducing risk of long-term complications through control of glycemia and blood pressure, and reducing CV risk through control of lipids and hypertension, through a combination of lifestyle modifications and pharmacotherapeutic approaches. The key unmet need highlighted by the clinical expert was a lack of widespread access to primary care and therefore lack of access to diabetes prevention, detection, and treatment. In particular, access to diabetes education and specialist care varies greatly across the country.

The clinical expert highlighted that insulin icodec would fit into the current paradigm for introduction of basal insulin in the management of T2DM. This includes patients who are not meeting glycemic targets despite lifestyle modification, and use of or intolerance or contraindication to metformin, GLP-1 RAs, and/or SGLT2 inhibitors. It may also be used as a first-line therapy in patients with T2DM who present with symptomatic hyperglycemia and/or metabolic decompensation with or without metformin. The clinical expert noted that insulin icodec may be preferred over daily basal insulins by some patients who are unable or unwilling to take daily basal insulin, or who would prefer a lower burden related to administration frequency.

The clinical expert consulted by CADTH noted that the timing of assessments varies substantially between physicians and between patients, but ideally a patient would be supported through phone and email to adjust dosing during the first 2 months to 3 months of insulin therapy, followed by an assessment of treatment suitability after 3 months to 6 months of therapy. Diabetes management is complex and individualized. As such, there are several factors a monitoring physician or nurse practitioner will assess with regard to insulin icodec, including: treatment acceptance, treatment adherence, hemoglobin A1C target achievement, time in range with continuous glucose monitoring (CGM) of more than 70%, time below range less than 4%, and no severe hypoglycemic episodes. A sign of positive response to insulin icodec would also be improved HRQoL, including but not limited to less diabetes distress and more treatment satisfaction. Factors influencing a decision to discontinue insulin icodec would include allergy, nonadherence, or diabetes remission or glycemic control improvement through weight loss or use of other antihyperglycemic drugs or bariatric surgery. The clinical expert stated that diagnosis, prescribing of therapies, and management of treatment for patients with T2DM may occur in primary care. The diagnosis of T2DM and the use of insulin icodec were described by the clinical expert to be uncomplicated and do not necessarily require specialist care. Additionally, there are limitations to access of specialist care due to the low number of endocrinologists in Canada and the high number of patients with T2DM.

Clinician Group Input

No clinician group feedback was received by the deadline for the call for input.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH reimbursement review process. The following were identified as key factors that could potentially impact the implementation of a CADTH recommendation for insulin icodec:

• consideration for initiation of therapy

• consideration for prescribing of therapy

• care provision issues

• system and economic issues.

The clinical expert consulted by CADTH provided advice on the potential implementation issues raised by the drug programs (refer to Table 5).

Clinical Evidence

Systematic Review

Description of Studies

Five active-controlled, multicentre randomized controlled trials (RCTs) were included in this review, all of which compared insulin icodec once weekly to once-daily basal insulins (insulin degludec and/or insulin glargine). Three of the included studies enrolled adult patients with T2DM who were insulin naive (ONWARDS 1, 3, and 5); of these, the ONWARDS 1 study was 52 weeks in duration, while the ONWARDS 3 and ONWARDS 5 studies were 26 weeks in duration. The ONWARDS 5 study additionally included the DoseGuide app to inform dosing choices in the insulin icodec arm. The remaining 2 included studies enrolled adult patients with T2DM who were insulin experienced. In the ONWARDS 2 study, patients were experienced with basal insulin, and in the ONWARDS 4 study, patients were experienced with basal and bolus insulin. Both of these studies were 26 weeks in duration. The ONWARDS 4 study additionally included insulin aspart (bolus) administered 2 times to 4 times per day before mealtimes. In all 5 studies, the primary outcome was an assessment of noninferiority of insulin icodec once weekly compared to the once-daily comparator for the outcome of change in hemoglobin A1C from baseline. In all but the ONWARDS 4 study, a secondary confirmatory analysis of superiority was also conducted for this outcome. Secondary outcomes varied between the studies and included percent time in range (3.9 mmol/L to 10.0 mmol/L blood glucose using CGM), time below range (< 3 mmol/L blood glucose using CGM), and time above range (> 10.0 mmol/L blood glucose using CGM) in the ONWARDS 1, 2, and 4 studies; change in body weight in all included studies; the number of clinically significant (level 2; < 3.0 mmol/L confirmed by blood glucose meter) or severe hypoglycemic episodes (level 3; any hypoglycemic event requiring active assistance of another person, for instance to administer corrective actions or receive medical care) in all studies; change in treatment satisfaction (measured via Diabetes Treatment Satisfaction Questionnaire [DTSQ] in which higher scores [ranging from 0 to 36] represent higher satisfaction with diabetes treatment) in the ONWARDS 5 and ONWARDS 2 studies (exploratory); and treatment compliance (measured via the Treatment-Related Impact Measure for Diabetes [TRIM-D] compliance domain, in which higher scores [ranging from 4 to 20] represent higher compliance with treatment) in the ONWARDS 5 study. All-cause mortality was a safety outcome in all trials. Additional outcomes of interest that were not reported include the long-term efficacy regarding CV death, nonfatal myocardial infarction (MI), nonfatal stroke, and other microvascular or macrovascular complications of T2DM.

At baseline, patients in the ONWARDS studies treatment arms had mean ages ranging from 58 years to 62 years, and 53% to 63% were male while 37% to 47% were female. Across all studies, the majority of patients were white (60% to 90%) followed by Asian (4% to 42%), Black or African American (2% to 5%), Other (< 1% to 4%), American Indian or Alaska Native (0 to < 1%), and Native Hawaiian or Other Pacific Islander (0 to < 1%). In the insulin-naive populations of the ONWARDS 1, 3, and 5 studies, the mean duration of diabetes was 11 years to 12 years and the mean hemoglobin A1C was 8.44% to 8.88% at baseline. In the insulin-experienced populations of the ONWARDS 2 and 4 studies, the mean duration of diabetes was 16 years to 18 years, and the mean hemoglobin A1C was 8.17% to 8.31% at baseline. Approximately 90% of patients were receiving metformin at baseline, and other common (> 15%) antihyperglycemic background medications included sulfonylureas, SGLT2is, DPP-4is, and GLP-1 RAs. Uncommon antihyperglycemic background medications included thiazolidinediones, alpha-glucosidase inhibitors, and glinides. In the ONWARDS 2 and 4 studies, insulin glargine U100 (approximately 41% to 50%) followed by insulin degludec (approximately 23% to 29%) were the most common basal insulins in use at screening; most patients in the ONWARDS 4 study were receiving basal insulin once daily and bolus insulin 3 times daily (approximately 75%).

Efficacy Results

Change in Hemoglobin A1C From Baseline

In the primary analyses for noninferiority among patients who were insulin naive (the ONWARDS 1, ONWARDS 3, and ONWARDS 5 studies), the between-group differences in mean change from baseline in hemoglobin A1C were –0.19% points (95% confidence interval [CI], –0.36% to –0.03%; P < 0.0001) in the ONWARDS 1 study at 52 weeks, –0.21% points (95% CI, –0.34% to –0.08%; P < 0.0001) in the ONWARDS 3 study at 26 weeks, and –0.38% points (95% CI, –0.66% to –0.09%; P < 0.0001) in the ONWARDS 5 study at 52 weeks, indicating that insulin icodec once weekly is noninferior to the once-daily comparator for the outcome of change in hemoglobin A1C from baseline in patients who were insulin naive. In the secondary analyses for superiority, the P values were 0.0210, 0.0016, and 0.0092, respectively, indicating that insulin icodec once weekly is superior to the once-daily comparator for the outcome of change in hemoglobin A1C from baseline in patients who were insulin naive.

In the primary analyses for noninferiority among patients who were insulin experienced (ONWARDS 2 and ONWARDS 4), the between-group differences in mean change from baseline in hemoglobin A1C was –0.22% points (95% CI, –0.37% to –0.08%; P < 0.0001) in the ONWARDS 2 study and 0.02% points (95% CI, –0.11% to 0.15%; P < 0.0001) for the ONWARDS 4 study, indicating that insulin icodec once weekly is noninferior to the once-daily comparator for the outcome of change in hemoglobin A1C from baseline in patients who were insulin experienced. In the ONWARDS 2 study, a secondary analysis for superiority was also conducted (P = 0.0028), indicating that insulin icodec once weekly is noninferior to the once-daily insulin glargine for the outcome of change in hemoglobin A1C from baseline in patients who were insulin experienced. No superiority analysis was conducted in the ONWARDS 4 study.

For each of the ONWARDS 1, 3, 5, 2, and 4 studies, a 2-dimensional tipping point sensitivity analysis was performed to evaluate the robustness of the assumptions regarding missing data; the results were consistent with the primary analysis for noninferiority of hemoglobin A1C.

Time in Range (3.9 mmol/L to 10.0 mmol/L)

In the ONWARDS 1 study from week 48 to week 52, the least squares (LS) mean time in glycemic range was 71.27% (standard error [SE] = 0.85) for insulin icodec and 67.00% (SE = 0.85) for insulin glargine, representing an estimated treatment difference of 4.27% points (95% CI, 1.92% to 6.62%; P = 0.0004). A 2-dimensional tipping point sensitivity analysis was conducted, which aligned with the primary analysis for time in range. This outcome was not assessed in the ONWARDS 3 or 5 studies.

In the ONWARDS 2 study from week 22 to week 26, the LS mean time in glycemic range between 3.9 mmol/L and 10.0 mmol/L was 62.34% (SE = 1.16) for the insulin icodec group and 59.93% (SE = 1.16) for the insulin degludec group. The estimated treatment difference between insulin icodec and insulin degludec was 2.41% (95% CI, –0.84% to 5.56%; P = 0.1461). In the ONWARDS 4 study from week 22 to week 26, the LS mean time in glycemic range between 3.9 mmol/L and 10.0 mmol/L was 66.75% (SE = 1.00) for the insulin icodec group and 66.46% (SE = 1.02) for the insulin glargine group. The estimated treatment difference between insulin icodec and insulin glargine was 0.29% (95% CI, –2.52% to 3.09%; P = 0.8406).

Time Spent at Less Than 3.0 mmol/L

In the ONWARDS 1 study from week 48 to week 52, the LS mean time in glycemic range below 3 mmol/L was 0.21% (SE = not reported) for insulin icodec and 0.16% for insulin glargine (SE = not reported), representing an estimated treatment ratio (insulin icodec:insulin glargine) of 1.27 (95% CI, 0.94 to 1.71; P = 0.1134). This outcome was not assessed in the ONWARDS 3 or 5 studies.

In the ONWARDS 2 study from week 22 to week 26, the LS mean time in glycemic range below 3.0 mmol/L was 0.33% in the insulin icodec group and 0.24% in the insulin degludec group. The estimated treatment ratio of insulin icodec and insulin degludec was 1.37 (95% CI, 0.92 to 2.04; P = 0.1180). In the ONWARDS 4 study from week 22 to week 26, the LS mean time in glycemic range below 3.0 mmol/L was 0.69% in the insulin icodec group and 0.58% in the insulin glargine group. The estimated treatment ratio of insulin icodec and insulin glargine was 1.20 (95% CI, 0.91 to 1.58; P = 0.2050).

Time Spent at Greater Than 10.0 mmol/L

In the ONWARDS 1 study from week 48 to week 52, the LS mean time in glycemic range above 10 mmol/L was 27.56% for insulin icodec and 32.13% for insulin glargine, representing an estimated treatment difference of –4.58% (95% CI, –6.99% to –2.17%; P = 0.0002). This outcome was not assessed in the ONWARDS 3 or 5 studies.

In the ONWARDS 2 study from week 22 to week 26, the LS mean time in glycemic range above 10 mmol/L was 36.34% (SE = 1.19) for insulin icodec and 39.28% (SE = 1.19) for insulin degludec, representing an estimated treatment difference of –2.93% (95% CI, –6.25% to 0.39%; P = 0.0833). In the ONWARDS 4 study from week 22 to week 26, the LS mean time in glycemic range above 10 mmol/L was 30.64% for insulin icodec (SE = 1.03) and 31.24% (SE = 1.04) for insulin glargine, representing an estimated treatment difference of –0.60% (95% CI, –3.47% to 2.28%; P = 0.6826).

Change in Body Weight

In the ONWARDS 1 (at 52 weeks), ONWARDS 3 (at 26 weeks), and ONWARDS 5 (at 52 weeks) studies, the between-group differences in change in body weight from baseline were 0.46 kg (95% CI, –0.12 kg to 1.04 kg; P = 0.1187), 0.46 kg (95% CI, –0.19 kg to 1.10 kg; P = 0.1657), and 0.83 kg (95% CI, –0.37 kg to 2.02 kg; P = 0.1747), respectively.

In the ONWARDS 2 (at 26 weeks) and ONWARDS 4 (at 26 weeks) studies, the between-group differences in change in body weight from baseline were 1.70 kg (95% CI, 0.76 kg to 2.63 kg; P = 0.0004) and 0.57 kg (95% CI, –0.39 kg to 1.54 kg; P = 0.2444), respectively.

Number of Level 2 or Level 3 Hypoglycemic Episodes

In the ONWARDS 1 study, a similar number of patients experienced level 2 hypoglycemic events in the 2 groups, but there were numerically more level 2 events in the insulin icodec group. There were 143 events in 48 patients (9.8%) in the insulin icodec group and 75 events occurring in 49 patients (10.0%) in the insulin glargine group. In the insulin icodec group, 3 of the 492 patients (0.6%) experienced 61 of the 143 clinically significant hypoglycemic events. The remaining patients in the insulin icodec group and all of the patients in the insulin glargine groups experienced between 1 episode and 5 episodes of level 2 hypoglycemic events. The estimated treatment ratio for level 2 events (insulin icodec:insulin glargine) was 1.67 (95% CI, 0.99 to 2.84; P = 0.0561). Severe (level 3) hypoglycemic events occurred in 1 patient (0.2%) in the insulin icodec group, and 3 patients (0.6%) in the insulin glargine group.

In the ONWARDS 3 study, there were 53 clinically significant events of hypoglycemia (level 2) reported in 26 patients (8.9%) in the insulin icodec group, and 23 events occurring in 17 patients (22.1%) in the insulin degludec group. In the insulin icodec group, 2 patients (0.7%) experienced 15 of the 53 clinically significant hypoglycemic events. The remaining patients in the insulin icodec group experienced between 1 episode and 4 episodes of level 2 hypoglycemic events. Patients in the insulin degludec group experienced between 1 episode and 3 episodes of level 2 hypoglycemic events. The estimated treatment ratio for level 2 events (insulin icodec:insulin degludec) was 2.09 (95% CI, 0.99 to 4.41; P = 0.0536). Severe hypoglycemic events occurred in zero patients in the insulin icodec group and 2 patients (0.7%) in the insulin degludec group.

In the ONWARDS 5 study, there were 104 clinically significant (level 2) hypoglycemic events that were reported in 64 patients (11.8%) in the insulin icodec group, and 81 events occurring in 45 patients (8.4%) in the once-daily analogues group. The estimated treatment ratio for level 2 events (insulin icodec:once-daily analogues) was 1.23 (95% CI, 0.77 to 1.98; P = 0.3928). Severe (level 3) hypoglycemic events occurred in zero patients in the insulin icodec group and 4 patients (0.7%) in the insulin glargine group.

In the ONWARDS 2 study, there were 113 clinically significant events of hypoglycemia (level 2) reported in 37 patients (14.1%) in the insulin icodec group, and 41 events occurring in 19 patients (7.2%) in the insulin degludec group. The estimated treatment ratio for level 2 events (insulin icodec:insulin degludec) was 1.98 (95% CI, 0.95 to 4.12; P = 0.0677). Severe hypoglycemic events occurred in zero patients in the insulin icodec group and 1 patient (0.4%) in the insulin degludec group.

In the ONWARDS 4 study, clinically significant events of hypoglycemia (level 2) were reported in 148 patients (50.9%) in the insulin icodec group, and 160 patients (55.0%) in the insulin glargine group. The estimated treatment ratio for level 2 events (insulin icodec:insulin glargine) was 0.99 (95% CI, 0.73 to 1.34; P = 0.9274). Severe (level 3) hypoglycemic events occurred in 4 patients (1.4%) in the insulin icodec group and 2 patients (0.7%) in the insulin glargine group.

Diabetes Treatment Satisfaction Questionnaire

This outcome was assessed only in the ONWARDS 5 and ONWARDS 2 studies.

In the ONWARDS 5 study, the observed mean DTSQ total score at baseline was 26.15 in the insulin icodec plus DoseGuide group and 26.77 in the once-daily analogues group. The estimated LS mean DTSQ total score at week 52 was 31.13 (SE = 0.25) in the insulin icodec group and 30.35 (SE = 0.25) in the once-daily analogues group, representing a LS mean change from baseline in DTSQ total satisfaction score of 4.68 (SE = 0.25) and 3.90 (SE = 0.25), respectively. The LS mean difference between groups was 0.78 (95% CI, 0.10 to 1.47; P = 0.0247).

In the ONWARDS 2 study, the observed mean DTSQ total score at baseline was 26.76 in the insulin icodec group and 26.69 in the insulin degludec group. The estimated LS mean DTSQ total score at week 26 was 30.95 (SE = 0.30) in the insulin icodec group and 29.69 (SE = 0.31) in the insulin degludec group, representing a LS mean change from baseline in DTSQ total satisfaction score of 4.22 (SE = 0.30) and 2.96 (SE = 0.31), respectively. The estimated LS mean treatment difference between insulin icodec and insulin degludec was 1.25 (95% CI, 0.41 to 2.10; P = 0.0036).

TRIM-D Compliance Domain

This outcome was assessed only in the ONWARDS 5 study. The estimated treatment difference was 3.04 (95% CI, 1.28 to 4.81; P = 0.0007) at 52 weeks.

CV Death

CV death was not measured as an outcome in the included trials.

Nonfatal MI

Nonfatal MI was not measured as an outcome in the included trials.

Nonfatal Stroke

Nonfatal stroke was not measured as an outcome in the included trials.

Other Microvascular and Macrovascular Complications of T2DM

Other microvascular and macrovascular complications of T2DM were not measured as outcomes in the included trials.

