The place of DPP-4 inhibitors in the treatment algorithm ...€¦ · the country and year of each study, DPP-4 inhibitors were highly cost-effective as second-line, as add-ons to

ORIGINAL PAPER

The place of DPP-4 inhibitors in the treatment algorithmof diabetes type 2: a systematic review of cost-effectiveness studies

Alexandre Baptista1• Ines Teixeira2

• Sonia Romano2• Antonio Vaz Carneiro3

•

Julian Perelman4

Received: 9 October 2015 / Accepted: 30 September 2016 / Published online: 17 October 2016

� Springer-Verlag Berlin Heidelberg 2016

Abstract

Objective To conduct a systematic review of cost-effec-

tiveness, cost-utility, and cost-benefit studies of DPP-4

inhibitors for diabetes treatment versus other antidiabetics.

Methods Three investigators searched the CRD York,

Tufts CEA Registry, and MEDLINE databases through

2015. We reviewed all potentially relevant titles and

abstracts, and screened full-text articles, according to

inclusion criteria. We established a quality score for each

study based on a 35-item list.

Results A total of 295 studies were identified, of which 20

were included. The average quality score was 0.720 on a

0–1 scale. All studies were performed in high- and middle-

income countries, using a 3rd-party payer perspective and

randomized clinical trials to measure effectiveness. Sita-

gliptin, saxagliptin and vildagliptin had an ICER below

25,000 €/QALY, as second-line and as add-ons to met-

formin, in comparison to sulfonylureas. When compared

with sitagliptin, liraglutide (GLP-1 receptor agonist) had an

ICER of up to 22,724 €/QALY for the 1.2-mg dosage, and

up to 32,869 €/QALY for the 1.8-mg dosage. Insulin

glargine was dominant when compared with sitagliptin.

Conclusions According to the WHO threshold applied to

the country and year of each study, DPP-4 inhibitors were

highly cost-effective as second-line, as add-ons to met-

formin, in comparison with sulfonylureas. More recent

therapies (GLP-1 receptor agonists and insulin glargine)

were highly cost-effective in comparison to DPP-4 inhi-

bitors. These results were obtained, however, on the basis

of a limited number of studies, relying on the same few

clinical trials, and financed by manufacturers. Further

independent research is needed to confirm these findings.

Keywords Diabetes � Cost-effectiveness studies � DPP-4inhibitors � Systematic review

JEL Classification I1

Introduction

Diabetes is one of the largest global health emergencies of the

twenty-first century. In addition to the 415 million adults

estimated to have diabetes in 2015, there are 318 million

adults with impaired glucose tolerance, which puts them at

high risk of developing the disease in the future [1]. Also,

according to the International Diabetes Federation (IDF), 5

million people died of diabetes in 2015.Uppermiddle-income

and high-income countries are those where the prevalence of

diabetes is highest amongst those 20–79 years old [1].

Regarding the economic burden, the majority of high-in-

come countries spend between 5 and 20 % of their total health

expenditure on diabetes [1]. In France total direct costs reached

12.9 € billion in 2010. In Germany the total direct cost burden

& Alexandre Baptista

[email protected]

1 Unit of Epidemiology of the Faculty of Medicine of Lisbon,

Edifıcio Egas Moniz, Faculdade de Medicina da

Universidade de Lisboa, Av. Professor Egas Moniz,

1649-028 Lisbon, Portugal

2 Centre for Health Evaluation and Research (CEFAR),

National Association of Pharmacies Group, R. Marechal

Saldanha, 1., 1249-069 Lisbon, Portugal

3 Center for Evidence-Based Medicine (CEMBE) of the

Faculty of Medicine at the University of Lisbon, Faculdade

de Medicina da Universidade de Lisboa, Av. Professor Egas

Moniz, 1649-028 Lisbon, Portugal

4 Escola Nacional de Saude Publica and Centro de

Investigacao em Saude Publica, Universidade Nova de

Lisboa, Avenida Padre Cruz, 1600-5605 Lisbon, Portugal

123

Eur J Health Econ (2017) 18:937–965

DOI 10.1007/s10198-016-0837-7

http://crossmark.crossref.org/dialog/?doi=10.1007/s10198-016-0837-7&domain=pdf

http://crossmark.crossref.org/dialog/?doi=10.1007/s10198-016-0837-7&domain=pdf

arising from treatment has been estimated at 43.2 € billion for2010. Total direct cost in the UK has been estimated at £13.8

billion (20.2 € billion using the base year rate of exchange) [2].In 2012 the total estimated cost of diagnosed diabetes in the

United States was $245 billion, including $176 billion in direct

medical costs and $68 billion in indirect costs [3]. These indi-

rect costs included increased absenteeism ($5 billion) and

reduced productivity while at work ($20.8 billion) for the

employed population, reduced productivity for those not in the

labour force ($2.7 billion), inability to work as a result of dis-

ease-related disability ($21.6 billion), and lost productive

capacity due to early mortality ($18.5 billion). On average,

medical expenditures among people with diabetes are approx-

imately 2.3 times higher than among thosewithout diabetes [3].

As a result, diabetes has been considered to be a research

priority, leading to a large increase in recent years in the

number of glucose lowering medicines for treating type 2

diabetes. Dipeptidyl peptidase-4 inhibitors (DPP-4 inhibi-

tors) are a relatively new oral hypoglycaemic drugs group.

Among these, sitagliptin, vildagliptin, saxagliptin, linaglip-

tin, and alogliptin are currently approved by theUS Food and

Drug Administration and the European Medicines Agency,

while others are awaiting approval or are in development. In

practice, the American Diabetes Association (ADA) and the

European Association for the Study of Diabetes (EASD)

clinical guideline for type 2 diabetes, suggested adding a

DPP-4 inhibitor, insulin, glucagon-like peptide 1 (GLP-1)

receptor agonists, sodium-glucose cotransporter 2 (SGLT2)

inhibitors, thiazolidinedione, or a sulfonylurea as second-

line treatment to first-line metformin [4].

Drug choice is based on patient preferences as well as

various patient, disease, and drug characteristics,with the goal

being to reduce glucose concentrationswhile minimizing side

effects, especially hypoglycemia [4]. However, the guideline

did not incorporate cost-effectiveness considerations regard-

ing the newest medicines, such as DPP-4 inhibitors or SGLT2

inhibitors, which are nevertheless essential as an instrument to

help resource allocation decisions. The importance is also

emphasized in the current context of economic recession and

pressure on tight public budgets, and considering the high

epidemiological and economic burden of the disease.

The objective of this study was to conduct a systematic

review of cost-effectiveness, cost-utility, and cost-benefit

studies of DPP-4 inhibitors versus other antidiabetics as

treatment of type 2 diabetesmellitus (T2DM), and understand

the implications for guidelines, policy, and further research.

Methods

This review followed the methodology recommended by

the Preferred Reporting Items for Systematic Reviews and

Meta-Analyses (PRISMA) statement [5], and by the Centre

for Reviews and Dissemination (CRD) of the University of

York for systematic reviews of economic evaluations [6].

The literature review was based on a search for journal

articles and abstracts in Medline and the CRD database

from 1996 to 2015, and NHS EED, the Health Economic

and Evaluations Database (HEED), and the Tufts CEA

Registry to 2015. Google scholar was also searched. In

addition, relevant grey literature, including models pre-

sented at recent professional meetings available solely as

abstracts in conference proceedings, were also explored.

Searches for economic outcomes were conducted using

a variety of terms from the medical literature to describe

the intervention, the comparator, the target patient popu-

lation, the outcomes, and the study design. A combination

of these search terms was also used for the survey. The

search terms were: sitagliptin, vildagliptin, saxagliptin,

linagliptin, alogliptin, DPP-4 inhibitors, cost-effectiveness,

cost-utility, and cost-benefit. The search strings used for

the PubMed searches are available in the Appendix.

Firstly, three investigators independently reviewed all

potentially relevant titles and abstracts (1st screening) and

subsequently screened full-text articles (2nd screening),

according to pre-established inclusion criteria. Inclusion

criteria followed the PICOS approach [7]. PICOS is the

acronym for population, intervention, comparator, out-

comes, and study design. We followed the PRISMA

flowchart in reporting study selection, as suggested by the

PRISMA statement [5].

Cost-effectiveness, cost-utility studies, and cost-benefit

studies should be available as a full-text publication and

published in English, French, Spanish, or Portuguese lan-

guages. We excluded the incomplete economic evalua-

tions, namely: cost consequence analyses (4 studies);

patient reported outcomes (PRO) studies (1 study); the

studies on sub-populations that cannot be generalized (2

studies); the health technology assessment agencies reports

that were not submitted to peer-review (36 studies); studies

out of scope (162), and only abstracts (66).

Secondly, the three investigators used a standardized

data abstraction template, as recommended by the CRD of

the University of York for systematic reviews of economic

evaluations [6], to independently extract data from each

study, with disagreements being resolved by discussion.

For each study, the information was extracted and recorded

in a specific template, provided in the Appendix.

Thirdly, a critical appraisal of the methodology and

reporting was performed focusing on key quality issues,

such as: methods of deriving the effectiveness data; mea-

surement and valuation of resource data; measurement and

valuation of health benefits (utilities); method of synthe-

sizing the costs and effects; analysis of uncertainty; and

external validity. To do so, we used the 35-item version of

the BMJ checklist [7]. A score in percentage was attributed

938 A. Baptista et al.

123

to each study, calculating the affirmative answers in the

checklist. Three investigators performed all quality

assessments independently, with disagreement resolved

through discussion.

Finally, we reported summary statistics and qualitative

(descriptive) syntheses of identified cost-effectiveness,

cost-utility, and cost-benefit studies in the form of sum-

mary tables. Categorical data were reported as percentages,

while continuous data were reported as means with confi-

dence intervals or standard deviations. A comparative

qualitative synthesis was performed to explore relation-

ships within and between studies.

Results

The literature search initially identified 295 citations

(Fig. 1). Of these, 24 cost-effectiveness studies were

accepted after the 1st screening, and 20 were accepted after

the 2nd screening.

Characteristics of the selected publications

Main details of the selected studies are presented in

Table 1. We organized these results according to the tim-

ing of the market introduction, that is: (1) sitagliptin versus

sulfonylurea; (2) saxagliptin versus sulfonylurea; (3) sax-

agliptin versus insulin; (4) saxagliptin versus pioglitazone;

(5) vildagliptin versus sulfonylurea; (6) liraglutide versus

sitagliptin; (7) insulin glargine versus sitagliptin; and (8)

GLP-1 receptor agonists versus DPP-4 inhibitors.

The studies were all performed in high-income and

middle-income countries (using the World Bank classifi-

cation) [8]. The high-income countries were Austria,

Canada, Finland, Germany, Greece, Poland, Portugal,

Spain, Sweden, United Kingdom (UK), and United States

(US). The middle-income countries were Argentina and

Brazil. These are very recent studies with retrieval of data:

one study from 2014, four studies from 2013, five studies

from 2012, two studies from 2011, one study from 2010,

four studies from 2009, two studies from 2008, and one

1st screening

2nd screening

Relevant titles and abstracts identified after duplicated removed n= 295

Full copies assessed and retrieved for eligibility n= 24

Excluded n = 271 Agencies reports = 36 Sub-populations= 2 PRO studies = 1 Cost consequence study =4 Out of scope= 162 Only abstracts =66

Excluded n= 4 Out of scope = 4

Studies included in the review n= 20

Relevant titles and abstracts identified n= 303

Flowchart of the literature search

Fig. 1 Flowchart of the literature search

The place of DPP-4 inhibitors in the treatment algorithm of diabetes type 2: a systematic… 939

123

Ta

ble

1Characteristicsoftheselected

publications

Reference/country/authors

(publicationyear)

Yearof

retrieval

ofdata

Population

Interventions,designandtime-horizon

Costs,benefits,discountrate

andperspective

Score

[12]/6Europeancountries

(Austria,

Finland,

Portugal,Spain,Scotland-

UK,Sweden)/Schwarz

etal.(2008)

2007

Meanage56.7–64.9

years

MeanHbA1c7.5–8.09%

Sitagliptin(?

Met)vsrosiglitazoneor

sulfonylurea(?

