Short‐term cost‐utility of degludec versus glargine U100 for patients with type 2 diabetes at high risk of hypoglycaemia and cardiovascular events: A Canadian setting (DEVOTE 9)

Aims To evaluate the short‐term cost‐effectiveness of insulin degludec (degludec) vs insulin glargine 100 units/mL (glargine U100) from a Canadian public healthcare payer perspective in patients with type 2 diabetes (T2D) who are at high risk of cardiovascular events and hypoglycaemia. Materials and methods A decision analytic model was developed to estimate costs (2017 Canadian dollars [CAD]) and clinical outcomes (quality‐adjusted life years [QALYs]) with degludec vs glargine U100 over a 2‐year time horizon. The model captured first major adverse cardiovascular event, death, severe hypoglycaemia and insulin dosing. Clinical outcomes were informed by a post hoc subgroup analysis of the DEVOTE trial (NCT01959529), which compared the cardiovascular safety of degludec and glargine U100 in patients with T2D who are at high cardiovascular risk. High hypoglycaemia risk was defined as the top quartile of patients (n = 1887) based on an index of baseline hypoglycaemia risk factors. Results In patients at high hypoglycaemia risk, degludec was associated with mean cost savings (CAD 129 per patient) relative to glargine U100, driven by a lower incidence of non‐fatal myocardial infarction, non‐fatal stroke and severe hypoglycaemia, which offset the slightly higher cost of treatment with degludec. A reduced risk of cardiovascular death and severe hypoglycaemia resulted in improved effectiveness (+0.0132 QALYs) with degludec relative to glargine U100. In sensitivity analyses, changes to the vast majority of model parameters did not materially affect model outcomes. Conclusion Over a 2‐year period, degludec improved clinical outcomes at a lower cost as compared to glargine U100 in patients with T2D at high risk of cardiovascular events and hypoglycaemia.


| INTRODUCTION
Diabetes is a major global public health concern because of its high prevalence and association with morbidity, mortality and economic burden. 1 In the past decade, the age-standardized prevalence of diabetes in Canada has increased by 37% and is expected to continue to rise, reaching 5 million Canadians (12.1%) by 2025, driven by the increase in and the aging of the Canadian population. 2,3 As a result, approximately 2.6 million Canadians are currently living with diagnosed diabetes. 1 In the 5 years between 2004 and 2008, diabetes was reported as an underlying or contributing cause of death in more than 120 000 Canadians, corresponding to 10.6% of all deaths reported during that period. 4 Cardiovascular disease has been identified as the most common co-existing condition reported on death certificates, appearing nine out of ten times when diabetes was listed as the underlying cause, and underpinning four out of ten deaths where diabetes contributed. 4 Canada is currently one of the top ten countries in terms of expenditure on diabetes, when considering both total expenditure (15 billion international dollars [ID], used by economists to compare values of different currencies) and mean expenditure per person (over 5700 ID). 1 Hypoglycaemia has considerable negative effects on patient quality-of-life, while posing a significant economic burden through increased healthcare resource utilization and loss of productivity. 5,6 The unpleasant symptoms of hypoglycaemia can result in significant anxiety about the possibility of future events, and fear of hypoglycaemia among both patients and clinicians can adversely affect diabetes management and clinical outcomes. 7,8 Severe hypoglycaemic events are associated with an increased risk of adverse clinical outcomes, including major adverse cardiovascular events (MACE), major microvascular events, dementia (in elderly patients) and death. 9,10 Several factors increase the risk of severe hypoglycaemia, including age, diabetes treatment regimen, duration of diabetes, presence of comorbidities, history of previous severe hypoglycaemic episodes and impaired awareness of hypoglycaemia. 11,12 Insulin degludec (degludec) is a basal insulin with an ultra-long duration of action and a stable glucose-lowering profile. 13 Randomized controlled trials have confirmed that similar improvements in glycaemic control can be achieved, with fewer hypoglycaemic episodes, across a broad spectrum of patients with diabetes receiving degludec vs insulin glargine 100 units/mL (glargine U100). 14,15 Thus, it is of interest to investigate cost-effectiveness in patients at high risk of hypoglycaemia who might be expected to benefit most from treatment with degludec compared with glargine U100. In the DEVOTE trial, a cardiovascular outcomes trial (CVOT) that compared the cardiovascular safety of degludec with that of glargine U100, degludec was non-inferior to glargine U100 with respect to the incidence of MACE, but was associated with significantly fewer episodes of severe hypoglycaemia at similar levels of glycaemic control. 16 A cost-utility analysis (CUA) is a cost-effectiveness analysis (CEA) that compares the costs of new interventions with their outcomes measured in utility units, most commonly, the quality-adjusted life year (QALY), and captures both quantity-and quality-of-life over a specified time horizon. Health economic analyses, including CUAs, are an important tool to assist healthcare professionals in deciding how best to allocate resources efficiently between therapy areas and interventions to achieve maximum healthcare gains within a limited budget. 17 Prior to the DEVOTE trial, CEAs of degludec compared with glargine U100 captured the effects of hypoglycaemia rates and insulin dosing over a short-term (1-year) time horizon in patients with type 1 (T1D) or type 2 (T2D) diabetes, based on the phase 3 clinical trial programme. [18][19][20][21][22][23] The DEVOTE trial provided an opportunity to evaluate randomized, double-blind clinical trial data, including cardiovascular endpoints and death, in addition to severe hypoglycaemia rates and insulin dosing, to provide health economic analyses of degludec vs glargine U100 over a 2-year time horizon without extrapolation.
The aim of the present post hoc analysis was to evaluate, from a Canadian public healthcare payer perspective, the short-term costutility of degludec vs glargine U100 in patients with T2D who are at high risk of hypoglycaemia and cardiovascular events.