Harms Results

Adverse Events

The proportion of patients who had adverse events (AEs) was similar between the insulin icodec and once-daily insulin analogue comparator groups in all ONWARDS studies. The most common AEs were COVID-19, nasopharyngitis, diarrhea, and back pain. The majority of AEs were determined by the study investigators to be nonserious, mild to moderate in severity, unlikely related to trial products, and recovered or recovering by the end of the trial duration in each trial.

In the insulin-naive populations (ONWARDS 1, 3, and 5), 50% to 71% of patients across each treatment arm experienced at least 1 AE.

In the insulin-experienced populations (ONWARDS 2 and 4), 51% to 62% of patients across each treatment arm experienced at least 1 AE.

Serious Adverse Events

Serious AEs (SAEs) occurred in similar proportions across both the insulin icodec groups and the once-daily analogues groups in each trial.

In the insulin-naive populations (ONWARDS 1, 3, and 5), among patients treated with insulin icodec, 5.1% to 10.4% of patients had at least 1 SAE. In these same trials, among patients treated with the comparator once-daily analogues, 5.1% to 10.6% of patients had at least 1 SAE.

In the insulin-experienced populations (ONWARDS 2 and 4), among patients treated with insulin icodec or once-daily analogues, 7.6% to 8.4% and 6.1% to 8.6% had at least 1 SAE, respectively.

Reported SAEs included: cardiac disorders; infections and infestations; injury, poisoning, and procedural; vascular disorders; musculoskeletal and connective tissue disorders; neoplasms benign, malignant and unspecified nervous system disorders; eye disorders; reproductive system and breast disorders; respiratory, thoracic, and mediastinal disorders; blood and lymphatic system disorders; congenital, familial, and genetic disorders; gastrointestinal disorders; metabolism and nutrition disorders; and renal and urinary disorders. Each of these SAE categories occurred in 0% to less than 5% of patients. The most frequent category of SAE observed was cardiac disorders, which ranged from approximately 3% to 4% of patients in the included studies, followed by infections and infestations (in approximately 2% to 3% of patients). There was no 1 most common event observed.

Withdrawals Due to AEs

In the insulin-naive populations (ONWARDS 1, 3, and 5), permanent discontinuation of the study drug due to AE occurred in 0.7% to 1.2% of patients treated with insulin icodec, and 0.8% to 1.3% of patients treated with comparators (insulin glargine or insulin degludec). In the insulin-experienced populations (ONWARDS 2 and 4), permanent discontinuation of the study drug due to AE occurred in 1.0% to 1.9% of patients treated with insulin icodec, and 1.0% to 1.1% of patients treated with insulin degludec or insulin glargine. Temporary discontinuation was similarly uncommon, as were AEs leading to dose increases or dose decreases.

Mortality

In the ONWARDS 1 study, there were 6 (0.6%) patients with fatal outcomes, of which 4 (0.8%) patients died in the insulin icodec treatment group and 2 (0.4%) patients died in the insulin glargine treatment group. The events (of which some patients may have had multiple) included infections and infestations (n = 2) and 1 each of COVID-19, cardiac disorders (angina pectoris), postoperative infection, pancreatic neoplasm, glioblastoma, unknown cause, and acute coronary syndrome. The death due to unknown cause in the insulin glargine treatment group was judged by investigators as “possibly” related to the trial product.

In the ONWARDS 3 study, there were 2 patients (0.7%) in the insulin icodec group and 1 patient (0.3%) in the insulin degludec group with fatal outcomes. In the insulin icodec group, deaths were due to malignancy and an undetermined cause (n = 1 for each). In the insulin degludec group, death was due to acute MI (n = 1).

In the ONWARDS 5 study, there were 3 patients (0.6%) who died in the insulin icodec plus DoseGuide group, and 7 patients (1.3%) in the once-daily analogues group. In the insulin icodec plus DoseGuide group, deaths were due to an undetermined cause (n = 2) and malignancy (n = 1). In the once-daily analogues group, deaths were due to pulmonary causes (n = 2), undetermined causes (n = 2), malignancy, sudden cardiac death, and heart failure (n = 1 for each).

In the ONWARDS 2 study, there were 2 patients (0.8%) who died in the insulin icodec group and 2 patients (0.8%) who died in the insulin degludec group. In the insulin icodec group, deaths were due to infection (n = 2). In the insulin degludec group, deaths were due to CV procedures and malignancy (n = 1 for each).

In the ONWARDS 4 study, there were 2 patients (0.7%) in the insulin icodec group and 1 patient (0.3%) in the insulin glargine group who died. In the insulin icodec group, deaths were due to other CV causes and infection (including sepsis) (n = 1 for each). In the insulin glargine group, there was 1 instance of gastrointestinal bleeding which resulted in death.

Notable Harms

Prespecified notable harms included hypersensitivity, injection site reactions, hypoglycemia, and nocturnal hypoglycemia.

Events of hypersensitivity were reported among less than 7% patients during all ONWARDS studies and were similar between treatment groups in each trial. Serious events were rare.

Injection site reactions occurred among less than 9% of patients across all ONWARDS studies. In the ONWARDS 1 study, 6 patients (1.2%) experienced 6 events in the insulin icodec group compared to 12 patients (2.4%) experiencing 12 events in the insulin glargine group. All events were considered mild or moderate in severity. In the ONWARDS 3 study, 25 patients (8.5%) experienced 62 events in the insulin icodec group compared to 13 patients (4.4%) who experienced 22 events in the insulin degludec group. Of the 62 injection site reactions reported in the insulin icodec group, 24 events were reported by only 2 patients. No events were considered serious. In the ONWARDS 5 study, 5 patients (0.9%) experienced 6 events in the insulin icodec plus DoseGuide group compared to 7 patients (1.3%) who experienced 28 events in the once-daily insulin analogue group. No events of injection site reactions were considered serious. In the ONWARDS 2 study, 3 patients (1.1%) experienced 3 events in the insulin icodec group compared to 1 patient (0.4%) who experienced 1 event in the insulin degludec group. All events of injection site reactions were considered mild or moderate in severity. In the ONWARDS 4 study, 2 patients (0.7%) experienced 2 events in both the insulin icodec and insulin glargine groups. No events of injection site reactions were considered serious, and all were mild in severity.

Nocturnal Hypoglycemia

In the ONWARDS 1 study, level 1 nocturnal hypoglycemic events occurred in 67 patients (13.6%) in the insulin icodec group and 58 patients (11.8%) in the insulin glargine group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 9 patients (1.8%) in the insulin icodec group and 10 patients (2.0%) in the insulin glargine group, and severe (level 3) nocturnal hypoglycemic events occurred in zero patients and 1 patient (0.2%) in the insulin icodec and insulin glargine groups, respectively. The estimated treatment ratio between insulin icodec and insulin glargine for clinically significant (level 2) nocturnal hypoglycemic events was 0.92 (95% CI, 0.29 to 2.86; P = 0.8816). The estimated treatment ratio between insulin icodec and insulin glargine for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 0.88 (95% CI, 0.29 to 2.64; P = 0.8189).

In the ONWARDS 3 study, level 1 nocturnal hypoglycemic events occurred in 24 patients (8.2%) in the insulin icodec group and 23 patients (7.8%) in the insulin degludec group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 1 patient (0.3%) in the insulin icodec group and 4 patients (1.4%) in the insulin degludec group. There were no severe (level 3) nocturnal hypoglycemic events in either treatment group. The estimated treatment ratio between insulin icodec and insulin degludec for clinically significant (level 2) nocturnal hypoglycemic events was 2.09 (95% CI, 0.99 to 4.41; P = 0.0536). The estimated treatment ratio between insulin icodec and insulin degludec for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 1.82 (95% CI, 0.87 to 3.80; P = 0.1091).

In the ONWARDS 5 study, level 1 nocturnal hypoglycemic events occurred in 48 patients (8.9%) in the insulin icodec group and 46 patients (8.6%) in the once-daily analogues group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 11 patients (2.0%) in both treatment groups, and severe (level 3) nocturnal hypoglycemic events occurred in zero patients in the insulin icodec group and 1 patient (0.2%) in the once-daily analogues groups. The estimated treatment ratio between insulin icodec and once-daily analogues for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 1.36 (95% CI, 0.82 to 2.27; P = 0.02396).

In the ONWARDS 2 study, level 1 nocturnal hypoglycemic events occurred in 60 patients (22.9%) in the insulin icodec group and 35 patients (13.3%) in the insulin degludec group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 16 patients (6.1%) in the insulin icodec group and 9 patients (3.4%) in the insulin degludec group. Severe (level 3) nocturnal hypoglycemic events occurred in zero patients in both treatment groups. The estimated treatment ratio between insulin icodec and insulin degludec for clinically significant (level 2) nocturnal hypoglycemic events was 1.98 (95% CI, 0.95 to 4.12; P = 0.0677). The estimated treatment ratio between insulin icodec and insulin degludec for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 1.93 (95% CI, 0.93 to 4.02; P = 0.0782).

In the ONWARDS 4 study, level 1 nocturnal hypoglycemic events occurred in 108 patients (37.1%) in the insulin icodec group and 132 patients (45.4%) in the insulin glargine group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 54 patients (18.6%) in the insulin icodec group and 71 patients (24.4%) in the insulin glargine group, and severe (level 3) nocturnal hypoglycemic events occurred in zero patients and 1 patient (0.3%) in the insulin icodec and insulin glargine groups, respectively. The estimated treatment ratio between insulin icodec and insulin glargine for clinically significant (level 2) nocturnal hypoglycemic events was 0.74 (95% CI, 0.47 to 1.15; P = 0.1818). The estimated treatment ratio between insulin icodec and insulin glargine for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 0.73 (95% CI, 0.47 to 1.14; P = 0.1694).

Critical Appraisal

All of the ONWARDS trials were randomized, active-controlled trials with adequate methodology related to randomization and allocation concealment, and there were no concerning between-arm imbalances in patient characteristics at baseline, nor in diabetes-related background medications. As such, the risk of bias arising from the randomization process is low in all trials. Each trial was adequately powered for the purpose of their primary hypotheses. The ONWARDS 1, ONWARDS 2, ONWARDS 4, and ONWARDS 5 studies were open-label trials, which is associated with a risk of bias in subjective and self-reported outcomes, while the ONWARDS 3 study was double-blinded with adequate blinding and concealment procedures including placebos matched in visual quality and administration methods to the active trial products.

The primary outcome in each trial was the change in hemoglobin A1C from baseline, and the noninferiority margin of 0.3% points was chosen based on established FDA guidance¹⁷ and previous trials of insulin products in the treatment of T2DM. Change in hemoglobin A1C from baseline was considered a clinically relevant outcome by the clinical expert consulted by CADTH. This outcome is considered acceptable by the FDA¹⁷ for trials of new antihyperglycemic therapies seeking a glycemic control indication, with the rationale being that it is a validated surrogate of microvascular disease risk reduction,¹⁷ and further it is currently recognized as the key surrogate marker for the development of long-term diabetes complications in people with type 1 diabetes mellitus or T2DM.¹⁸ The selection of this noninferiority margin was determined based on FDA guidance¹⁷ as previously described, and was considered clinically relevant as a threshold of minimal important difference (MID) according to the clinical expert consulted by CADTH. However, hemoglobin A1C is ultimately a surrogate biomarker, and there is evidence to suggest that hemoglobin A1C may not be appropriate as a surrogate outcome for downstream complications in diabetes mellitus trials due to poor associations with mortality,^19,20 CV mortality,²¹ MI,¹⁹ heart failure,¹⁹ kidney injury,¹⁹ and stroke.¹⁹ Other limitations of hemoglobin A1C include a lack of information about acute glycemic events (i.e., hypo- or hyperglycemia) and insensitivity regarding day-to-day variations of glucose. As well, measurement of hemoglobin A1C can be confounded by other conditions such as anemia, hemoglobinopathies, iron deficiency, and pregnancy.¹⁸

Use of CGM allows for observation of time in and outside of range and daily glycemic variability,¹⁸ and the clinical expert consulted by CADTH indicated that this is of growing importance in clinical trials of glycemic control in patients with T2DM in addition to hemoglobin A1C.¹⁸ Time in range as measured by CGM is useful as a measure of short-term glycemic control,²² and there is good correlation between time in range and hemoglobin A1C.²² Time in range has been demonstrated to be associated with diabetic retinopathy^23,24 and microalbuminuria²³ but publications assessing this outcome as a surrogate for other diabetes-related complications (e.g., mortality, MI, and other major CV or renal events) were not identified.

The primary outcome in all trials was adjusted for multiple comparisons. Additionally, in the ONWARDS 1 study, the outcome of time in range (3.9 mmol/L to 10 mmol/L) was also adjusted for multiple comparisons. As the remaining outcomes were not adjusted for multiplicity, there is an increased risk of type I error (i.e., false-positive results) for statistically significant results for those outcomes.

Multiple imputation was used for all outcomes to account for missing data. Multiple imputation methods will not remove or reduce bias that occurs when missingness is not random, but the proportion of missing data in each case was low, so this was not considered cause for concern. Additionally, sensitivity analyses were conducted for the primary outcome which bolstered confidence in the primary analyses.

The study designs with respect to patient eligibility criteria and characteristics at baseline were appropriately reflective of the target population in Canada, with the exception that there is a notable lack of inclusion of Indigenous Peoples, who are at higher risk of developing T2DM and its complications.^6,8 The selected comparators, medications at baseline among included patients, and concomitant mediations during the trials were considered by the consulted clinical expert to be appropriate and to reflect clinical practice in Canada.

The impact of insulin icodec on patients’ HRQoL was not measured in the ONWARDS trials. Although the DTSQ and TRIM-D compliance domain provide information about treatment satisfaction and compliance, they are not comprehensive measures of HRQoL. As such, the influence of insulin icodec on HRQoL as compared with insulin degludec or insulin glargine is not known. Additionally, there were no compliance data reported for the insulin-experienced populations.

There is a data gap regarding the long-term effect of insulin icodec versus daily insulins on outcomes such as CV death, nonfatal MI, nonfatal stroke, and long-term all-cause mortality beyond the duration of the included clinical trials. Additionally, the clinical trials did not evaluate any global HRQoL measures.

GRADE Summary of Findings and Certainty of the Evidence

The selection of outcomes for Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) assessment was based on the sponsor’s Summary of Clinical Evidence, consultation with clinical experts, and input received from patient and clinician groups and public drug plans. The following list of outcomes was finalized in consultation with expert committee members:

• outcomes related to blood glucose (percent change in hemoglobin A1C from baseline, percent time in range [3.9 mmol/L to 10.0 mmol/L], percent time spent below range, percent time spent above range)

• mortality and morbidity (all-cause mortality, CV death, nonfatal MI, nonfatal stroke, microvascular and macrovascular complications of T2DM)

• change in body weight from baseline

• treatment satisfaction (DTSQ) and compliance (TRIM-D compliance domain)

• proportion of patients with clinically significant or severe hypoglycemic events.

Table 2: Summary of Findings for Insulin Icodec Versus Daily Basal Insulins^a for Patients With T2DM — Patients Who Were Insulin Naive

CI = confidence interval; DTSQ = Diabetes Satisfaction Treatment Questionnaire; hemoglobin A1C = glycated hemoglobin; LS = least squares; MI = myocardial infarction; NA = not applicable; NR = not reported; RCT = randomized controlled trial; T2DM = type 2 diabetes mellitus; TRIM-D = Treatment-Related Impact Measure for Diabetes.

Note: Study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the rating down a level of certainty are documented in the table footnotes.

^aThe comparator for ONWARDS 1 was insulin glargine, the comparator for ONWARDS 3 was insulin degludec, and the comparators for ONWARDS 5 were insulin glargine and insulin degludec.

^bAdditional information was requested from the sponsor to obtain 95% CIs for the LS mean estimates in each treatment group within the trials, and to obtain between-group differences with 95% CIs for hypoglycemia outcomes. This information was not necessarily part of the sponsor’s statistical analysis plan and is considered exploratory evidence.

^cRated down 1 level for serious imprecision. The target of the certainty appraisal is little to no difference based on a threshold of 0.3% points for a clinically important between-group difference (the noninferiority margin). The 95% CI for all trials includes the potential for important benefit. There is high certainty that insulin icodec is noninferior to insulin glargine or insulin degludec with respect to change from baseline in hemoglobin A1C.

^dRated down 1 level for serious imprecision. The CI for the percent time in range (3.9 mmol/L to 10.0 mmol/L) included a potential benefit (based on a threshold of importance of 5% provided by the clinical expert).

^eIn the trial, statistical testing for this outcome was not adjusted for multiplicity. The results are considered as supportive evidence.

^fRated down 1 level for serious imprecision. The CI for the percent time > 10.0 mmol/L included a potential benefit (based on a threshold of importance of 5% provided by the clinical expert).

^gRated down 1 level for serious indirectness due to the short follow-up length in the trials, and –2 levels for very serious imprecision due to the small number of events.

^hRated down 1 level for serious study limitations. The open-label design may bias reporting of subjective measures because patients were aware of the treatment they were receiving.

ⁱRated down 1 level for serious imprecision. The CI for difference between groups in all trials included a potentially important increase (based on a threshold of importance of 3% provided by the clinical expert).

^jRated down 1 level for serious imprecision. The CI for the proportion of patients experiencing level 2 or 3 hypoglycemia included potential harm (based on a threshold of importance of 3% provided by the clinical expert).

Sources: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 3 Clinical Study Report,²⁶ ONWARDS 5 Clinical Study Report,²⁷ ONWARDS 2 Clinical Study Report,²⁸ ONWARDS 4 Clinical Study Report,²⁹ and additional information provided by the sponsor at CADTH request.³⁰

Table 3: Summary of Findings for Insulin Icodec Versus Daily Basal Insulins for Patients With T2DM — Patients Who Were Insulin Experienced^a

CI = confidence interval; DTSQ = Diabetes Satisfaction Treatment Questionnaire; hemoglobin A1C = glycated hemoglobin; LS = least squares; MI = myocardial infarction; NA = not applicable; NR = not reported; RCT = randomized controlled trial; T2DM = type 2 diabetes mellitus.

Note: Study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the rating down of the level of certainty are documented in the table footnotes.