Met)

Cost-utility

analysis:

discreteevent

simulation(JADEmodel)

Tim

e-horizon:patientlifetime

Directcost:drugs,adverse

events,

(hypoglycaem

ia,weight),macro-and

microvascularcomplications

Benefits:QALY

Discountrate:Austria/Sweden

3%;Spain6%;

Finland/Portugal

5%;Scotland3.5

%(costs

andbenefits)

Perspective:

3rd

party

payer

0.714

[13]/Portugal

/Pereira

etal.

(2012)

2010

MeanHbA1c6.5–11%

Sitagliptin(?

Met)vssulfonylurea

(?Met)

Cost-utility

analysis:

discreteevent

simulation(JADEmodel)

Tim

e-horizon:50years

(patientlifetime)

Directcost:drugs,macro-andmicrovascular

complications,hypoglycaem

ia

Benefits:QALY

Discountrate:5%

costsandbenefits

(0–3%

SA)

Perspective:

societal

(butonly

withdirectcosts)

0.629

[14]/Argentina/Elgartet

al.

(2013)

2009

Meanage64years

Proportionofmen

53%

MeanHbA1c7.7

%

Meandurationofdiabetes

10.5

years

Saxagliptin(?

Met)vssulfonylurea

(?Met)

Cost-utility

analysis:

discreteevent

simulation(Cardiffdiabetes

model)

Tim

e-horizon:20years

Directcost:drugs,adverse

events,macro-and

microvascularcomplications

Benefits:QALY

andLYG

Discountrate:3.5

%costsandbenefits

Perspective:

3rd

party

payer

0.829

[15]/Germany

/Erhardtet

al.

(2012)

2009

Meanage57.55years

Proportionofmen

52%

MeanHbA1c7.65%


5.4

years

Saxagliptin(?

Met)vssulfonylurea

(?Met)

Cost-utility

analysis:

discreteevent


model)

Tim

e-horizon:40years

(patientlifetime)

Directcost:medicines,hypoglycaem

ia

treatm

ent,diabetes

relatedcomplications

Benefits:QALY

andLYG

Discountrate:3%

costsandbenefits(0,5,7and

10%

SA)

Perspective:

3rd

party

payer

0.714

[16]/Sweden/Granstrom

etal.(2012)

2008

Meanage57.55years

Proportionofmen

52%

MeanHbA1c7.65%


5.4

years

Saxagliptin(?

Met)vssulfonylurea

(?Met)

Cost-utility

analysis:

discreteevent

simulationmodel

Tim


Directcost:drugs,BGSM,macro-and

microvascularcomplications,hypoglycaem

ia

Benefits:QALY

andLYG

Discountrate:3%

costsandbenefits

Perspective:

3rd

party

payer

0.771

[17]/US/Bergenheim

etal.

(2012)

2009

Meanage60years

Proportionofmen

48%

MeanHbA1cNA


5.4

years

Saxagliptin(?

Met)vssulfonylurea

(?Met)

Cost-utility

analysis:

discreteevent


model)

Tim

e-horizon:5–40years

(patient

lifetime)


complications

Benefits:QALY

Discountrate:3%

costsandbenefits

Perspective:

3rd

party

payer

0.600


123

Ta

ble

1continued


(publicationyear)

Yearof

retrieval

ofdata

Population



andperspective

Score

[18]/Portugal/Carvalho

etal.(2014)

2014

Meanage53years

Proportionofmen

64.4

%

MeanHbA1c7.0

%


0years

Saxagliptin(?

Met)vssulfonylurea

(?Met)

Cost-utility

analysis:

discreteevent


model)

Tim

e-horizon:40-yeartime-horizon

(patientlifetime)

Directcost:medicines,hypoglycaem

ia

treatm

ent,diabetes


Benefits:QALY

andLYG

Discountrate:5%

costsandbenefits

Perspective:

societal

(productivityloss

included

for1styear)

0.829

[19]/Poland/Grzeszczak

etal.(2012)

2009

Meanage52years

Proportionofmen

48%

MeanHbA1c7.9

%


1.7

years

Saxagliptin(?

Met/sulfonylurea)

vsNPH

insulin(?

Met/sulfonylurea)

Cost-utility

analysis:

model

(long-term

Cardiffdiabetes

model)

Tim

e-horizon:40-year(patientlifetime)



ia

Benefits:QALY

Discountrate:5%

costsand3.5

%benefits

Perspective:

3rd

party

payer

0.686

[20]/Brazil/Nitaet

al.

(2012)

2011

Meanage59.77years

Proportionofmen

42%

MeanHbA1c6.47%


7.27years

Saxagliptin(?

Met)vsrosiglitazoneor

pioglitazone(?

Met)

Cost-utility

analysis:

discreteevent

simulationmodel

(UKPDS)

Tim


Directcost:hospitalizationandtreatm

entof

adverse

advents

Benefits:QALY

andLYG

Discountrate:5%

costs

Perspective:

privatehealthcare

system

0.571

[21]/Portugal/Viriato

etal.

(2014)

2013

Meanage63years

Proportionofmen

52%

MeanHbA1c7.2

%


9.13years

MeanBMI31.39kg/m

2

Vildagliptin(?

Met)vssulfonylurea

(?Met)

Cost-utility

analysis:

patientlevel

simulationmodel

(UKPDS)

Tim

e-horizon:40years

(patientlifetime)



ia

Benefits:QALY

andLYG

Discountrate:5%

costsandbenefits

(0–8%

SA)

Perspective:

3rd

party

payer

0.771

[22]/US/Liet

al.(2014)

2010-2012

Meanage54years

(liraglutide),58years

(sitagliptin)

Proportionofmen

43.9

%

(liraglutide),61.8

%

(sitagliptin)

Liraglutidevssitagliptin

Retrospectiveobservational

study

Tim

e-horizon:3years

Directcost:medicines,diabetes

related

complications

Benefits:%

pointsin

reductionofHbA1c

(liraglutide-0.95%,sitagliptin-0.63%,

Difference

0.31%)

Discountrate:N/A

Perspective:

3rd

party

payer

0.485

[23]/Sweden/CarlssonK,

PerssonU.(2014)

2013

Meanage56years

MeanHbA1c8.4

%


6years

MeanBMI32.6

kg/m

2

Liragutide1.2

mg(?

Met)vssitagliptin

100mg(?

Met)

Costutility

analysis:Markovmodel

(IHE

cohortmodel

oftype2

diabetes)

Tim

e-horizon:40years

(patientlifetime)

Directcosts:

medicines,diabetes

related


ia

Productionloss:dueto

hypoglycaem

iaand

diabetic

complications

Discountrate:3%

costsandbenefits

Perspective:

societal

0.800


123

Ta

ble

1continued


(publicationyear)

Yearof

retrieval

ofdata

Population



andperspective

Score

[24]/Greece/Tzanetakos

etal.(2014

)2013

Meanage64.5

years

Proportionofmen

51.5

%

MeanHbA1c8.2

%


10.4

years

MeanBMI30.4

kg/m

2

Liragutide1.2

mg(?

Met)vssitagliptin

100mg(?

Met)

Cost-utility

analysis:

Markov(CORE

diabetes

model)

Tim



related


ia

Benefits:QALY

Discountrate:3.5

%costsandbenefits

(0–6%

SA)

Perspective:

3rd

party

payer

0.828

[25]/US/Langer

etal.

(2013

)2012

Meanage55.3

years

Proportionofmen

52.9

%

MeanHbA1c8.4

%


6.0

years

MeanBMI32.8

kg/m

2

Liraglutide1.2

mgand1.8

mg(?

Met)vs

sitagliptin100mg(?

Met)

Cost

effectivenessanalysis

Tim

e-horizon:1year

Directcost:medicines

Benefits:%

patients

reachingthecomposite

endpoint(H

bA1c\7.0

%,nohypoglycaem

ia,

noweightgain)

Discountrate:0%

(1yeartime-horizon)

Perspective:

3rd

party

payer

0.657

[26]/Spain/Perez

etal.

(2015)

2012

Meanage55.3

years

MeanHbA1c8.4

%


6.0

years

MeanBMI32.8

kg/m

2

Liraglutide1.8

mg(?

Met)

vssitagliptin100mg(?

Met)

Cost-utility

analysis:

Markov(CORE

diabetes

model)

Tim


Directcosts:medicines,self-monitoringof

bloodglucose,diabetes


Benefits:QALY

Discountrate:3%

costsandbenefits

(0–5%

SA)

Perspective:

3rd

party

payer

0.685

[27]/Spain/Rayaet

al.

(2013)

2012

Meanage55.3

years

Proportionofmen

52.9

%

MeanHbA1c8.4

%


6.0

years

MeanBMI32.8

kg/m

2

Liraglutide1.2

mg(?

Met)


Met)

Cost-utility

analysis:

Markov(CORE

diabetes

model)

Tim


Directcost:medicines,self-monitoringofblood

glucose,diabetes


Benefits:QALY

Discountrate:3%

costsandbenefits

(0–5%

SA)

Perspective:

3rd

party

payer

0.714

[28]/UK/Davieset

al.

(2012)

2008

Meanage55.3

years

Proportionofmen

52.9

%

MeanHbA1c8.4

%


6.0

years

MeanBMI32.8

kg/m

2

Liraglutide1.2

mgand1.8

mg(?

Met)vs

sitagliptin100mg(?

Met)

Cost-utility

analysis:

Markov(CORE

diabetes

model)

Tim


Directcost:medicines,self-monitoringofblood

glucose,diabetes


Benefits:QALY

Discountrate:3.5

%costsandbenefits

(0–6%

SA)

Perspective:

3rd

party

payer

0.686

[29]/US/Lee

etal.(2012)

2011

Meanage55.3

years

Proportionofmen

52.9

%

MeanHbA1c8.4

%


6.0

years

MeanBMI32.8

kg/m

2

Liraglutide1.2

mgand1.8

mg(?

Met)


Met)

Cost-utility

analysis:

Markov(CORE

diabetes

model)

Tim

e-horizon:35years

Directcost:medicines,macro-and

microvascularcomplications,hypoglycaem

ia

Benefits:QALY

andLYG

Discountrate:3%

costsandbenefits

(0–6%

SA)

Perspective:

3rd

party

payer

0.743


123

study from 2007. There were four studies with a societal

perspective and sixteen studies with a 3rd-party payer

perspective. The societal perspective is the broadest and

most complex perspective, in which ideally the estimated

costs reflect the true social opportunity costs, while the 3rd-

party payer perspective reflects only the costs borne by that

payer. Costs measurements should be fully transparent and

the perspective adopted defines the scope of the analysis,

with different countries having different guidelines for this

issue [9].

One single study used an observational study providing

real-world effectiveness data. The majority of the studies

were based on one randomized controlled trial. Sitagliptin

was assessed in eleven studies, saxagliptin in seven studies,

vildagliptin in one study, and the three DPP-4 inhibitors

(sitagliptin, saxagliptin, and vildagliptin) were also asses-

sed jointly versus two GLP-1 receptor agonists (liraglutide

and exenatide) in one study.

All but one study evaluated these drugs as add-on to

metformin monotherapy as second-line, that is, for those

patients who did not achieve glycaemic control with met-

formin as first-line. One study did not specify clearly

whether the drug was assessed as second-line. Finally, the

majority of studies included in this review were of high

quality (average score 0.720 on a 0–1 scale, with a range

between 0.485 and 0.885).

The World Health Organization (WHO) threshold for

cost-effectiveness was used as Ref. [10]. This criterion

defined a strategy as cost-effective when the cost per

DALY averted or QALY gained was less than 3 times the

gross domestic product (GDP) per capita, and as very cost-

effective if it was less than the GDP per capita [11]. The

threshold was calculated for each study on the basis of the

year and country where the study was performed.

Main results of the selected publications

These findings are reported in Table 2, according to the

timing of the market introduction mentioned in the Meth-

ods section. More complete details of the studies, including

the characteristics of the base population, are shown in

Table A.1, in the Appendix.

