Cost-effectiveness of dapagliflozin (Forxiga®) added to metformin compared with sulfonylurea added to metformin in type 2 diabetes in the Nordic countries.

ARTICLE IN PRESS

PCD-392; No. of Pages 9

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 4 ) xxx–xxx

Contents lists available at ScienceDirect

Primary Care Diabetes journal homepage: http://www.elsevier.com/locate/pcd

Original research

Cost-effectiveness of dapagliflozin (Forxiga® ) added to metformin compared with sulfonylurea added to metformin in type 2 diabetes in the Nordic countries Ugne Sabale a,∗ , Mattias Ekman a , Ola Granström a,1 , Klas Bergenheim b , Phil McEwan c,d a

AstraZeneca, Södertälje, Sweden AstraZeneca, Mölndal, Sweden c Health Economics and Outcomes Research Ltd., Monmouth, UK d Centre for Health Economics, Swansea University, UK b

a r t i c l e

i n f o

a b s t r a c t

Article history:

Aims: The aim of this study was to assess the long-term cost-effectiveness of dapagliflozin

Received 3 March 2014

(Forxiga® ) added to metformin, compared with sulfonylurea (SU) added to metformin, in

Received in revised form

Nordic Type 2 diabetes mellitus (T2DM) patients inadequately controlled on metformin.

15 April 2014

Methods: Data from a 52-week clinical trial comparing dapagliflozin and SU in combination

Accepted 17 April 2014

with metformin was used in a Cardiff simulation model to estimate long term diabetes-

Available online xxx

related complications in a cohort of T2DM patients. Costs and QALYs were calculated from

Keywords:

Results:

Dapagliflozin

dapagliflozin + metformin was D 7944 in Denmark, D 5424 in Finland, D 4769 in Norway, and

Cost-effectiveness

D 6093 in Sweden. Metformin + dapagliflozin was associated with QALY gains ranging from

a healthcare provider perspective and estimated over a patient’s lifetime. Compared

with

metformin + SU,

the

cost

per

QALY

gained

with

Type 2 diabetes mellitus

0.236 in Norway to 0.278 in Sweden and incremental cost ranging from D 1125 in Norway to

Sulfonylurea

D 1962 in Denmark. Results were robust across both one-way and probabilistic sensitivity

Economic modeling

analyses. Results were driven by weight changes associated with each treatment. Conclusions: Results indicate that metformin + dapagliflozin is associated with gains in QALY compared with metformin + SU in Nordic T2DM patients inadequately controlled on metformin. Dapagliflozin treatment is a cost-effective treatment alternative for Type 2 diabetes in all four Nordic countries. © 2014 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.

∗ Corresponding author at: AstraZeneca Nordic-Baltic, Department of Health Economics, Astraallén, B674, SE-151 85 Södertälje, Sweden. Mobile: +46 705 66 17 91. E-mail address: [email protected] (U. Sabale). 1 Now at: Gilead Sciences, United States. http://dx.doi.org/10.1016/j.pcd.2014.04.007 1751-9918/© 2014 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: U. Sabale, et al., Cost-effectiveness of dapagliflozin (Forxiga® ) added to metformin compared with sulfonylurea added to metformin in type 2 diabetes in the Nordic countries, Prim. Care Diab. (2014), http://dx.doi.org/10.1016/j.pcd.2014.04.007


ARTICLE IN PRESS

2


1.

Introduction

Rapidly increasing prevalence of Type 2 diabetes mellitus has developed into a major health concern worldwide [1–3]. The treatment goal in T2DM is to achieve blood glucose (HbA1c) control and to avoid diabetes-related complications. Although ensuring adequate long-term blood glucose control is emphasized as one of the key goals of T2DM treatment [4,5], minimizing the risk for hypoglycemia and avoiding weight gain are also important considerations [6]. Yet, hypoglycemia is a commonly observed side effect of diabetes treatments such as sulfonylurea and insulin [6]. Experiencing hypoglycemic events negatively affects patients’ Health Related Quality of Life (HRQoL) [7–9]. Another common side effect of some diabetes treatments (e.g. sulfonylurea and insulin) is weight gain [6,10]. Excess weight is directly linked to worsened insulin resistance, high cardiovascular morbidity and mortality risks, and continually increasing economic burden [3,10–14]. However, achieving and maintaining the treatment goals remains challenging as around 80% of T2DM patients suffer from overweight [9]. Optimizing diabetes management therefore requires a multifactorial approach to treatment which goes beyond glycemic control and encompasses other risk factors such as reduction of blood pressure, blood lipid levels, and weight. Dapagliflozin (Forxiga® ) is the first compound of the new class of sodium-glucose co-transporter 2 (SGLT-2) inhibitors for the oral treatment of T2DM. Dapagliflozin reduces HbA1c and body weight significantly and has a positive impact on blood pressure [15]. Compared with sulfonylureas, the most commonly used oral antidiabetics beside metformin, dapagliflozin showed similar HbA1c reduction, but was associated with significant weight loss and significantly fewer hypoglycemic events. Dapagliflozin is thus an alternative to sulfonylureas for T2DM patients inadequately controlled on metformin plus diet/exercise [15]. This study examines the long-term cost-effectiveness of metformin + dapagliflozin compared with metformin + SU in T2DM patients in Denmark, Finland, Norway, and Sweden.

2.

Materials and methods

2.1.

