Modeling the lifetime costs of insulin glargine and insulin detemir in type 1 and type 2 diabetes patients in Canada: a meta-analysis and a cost-minimization analysis

Abstract

Background:

Two basal insulin analogues, insulin glargine once daily and insulin detemir once or twice daily, are marketed in Canada.

Objective:

To estimate the long-term costs of insulin glargine once daily (QD) versus insulin detemir once or twice daily (QD or BID) for type 1 (T1DM) and type 2 (T2DM) diabetes mellitus from a Canadian provincial government’s perspective.

Methods:

A cost-minimization analysis comparing insulin glargine (IGlarg) to insulin detemir (IDet) was conducted using a validated computer simulation model, the CORE Diabetes Model. Lifetime direct medical costs including costs of insulin treatment and diabetes complications were projected. T1DM and T2DM patients’ daily insulin dose (T1DM: IGlarg QD 26.2 IU; IDet BID 33.6 IU; T2DM: IGlarg QD 47.2 IU; IDet QD 65.7 IU or IDet BID 80.4 IU) was derived from a meta-analysis of randomized trials. All patients were assumed to stay on the same treatment for life. Costs were discounted at 5% per annum and reported in 2010 Canadian Dollars.

Results:

The meta-analysis showed T1DM and T2DM patients had similar HbA_1c change from baseline when receiving IGlarg compared to IDet (T1DM: 0.002%-points; p = 0.97; T2DM: −0.05%-points; p = 0.28). Treatment of T1DM patients with IGlarg versus IDet BID resulted in lifetime cost savings of $4231 per patient. Treatment of T2DM patients with IGlarg resulted in lifetime cost savings of $4659 per patient versus IDet QD and cost savings of $8709 per patient versus IDet BID.

Conclusions:

Similar HbA_1c change from baseline can be achieved with a lower IGlarg than IDet dose. From the perspective of a Canadian provincial government, treatment of T1DM and T2DM patients with IGlarg instead of IDet can generate long-term cost savings. Main limitations include trial data were derived from multi-country studies rather than the Canadian population and self-monitoring blood glucose costs were not included.

Key words::

• Canada
• Costs
• Diabetes
• Insulin analogue
• Insulin detemir
• Insulin glargine
• Modeling

Introduction

Diabetes mellitus is a chronic disease that occurs either when the pancreas does not produce enough insulin or when the body cannot effectively use the insulin it produces. Over 90% diagnosed diabetes patients have type 2 diabetes (T2DM), a disease characterized by insulin resistance in peripheral tissue and the progressive decline of insulin secretion by the pancreas in response to plasma glucose levels1. The remaining patients have type 1 diabetes mellitus (T1DM) and require exogenous insulin from diagnosis.

The global prevalence of T1DM and T2DM is estimated to be 285 million (6.6% of the world population) and is projected to reach 439 million patients (7.8%) by 20302. The rise in T2DM prevalence mainly results from the ageing population but also from an increasing obesity and more sedentary lifestyles3. In Canada, the prevalence of diabetes in people over 20 years of age is 2.8 million (9.2%), with an estimated increase to 4 million (10.9%) by 20304.

Both T1DM and T2DM significantly diminish patients’ duration and quality of life. Patients with diabetes are at a higher risk of developing microvascular (e.g., retinopathy, neuropathy) and macrovascular (e.g., stroke, myocardial infarction) diseases compared to the non-diabetic population5,6. In addition, the life expectancy may be shortened by as much as 15 years in T1DM patients and by 5–10 years in T2DM patients2. In North America, 15.7% of adult deaths were found to be attributable to diabetes7. Healthcare expenditure resulting from the treatment of diabetes and its associated complications are estimated to be high and are increasing worldwide. In Canada, they were estimated at 11 billion Canadian dollars for 2010 and are expected to rise to over $15 billion in 20302.

Insulin preparations are a key component of diabetes therapy, as they can supplement insufficient or absent endogenous insulin production as required for glycemic control. For T1DM patients, insulin is the cornerstone of diabetes therapy8, whereas for T2DM patients, insulin is mainly used once adequate glycemic control is not achieved by other measures such as early lifestyle changes or therapy with oral antidiabetic agents8. Insulin analogues target the basal-bolus insulin regimen components separately with a combination of prandial and basal insulins, allowing for more flexible treatment schedules. They are also known to be associated with fewer hypoglycemic events than conventional basal insulins, but are generally more costly alternatives9.