First, we compared sitagliptin with sulfonylureas

[12, 13]. The studies used the Januvia Diabetes Economic

model (JADE). The JADE model is a discrete event sim-

ulation model developed to project the long-term impacts

of different interventions on diabetes related outcomes

[32]. It is an extension of the more well-known United

Kingdom Prospective Diabetes Study (UKPDS) outcomes

model, to which it adds the possibility of modelling the

effect of different treatment regimes in terms of health

outcomes, costs, and quality of life. The UKPDS appears as

less complete than the IMS core model—the most well-Ta

ble

1continued


(publicationyear)

Yearof

retrieval

ofdata

Population



andperspective

Score

[30]/Canada/Brownet

al.

(2014)

2012

Meanage54years

Proportionofmen

51%

MeanHbA1c8.5

%


4.5

years

MeanBMI31.1

kg/m

2

Insulinglargine(?

Met)vssitagliptin

(?Met)

Cost–utility

analysis:

Markov(CORE

diabetes

model)

Tim



related


ia

Benefits:QALY

Discountrate:5%

costsandbenefits

(0%

SA)

Perspective:

3rd

party

payer

0.800

[31]/Sweden/Kiadaliriet

al.

(2014)

2013

Meanage64.7

years

Proportionofmen

57.5

%

MeanHbA1c7.7

%


5.6

years

MeanBMI30.9

kg/m

2

GLP-1

agonists

(?Met)vsDPP-4

inhibitors

(?Met)vsNPH

insulin

(?Met)

Cost-utility

analysis:

Markovmodel

(IHECM-T2DM)

Tim

e-horizon:35years

(patientlifetime)

Directcosts:healthcare

costs

Productivitylosses,net

consumptionlosses

Benefits:QALY

Discountrate:3%

costsandbenefits

(0–5%

SA)

Perspective:

societal

0.885

N/A

notapplicable,Met

metform

in,SAsensitivityanalysis,JA

DEJanuvia

diabetes

economic

model,UKPDSUnited

Kingdom

prospectivediabetes

study,IH

ECM-T2DM

Institute

forHealth

Economicscohortmodel

forT2DM,BGSM

bloodglucose

self-m

onitoring


123

Ta

ble

2Resultsoftheselected

publications


(publicationyear)

Interventionvscomparator

Costs(€)a

Clinical

outcomes/QALYs

Increm

entalcost-effectiveness

ratio(ICER)

[12]/6Europeancountries

(Austria,

Finland,Portugal,

Spain,UK,Sweden)/Schwarz

etal.(2008)

Sitagliptin(?

Met)vs

rosiglitazoneorsulfonylurea

(?Met)

Sitagliptinvssulfonylurea

Difference

331€to

1097€

across

countries

Difference

0.037–0.095

QALYsacross

countries

Austria:

20,350€/QALY

Finland:13,737€/QALY

Portugal:5949€/QALY

Spain:13,440€/QALY

UK:11,547€/QALY

Sweden:12,219€/QALY

5949€–20,350€/QALYacross

countries

[13]/Portugal/Pereira

etal.

(2012)

Sitagliptin(?

Met)vs

sulfonylurea(?

Met)

Sitagliptin44,821€

Sulfonylurea44,283€

Difference

538€

Sitagliptin8.222QALYs

Sulfonylurea8.173QALYs

Difference

0.048QALYs

11,198€/QALY

[14]/Argentina/Elgartet

al.

(2013)

Saxagliptin(?

Met)vs

sulfonylurea(?

Met)

Saxagliptin10,883€

[US$12,327.7]

Sulfonylurea9441€

[US$10,694.8]

Difference

1441€

[US$1632.9]

Saxagliptin9.54QALYs


Difference

0.22QALYs

6510€[U

S$7374]/QALY

[15]/Germany/Erhardtet

al.

(2012)

Saxagliptin(?

Met)vs

sulfonylurea(?

Met)



Difference

1613€



Difference

0.12QALYs

13,931€/QALY

[16]/Sweden/Granstrom

etal.

(2012)

Saxagliptin(?

Met)vs

sulfonylurea(?

Met)

Saxagliptin12,328€[116,221

SEK]


[106,727SEK]

Difference

1006€[9,484SEK]



Difference

0.10QALYs

9608€[91,260SEK]/QALY

[17]/US/Bergenheim

etal.

(2012)

Saxagliptin(?

Met)vs

sulfonylurea(?

Met)


[US$65,139]


[US$62,367]

Difference

2447€[U

S$2772]



Difference

2.65QALYs

924€[U

S$1047]/QALY

[18]/Portugal/Carvalhoet

al.

(2014)

Saxagliptin(?

Met)vs

sulfonylurea(?

Met)



Difference

761€



Difference

0.14QALYs

5307€/QALY


123

Ta

ble

2continued


(publicationyear)


Costs(€)a

Clinical

outcomes/QALYs

Increm


ratio(ICER)

[19]/Poland/Grzeszczaket

al.

(2012)

Saxagliptin(?

Met/SU)vs

NPH

insulin(?

Met/SU)

Saxagliptin(?

Met)7765€

[PLN

31,394]

Insulin(?

Met)6858€[PLN

27,730]

Difference

906€[PLN

3663]

Saxagliptin(?

SU)7963€

[PLN

32,198]

Insulin(?

SU)7090€[PLN

28,668]

Difference

873€[PLN

3,529]

Saxagliptin?

Met

13.33

QALYs

Insulin?

Met

13.20QALYs

Difference

0.13QALYs

Saxagliptin?

SU

13.32

QALYs

Insulin?

SU

13.18QALYs

Difference

0.14QALYs

(?Met)6790€[PLN

27,454]/

QALY

(?SU)6100€[PLN

24,663]/

QALY

[20]/Brazil/Nitaet

al.(2012)

Saxagliptin(?

Met)vs

rosiglitazoneorpioglitazone

(?Met)

Saxagliptin9679€[R$33,023]

Pioglitazone10,850€

[R$37,019]

Difference

-1171€[R$3996]


Pioglitazone10.42QALYs

Difference

0.13QALYs

Dominant

[21]/Portugal/Viriato

etal.

(2014)

Vildagliptin(?

Met)vs

sulfonylurea(?

Met)

Vildagliptin14,409€


Difference

1161€

Vildagliptin5.7681QALYs


Difference

0.1279QALYs

9072€/QALY

[22]/US/Liet

al.(2014)


Liraglutide1403€[U

S$1589]

Sitagliptin1809€[U

S$2049]

Difference

-406€[U

S$-460]

Liraglutide-0.95%

Sitagliptin-0.63%

Difference

-0.31%

%points

inreductionof

HbA1c

NA

cost

per

patient

successfullytreatedto

the

target

composite

endpoint

[23]/Sweden/Carlssonand

Persson(2014)

Liraglutide1.2

mg(?

Met)vs

sitagliptin100mg(?

Met)

Liraglutide101,052€–

149,673€

[952,648–1411,014SEK]

Sitagliptin95,547€–

143,563€

[900,750–1353,411SEK]

Difference

5505€–

6110€

[51,898-57,603SEK]

Liraglutide8.55–10.53QALYs

Sitagliptin8.21–10.15QALYs

Difference

0.33–0.38QALYs

15,780€-17,060€[148,766-

160,827SEK]/QALY

[24]/Greece/Tzanetakoset

al.

(2014)

Liragutide1.2

mg(?

Met)vs

sitagliptin100mg(?

Met)

Liraglutide39,524€


Difference

2797€

Liraglutide9.24QALYs


Difference

0.19QALYs

15,101€/QALY


123

Ta

ble

2continued


(publicationyear)


Costs(€)a

Clinical

outcomes/QALYs

Increm


ratio(ICER)

[25]/US/Langer

etal.(2013)

Liraglutide1.2

and1.8

mg

(?Met)vssitagliptin100mg

Liraglutide1.2

mg9123€

[US$10,335]

Liraglutide1.8

mg10,377€

[US$11,755]


[US$16,858]

Difference

1.2

mg5759€

Difference

1.8

mg4505€

Liraglutide1.2

mg38.9

%

Liraglutide1.8

mg49.9

%

Sitagliptin100mg18.6

%

%ofpatientsreachingendpoint

NA,cost

per

patient

successfullytreatedto

the

target

composite

endpoint

[26]/Spain/Perez

etal.(2015)

Liraglutide1.8

mg(?

Met)vs

sitagliptin100mg(?

Met)



Difference

4177€



Difference

0.4

QALYs

10,436€/QALY

[27]/Spain/Rayaet

al.(2013)

Liraglutide1.2

mg(?

Met)vs

sitagliptin100mg(?

Met)



Difference

2297€



Difference

0.17QALYs

13,266€/QALY

[28]/UK/Davieset

al.(2012)

Liraglutide1.2

and1.8

mg


(?Met)

Liraglutide1.2

mg30,222€

[£21,793]

Liraglutide1.8

mg32,138€

[£23,175]

Sitagliptin27,667€[£19,951]

Difference

(1.2

mg)2554€

[£1842]

Difference

(1.8

mg)4471€

[£3224]

Liraglutide1.2

mg7.52

QALYs

Liraglutide1.8

mg7.64

QALYs


QALYs

Difference

(1.2

mg)0.19

QALYs

Difference

(1.8

mg)0.31

QALYs

(1.2

mg)13,661€[£9851]/

QALY

(1.8

mg)14,513€[£10,465]/

QALY

[29]/US/Lee

etal.(2012)

Liraglutide1.2

and1.8

mg


(?Met)

Liraglutide1.2

mg71,896€

[US$81,444]

Liraglutide1.8

mg78,127€

[US$88,502]

Sitagliptin66,438€[$76,262]

Difference

(1.2

mg)4575€

[US$5182]

Difference

(1.8

mg)11,689€

[US$13,241]

Liraglutide1.2

mg8.83

QALYs

Liraglutide1.8

mg8.98

QALYs


QALYs

Difference

(1.2

mg)0.20

QALYs

Difference

(1.8

mg)0.36

QALYs

(1.2

mg)22,724€

[US$25,742]/QALY

(1.8

mg)32,869€

[US$37,234]/QALY


123

Ta

ble

2continued


(publicationyear)


Costs(€)a

Clinical

outcomes/QALYs

Increm


ratio(ICER)

[30]/Canada/Brownet

al.

(2014)

Insulinglargine(?

Met)vs

sitagliptin(?

Met)

Insulinglargine33,409€

[$CAD

45,556]

Sitagliptin33,469€[$CAD

45,638]

Difference

-1052€[$CAD

-1434]

Insulinglargine8.815QALYs


Difference

0.076QALYs

Dominant

–13,847€[$CAD

-

18,882]/QALY

[31]/Sweden/Kiadaliriet

al.

(2014)

GLP-1

agonists

(?Met)vs

DPP-4

inhibitors

(?Met)and

DPP-4

inhibitors

(?Met)vs

NPH

insulin(?

Met)

GLP-1

agonists

229,430€

[2162,907SEK]

DPP-4

inhibitors

225,732€

[2128,042SEK]

NPH

Insulin225,102€

[2122,105SEK]

Difference

(GLP-1

vsDPP-4

inhibitors)3698€[34,865

SEK]

Difference

(DPP-4

inhibitors

vsNPH

Insulin630€[5936

SEK]

GLP-1

agonists

4.75QALYs

DPP-4

inhibitors

4.65QALYs

NPH

Insulin4.50QALYs

Difference

(GLP-1

vsDPP-4)

0.10QALYs

Difference

(DPP-4

vsNPH

Insulin)0.15QALYs

GLP-1

vsDPP-4:37,463€

[353,172SEK]/QALY

DPP-4

vsNPH

insulin

3824€[36,050SEK]/QALY

Met

metform

in,SU

sulfonylurea,

NAnotavailable,PSAprobabilisticsensitivityanalysis,TZD

thiazolidinedione

aExchangerateswereassessed

on15May

2015usingtheconvertingtoolofthePortugueseNational

Bank(https://www.bportugal.pt/en-U

S/Estatisticas/Dominios%

20Estatisticos/Estatisti

casC

ambiais/Pages/Taxasdereferenciadiarias.aspx)


123

https://www.bportugal.pt/en-US/Estatisticas/Dominios%20Estatisticos/EstatisticasCambiais/Pages/Taxasdereferenciadiarias.aspx


Ta

ble

3Detaileddescriptionsofstudyresults

References

Yearofretrieval

ofdata

Country/Authors

(pubyear)

Interventions

Perspective

Populationstudied

Sitagliptin(?

metform

in)vssulfonylurea(?

metform

in)

Schwarzet

al.