^aThe patient population for ONWARDS 2 was patients experienced with basal insulin, and the patient population for ONWARDS 4 was patients experienced with basal and bolus insulin.

^bAdditional information was requested from the sponsor to obtain 95% CIs for the LS mean estimates in each treatment group within the trials, and to obtain between-group differences with 95% CIs for hypoglycemia outcomes. This information was not necessarily part of the sponsor’s statistical analysis plan and is considered exploratory evidence.

^cRated down 1 level for serious imprecision. The target of the certainty appraisal is little to no difference based on a threshold of 0.3% points for a clinically important between-group difference (the noninferiority margin). The 95% CI for all trials includes the potential for important benefit. There is high certainty that insulin icodec is noninferior to insulin glargine or insulin degludec with respect to change from baseline in hemoglobin A1C.

^dRated down 1 level for serious indirectness; the short follow-up length in the trials is insufficient to fully capture this outcome. –2 levels for very serious imprecision; there is a very small number of events captured.

^eIn the trial, statistical testing for this outcome was not adjusted for multiplicity. The results are considered as supportive evidence.

^fRated down 1 level for serious study limitations. The open-label design may bias reporting of subjective measures because patients were aware of the treatment they were receiving.

^gRated down 1 level for serious imprecision. The target of the certainty appraisal is an increase based on a threshold for a clinically important between-group difference of 3% as informed by the clinical expert. The 95% CI includes the possibility of little to no difference.

^hRated down 2 levels for very serious imprecision. The target of the certainty appraisal is a decrease based on a threshold for a clinically important between-group difference of 3% as informed by the clinical expert. The 95% CI includes the possibility of little to no difference and an increase.

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence, ONWARDS 2 Clinical Study Report,²⁸ ONWARDS 4 Clinical Study Report,²⁹ and additional information provided by the sponsor.³

Long-Term Extension Studies

Description of Studies

The sponsor submitted the long-term extension (LTE) phase of the ONWARDS 1 trial, which extended the original open-label trial design an additional 26 weeks to provide 78 weeks of data. Patients originally randomized to either insulin icodec or insulin glargine continued their treatment as per the protocol of the ONWARDS 1 study until the end of the LTE phase. Patient population, interventions, comparators, and trial design elements remained the same. The same efficacy and safety outcomes were also assessed using the same statistical methods with some exceptions; the efficacy outcomes were not controlled for multiplicity and there was no hierarchical testing procedure for the primary outcome.

Efficacy outcomes summarized by CADTH included change in hemoglobin A1C from baseline, change in body weight from baseline, proportion of patients with level 2 or 3 hypoglycemic events, as well as time spent in range (3.9 mmol/L to 10.0 mmol/L), time spent at less than 3.0 mmol/L, and time spent at greater than 10.0 mmol/L, all between weeks 74 and 78.

Efficacy Results

Briefly, similar to the 52-week mark of the ONWARDS 1 study, there was little to no difference between insulin icodec and insulin glargine in terms of change in hemoglobin A1C from baseline to week 78, change in body weight from baseline to week 78, or time spent at less than 3.0 mmol/L. Similar to the 52-week mark of the ONWARDS 1 study, insulin icodec was statistically favoured for time spent in range from 3.9 mmol/L to 10.0 mmol/L (treatment difference = 4.41; 95% CI, 1.92 to 6.90; P = 0.0005) and time spent at greater than 10.0 mmol/L (treatment difference = –4.65; 95% CI, –7.20 to –2.10; P = 0.0004) between weeks 74 and 78. The treatment ratio for level 2 or 3 hypoglycemic in the LTE phase was 1.63 (95% CI, 1.02 to 2.61).

Harms Results

Harms in the 2 study arms were broadly similar during the LTE phase of the study, with some exceptions. Patients in the insulin glargine arm had a numerically higher incidence of AEs requiring temporary discontinuation, although the proportion was low in each group (< 5%). There were a numerically higher proportion of patients in the insulin icodec arm who experienced level 1 hypoglycemic events (55.9% versus 48.2%). Level 2 events occurred in the same proportion of patients (12.4% in both arms) and level 3 hypoglycemic events were rare in both arms (1 patient [0.2%] in the insulin icodec arm and 5 patients [1.0%] in the insulin glargine arm). All-cause mortality was similar between treatment arms with 5 patients (1.0%) in the insulin icodec arm and 3 patients (0.6%) in the insulin glargine arm (1 additional death per group relative to the 52-week mark of ONWARDS 1).

Critical Appraisal

All appraisal points pertaining to the main phase of the ONWARDS 1 study also pertain here as this LTE was a continuation of the same study design, patients, and outcomes. In addition, the fact that all efficacy outcomes here are exploratory and not adjusted for multiplicity, resulting in an increased risk of type I error (false-positive conclusions) for statistically significant results, is an additional internal validity limitation. Regarding external validity, the LTE results are only applicable to patients who were insulin naive as this was the only patient population included in the ONWARDS 1 study, leaving a knowledge gap for these outcomes in patients who are insulin experienced. Data on all-cause mortality are only provided during the LTE phase and thus information on mortality beyond 78 weeks is lacking. The comparison was also based on a small number of events, limiting a conclusion as to which treatment may be favoured. Lastly, results on long-term treatment adherence or satisfaction and clinical outcomes such as microvascular and macrovascular complications (e.g., nonfatal MI, stroke) were not assessed.

Indirect Comparisons

A network meta-analysis (NMA) was submitted with the objectives of assessing the relative efficacy and safety of insulin icodec compared to other basal insulin analogues used by patients in Canada. Analyses were conducted for patients who were insulin naive, patients who were basal insulin experienced, and patients who were basal and bolus insulin experienced . Outcomes of interest appraised by CADTH were change in hemoglobin A1C, overall hypoglycemia, level 2 and 3 hypoglycemia, and nocturnal hypoglycemia. Relevant comparators were insulin glargine U100 and U300, insulin degludec U100 and U200, and insulin detemir.

Description of Studies

The literature search yielded a total of 8,760 citations which were screened at the title and abstract phase. Of these, 22 studies were considered for data extraction for the feasibility assessment of the NMA. For patients who were insulin naive, the NMAs contained 14 studies (11 phase III and IV studies, 1 phase II trial,³¹ and 2 studies with unreported trial phases). Trials were either open label (n = 13), or double blinded (n = 1). For the patients who were basal insulin experienced, there were a total of 5 unique trials contributing to the NMA. All trials were multicentre, open-label, phase III trials. For the patients who were basal and bolus insulin experienced, a total of 3 unique trials contributed to the NMA. Two studies were multicentre, multinational, open-label trials. One trial was a phase III study, and the phase was not reported for the other.

Efficacy Results

||| ||||||||||||| ||||||||| || || ||||| ||||||| |||||| ||| |||||||| ||| |||||| || ||| |||| |||||||| |||||||| || ||||||| |||||||| |||| ||||||| ||||||||| |||||||||| |||| |||| |||| ||||| |||||| || ||||||||| |||||||| ||| ||||| |||||||| ||| ||| ||||||| | |||||||||| ||||||||| ||||||| ||| ||| ||| ||||||||| ||||| |||| |||| || |||||| ||||||| |||||| ||| |||||||| ||| ||||||| |||||| || ||| |||| |||||||| |||||||| || ||||||| |||||||| |||| ||| |||| |||| ||||| ||||||| |||||||| |||||||| |||| ||| |||| |||| ||||| |||||| || |||||||| ||| ||||||| ||| |||| |||| ||||| |||||| || ||||||| |||||||||||| |||||||| || || |||||||||| ||| ||| ||||| |||||||| ||||||| | |||||||||| ||||||||| ||||||| ||||| ||| |||||||||||| |||||||| || |||| | |||||||||| ||||||| ||||||| |||||| ||| ||| ||||| |||||||||| ||| ||||| |||||||| || |||||| |||||||||||||| ||||| ||||||||||||||||||| ||||||||| ||||||| |||||| ||| |||||||| ||| |||||| |||| |||||||| || ||| |||||||| || ||||||| |||||||| |||| ||||||| || |||| |||| ||||| |||||| || |||||||| |||||||| ||| ||||| |||||||| ||| ||| ||||||| | |||||||||| ||||||||| ||||||| || ||| |||||||||| || ||||||| |||||| ||| ||||||| |||||||| |||| ||| ||||||| ||||||||||||| ||||||| |||||||| |||| ||| |||||||| ||||||| || |||| |||| |||| ||||| || ||||||| ||||| ||| |||||||||||| |||||||| || |||| | |||||||||| ||||||| ||||||| |||||| ||| ||| ||||| |||||||||| ||| ||| ||||| ||||||||| |||||||| || || ||||| |||| ||| |||||||||||||| ||||| | ||||| ||||||||||||||||||| ||||||||| ||| |||||||| |||| |||||| ||| ||| ||| |||||| |||||||||| || ||| |||||| ||||||||| |||||||| ||| ||| ||||||||||| |||| |||| ||||| || || ||||| || |||||| ||| ||||| ||| |||||||||||| |||||||| || |||| | |||||||||| ||||||| ||||||| |||||| ||| ||| |||||||||| ||| ||| |||||||| |||||||| || || ||||| |||| ||| ||||||||||.

Harms Results

No safety analysis was run beyond the NMAs for hypoglycemia outcomes.

Critical Appraisal

The systematic literature review (SLR) which informed the NMA did not specify which comorbidities were used for the exclusion criteria of “patients with comorbidities.” This could impact the generalizability of the NMA results and affect confidence in the transitivity assumption if patient populations with different comorbidities are included.³²

With regards to the feasibility assessment in the NMA, the risk of bias appraisals was undertaken at the level of the trial, rather than at the level of the reported result (within each trial), ignoring that risk of bias can differ across outcomes within the same trial. Furthermore, the methods for appraising risk of bias were not reported. In addition, there was no discussion of how the treatment effect modifiers were chosen for the feasibility assessment, or how the assessment ensured that the list of treatment effect modifiers was comprehensive. There are also concerns with unmeasured treatment effect modifiers and heterogeneity across trials in treatment effect modifiers (e.g., a paucity of studies reported ethnicity and the ranges reported were wide). Overall, there remains uncertainty in the plausibility of the transitivity assumption underpinning the NMA.

Small treatment networks, particularly for the hypoglycemia outcomes and patients who were insulin experienced, necessitated the selection of fixed-effect models for most comparison outcomes as the SE was unstable to estimate with such a small network; however, these models do not account for between-study variance and this adds some uncertainty to the results. Furthermore, the submission did not contain any consistency assessments for the instances where there were closed loops in the network, which limits assessing the consistency of the results in the NMA with results from the individual trials.

In addition, in several analyses, the proportion of patients experiencing hypoglycemic events was much lower in the insulin icodec studies (ONWARDS trials) than the comparator studies. The submission raised the question whether the comparison was appropriate but did not adjust for these differences in any way or explore them in sensitivity analyses. For nearly all hypoglycemia outcome comparisons, the effect estimates were also affected by imprecision due to wide credible intervals, precluding any conclusions regarding which treatment in the comparison may be favoured.

Furthermore, the NMA is subject to some limitations in clinical meaningfulness. The clinical expert consulted by CADTH noted that while the results for hemoglobin A1C change from baseline across study populations may attain statistical significance, they overall do not provide an important clinical benefit. In addition, while the rationale for the NMA was to include insulin detemir and provide data for insulin icodec compared to insulin detemir, a lack of available results limited the outcomes for which insulin icodec could be compared to insulin detemir. Lastly the NMA is limited in its generalizability as the included studies primarily represent white patients younger than 70 years with relatively well-controlled diabetes, and the impact of insulin icodec on the long-term control of blood glucose and the long-term safety relative to daily insulin comparators remains unknown.

Studies Addressing Gaps in the Evidence From the Systematic Review

No studies were submitted addressing gaps in the evidence.

Conclusions

Based on evidence from 5 active-controlled RCTs which compared insulin icodec to insulin glargine and/or insulin degludec in adults with T2DM who were either insulin naive or insulin experienced, insulin icodec once-weekly injections were noninferior to the included once-daily basal insulin comparators for the change from baseline in hemoglobin A1C at week 26 or week 52 of treatment, and likely result in little to no difference in this outcome. Secondary analyses of superiority showed that insulin icodec is statistically superior compared with once-daily insulin analogues evaluated for this outcome, but the magnitude of the difference was not likely to be clinically important. Secondary and exploratory analyses of CGM outcomes (time in range, time above range, and time below range) and change in weight demonstrated that there is little to no difference associated with insulin icodec compared with insulin degludec or insulin glargine across the included studies.

Among patients who were insulin naive and insulin experienced, respectively, results of the ONWARDS 5 and ONWARDS 2 studies showed that insulin icodec likely results in little to no difference in treatment satisfaction per the DTSQ compared with insulin degludec or insulin glargine; in the absence of known MIDs, there is uncertainty in the clinical importance of the observed between-group difference, but all point estimates appeared small (0.78 points in ONWARDS 5 and 1.25 points in ONWARDS 2).

Treatment compliance was measured using the TRIM-D compliance domain only in patients who were insulin naive, and the results of the ONWARDS 5 study showed that insulin icodec likely results in increased compliance when compared to insulin degludec or insulin glargine; however, the clinical importance of the increase is uncertain. The effect of insulin icodec compared with daily insulins on treatment compliance among patients who were insulin experienced is not known.

The studies were generalizable to populations in Canada with T2DM and provided robust evidence versus key active comparators. Per the clinical expert consulted by CADTH, the uptake of insulin icodec is likely to be highest among eligible patients starting on basal insulin for the first time, and may also be preferred over daily basal insulins by patients who would prefer a lower burden of administration frequency.

The safety profile of insulin icodec was generally comparable to that of insulin glargine and degludec. In the insulin-naive populations, insulin icodec resulted in little to no difference in the proportion of patients experiencing hypoglycemia (level 2 or 3). In the insulin-experienced populations, the proportion of patients with level 2 or 3 hypoglycemic events was higher in the ONWARDS 2 study but lower in the ONWARDS 4 study in the insulin icodec treatment arms than the comparator arms. Additionally, there were more level 2 hypoglycemic events among patients treated with insulin icodec than the comparators in 4 of the 5 trials, although a few patients accounted for many events in the ONWARDS 1 and ONWARDS 3 studies. Evidence from the LTE was generally consistent with the results of the main trials. The NMA was inconclusive either due to substantial imprecision or inadequate evidence to make a comparison, or the results did not demonstrate clinically meaningful impact (in the case of hemoglobin A1C).

There is a data gap regarding long-term outcomes such as the downstream consequences of T2DM, including nonfatal CV events such as MI and stroke, CV deaths, and long-term all-cause mortality beyond the duration of the clinical trials. Additionally, there is a data gap regarding the effect of insulin icodec on patients’ HRQoL using more comprehensive global measures.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of insulin icodec (Awiqli) once-weekly SC insulin injection in the treatment of T2DM in adults to improve glycemic control.

Disease Background

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following have been summarized and validated by the CADTH review team.

Diabetes mellitus is a chronic health condition affecting the body’s ability to convert sugar, or glucose, into energy. Insulin is released from beta cells in the pancreas when food is ingested, and under normal circumstances is involved in transferring glucose into cells to use as energy.² T2DM develops when the body is no longer able to use insulin efficiently or produce enough insulin to manage blood glucose levels within a normal range.³ This persistent hyperglycemia results in a constellation of symptoms and downstream impacts on the body.³

Diagnosis of T2DM is made using blood tests and confirmed by a clinician. Diagnostic criteria are 1 of 3 results on a blood test:³³

• fasting blood glucose of 7.0 mmol/L or greater

• hemoglobin A1C of 6.5% or greater

• 2-hour plasma glucose of 11.1 mmol/L or greater during an oral glucose tolerance test.

The vast majority of diabetes mellitus cases are T2DM, accounting for approximately 90% of patients.⁵ T2DM is more often diagnosed in adulthood with many being diagnosed after the age of 40;^34,35 however, rates of diagnosis are increasing in younger ages.³³ Diabetes Canada estimates that more than 4 million people living in Canada, representing approximately 10% of the population, will be diagnosed with diabetes mellitus in 2023, and that this will increase to more than 5 million (12%) by 2033.⁴

The prevalence of T2DM may be higher in racialized and minority groups.^4,5 Indigenous Peoples in Canada have higher risk for diabetes and diabetes-related complications,⁶ and self-reported data from Ontario showed that the prevalence of T2DM was higher for South Asian (8.1%) and Black (8.5%) populations than white (4.2%) populations.⁷ Patients with COVID-19 are also at increased risk of developing T2DM, especially marginalized communities which are more likely to be impacted by the health, social, and economic effects caused by the pandemic.⁵ Insufficient access to primary care providers are also barriers which may impact T2DM prevention, diagnosis, and care.

While T2DM does not directly result in death, it is associated with increased mortality from ischemic heart disease, stroke, chronic liver disease, infections, and cancer of the liver, pancreas, breast, and female reproductive system.⁹ Similar associations are reported in patients living with undiagnosed diabetes, where age- and sex-standardized excess risk of all-cause mortality is at least as high as among persons with diagnosed T2DM, and approximately 2-fold higher compared with persons without diabetes.³⁶

In addition to increased all-cause mortality, patients with poorly controlled T2DM are at increased risk of developing diabetes-related microvascular and macrovascular complications such as retinopathy, nephropathy, neuropathy, stroke, heart failure, MI, or other CV events.³⁷ In Canada, diabetes mellitus is associated with a 20-times increased risk of hospitalization for nontraumatic lower limb amputations, with an estimated 65% of lower limb amputations due to diabetes.¹⁰

T2DM also has multifaceted impacts on patients’ HRQoL. Patients with T2DM may feel stigma from socially identifiable characteristics related to their condition, for example, insulin injections, blood glucose monitoring, dietary restrictions, having obesity, and experiences of hypoglycemic episodes.¹¹ Of note, insulin injections and increased frequency of injections are associated with increased perceptions of burden.^38,39 Hypoglycemia due to diabetes may result in both long- and short-term complications which can negatively impact patients’ HRQoL.¹³ A systematic review concluded that depression is associated with not only worse mental health, but also worse physical health in patients with diabetes mellitus; people living with both diabetes mellitus and depressive symptoms were less satisfied with their treatment, and worried more about the impact of diabetes in the future and about the social and vocational impact of diabetes.¹²

Risk factors for T2DM include age, overweight, lack of physical activity, and genetics.⁴⁰ There is currently no cure for T2DM, but it is typically managed with a combination of nonpharmacological (i.e., weight management, nutrition, and physical activity) and pharmacological (i.e., noninsulin and insulin drugs) interventions.³

Standards of Therapy

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following have been summarized and validated by the CADTH review team.

