Anticoagulant selection for patients with VTE—Evidence from a systematic literature review of network meta-analyses

Accepted Manuscript Title: Anticoagulant Selection for Patients with VTE—Evidence from A Systematic Literature Review of Network Meta-analyses Authors: Alexander T. Cohen, Samantha E. Berger, Duˇsan Milenkovi´c, Nathan R. Hill, Steven Lister PII: DOI: Reference:

S1043-6618(18)31842-5 https://doi.org/10.1016/j.phrs.2019.03.017 YPHRS 4208

To appear in:

Pharmacological Research

Received date: Revised date: Accepted date:

12 December 2018 5 February 2019 20 March 2019

Please cite this article as: Cohen AT, Berger SE, Milenkovi´c D, Hill NR, Lister S, Anticoagulant Selection for Patients with VTE—Evidence from A Systematic Literature Review of Network Meta-analyses, Pharmacological Research (2019), https://doi.org/10.1016/j.phrs.2019.03.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Anticoagulant Selection for Patients with VTE—Evidence from A Systematic Literature Review of Network Meta-analyses

SC RI PT

Authors: Alexander T. Cohena; Samantha E. Bergerb; Dušan Milenkovićc; Nathan R. Hilld; Steven Listere* Affiliations: a

Department of Haematology, Guy’s and St. Thomas’ Hospitals, King’s College, Westminster

Bridge Rd, Lambeth, London SE1 7EH, UK; [email protected]

Meta Research, Evidera, 500 Totten Pond Rd., Fifth Floor, Waltham, MA, USA;

U

b

[email protected]

Department of Innovative Medicine (Cardiovascular and Immunology), Bristol-Myers Squibb

D

d

A

Meta Research, Evidera, Metro Building, 6th Floor, 1 Butterwick, London W6 8DL, UK;

M

c

N

[email protected]

TE

Pharmaceuticals Ltd, Uxbridge Business Park, Sanderson Road, Uxbridge, Middlesex UB8 1DH, UK; [email protected]

Health Economics and Outcomes Research, Bristol-Myers Squibb Pharmaceuticals Ltd,

EP

e

CC

Uxbridge Business Park, Sanderson Road, Uxbridge, Middlesex UB8 1DH, UK; [email protected]

A

*Corresponding Author Steven Lister, MA cantab Health Economics and Outcomes Research, Bristol-Myers Squibb Pharmaceuticals Ltd

1

Uxbridge Business Park, Sanderson Road Uxbridge, Middlesex UB8 1DH, UK

SC RI PT

Tel: +44 (0)189.552.3744 | Fax: 01895 523010 | E-mail: [email protected]

A

CC

EP

TE

D

M

A

N

U

Graphical abstract

2

Abstract The aim of this study was to systematically review published network meta-analyses (NMAs) that compare venous thromboembolism (VTE) treatments.

SC RI PT

A systematic literature search (in MEDLINE, Embase, and Cochrane Database of Systematic

Reviews through September 2017) was conducted to identify NMAs that compared the safety and efficacy of direct oral anticoagulants (DOACs) for the treatment of VTE in the acute and extended treatment settings. The NMAs included randomized controlled trials comparing

U

multiple DOACs, low-molecular weight heparin, unfractionated heparin, and vitamin K

N

antagonists (VKAs). The quality of the NMA results were evaluated using the Grading of

A

Recommendations and Evaluation (GRADE) assessment.

M

The SLR identified 294 records and nine NMAs (68 trials). Among the NMAs, three evaluated the acute treatment setting, five the extended, and one in both treatment settings. The NMAs

D

showed a significant reduction in major bleeding and clinically relevant bleeding (CRB) with

TE

apixaban compared to other DOACs. Major bleeding with apixaban was reduced compared to

EP

dabigatran, edoxaban, and fondaparinux-VKA combination in all comparisons in the acute setting (range of effect estimates: 0.30–0.43). CRB was reduced with apixaban compared to

CC

dabigatran, edoxaban, and rivaroxaban in the acute and extended settings (range of effect estimates: 0.23–0.72). No significant differences were seen in efficacy outcomes between the

A

DOACs.

This SLR of NMAs systematically collected all indirect evidence of the impact of apixaban compared to other anticoagulants in patients with VTE. In the absence of head-to-head trials, well-conducted NMAs provide the best evidence.