Model

Sweden, Finland and Denmark and 4% in Norway. The discount rate was varied in the one-way sensitivity analysis. Baseline demographics and modifiable risk factors define a time-dependent risk-factor profile, which determines the patient’s risk of developing diabetes-related complications. The baseline demographics are set at the start of the analysis, and are updated with time in the model, while the modifiable risk factors (HbA1c, weight, etc.) are altered by drug treatment over time. Diabetes-related events have a direct impact on costs and utilities. Utility weights are applied to the event both in the year of occurrence and where relevant also in subsequent years. The impact of uncertainty around model inputs is assessed both in a one-way sensitivity analysis and a probabilistic sensitivity analysis (PSA) performed as a second order MonteCarlo simulation. HbA1c lowering effects associated with treatment and control, weight changes, and symptomatic hypoglycemic events are sampled from a normal distribution, the probability of a severe hypoglycemic event and utility decrements from a beta distribution, and costs from a gamma distribution.

2.2. Patient population, treatment strategies, HbA1c threshold, and outcomes The analyzed patient population is characterized by a demographic and risk factor profile taken from the study of Nauck et al. [15] (Table 1), a randomized controlled trial where 801 individuals received either metformin + dapagliflozin or metformin + SU. Diabetes is a progressive disease, meaning that HbA1clevels increase over time also in patients who receive treatment. In our analysis, the HbA1c progression is modeled based on the glucose profile observed in the UKPDS trial [18]. The slopes of the HbA1c curves, which determine the pace at which the blood glucose increases over time, are assumed to be identical for both arms. When HbA1c reaches a threshold level of 7.5%, treatment intensification is needed and a patient is switched to rescue therapy – NPH insulin [5]. An HbA1c threshold of 8% is tested in one-way sensitivity analyses.

2.3.

This analysis uses a simulation model based on a previously developed model designed for evaluating treatment regimens in T2DM [16,17]. Our model uses the UKPDS 68 equations to forecast the occurrence of seven diabetes-related events and death [18]. We simulate a T2DM patient cohort consisting of 10 000 individuals over 40 years (a lifetime horizon as the baseline age is 58 years). The model reports the number of cumulative diabetes-related complications (macro- and microvascular events), hypoglycemia events avoided, diabetes mortality, non-diabetes mortality, and cost-effectiveness results. The costs and QALYs associated with each treatment are calculated from a healthcare provider perspective. Based on country-specific pharmacoeconomic guidelines, future costs and benefits were discounted at the rate of 3% annually in

Weight gain

The utility impact of weight change is modeled linearly, so that each unit change in BMI is associated with an identical utility weight. This linear approach gives a conservative estimate of the utility decrements associated with weight gain, as the disutility of weight gain increases with body weight and has been found to be larger than the positive impact of weight loss [7]. Body weight changes also have an impact on the long-term risk of cardiovascular complications [11].

2.4.

Hypoglycemia

We differentiate between symptomatic and severe hypoglycemic events, where a severe event is defined as a major symptomatic episode requiring 3rd party assistance due to severe impairment in consciousness [15]. Each drug therapy is



ARTICLE IN PRESS 3


Table 1 – Clinical input variables, demographics and risk factors 52 week follow-up. Variable Clinical inputa HbA1c reduction, year 1 (%) Weight change (kg) Number symptomatic hypoglycemic events Probability of severe hypoglycemic events Systolic blood pressure reduction year 1(mm HG) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) Number of UTIs and GIs Demographicsb Baseline age (years) Proportion female (%) Duration diabetes (years) Height (m) Proportion smokers (%) Risk factorsb HbA1C (%) Total cholesterol (mg/DL) HDL cholesterol (mg/DL) SBP (mmHg) Weight (kg)

Forxiga + met −0.52 −3.22 0.035 0.000 −4.3 2.79 2.65 0.23

SU + met −0.52 1.44 0.404 0.0075 0.8 −1.06 −0.07 0.09

Reference

Rescue therapy

Reference

[15] [15] [15] [15] [15] [15] [15] [15]

−1.11 +1.9 0.63 0.0213

[20] [31] [20] [20]

58.4 44.9 6.32 1.67 17.6

[15] [15] [15] [15] [15]

7.72 182.5 45.87 133.3 88.02

[15] [15] [15] [15] [15]

HbA1c – glycosylated hemoglobin; HDL – high-density lipoprotein; SBP – systolic blood pressure; UTI – urinary tract infection; GI – genital infection. a All efficacy variables are taken from the intention to treat analysis. Where these were not available in Nauck et al. [15] they have been obtained from the clinical trial report. b Based on all patients in the Nauck et al. study [15].

associated with a fixed number of symptomatic hypoglycemic events per patient in every cycle, as well as a probability of experiencing a severe hypoglycemic event. Due to the transient nature and relative infrequency of symptomatic and severe hypoglycemic events in patients treated with oral anti-diabetic therapies, fear of hypoglycemia is an important driver of disutility and is strongly associated with severe events [7,19]. The utility decrements associated with fear of hypoglycemia are applied to the year of occurrence and are based on the Hypoglycemia Fear Survey Data [8].

2.5.

Other adverse events

Besides weight increase and hypoglycemia, we model genital infections (GIs) and urinary tract infections (UTIs), which are the most common adverse events with dapagliflozin treatment. A GI or UTI event is associated with a utility decrement and a cost in the year that they occur.

3.

Input data

3.1.

Clinical trial data

The primary efficacy parameter applied for treatment and control is the initial reduction in HbA1c (Table 1), which is identical in both arms after 52 weeks [15]. Metformin + dapagliflozin is associated with weight loss whereas SU + metformin is associated with weight gain. In the base case analysis, the weight loss with dapagliflozin is assumed to be maintained for two years based on the observed

duration of weight loss in the trial [15]. After that, it was assumed that the patient’s weight gradually increases such that the treatment-related weight loss is fully regained upon the initiation of insulin (Fig. 1). When patients reach the HbA1c threshold of 7.5%, they are assumed to be switched to NPH insulin, which leads to additional weight gain of 1.9 kg [20]. The weight increases in accordance with the natural weight progression, i.e. an increase of 0.1 kg per year [4]. Clinical input data for NPH insulin are presented in Table 1.