The main goals of treatment for patients with T2DM are to reduce the risk of long-term complications through control of glycemia and blood pressure, and CV risk reduction through control of lipids and hypertension. Clinicians also aim to maintain or improve HRQoL for people living with T2DM^41-45 and to reduce the symptoms of hyperglycemia.¹⁴ Treatment must be individualized, and it is recommended to combine lifestyle changes (e.g., dietary modification, exercise, quitting smoking) and pharmacological interventions as appropriate.¹⁴

Early intervention may include lifestyle modification alone, such as increased physical activity, dietary changes, and weight loss, or it may include these modifications in addition to pharmacotherapy, depending on the patient’s initial presentation.¹⁴ As T2DM is a progressive condition, eventually most patients will require pharmacological intervention. Metformin is a noninsulin antihyperglycemic drug and is usually the first pharmacological intervention that patients will be prescribed.¹⁴ If a patient is unable to lower or maintain their hemoglobin A1C or blood glucose levels with metformin treatment alone, additional therapies may be combined with continued metformin therapy, such as sulfonylureas, DPP-4is, SGLT2is, GLP-1 RAs, and insulin.^14,46

According to Canadian diabetes treatment guidelines, insulin (in combination with metformin) should be initiated in the event of a patient in whom fasting glucose levels and/or hemoglobin A1C are not at target on current antihyperglycemic drugs, or with symptomatic hyperglycemia and/or metabolic decompensation.¹⁴ Basal insulin should be initiated and titrated to achieve fasting glucose targets, and metformin should be continued unless contraindicated. Other antihyperglycemic drugs may also be used in combination with these therapies as needed, and therapy should be advanced if the patient’s hemoglobin A1C is not at target within 3 months to 6 months despite adequate titration of basal insulin and supports for lifestyle and other pharmacotherapeutic interventions.¹⁴ Basal insulins for treatment of T2DM can include long-acting or intermediate-acting insulins.¹⁴ Currently available long-acting insulins include insulin degludec (U100 or U200), insulin glargine (U100 or U300), and insulin detemir, while neutral protamine Hagedorn insulin is an intermediate-acting insulin.¹⁴

For patients who require further treatment intensification beyond the use of basal insulin, a bolus (i.e., rapid-acting) insulin regimen may be added in a step-wise fashion, beginning with 1 mealtime injection per day followed by additional injections as recommended by a health care practitioner.¹⁴

Drug Under Review

Insulin icodec (Awiqli) is approved by Health Canada for the once-weekly treatment of adults with diabetes mellitus to improve glycemic control. The reimbursement request is for the once-weekly treatment of adults with T2DM to improve glycemic control. This drug has not been previously reviewed by CADTH. It is currently under review with the FDA, European Medicines Agency, and the Therapeutic Goods Administration (Australia).

Insulin icodec is recommended to be administered SC once weekly on any day of the week, but preferably on the same day each week.¹⁶ The dose of insulin icodec is to be titrated based on a patient’s metabolic needs, blood glucose monitoring results, and glycemic control goal.¹⁶ The recommended starting dose for patients with T2DM who are insulin naive is 70 units administered once weekly.¹⁶ Recommended doses for patients who switch from a daily basal insulin to insulin icodec are described separately in the Health Canada product monograph.¹⁶ Insulin icodec is not to be used in combination with other long-acting basal insulins.¹⁶

Insulin and insulin analogues work to lower blood glucose by stimulating peripheral glucose uptake, especially by skeletal muscle and fat, and by inhibiting hepatic glucose production.¹⁵ Insulin also inhibits lipolysis and proteolysis and enhances protein synthesis. Insulin icodec was developed by re-engineering the ultra-long basal insulin OI338.¹⁵ Three amino acids were changed and a C20 icosane fatty diacid chain was attached, resulting in molecular stability, reduced enzymatic degradation, strong reversible binding to albumin, and slow receptor-mediated clearance which allows for continuous insulin action.¹⁵ A depot of essentially inactive insulin icodec is formed in the circulation and interstitial compartment, from which insulin icodec is slowly and continuously released and binds specifically to the insulin receptor. Insulin icodec has a plasma half-life of 196 hours (approximately 8 days) while retaining the same biological properties as natural human insulin.¹⁵

Key characteristics of insulin icodec are summarized in Table 4 with other injectable long-acting insulin therapies available for the treatment of T2DM.

Table 4: Key Characteristics of Insulin Icodec and Relevant Comparators

DM = diabetes mellitus; GLP-1 = glucagon-like peptide-1; IM = intramuscular; OAD = oral antihyperglycemic drug; SC = subcutaneous; T1DM = type 1 diabetes mellitus; T2DM = type 2 diabetes mellitus.

^aHealth Canada–approved indication.

Source: Product monographs for insulin icodec,¹⁶ Lantus,⁴⁷ Basaglar,⁴⁸ Semglee,⁴⁹ Toujeo,⁵⁰ Levemir,⁵¹ and Tresiba.⁵²

Stakeholder Perspectives

Patient Group Input

This section was prepared by the CADTH review team based on the input provided by patient groups. The full original patient input received by CADTH have been included in the Stakeholder section of this report.

Patient input was submitted for this CADTH review by Diabetes Canada, which fielded a survey between October 3, 2023, and October 23, 2023, directed at people living with T2DM and their caregivers across Canada. The survey consisted of a self-directed questionnaire which inquired about respondents’ lived experiences with diabetes and contained several questions pertaining to insulin icodec. Of the 21 respondents, 13 identified as living with T2DM and 1 identified as a caregiver (the remainder did not self-identify whether they were a patient or a caregiver). A total of 93% (of 14 respondents for the question) were aged older than 55 years, 35% were aged 75 years to 84 years, and 71% reported living with T2DM for more than 10 years (of which 29% reported living with T2DM for more than 20 years). Respondents lived in Ontario, British Columbia, Prince Edward Island, Quebec, Manitoba, Alberta, and Nova Scotia. A total of 78% of respondents lived in urban or suburban areas and 79% identified as white, with 1 respondent identifying as being either South Asian, Southeast Asian, or Chinese.

Most respondents (n not specified) indicated that living with T2DM was preoccupying, inconvenient, and burdensome, with constant management requiring foresight and planning. Most respondents (n not specified) mentioned challenges in managing their blood glucose in relation to diet and food choices. A total of 24% (of 20 respondents) reported experiencing hyperglycemia more than once per day and 10% reported experiencing it more than once per week (3 respondents indicated no hyperglycemia or unknown). A total of 43% (of 20 respondents) indicated they did not experience hypoglycemia or experienced it in the past but not currently, while 14% experienced it more than once per week — none reported experiencing hypoglycemia daily.

All respondents who provided data on diabetes-related drug use (n = 19) reported taking antihyperglycemic medication (2 did not respond). Medications being taken included long-, short-, and rapid-acting insulin, insulin icodec, and other noninsulin antihyperglycemic drugs, either as single-drug products or combined with metformin. A total of 18 respondents also reported nonpharmacological interventions to manage their diabetes, with 50% reporting healthy eating and 28% reporting taking herbal remedies or over-the-counter supplements. A total of 5 of 18 (28%) respondents reported current insulin icodec use.

A total of 61% (of 18) respondents said they were very satisfied or satisfied with their medication; no respondents indicated dissatisfaction. Respondents indicated that ease of use, lack of side effects, and helping to lower hemoglobin A1C were aspects they liked about their medications. Quantity and/or timing of the dosing of different medications (e.g., number of pills or needles, managing weekly versus daily dosing), fluctuations to blood glucose levels, the need for antihyperglycemic medication, and medication cost were factors which respondents indicated they disliked.

When choosing a medication for diabetes management, several considerations were important to respondents including avoiding hypoglycemia and hyperglycemia, reducing the risk of heart problems, reducing high blood pressure, maintaining satisfactory blood glucose levels throughout the day, and avoiding yeast infections, urinary tract infections, fluid retention, or weight gain. Affordability was also highlighted as an important consideration. Improvements that respondents wished to see in a new treatment that are not currently being achieved with available therapies included fewer side effects, blood flow improvement to extremities, weight control, and better hemoglobin A1C results. A handful of respondents (n not specified) noted ways in which their daily life and overall quality of life would be different if a new treatment provided their desired improvements: reduced stress, staying healthy, reducing the number of medications, reducing insulin, reducing weight gain, better nutrition counselling, and improved mental health were highlighted.

Clinician Input

Input From Clinical Expert 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

Treatment goals for patients with T2DM include reducing symptoms of hyperglycemia, reducing risk of long-term complications through control of glycemia and blood pressure, and reducing CV risk through control of lipids and hypertension, through a combination of lifestyle modifications and pharmacotherapeutic approaches. The clinical expert consulted by CADTH highlighted that treatment must be individualized, provided in a culturally appropriate manner, and equitably and affordably accessible across Canada. The bulk of care for patients with T2DM is provided in primary care but requires a multidisciplinary approach with nursing and nutrition services, and access to specialist care when required. The key unmet need highlighted by the clinical expert was a lack of widespread access to primary care and therefore lack of access to diabetes prevention, detection, and treatment. In particular, access to diabetes education and specialist care varies greatly across the country. Many medications and technologies for diabetes are not easily affordable and coverage on provincial formularies varies from province to province.

Place in Therapy

The clinical expert described that insulin icodec would be used in patients not meeting glycemic targets despite lifestyle modification, and use of or intolerance or contraindication to metformin, GLP-1 RAs, and/or SGLT2is. It may also be used as a first-line therapy in patients with T2DM who present with symptomatic hyperglycemia and/or metabolic decompensation with or without metformin.

The clinical expert highlighted that insulin icodec would fit into the current paradigm for introduction of basal insulin in the management of T2DM. Insulin icodec may be preferred by some patients who are unable or unwilling to take daily basal insulin, or who would prefer a lower burden related to administration frequency.

Patient Population

The clinical expert indicated that patients most suited for treatment with insulin icodec would be those with T2DM who are not meeting glycemic targets (hemoglobin A1C or blood glucose) despite lifestyle modification, with use of or intolerance or contraindication to metformin, GLP-1 RAs, and SGLT2is, or patients with T2DM with symptomatic hyperglycemia and/or metabolic decompensation with or without metformin. In the current treatment paradigm, these are the same patients who would be introducing a daily basal insulin to their diabetes management strategy. Patients who are unwilling or unable to take daily basal insulin, or who are reluctant to start insulin therapy due to concerns about the burden of daily injections, may be the most likely to benefit from insulin icodec compared to the daily basal insulin options.

Assessing the Response Treatment

The clinical expert consulted by CADTH noted that the timing of assessments varies substantially between monitoring clinicians and between patients, but ideally a patient would be supported through phone and email support to adjust dosing during the first 2 months to 3 months of insulin therapy, followed by an assessment of treatment suitability after 3 months to 6 months of therapy. Diabetes management is complex, individualized, and ideally holistic. As such, there are several factors a monitoring physician or nurse practitioner will assess with regards to insulin icodec, including: treatment acceptance, treatment adherence, hemoglobin A1C target achievement, time in range with CGM of more than 70%, time below range less than 4%, and no severe hypoglycemic episodes. A sign of positive response to insulin icodec would also be increased HRQoL, such as less diabetes distress and more treatment satisfaction.

Discontinuing Treatment

Factors influencing a decision to discontinue insulin icodec would include allergy, nonadherence, or diabetes remission or glycemic control improvement through weight loss or use of other antihyperglycemic drugs or bariatric surgery.

Prescribing Considerations

Diagnosis, prescribing, and management of treatment for patients with T2DM may occur in primary care. The diagnosis of T2DM and the use of insulin icodec were described by the clinical expert to be uncomplicated and do not necessarily require specialist care. There are limitations to access of specialist care due to the low number of endocrinologists in Canada and the high number of patients with T2DM.

Additional Considerations

According to the clinical expert, the use of sulfonylureas with insulin increases the risk of hypoglycemia, so combination of sulfonylureas with insulin icodec should likely be avoided.

Clinician Group Input

No clinician group feedback was received by the deadline of the call for input.

Drug Program Input

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

Table 5: Summary of Drug Plan Input and Clinical Expert Response

T2DM = type 2 diabetes mellitus.

Clinical Evidence

The objective of CADTH’s Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of insulin icodec once-weekly injection in the treatment of T2DM in adults. The focus will be placed on comparing insulin icodec to relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of insulin icodec is presented in 4 sections with CADTH’s critical appraisal of the evidence included at the end of each section. The first section, the systematic review, includes pivotal studies and RCTs that were selected according to the sponsor’s systematic review protocol. CADTH’s assessment of the certainty of the evidence in this first section using the GRADE approach follows the critical appraisal of the evidence. The second section includes sponsor-submitted LTE studies. The third section includes indirect evidence from the sponsor. The fourth section would include additional studies that were considered by the sponsor to address important gaps in the systematic review evidence; however, none were submitted for this review.

Included Studies

Clinical evidence from the following is included in the CADTH review and appraised in this document:

• 5 RCTs identified in systematic review

• 1 LTE study

• 1 indirect treatment comparison.

Systematic Review

Contents within this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CADTH review team.

Description of Studies

Characteristics of the included studies are summarized in Table 6.

Table 6: Details of Studies Included in the Systematic Review

BMI = body mass index; DPP = dipeptidyl peptidase; DTSQ = Diabetes Treatment Satisfaction Questionnaire; FPFV = first patient first visit; FPG = fasting plasma glucose; GLP-1 = glucagon-like peptide-1; hemoglobin A1C = glycated hemoglobin; LPLV = last patient last visit; NA = not applicable; q.d. = once a day; RCT = randomized controlled trial; SGLT2 = sodium-glucose cotransporter-2; T2DM = type 2 diabetes mellitus; TRIM-D = Treatment-Related Impact Measure for Diabetes.

^aSimultaneous participation in a trial with the primary objective of evaluating an approved or nonapproved investigational medicinal product for prevention or treatment of COVID-19 disease or postinfectious conditions was allowed at the investigator’s discretion without discontinuing trial product.

^bWeekly placebo in the intervention group or daily placebo in the comparator group of the ONWARDS 3 trial were administered to patients to maintain blinding.

Sources: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 3 Clinical Study Report,²⁶ ONWARDS 5 Clinical Study Report,²⁷ ONWARDS 2 Clinical Study Report,²⁸ and ONWARDS 4 Clinical Study Report.²⁹ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

All 5 of the ONWARDS studies were multicentre, phase III RCTs comparing the efficacy and safety of insulin icodec to once-daily insulin analogues among patients with diabetes mellitus; all studies are open label, except the ONWARDS 3 study in which a double-blind, double-dummy design was used to compare insulin icodec to insulin degludec. The ONWARDS 1 through 5 studies randomized patients with T2DM; of these studies, the ONWARDS 1, ONWARDS 3, and ONWARDS 5 studies recruited patients who were insulin naive, while the ONWARDS 2 and ONWARDS 4 trials recruited patients who had previous experience with insulin. Specifically, the ONWARDS 2 trial enrolled patients who had previously received basal insulin, and the ONWARDS 4 trial enrolled patients with previous treatment with basal and bolus insulin. All studies consisted of a 2-week screening period followed by a main on-treatment phase (consisting of either 26 weeks or 52 weeks), and a follow-up phase of 5 weeks. The ONWARDS 1 trial also included a 26-week LTE phase. All patients were randomized at a 1:1 ratio to receive either insulin icodec or a once-daily basal insulin. Randomization was centrally conducted and adjudicated using an interactive web response system. In the ONWARDS 5 study, patients in the once-daily analogues treatment arm received either insulin degludec or insulin glargine as selected by the investigator before randomization per standard of care. No stratification was described in the ONWARDS 1, ONWARDS 2, ONWARDS 4, or ONWARDS 5 studies. However, in the ONWARDS 3 study, randomization was stratified by region and sulfonylurea or glinide use. Events for adjudication were evaluated by an independent external event adjudication committee in a blinded manner. All trials used a treat-to-target approach ensuring that patients received optimal titration of insulin to improve their hemoglobin A1C. Insulin titration was based on prebreakfast self-measured plasma glucose values. CGM was also employed in the ONWARDS 1, ONWARDS 2, and ONWARDS 4 trials based on patient preference to generate profiles for evaluating the glycemic control during the trial; the CGM data were blinded for both patients and the investigator to avoid influence on titration and the glycemic control.

Populations

Inclusion and Exclusion Criteria

Eligibility criteria were similar across the trials. Key inclusion criteria included adult patients (aged ≥ 18 years) diagnosed with T2DM a minimum of 180 days before baseline. All patients had to have a hemoglobin A1C from 7.0% to 11.0% (53.0 mmol/mol to approximately 96.7 mmol/mol) confirmed by central laboratory analysis. The ONWARDS 1, ONWARDS 3, and ONWARDS 5 trials specified inclusion of patients with T2DM who were naive to previous insulin therapy. The ONWARDS 2 and ONWARDS 4 trials included patients who were previously treated with basal insulin or basal and bolus insulin, respectively. Patients in the ONWARDS 2 and ONWARDS 4 trials had to have been treated with insulin with or without antihyperglycemic drug regimens at least 90 days before screening. Key exclusion criteria across the ONWARDS studies were also generally consistent. Exclusion criteria across the ONWARDS 1 to ONWARDS 5 studies specified patients with uncontrolled diabetes or other conditions, such as chronic heart failure or MI. Patients with any other safety concerns related to the trial product or unacceptable intolerability were also excluded.

Interventions

All patients who were insulin naive (ONWARDS 1, 3, and 5) were administered insulin icodec via SC injection at a dose of 70 U (700 units/mL) once weekly. Patients who were previously receiving insulin (ONWARDS 2 and 4) determined their weekly dose of insulin icodec by multiplying their daily basal insulin dose before randomization by 7. Patients switching from a daily insulin analogue treatment to insulin icodec were also required to have a 1-time dose which consisted of the total daily basal insulin dose before randomization multiplied by 7 plus an additional 50% of that dose.