3

Key Words: Venous thromboembolism, deep vein thrombosis, pulmonary embolism, direct oral anticoagulants, systematic literature review, network meta-analysis

Introduction

SC RI PT

1

Venous thromboembolism (VTE), a condition that includes deep venous thrombosis (DVT) and pulmonary embolism (PE), is associated with significant morbidity and mortality. Although

warfarin has been the standard oral anticoagulant VTE treatment for several years, direct oral

U

anticoagulants (DOACs), either preferred to vitamin K antagonists (VKAs) or as an alternative to

N

VKAs, have been currently recommended by several clinical guidelines because of their superior

A

safety profile with respect to bleeding.1,2 Reduced rates of fatal bleeding, major bleeding, and

M

clinically relevant non-major bleeding (CRNMB) with DOACs compared to VKAs have been reported in a prior meta-analysis.3 In managing patients with VTE, healthcare providers face the

D

challenge of balancing the benefits of DOACs, such as their efficacy and safety in reducing risk

TE

of recurrent VTE and other thromboembolic phenomena, and the risk of side effects, mainly

EP

bleeding.4,5 DOACs have fewer treatment management complexities and are more convenient for patients and clinicians, as they do not require the regular monitoring of international

CC

normalization ratio (INR) or have the drug interactions associated with use of VKAs.6

A

The comparative efficacy and safety of DOACs versus VKAs or placebo for the management of VTE in the acute and extended treatment settings have been demonstrated in phase II and III randomized controlled trials. Key trials in the treatment of acute VTE by DOACs include: apixaban (AMPLIFY, BOTICELLI -DVT),7,8 dabigatran (RE-COVER, RE-COVER II),9,10 edoxaban (HOKUSAI-VTE),11 and rivaroxaban (EINSTEIN DVT, EINSTEIN DVT-dose

4

ranging, EINSTEN PE, ODiXa-DVT).12-14 The four DOAC trials in the extended treatment settings include three of the four comparators as the acute setting - apixaban (AMPLIFY-EXT),15 dabigatran (RE-SONATE, RE-MEDY),16 and rivaroxaban (EINSTEIN-EXT).13,17

SC RI PT

Major bleeding is not uncommon in the treatment of acute VTE where the incidence of bleeding is highest and anticoagulant interruption may have significant clinical consequences.18 However in the extended setting, the burden of bleeding continues but major bleeding occurs at a

significantly lower incidence rate and therefore fewer events are seen. Hence in the extended

setting clinically relevant bleeding (CRB), the composite of major and CRNM bleeding becomes

U

the outcome of interest as it includes major bleeding and clinically relevant non-major bleeding.

A

N

These two components of CRB are of importance to clinicians and patients.

M

Better understanding of the best treatment to prevent or reduce the risk of bleeding events during treatment with DOACs for VTE may enable health professionals when the decision on intensity

D

and VTE treatment is reached. Due to the lack of direct comparisons of DOACs in patients with

TE

VTE in randomized controlled trials (RCTs), several indirect comparisons such as network metaanalyses (NMAs) have been undertaken to estimate the relative safety and efficacy among

EP

DOACs. In order to bring the evidence together from the different NMAs and summarize the

CC

results collectively, we undertook a systematic literature review (SLR) of published NMAs to examine the methodological approach and outcomes of the NMAs and to help answer the

A

important clinical questions concerning the efficacy and safety of DOACs in the treatment of VTE.

5

2

Methods

The systematic literature review (SLR) was conducted and reported according to the recommendations of the Cochrane Collaboration and the Preferred Reporting Items for

SC RI PT

Systematic Reviews and Meta-Analyses (PRISMA) Statement.19,20 MEDLINE, Embase, and the Cochrane Database of Systematic Reviews databases were searched to identify English-language NMAs published between January 1, 2010 and September 29, 2017. The full search strategies are reported in the S1 Appendix (S1 Table 2).