(2008)[12]

2007

6Europeancountries

(Austria,

Finland,

Portugal,Spain,Scotland-

UK,Sweden)

Schwarzet

al.(2008)

Sitagliptinvsrosiglitazoneor

sulfonylureaallas

add-onsto

metform

in

3rd

party

payer

in6European

countries

Meanagefrom

56.7

(Finnish

women)to

64.9

years

(Scottishmen

andwomen)

MeanHbA1cfrom

7.5

(Portugal

andFinland)to

8.09%

(Spanishmen)

BMIfrom

26.1

(Austrian

men)

to34.6

(Portuguesemen)

Pereira

etal.

(2012)[13]

2010

Portugal

Pereira

etal.(2012)

Sitagliptinvssulfonylureaall

asadd-onsto

metform

in

Societal

(butonly

withdirect

costs)

MeanHbA1c6.5–11%

Saxagliptin(?

metform

in)vssulfonylurea(?

metform

in)

Elgartet

al.

(2013)[14]

2009

Argentina

Elgartet

al.(2013)

Saxagliptinvssulfonylureaall

asadd-onsto

metform

in

(dose

notreferred)

3rd

party

payer

(Argentina

social

security

healthcare

system

)

Meanage64years

Proportionofmen

53%

MeanHbA1c7.7

%


10.5

years

Erhardtet

al.

(2012)[15]

2009

Germany

Erhardtet

al.(2012)


asadd-onsto

metform

in

(dose

notreferred)

3rd

party

payer

(National

sick

funds)

Meanage57.55years

Proportionofmen

52%

MeanHbA1c7.65%


5.4

years

Granstrom

etal.(2012)

[16]

2008

Sweden

Granstrom

etal.(2012)


asadd-onsto

metform

in

3rd

party

payer

inSweden

Meanage57.55years

Proportionofmen

52%

MeanHbA1c7.65%


5.4

years

Bergenheim

etal.(2012)

[17]

2009

US

Bergenheim

etal.(2012)


asadd-onsto

metform

in

(dose

notreferred)

3rd

party

payer

intheUS

Meanage60years

Proportionofmen

48%


5.4

years

Carvalhoet

al.

(2014)[18]

2014

Portugal

Carvalhoet

al.(2014)


asadd-onsto

metform

in

(dose

notreferred)

Societal

(Portuguese

perspective)

Meanage53years

Proportionofmen

64.4

%

MeanHbA1c7.0

%


0years


123

Ta

ble

3continued

References

Yearofretrieval

ofdata

Country/Authors

(pubyear)

Interventions

Perspective

Populationstudied

Saxagliptin(?

metform

in/sulfonylurea)

vsinsulin(?

metform

in/sulfonylurea)

Grzeszczak

etal.(2012)

[19]

2009

PolandGrzeszczaket

al.

(2012)

SaxagliptinvsNPH

insulin

when

usedin

combination

withmetform

inor

sulfonylurea

3rd

party

payer

(Polish

National

HealthFund)

Meanage52years

Proportionofmen

48%

MeanHbA1c7.9

%


1.7

years

Saxagliptin(?

metform

in)vspioglitazone(?

metform

in)

Nitaet

al.

(2011)[20]

2011

Brazil

Nitaet

al.(2012)

Saxagliptinvsrosiglitazoneor

pioglitazoneallas

add-onsto

metform

in(dose

notreferred)

Privatehealthcare

system

Meanage59.77years

Proportionofmen

42%

MeanHbA1c6.47%


7.27years

Vildagliptin(?

metform

in)vssulfonylurea(?

metform

in)

Viriato

etal.

(2014)[21]

2013

Portugal

Viriato

etal.(2014)

Vildagliptinvssulfonylureaall

asadd-onsto

metform

in

3rd

party

payer

(Portuguese

healthcare

system

perspective)

Meanage63years

Proportionofmen

52%

MeanHbA1c7.2

%


9.13years

MeanBMI31.39kg/m

2


Liet

al.(2014)

[22]

2010–2012

US

Liet

al.(2014)


3rd

party

payer

Meanage54years

(liraglutide),58years

(sitagliptin)

Proportionofmen

43.9

%

(liraglutide),61.8

%

(sitagliptin)

Liraglutide(?

metform

in)vssitagliptin(?

metform

in)

Carlssonand

Persson

(2014)[23]

2013

Sweden

CarlssonK,PerssonU.

(2014)

Liraglutide1.2

mgvs

sitagliptin100mgallas

add-

onsto

metform

in

Societal

Meanage56years

MeanHbA1c8.4

%


6years

MeanBMI32.6

kg/m

2


123

Ta

ble

3continued

References

Yearofretrieval

ofdata

Country/Authors

(pubyear)

Interventions

Perspective

Populationstudied

Tzanetakos

etal.(2014)

[24]

2013

Greece

Tzanetakoset

al.(2014)

Liraglutide1.2

mgvs

sitagliptin100mg

allas

add-onsto

metform

in

3rd

party

payer

Meanage64.5

years

Proportionofmen

51.5

%

MeanHbA1c8.2

%


10.4

years

MeanBMI

30.4

kg/m

2

Langer

etal.

(2013)[25]

2012

US

Langer

etal.(2013)

Liraglutide1.2

and1.8

mgvs


add-

onsto

metform

in

3rd

party

payer

Meanage55.3

years

Proportionofmen

52.9

%

MeanHbA1c8.4

%


6.0

yearsMeanBMI32.8

kg/

m2

Perez

etal.

(2015)[26]

2012

Spain

Perez

etal.(2015)

Liraglutide1.8

mgvs


add-

onsto

metform

in

3rd

party

payer

(Spanish

Healthcare

payer

perspective)

Meanage55.3

years

MeanHbA1c8.4

%


6.0

yearsMeanBMI32.8

kg/

m2

Rayaet

al.

(2013)[27]

2012

Spain

Rayaet

al.(2013)

Liraglutide1.2

mgvs


add-

onsto

metform

in

3rd

party

payer

(Spanish

Healthcare

payer

perspective)

Meanage55.3

years

Proportionofmen

52.9

%

MeanHbA1c8.4

%


6.0

years

MeanBMI32.8

kg/m

2

Davieset

al.

(2012)[28]

2008

UK

Davieset

al.(2012)

Liraglutide1.2

and1.8

mgvs


add-

onsto

metform

in

3rd

party

payer

(NHS

perspective)

Meanage55.3

years

Proportionofmen

52.9

%

MeanHbA1c8.4

%


6.0

years

MeanBMI32.8

kg/m

2

Lee

etal.

(2012)[29]

2011

US

Lee

etal.(2012)

Liraglutide1.2

and1.8

mgvs


add-

onsto

metform

in

3rd

party

payer

inUS

Meanage55.3

years

Proportionofmen

52.9

%

MeanHbA1c8.4

%


6.0

years

MeanBMI32.8

kg/m

2


123

Ta

ble

3continued

References

Yearofretrieval

ofdata

Country/Authors

(pubyear)

Interventions

Perspective

Populationstudied

Insulinglargineandsitagliptin

Brownet

al.(2014)

[30]

2012

Canada

Brownet

al.

(2014)

Insulinglarginevssitagliptin

allas

add-onsto

metform

in

3rd

party

payer

Meanage54years

Proportionofmen

51%

MeanHbA1c8.5

%


4.5

years

MeanBMI31.1

kg/m

2

GLP-1

agonists

(?metform

in)vsDPP-4

inhibitors

(?metform

in)vsNPH

insulin(?

metform

in)

Kiadalirietyal.(2014)

[31]

2013

Sweden

Kiadaliriet

al.

(2014)

GLP-1

agonists

vsDPP-4

inhibitors

vs

NPHinsulinallas

add-onsto

metform

in

Societal

Meanage64.7

years

Proportionofmen

57.5

%

MeanHbA1c7.7

%


5.6

years

MeanBMI30.9

kg/m

2

References

Datasourceto

Effectiveness

measurement

Outcomes/consequences

measurement/utilities

source/Baseutilities

reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Sitagliptin(?

metform

in)vssulfonylurea(?

metform

in)

Schwarzet

al.

(2008)[12]

Nauck

etal.

(2007)

QALY

UKPDS

0.78UKPDS

MeanQALYs:

Sitagliptinvssulfonylurea

Difference

0.037–0.095QALYsacross

countries

Meandirectcosts:sitagliptinvssulfonylurea

Difference

331–1097€across

countries

ICER(sitagliptinvssulfonylurea)

5949–20,350€/QALY

across

countries

Irrespectiveof20%

variationsin

thecost

andutility

weights

associated

withdiabetes-related

complications,discountedICER

values

remained

within

anarrow

range(4060–5473€),as

did

values

associated

with50%

variationsin

costsandutility

weights

associated

with

hypoglycemia

(5040–5256€)

0.714

Pereira

etal.

(2012)[13]

Nauck

etal.

(2006),

Goldsteinet

al.

(2007),

Charbonnel

etal.(2006)

QALY

UKPDS

0.785UKPDS

MeanQALYs:



Difference

0.048QALYs

Meandirectcosts:

Sitagliptin€4

4,821

Sulfonylurea€44,283

Difference

€538

ICER(sitagliptinvssulfonylurea)

11,198€/QALY

Globally

inoneway

sensitivity

analysisandmulti-way

sensitivityanalysistheICERs

arerobust

relativeto

the

differencesin

utilities

orin

the

costsofthedifferentparam

eters

aswellas

insomeparam

etersof

efficacy

0.629


123

Ta

ble

3continued

References

Datasourceto

Effectiveness

measurement




reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Saxagliptin(?

metform

in)vssulfonylurea(?

metform

in)

Elgartet

al.

(2013)[14]

RCT

D1680C00001-

52-w

eektrial

Gokeet

al.

(2010)

QALY

andLYG

Meandiscountedlife

expectancy:

Saxagliptin20.84years

Sulfonylurea20.76years

Difference

years

0.08years

MeandiscountedQALYs:Saxagliptin

9.54QALYs


Difference

0.22QALYs

Meandiscounteddirectcosts:

Saxagliptin10,883€[$12,327]

Sulfonylurea9441€[$10,694]

Difference

1441€[$1632.9]

ICER(saxagliptinvssulfonylurea)

6510€[$7374]/QALY

Cost-effectivenessacceptabilitycurve

illustratesaprobabilityofless

than

58%

that

saxagliptin?

metform

inis

cost-effectivecompared

with

sulfonylurea?

metform

in,considering

awillingnessto

pay

of6732€[$7626]/

(GDPper

capitaforArgentina)/QALY

0.829

Erhardtet

al.

(2012)[15]

RCT

D1680C00001

–52-w

eektrial

Gokeet

al.

(2010)

QALY

andLYG

UKPDS

Meanlife

expectancy:Saxagliptin

15.63years


Difference

years

0.01years

MeanQALYs:

Saxagliptin13.42

QALYs


Difference

0.12QALYs

Meandirectcosts:

Saxagliptin

38,163€


Difference

1613€

ICER(saxagliptinvssulfonylurea)

13,931€/QALY

Univariatedsensitivityanalysesshowthat

akey

driver

oftheresultswas

the

assumptionthat

patients

received

combinationtherapyandnotmetform

in

alonefrom

model

entry.ICER

of

saxagliptinvssulfonylureafellto

2372€/QALY

(an83.2

%reduction

from

basecase).In

ascenario

where

patients

enteredthemodel

atage

71.94years

[anincrease

of25%

over

thebasecase

(57.55years)],theICER

rose

by63.8

%to

23,175€/QALY.

Mean(H

bA1c)

level

atbaselinewas

also

akey

model

driver;values

both

higher

andlower

than

thebasecase

resulted

inhigher

ICERs:

HbA1c,

7.65%–ICER

14,147€/QALY

HbA1c,

7.15%–ICER17,840€/QALY

HbA1c,

8.15%–ICER15,155€/QALY

ICER10,329€in

thePSA

0.714


123

Ta

ble

3continued

References

Datasourceto

Effectiveness

measurement




reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Granstrom

etal.(2012)

[16]

RCT

D1680C00001-

52-w

eektrial

Gokeet

al.