In the ONWARDS 5 study, patients in the insulin icodec treatment arm also used an investigational device called the DoseGuide system. The DoseGuide system consists of the DoseGuide app for patients and the DoseGuide Portal for investigators, which are integrated with the DoseGuide Cloud, where dose recommendations are calculated. The system utilizes measurements from a blood glucose meter via Bluetooth, injection history, and prebreakfast fasting self-measured plasma glucose provided by the patient, and provides the patient with dosing recommendations through the app. Patients were trained on using the DoseGuide system by the investigators.

Comparators in the ONWARDS studies included insulin degludec or insulin glargine. Insulin degludec was administered to patients via SC injection at a concentration of 100 units/mL once daily. Insulin glargine was administered to patients via SC injection at a concentration of either 100 units/mL or 300 units/mL. Once-daily insulin analogues in all ONWARDS studies (degludec, glargine U100, or glargine U300), were initiated and titrated as per the local label and standard clinical practice. For patients enrolled in the ONWARDS 2 and ONWARDS 4 trials switching from a previous once-daily basal insulin, the dose of trial once-daily basal insulin was in accordance with the local label.

Patients’ doses of insulin could be adjusted during the treatment periods of the trials based on 3 prebreakfast values. In the ONWARDS 4 study, patients in the insulin icodec and comparator treatment groups were also administered insulin aspart at mealtimes. Insulin aspart was administered via SC injection (100 units/mL) 2 times to 4 times daily before consumption of food.

Any background medications taken during the trial were considered noninvestigational during the trial and were not provided by Novo Nordisk; patients purchased these medications themselves and were to take them as they were prescribed by their health care professional. Concomitant medications included any antihyperglycemic medication other than trial products that patients were receiving from screening until the end-of-trial visit. Until the end of treatment, only randomized trial products and pretrial noninsulin antihyperglycemic background medications were allowed, except for sulfonylureas and glinides which had to be discontinued (ONWARDS 1, 2, and 4) or reduced by approximately 50% (ONWARDS 3 and 5) at randomization; these medications could be discontinued at the discretion of the investigator if there were any safety concerns. If required, the investigators were also able to initiate antihyperglycemic medications before the end of the trial treatment period.

Outcomes

A list of efficacy end points assessed in this Clinical Review Report is provided in Table 7, followed by descriptions of the outcome measures. Summarized end points are based on outcomes included in the sponsor’s Summary of Clinical Evidence as well as any outcomes identified as important to this review according to the clinical expert consulted by CADTH and stakeholder input from patient and clinician groups and public drug plans. Using the same considerations, the CADTH review team selected end points that were considered to be most relevant to inform CADTH’s expert committee deliberations and finalized this list of end points in consultation with members of the expert committee. All summarized efficacy end points were assessed using GRADE. Select notable harms outcomes considered important for informing CADTH’s expert committee deliberations were also assessed using GRADE.

Table 7: Outcomes Summarized From the Studies Included in the Systematic Review

BG = blood glucose; CV = cardiovascular; DTSQ = Diabetes Treatment Satisfaction Questionnaire; hemoglobin A1C = glycated hemoglobin; MI = myocardial infarction; NR = not reported; T2DM = type 2 diabetes mellitus; TRIM-D = Treatment-Related Impact Measure for Diabetes.

^aStatistical testing for these end points was adjusted for multiple comparisons (e.g., hierarchal testing).

^bLevel 2 hypoglycemic episodes are defined as “clinically significant” and level 3 as “severe.”

Source: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 2 Clinical Study Report,²⁸ ONWARDS 3 Clinical Study Report,²⁶ ONWARDS 4 Clinical Study Report,²⁹ and ONWARDS 5 Clinical Study Report.²⁷ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Change in Hemoglobin A1C From Baseline

This end point was calculated as the baseline value of hemoglobin A1C at week 0 subtracted from the hemoglobin A1C value at the last week of assessment. Testing for hemoglobin A1C was performed in a central laboratory. A noninferiority margin of 0.3% points was chosen based on the recommendation in the FDA guidance on developing drugs for the treatment of diabetes.¹⁷ The clinical expert consulted by CADTH agreed that this was a reasonable margin for the MID between groups.

Time in Range (3.9 mmol/L to 10.0 mmol/L)

Patients were equipped with CGM devices in the relevant trials in several 4-week intervals. In the ONWARDS 1 study, these were weeks 1 to 4, 22 to 26, 48 to 52, 74 to 78, and during follow-up from week 78 to 83. In the ONWARDS 2 study, these were weeks 0 to 4 and weeks 22 to 26. In the ONWARDS 4 study, these were weeks 0 to 4, weeks 22 to 26, and during follow-up from week 26 to 31. There was no CGM in the ONWARDS 3 or ONWARDS 5 studies. In the relevant trials (ONWARDS 1, 2, and 4), the system used was the Dexcom G6 and readings were blinded to both the patient and the investigator and were not used for dose titration or hypoglycemic episode reporting. The sensor and transmitter were fitted, verified, and uploaded by site staff on the patient during site visits and relevant training was the responsibility of the investigator or site staff.

This end point was calculated as 100 times the number of recorded measurements in the glycemic range of 3.9 mmol/L to 10.0 mmol/L, divided by the total number of measurements. International consensus criteria⁵⁴ required that at least 70% of the planned CGM measurements during the last 4 weeks of treatment are available for end point data to be included in the analysis; the ONWARDS studies were in alignment with international consensus criteria for time in range. A 2019 International Consensus on Time in Range¹⁸ identified that a 5% increase in time in range is associated with clinically significant benefits for individuals with type 1 diabetes mellitus or T2DM, and based on this, a between-group difference of approximately 5% may be an appropriate threshold for a clinically important effect.

Time Spent Less Than 3.0 mmol/L

The measurement of blood glucose for the purpose of calculating this outcome was the same as for time in range (3.9 mmol/L to 10.0 mmol/L).

This end point was calculated as 100 times the number of recorded measurements below 3.0 mmol/L (54 mg/dL) divided by the total number of recorded measurements. Following international consensus criteria, it was required that at least 70% of the planned CGM measurements during the last 4 weeks of treatment be available for end point data to be included in the analysis. There was no MID identified in the literature, but the clinical expert consulted by CADTH indicated that a between-arm difference of approximately 1% may be an appropriate threshold for a clinically important effect, based on the fact that 1% of the day is equal to 15 minutes.¹⁸

Time Spent Greater Than 10.0 mmol/L

The measurement of blood glucose for the purpose of calculating this outcome was the same as for time in range (3.9 mmol/L to 10.0 mmol/L).

This end point was calculated as 100 times the number of recorded measurements above 10 mmol/L (180 mg/dL) divided by the total number of recorded measurements. Following international consensus criteria, it was required that at least 70% of the planned CGM measurements during the last 4 weeks of treatment be available for end point data to be included in the analysis. The clinical expert consulted by CADTH indicated that a between-group difference of approximately 5% may be an appropriate threshold for a clinically important effect, to align with the MID selected for time in range.¹⁸

Change in Body Weight

Body weight was measured by study investigators on predefined days in kilograms or pounds and recorded to 1 decimal place. In the ONWARDS 1, ONWARDS 3, ONWARDS 2, and ONWARDS 4 studies, it was specified in the protocol that weight should be measured without coat and shoes, wearing only light clothing, and assessed with the same equipment throughout the trial if possible. In the ONWARDS 5 study, this detail was not specified.

The body weight value at baseline (week 0) was subtracted from the body weight value measured at the last week of the main trial period (i.e., week 26 or week 52 as applicable). Based on clinical expert opinion, a threshold of clinical importance of 5 kg was selected by CADTH to assess the clinical importance of the results.

Number of Level 2 or Level 3 Hypoglycemic Episodes

Clinically significant (level 2) hypoglycemic events were defined as a reading of less than 3.0 mmol/L (54 mg/dL) confirmed by the trial-supplied blood glucose meter. Severe (level 3) hypoglycemic levels were defined as any event requiring assistance of another person to actively administer carbohydrates, glucagon, or take other corrective actions, whether or not plasma glucose could be recorded to confirm.

A count was conducted of all clinically significant hypoglycemic episodes (level 2) or severe hypoglycemic episodes (level 3) within the main trial time frame, and during the extension phase where appropriate (ONWARDS 1).

Patients were supplied a blood glucose meter including auxiliaries, which used test strips calibrated to plasma values. Patients were instructed to record their self-measured plasma glucose before breakfast daily throughout each of the trials (ONWARDS 1, 2, 3, 4, and 5). In the ONWARDS 4 study, patients were additionally instructed to measure their plasma glucose daily before lunch, before dinner, and at bedtime. These data were used to record relevant hypoglycemic events.

Diagnosis of hypoglycemic episodes included readings of fasting plasma glucose (where possible), neurologic examination after the event (for severe episodes in which readings of fasting plasma glucose were not possible), and patient recordings in an e-diary. Any event of plasma glucose recording under 3.9 mmol/L (70 mg/dL) on the trial-supplied blood glucose meter was to be recorded as a hypoglycemic event in the e-diary. In the event a patient was not able to fill in the e-diary, (e.g., in the case of hospitalization at the time of the episode), it could be reported by the patient retrospectively. If the event fulfilled the criteria for a SAE in addition to the previously mentioned criteria, then an AE form and safety information form were also to be completed.

A 3% between-group threshold of clinically important difference was suggested by the clinical expert consulted by CADTH.

Diabetes Treatment Satisfaction Questionnaire

The DTSQ was developed to assess patient satisfaction with diabetes treatment.^55,56 The questionnaire consisted of 6 items which were scored on a scale of 0 to 6, with a total questionnaire score of 0 to 36. Higher scores indicated greater satisfaction with treatment. Patient’s scores in total treatment satisfaction from baseline were subtracted from their score at the final week of analysis. The DTSQ is widely used in the context of clinical trials, but publications on the validation of its psychometric properties in general or specifically for the English version were not identified. Further, it is unclear whether the ONWARDS trials used the Status(S) or the ChangeI version of the DTSQ, the latter of which may be more sensitive to change for patients closer to the ceiling level of satisfaction at baseline.⁵⁷ No MID was identified in the literature and as this is not a measure used in clinical practice, the clinical expert consulted by CADTH did not provide a clinically informed threshold for a clinically important between-group difference in the change from baseline.

TRIM-D Compliance Domain

The TRIM-D measures patients’ perceptions of the impact of treatment on functioning and well-being.⁵⁸ Only the compliance domain was reported. The questionnaire was scored with a total value ranging from 4 to 20. The questionnaire consisted of 4 items which were scored on a scale of 1 to 5 and the transformed to a 0 to 100 scale. In the compliance domain, higher scores indicated better compliance. The compliance domain of the TRIM-D was measured at the final week of analysis.

The TRIM-D has been demonstrated to be significantly responsive to change over time⁵⁹ and demonstrated adequate internal consistency reliability and test-retest reliability.⁵⁸ However, it is unknown whether these assessments of responsiveness and reliability are relevant to the compliance domain in isolation. No MID was identified in the literature and as this is not a measure used in clinical practice, the clinical expert consulted by CADTH did not provide a clinically informed threshold for a clinically important between-group difference in the change from baseline.

CV Death

This outcome was identified as an outcome of interest by the Canadian Drug Expert Committee (CDEC) presenters for patients with T2DM. However, it was not an outcome assessed in the included studies.

Nonfatal MI

This outcome was identified as an outcome of interest by the CDEC presenters for patients with T2DM. However, it was not an outcome assessed in the included studies.

Nonfatal Stroke

This outcome was identified as an outcome of interest by the CDEC presenters for patients with T2DM. However, it was not an outcome assessed in the included studies.

All-Cause Mortality

This outcome was identified as an outcome of interest by the CDEC presenters for patients with T2DM. It was assessed as a safety outcome (i.e., number of deaths).

Other Microvascular and Macrovascular Complications of T2DM

This outcome was identified as an outcome of interest for patients with T2DM. However, it was not an outcome assessed in the included studies.

Statistical Analysis

A summary of the statistical analyses conducted for each outcome is presented in Table 8.

Sample Size and Power Calculation

ONWARDS 1

Assumptions based on the NN1250-4252 trial,⁶⁰ which compared insulin degludec 200 U/mL to insulin glargine 300 U/mL among patients with T2DM treated with basal insulin, were used to inform the sample size and power calculations. The sample size of the ONWARDS 1 study was determined based on FDA requirements to have a minimum of 300 patients completing 78 weeks of treatment with insulin icodec, as well as having sufficient marginal power for the confirmatory secondary hypothesis. Assuming 85% of patients would complete treatment, 353 patients per treatment group (i.e., 706 randomized patients) would be required.

The sample size calculation also considered the power for declaring noninferiority with a noninferiority margin of 0.3% points with respect to the primary end point, change in hemoglobin A1C from baseline to week 52, for the specified estimand and the full analysis set (primary analysis set). With 970 patients and the assumption of a mean treatment difference of 0.03% points, there would be 99% power for declaring noninferiority. Assuming 25% of patients will fail screening, approximately 1,293 patients had to be enrolled to randomize 485 patients to each treatment group.

Per assumptions from the NN1250-4252 trial,⁶⁰ there would be 98% power to detect a true difference of 4% for time-in-target range (3.9 mmol/L to 10.0 mmol/L) from week 48 to 52.

ONWARDS 2

The sample size was calculated with the aim of having 90% power for declaring noninferiority with a noninferiority margin of 0.3% points with respect to change in hemoglobin A1C. Observations from recent studies with basal insulins (NN9068-4185⁶¹ and NN9068-4229) were used to determine the power required and expected treatment difference for the primary end point. Therefore, based on the NN9068-4185 and NN9068-4229 studies, it was assumed that the ONWARDS 2 study would require 520 patients randomly assigned to the trial product to have 90% power to confirm noninferiority and detect a mean treatment difference of 0.015% points for the primary end point in the overall population. Assuming 25% of patients would be classified as screen failures, approximately 693 patients were required to be screened to achieve randomization of 520 patients. For the hypothesis of superiority of insulin icodec compared to insulin degludec in terms of change from baseline to week 26 in hemoglobin A1C, there would be 49% marginal power and also 49% combined power to detect if the true treatment difference in hemoglobin A1C was 0.18% points with the sample size of 520.

ONWARDS 3

The sample size was calculated with the aim of having 90% power for declaring noninferiority with a noninferiority margin of 0.3% points with respect to change in hemoglobin A1C. Observations from a study with basal insulin analogues in patients with T2DM (NN1250-3998,⁶³ NN1250-3579,⁶⁴ NN1250-3643,⁶⁵ and NN1250-3672⁶⁶) were used to determine the number of patients and expected treatment difference in the primary end point for the ONWARDS 3 study. Therefore, there was an assumed mean treatment difference of 0.03% points for the primary end point in the overall population. A total of 580 patients were determined to be required to be randomly assigned to the trial product to have 90% power to confirm noninferiority. For the hypothesis of superiority of insulin icodec compared to insulin degludec in terms of change from baseline to week 26 in hemoglobin A1C there would be 50% marginal power and also 50% combined power to detect if the true treatment difference in hemoglobin A1C is 0.18% points with the sample size of 580.

ONWARDS 4

The sample size was calculated with the aim of having 90% power for declaring noninferiority with a noninferiority margin of 0.3% points with respect to change in hemoglobin A1C. Observations from studies with basal and bolus treatment (NN9068-4185⁶¹ and NN1250-3582)⁶⁷ were used to determine the number of patients required to be randomized and the expected treatment difference in the primary end point. Therefore, a total of 580 patients were required to be randomized to ensure 90% power for confirming noninferiority with an assumption of a mean treatment difference of 0.03% points for the primary end point in the overall population. A screen failure rate of 25% was expected, resulting in the need to screen approximately 774 patients to achieve 580 patients.

ONWARDS 5

The sample size was calculated with the aim of having 90% power for declaring noninferiority with a noninferiority margin of 0.3% points with respect to change in hemoglobin A1C. Assumptions for sample size and expected treatment difference for the primary end point in the ONWARDS 5 trial were based on observations from the CONFIRM study⁶⁸ (a study that compared the real-world effectiveness of insulin degludec and insulin glargine U300 in patients who were insulin naive with T2DM in routine US clinical practice), market analytics on retention and switching from a basal insulin, phase III RCTs (NN1250,⁶⁷ NN9068⁶¹), and the Explorys database (a database of electronic medical records representing 39 integrated health care delivery networks in the US). Therefore, a treatment difference of 0.0% points in patients completing 52 weeks of treatment, and a treatment difference of 0.045% points in favour of basal insulin analogues for patients discontinuing randomized treatment before week 52 were expected. The investigators aimed for the trial to be powered to demonstrate noninferiority of insulin icodec used with DoseGuide versus once-daily basal insulin analogues. They also wanted the trial to have reasonable power to detect a treatment difference of 0.3% points in favour of insulin icodec used with DoseGuide in all randomized patients as being statistically significantly different from 0.0% points for the primary end point of change in hemoglobin A1C from baseline to week 52. A total of 1,092 patients were planned to be randomized to the trial to have 90% power to confirm noninferiority and to have 97% marginal power to detect a treatment difference of 0.3% points in favour of insulin icodec used with DoseGuide as being statistically significant different from 0.0% points. A screen failure rate of 9% was expected resulting in the need to screen approximately 1,200 patients to have 1,096 patients randomly assigned to the trial product.

Statistical Testing

The primary estimand for all ONWARDS studies was defined as the treatment difference in change in hemoglobin A1C from baseline to week 26 or week 52 between insulin icodec and once-daily insulin analogues for all randomized patients, regardless of adherence to randomized treatment and changes made to antihyperglycemic background medication. An intent-to-treat approach was used for assessment of the primary estimand.

For all trials, a noninferiority margin of 0.3% points was chosen based on the recommendation from the FDA guidance for industry on developing drugs for treatment of diabetes.¹⁷ This margin is also considered to provide sufficient assay sensitivity based on the following considerations.