Two basal insulin analogues are currently marketed in Canada, Lantus (insulin glargine; IGlarg) once daily10, and Levemir (insulin detemir; IDet) once or twice daily11. Both these long-acting insulin analogues have been shown to be effective treatment options for patients with diabetes, by being able to achieve lower HbA_1c levels while reducing hypoglycemia compared to intermediate-acting insulin preparations, such as insulin neutral protamine Hagedorn (NPH)12–14. However, due to their higher cost, insulin analogues do not have the same level of reimbursement as NPH across all public drug plans in Canada15. In the past years, the expenditures associated with insulin analogues have increased for the provincial Ministers of Health, resulting from both an increased diabetes prevalence and higher rates of use of these products9,16.

Recent cost-effectiveness analyses have estimated the lifetime costs associated with treatment with IGlarg and IDet in Canada9,17. However, these analyses did not directly compare IGlarg to IDet. Furthermore, recent head-to-head randomized trials have compared the efficacy and safety of IGlarg and IDet in T1DM18,19 and T2DM patients20–23. Previously, relative efficacy estimates could only be derived from indirect comparisons. The lack of a Canada-specific study comparing the cost of IGlarg and IDet raises the need to assess the cost of each insulin analogues treatment strategy in light of the recently published head-to-head clinical trial results.

This study has two subsequent objectives: first, to determine the most appropriate type of economic analysis to estimate the cost of IGlarg and IDet for T1DM and T2DM patients based on the direct comparison trial data; and second, to estimate the most accurate long-term costs of these two drugs from a Canadian provincial government’s perspective.

Methods

Meta-analysis

A meta-analysis was conducted to assess the two products’ relative efficacy and safety. The meta-analysis additionally provided data for the T1DM and T2DM cohorts’ characteristics and estimates of the daily insulin dose per patient. As reported in Table 1, the meta-analysis was based on the results of six recently published head-to-head trials comparing IGlarg with IDet in T1DM18,19 and T2DM20–23. The outcomes of interest were: change from baseline in hemoglobin A_1c (HbA_1c), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), total cholesterol (TC), triglycerides, body weight, and rates of hypoglycemic events. Results were reported either as weighted mean difference (WMD) or in relative risk (RR) of IGlarg compared to IDet. The statistical method to generate weighted mean of the results was based on a random-effect model analysis in order to take into consideration the heterogeneity among the studies. A mean insulin daily dose for T2DM patients treated with detemir once and twice daily (‘mixed’ detemir patient group) was obtained from the mean insulin daily dose weighted by the number of patients in each administration frequency group from three studies20–22 (the remaining T2DM study23 did not report once-daily insulin detemir dose). Missing data were handled with the multiple imputations method that allows for any type and number of variables and give the most reliable variance estimates. Two separate meta-analyses were conducted for T1DM and T2DM based on two18,19 and four open-label clinical trials20–23, respectively in order to provide data for each cost analysis.

Table 1.  Main studies’ characteristics and reported insulin daily doses.

First author	Year	Trial characteristics	Trial duration	Insulin type	Frequency of injection	Mean insulin daily dose (IU/kg/day)*	Number of patients at end of trial (%)†
Type 1 diabetes mellitus
Heller‡	2009	Multinational, open-label, parallel-group, non-inferiority	52 weeks	Glargine	QD	0.33	122 (100%)
				Detemir	QD	0.33	90 (34%)
				Detemir	BID	0.47	173 (66%)
Pieber	2007	Multinational, open-label, parallel-group, non-inferiority	26 weeks	Glargine	QD	0.35	124 (100%)
				Detemir	BID	0.47	128 (100%)
Type 2 diabetes mellitus
Rosenstock§	2007	Multinational, open-label, parallel-group, non-inferiority	52 weeks	Glargine	QD	0.44	248 (100%)
				Detemir	QD	0.52	102 (45 %)
				Detemir	BID	1.00	125 (55%)
Swinnen¶	2010	Multinational, open-label, non-inferiority	24 weeks	Glargine	QD	0.52	456 (100%)
				Detemir	BID	0.90	437 (100%)
Hollander**	2008	Multinational, open-label, parallel-group, non-inferiority	52 weeks	Glargine	QD	0.59	84 (100%)
				Detemir	QD	0.69	74 (42.8%)
				Detemir	BID	0.95	99 (57.2%)
Raskin††	2009	Multi-center, open-label, parallel-group, non-inferiority	26 weeks	Glargine	QD	0.75	113 (100%)
				Detemir	QD	0.80	182 (87.4%)
				Detemir	BID	0.89	26 (12.6%)

*Insulin daily dose at study termination.