3.2.

Costs

The analysis includes country-specific direct costs for medications, fatal and non-fatal complications, maintenance treatment, and treatment-related adverse-events (Table 2). For NPH insulin, a cost per kilo body weight per day was used as the dose is weight-dependent in clinical practice. Medication retail price per unit was applied in Sweden and Denmark, whereas retail price excluding value-added tax (VAT) was applied in Norway and Finland (25% and 9% respectively) in line with country-specific guidelines. In Sweden, the costs for complications are mainly taken from the Swedish study of Gerdtham et al. [21], where costs were controlled for confounding factors (gender, age, and comorbidities). For blindness, the non-fatal and maintenance costs were taken from the study of Henriksson et al. [22]. In Norway and Denmark, the costs of diabetes-related events were based on the DRG (diagnosis related group) systems for 2010.2 In Finland the cost data were collected from the

2 Diagnosis related group (DRG) system in Norway: http://helsedirektoratet.no/finansiering/drg/Sider/default.aspx.



ARTICLE IN PRESS

4


Fig. 1 – Simulated progression of weight in the treatment (metformin + dapagliflozin) and control (metformin + SU) arms over 40 years.

Table 2 – Yearly direct costs for diabetes-related events, side-effects and medications (D 2013).

Drug cost Forxiga + metformin SU (glipizide) + metformin NPH insulin (price/kg/day) Event cost (fatal) Myocardial infarction Congestive heart failureg Stroke Amputationk Event cost (non-fatal) Ischemic heart disease Myocardial infarction Congestive heart failure Stroke Amputation Blindness ESRD Event cost (maintenance) Ischemic heart disease Myocardial infarction Congestive heart failure Stroke Amputation Blindness ESRD Cost of adverse events UTI/GIh Severe hypoglycemia a b c d e

f g h i j k

Denmark

Finland

Norway

Sweden

816.13a 99.30a 0.0195a

715.29b 180.17b 0.0146b

613.28c 171.61c 0.0121c

741.03d 118.41d 0.0139d

4677e 7323e 5051e 12,261e

2661 4068 5741 9834

1726e 6020e 6368e 14,675e

5491 8158 9099 17,804

5345 8862 7323 5192 12,261 4237 25,511

3695 5041 4068 5902 9834 14,281 53,814

2336 5543 6020 6545 14,675 4327 66,597

12,339 10,405 8158 9353 17,804 4419 13,591

236e 208e 693e 219e 199e 135f,i 1663e

164 119 385 249 160 456f,j 3509

103e 130e 570e 277e 238e 138f,i 4342e

545 244 772 395 288 141 886

34a 394

72b 397

66c 400

257d 395

Pharmaceutical prices extracted from http://www.medicinpriser.dk/. Pharmaceutical prices extracted from http://www.stm.fi/en/ministry/boards/pharmaboard/notices/2013. Pharmaceutical prices extracted from http://www.legemiddelverket.no/English/about-norwegian-medicines-agency/Sider/default.aspx. Pharmaceutical prices extracted from http://www.tlv.se/beslut/sok/lakemedel/. Where country-specific costs were not available (i.e. costs of fatal events and maintenance costs), non-fatal costs were multiplied by a ratio between year 1 non-fatal and fatal events as well as year 1 non-fatal events and maintenance costs from Gerdtham et al. [21]. Cost calculation was based on cost ratios from Henriksson et al. [22]. Non-fatal event costs were applied. Cost includes one primary care visit and medications. Cost was sourced from Stewart et al. [23]. Cost was sourced from Schwartz et al. [24]. Currency exchange rate (2013): D 1 = DKK 7.4571; NOK 7.7969; SEK 8.6254.


ARTICLE IN PRESS



Table 3 – Utility decrements used in base case analysis. Utility decrement Event

First year

Subsequent year

Reference

Ischemic heart disease Myocardial infarction Congestive heart failure Stroke Pre-blind¥ Blind¥ ESRD¥ Transplant Amputation UTI/GI Hypoglycemia Severe Symptomatic BMI (per unit increase) BMI (per unit decrease)

0.090 0.055 0.108 0.164 0.029 0.074 0.263 0.075 0.280 0.003

0.090 0.055 0.108 0.164 0.029 0.074 0.263 0.075 0.280 0.003

[26] [26] [26] [26] [32] [32] [32] [26] [26] [33]

0.047 0.0142 0.014 0.014

0.000 0.000 0.014 0.014

[8] [8] [34] [34]

ESRD – end stage renal disease; UTI – urinary tract infection; GI – genital infection.

National Institute for Health and Welfare.3 For blindness, the non-fatal and maintenance costs were sourced from Stewart et al. [23] and Schwarz et al. [24] in respective countries. Only severe hypoglycemia is associated with a cost in the model [25]. For urinary tract and genital infections, costs were based on one primary care visit plus relevant medications. All costs were converted into local currencies using the relevant exchange rates and inflated to the 2013 price level by applying consumer price indices where needed.

3.3.

Health state utilities

Utility decrements associated with diabetes-related complications and side-effects are presented in Table 3. Given the absence of country specific utility decrements for diabetesrelated events, data from the UKPDS 62 study were used [26]. Utilities for end-stage renal disease, pre-blindness, and blindness, and side-effects were obtained from other sources (Table 3).