U

Studies that met the following predefined eligibility criteria were systematically identified for

N

inclusion in the review: NMAs including RCTs evaluating multiple comparisons of DOACs,

A

low-molecular weight heparin (LMWH), unfractionated heparin (UFH), and VKAs as VTE

M

therapies in the acute or extended treatment settings and reporting at least one outcome of interest. Outcomes of interest included recurrent/symptomatic VTE (DVT, PE, or both),

D

mortality, and bleeding (major bleeding (which includes fatal bleeding), clinically relevant non-

TE

major bleeding (CRNMB), the composite of these, clinically relevant bleeding (CRB) or any other bleeding types). NMAs were required to have included ≥ 90% patients with VTE (DVT or

EP

PE) or report outcomes data separately for VTE subjects. NMAs of observational studies, NMAs

CC

of RCTs evaluating DOACs for primary prevention, and meta-analyses that pooled data from RCTs evaluating similar comparisons were excluded.

A

Titles and abstracts of citations identified from the searches were evaluated using the abovementioned pre-specified criteria. Citations at the title and abstract level were excluded if they clearly did not meet inclusion/exclusion criteria. Full texts were obtained for the citations that were included after this first screening stage, and these were evaluated again based on the inclusion/exclusion criteria. Screening was conducted by two investigators who were trained at 6

the objectives of the SLR, and disagreements were resolved by a third senior investigator. For each NMA included in the review, data on study characteristics, methodology, and outcomes of interest were extracted by one investigator and validated for accuracy by a second investigator.

SC RI PT

The quality of evidence (effect estimates) from the identified NMAs was rated for each comparison and outcome based on the guidance by the Grading of Recommendations and

Evaluation (GRADE) Working Group.21. Direct, indirect and NMA effect estimates were

presented and its quality were rated using the recommended GRADE domains: risk of bias, inconsistency, indirectness, incoherence, and imprecision.21 Of note, indirect evidence was

U

examined for the RCTs in the first order loop of the evidence network via the common

N

comparator of VKA or placebo (network comparisons involving only a single additional

M

A

intervention) as suggested by this GRADE guidance.

Evidence from the identified NMAs comparing DOACs versus other DOACs or non-DOACs

D

was synthesized. Results are presented for all other DOACs and non-DOACs compared to

TE

apixaban as only apixaban had demonstrated reductions in major bleeding and clinically relevant non-major bleeding versus VKA.7 Estimates of treatment effects and their corresponding

EP

credible intervals were compiled and presented in forest plots per each outcome of interest,

CC

namely bleeding (safety) and recurrent VTE (DVT and or PE) (efficacy). Results for other

A

outcomes are summarized in figures and presented in S1 Appendix.

7

3

Results

3.1

SLR Results

The search identified 294 unique records and these were reviewed at the title/abstract level. We

SC RI PT

included 22 records for review at the full-text level, and identified nine NMAs for inclusion that reported data across 68 primary trials for comparisons among apixaban, other DOACs, LMWH, and UFH.22-30 A PRISMA diagram detailing the flow of NMAs during the screening process is

A

CC

EP

TE

D

M

A

N

U

shown in Figure 1.

8

Figure 1. PRISMA Flow Chart

109 DUPLICATES were removed

294 RECORDS were reviewed after duplicate citations were removed

M

A

N

U

272 ABSTRACTS were excluded

SC RI PT

Records identified through database (PubMed, Embase, CDSR) searches N=403

22 RECORDS had a full-text review

TE

 6: not a study design of interest  6: not a population of interest

D

13 PUBLICATIONS were excluded:

 1: did not report outcomes

9 NMAs in 9 publications were included in the review

A

CC

EP

for comparisons of interest

Abbreviations: CDSR = Cochrane Database of Systematic Reviews; NMA = network meta-analysis; PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses

9

A summary of the characteristics of the included NMAs are presented in Table 1. Of the nine included NMAs, three of them evaluated outcomes related to the acute treatment setting,24,26,27 five evaluated outcomes for the extended-treatment setting,22,23,25,28,29 and one NMA evaluated

SC RI PT

outcomes in both settings.30 Across all nine NMAs, the population examined was patients with symptomatic VTE (DVT or PE). Across the four NMAs evaluating DOACs in the acute treatment setting, all NMAs included the AMPLIFY, EINSTEIN-DVT, EINSTEIN-PE,

RECOVER I and RECOVER II RCTs. Three of the four NMAs also included the HOKUSAIVTE trial. One NMA in the acute setting had broader inclusion criteria of anticoagulant

U

comparisons thus included a larger evidence network of 45 trials.23 Similarly, the trials

N

overlapped across the NMAs in the extended treatment setting. The specific trials included in

A

each NMA are summarized in Table 2. The AMPLIFY-EXT, EINSTEIN-EXT and RESONATE

A

CC

EP

TE

D

M

trials were included in all six NMAs evaluating DOACs in the extended treatment setting.