(2010)

QALY

andLYG

UKPDS

Meandiscountedlife


14.72years


Difference

years

0.0

years

MeandiscountedQALYs:

Saxagliptin12.56

QALYs


Difference

0.10QALYs

Meandiscounteddirectcosts:Saxagliptin12,328€

[116,211SEK]

Sulfonylurea11,321€[106,727SEK]

Difference

1006€[9484SEK]

ICER

(saxagliptinvssulfonylurea)

9608€[91,260SEK]/QALY

Weight,anditsassociated

HRQoLdecrement

andim

pactondiabetes-related

events,was

anim

portantparam

eter

inthemodel.The

highestcost

per

QALY

29,320€[276,408

SEK]isobtained

under

theextrem

e

assumptionthat

theHRQoLdecrement(for

thefirstandsubsequentyears)per

unitBMI

gainis

reducedby75%

(0.0035).As

expected,when

thisHRQoLdecrementis

reducedbyhalf,theresultingcost

per

QALY

islower,17,652€[166,408SEK].

RaisingtheHbA1cthreshold

when

insulinis

initiatedto

8.0

%,increasesthecost

per

QALY

to18,156€[171,162SEK].The

overallcost

per

QALY

inthePSA

was:

11,578€[109,152SEK]

0.771

Bergenheim

etal.(2012)

[17]

RCT

D1680C00001-

52-w

eektrial

Gokeet

al.

(2010)

QALY

UKPDS

MeanQALYs:



Difference

2.65QALYs

Meandirectcosts:Saxagliptin57,503€[$65,139]

Sulfonylurea55,056€[$62,367]

Difference

2447€[$2772]

ICER


924€[$

1047]/QALY

When

evaluatingsensitivityaroundthe

subgroupofhypoglycaem

icevents

accruingcost,attributingacost

value

only

toevents

requiringmedical

assistance

did

notsignificantlychangetheoutcome.

PSAresultsshowed

ameanincrem

entalcost-

effectivenessratio(ICER)of(-€1748)

[-$1980]withsaxagliptinplusmetform

in

appearingto

bedominant.

0.600

Carvalhoet

al.

(2014)[18]

RCT

D1680C00001-

52-w

eektrial

Gokeet

al.

(2010)

QALY

andLYG

0884HealthSurvey

for

England2003

Meanlife

expectancy:Saxagliptin13.58years


Difference

years

0.01years

MeanQALYs:



Difference

0.14QALYs

Meancosts:Saxagliptin21,959€


Difference

761€

ICER


5307€/QALY

Inone-way

sensitivityanalysisthevalueof

HbA1cis

avariable

withaconsiderable

impactin

theresultsofthestudy.

InPSAtheprobabilityofbeingcost-effective

is84.5

%to

aWTPof20,000€per

QALY

and87%

forathreshold

of30,000€per

QALY

0.829


123

Ta

ble

3continued

References

Datasourceto

Effectiveness

measurement




reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Saxagliptin(?

metform

in/sulfonylurea)

vsinsulin(?

metform

in/sulfonylurea)

Grzeszczak

etal.(2012)

[19]

RCTJadzinsky

etal.(2009)

andRCT

Nauck

etal.

(2007)

QALY

UKPDS

Scenario

1(m

etform

in?

insulinvs

metform

in?

saxagliptin)

Meanlife

expectancy:

Saxagliptin?

metform

in22.58years

Insulin?

metform

in22.58years

Difference

years

0.00years

MeanQALYs:

Saxagliptin?

metform

in

13.33QALYs

Insulin?

metform

in13.20QALYs

Difference

0.13QALYs

Meandirectcosts:

Saxagliptin?

metform

in

7765€[PLN

31,394]

Insulin?

metform

in6858€[PLN

27,730]

Difference

906€[PLN

3663]

ICERsaxagliptinvsinsulin(?

metform

in)

6790€[PLN

27,454]/QALY

Scenario

2(SU

?saxagliptinvs

SU

?insulin)

Meanlife

expectancy:Saxagliptin?

SU

22.53years

Insulin?

SU

22.53years

Difference

years

0.00years

MeanQALYs:

Saxagliptin?

SU

13.32

QALYs

Insulin?

SU

13.18QALYs

Difference

0.14QALYs

Meandirectcosts:

Saxagliptin?

SU

7963€[PLN

32,198]

Insulin?

SU

7090€[PLN

28,668]

Difference

873€[PLN

3529]

ICERsaxagliptinvsinsulin(?

SU)6100€

[PLN

24,663]/QALY

Theresultswerefoundto

be

sensitiveto

someofthebasic

model

assumptions,althoughthe

ICERremained

below

12,366€

[PLN

50,000]per

QALY

gained

inallcases

0.686


123

Ta

ble

3continued

References

Datasourceto

Effectiveness

measurement




reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Saxagliptin(?

metform

in)vspioglitazone(?

metform

in)

Nitaet

al.

(2011)[20]

DIA

PS79StudyGroup

(2010)

QALY

andLYG

UKPDS

0885UKPDS

Meanlife


12.17years

Pioglitazone12.16years

Difference

years

0.01

MeanQALYs:

Saxagliptin10.55

QALYs

Pioglitazone10.42QALYs

Difference

0.13QALYs

Meandirectcosts:Saxagliptin

9679€[R$33,023]

Pioglitazone10,850€[R$37,019]

Difference

-1171€[R$-3996]

ICER

(saxagliptinvs

pioglitazone)—

Dominant

Intheunivariate

sensitivity

analysis,saxagliptinremained

dominantcompared

withTZDs

afteravariationof±15%

onall

selected

param

eters.

InPSA,addingsaxagliptinto

the

metform

intherapywas

dominant

in62.1

%ofallscenariosversus

theadditionofpioglitazone.

Only

in2.2

%ofthesimulations

did

saxagliptinshow

less

effectivenessandhigher

costs

0.571

Vildagliptin(?

metform

in)vssulfonylurea(?

metform

in)

Viriato

etal.

(2014)[21]

Ferranniniet

al.(2009)

RCT

QALY

andLYG

UKPDS

0.78Clarkeet

al.(2002)

Meanlife

expectancy:Vildagliptin

7.7486years


Difference

years

0.0896years

MeanQALYs:Vildagliptin5.7681

QALYs


Difference

0.1279QALYs

Meandirectcosts:Vildagliptin

14,409€


Difference

1161€

ICER

(vildagliptinvs

sulfonylurea)

9072€/QALY

Univariate

analysesshowed

that

ICERvalues

wererobustand

ranged

from

4195€to

16,052€

per

QALY

when

different

param

eterswerevaried.

ThePSA

of100simulated

interactionssuggestedthat

fora

WTPof30,000€per

QALY

treatm

entwithmetform

inplus

vildagliptinhad

a79%

probabilityofbeingcost-

effectivecompared

with

metform

inplussulfonylurea

0.771


123

Ta

ble

3continued

References

Datasourceto

Effectiveness

measurement




reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification


Liet

al.(2014)

[22]

Observational

longitudinal

retrospectivestudy

HbA1cchange

%points

Meanlife

expectancy:N/A

(6monthsfollow-up,without

extrapolation)

%pointsin

reductionofHbA1c

Liraglutide-0.95%

Sitagliptin-0.63%

Difference

0.31%

Liraglutide1403€[$1589]

Sitagliptin1809€[$2049]

Difference

-406€[$-460]

Liraglutideprovides

greater

benefits

atlower

cost,when

compared

withsitagliptin

(dominance)

Notapplicable

0.485

Liraglutide(?

metform

in)vssitagliptin(?

metform

in)

Carlssonand

Persson

(2014)[23]

RCTPratley

etal.(2010)

(ClinicalTrials.gov

Identifier:NCT00700817)

HbA1cchange

%points

UKPDS

Men,nonsm

oker

MeanQALYs:

Liraglutide

8.55–10.53QALYs

Sitagliptin8.21–10.15QALYs

Difference

0.33–0.38QALYs

Meandirectcosts:Liraglutide

101,052–149,673€

[952,648–1411,014SEK]

Sitagliptin95,547–143,563€

[900,750–1353,411SEK]

Difference

5505-6110€

[51,898–57,603SEK]

ICER

(liraglutidevssitagliptin)

15,780–17,060€

[148,766–160,827SEK]/QALY

Ninetyper

centofthepredicted

cost

increm

ents

andQALY

increm

entsforthecomparisonof

liraglutide1.2

mgvsSU

were

within

theintervals2408€

22,701.SEK–14,175€133,633

SEK

and0.15–1.27QALYs,

respectively.Theprobabilitythat

liraglutidewould

beconsidered

cost-effectivecompared

with

sitagliptinwas

89%

ataWTP

per

QALY

of53,037€500,000

SEK

0.800


123

Ta

ble

3continued

References

Datasourceto

Effectivenessmeasurement




reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Tzanetakos

etal.(2014)

[24]

QALY

UKPDS

Meanlife

expectancy:Liraglutide

14.22years

Sitagliptin14.09years

Difference

0.13years

MeanQALYs:

Liraglutide9.24

QALYs


Difference

0.19QALYs

Meandirectcosts:Liraglutide

39,524€


Difference

2797€

ICER


15,101€/QALY

Sim

ulationresultswerequite

sensitiveto

thegradual

shorteningofmodel

time-

horizonresultingin

anincrease

ofbasecase

ICERforliraglutide

bymore

than

600%

at5years

simulation.Sim

ulationresults

werequitesensitiveto

patients’

HbA1cvalues,underlining,as

such,theim

portance

ofthis

biochem

ical

param

eter

tohealth

andcost

outcomes

ofmodel

analysis

0.828

Langer

etal.

(2013)[25]

RCTPratley

etal.(2010)

(ClinicalTrials.gov


HbA1c

%ofpatientsreachingendpoint

Liraglutide1.2

mg38.9%

Liraglutide1.8

mg49.9%

Sitagliptin100mg18.6%

Difference

(liraglutide1.2

mg)

(20.3%)

Difference

(liraglutide1.8

mg)

(31.3%)

Meandirectcosts(SD):

Liraglutide1.2

mg9123€

[US$10,335]

Liraglutide1.8

mg10,377€

[US$11,755]

Sitagliptin14,882€[U

S$16,858]

Variationin

cost

assumptionsby

±20%

andvariationin

clinical

inputsparam

eterswithin

the

rangeofthe95%

CIdid

not

changethefindingsthat

thecost

ofreachingthecomposite

endpoint(cost

ofcontrol)with

liraglutide1.2

mgand1.8

mg

was

lower

than

thecost

with

sitagliptinafter52weeksof

treatm

ent

0.657


123

Ta

ble

3continued

References

Datasourceto





reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Perez

etal.

(2015)[26]

RCTPratley

etal.(2010)

(ClinicalTrials.gov


QALYs

Meanundiscountedlife

expectancy

(SD):Liraglutide

14.241years

(0.183)


(0.185)

Difference

0.368years

MeanQALY

(SD):

Liraglutide9.24QALYs(0.121)

Sitagliptin8.84QALYs(0.121)

Difference

0.40QALYs

Meandiscounteddirectcosts(SD):

Liraglutide56,628€(1323)

Sitagliptin52,450€(1394)

Difference

4177€

ICER

of10,436€/QALY

Cost

effectivenessoutcomes

were

most

sensitiveto

changes

in

time-horizonofthemodelling

analysis,withliraglutideless

cost-effectiveover

shorter

time-

horizons.Thiswas

primarilydue

totheim

provem

ents

in

physiological

param

eters

associated

withliraglutide

resultingin

reducedrisk

oflong-

term

complications,withthe

benefitsofthisnotfullyrealized

over

shorter

time-horizons

0.685

Rayaet

al.

(2013)[27]

RCTPratley

etal.(2010)

(ClinicalTrials.gov


QALY

Meanundiscountedlife

expectancy

(SD):Liraglutide

20.00years

(0.33)


(0.30)

Difference

0.28years

MeanQALY

(SD):

Liraglutide9.04QALYs(0.13)

Sitagliptin8.87QALYs(0.11)

Difference

0.17QALYs

Meandiscounteddirectcosts(SD):

Liraglutide54,684€(1250)

Sitagliptin52,387€(1346)

Difference

2297€

ICER

of13,266€/QALY

Cost

effectivenessoutcomes

were

most

sensitiveto

changes

in

shorteningthetime-horizon

(5–10years)andin

changes

of

HbA1cbenefitassociated

with

liraglutide.