4.

Results

4.1.

Base case analysis

In the base case analysis metformin + dapagliflozin was associated with fewer diabetes-related complications compared to metformin + SU. The difference is more pronounced for macrovascular (cardiovascular) complications. The costs for most events and hypoglycemia are lower for dapagliflozin.

Diagnosis related group (DRG) system in Denmark: http://www.ssi.dk/Sundhedsdataogit/Sundhedsoekonomi/ Takster.aspx. 3 Cost data for the Finnish population extracted from the Information Management System database (HILMO) in 2011. The cost data were gathered from the diabetic population and show the yearly cost of diabetes-related events based on ICD-10.

5

However, costs associated with experiencing UTI/GIs as well as medication costs are higher. The mean lifetime QALY gain for metformin + dapagliflozin compared to metformin + SU is 0.247 in Denmark, 0.270 in Finland, 0.236 in Norway, and 0.278 in Sweden. The mean lifetime incremental cost for metformin + dapagliflozin compared to metformin + SU is D 1962, D 1467, D 1125, and D 1695 in respective countries. Hence the cost per QALY gained is D 7944 in Denmark, D 5424 in Finland, D 4769 in Norway, and D 6093 in Sweden. More detailed results for each country are presented in Table 4.

4.2.

Sensitivity analyses

The cost-effectiveness results were robust to changes in model input parameters, including costs and utilities of diabetes related and hypoglycemic events, efficacy parameters for the rescue therapy, and other variables (e.g. discount rate, clinical history, and alternative utility weights). Results of the one-way sensitivity analyses are reported for Sweden and presented in Table 5. In the one-way sensitivity analyses, the cost per QALY gained with metformin + dapagliflozin versus metformin + SU ranged from D 3704 to D 13,514 in Denmark, from D 2607 to D 9355 in Finland, from D 2276 to D 8167 in Norway, and from D 2940 to D 10,458 in Sweden. The results were most sensitive to the assumptions on body weight progression over time and utility weights associated with this change. A scenario where the mean body weight is set equal in the treatment and control arms in the year where insulin rescue therapy is initiated (e.g. the body weight advantage of patients formerly on dapagliflozin is removed as insulin is initiated) resulted in one of the highest incremental cost-effectiveness ratios (ICERs) in the sensitivity analysis. Applying utility weights related to body weight changes from Lane et al. [27], where the absolute value of the disutility associated with a BMI unit of weight increase is larger than the utility gain associated with a BMI unit of weight decrease (0.0472 and 0.0171 respectively), yielded the lowest cost per QALY gained with dapagliflozin. Raising the HbA1c threshold to 8% increased the mean cost per QALY gained. This is explained by the prolonged duration of the add-on treatments which results in higher medication costs and thus higher cost difference between the treatment and control arms. In the probabilistic sensitivity analysis (PSA), metformin + dapagliflozin generates a mean QALY gain of 0.264 in Denmark, 0.287 in Finland, 0.247 in Norway, and 0.297 in Sweden over a patient’s life-time at an average incremental cost of D 1752, D 1286, D 958, D 1458 in respective countries. Hence the mean cost per QALY is D 6627 in Denmark, D 4474 in Finland, D 3876 in Norway, and D 4906 with metformin + dapagliflozin compared to metformin + SU in respective countries. Fig. 2 presents the cost-effectiveness acceptability curve (CEAC) for the base case analysis in Finland based on the PSA. At a cost-effectiveness threshold value of D 50,000, for example, metformin + dapagliflozin is costeffective in 99% of the simulations in all four countries. Cost-effectiveness acceptability curves for the rest of the countries resemble the CEAC presented in Fig. 2.



ARTICLE IN PRESS

6


Table 4 – Results for Forxiga + metformin compared to SU + metformin. Denmark Total events predicted Macrovascular events IHD MI CHF Stroke Microvascular events Blindness Nephropathy Amputation Fatal events Macrovascular Microvascular Non-diabetes

Met + dapa

Met + SU

1265 3181 1036 1063

1300 3270 1050 1127

798 308 559 1773 449 7338

Finland Diff.

Met + dapa

Met + SU

Diff.

Met + dapa

Met + SU

Diff.

Met + dapa

Met + SU

−35 −89 −14 −64

1393 3605 1198 1241

1428 3702 1209 1308

−35 −97 −11 −67

1362 3489 1153 1189

1398 3586 1164 1257

−36 −97 −11 −67

1438 3733 1243 1289

1472 3831 1253 1358

−35 −98 −11 −70

792 332 579

6 −24 −20

898 375 669

890 404 693

8 −29 −23

869 353 633

864 380 658

5 −27 −25

930 392 699

920 421 722

10 −29 −24

1853 472 7248

−80 −23 90

2056 566 6822

2145 594 6722

−88 −28 100

1973 526 7079

2060 553 6976

−87 −28 103

2135 593 6647

2226 622 6544

−92 −29 102

Finland

Met + SU

Macrovascular events 6,660,748 IHD 17,887,168 MI 6,883,578 CHF Stroke 4,482,483 Microvascular events 2,907,525 Blindness 3,627,815 Nephropathy 4,355,943 Amputation 1,077,461 Hypoglycemia 113,107,749 Treatmenta 224,400 Adverse events 161,214,870 Total

Sweden

Met + dapa

Denmark Total costs (D )

Norway

Met + dapa

Norway

Met + SU

Diff.