10

Table 1. Characteristics of Network Meta-analyses RCTs Included

Population (Treatment Setting)

Outcomes Assessed

Statistical Approach (framework, model type etc.)

Drugs and Doses Compared with Apixaban for Primary Outcomes

Alotaibi et al, 2014

AMPLIFYEXT EINSTEINEXT RE-SONATE

Subjects with VTE (extended treatment)

Bleeding: • Major • CRB

NR

All-cause Mortality: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

EP

A

CC

Castellucci et al, 2014

Subjects with VTE (acute treatment)

Recurrent VTE: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

Bayesian NMA and direct, frequentist, pairwise metaanalysis

N

Other: • Fatal recurrent VTE • Recurrent VTE

M

AmericanCanadian Thrombosis Study AMPLIFY CORTES EINSTEIN EINSTEIN-PE FIDO IRIS Main-LITE MATISSE-PE MATISSEDVT RE-COVER RE-COVER II

Major Bleeding: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

U

Bleeding: • Major

A

Subjects with symptomatic DVT or PE (extended treatment)

D

ASPIRE AMPLIFYEXT DURAC II EISTEINEXT ELATE LAFIT PREVENT RESONATE REMEDY Thrive III WARFASA WODIT DVT WODIT PE

CRNMB: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

Other: • Recurrent DVT • Recurrent PE • Recurrent VTE • ACS • All-cause mortality

TE

Castellucci et al, 2013

SC RI PT

Study

Bleeding: • Major Other: • Recurrent DVT • Recurrent PE • Recurrent VTE • Fatal recurrent VTE

11

Binomial likelihood model, random effects model for informative priors and fixed- and randomeffects models for vague priors Bayesian NMA and direct, frequentist, pairwise metaanalysis Poisson likelihood model, randomeffects model for vague priors, fixedeffects model used with binomial likelihood model and

Major Bleeding: Apixaban 2.5 mg BID Apixaban 5 mg BID Dabigatran 150 mg BID Rivaroxaban 20 mg OD Recurrent VTE: Apixaban 2.5 mg BID Apixaban 5 mg BID Dabigatran 150 mg BID Rivaroxaban 20 mg OD

Major Bleeding: Fondaparinux–VKA LMWH alone LMWH–Dabigatran 150 mg BID LMWH–Edoxaban 60 mg OD LMWH–VKA Rivaroxaban 15 mg BID then 20 mg OD UFH–VKA combination Recurrent VTE: Fondaparinux–VKA LMWH alone LMWH–Edoxaban 60 mg or 30 mg OD LMWH–VKA

RCTs Included

Population (Treatment Setting)

Outcomes Assessed

Statistical Approach (framework, model type etc.) randomeffects models with informative priors

TASMAN THESEE 30 additional trials with no NCT# or Trial name Subjects with VTE (extended treatment)

Bayesian fixed-effects NMA

Bleeding: • Major • CRNMB • CRB • Intracranial Other: • Recurrent DVT • Recurrent PE • Recurrent VTE • VTE and fatal VTE • Non-fatal VTE • MI • Overall treatment discontinuation • All-cause mortality

N

A D

TE

AMPLIFY EINSTEINDVT EINSTEIN-PE HOKUSAIVTE RECOVER RE-COVER II

A

CC

EP

Hirschl et al, 2014

Subjects with VTE (acute treatment)

Bleeding: • Major • CRB • Fatal • Intracranial Other: • Recurrent VTE • Fatal or nonfatal PE or DVT • All-cause mortality

All-cause mortality: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

Fatal VTE: Dabigatran 150 mg BID Rivaroxaban 20 mg OD Major Bleeding: Dabigatran 150 mg BID Rivaroxaban 20 mg OD Recurrent VTE: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

A metaanalytical estimate of overall efficacy of DOACs with respect to VKA and of the safety endpoint ‘major bleeding’ was computed applying the random-effect model of Der SimonianLaird A fixed-effect approach was used for other outcomes

12

Rivaroxaban 15 mg BID then 20 mg OD UFH–VKA combination

CRNMB: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

U

AMPLIFYEXT ASPIRE EINSTEINEXT ELATE LAFIT PREVENT RESONATE RE-MEDY WARFASA WODIT DVT WODIT PE

M

Cohen et al, 2016

Drugs and Doses Compared with Apixaban for Primary Outcomes

SC RI PT

Study

All-cause Mortality: LMWH–Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD Intracranial Bleeding: LMWH–Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD Major Bleeding: LMWH–Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD

RCTs Included

Population (Treatment Setting)

Outcomes Assessed

Statistical Approach (framework, model type etc.)