When

thetime-

horizonwas

changed

to10years,

theICER

increasedto

58,433€/

QALY

andwhen

changed

to

5years

theICER

increasedto

102,605€/QALY.Withthe

abolishmentoftheHbA1c

benefittheICERincreasedto

199,114€/QALY

0.714


123

Ta

ble

3continued

References

Datasourceto





reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Davieset

al.

(2012)[28]

RCTPratley

etal.(2010)

(ClinicalTrials.gov


QALY

MeanQALYs(SD):Liraglutide1.2

mg7.52

QALYs(0.11)

Liraglutide1.8

mg7.64QALYs(0.11)

Sitagliptin100mg7.34QALYs(0.11)

Difference

(1.2

mg)0.19QALYs(0.15)

Difference

(1.8

mg)0.31QALYs(0.15)

Meandirectcosts:

Liraglutide1.2

mg

30,222€[£21,793]

Liraglutide1.8

mg32,138€[£23,175]

Sitagliptin27,667€[£19,951]

Difference

(1.2

mg)2554€£1842]

Difference

(1.8

mg)4471€[£3224]

ICER


1.2

mg13,661€[£9851]/QALY

1.8

mg14,513€[£10,465]/QALY

Variationsin

key

param

etersare

allcost-effectiveat

athreshold

of£20000.Thegainin

QALYswithliraglutide1.2

mg

over

sitagliptinarises

mainly

from

improvem

ents

inHbA1c

(54%)andweight(44%)

0.686

Lee

etal.

(2012)[29]

RCTPratley

etal.(2010)

(ClinicalTrials.gov


QALY

andLYG

Meanlife

expectancy:Liraglutide1.2

mg

13.003years

Liraglutide1.8

mg13.189years


Difference

(1.2

mg)0.163years

Difference

(1.8

mg)0.348years

MeanQALYs:

Liraglutide1.2

mg8.825QALYs

Liraglutide1.8

mg8.979QALYs

Sitagliptin100mg8.624QALYs

Difference

(1.2

mg)0.201QALYs

Difference

(1.8

mg)0.356QALYs

Meandirectcosts:

Liraglutide1.2

mg

1,896€[$81.444]

Liraglutide1.8

mg79,010€[$89.502]

Sitagliptin7322€[$76.262]

Difference

(1.2

mg)4575€[$5182]

Difference

(1.8

mg)11,689€[$13,241]

ICER


1.2

mg22,724€[$25,742]/QALY

1.8

mg32,869€[$37,234]/QALY

Aseries

ofsensitivityanalyses

indicated

that

liraglutidewould

becost-effectiveat

both

1.8-

and1.2-m

gdosages

compared

withsitagliptinover

arangeof

plausible

inputparam

eters.

Cost-effectivenessresults(for

liraglutide1.2

and1.8

mg)

weremost

sensitiveto

the

time-horizonofthe

simulations;when

a10-year

time-horizonwas

usedthe

ICERwas

100,820€

[$114,209]/QALY

for

liraglutide1.8

mgand58,939

€[$66,766]/QALY

for

liraglutide1.2

mg.When

PSA

was

conducted:ICER

for

liraglutide1.8

mgvssitagliptin

was

37,856€[$42,883]/QALY

andforliraglutide1.2

mgvs

sitagliptinwas

31,238

€[$35,386]/QALY

0.743


123

Ta

ble

3continued

References

Datasourceto





reference

Totalandincrem

entalanalysis

Sensitivityanalysis

Classification

Insulinglargineandsitagliptin

Brownet

al.

(2014)[30]

EASIE

RCT

QALY

UKPDS

0.80Cam

eronandBennett,

2009

Meanlife

expectancy:Insulinglargine11.47years


Difference

years

0.09years

MeanQALYs:

Insulinglargine8.815QALYs


Difference

0.076QALYs

Meandirectcosts:Insulinglargine33,409€

[$CAD45,556]

Sitagliptin33,469€[$CAD45,638]

Difference

-1052€[$CAD1,434]

ICER(insulinglarginevssitagliptin)-13,847€

[$CAD

-18,882]/QALY

Dominant

Changes

inthenumber

ofteststrips

usedforsitagliptinandinsulin

glargineresulted

inthelargest

impactontheICER.Depending

onthenumber

ofteststripsused,

theICERvariedfrom

13,692€to

28,667€

[$CAD18,670–$CAD39,091]

0.800

GLP-1

agonists

(?metform

in)vsDPP-4

inhibitors

(?metform

in)vsNPH

insulin(?

metform

in)

Kiadalirietyal.

(2014)[31]

Ekstrom

etal.(2012)

Observational

study

QALY

EQ-5D

UKPDS

SwedishNational

Diabetes

Register(N

DR)

MeanQALYs:

GLP-1

agonists

4.75QALYs

DPP-4

inhibitors

4.65QALYs

Difference

(GLP-1

vsDPP-4)0,10QALYs

NPH

insulin4.50QALYs

Difference

(DPP-4

vsNPH

insulin)0.15QALYs

Meandirectcosts:GLP-1

agonists

229,430€[SEK

2,162.907]

DPP-4inhibitors

225,732€[SEK

2,128.042]

NPH

insulin225,102€[SEK

2,122.105]

Difference

(GLP-1

vsDPP-4)3,698€[34,865SEK]

Difference

(DPP-4

vsNPH

insulin)

630€[5,936SEK]

ICER(G

LP-1

vsDPP-4):

37,463€[353,172SEK]/QALY

ICER(D

PP-4

vsNPH

insulin)

3,824€[36,050SEK]/QALY

InPSA,theestimated

ICERwas

319,217.AssumingaWTPof

500,000SEK

per

QALY

gained,

strategy1had

a74.7

%

likelihoodofbeingconsidered

cost-effectivein

comparisonto

strategy2

0.885

Total

0.720

Exchangerateswereassessed

on15May

2015usingtheconvertingtoolofthePortugueseNational

Bank(https://www.bportugal.pt/en-U

S/Estatisticas/Dominios%

20Estatisticos/Estatisti

casC

ambiais/Pages/Taxasdereferenciadiarias.aspx)


123



known and widely used model, presented below—because

it includes fewer complications and does not encompass

the interactions between them, and also because the disease

progression is modelled in a limited way for some com-

plications; although, it has also been shown to perform

quite similarly in simulating trial outcomes [33]. The main

contribution was an international comparison including six

countries, which allowed determining the possible dis-

crepancies in cost-effectiveness ratios.

Sitagliptin was beneficial as compared with sulfony-

lureas by the reduction of hypoglycemia episodes and

avoidance of weight gain [4]. The ICER of sitagliptin

versus the sulfonylureas varied from 5949 €/QALY in

Portugal to 20,350 €/QALY in Austria [12]. Discrepancies

in incremental effectiveness may contribute to these dif-

ferences, from 0.037 QALYs in Austria to 0.056 QALYs in

Portugal. The Portuguese population considered in this

study had a larger body mass index (BMI) (34.77 kg/m2,

for men) as compared with Austria (26.12 kg/m2, for men),

possibly explaining the higher benefits of sitagliptin. There

were also discrepancies in incremental costs, from 331 € in

Portugal to 760 € in Austria. This was partly explained by

the difference in price between sitagliptin and sulfony-

lureas, from 1.55 € in Portugal to 1.78 € in Austria. The

higher costs of macro- and micro-vascular complications in

Austria might have been compensated for by their lower

probability of occurrence, given the lower BMI and zero-

percent smoking rate in the sample. Note that a more recent

study for Portugal, which replicated the same analysis,

found a value of 11,198 €/QALY [13]. The higher ICER,

compared with the value obtained for Portugal in the

international comparison, is explained by the higher treat-

ment costs.

Second, we analysed the cost-effectiveness of sax-

agliptin compared with sulfonylureas [14–18], mainly as

add-ons to metformin, in patients not controlled with

metformin monotherapy. The studies used a stochastic

simulation model especially designed to evaluate the

impact of new therapies in people with T2DM (Cardiff

Diabetes model) [34–36]. Although possibly less compre-

hensive than the IMS core model, this model performed

comparably when compared with trials outcomes [34]. It

provided a reasonable prediction of treatment effects but

tended to underestimate or overestimate the risks of

complications.

The major advantages of saxagliptin as compared with

sulfonylurea were the reduction in the number of hypo-

glycemia episodes, which influences the quality of life, and

the avoidance of weight increases [16]. The incremental

effectiveness was rather similar across studies, with values

of 0.10 QALYs in Sweden [16], 0.12 in Germany [15], and

0.14 in Portugal [18]. A higher value was observed for

Argentina [14], of 0.22 QALYs, possibly due to the older

base population (64 years old, compared with 60 or

younger in the other studies), and longer duration of dia-

betes (more than 10 years, compared with less than 6 in

other studies).

The US study was a clear outlier, indicating an incre-

mental effectiveness of 2.65 QALYs [17]. This very

favourable effect of saxagliptin in the US study was related

to the much lower QALYs for the patients treated with

sulfonylureas, of 8.37, compared with values above 9.32 in

the other studies. We interpret this discrepancy through the

base characteristics of the population considered in the US

study, which had a much higher average BMI (34.01 kg/m2

versus lower than 31.20 kg/m2 in the other studies, from

our own calculations), explaining the largely unfavourable

outcomes for the patients treated with sulfonylureas.

The incremental costs varied little across studies, from

761 € in Portugal [18] to 2,447 € in the US [17]. The ICER

values ranged from 5,307 €/QALY to 13,931 €/QALY.The US case was an outlier in terms of ICER, as expected

from the incremental effectiveness outcome, with a much

lower value of 924 €/QALY.Third, we observed that saxagliptin in combination with

metformin had an ICER of 6790 €/QALY when compared

with insulin; the ICER was 6100 €/QALY in combination

with sulfonylurea [19]. The benefits were driven by the

lower risk of hypoglycaemic events and the neutral effect

on weight [19]. A quite similar ratio was obtained in

Sweden, by Kiadaliri et al. [31], which compared the DPP-

4 inhibitors to insulin, finding an ICER of 3824 €/QALY.Fourth, a single study compared saxagliptin with

pioglitazone, in combination with metformin, and found

that saxagliptin was dominant (greater effectiveness and

lower cost) [20]. The price of saxagliptin is lower, while

the benefits are greater, namely the higher glycaemic

control.

Fifth, one single study was obtained for vildagliptin,

compared with sulfonylureas [21], all as add-ons to met-

formin, in patients not controlled with metformin

monotherapy. The vildagliptin versus sulfonylurea model

was constructed as a patient-level simulation model, uti-

lizing the risk equations from the UKPDS outcomes model

to predict microvascular and macrovascular complications

and mortality (both disease-specific and all-cause) over a

40-year horizon [21]. Vildagliptin had a similar clinical

benefit in comparison to sulfonylurea; namely it avoids the

weight gain and reduces the risk of hypoglycemia [4].

Vildagliptin had an ICER of 9072 €/QALY compared with

sulfonylureas, using the Portuguese National Health Ser-

vice perspective.

Sixth, we examined the comparison between liraglutide

and sitagliptin [22–29]. The studies that compared liraglu-

tide with sitagliptin, all as add-ons to metformin as second-

line therapy, used the CORE diabetes model (IMS Health,


123

Basel, Switzerland). This model, widely used and validated,

allows estimating the long-term health outcomes and eco-

nomic consequences of diabetes, with the possibility also of

measuring the impact of different treatments and patient

management strategies [37]. It has become a reference in the

field because it integrates a comprehensive set of compli-

cations, and accounts for multiple risk factors. Also, it

combines aMarkov structure with Monte Carlo simulations,

which allows measuring the joint progression and interac-

tions between complications, and the effect of event history.

On the one hand, most of the studies were based on the

same measurements of benefits, so that the incremental

effectiveness was quite similar across studies. In particular,

six out of eight studies were based on the same random

clinical trial, by Pratley et al. [38]. According to Li et al.