†Number of patients treated with each insulin type at study termination, with a breakdown, in some studies, between insulin detemir once and twice-daily (for patients switched to twice-daily dosing during the study).

‡Patients were randomized in a 2:1 ratio to receive either detemir or glargine once daily. If, during the trial, patients from the detemir group were achieving plasma glucose target before breakfast but not before dinner, they were switched to twice-daily administration.

§Patients were randomized 1:1 to receive insulin glargine or insulin detemir once daily. The majority of patients administering insulin detemir twice daily at study completion were transferred to this regimen within 12 weeks of treatment.

¶Insulin daily dose per kg was estimated for each treatment group, based on the insulin daily dose (43.5 and 76.5 units) and the baseline weight (83.65 kg and 85.50 kg) reported for glargine and detemir twice daily, respectively in the study.

**Patients randomized to receive detemir could be switched to a second dose given in the morning based on failure to achieve a pre-specified pre-dinner plasma glucose target. The absolute number of patients on detemir once or twice daily was estimated based on the proportion of 57.2% of detemir patients switched to twice-daily dosing that was reported in the study.

††Patients were randomized 2:1 (insulin detemir : insulin glargine). The absolute number of patients on detemir once or twice daily was estimated based on the proportion of 12.6% of detemir patients switched to a twice-daily administration that was reported in the study.

Of the two T1DM studies, Pieber et al.19 conducted a 26-week trial comparing IGlarg QD to IDet BID, while Heller et al.18 conducted a 52-week trial comparing IGlarg QD to IDet QD and BID (Table 1). In the latter, the majority of patients (65.8%) required two injections of IDet per day by the end of the study. In both trials, patients were administered IGlarg and IDet in combination with pre-meal insulin aspart. Similar reductions in HbA_1c level and comparable rates of all hypoglycemic events were observed with each insulin treatment by the end of both studies.

Of the four T2DM studies, Rosenstock et al.22 and Hollander et al.20 conducted 52-week trials, whereas Swinnen et al.23 and Raskin et al.21 conducted a 24- and a 26-week trial, respectively (Table 1). In all trials, both IGlarg and IDet were associated with similar reductions in HbA_1c level. The rates of all hypoglycemic events were also found to be comparable between groups. Interestingly, Rosenstock et al.22 as well as Hollander et al.20 reported that over 50% of patients treated with IDet required two injections per day.

All six studies consistently reported that a higher dose of IDet was needed in patients to achieve the same level of efficacy as those treated with IGlarg. As the meta-analysis showed that IGlarg and IDet had a similar efficacy and safety profile, a cost-minimization analysis was deemed appropriate to achieve the study’s ultimate objective.

CORE diabetes model

A cost-minimization analysis was conducted using a validated computer simulation model, the CORE Diabetes Model (CDM). Details of this model are described elsewhere24,25. The CDM was developed to estimate the long-term health outcomes and economic consequences of interventions in T1DM and T2DM. It takes into account baseline cohort characteristics, history of complications, current and future management of diabetes and concomitant medications, treatment effects and changes in physiological parameters over time. A series of sub-models within the CDM also simulate the major complications of diabetes, including myocardial infarction, congestive heart failure, stroke, peripheral vascular disease, neuropathy, foot ulcer, amputation, retinopathy, macular edema, cataract and nephropathy. Each of these sub-models uses Markov techniques combined with Monte Carlo simulation. Transition probabilities for the model were obtained from epidemiological and clinical studies in the published literature, including the Diabetes Control and Complications Trial (DCCT) and the United Kingdom Prospective Diabetes Study (UKPDS)5,6. The CDM is consistent with published American Diabetes Association computer-based modeling guidelines for assessing the long-term costs and clinical outcomes for various diabetes treatments26. In addition, the CDM has been used for analyses published in various journals as well as in governmental publications9,27–30.

Simulation cohorts

Cohorts of T1DM and T2DM patients were primarily defined with baseline demographics, clinical parameters and racial characteristics representative of the head-to-head trials populations explored in the meta-analysis. At the start of treatment, in the T1DM and T2DM cohorts, the mean age was 41.7 and 58.1 years, the proportion of male was 54% and 56%, and duration of diabetes was 17.2 and 10.6 years, respectively. The meta-analysis provided data for HbA_1c level (8.1% and 8.6%, for T1DM and T2DM, respectively). Data for systolic blood pressure (SBP), total cholesterol (TC), high density lipoprotein cholesterol (HDL-C), low density lipoprotein cholesterol (LDL-C), and triglyceride were derived from the 2008 Canadian Optimal Medication Prescribing and Utilization Service report (COMPUS)15. Racial characteristics were taken from the COMPUS report for T1DM and from the meta-analysis for the T2DM cohort15. The COMPUS report was also the main source used for both cohorts’ complications rates. Baseline cohort characteristics and complications are presented in Table 2.