Sweden

Met + dapa

Met + SU

Met + dapa

Met + SU

6,856,112 18,340,903 6,967,028 4,758,086

5,059,161 11,319,762 4,332,625 5,847,214

5,196,913 11,595,736 4,367,485 6,173,870

2,747,380 9,665,333 5,313,417 5,317,583

2,827,733 9,889,089 5,362,973 5,630,782

17,413,836 24,118,985 8,993,445 9,605,651

17,860,297 24,688,178 9,061,106 10,146,209

2,882,114 3,947,556 4,529,279 1,156,316 92,070,117 87,790 141,595,301

10,834,643 8,849,348 4,036,200 1,178,971 95,259,831 479,444 147,197,199

10,723,293 9,561,527 4,189,863 1,256,738 79,274,026 187,724 132,527,175

2,791,150 8,889,512 4,880,301 1,030,494 68,959,373 431,929 110,026,471

2,773,243 9,609,158 5,092,529 1,109,911 56,316,701 169,028 98,781,147

3,459,487 2,320,829 7,599,654 1,208,499 90,889,061 1,711,700 167,321,147

3,422,891 2,503,651 7,882,941 1,284,997 72,855,550 669,260 150,375,080

Denmark

Finland

Norway

Sweden

Hypoglycemiab

Met + dapa

Met + SU

Met + dapa

Met + SU

Met + dapa

Met + SU

Met + dapa

Met + SU

Symptomatic Severe

121,914 4087

132,030 4281

136,079 4570

145,959 4755

132,291 4439

142,297 4626

140,537 4719

150,383 4899

Denmark Cost-effectiveness Per patient Discounted costs (D ) Discounted QALYs Discounted LYG Cost/QALY (D )

Finland

Norway

Met + dapa

Met + SU

Diff.

Met + dapa

Met + SU

Diff.

Met + dapa

16,121 13.068 15.860

14,160 12.821 15.788

1962 14,720 0.24713.924 0.07216.951 7944

13,253 13.654 16.861

1467 11,003 0.27012.434 0.09015.057 5424

Sweden

Met + SU

Diff.

Met + dapa

Met + SU

9878 12.198 14.990

1125 16,732 0.23614.253 0.06717.358 4769

15,038 13.975 17.263

Diff.

1695 0.278 0.095 6093

Analysis based on 1000 runs (simulations) for 10,000 patients. Treatment cost includes rescue therapy (insulin) drug costs. b Number of hypoglycemia events includes those experienced on insulin rescue therapy. IHD – ischemic heart disease; MI – myocardial infarction; CHF – congestive heart failure; ESRD – end stage renal disease. a

5.

Discussion

Our analysis indicates that for Nordic T2DM patients who are inadequately controlled on metformin alone, metformin + dapagliflozin is a cost-effective treatment strategy compared to metformin + SU. The model estimated that adding dapagliflozin to metformin is associated with an incremental benefit of 0.236–0.278 QALYs at an incremental cost

raging from D 1125 to D 1962, which results in an ICER ranging from D 4769 to D 7944. This indicates that adding dapagliflozin to metformin produces health outcomes at an acceptable cost in all Nordic countries. In addition, these results were robust to changes in relevant model input parameters. The comparison between metformin + dapagliflozin and metformin + SU is based on a randomized, 52-week noninferiority HbA1c trial [15] where dapagliflozin was associated with weight loss and significantly decreased proportion of



ARTICLE IN PRESS 7


Table 5 – One-way sensitivity analysis for Forxiga + metformin versus SU + metformin in Sweden (results per patient).

Utility weight 0.0171 per unit BMI decrease and 0.0472 per unit BMI increase in all years [27] Utility weights per unit BMI change reduced by 75% in subsequent years (0.0035) Forxiga weight effect maintained for 3 years Mean weight equal for treatment and control arm in the year where rescue therapy is initiated (i.e. no weight advantage for dapagliflozin cohort as NPH insulin is initiated) HbA1c threshold value for rescue therapy (insulin) 64 mmol/mol (8%) SU no. symptomatic hypoglycemias as dapagliflozin (0.035) Cost severe hypoglycemia SEK 256 All event costsa 30% lower All event costsa 30% higher Alternative utility weightsb applied Clinical history: 20% with all the conditions Discount rate (costs and benefits) 5%

QALY

Cost (D )

0.575

1690

2940

ICER

0.166 0.288 0.172

1690 1690 1799

10,193 5875 10,458

0.291 0.250 0.276 0.276 0.276 0.262 0.287 0.223

2806 1690 1697 1753 1627 1690 1787 1705

9643 6765 6156 6358 5904 6440 6233 7649

QALYs gained, discounted total cost difference, and cost/QALY, refer to the metformin + dapagliflozin treatment (e.g. QALYs with metformin + dapagliflozin as compared to metformin +SU). a Analysis based on 1000 runs for 1000 patients. Everything else as in base case. b Includes utilities for all macro and microvascular events but not for hypoglycemia and GI/UTIs.