• ONWARDS 1, ONWARDS 3, ONWARDS 5 (insulin naive): The margin represents less than 50% of a suitable conservative estimate of insulin glargine’s treatment effect on hemoglobin A1C in a placebo-controlled trial (–0.85% points; 95% CI, −1.04% to −0.66%) in patients who were insulin naive.⁶⁹ As only insulin glargine (not insulin degludec) has been compared to placebo in previous trials, the above-mentioned insulin glargine results were used as reference for consideration of the margin.

• ONWARDS 2, ONWARDS 4 (insulin experienced): The treatment effect of insulin degludec relative to placebo among patients previously treated with basal insulin has not been investigated in trials. Among a similarly progressed patient population (previously treated with liraglutide) at 52 weeks, the between-group difference in change from baseline in hemoglobin A1C for patients randomized to insulin degludec versus placebo was 0.92% points (95% CI, −1.00% to −0.75%).⁷⁰ A 0.30%-point noninferiority margin preserves less than 50% of this estimate of insulin degludec’s effect on hemoglobin A1C relative to placebo. The treatment effect of insulin glargine versus placebo in a basal and bolus insulin treated population is unknown.

In the ONWARDS 1 trial, a 2-sided hierarchical testing procedure was used to control the overall type I error at a 5% level. If the primary end point (noninferiority in glycemic control) was concluded, then confirmatory testing was conducted down the following hierarchy as long as the following confirmatory secondary hypothesis at a given stage was confirmed:

• Insulin icodec is superior to insulin glargine in terms of time in range (3.9 mmol/L to 10.0 mmol/L) from week 48 to week 52.

• Insulin icodec is superior to insulin glargine in terms of change from baseline to week 52 in hemoglobin A1C.

In the ONWARDS 2, ONWARDS 3, and ONWARDS 5 trials, a 2-sided hierarchical testing procedure was used to control the overall type I error at the 5% level. If noninferiority in glycemic control was demonstrated, then confirmatory testing was conducted to test superiority of insulin icodec (plus DoseGuide in ONWARDS 5) over insulin glargine (ONWARDS 2), insulin degludec (ONWARDS 3), or once-daily basal insulin analogues (ONWARDS 5) in terms of change from baseline in hemoglobin A1C at week 26 (ONWARDS 2 and 3) or week 52 (ONWARDS 5).

No hierarchical testing procedure was conducted in the ONWARDS 4 trial.

Table 8: Statistical Analysis of Efficacy End Points

ANCOVA = analysis of covariance; BG = blood glucose; CGM: continuous glucose monitoring; DTSQ = Diabetes Treatment Satisfaction Questionnaire; FPG = fasting plasma glucose; hemoglobin A1C = glycated hemoglobin; SU = sulfonylurea; TRIM-D = Treatment-Related Impact Measure for Diabetes.

^aAchievement of hemoglobin A1C targets was not a formal end point as part of the ONWARDS studies, although it was prespecified in the statistical analysis plans. This end point was considered to be informative to clinicians treating diabetes. The primary reason this was not included in the ONWARDS studies as a formal end point was to limit the number of end points to reduce biases related to multiplicity and to avoid regulatory guidelines surrounding the acceptance of exploratory end points for approval of new therapies in some countries.

Source: ONWARDS 1 protocol, ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 2 protocol, ONWARDS 2 Clinical Study Report,²⁸ ONWARDS 3 protocol, ONWARDS 3 Clinical Study Report,²⁶ ONWARDS 4 protocol, ONWARDS 4 Clinical Study Report, ONWARDS 5 protocol, and ONWARDS 5 Clinical Study Report.²⁷ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Analysis Populations

The definitions of the analysis sets were the same across all studies (refer to Table 9). In certain exceptional cases, patients or observations could be eliminated from the full analysis set; in these cases, the reasons for the exclusions were documented before unblinding.

No subgroup analyses were conducted.

Table 9: Analysis Populations of the ONWARDS Studies

Source: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 2 Clinical Study Report,²⁸ ONWARDS 3 Clinical Study Report,²⁶ ONWARDS 4 Clinical Study Report,²⁹ and ONWARDS 5 Clinical Study Report.²⁷ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Results

Patient Disposition

Patient disposition of the ONWARDS 1, ONWARDS 3, and ONWARDS 5 studies is summarized in Table 10, and patient disposition of the ONWARDS 2 and ONWARDS 4 studies is summarized in Table 11. Patient disposition (e.g., percent who completed study, discontinuations) was generally balanced between the treatment arms of each trial. While permanent discontinuation of treatment (< 5%) or withdrawal from the trial (< 3%) was uncommon in most of the studies, approximately 10% of patients in each treatment group in the ONWARDS 5 study permanently discontinued the trial product, and approximately 9% of patients in each group withdrew from the trial. In the ONWARDS 4 study, across the treatment groups, 6% to 8% of patients discontinued the trial product and 8% to 9% of patients withdrew from the trial.

Table 10: Summary of Patient Disposition From ONWARDS 1, 3, and 5 — Patients With T2DM Who Were Insulin Naive

FAS = full analysis set; NR = not reported; T2DM = type 2 diabetes mellitus.

^aOnce-daily basal insulin analogues: insulin degludec, insulin glargine (U100 and U300).

^bOne patient was classified as “withdrawal of consent” but returned for follow-up and end-of-trial visit (n = 1).

^cWeek 26 visit for ONWARDS 3.

^dWeek 52 visit for ONWARDS 1 and ONWARDS 5.

Source: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 3 Clinical Study Report,²⁶ and ONWARDS 5 Clinical Study Report.²⁷ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Table 11: Summary of Patient Disposition From ONWARDS 2 and 4 — Patients With T2DM Who Were Basal or Bolus Insulin Experienced

FAS = full analysis set; T2DM = type 2 diabetes mellitus.

Source: ONWARDS 2 Clinical Study Report²⁸ and ONWARDS 4 Clinical Study Report.²⁹ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Baseline Characteristics

Relevant baseline demographic and disease characteristics of patients randomized into the included trials are summarized for the insulin-naive populations (ONWARDS 1, 3, and 5) in Table 12 and for the insulin-experienced populations (ONWARDS 2 and 4) in Table 13.

In the insulin-naive populations (ONWARDS 1, 3, and 5), the mean age of patients in each treatment arm was approximately 58 years to 59 years, and a minority of patients were aged 75 years or older. More patients were male than female in each trial, and the percent male ranged from 54% to 63% across treatment arms. The ONWARDS 1 and ONWARDS 3 studies included sites in Asia, Europe, North America, and South America, while the ONWARDS 5 study included sites in Europe and North America only. No sites were located in Africa. The majority of patients were of white race in all trials; in the ONWARDS 1 and ONWARDS 3 studies, 61% to 68% of each treatment arm were white, while in the ONWARDS 5 study, 88% to 90% were white. In the ONWARDS 1 and ONWARDS 3 studies, the next most represented race was Asian at approximately 26% to 29%, followed by Black or African American, Other, and American Indian or Alaska Native. In the ONWARDS 5 study, the second most represented race was Black or African American (4.4% to 5.2%) followed by Asian (3.5% to 5.2%), while 0.2% to 0.4% were American Indian or Alaska Native, and 0.2% to 1.3% were Other. In each trial, more than half of patients had never smoked, while approximately one-quarter to one-third of patients were previous smokers, and the remainder (< 20%) were current smokers. On average, patients had between 11 years and 12 years of diabetes history and a mean hemoglobin A1C of 8.44% to 8.96% at baseline. Approximately 90% of patients were receiving metformin at baseline, and other common (> 15%) antihyperglycemic background medications included sulfonylureas, SGLT2is, DPP-4is, and GLP-1 RAs. Uncommon antihyperglycemic background medications included thiazolidinediones, alpha-glucosidase inhibitors, and glinides. There were no notable between-group differences in antihyperglycemic background medication at screening.

In the insulin-experienced populations (ONWARDS 2 and 4), the mean age of patients in each treatment arm was approximately 60 years to 62 years, and 3.4% to 7.2% were aged 75 years or older. The percentage who were male was 53% to 62% while 37% to 47% were female. Both studies included Asian, European, and North American sites; the ONWARDS 2 study also included African sites, while the ONWARDS 4 study also included South American sites. The racial distributions were similar to those described in the ONWARDS 1 and ONWARDS 3 studies, where the majority of patients were white (> 50% in each treatment arm), followed in sequence by Asian (32% to 42%), Black or African American (3% to 5%), and finally Other (0% to 2%) or American Indian or Alaska Native (0% to 1%). Similar to the insulin-naive populations, more than half of the patients had never smoked, while approximately one-quarter to one-third of patients were previous smokers, and fewer than 15% per treatment arm were current smokers. The mean duration of diabetes in each treatment arm ranged from 16 years to 18 years, and the mean hemoglobin A1C was 8.10% to 8.31%. Patients in the ONWARDS 2 and ONWARDS 4 studies received prior basal insulins, while patients in the ONWARDS 4 study also received basal and bolus insulins. Across both the ONWARDS 2 and ONWARDS 4 trials, most patients received once-daily regimens of basal insulins, most commonly insulin glargine U100 (43%), insulin degludec (26%), and insulin glargine U300 (18%), with few patients receiving neutral protamine Hagedorn insulin (5%) and insulin detemir (4%). In the ONWARDS 4 trial, the most commonly administered combination basal-bolus therapy was once-daily basal insulin with bolus insulin 3 times daily (75%) followed by once-daily basal insulin with twice-daily bolus insulin (20%).

Table 12: Summary of Baseline Characteristics in Insulin-Naive Populations — ONWARDS 1, 3, and 5 (FAS)

BMI = body mass index; eGFR = estimated glomerular filtration rate; FAS = full analysis set; FPG = fasting plasma glucose; hemoglobin A1C = glycated hemoglobin NA = not applicable; NR = not reported; SD = standard deviation.

Source: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 3 Clinical Study Report,²⁶ and ONWARDS 5 Clinical Study Report.²⁷ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Table 13: Summary of Baseline Characteristics in Insulin-Experienced Populations — ONWARDS 2 and 4 (FAS)

b.i.d. = twice a day; BMI = body mass index; CGM = continuous glucose monitoring; eGFR = estimated glomerular filtration rate; FAS = full analysis set; FGM = flash glucose monitoring; FPG = fasting plasma glucose; hemoglobin A1C = glycated hemoglobin; NA = not applicable; q.d. = once a day; q.i.d. = 4 times a day; SD = standard deviation; t.i.d. = 3 times a day.

Source: ONWARDS 2 Clinical Study Report²⁸ and ONWARDS 4 Clinical Study Report.²⁹ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Exposure to Study Treatments

Exposure to study treatments is summarized in Table 14 (ONWARDS 1, 3, and 5) and Table 15 (ONWARDS 2 and 4).

Table 14: Exposure to Study Treatments for ONWARDS 1, 3, and 5 — Patients Who Were Insulin Naive

ANOVA = analysis of variance; CI = confidence interval; FAS = full analysis set; LS = least squares; SAS = safety analysis set; SD = standard deviation.

^aOne patient-year of exposure = 365.25 days.

^bReported from SAS in ONWARDS 3, and FAS in ONWARDS 1 and ONWARDS 5.

^cP value: Two-sided P value for test of no treatment difference. No correction for multiplicity. The log-transformed response from week 50 to week 52 or week 24 to 26 is analyzed using an ANOVA model with treatment and region as fixed factors. Missing mean values from week 50 to week 52 or week 24 to 26 are imputed using multiple imputation based on values for participants from the comparator group who completed randomized treatment without initiation of bolus insulin for more than 2 weeks at any time before the week 52 visit. Each imputed dataset is analyzed separately, and estimates are combined using Rubin’s rules.

Source: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 3 Clinical Study Report,²⁶ and ONWARDS 5 Clinical Study Report.²⁷

Table 15: Exposure to Study Treatments for ONWARDS 2 and 4 — Patients Who Were Insulin Experienced (FAS)

ANOVA = analysis of variance; CI = confidence interval; FAS = full analysis set; LS = least squares; NA = not applicable; NR = not reported; SAS = safety analysis set; SD = standard deviation.

^aOne patient-year of exposure = 365.25 days.

^bP value: Two-sided P value for test of no treatment difference. No correction for multiplicity. The log-transformed response from week 50 to week 52 or week 24 to 26 is analyzed using an ANOVA model with treatment and region as fixed factors. Missing mean values from week 50 to week 52 or week 24 to 26 are imputed using multiple imputation based on values for participants from the comparator group who completed randomized treatment without initiation of bolus insulin for more than 2 weeks at any time before the week 52 visit. Each imputed dataset is analyzed separately, and estimates are combined using Rubin's rules.

Source: ONWARDS 2 Clinical Study Report²⁸ and ONWARDS 4 Clinical Study Report.²⁹

Efficacy

Key efficacy outcomes are summarized in Table 16 for insulin-naive populations (ONWARDS 1, 3, and 5), and in Table 17 for insulin-experienced populations (ONWARDS 2 and 4).

Change in Hemoglobin A1C From Baseline

Patients Who Were Insulin Naive (ONWARDS 1, ONWARDS 3, and ONWARDS 5)

In the primary analyses for noninferiority, the between-group differences in change from baseline in hemoglobin A1C was –0.19% points (95% CI, –0.36% to –0.03%; P < 0.0001) in the ONWARDS 1 study at 52 weeks, –0.21% points (95% CI, –0.34% to –0.08%; P < 0.0001) in the ONWARDS 3 study at 26 weeks, and –0.38% points (95% CI, –0.66% to –0.09%; P < 0.0001) in the ONWARDS 5 study at 52 weeks.

In each of the ONWARDS 1, ONWARDS 3, and ONWARDS 5 studies, a 2-dimensional tipping point sensitivity analysis was performed to evaluate the robustness of the assumptions regarding missing data; the results were consistent with the primary analysis for noninferiority of hemoglobin A1C and confirmed that insulin icodec was noninferior to the comparator in question.

In the secondary analyses for superiority, the P values were 0.0210, 0.0016, and 0.0092, respectively.

Insulin-Experienced Patients Who Were Insulin Experienced (ONWARDS 2 and ONWARDS 4)

In the primary analyses for noninferiority, estimated treatment difference of change from baseline in hemoglobin A1C was –0.22% points (95% CI, –0.37% to –0.08%; P < 0.0001) in the ONWARDS 2 study and 0.02% points (95% CI, –0.11% to 0.15%; P < 0.0001) for the ONWARDS 4 study.

For both the ONWARDS 2 and 4 studies, a 2-dimensional tipping point sensitivity analysis was performed to evaluate the robustness of the assumptions regarding missing data; the results were consistent with the primary analysis for noninferiority of hemoglobin A1C and confirmed that insulin icodec was noninferior to the comparator in question.

In the ONWARDS 2 study, a secondary analysis for superiority was also conducted (P = 0.0028). No superiority analysis was conducted in the ONWARDS 4 study.

Time in Range (3.9 mmol/L to 10.0 mmol/L)

Patients Who Were Insulin Naive (ONWARDS 1, ONWARDS 3, and ONWARDS 5)

In the ONWARDS 1 study from week 48 to week 52, the LS mean time in glycemic range was 71.27% (SE = 0.85) for insulin icodec and 67.00% (SE = 0.85) for insulin glargine, representing an estimated treatment difference of 4.27% points (95% CI, 1.92% to 6.62%; P = 0.0004). A 2-dimensional tipping point sensitivity analysis was conducted, which aligned with the primary analysis for time in range.

This outcome was not assessed in the ONWARDS 3 or ONWARDS 5 studies.

Patients Who Were Insulin Experienced (ONWARDS 2 and ONWARDS 4)

In the ONWARDS 2 study from week 22 to week 26, the LS mean time in glycemic range between 3.9 mmol/L and 10.0 mmol/L was 62.34% (SE = 1.16) for the insulin icodec group and 59.93% (SE = 1.16) for the insulin degludec group. The estimated treatment difference between insulin icodec and insulin degludec was 2.41% (95% CI, –0.84% to 5.56%; P = 0.1461).

In the ONWARDS 4 study from week 22 to week 26, the LS mean time in glycemic range between 3.9 mmol/L and 10.0 mmol/L was 66.75% (SE = 1.00) for the insulin icodec group and 66.46% (SE = 1.02) for the insulin glargine group. The estimated treatment difference between insulin icodec and insulin glargine was 0.29% (95% CI, –2.52% to 3.09%; P = 0.8406).

Time Spent Less Than 3.0 mmol/L

Patients Who Were Insulin Naive (ONWARDS 1, ONWARDS 3, and ONWARDS 5)

In the ONWARDS 1 study from week 48 to week 52, the LS mean time in glycemic range below 3 mmol/L was 0.21% (SE = not reported) for insulin icodec and 0.16% for insulin glargine (SE = not reported), representing an estimated treatment ratio (insulin icodec:insulin glargine) of 1.27 (95% CI, 0.94 to 1.71; P = 0.1134).

This outcome was not assessed in the ONWARDS 3 or 5 studies.

Patients Who Were Insulin Experienced (ONWARDS 2 and ONWARDS 4)

In the ONWARDS 2 study from week 22 to week 26, the LS mean time in glycemic range below 3.0 mmol/L was 0.33% in the insulin icodec group and 0.24% in the insulin degludec group. The estimated treatment ratio of insulin icodec and insulin degludec was 1.37 (95% CI, 0.92 to 2.04; P = 0.1180).

In the ONWARDS 4 study from week 22 to week 26, the LS mean time in glycemic range below 3.0 mmol/L was 0.69% in the insulin icodec group and 0.58% in the insulin glargine group. The estimated treatment ratio of insulin icodec and insulin glargine was 1.20 (95% CI, 0.91 to 1.58; P = 0.2050).

Time Spent Greater Than 10.0 mmol/L

Patients Who Were Insulin Naive (ONWARDS 1, ONWARDS 3, and ONWARDS 5)

In the ONWARDS 1 study from week 48 to week 52, the LS mean time in glycemic range above 10 mmol/L was 27.56% for insulin icodec and 32.13% for insulin glargine, representing an estimated treatment difference of –4.58% (95% CI, –6.99% to –2.17%; P = 0.0002).

This outcome was not assessed in the ONWARDS 3 or ONWARDS 5 studies.

Patients Who Were Insulin Experienced (ONWARDS 2 and ONWARDS 4)

In the ONWARDS 2 study from week 22 to week 26, the LS mean time in glycemic range above 10 mmol/L was 36.34% (SE = 1.19) for insulin icodec and 39.28% (SE = 1.19) for insulin degludec, representing an estimated treatment difference of –2.93% (95% CI, –6.25% to 0.39%; P = 0.0833).