Table 2.  Baseline patient characteristics for type 1 and type 2 diabetes mellitus.

Baseline characteristic	T1DM cohort	T2DM cohort
Patient demographics
Mean age in years	41.7 (12.7)^a	58.1 (9.5)^a
Duration of diabetes in years	17.2 (11.4)^a	10.6 (6.5)^a
Proportion male (%)	54.0^a	56.0^a
Ethnic group
Caucasian (%)	88.015	81.3^a
Black (%)	3.015	5.3^a
Hispanic (%)	1.015	3.8^b
Native American (%)	1.0^b	0.2^a
Asian/Pacific Islander (%)	7.015	9.4^a
Risk factors
HbA1c (%)	8.1 (1.1)^a	8.6 (0.9)^a
Systolic blood pressure (mmHg)	11515	13715
Body mass index (kg/m^2)	26.4 (4.0)^a	30.9 (4.7)^a
Total cholesterol (mg/dL)	188.715	221.9615
High density lipoprotein  cholesterol (mg/dL)	48.715	43.715
Low density lipoprotein  cholesterol (mg/dL)	110.915	121.015
Triglycerides (mg/dL)	86.815	235.615
Cardiovascular disease
Myocardial infarction (%)	0.0^c	11.015
Angina (%)	2.015	11.215
Peripheral vascular disease	0.0^c	8.715
Stroke (%)	0.0^c	8.115
Congestive heart failure (%)	0.0^c	9.215
Atrial fibrillation (%)	3.0^c	3.0^d
Left ventricular hypertrophy (%)	3.0^c	3.0^d
Baseline renal complications
Microalbuminuria (%)	10.015	28.215
Gross proteinuria (%)	0.0^c	7.615
End-stage renal disease (%)	0.0^c	0.415
Baseline retinopathy complications
Background of diabetic  retinopathy (%)	100.015	39.015
Proliferative diabetic retinopathy (%)	0.0^c	3.015
Severe vision loss (%)	0.0^c	1.315
Baseline macular edema
Macular edema (%)	0.0^c	4.015
Baseline cataract
Cataract (%)	0.0^c	14.015
Baseline foot ulcer complications
Uninfected ulcer (%)	0.0^c	10.515
Infected ulcer (%)	0.0^d	0.0^c
Healed ulcer (%)	0.0^d	0.0^c
History of amputation (%)	0.0^c	1.015
Baseline neuropathy
Neuropathy (%)	0.0^d	30.015

*All values are expressed in means (standard deviation).

†Additional sources: ^aMeta-analysis; ^bEstimate; ^cPrevious analysis; ^dCommon Core Model assumption.

Perspective and costs

The analysis was conducted from the perspective of a Canadian provincial government, therefore capturing direct medical costs only. Direct medical costs included costs associated with diabetes treatment (for each insulin treatment), diabetes complications (cardiovascular, renal, ulcer/amputation/neuropathy and eye diseases) management and treatment, as well as insulin adverse events treatment (severe hypoglycemia). All unit costs except treatment costs were deemed to be similar for T1DM and T2DM patients.

Both T1DM and T2DM patients were considered to stay on the same insulin treatment for life. Insulin treatment costs were estimated per year and were assumed to be constant over time. Treatment costs encompassed drug costs including a 6% wholesaler mark-up, $8.44 pharmacist dispensing fee31, disposable syringes and needles for IGlarg vials (a 65:35 cartridge-to-vial ratio was used for IGlarg15). IDet is only available in cartridge15. Drug costs were estimated based on the daily insulin dose per patient generated by the meta-analysis (Table 3) and on the IGlarg and IDet prices from the Quebec Formulary List (June 2010)31. Costs associated with self-monitoring of blood glucose (SMBG) were not included as they were assumed to be similar in both treatment groups.

Table 3.  Pooled meta-analysis mean daily insulin doses per patient per type of diabetes.