| DEVOTE trial design
The trial design and primary results of the DEVOTE trial have been published previously. 16,24 In brief, the DEVOTE trial was a randomized, double-blind CVOT that compared the cardiovascular safety of degludec and glargine U100 in patients with T2D at high cardiovascular risk (refer to Supporting Information, Methods for additional details). Patients (N = 7637) were randomly assigned 1:1 to receive once-daily degludec (100 units/mL) or glargine U100, in addition to standard care. The primary endpoint in the DEVOTE trial was time to first MACE, a composite of cardiovascular death, non-fatal myocardial infarction (MI) or non-fatal stroke. Severe hypoglycaemia was selfrecorded in the DEVOTE trial and was defined according to the American Diabetes Association definition as an event requiring third-party assistance. 25 Non-severe hypoglycaemia was not recorded. Events of MI, stroke, death and severe hypoglycaemia were independently adjudicated by the DEVOTE Event Adjudication Committee.
The DEVOTE trial was conducted in accordance with the provisions of the Declaration of Helsinki and the International Conference on Harmonisation Good Clinical Practice Guidelines, the DEVOTE protocol was approved by the independent ethics committee or institutional review board at each trial centre and written informed consent was obtained from each patient before any trial-related activities.  Table 1 and Figure S1). Treatment differences were estimated using hazard, rate and dose ratios from regression analyses. Treatment effects and dose differences were captured only in the case of a statistically significant difference between treatment arms; otherwise, event rates and doses from the glargine U100 arm were modelled in both treatment arms. An annual discount rate of 1.5% was applied to costs and clinical outcomes in the second annual time period, in line with recommendations from the Canadian Agency for Drugs and Technologies in Health. 27

| Simulated cohort and treatment effects
In this analysis, the subgroup at high risk of hypoglycaemia was defined as the top quartile of patients in the DEVOTE trial, based on an index of established risk factors for severe hypoglycaemia at baseline, including age, diabetes duration, HbA1c, sex and insulin regimen.
This internal index of high risk of severe hypoglycaemia was estimated using a Cox regression model of time to first event of severe hypoglycaemia for the full DEVOTE trial population (Table S1)

| Costs, utilities and time horizon
Treatment unit costs were based on Canadian list prices (

| Sensitivity analyses
Deterministic sensitivity analyses were conducted to identify key drivers of outcomes in the base case analysis. Annual discount rates of 0%, 3% or 5% were explored, along with variations in cardiovascular costs (20% lower, 40% lower or UK costs) or severe hypoglycaemia costs (20% lower, 40% lower or UK costs). Additional sensitivity analyses were performed, in which treatment effects (dose, hazard and rate ratios) were applied, regardless of whether they were statistically significant or not, and the glargine U100 distribution of individual MACE components was applied to both arms. Alternative disutilities were explored for cardiovascular events and severe hypoglycaemia, and analyses were conducted in which a utility associated with the flexible dosing of degludec was captured. [32][33][34][35] In an additional sensitivity analysis, long-term costs and clinical outcomes with degludec vs glargine U100 were simulated over a  Table S4. 36 Cost and effectiveness outcomes, the latter expressed in QALYs, were attached to each of the four scenarios detailed in Figure S2 and did not vary between treatment arms. Distribution of patients between scenarios in the two arms was informed by the proportion of the cohort in each state at the end of the base case analysis (Table S5). The same assumptions regarding use of rescue medication were employed in both arms.
A probabilistic sensitivity analysis (PSA) was conducted to quantify the effect of statistical uncertainty around all relevant input parameters concerning cost and effectiveness outcomes (