In the ONWARDS 4 study from week 22 to week 26, the LS mean time in glycemic range above 10 mmol/L was 30.64% for insulin icodec (SE = 1.03) and 31.24% (SE = 1.04) for insulin glargine, representing an estimated treatment difference of –0.60% (95% CI, –3.47% to 2.28%; P = 0.6826).

Change in Body Weight

Patients Who Were Insulin Naive (ONWARDS 1, ONWARDS 3, and ONWARDS 5)

In the ONWARDS 1 study (at 52 weeks), ONWARDS 3 study (at 26 weeks), and ONWARDS 5 study (at 52 weeks), the between-group differences in change in body weight from baseline were 0.46 kg (95% CI, –0.12 kg to 1.04 kg; P = 0.1187), 0.46 kg (95% CI, –0.19 kg to 1.10 kg; P = 0.1657), and 0.83 kg (95% CI, –0.37 kg to 2.02 kg; P = 0.1747), respectively.

Patients Who Were Insulin Experienced (ONWARDS 2 and ONWARDS 4)

In the ONWARDS 2 study (at 26 weeks) and ONWARDS 4 study (at 26 weeks), the between-group differences in change in body weight from baseline were 1.70 kg (95% CI, 0.76 kg to 2.63 kg; P = 0.0004) and 0.57 kg (95% CI, –0.39 kg to 1.54 kg; P = 0.2444), respectively.

Number of Clinically Significant Level 2 or Level 3 Hypoglycemic Episodes

Patients Who Were Insulin Naive (ONWARDS 1, ONWARDS 3, and ONWARDS 5)

In the ONWARDS 1 study, a similar number of patients experienced level 2 hypoglycemic events in the 2 groups, but there were numerically more level 2 events in the insulin icodec group. There were 143 events in 48 patients (9.8%) in the insulin icodec group and 75 events occurring in 49 patients (10.0%) in the insulin glargine group. In the insulin icodec group, 3 of the 492 patients (0.6%) experienced 61 of the 143 clinically significant hypoglycemic events. The remaining patients in the insulin icodec group and all of the patients in the insulin glargine groups experienced between 1 and 5 episodes of level 2 hypoglycemic events. The estimated treatment ratio for level 2 events (insulin icodec:insulin glargine) was 1.67 (95% CI, 0.99 to 2.84; P = 0.0561). Severe (level 3) hypoglycemic events occurred in 1 patient (0.2%) in the insulin icodec group, and 3 patients (0.6%) in the insulin glargine group.

In the ONWARDS 3 study, there were 53 clinically significant events of hypoglycemia (level 2) reported in 26 patients (8.9%) in the insulin icodec group, and 23 events occurring in 17 patients (22.1%) in the insulin degludec group. In the insulin icodec group, 2 patients (0.7%) experienced 15 of the 53 clinically significant hypoglycemic events. The remaining patients in the insulin icodec group experienced between 1 and 4 episodes of level 2 hypoglycemic events. Patients in the insulin degludec group experienced between 1 and 3 episodes of level 2 hypoglycemic events. The estimated treatment ratio for level 2 events (insulin icodec:insulin degludec) was 2.09 (95% CI, 0.99 to 4.41; P = 0.0536). Severe hypoglycemic events occurred in zero patients in the insulin icodec group, and 2 patients (0.7%) in the insulin degludec group.

In the ONWARDS 5 study, there were 104 clinically significant (level 2) hypoglycemic events that were reported in 64 patients (11.8%) in the insulin icodec group, and 81 events occurring in 45 patients (8.4%) in the once-daily analogues group. The estimated treatment ratio for level 2 events (insulin icodec:once-daily analogues) was 1.23 (95% CI, 0.77 to 1.98; P = 0.3928). Severe (level 3) hypoglycemic events occurred in zero patients in the insulin icodec group, and 4 patients (0.7%) in the insulin glargine group.

Patients Who Were Insulin Naive (ONWARDS 2 and ONWARDS 4)

In the ONWARDS 2 study, there were 113 clinically significant events of hypoglycemia (level 2) reported in 37 patients (14.1%) in the insulin icodec group, and 41 events occurring in 19 patients (7.2%) in the insulin degludec group. The estimated treatment ratio for level 2 events (insulin icodec:insulin degludec) was 1.98 (95% CI, 0.95 to 4.12; P = 0.0677). Severe hypoglycemic events occurred in zero patients in the insulin icodec group, and 1 patient (0.4%) in the insulin degludec group.

In the ONWARDS 4 study, clinically significant events of hypoglycemia (level 2) were reported in 148 patients (50.9%) in the insulin icodec group, and 160 patients (55.0%) in the insulin glargine group. The estimated treatment ratio for level 2 events (insulin icodec/insulin glargine) was 0.99 (95% CI, 0.73 to 1.34; P = 0.9274). Severe (level 3) hypoglycemic events occurred in 4 patients (1.4%) in the insulin icodec group, and 2 patients (0.7%) in the insulin glargine group.

Diabetes Treatment Satisfaction Questionnaire

This outcome was assessed only in the ONWARDS 5 and ONWARDS 2 studies.

Patients Who Were Insulin Naive (ONWARDS 5)

In the ONWARDS 5 study, the observed mean DTSQ total score at baseline was 26.15 in the insulin icodec plus DoseGuide group and 26.77 in the once-daily analogues group. The estimated LS mean DTSQ total score at week 52 was 31.13 (SE = 0.25) in the insulin icodec group and 30.35 (SE = 0.25) in the once-daily analogues group, representing a LS mean change from baseline in DTSQ total satisfaction score of 4.68 (SE = 0.25) and 3.90 (SE = 0.25), respectively. The LS mean difference between groups was 0.78 (95% CI, 0.10 to 1.47; P = 0.0247).

Patients Who Were Insulin Experienced (ONWARDS 2)

In the ONWARDS 2 study, the observed mean DTSQ total score at baseline was 26.76 in the insulin icodec group and 26.69 in the insulin degludec group. The estimated LS mean DTSQ total score at week 26 was 30.95 (SE = 0.30) in the insulin icodec group and 29.69 (SE = 0.31) in the insulin degludec group, representing a LS mean change from baseline in DTSQ total satisfaction score of 4.22 (SE = 0.30) and 2.96 (SE = 0.31), respectively. The estimated LS mean treatment difference between insulin icodec and insulin degludec was 1.25 (95% CI, 0.41 to 2.10; P = 0.0036).

TRIM-D Compliance Domain

This outcome was assessed only in the ONWARDS 5 study. The estimated treatment difference was 3.04 (95% CI, 1.28 to 4.81; P = 0.0007) at 52 weeks.

CV Death

CV death was not measured as an outcome in the included trials.

Nonfatal MI

Nonfatal MI was not measured as an outcome in the included trials.

Nonfatal Stroke

Nonfatal stroke was not measured as an outcome in the included trials.

Other Microvascular and Macrovascular Complications of T2DM

Other microvascular and macrovascular complications of T2DM were not measured as outcomes in the included trials.

Table 16: Summary of Key Efficacy End Points for ONWARDS 1, 3, and 5 — Patients Who Were Insulin Naive

BG = blood glucose; CI = confidence interval; DTSQ = Diabetes Treatment Satisfaction Questionnaire; FAS = full analysis set; hemoglobin A1C = glycated hemoglobin; LS = least squares; NR = not reported; PYE = patient-years of exposure; SAS = safety analysis set; SE = standard error.

^aP value has been adjusted for multiple testing. P values not accompanied by this footnote are not adjusted for multiple testing.

^bLevel 2 hypoglycemic episodes are defined as “clinically significant” episodes and level 3 hypoglycemic episodes are defined as “severe” episodes.

^cOne PYE = 365.25 days.

Source: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 3 Clinical Study Report,²⁶ and ONWARDS 5 Clinical Study Report.²⁷ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Table 17: Summary of Key Efficacy End Points for ONWARDS 2 and 4 — Patients Who Were Insulin Experienced

BG = blood glucose; CI = confidence interval; DTSQ = Diabetes Treatment Satisfaction Questionnaire; FAS = full analysis set; hemoglobin A1C = glycated hemoglobin; LS = least squares; NR = not reported; PYE = patient-years of exposure; SAS = safety analysis set; SE = standard error.

^aP value has been adjusted for multiple testing. P values not accompanied by this footnote are not adjusted for multiple testing.

^bLevel 2 hypoglycemic episodes are defined as “clinically significant” episodes and level 3 hypoglycemic episodes are defined as “severe” episodes.

^cOne PYE = 365.25 days.

Source: ONWARDS 2 Clinical Study Report²⁸ and ONWARDS 4 Clinical Study Report.²⁹ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Harms

Refer to Table 18 and Table 19 for harms data in the insulin-naive populations (ONWARDS 1, 3, and 5) and insulin-experienced populations (ONWARDS 2 and 4), respectively.

Adverse Events

The proportion of AEs was similar between the insulin icodec and once-daily insulin analogue comparator groups in all ONWARDS studies. The most common AEs were COVID-19, nasopharyngitis, diarrhea, and back pain. The majority of AEs were determined by the study investigators to be nonserious, mild to moderate in severity, unlikely related to trial products, and recovered or recovering by the end of the trial duration in each trial.

In the insulin-naive populations (ONWARDS 1, 3, and 5), 50% to 71% of patients across each treatment arm experienced at least 1 AE.

In the insulin-experienced populations (ONWARDS 2 and 4), 51% to 62% of patients across each treatment arm experienced at least 1 AE.

Serious AEs

SAEs occurred in similar proportions across both the insulin icodec groups and the once-daily analogues groups in each trial.

In the insulin-naive populations (ONWARDS 1, 3, and 5),among patients treated with insulin icodec, 5.1% to 10.4% of patients had at least 1 SAE. In these same trials, among patients treated with the comparator once-daily analogues, 5.1% to 10.6% of patients had at least 1 SAE.

In the insulin-experienced populations (ONWARDS 2 and 4), among patients treated with insulin icodec or once-daily analogues, 7.6% to 8.4% and 6.1% to 8.6% had at least 1 SAE, respectively.

Reported SAEs included: cardiac disorders; infections and infestations; injury, poisoning, and procedural; vascular disorders; musculoskeletal and connective tissue disorders; neoplasms benign, malignant and unspecified nervous system disorders; eye disorders; reproductive system and breast disorders; respiratory, thoracic, and mediastinal disorders; blood and lymphatic system disorders; congenital, familial, and genetic disorders; gastrointestinal disorders; and metabolism and nutrition disorders; renal and urinary disorders. Each of these SAE categories occurred in 0 to less than 5% of patients. The most frequent category of SAE observed was cardiac disorders, which ranged from approximately 3% to 4% of patients in the included studies, followed by infections and infestations (in approximately 2% to 3% of patients). There was no 1 most common event observed.

Withdrawals Due to AEs

In the insulin-naive populations (ONWARDS 1, 3, and 5), permanent discontinuation of the study drug due to AE occurred in 0.7% to 1.2% of patients treated with insulin icodec, and 0.8% to 1.3% of patients treated with comparators (insulin glargine or insulin degludec). In the insulin-experienced populations (ONWARDS 2 and 4), permanent discontinuation of the study drug due to AEs occurred in 1.0% to 1.9% of patients treated with insulin icodec, and 1.0 to 1.1% of patients treated with insulin degludec or insulin glargine. Temporary discontinuation was similarly uncommon, as were AEs leading to dose increases or dose decreases.

Mortality

In the ONWARDS 1 study, there were a total of 8 events in 6 (0.6%) patients with fatal outcomes, with 5 events in 4 (0.8%) patients in the insulin icodec treatment group and 3 events in 2 (0.4%) patients in the insulin glargine treatment group. The events included infections and infestations (n = 2) and 1 each of COVID-19, cardiac disorders (angina pectoris), postoperative infection, pancreatic neoplasm, glioblastoma, unknown cause, and acute coronary syndrome. The death due to unknown cause in the insulin glargine treatment group was judged by investigators as “possibly” related to the trial product.

In the ONWARDS 3 study, there were a total of 2 events occurring among 2 patients (0.7%) in the insulin icodec group, and 1 event in 1 patient (0.3%) in the insulin degludec group leading to fatal outcomes. In the insulin icodec group, deaths were due to malignancy and an undetermined cause (n = 1 for each). In the insulin degludec group, death was due to acute MI (n = 1).

In the ONWARDS 5 study, there were a total of 3 events occurring among 3 patients (0.6%) in the insulin icodec plus DoseGuide group, and 8 events in 7 patients (1.3%) in the once-daily analogues group leading to fatal outcomes. In the insulin icodec plus DoseGuide group, deaths were due to an undetermined cause (n = 2) and malignancy (n = 1). In the once-daily analogues group, deaths were due to pulmonary causes (n = 2), undetermined causes (n = 2), malignancy, sudden cardiac death, and heart failure (n = 1 for each).

In the ONWARDS 2 study, there were a total of 3 events in 2 patients (0.8%) in the insulin icodec group and 2 events in 2 patients (0.8%) in the insulin degludec group leading to fatal outcomes. In the insulin icodec group, deaths were due to infection (n = 2). In the insulin degludec group, deaths were due to CV procedures and malignancy (n = 1 for each).

In the ONWARDS 4 study, there were a total of 4 events in 2 patients (0.7%) in the insulin icodec group and 1 event in 1 patient (0.3%) in the insulin glargine group leading to fatal outcomes. In the insulin icodec group, deaths were due to other CV causes and infection (including sepsis) (n = 1 for each). In the insulin glargine group, there was 1 instance of gastrointestinal bleeding which resulted in death.

Notable Harms

Prespecified notable harms included hypersensitivity, injection site reactions, hypoglycemia, and nocturnal hypoglycemia.

Hypersensitivity

Events of hypersensitivity were reported among less than 7% patients during all ONWARDS studies.

In the ONWARDS 1 study, a total of 23 patients (4.7%) experienced 29 events of hypersensitivity in the insulin icodec group, compared to 32 patients (6.5%) who experienced 45 events in the insulin glargine group. No events were considered serious, and all events were considered moderate or mild in severity. In the ONWARDS 3 study, a total of 7 patients (2.4%) experienced 7 events of hypersensitivity in the insulin icodec group, compared to 13 patients (4.4%) who experienced 15 events in the insulin degludec group. No events were considered serious. In the ONWARDS 5 study, a total of 14 patients (2.6%) experienced 29 events of hypersensitivity in the insulin icodec plus DoseGuide group, compared to 14 patients (2.6%) who experienced 16 events in the once-daily insulin analogues group. One patient (0.2%) in the insulin icodec plus DoseGuide group experienced 1 event which was considered serious; no serious hypersensitivity events occurred in the once-daily insulin analogue comparator group.

In the ONWARDS 2 study, a total of 9 patients (3.4%) experienced 9 events of hypersensitivity in the insulin icodec group, compared to 5 patients (1.9%) who experienced 5 events in the insulin degludec group. One patient in each treatment group (0.4%) experienced 1 hypersensitivity event which was considered serious. In the ONWARDS 4 study, a total of 6 patients (2.1%) experienced 6 events of hypersensitivity in the insulin icodec group, compared to 7 patients (2.4%) who experienced 7 events in the insulin glargine group. No events were considered serious.

Injection Site Reactions

Injection site reactions occurred among less than 9% of patients across all ONWARDS studies.

In the ONWARDS 1 study, 6 patients (1.2%) experienced 6 events in the insulin icodec group compared to 12 patients (2.4%) experiencing 12 events in the insulin glargine group. All events were considered mild or moderate in severity. In the ONWARDS 3 study, 25 patients (8.5%) experienced 62 events in the insulin icodec group compared to 13 patients (4.4%) who experienced 22 events in the insulin degludec group. Of the 62 injection site reactions reported in the insulin icodec group, 24 events were reported by only 2 patients. No events were considered serious. In the ONWARDS 5 study, 5 patients (0.9%) experienced 6 events in the insulin icodec plus DoseGuide group compared to 7 patients (1.3%) who experienced 28 events in the once-daily insulin analogue group. No events of injection site reactions were considered serious.

In the ONWARDS 2 study, 3 patients (1.1%) experienced 3 events in the insulin icodec group compared to 1 patient (0.4%) who experienced 1 event in the insulin degludec group. All events of injection site reactions were considered mild or moderate in severity. In the ONWARDS 4 study, 2 patients (0.7%) experienced 2 events in both the insulin icodec and insulin glargine groups. No events of injection site reactions were considered serious, and all were mild in severity.

Hypoglycemia

In the ONWARDS 1 study, there was a greater proportion of level 1 hypoglycemic events in the insulin icodec group with a total of 232 patients (47.2%) compared to 191 patients (38.8%) in the insulin glargine group. There were similar proportions of patients who experienced level 2 events, but more level 2 events in total occurred in the insulin icodec group than the insulin glargine group, and a small number of patients experienced numerous level 2 events; level 3 events were rare in both groups (refer to hypoglycemia efficacy outcomes previously reported for additional detail on level 2 and 3 events).

In the ONWARDS 3 study, there was a greater proportion of level 1 hypoglycemic events in the insulin icodec group with a total of 145 patients (55.3%) compared to 118 patients (44.9%) in the insulin degludec group. More patients in the insulin icodec group experienced level 2 events, and there was a higher total number of level 2 events in the insulin icodec arm compared to the insulin degludec group, and a small number of patients experienced numerous level 2 events; level 3 events did not occur in the insulin icodec group and were rare in the insulin degludec group (refer to hypoglycemia efficacy outcomes previously reported for additional detail on level 2 and 3 events).

In the ONWARDS 5 study, there was a greater proportion of level 1 hypoglycemic events in the insulin icodec group with a total of 200 patients (36.9%) compared to 153 patients (28.4%) in the once-daily analogues group. There was a larger proportion of patients who experienced level 2 events and more total level 2 events in the insulin icodec group compared to the once-daily analogues group. Severe (level 3) events occurred in no patients in the insulin icodec group and were rare in the insulin glargine group (refer to hypoglycemia efficacy outcomes previously reported for additional detail on level 2 and 3 events).