Insulin treatment	Mean dose of daily insulin per patient with T1DM (I U/day)	Mean dose of daily insulin per patient with T2DM (I U/day)
Glargine QD	26.2	47.2
Detemir QD	–	65.7
Detemir BID	33.6	80.4
Detemir mixed* QD and BID	–	74.5

*‘Mixed’ refers to a combined population using either detemir once daily or detemir twice daily.

QD, once daily; BID, twice daily.

Costs resulting from diabetes complications management (i.e., use of concomitant medications) and treatment, as well as from insulin adverse events treatment, were derived from the published literature, mainly from sources used in the 2008 COMPUS report32–44; diabetes complication and insulin adverse events treatment costs were adjusted to 2010 Canadian dollars using the health component of the Consumer Price Index (Table 4)45.

Table 4.  Direct unit of diabetes complications or events, adjusted to $CAN 2010 (type 1 and type 2 diabetes mellitus).

Description of event or state	Adjusted complication costs ($CAN 2010)
Management costs
Annual cost of statins (assume atorvastatin)	34854
Annual cost of aspirin	12255
Annual cost of ACE inhibitors  (assume Apo-Ramipril)	23754
Annual cost of microalbuminuria screening	2938
Annual cost for gross proteinuria screening	2938
Annual cost for retinopathy	8738
Non-standard ulcer treatment	17236
Direct costs cardiovascular complications
Myocardial infarction year 1	19,23239
Myocardial infarction, each subsequent year	300739
Angina year 1	602039
Angina, each subsequent year	347539
Congestive heart failure year 1	17,59339
Congestive heart failure, each subsequent year	493339
Stroke year 1	26,19639
Stroke, each subsequent year	363539
Stroke death within 30 days	949239
Peripheral vascular disease year 1	165734
Peripheral vascular disease, each subsequent year	0
Direct costs for renal complications
Hemodialysis costs year 1	83,28137
Hemodialysis costs, each subsequent year	83,281*
Peritoneal dialysis, year 1	83,281*
Peritoneal dialysis, each subsequent year	83,281*
Renal transplant costs in year 1	95,42037
Renal transplant costs, each subsequent year	40,12437
Direct costs for eye disease
Retinal photocoagulation	96638
Severe vision loss/blindness, first year	321839
Severe vision loss/blindness, each  subsequent year	229339
Cataract extraction	425539
Cataract extraction, each subsequent year	267839
Direct costs for other complications
Neuropathy, onset	18138
Uninfected ulcer (monthly based)	133641
Infected ulcer (monthly based)	267241
Gangrene (monthly based)	954841
Amputation, year of event	40,63739
Amputation, subsequent years	556639
Direct costs acute events
Major hypoglycemic event	14238
Ketoacidosis event	428141
Lactic acidosis event	428141
Edema onset	342540

*Assumed same as Hemodialysis costs year 1.

ACE, angiotensin-converting enzyme.

Time horizon and discounting

A time horizon of 60 years and of 50 years (lifetime perspective) was used for the respective T1DM and T2DM base-case analyses in order to capture long-term complications and their associated costs for each type of diabetes. Future costs were discounted at a rate of 5% per annum in the base-case analyses, in line with the Canadian Agency for Drugs and Technologies in Health46.

Sensitivity analysis

Due to the nature of cost-minimization analysis, a limited range of one-way sensitivity analyses (SAs) was performed. The discount rate was varied using 0% and 3% annual rates, compared to 5% in the base-case. The impact of time horizon was explored by adjusting the time to a short-term (5 years) and middle-term (10 years) perspective. For T2DM, additional SAs were conducted on the frequency of IDet injection (QD and BID separately) to examine the robustness of base-case results.

Statistical methodology

Uncertainty in the estimated cost minimization in this analysis was evaluated using non-parametric bootstrapping methods47. Each probability used in the model was sampled during first-order Monte Carlo simulation; this provided a point estimate for each parameter to be used in the second-order Monte Carlo simulation where sampling around input distributions was performed where estimates of their uncertainty were available, with mean costs and outcomes generated for 1000 cohorts of 1000 theoretical patients24,25.

Results

Main findings

The meta-analysis results showed that a similar efficacy can be achieved with both drugs using a significantly lower daily dose of IGlarg than IDet for both T1DM and T2DM patients. Long-term modeling using daily insulin doses estimated by the meta-analysis subsequently indicated that IGlarg QD may be cost saving by $4231 per T1DM patient compared to using IDet BID and from $4659 to $8709 per T2DM patient compared to using IDet QD and BID, respectively, over a lifetime from a Canadian provincial government perspective.