Fig. 2 – Cost-Effectiveness Acceptability Curve (CEAC) for metformin + dapagliflozin versus metformin + SU in Finland.

hypoglycemic events versus SU associated with weight increase and more frequent hypoglycemic events [15]. Therefore, the most important drivers of the cost-effectiveness results are side-effects associated with the treatment strategies, particularly weight changes as adding dapagliflozin to metformin results in a QALY gain as dapagliflozin is associated with weight loss and SU with weight gain. Hence, utility values associated with a unit change in BMI have implications for the results. The values used in our analysis are in line with those in other studies [28]. We assume that the increase in utility associated with a unit weight loss is equal to the disutility associated with a unit weight gain. However, in a Canadian study weight gain was shown to have a larger impact on patient utility than weight loss [27], which indicates that our analysis may be conservative. Additionally, HRQoL decrements associated with weight changes are modeled linearly. This approach is simple and transparent, but also conservative

given that the disutility of weight gain increases with body weight [28]. The duration of weight loss has implications for the costeffectiveness results. In the study of Nauck et al. [15], which provides efficacy and safety data for the model, the weight reduction in the dapagliflozin arm was maintained over the two-year duration of the study. In the base case analysis we assumed that patients treated with metformin + dapagliflozin return to their baseline weight after two years. Given a noninsulin dependent mode of action, dapagliflozin improves glycemic control in T2DM patients by reducing renal glucose reabsorption leading to urinary excretion of excess glucose which is sustained for the duration of the treatment. Therefore it is reasonable to assume that the placebo-corrected weight reduction seen with dapagliflozin would be maintained throughout the course of treatment, as sustainability of the weight effect was observed in other dapagliflozin clinical



ARTICLE IN PRESS

8


studies [29]. If the weight effect were to be maintained beyond 2 years, it would further improve the dapagliflozin results. In this respect, our cost-effectiveness analysis seems to be conservative. In addition to weight changes, the results are also affected by the occurrence of diabetes-related events and to a lesser extent by other side effects such as hypoglycemia and UTIs/GIs. The direct utility effect as well as cost of hypoglycemic events and UTI/GIs affect the results relatively less than weight loss as only a fraction of the patients are affected each year and the rates do not differ much between the arms. Hence the impact on the results is much smaller. Applying alternative utility weights and costs also results in only marginal changes in the results. Our analysis highlights that it is important to consider not only a drug’s effect on HbA1c in the treatment of T2DM, but also its effect on weight. Type 2 diabetes is associated with obesity and achieving weight control is often challenging given that commonly used diabetes drugs (sulfonylurea and insulin) are associated with weight gain. Not only does weight gain have negative effects on patients’ HRQoL, but it also increases the risk for cardiovascular events and subsequent mortality [11]. In addition, it has been shown that even a small reduction in body weight is meaningful and leads to better treatment outcomes. Weight loss of 2–5% of body weight increased the odds of having improvements in systolic blood pressure, HbA1c, and triglycerides [30]. As strength of our analysis is that it is performed by a well-validated model which is a refined version of the UKPDS Outcomes Model [17] and was also based on the UKPDS risk equations [18]. Different from other models, our model allows differentiation between different types of hypoglycemic events, which makes it possible to assign relevant costs and utility weights. An additional strength is that the study was conducted in four Nordic countries. The present study has some limitations. Similar to other studies assessing the cost-effectiveness of diabetes treatments [16,24], this analysis is based on short-term data providing surrogate markers (i.e. HbA1c, blood pressure, total cholesterol) that are used to predict the occurrence of future diabetes-related events and death by using UKPDS equations. However, even if no long-term data are available and the UKPDS equations are used to simulate a range of longterm outcomes, the model is able to simulate event histories that closely match event rates observed in the UKPDS study [18]. Ideally, country specific utilities should be used for the diabetes-related events but since there are no studies covering all diabetes-related events for the Nordic countries that would require combining utilities from a diverse set of sources. The UKPDS utilities have the advantage of being estimated with a consistent methodology which makes the utilities comparable across different diabetes-related events. In addition, as the analyses were performed for four countries, input parameters (medication and event costs) and all-cause mortality varied across the countries, particularly, costs of diabetes-related events. For instance, using DRGs may underestimate the annual direct medical costs of diabetesrelated events since DRGs do not capture the costs associated with rehabilitation, outpatient visits, and subsequent medication use. The fact that Denmark had a higher cost per QALY

gained than the other countries is explained by differences in unit costs in combination with no VAT deduction for the medication costs. The effect of varying event costs is tested in a one-way sensitivity analysis which affected the results only marginally. Although direct comparisons between countries should be done with caution, one may notice that the results were fairly similar in all four countries. Due to a lack of data, neither indirect costs nor direct medical costs due to overweight (unless related to macroand microvascular events) are included in our analyses. However, being overweight is associated with considerable direct healthcare costs as weight gainers incur significantly higher mean diabetes-related costs than those who do not gain weight [2]. Furthermore, diabetes-related events, weight gain and hypoglycemia should have considerable impact on productivity (indirect costs) which was not considered in the analysis. Hence, we may underestimate the societal benefit associated with dapagliflozin treatment.

6.

Conclusions

To conclude, across four Nordic countries, adding dapagliflozin to metformin is a cost-effective treatment alternative for Type 2 diabetes patients who are inadequately controlled on metformin alone.

Conflicts of interest US, ME, and KB are employees of AstraZeneca. OG was an employee of AZ at the time of his work on this study. PM acts as a consultant for AZ and BMS.

Acknowledgement The manuscript was funded by AstraZeneca Nordic-Baltic. Study was performed by AZ employees. Phil McEwan was acting as a consultant. Hence, not grant was received for this publication.

references

[1] M. Bogelund, T. Vilsboll, J. Faber, J.E. Henriksen, R.P. Gjesing, M. Lammert, Patient preferences for diabetes management among people with type 2 diabetes in Denmark – a discrete choice experiment, Curr. Med. Res. Opin. 27 (November (11)) (2011) 2175–2183. [2] A.P. Yu, E.Q. Wu, H.G. Birnbaum, S. Emani, M. Fay, G. Pohl, et al., Short-term economic impact of body weight change among patients with type 2 diabetes treated with antidiabetic agents: analysis using claims, laboratory, and medical record data, Curr. Med. Res. Opin. 23 (September (9)) (2007) 2157–2169. [3] P. Zhang, X. Zhang, J. Brown, D. Vistisen, R. Sicree, J. Shaw, et al., Global healthcare expenditure on diabetes for 2010 and 2030, Diabetes Res. Clin. Pract. 87 (March (3)) (2010) 293–301. [4] UKPDS 33, Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2