In the ONWARDS 2 study, there was a greater proportion of level 1 hypoglycemic events in the insulin icodec group with a total of 145 patients (55.3%) compared to 118 patients (44.9%) in the insulin degludec group. There was a larger proportion of patients who experienced level 2 events and more total level 2 events in the insulin icodec group compared to the insulin degludec group. Severe hypoglycemic events occurred in zero patients in the insulin icodec group, and 1 patient (0.4%) in the insulin degludec group (refer to hypoglycemia efficacy outcomes previously reported for additional detail on level 2 and 3 events).

In the ONWARDS 4 study, level 1 hypoglycemic events occurred among 244 patients (83.8%) in the insulin icodec group compared to 251 patients (86.3%) in the insulin glargine group. There was a lower proportion of patients who experienced level 2 events and a lower total number of level 2 events in the insulin icodec group compared to the insulin glargine group. Level 3 events were rare in both groups (refer to hypoglycemia efficacy outcomes previously reported for additional detail on level 2 and 3 events).

Nocturnal Hypoglycemia

In the ONWARDS 1 study, level 1 nocturnal hypoglycemic events occurred in 67 patients (13.6%) in the insulin icodec group and 58 patients (11.8%) in the insulin glargine group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 9 patients (1.8%) in the insulin icodec group and 10 patients (2.0%) in the insulin glargine group, and severe (level 3) nocturnal hypoglycemic events occurred in zero patients and 1 patient (0.2%) in the insulin icodec and insulin glargine groups, respectively. The estimated treatment ratio between insulin icodec and insulin glargine for clinically significant (level 2) nocturnal hypoglycemic events was 0.92 (95% CI, 0.29 to 2.86; P = 0.8816). The estimated treatment ratio between insulin icodec and insulin glargine for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 0.88 (95% CI, 0.29 to 2.64; P = 0.8189).

In the ONWARDS 3 study, level 1 nocturnal hypoglycemic events occurred in 24 patients (8.2%) in the insulin icodec group and 23 patients (7.8%) in the insulin degludec group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 1 patient (0.3%) in the insulin icodec group and 4 patients (1.4%) in the insulin degludec group. There were no severe (level 3) nocturnal hypoglycemic events in either treatment group. The estimated treatment ratio between insulin icodec and insulin degludec for clinically significant (level 2) nocturnal hypoglycemic events was 2.09 (95% CI, 0.99 to 4.41; P = 0.0536). The estimated treatment ratio between insulin icodec and insulin degludec for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 1.82 (95% CI, 0.87 to 3.80; P = 0.1091).

In the ONWARDS 5 study, level 1 nocturnal hypoglycemic events occurred in 48 patients (8.9%) in the insulin icodec group and 46 patients (8.6%) in the once-daily analogues group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 11 patients (2.0%) in both treatment groups, and severe (level 3) nocturnal hypoglycemic events occurred in zero patients in the insulin icodec group and 1 patient (0.2%) in the once-daily analogues groups. The estimated treatment ratio between insulin icodec and once-daily analogues for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 1.36 (95% CI, 0.82 to 2.27; P = 0.02396).

In the ONWARDS 2 study, level 1 nocturnal hypoglycemic events occurred in 60 patients (22.9%) in the insulin icodec group and 35 patients (13.3%) in the insulin degludec group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 16 patients (6.1%) in the insulin icodec group and 9 patients (3.4%) in the insulin degludec group. Severe (level 3) nocturnal hypoglycemic events occurred in zero patients in both treatment groups. The estimated treatment ratio between insulin icodec and insulin degludec for clinically significant (level 2) nocturnal hypoglycemic events was 1.98 (95% CI, 0.95 to 4.12; P = 0.0677). The estimated treatment ratio between insulin icodec and insulin degludec for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 1.93 (95% CI, 0.93 to 4.02; P = 0.0782).

In the ONWARDS 4 study, level 1 nocturnal hypoglycemic events occurred in 108 patients (37.1%) in the insulin icodec group and 132 patients (45.4%) in the insulin glargine group. Clinically significant (level 2) nocturnal hypoglycemic events occurred in 54 patients (18.6%) in the insulin icodec group and 71 patients (24.4%) in the insulin glargine group, and severe (level 3) nocturnal hypoglycemic events occurred in zero patients and 1 patient (0.3%) in the insulin icodec and insulin glargine groups, respectively. The estimated treatment ratio between insulin icodec and insulin glargine for clinically significant (level 2) nocturnal hypoglycemic events was 0.74 (95% CI, 0.47 to 1.15; P = 0.1818). The estimated treatment ratio between insulin icodec and insulin glargine for clinically significant (level 2) or severe (level 3) nocturnal hypoglycemic events was 0.73 (95% CI, 0.47 to 1.14; P = 0.1694).

Table 18: Summary of Harms Results in ONWARDS 1, 3, and 5 — Patients Who Were Insulin Naive (SAS)

PYE = patient-year of exposure; SAE = serious adverse event; SAS = safety analysis set.

^a1 PYE = 365.25 days.

^bMore than 5% in any treatment group.

Source: ONWARDS 1 Clinical Study Report,²⁵ ONWARDS 3 Clinical Study Report,²⁶ and ONWARDS 5 Clinical Study Report.²⁷ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Table 19: Summary of Harms Results in ONWARDS 2 and 4 — Patients Who Were Insulin Naive (SAS)

PYE = patient-years of exposure; SAE = serious adverse event; SAS = safety analysis set.

Note: Values are n (%) unless otherwise indicated.

^aOne PYE = 365.25 days.

^bMore than 5% in any treatment group.

Source: ONWARDS 2 Clinical Study Report²⁸ and ONWARDS 4 Clinical Study Report.²⁹ Details included in the table are from the sponsor’s Summary of Clinical Evidence.¹

Critical Appraisal

Internal Validity

All of the ONWARDS trials were randomized, active-controlled trials with adequate methodology related to randomization and allocation concealment, and there were no concerning between-arm imbalances in patient characteristics at baseline, nor in diabetes-related background medications. As such, the risk of bias arising from the randomization process is low in all trials. Each trial was adequately powered for the purpose of their primary hypotheses. The ONWARDS 1, ONWARDS 2, ONWARDS 4, and ONWARDS 5 studies were open-label trials, while the ONWARDS 3 study was double-blinded with adequate blinding and concealment procedures including placebos matched in visual quality and administration methods to the active trial products.

There is an increased risk of bias associated with the open-label design of the ONWARDS 1, ONWARDS 2, ONWARDS 4 and ONWARDS 5 studies with respect to the measurement of the subjective outcomes due to the impact of knowledge of treatment assignment on both the patient (for self-reported outcomes) and the evaluating clinicians; however, this does not apply to objective measures such as the primary outcome (change in hemoglobin A1C from baseline), and other objective outcomes related to, for example, blood glucose monitoring. This may be a concern for subjective assessments, which in this case includes the DTSQ, TRIM-D, and subjective harms. As the intention of the DTSQ is to evaluate satisfaction with treatment and a key difference of the intervention versus comparators is the administration schedule, a blinded assessment of DTSQ with matched placebos would have been uninformative for this research question. The extent and direction of the potential bias could not be ascertained.

The intervention in each trial was insulin icodec; the comparator product was 1 or either of the once-daily insulins, insulin degludec or insulin glargine. Additionally, in the ONWARDS 5 study, patients in the insulin icodec arm used the DoseGuide app to inform dosage adjustments; however, there was no notable inconsistency in the results of the ONWARDS 5 study compared to the trials that did not use the DoseGuide app. The ONWARDS 4 study also differed from the other trials in that patients in both treatment groups also received insulin aspart after mealtimes (2 times to 4 times daily), but again there was no notable inconsistency in the results of the ONWARDS 4 study compared to other trials, and this trial feature was appropriate for the patient subpopulation recruited to the ONWARDS 4 study.

The primary outcome in each trial was the change in hemoglobin A1C from baseline, and the noninferiority margin of 0.3% points was chosen based on established FDA guidance¹⁷ and previous trials of insulin products in the treatment of T2DM. Change in hemoglobin A1C from baseline was considered a clinically relevant outcome by the clinical expert consulted by CADTH. This outcome is considered acceptable by the FDA¹⁷ for trials of new antihyperglycemic therapies seeking a glycemic control indication, the rationale being that it is a validated surrogate of microvascular disease risk reduction,¹⁷ and further it is currently recognized as the key surrogate marker for the development of long-term diabetes complications in people with type 1 diabetes mellitus or T2DM.¹⁸ The selection of this noninferiority margin was determined based on FDA guidance¹⁷ as previously described, and was considered clinically relevant as a threshold of MID according to the clinical expert consulted by CADTH. However, hemoglobin A1C is ultimately a surrogate biomarker, and there is evidence to suggest that hemoglobin A1C may not be appropriate as a surrogate outcome for downstream complications in diabetes trials due to poor associations with mortality,^19,20 CV mortality,²¹ MI,¹⁹ heart failure,¹⁹ kidney injury,¹⁹ and stroke.¹⁹ Other limitations of hemoglobin A1C include a lack of information about acute glycemic events (i.e., hypoglycemia or hyperglycemia) and insensitivity regarding day-to-day variations of glucose, and measurement of hemoglobin A1C can be confounded by other conditions such as anemia, hemoglobinopathies, iron deficiency, and pregnancy.¹⁸

Use of CGM allows for observation of time in and outside of range and daily glycemic variability,¹⁸ and the clinical expert consulted by CADTH indicated that this is of growing importance in clinical trials of glycemic control in patients with T2DM in addition to hemoglobin A1C.¹⁸ Time in range as measured by CGM is a useful as a measure of short-term glycemic control,²² and there is good correlation between time in range and hemoglobin A1C.²² Time in range has been demonstrated to be associated with diabetic retinopathy^23,24 and microalbuminuria²³ but publications assessing this outcome as a surrogate for other diabetes-related complications (e.g., mortality, MI, and other major CV or renal events) were not identified.

The primary outcome in all trials was adjusted for multiple comparisons. Additionally, in the ONWARDS 1 study, the outcome of time in range (3.9 mmol/L to 10 mmol/L) was also adjusted for multiple comparisons. As the remaining outcomes were not adjusted for multiplicity, there is an increased risk of type I error (i.e., false-positive results) for statistically significant results for those outcomes.

In each trial, there were no obvious or substantial imbalances between groups with respect to treatment discontinuation. Notably, however, there was a higher rate of discontinuations in the ONWARDS 5 study overall (approximately 9% to 11% in each arm), compared to less than 5% in each arm in the other ONWARDS trials. The reason for the higher rate in the ONWARDS 5 study compared to other ONWARDS trials is unknown. In the ONWARDS 5 study, the most common reasons for permanent discontinuation were withdrawal of consent or loss to follow-up.

Multiple imputation was used for all outcomes to account for missing data. Multiple imputation methods will not remove or reduce bias that occurs when missingness is not random, but the proportion of missing data in each case was low, so this was not considered cause for concern. Additionally, sensitivity analyses were conducted for the primary outcome which bolstered confidence in the primary analyses.

External Validity

The study designs with respect to patient eligibility criteria and characteristics at baseline were appropriately reflective of the target population in Canada, with the exception that there is a notable lack of inclusion of Indigenous Peoples, who are at higher risk of T2DM and its complications.^6,8 The selected comparators, medications at baseline among included patients, and concomitant mediations during the trials were considered by the consulted clinical expert to be appropriate and to reflect clinical practice in Canada.

The primary, secondary, and exploratory efficacy outcomes were considered appropriate for clinical trial evaluation in this population and the follow-up time was considered adequate, with the exception of all-cause mortality, for which a longer follow-up time would be necessary to compare between treatment arms. As is common in clinical trials, long-term clinical outcomes (i.e., microvascular and macrovascular complications of T2DM) relevant to T2DM were not directly measured due to the limited length of follow-up. Thus, the impact on these outcomes relative to the daily insulins is not known.

The impact of insulin icodec on patients’ HRQoL was not measured in the ONWARDS trials. Although the DTSQ and TRIM-D provide information about treatment satisfaction and compliance, they are not comprehensive measures of HRQoL. As such, the influence of insulin icodec on HRQoL as compared with insulin degludec or insulin glargine is not known. Additionally, there were no compliance data reported for the insulin-experienced populations.

The DoseGuide app used in the intervention arm of the ONWARDS 5 study is not publicly available to patients, so hypothetically could have impacted dosing decisions in a way that is less generalizable to clinical practice in Canada. However, this was not considered to be a major concern, in part because dosing of insulin icodec already differs from dosing of the daily insulins by design and, according to the clinical expert consulted by CADTH, typical practices around dosing adjustments for insulin icodec will be quickly evolving in the event of this therapy becoming available to patients with T2DM. Moreover, the results of the ONWARDS 5 study were aligned with those of the ONWARDS 1 and ONWARDS 3 studies, in which DoseGuide was not used.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for outcomes considered most relevant to inform CADTH’s expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group.^71,72

• High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: We are moderately confident in the effect estimate — the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. We use the word “likely” for evidence of moderate certainty (e.g., “X intervention likely results in Y outcome”).

• Low certainty: Our confidence in the effect estimate is limited — the true effect may be substantially different from the estimate of the effect. We use the word “may” for evidence of low certainty (e.g., “X intervention may result in Y outcome”).

• Very low certainty: We have very little confidence in the effect estimate — the true effect is likely to be substantially different from the estimate of effect. We describe evidence of very low certainty as “very uncertain.”

Following the GRADE approach, evidence from RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty of evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The target of the certainty of evidence assessment was the presence or absence of important effect based on thresholds informed by the clinical expert consulted for this review.

For the insulin-naive population, findings from the ONWARDS 1, ONWARDS 3, and ONWARDS 5 studies were summarized narratively for the hemoglobin A1C, body weight, and hypoglycemia outcomes because these studies were similar in population, interventions, and outcome measures. For all other outcomes only 1 of the ONWARDS studies provided outcome data and as such each was assessed individually. For the insulin-experienced population, the ONWARDS 2 and ONWARDS 4 studies were summarized narratively for all outcomes with the exception of level 2 and 3 hypoglycemia as the point estimates for the between-group differences in the trials suggested different directions of effects. Although the study populations differed in their previous insulin treatment history (the ONWARDS 4 study was basal-bolus while the ONWARDS 2 study was basal only), there was an absence of notable heterogeneity in observed effects. Additional information was requested from the sponsor to obtain 95% CIs for the LS mean estimates in each treatment group within the trials, and to obtain between-group differences with 95% CI for hypoglycemia outcomes.

Results of GRADE Assessments in Table 2 present the GRADE summary of findings for insulin icodec versus insulin glargine or insulin degludec in patients with T2DM who were insulin naive. Table 3 presents the GRADE summary of findings for insulin icodec versus insulin degludec (ONWARDS 2) or insulin glargine (ONWARDS 4) in patients who were insulin experienced with T2DM.

Long-Term Extension Studies

Contents within this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CADTH review team.

Description of Studies

One LTE of the ONWARDS 1 trial was included in the sponsor submission. The purpose of the 26-week extension phase was to continue to compare glycemic control and safety of insulin icodec versus insulin glargine among patients with T2DM who were previously insulin naive, with the primary focus of evaluating long-term safety data. The study design for the on-trial period and LTE phase are shown in Figure 1.

Figure 1: ONWARDS 1 Trial and LTE Study Design

LTE = long-term extension; V = visit.

Source: Details included in the figure are from the sponsor’s ONWARDS 1 Clinical Trial Report.⁷³

Populations

Eligibility criteria for the LTE study were the same as for the ONWARDS 1 trial and are described in the systematic review section of this report. Patient baseline characteristics for the ONWARDS trial overall are also described in the systematic review section of this report; specific baseline characteristics for the sample which completed the LTE phase were not provided in the submission.

Interventions

The experimental and control interventions during the LTE phase of the ONWARDS 1 study were the same as those assigned to patients during the main 52-week trial period. Briefly, patients initially randomized to receive insulin icodec or insulin glargine in the open-label ONWARDS 1 trial continued to receive the same treatment into the LTE phase; doses and administration are described in the systematic review section.

Outcomes

Outcomes during the LTE phase of ONWARDS 1 were assessed in the same manner as those assessed during the main trial phase and are described in the systematic review section of this report. The following efficacy outcomes were summarized by CADTH: change from baseline in hemoglobin A1C, time in range (3.9 mmol/L to 10.0 mmol/L), time spent at less than 3.0 mmol/L, time spent at greater than 10.0 mmol/L, change in body weight, and proportion of patients with level 2 or 3 hypoglycemic events.

Statistical Analysis

The statistical analyses used for the outcomes assessed during the LTE phase of the trial were similar to those used for the primary analyses, with the exception that there was no control for multiplicity of the efficacy outcomes. All efficacy end points apart from hypoglycemic events were assessed for a total of 78 weeks; hypoglycemic events were assessed for 83 weeks until the end of the follow-up period. Time in range (3.9 mmol/L to 10.0 mmol/L), time spent at less than 3.0 mmol/L, and time spent at greater than 10.0 mmol/L were reported as change from week 74 to week 78; all other outcomes were reported as changes from baseline.

Missing data for hemoglobin A1C and body weight were imputed using multiple imputation based on the change from last available value on treatment without initiation of bolus insulin for more than 2 weeks. Missing data for time-in-target range and time spent above or below range were imputed using multiple imputation from patients in the insulin glargine arm who completed their randomized insulin treatment.

Results

Patient Disposition

A summary of the patient disposition after the completion of the LTE phase of the ONWARDS 1 study is provided in Table 20. During the main phase of the ONWARDS 1 study, the majority of patients (98%) completed their final trial visit at week 52, and 97% of patients continued to complete their final week 78 visit. A total of 984 patients (492 patients in each arm) were randomized for the ONWARDS 1 study; 476 (97%) in the insulin icodec arm and 477 (97%) in the insulin glargine arm completed the LTE phase, respectively.

The rates and reasons for trial product discontinuation and trial withdrawal with or without trial product discontinuation were similar between treatment arms, with all withdrawals and/or discontinuations occurring in less than 6% of patients in each arm. Fewer than 1% of patients in each arm were lost to follow-up during the LTE phase. There were 2 discontinuation and withdrawals due to hypoglycemic episodes in the insulin glargine arm and none in the insulin icodec arm.

Table 20: Patient Disposition