Type 1 diabetes

The results from the meta-analysis including the findings from two recent head-to-head studies18,19 showed no difference for the change in HbA_1c level (WMD 0.002%-points; p = 0.97).

The cost-minimization analysis projected a long-term treatment cost of $52,389 (±1703) for T1DM patients treated with IGlarg QD compared to $56,620 (±1715) for T1DM patients taking IDet BID (Table 5). Using IGlarg QD would save $4231 per patient compared to using IDet over a lifetime. This difference is mainly driven by the lower dose of insulin required for IGlarg compared to IDet. The costs associated with diabetes events (both complications and adverse effects) were equal, based on the absence of differences detected in the meta-analysis of T1DM studies. Consequently, the lower lifetime costs associated with IGlarg QD were attributable to both a 22% smaller daily insulin dose needed to achieve similar HbA_1c level control (26.2 units IGlarg vs. 33.6 units IDet) and a lower IGlarg cost ($1.95 per day) compared to IDet BID ($2.82 per day) (including the 6% wholesaler mark-up, pharmacist dispensing fees, syringes and needles).

Table 5.  Base-case direct medical costs associated with insulin glargine and insulin detemir treatment strategies in type 1 diabetes mellitus ($CAN 2010).*

	Glargine QD	Detemir BID	Incremental difference
Total costs	52,389	56,620	−4231
Treatment	10,927	15,158	−4231
Management	5077	5077	0
Cardiovascular diseases	11,329	11,329	0
Renal complications	13,176	13,176	0
Ulcer/amputation/neuropathy	4360	4360	0
Eye complications	7521	7521	0

*Differences between total costs and treatment costs result from rounding.

QD, once daily; BID, twice daily.

Type 2 diabetes

The results from the meta-analysis based on findings from four recent head-to-head studies20–23 showed no difference in HbA_1c change from baseline between treatment groups (WDM –0.05%-points; p = 0.28).

The economic analysis estimated a long-term treatment cost of $57,642 (±1828) for T2DM patients treated with IGlarg QD compared to $64,399 (±1885) for T2DM patients on IDet, irrespectively of IDet administration schedule (weighted average from QD and BID users) (Table 6). Using IGlarg QD instead of IDet QD or BID would save $6757 per patient over a lifetime. Similarly to the findings for T1DM, the lower costs associated with IGlarg were attributable to a smaller daily insulin dose of IGlarg (36.6%) needed to achieve equivalent HbA_1c level control (47.2 units IGlarg vs. 74.5 units IDet) and a lower IGlarg cost ($3.23 per day) compared to IDet ($5.35 per day) (also including 6% wholesaler mark-up, pharmacist fees, syringes and needles).

Table 6.  Base-case direct medical costs associated with insulin glargine and insulin detemir treatment strategies in type 2 diabetes mellitus ($CAN 2010).*

	Glargine QD	Detemir ‘mixed’*	Incremental difference
Total costs	57,642	64,399	−6757
Treatment	11,198	17,954	−6756
Management	4122	4122	0
Cardiovascular diseases	26,354	26,354	0
Renal complications	6040	6040	0
Ulcer/amputation/neuropathy	4149	4149	0
Eye complications	5780	5780	0

*Differences between total costs and treatment costs result from rounding.

‘Mixed’ refers to a combined population using either detemir once daily or detemir twice daily.

QD, once daily; BID, twice daily.

Sensitivity analyses

One-way sensitivity analyses were performed in order to identify the magnitude and direction of the impact of key model parameters and assumptions used in the base-case analyses on cost estimates and in order to test the robustness of the model results.

Type 1 diabetes

In the base-case, the total lifetime costs associated with treatment with IGlarg and IDet were estimated at $52,389 ± 1703 and $56,620 ± 1715 per patient, respectively, resulting in a saving of $4231 per T1DM patient receiving IGlarg QD compared to IDet BID (Table 7). When the discount rate for costs was set to 0% per annum (compared to 5% in the base-case), the total lifetime costs savings increased to $7990 per patient with IGlarg ($137,088 ± 4865 with IGlarg vs. $145,078 ± 4921 with IDet). When the discount rate for costs was set to 3% per annum, the total lifetime cost savings were $5307 per patient with IGlarg ($74,279 ± 2452 with IGlarg vs. $79,586 ± 2474 with IDet).