| Sensitivity analyses
Favourable cost-utility results with degludec as compared to glargine U100 were insensitive to changes in most parameters (Table 4).
Degludec was dominant relative to glargine U100, improving effectiveness while reducing costs, in all but four sensitivity analyses: those in which cardiovascular costs were reduced by 40%, those in which UK costs were substituted for cardiovascular complications or for severe hypoglycaemic events, and those in which additional long-term costs and effects were included in the model. In these analyses, degludec remained highly cost-effective as compared to glargine U100, with incremental cost-utility ratios well below the commonly used Canadian willingness-to-pay threshold of CAD 50000 per QALY. 37 The vast majority of points on the incremental cost-utility scatterplot were on the right-hand side, indicating improved effectiveness with degludec as compared to glargine U100, with most points specifically in the lower right quadrant, demonstrating that costs were also lower with degludec over the 2-year time horizon (Figure 1(A)).
Using PSA results to generate a cost-utility acceptability curve showed that there would be a 91.1% likelihood that treatment with degludec as compared to treatment with glargine U100 would be cost-effective at a willingness-to-pay threshold of CAD 50000 per QALY (Figure 1(B)).

| DISCUSSION
Our short-term modelling analysis suggested that, from a Canadian public healthcare payer perspective, treatment with degludec over Our findings share several similarities with the results of a shortterm CUA in the UK setting that was based on clinical outcomes from the DEVOTE basal-bolus subgroup. 26  Severe hypoglycaemia is associated with higher incidence of various adverse outcomes in patients with diabetes, including MACE, major microvascular complications and death. 9 It is not currently clear if there is a direct causal link between severe hypoglycaemia and adverse outcomes, or whether severe hypoglycaemia is a marker of vulnerability to a range of poor clinical outcomes. 9 Irrespective of the nature of this relationship, our analysis demonstrates that, based on data from a subgroup analysis of DEVOTE data, treatment with degludec as compared to treatment with glargine U100 would improve quality-adjusted life expectancy through a reduced incidence of both severe hypoglycaemia and MACE, at lower cost, in patients at high risk of hypoglycaemia in a Canadian setting.
One of the key advantages of the present analysis is its simplicity and transparency. For instance, clinical outcomes and patient characteristics were taken from a single, high-quality data source. The base case was highly conservative, using only endpoints with significant differences between treatment arms. In addition to severe hypoglycaemia rates and insulin dosing, our short-term CUA also captured MACE and death from other causes to provide a more in-depth evaluation of costs and clinical outcomes without extrapolation. A further strength of this analysis was that it did not rely on long-term cost-effectiveness modelling and the associated assumptions concerning the progression of risk factors for diabetes-related complications or predictions over long timeframes. Given that most complications develop over the course of decades rather than years, however, the short-term time horizon could also be considered a limitation.
The treat-to-target clinical trial design renders the modelling of long-term clinical outcomes as a function of glycaemic control inappropriate, as end-of-trial HbA1c tends to be similar across treatment arms. In contrast, real-world data have demonstrated that switching to degludec from other basal insulins, including glargine U100, under conditions of routine clinical care, results in significant improvements in glycaemic control alongside other clinical benefits for T1D or T2D. 29 By using treat-to-target data from a clinical trial setting, the present analysis may, therefore, underestimate the clinical benefits of degludec experienced in the real-world clinical setting. One final limitation is that our analysis omitted microvascular complications, which can exert a large influence on healthcare costs and patient quality-oflife. 30,38 Microvascular events were not included as defined endpoints in the DEVOTE trial and, therefore, were omitted from our analysis in order to maintain homogeneity of clinical data. While not a limitation per se, it is worth reiterating that the present analysis was conducted in the subgroup of patients in the DEVOTE trial who were at high risk of hypoglycaemia; consequently, the results are unlikely to be applicable to other patient populations.
Our findings suggest that, in Canada, treatment with degludec over a 2-year period is associated with improved effectiveness at lower cost as compared to treatment with glargine U100 in patients with T2D who are at high risk of hypoglycaemia and cardiovascular events. As such, our short-term modelling analysis indicates that degludec would be an efficient use of Canadian public healthcare resources in this patient population.