[22], using effectiveness measures (i.e., from the real

world), the benefits of liraglutide compared with sitagliptin

are driven by improved HbA1c and by greater weight loss.

Noticeably, all studies considered a population with a BMI

higher than 30 kg/m2, so that the results must be inter-

preted as relevant for this specific sub-population. For the

1.2-mg dosage, the incremental effectiveness varied from

0.17 QALYs in Spain [27] to 0.20 in the US [29]; for the

1.8-mg dosage, it varied from 0.31 QALYs in the UK [28]

to 0.40 in Spain [26].

On the other hand, the variations in incremental costs

ranged between 2297 € in Spain [27] to 6110 € in Sweden

[23], for the 1.2-mg dosage. This resulted in ICERs of

relatively similar values, from 13,266 €/QALY in Spain

[27] to 22,724 €/QALY in the US [29]. For the 1.8-mg

dosage, the values varied between 4000 € and 5000 €except for the US, where the incremental cost rose to

11,689 € [29]. This resulted in a much higher ICER in the

US, of 32,869 €/QALY, compared with values below

15,000 €/QALY in other countries. This much higher

incremental cost in the US, driving a much less favourable

ICER, was mainly due to the discrepancy in treatment

costs: in the US, the liraglutide treatment is twice as costly

as the sitagliptin, while it is only 16 % higher in the UK

(according to our own calculation) [28]. Note also that the

US study adopted a time-horizon of 3 years only, possibly

omitting potential long-term consequences of treatments,

so that its findings are hardly comparable.

Seventh, insulin glargine was compared with sitagliptin

as add-ons to metformin [30], and was found to be domi-

nant in Canadian settings. The model also used the IMS

CORE diabetes model [37]. The model considered a time-

horizon of 50 years, and was populated with data from the

EASIE trial. The insulin glargine increased the number of

hypoglycaemic episodes but achieved a greater reduction

of HbA1c [30]. This resulted in a slight benefit, of 0.08

QALYs, obtained with a lower treatment cost, due to the

lower costs of treatment and disease complications.

Eighth, the GLP-1 receptor agonists (exenatide and

liraglutide) were compared with DPP-4 inhibitors as add-

ons to metformin [31]. The authors used the IHECM-

T2DM model, which was previously described and used to

compare cost-effectiveness of one of the GLP-1 receptor

agonists, liraglutide, versus sulfonylureas or sitagliptin in

Sweden. This model includes yearly cycles and a time-

horizon of up to 40 years. The macrovascular health states

were based on the UKPDS models and on the Swedish

national diabetes register equations [31]. The authors cal-

culated a gain of 0.10 QALYs for GLP-1 receptor agonists,

driven by the better control of glycemia and weight

reductions. These benefits were achieved at an additional

cost of 3698 €, resulting in an ICER of 37,463 €/QALY in

Swedish settings.

Discussion

The DPP-4 inhibitors represent substantial benefits in the

treatment of type-2 diabetes, but also a considerable chal-

lenge for health systems due to their high prices. There is

thus a clear need to perform economic evaluations of these

new therapies, to guide policy-makers in their decisions

regarding co-payments, therapeutic guidelines, and reim-

bursement strategies.

Key findings

This systematic review of the literature shows first and

foremost that the evidence on cost-effectiveness is still

recent and scarce, but is based on studies of overall high

quality. In particular, most of the studies were based on

published clinical trials for the measurement of conse-

quences, and on official unit prices and country-specific

resource use for costs. The studies used different but val-

idated models, with comparable characteristics and pre-

dictive power, which all considered long time-horizons and

the various health consequences of diabetes.

According to our findings, the DPP-4 inhibitors

appear as highly cost-effective in most of the cases,

following the WHO criterion and using the Organisation

for Economic Co-operation and Development (OECD)

GDP per capita values for 2015 as Ref. [39]. All as add-

ons to metformin, sitagliptin, saxagliptin, and vilda-

gliptin had an ICER below 25,000 €/QALY when

compared with sulfonylureas. When compared with

NPH insulin, saxagliptin as add-on to metformin or

sulfonylurea had an ICER below 10,000 €/QALY.Saxagliptin was dominant as compared with pioglita-

zone. These ratios were driven mainly by the reduction

in the risk of hypoglycaemic events, and in the avoid-

ance of weight gain.


123

Some variations across studies are worth mentioning. In

particular, the ratios for sitagliptin and saxagliptin varied

due to differences in baseline patient characteristics. The

findings were more favourable, from an economic view-

point, for the sub-group of patients with a higher BMI, who

were older, and who had been suffering from long-standing

diabetes. By contrast, in most cases there was little varia-

tion in incremental costs, so that these may not be con-

sidered as a major cause of ICER discrepancies.

Some studies evaluated more recent alternative thera-

pies, namely the GLP-1 receptor agonists and insulin

glargine. When compared with sitagliptin, liraglutide was

highly cost-effective. Note that this result was obtained

only for populations with an average BMI higher than

30 kg/m2; the favourable results are mainly explained by

the reduction in weight and better control of glycemia

achieved by GLP-1 receptor agonists. The insulin glargine

was demonstrated to be dominant when compared with

sitagliptin, related to the higher decrease in HbA1c and the

lower price of the drug.

Comparison with earlier literature

To the best of our knowledge, two earlier systematic reviews

of economic evaluations of DPP-4 inhibitors have been

performed, in 2010 [40] and 2015 [41]. Given the rapid

introduction of new drugs, we considered only the last

review as a relevant comparison for our study. The major

difference is that we retrieved 20 studies, in comparison to 11

in the reviewbyGeng et al. [41],mainly due to the analysis of

recently introduced GLP-1 receptor agonists.

Our findings were quite similar in regard to the com-

parison between DPP-4 inhibitors and sulfonylureas, both

as add-ons to metformin, concluding that DPP-4 inhibitors

are cost-effective in patients who do not achieve glycaemic

targets with metformin. In comparison with thiazolidine-

diones, Geng et al. [41] found five papers that came to

uncertain conclusions; in comparison, we only retrieved

one single paper, because we did not consider comparisons

with rosiglitazone, whose marketing authorization was

suspended in Europe due to safety issues.

Limitations

This study has some limitations. First, a relatively small

number of articles met the inclusion criteria for some

comparisons, namely DPP-4 inhibitors versus thiazo-

lidinediones (one single study), and DPP-4 inhibitors ver-

sus insulin (two studies). Obviously, this represents too

little evidence to draw robust conclusions. Regarding thi-

azolidinediones, this does not represent a major problem

given their low utilization in current practice [42]. For

insulin, this is a more serious issue given the high number

of users; in this sense, the lack of economic evaluations

was quite surprising, and further research is needed to

elaborate solid recommendations.

Second, as mentioned by Waugh et al. [40], most studies

were funded by manufacturers. Even though these studies

were often carried out by independent consultants, this can

lead to a publication bias, the unfavourable findings not

being diffused [43]. More generally, these limitations

highlight the urgent need for independent economic eval-

uations of DPP-4 inhibitors and other new treatments, e.g.

by State agencies.

Third, the health consequences were measured in most

studies using the same randomized control trials. For

example, the D1680C00001 trial was used in the studies for

Argentina, Germany, Portugal Sweden, and the US. On the

one hand, observational studies should be carried out to

confirm the favourable outcomes in real-world practice,

which may produce different outcomes according to pop-

ulation characteristics, patients’ attitudes, and physicians’

practices. There is evidence, for example, of the use in

actual practice of DPP-4 inhibitors as single compounds or

in combination with sulfonylureas [44]. On the other hand,

the reliance on a few trials is problematic because it may

not encompass the possible different outcomes according

to populations’ characteristics. In particular, the ADA/

EASD guidelines mention that the choice of the second-

lines should be based on ‘‘a variety of patient and disease

specific factors’’ (p. 145) [4]. The lack of variety in the

populations under scrutiny limits the contributions of the

studies for drug evaluation across sub-populations.

Finally, relevant databases such as Econlit, CINAHL,

EMBASE, and Psychoinfo were not included in the search

strategy, because they were not accessible for free. Note,

however, that we have used the most commonly used

databases in systematic reviews of economic evaluations.

Future cost-effectiveness studies should be repeated during

the lifecycle of medicines under real-world utilization and

incorporate budget impact analysis.

Conclusions

According to WHO thresholds, there is consistent evidence

about the cost-effectiveness of DDP-4 inhibitors as second-

line, as add-ons to metformin, in comparison with sul-

fonylureas. More recent therapies, namely the GLP-1

receptor agonists, were, however, demonstrated to be

highly cost-effective in comparison to DPP-4 amongst

populations with an average BMI higher than 30 kg/m2.

Also, insulin glargine was demonstrated to be dominant

when compared with DPP-4 inhibitors.

These results were, however, obtained on the basis of

a limited number of studies, relying on the same few


123

clinical trials, and financed by manufacturers, raising the

issue of a possible publication bias. Further independent

research is needed, with more experimental but also

observational studies, to confirm these findings and

incorporate them into therapeutic guidelines.

Acknowledgments This research had no external funding.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict of

interest. The authors have no affiliations or financial involvement that

conflicts with the material presented in this paper. The authors are not

employed in the pharmaceutical industry and, at the time of this

research, had not performed nor were conducting any research study

pertaining to the DPP-4 inhibitors group, nor were conducting any

research for any of the marketing authorization holders of those

therapeutic agents.

Appendix

See Table 3.

Template for each study extracted and recorded

information:

• Title [incl. author(s) and journal]

• Type of economic evaluation

• Objective

• Interventions

• Comparators

• Methods (Analytical approach; Perspective; Time

Horizon; Population; Effectiveness data; Monetary

benefit and utility valuation; Measure of benefit; Cost

data; Discount rate; Analysis of uncertainty)

• Results

• Authors’ Conclusions

• Limitations

• Affiliation to pharmaceutical industry

Search strings for search:

• Sitagliptin OR Saxagliptin OR Vildagliptin OR Alo-

gliptin OR Linagliptin assessed in March 2015

• Sitatgliptin AND Saxagliptin assessed in March 2015

• Sitagliptin AND Vildagliptin assessed in March 2015

• Sitagliptin AND Alogliptin assessed in March 2015

• Sitagliptin AND Linagliptin assessed in March 2015

• Saxagliptin AND Vildagliptin assessed in March 2015

• Saxagliptin AND Alogliptin assessed in March 2015

• Saxagliptin AND Linagliptin assessed in March 2015

• Vildagliptin AND Alogliptin assessed in March 2015

• Vildagliptin AND Linagliptin assessed in March 2015

• Alogliptin AND Linagliptin assessed in March 2015

References

1. International Diabetes Federation. IDF Diabetes atlas, 7th edition.

Brussels, Belgium: International Diabetes Federation, 2015.

http://www.diabetesatlas.org. Accessed 20 Apr 2016

2. Kanavos, P., Aardweg, S., Schurer, W.: Diabetes expenditure,

burden of disease and management in 5 EU countries. LSE

Health, London School of Economics, London (2012)

3. American Diabetes Association. Economic costs of diabetes in the

U.S. in 2012. Diabetes Care 2013. http://care.diabetesjournals.org/

content/early/2013/03/05/dc12-2625.full.pdf. Accessed 20 Apr

2016

4. Inzucchi, S.E., Bergenstal, R.M., Buse, J.B., Diamant, M., Fer-

rannini, E., Nauck, M., Peters, A.L., Tsapas, A., Wender, R.,

Matthews, D.R.: Management of hyperglycemia in type 2 dia-

betes, 2015: a patient-centered approach. Update to a position

statement of the American Diabetes Association and the Euro-

pean Association for the study of diabetes. Diabetes Care 38,140–149 (2015). doi:10.2337/dc14-2441

5. Liberati, A., Altman, D.G., Tetzlaff, J., Mulrow, C., Gøtzsche,

P.C., Ioannidis, J.P.A., Clarke, M., Devereaux, P.J., Kleijnen, J.,

Moher, D.: The PRISMA statement for reporting systematic

reviews and meta-analyses of studies that evaluate health care

interventions: explanation and elaboration. Ann. Int. Med. 151,W65–W94 (2009)

6. Centre for Reviews and Dissemination: Systematic reviews.

CRD’s guidance for undertaking reviews in healthcare. CRD

University of York, York (2008)

7. Drummond, M., Jefferson, T.: Guidelines for authors and peer

reviewers of economic submissions to the BMJ. The BMJ Eco-

nomic Evaluation Working Party. BMJ 313, 275–283 (1996)

8. World Bank. World Bank Country and Lending Groups Data.

https://datahelpdesk.worldbank.org/knowledgebase/articles/9065

19#High_income (2016). Accessed 6 Jul 2016

9. Hay, J., Smeeding, J., Carrol, N.V., Drummond, M., Garrison,

L.P., Mansley, E.C., Mullins, C.D., Mycka, J.M., Seal, B., Shi, L.:

Good research practices for measuring drug costs in cost effec-

tiveness analyses: issues and recommendations: the ISPOR Drug

Cost Task Force report—part I. Value Health 13, 3–7 (2010)

10. Sachs, J.D.: Macroeconomics and health: investing in health for

economic development. Report of the Commission on Macroe-

conomics and Health. World Health Organization, Geneva (2003)

11. Baltussen, R.M.P.M., Adam, T., Tan-Torres, E.T., Hutubessy,

R.C.W., Acharya, A., Evans, D.B., Murray, C.J.L.: Making

choices in health: WHO guide to cost-effectiveness analysis.