Table 7.  Summary of sensitivity analysis results for patients with type 1 and type 2 diabetes mellitus in Canada ($CAN 2010).*

Analysis	Total lifetime costs		Incremental cost
	Glargine QD	Detemir BID
Base case	52,389 ± 1703	56,620 ± 1715	−4231
0% discount rate	137,088 ± 4865	145,078 ± 4921	−7990
3% discount rate	74,279 ± 2452	79,586 ± 2474	−5307
Time horizon 5 years	7,130 ± 241	8,486 ± 241	−1355
Time horizon 10 years	15,356 ± 624	17,733 ± 624	−2377

*Differences between total costs and incremental costs result from rounding.

QD, once daily; BID, twice daily.

When time horizon was set to 5 years, total costs associated with treatment with IGlarg and IDet were estimated at $7130 ± 241 and $8486 ± 241, respectively, resulting in a saving of $1355 per patient with IGlarg compared to IDet. When the time horizon was set to 10 years, total costs associated with treatment with IGlarg and IDet were estimated at $15,356 ± 624 and $17,733 ± 624, respectively, resulting in a saving of $2377 per patient receiving IGlarg.

Type 2 diabetes

In the base-case, the total lifetime costs associated with treatment with IGlarg and IDet were estimated at $57,642 ± 1828 and $64,399 ± 1885, respectively, resulting in a saving of $6757 per patient with IGlarg QD compared to IDet QD and BID combined (Table 7). When the discount rate for costs was set to 0% per annum (compared to 5% in the base-case), the total lifetime costs savings increased to $10,232 per patient with IGlarg ($95,724 ± 3404 with IGlarg vs. $105,955 ± 3546 with IDet). When the discount rate for costs was set to 3% per annum, the total lifetime cost savings were $7866 per patient with IGlarg ($69,362 ± 2263 with IGlarg vs. $77,228 ± 2343 with IDet).

When the time horizon was set to 5 years, total costs per patient associated with treatment with IGlarg and IDet were estimated at $20,989 ± 851 and $24,080 ± 854, respectively, resulting in a saving of $3091 per patient with IGlarg. When the time horizon was set to 10 years, total costs per patient associated with treatment with IGlarg and IDet were estimated at $36,742 ± 1298 and $41,740 ± 1310, respectively, resulting in a saving of $4998 per patient with IGlarg.

When IGlarg was compared to IDet BID only, the projected cost savings were $8709 per patient with IGlarg. When IGlarg was compared to IDet QD only, the projected cost savings were $4659 per patient with IGlarg.

Discussion

Several studies have analyzed the efficacy and safety of IDet compared to IGlarg in both T1DM13 and T2DM patients14–17. Additionally, a recent meta-analysis of 23 RCTs of IGlarg and IDet in T2DM insulin naïve patients48 estimated that change from baseline in HbA_1c level (−1.4%; statistically not significant) was similar with IGlarg and IDet. It also showed that patients receiving IDet required a 40.7% higher daily insulin dose than those taking IGlarg to achieve the same level of glycemic control (51.5 units versus 36.6 units for IDet and IGlarg, respectively)48. All but one22 of these RCTs involved indirect comparisons.

In this study, in contrast, the clinical input data were derived from a meta-analysis of published head-to-head trials18–23. These data were used in a cost-minimization analysis. The results demonstrated that, when considering that a similar level of HbA_1c control and comparable safety outcomes were achieved with a lower daily IGlarg dose compared to IDet, a Canadian provincial government could expect a long-term cost-saving of $4231 per T1DM patient and $6757 per T2DM patient when using IGlarg instead of IDet. The potential cost-saving of IGlarg compared to IDet is in line with findings from other studies49,50.

Beyond these numerical findings, the achievement of this study was three-fold: first, it ascertained the similarity of the efficacy profile between IGlarg and IDet based on clinical data from six recently published direct comparison trials, thus providing robust evidence justifying the use of a cost-minimization analysis; second, it estimated $CAN 2010 adjusted insulin and diabetes complications treatment costs as representative as possible of the costs that would be incurred by a Canadian provincial health authority; and third, it tested the variability of the cost of treatment with IDet when using the different daily injection frequency (QD or BID), reflecting the product’s licensed indication.

The results of the meta-analysis showed no difference in efficacy between both basal insulin analogues, IGlarg and IDet. They additionally demonstrated that equivalent glycemic control can be achieved with a reduction of 22% in the daily insulin dose for T1DM patients treated with IGlarg QD compared to IDet BID and a reduction of 36.6% in the daily insulin dose for T2DM patients taking IGlarg compared to IDet QD and BID combined. The lower requirement in the daily insulin dose with IGlarg could be explained in part by its longer half-life compared to IDet51.