ARTICLE IN PRESS p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 4 ) xxx–xxx

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group, Lancet 352 (September (9131)) (1998) 837–853. NICE, Type 2 diabetes: the management of type 2 diabetes, in: NICE Clinical Guideline 87, National Institute for Health and Clinical Excellence, 2009, May. S.E. Inzucchi, R.M. Bergenstal, J.B. Buse, M. Diamant, E. Ferrannini, M. Nauck, et al., Management of hyperglycaemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD), Diabetologia 55 (June (6)) (2012) 1577–1596. L.S. Matza, K.S. Boye, N. Yurgin, J. Brewster-Jordan, S. Mannix, J.M. Shorr, et al., Utilities and disutilities for type 2 diabetes treatment-related attributes, Qual. Life Res. 16 (September (7)) (2007) 1251–1265. C.J. Currie, C.L. Morgan, C.D. Poole, P. Sharplin, M. Lammert, P. McEwan, Multivariate models of health-related utility and the fear of hypoglycaemia in people with diabetes, Curr. Med. Res. Opin. 22 (August (8)) (2006) 1523–1534. J.W. Anderson, C.W. Kendall, D.J. Jenkins, Importance of weight management in type 2 diabetes: review with meta-analysis of clinical studies, J. Am. Coll. Nutr. 22 (October (5)) (2003) 331–339. A.L. Peters, Patient and treatment perspectives: revisiting the link between type 2 diabetes, weight gain, and cardiovascular risk, Cleve. Clin. J. Med. 76 (December (Suppl. 5)) (2009) S20–S27. J. Bodegard, J. Sundstrom, B. Svennblad, C.J. Ostgren, P.M. Nilsson, G. Johansson, Changes in body mass index following newly diagnosed type 2 diabetes and risk of cardiovascular mortality: a cohort study of 8486 primary-care patients, Diabetes Metab. 39 (September (4)) (2013) 306–313. K. Bolin, C. Gip, A.C. Mork, B. Lindgren, Diabetes, healthcare cost and loss of productivity in Sweden 1987 and 2005 – a register-based approach, Diabet. Med. 26 (September (9)) (2009) 928–934. F.X. Pi-Sunyer, The impact of weight gain on motivation, compliance, and metabolic control in patients with type 2 diabetes mellitus, Postgrad. Med. 121 (September (5)) (2009) 94–107. S.A. Ross, G. Dzida, J. Vora, K. Khunti, M. Kaiser, R.J. Ligthelm, Impact of weight gain on outcomes in type 2 diabetes, Curr. Med. Res. Opin. 27 (July (7)) (2011) 1431–1438. M.A. Nauck, S. Del Prato, J.J. Meier, S. Duran-Garcia, K. Rohwedder, M. Elze, et al., Dapagliflozin versus glipizide as add-on therapy in patients with type 2 diabetes who have inadequate glycemic control with metformin: a randomized, 52-week, double-blind, active-controlled noninferiority trial, Diabetes Care 34 (September (9)) (2011) 2015–2022. O. Granstrom, K. Bergenheim, P. McEwan, K. Sennfalt, M. Henriksson, Cost-effectiveness of saxagliptin (Onglyza(R)) in type 2 diabetes in Sweden, Prim. Care Diabetes 6 (July (2)) (2012) 127–136. Mount Hood 4 Modeling Group, Computer modeling of diabetes and its complications: a report on the Fourth Mount Hood Challenge Meeting, Diabetes Care 30 (June (6)) (2007) 1638–1646. P.M. Clarke, A.M. Gray, A. Briggs, A.J. Farmer, P. Fenn, R.J. Stevens, et al., A model to estimate the lifetime health outcomes of patients with type 2 diabetes: the United Kingdom Prospective Diabetes Study (UKPDS) Outcomes Model (UKPDS no. 68), Diabetologia 47 (October (10)) (2004) 1747–1759. J. Lundkvist, C. Berne, B. Bolinder, L. Jönsson, The economic and quality of life impact of hypoglycemia, Eur. J. Health Econ. 50 (2005) 197–202.