World Health Organization, Geneva (2003)

12. Schwarz, B., Gouveia, M., Chen, J., Nocea, G., Jameson, K.,

Cook, J., Krishnarajah, G., Alemao, E., Yin, D., Sintonen, H.:

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(Supplement 1), 43–55 (2008)

13. Pereira, R., Gouveia, M., Martins, A.P.: Analise custo-efectivi-

dade de sitagliptina quando adicionada a metformina em doentes

com diabetes tipo 2 em Portugal. Revista Portuguesa de diabetes

7(1), 13–23 (2012)

14. Elgart, J.F., Caporale, J.E., Gonzalez, L., Aiello, E., Waschbusch,

M., Gagliardino, J.J.: Treatment of type 2 diabetes with sax-

agliptin: a pharmacoeconomic evaluation in Argentina. Health

Econ. Rev. 3(1), 11 (2013)

15. Erhardt, W., Bergenheim, K., Duprat-Lomon, I., McEwan, P.:

Cost effectiveness of saxagliptin and metformin versus sulfony-

lurea and metformin in the treatment of type 2 diabetes mellitus


123

http://www.diabetesatlas.org

http://care.diabetesjournals.org/content/early/2013/03/05/dc12-2625.full.pdf

http://care.diabetesjournals.org/content/early/2013/03/05/dc12-2625.full.pdf

http://dx.doi.org/10.2337/dc14-2441

https://datahelpdesk.worldbank.org/knowledgebase/articles/906519%23High_income

https://datahelpdesk.worldbank.org/knowledgebase/articles/906519%23High_income

in Germany: a Cardiff Diabetes Model analysis. Clin. Drug

Investig. 32(3), 189–202 (2012)

16. Granstrom, O., Bergenheim, K., McEwan, P., Sennfalt, K.,

Henriksson, M.: Cost-effectiveness of saxagliptin (Onglyza) in

type 2 diabetes in Sweden. Primary Care Diabetes 6(2), 127–136(2012)

17. Bergenheim, K., Williams, S.A., Bergeson, J.G., Stern, L., Sri-

prasert, M.: US cost-effectiveness of saxagliptin in type 2 dia-

betes mellitus. Am. J. Pharm. Benefits 4(1), 20–28 (2012)

18. Carvalho, D., Contente, M., Silva, C., Trindade, R.: Saxagliptin

in the treatment of diabetes mellitus type 2 in Portugal: a study of

cost-utility in the perspective of society. Revista Portuguesa da

Diabetes 9(2), 60–72 (2014)

19. Grzeszczak, W., Czupryniak, L., Kolasa, K., Sciborski, C.,

Lomon, I.D., McEwan, P.: The cost-effectiveness of saxagliptin

versus NPH insulin when used in combination with other oral

antidiabetes agents in the treatment of type 2 diabetes mellitus in

Poland. Diabetes Technol. Therap. 14(1), 65–73 (2012)

20. Nita, M.E., Eliaschewitz, F.G., Ribeiro, E., Asano, E., Barbosa,

E., Takemoto, M., Donato, B., Rached, R., Rahal, E.: Cost-ef-

fectiveness and budget impact of saxagliptin as additional therapy

to metformin for the treatment of diabetes mellitus type 2 in the

Brazilian private health system. Rev. Assoc. Med. Bras. 58(3),294–301 (2011)

21. Viriato, D., Calado, F.; Gruenberger, J.B., Ong, S.H., Carvalho,

D., Silva-Nunes, J., Johal, S., Viana, R.: Cost-effectiveness of

metformin plus vildagliptin compared with metformin plus

sulphonylurea for the treatment of patients with type 2 diabetes

mellitus: a Portuguese healthcare system perspective. J. Med.

Econ. 17(7), 499–507 (2014)

22. Li, Q., Chitnis, A., Hammer, M., Langer, J.: Real-world clinical

and economic outcomes of liraglutide versus sitagliptin in

patients with type 2 diabetes mellitus in the United States. Dia-

betes Ther. 5(2), 579–590 (2014)

23. Carlsson, K.S., Persson, U.: Cost-effectiveness of add-on treat-

ments to metformin in a Swedish setting: liraglutide vs sulpho-

nylurea or sitagliptin J. Med. Econ. 17(9), 658–669 (2014).

doi:10.3111/13696998.2014.933110. (Epub 2014 Jun 26)24. Tzanetakos, C., Melidonis, A., Verras, C., Kourlaba, G., Mani-

adakis, N.: Cost-effectiveness analysis of liraglutide versus sita-

gliptin or exenatide in patients with inadequately controlled type

2 diabetes on oral antidiabetic drugs in Greece. BMC Health

Serv. Res. 14, 419 (2014)

25. Langer, J., Hunt, B., Valentine, W.J.: Evaluating the short-term

cost-effectiveness of liraglutide versus sitagliptin in patients with

type 2 diabetes failing metformin monotherapy in the United

States. J Manag. Care Pharm. 19(3), 237–246 (2013)

26. Perez, A., Raya, P.M., Arellano, A.R., Briones, T., Hunt, B.,

Valentine, W.J.: Cost-effectiveness analysis of incretin therapy

for type 2 diabetes in Spain:1.8 mg liraglutide vs. sitagliptin.

Diabetes Ther. 6, 61–74 (2015)

27. Raya, P.M., Perez, A., Arellano, A.R., Briones, T., Hunt, B.,

Valentine, W.J.: Incretin therapy for type 2 diabetes in Spain: a

cost-effectiveness analysis of liraglutide versus sitagliptin. Dia-

betes Ther. 4, 417–430 (2013)

28. Davies, M.J., Chubb, B.D., Smith, I.C., Valentine, W.J.: Cost-

utility analysis of liraglutide compared with sulphonylurea or

sitagliptin, all as add-on to metformin monotherapy in type 2

diabetes mellitus. Diabet. Med. 29(3), 313–320 (2012)

29. Lee, W.C., Samyshkin, Y., Langer, J., Palmer, J.L.: Long-term

clinical and economic outcomes associated with liraglutide ver-

sus sitagliptin therapy when added to metformin in the treatment

of type 2 diabetes: a CORE diabetes model analysis. J. Med.

Econ. 15(Supplement 2), 28–37 (2012)

30. Brown, S.T., Grima, D.G., Sauriol, L.: Cost-effectiveness of

insulin glargine versus sitagliptin in insulin-naive patients

with type 2 diabetes mellitus. Clin. Ther. 36(11), 1576–1587(2014)

31. Kiadaliri, A.A., Gerdtham, U.G., Eliasson, B., Carlsson, K.S.:

Cost-utility analysis of glucagon-like peptide-1 agonists com-

pared with dipeptidyl peptidase-4 inhibitors or neutral protamine

hagedorn basal insulin as add-on to metformin in type 2 diabetes

in Sweden. Diabetes Ther. 5, 591–607 (2014)

32. Chen, J., Alemao, E., Yin, D., Cook, J.: Development of a dia-

betes treatment simulation model: with application to assessing

alternative treatment intensification strategies on survival and

diabetes related complications. Diabetes Obes. Metab. 10(47),1747–1759 (2008)

33. Clarke, P.M., Gray, A.M., Briggs, A., Farmer, A.J., Fenn, P.,

Stevens, R.J., Matthews, D.R., Stratton, I.M., Holman, R.R.: A

model to estimate the lifetime health outcomes of patients with

diabetes type 2: the United Kingdom prospective diabetes study

(UKPDS) outcomes model (UKPDS no.68). Diabetologia 47,1747–1759 (2004)

34. Palmer, A.J., Mount Hood 5 Modeling Group.: Computer mod-

eling of diabetes and its complications: a report on the fifth

Mount Hood challenge meeting. Value Health 16(4), 670–685(2013)

35. McEwan, P., Peters, J.R., Bergenheim, K., Currie, C.J.: Evalua-

tion of the costs and outcomes from changes in risk factors in

type 2 diabetes using the Cardiff stochastic simulation cost-utility

model (DiabForecaster). Curr. Med. Res. Opin. 22, 121–129

(2006)

36. McEwan, P., Bergenheim, K., Yuan, Y., Tetlow, A.P., Gor-

don, J.P.: Assessing the relationship between computational

speed and precision: a case study comparing an interpreted

versus compiled programming language using a stochastic

simulation model in healthcare. Pharmacoeconomics 28,665–674 (2010)

37. Palmer, A.J., Roze, S., Valentine, W.J., Minshall, M.E., Foos, V.,

Lurati, F.M., Lammert, M., Spinas, G.A.: The core diabetes

model: projecting long-term clinical outcomes, costs and cost-

effectiveness of interventions in diabetes mellitus (types 1 and 2)

to support clinical and reimbursement decision-making. Curr.

Med. Res. Opin. 20(1 Suppl), S5–S26 (2004)

38. Pratley, R., Nauck, M., Bailey, T., Montanya, E., Cuddihy, R.,

Filetti, S., Garber, A., Thomses, A.B., Hartvig, H., Davies, M.:

One year of liraglutide treatment offers sustained and more

effective glycaemic control and weight reduction compared with

sitagliptin, both in combination with metformin, in patients with

type 2 diabetes: a randomised, parallel-group, open-label trial.

Int. J. Clin. Pract. 65, 397–407 (2011)

39. OECD. OECD.Stat. https://stats.oecd.org/Index.aspx?DataSet

Code=PDB_LV## (2015). Accessed 29 Jun 2016

40. Waugh, N., Cummins, E., Royle, P., Clar, C., Marien, M.,

Richter, B., Philip, S.: Newer agents for blood glucose control in

type 2 diabetes: systematic review and economic evaluation.

Health Technol. Assess. 14(36), 1–248 (2010)

41. Geng, J., Yu, H., Ma, O.Y., Zhang, P., Chen, Y.: Cost effec-

tiveness of dipeptidyl peptidase-4 inhibitors for type 2 diabetes.

Pharmacoeconomics 33, 581–597 (2015)

42. Ahuja, V., Chou, C.H.: Novel therapeutics for diabetes: uptake,

usage trends, and comparative effectiveness. Curr. Diabetes Rep.

16, 47 (2016)

43. Bell, C.H., Urbach, D.R., Ray, J.G., Bayoumi, A., Rosen, A.B.,

Greenberg, D., Neumann, P.J.: Bias in published cost effective-

ness studies: systematic review. BMJ 332(7543), 699–703 (2006)

44. Turner, L.W., Nartey, D., Stafford, R.S., Singh, S., Alexander,

G.C.: Ambulatory treatment of type 2 diabetes in the U.S.,

1997–2012. Diabetes Care 37(4), 985–992 (2014). doi:10.2337/

dc13-2097


123

http://dx.doi.org/10.3111/13696998.2014.933110



Documents

The place of DPP-4 inhibitors in the treatment algorithm ...€¦ · the country and year of each study, DPP-4 inhibitors were highly cost-effective as second-line, as add-ons to