Because of the nature of this study (a cost-minimization analysis), findings related to the relative safety of IGlar compared to IDet that were not reported in the results section. The meta-analyses indeed estimated a small increase in weight associated with IGlarg compared with IDet, of 0.282 kg in T1DM and 0.92 kg in T2DM. These results are in line with findings from other studies52,53. However, as reported by other authors, such small weight gains are unlikely to be considered clinically meaningful9,52.

The meta-analyses additionally estimated the relative risk of experiencing a hypoglycemic event when trial data were available. In the two T1DM studies, no minor hypoglycemic events were reported. The meta-analysis did not show any difference in the risk of having a major hypoglycemic event when using IGlarg instead of IDet in T1DM patients (RR 1.29; p = 0.486). It estimated that T2DM patients had a higher rate of minor hypoglycemic events (RR 1.14; p = 0.039) and no difference in the risk of having major hypoglycemic events (RR 1.21; p = 0.42) when treated with IGlarg compared to IDet.

The cost-minimization analysis has additionally demonstrated that lifetime treatment with IGlarg compared to IDet would generate savings of $4231 per patient in T1DM and savings from $4659 to $8709 per T2DM patient compared to using IDet QD and BID, respectively from the perspective of a Canadian provincial government. The total lifetime direct costs estimated for each diabetes treatment are believed to reflect as accurately as possible the respective insulin cost paid by a Canadian provincial government. Not only did the treatment cost estimates include the cost of disposable syringes and needles, wholesaler mark-ups and pharmacist fees, but daily insulin dosage also reflected the frequency of injection.

In the T2DM base-case analysis, a conservative comparator treatment was used, IDet considering both licensed frequencies of administration, once-daily and twice-daily injections. It is reported that twice-daily injection schedule is more common with IDet than a once-daily schedule20,22. The sensitivity analysis showed an additional saving of $1952 per patient ($8709–6757) to a Canadian provincial government when the comparison of IGlarg was made with IDet BID compared to the base case when the combination of QD and BID dosing schedule for IDet was used.

Limitations

This study encompasses a certain number of limitations. The representativeness of the Canadian diabetes population may be discussed in this analysis: most of the six trials are multinational studies and it did not appear that any of the sites involved were located in Canada. Consequently, the patients’ characteristics may not be accurately representative of the Canadian population with diabetes.

Furthermore, due to a lack of data in the two head-to-head trials involving T1DM patients, it was not possible to estimate total direct costs for IDet patients injecting once a day as it was done in the T2DM analysis. Therefore, in T1DM patients, this analysis did not project cost for IDet considering both QD and BID frequency of injection.

Lastly, it was assumed in the analysis that self-monitoring blood glucose (SMBG) was similar in both treatment strategies and therefore this parameter has not been captured in the projected lifetime direct costs. It could however be suggested that the frequency of SMBG would be higher in the IDet than in the IGlarg patients due to a more frequent insulin intake per day in the twice-daily injection schedule. However, this was not studied or compared.

Conclusion

This cost-minimization analysis was supported by a robust meta-analysis of six recently published head-to-head trials comparing IGlarg and IDet in T1DM and T2DM patients. It provided additional evidence of the similarity in the efficacy and safety profile of IGlarg and IDet, and on the long-term costs that a Canadian provincial government could expect to save when treating diabetes patients with IGlarg instead of IDet. The results of this analysis demonstrated that cost-saving is mainly driven by the significantly lower dose of insulin required for patients on IGlarg compared to IDet, in both T1DM and T2DM patients.

Transparency

Declaration of funding:

The funding for the study and manuscript preparation was provided by Sanofi-Aventis, Canada. The sponsor had no decisive role in the conduct of the cost-minimization analysis and in the meta-analysis. The sponsor did provide editorial assistance for the manuscript. Although results, interpretations and conclusions were not dictated by the sponsor, the paper did receive sponsor approval prior to submission.

Declaration of financial/other relationships

JV is an employee of Sanofi-Aventis, who sponsored the preparation of the manuscript. A-L.G., Y.S., D.W. and T.N. are employees of IMS Health and have no relationship (financial, employment, other significant/relevant relationships) with the sponsor.

Acknowledgments

The authors wish to thank Luc Sauriol from Sanofi-Aventis for his advice and Corinne Le Reun for performing the meta-analysis.

Notes

*Lantus is a registered trademark of Sanofi-Aventis Inc., USA.

†Levemir is a registered trademark of Novo Nordisk A/S, Bagsværd, Denmark.