9

[20] N. Waugh, E. Cummins, P. Royle, C. Clar, M. Marien, B. Richter, et al., Newer agents for blood glucose control in type 2 diabetes: systematic review and economic evaluation, Health Technol. Assess. 14 (July (36)) (2010) 1–248. [21] U.G. Gerdtham, P. Clarke, A. Hayes, S. Gudbjornsdottir, Estimating the cost of diabetes mellitus-related events from inpatient admissions in Sweden using administrative hospitalization data, Pharmacoeconomics 27 (1) (2009) 81–90. [22] F. Henriksson, Applications of economic models in healthcare: the introduction of pioglitazone in Sweden, Pharmacoeconomics 20 (Suppl. 1) (2002) 43–53. [23] W.C. Stewart, J.A. Stewart, M.A. Mychaskiw, Cost-effectiveness of latanoprost and timolol maleate for the treatment of glaucoma in Scandinavia and the United Kingdom, using a decision-analytic health economic model, Eye (Lond.) 23 (January (1)) (2009) 132–140. [24] B. Schwarz, M. Gouveia, J. Chen, G. Nocea, K. Jameson, J. Cook, et al., Cost-effectiveness of sitagliptin-based treatment regimens in European patients with type 2 diabetes and haemoglobin A1c above target on metformin monotherapy, Diabetes Obes. Metab. 10 (June (Suppl. 1)) (2008) 43–55. [25] L. Jonsson, B. Bolinder, J. Lundkvist, Cost of hypoglycemia in patients with Type 2 diabetes in Sweden, Value Health 9 (May–June (3)) (2006) 193–198. [26] P. Clarke, A. Gray, R. Holman, Estimating utility values for health states of type 2 diabetic patients using the EQ-5D (UKPDS 62), Med. Decis. Making 22 (July–August (4)) (2002) 340–349. [27] S. Lane, A. Levy, J. Sambrook, J. Mukherjee, M. Leahy, H. Tildesley, The impact on utilities of weight loss and weight gain among Canadian patients with type 2 diabetes, Value Health 15 (4) (2012) A183. [28] S.L. Dennett, K.S. Boye, N.R. Yurgin, The impact of body weight on patient utilities with or without type 2 diabetes: a review of the medical literature, Value Health 11 (May–June (3)) (2008) 478–486. [29] J.P. Wilding, V. Woo, N.G. Soler, A. Pahor, J. Sugg, K. Rohwedder, et al., Long-term efficacy of dapagliflozin in patients with type 2 diabetes mellitus receiving high doses of insulin: a randomized trial, Ann. Intern. Med. 156 (March (6)) (2012) 405–415. [30] R.R. Wing, W. Lang, T.A. Wadden, M. Safford, W.C. Knowler, A.G. Bertoni, et al., Benefits of modest weight loss in improving cardiovascular risk factors in overweight and obese individuals with type 2 diabetes, Diabetes Care 34 (July (7)) (2011) 1481–1486. [31] C. Fajardo Montanana, C. Hernandez Herrero, M. Rivas Fernandez, Less weight gain and hypoglycaemia with once-daily insulin detemir than NPH insulin in intensification of insulin therapy in overweight Type 2 diabetes patients: the PREDICTIVE BMI clinical trial, Diabet. Med. 25 (August (8)) (2008) 916–923. [32] C.J. Currie, P. McEwan, J.R. Peters, T.C. Patel, S. Dixon, The routine collation of health outcomes data from hospital treated subjects in the Health Outcomes Data Repository (HODaR): descriptive analysis from the first 20,000 subjects, Value Health 8 (September-October (5)) (2005) 581–590. [33] H.C. Barry, M.H. Ebell, J. Hickner, Evaluation of suspected urinary tract infection in ambulatory women: a cost-utility analysis of office-based strategies, J. Fam. Pract. 44 (January (1)) (1997) 49–60. [34] J.J. Caro, I.O. Stillman, A. Danel, D. Getsios, P. McEwan, Cost effectiveness of rimonabant use in patients at increased cardiometabolic risk: estimates from a Markov model, J. Med. Econ. 10 (2007) 239–254.


Sitagliptin added to stable insulin therapy with or without metformin in Chinese patients with type 2 diabetes.

Combination therapy for type 2 diabetes: dapagliflozin plus metformin.

The cost-effectiveness of dapagliflozin versus sulfonylurea as an add-on to metformin in the treatment of Type 2 diabetes mellitus.

Starting with low dose sulfonylurea and metformin in early stage type 2 diabetes mellitus.

Vildagliptin added to sulfonylurea improves glycemic control without hypoglycemia and weight gain in Chinese patients with type 2 diabetes mellitus.

The efficacy and safety of liraglutide added to metformin in patients with diabetes: a meta-analysis of randomized controlled trials.

Cardiovascular risks associated with second-line oral antidiabetic agents added to metformin in patients with Type 2 diabetes: a nationwide cohort study.

Linagliptin added to sulphonylurea in uncontrolled type 2 diabetes patients with moderate-to-severe renal impairment.

Treatment escalation options for patients with type 2 diabetes after failure of exenatide twice daily or glimepiride added to metformin: results from the prospective European Exenatide (EUREXA) study.

The target of metformin in type 2 diabetes.

Sitagliptin added to voglibose monotherapy improves glycemic control in patients with type 2 diabetes.

Metformin powder formulation compared to metformin tablets on glycemic control and on treatment satisfaction in subjects with type 2 diabetes mellitus.

Higher prevalence of metformin-induced vitamin B12 deficiency in sulfonylurea combination compared with insulin combination in patients with type 2 diabetes: a cross-sectional study.

Twice-daily dapagliflozin co-administered with metformin in type 2 diabetes: a 16-week randomized, placebo-controlled clinical trial.

Efficacy and safety of sitagliptin compared with sulfonylurea therapy in patients with type 2 diabetes showing inadequately controlled glycosylated hemoglobin with metformin monotherapy: A meta-analysis.

Mortality risk with sulphonylureas compared to metformin.

other oral antidiabetes agents in patients with type 2 diabetes and metabolic syndrome.

Empagliflozin as add-on to metformin in people with Type 2 diabetes.

Sodium-glucose cotransporter-2 inhibitor combination therapy to optimize glycemic control and tolerability in patients with type 2 diabetes: focus on dapagliflozin-metformin.

Metformin decreases glycated albumin to glycated haemoglobin ratio in patients with newly diagnosed type 2 diabetes.

Effects of metformin on endothelial function in type 2 diabetes.

Postprandial dynamics of plasma glucose, insulin, and glucagon in patients with type 2 diabetes treated with saxagliptin plus dapagliflozin add-on to metformin therapy.

Efficacy and safety of dulaglutide added onto pioglitazone and metformin versus exenatide in type 2 diabetes in a randomized controlled trial (AWARD-1).

Obese Adolescents With Type 1 Diabetes: A Randomized Clinical Trial.