Drugs and Doses Compared with Apixaban for Primary Outcomes

Kang et al, 2014

AMPLIFY EINSTEIN-PE EINSTEINDVT HOKUSAIVTE RECOVER RECOVER II

Subjects with VTE (acute treatment)

Bleeding: • Major

Buchers adjusted indirect comparison meta-analysis

All-cause Mortality: LMWH–Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD

No information on model type

EP

A

CC

Sobieraj et al, 2015

AMPLIFYEXT ASPIRE EINSTEINEXT Kearon, 1999 RESONATE REMEDY WARFASA WODIT-DVT WODIT-PE Van Gogh, 2007

Bleeding: • Major • CRNMB

M

Subjects with VTE (extended treatment)

D

AMPLIFY AMPLIFYEXT EINSTEINEXT ELATE LAFIT PREVENT RE-MEDY RE-SONATE WODIT-PE WODIT-DVT

TE

Rollins et al, 2014

A

N

U

Other: • Recurrent VTE • Recurrent PE • Recurrent DVT • All-cause mortality

SC RI PT

Study

Subjects with VTE (extended treatment)

Bayesian randomeffects NMA

Other: • Recurrent VTE

13

Recurrent VTE: LMWH–Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD CRNMB: Dabigatran 150 mg BID Rivaroxaban 20 mg OD Major Bleeding: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

Other: • Recurrent VTE or death • Non-fatal DVT • Non-fatal PE

Bleeding: • CRB

Major Bleeding: LMWH–Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD

Recurrent VTE: Dabigatran 150 mg BID Rivaroxaban 20 mg OD

Frequentist randomeffects NMA

Recurrent VTE: Apixaban 2.5 mg BID Apixaban 5 mg BID Dabigatran 150 mg BID Idraparinux 2.5 mg QW Rivaroxaban 20 mg OD

RCTs Included

Population (Treatment Setting)

Outcomes Assessed

Statistical Approach (framework, model type etc.)

Drugs and Doses Compared with Apixaban for Primary Outcomes

Sterne et al, 2017

Acute AMPLIFY BOTTICELLI DVT EINSTEIN DVT EINSTEIN DVT doseranging study EINSTEIN PE HOKUSAIVTE ODiXa-DVT RE-COVER RE-COVER II

Subjects with VTE (acute and extended treatment)

Bleeding: • Major • Major and CRNMB

Bayesian fixed-effects NMA

All-cause Mortality: Apixaban 5 mg BID Apixaban 10 mg BID LMWH–Dabigatran 150 mg BID Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD Rivaroxaban 20 mg OD

U

N A M D

EP

TE

Extended AMPLIFYEXT ASPIRE EINSTEINEXT LAFIT PREVENT RE-MEDY RE-SONATE WARFASA WODIT-PE WODIT-DVT

Other: • Symptomatic DVT • Symptomatic PE • Symptomatic VTE • Recurrent VTE • All-cause Mortality

SC RI PT

Study

CRB Apixaban 2.5 mg BID Apixaban 5 mg BID LMWH–Dabigatran 150 mg BID Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD Rivaroxaban 20 mg OD Major Bleeding: Apixaban 5 mg BID Apixaban 10 mg BID LMWH–Dabigatran 150 mg BID Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD Rivaroxaban 20 mg OD

A

CC

Recurrent VTE: Apixaban 2.5 mg BID Apixaban 5 mg BID LMWH–Dabigatran 150 mg BID Dabigatran 150 mg BID LMWH–Edoxaban 60 mg or 30 mg OD Rivaroxaban 15 mg BID then 20 mg OD Rivaroxaban 20 mg OD

Abbreviations: AE = adverse event; BID = twice per day; CRB = clinically relevant bleeding; CRNMB = clinically relevant nonmajor bleeding; DOAC = direct oral anticoagulant; DVT = deep venous thrombosis; ITC = indirect treatment comparison;

14

LMWH = low-molecular weight heparin; NMA = network meta-analysis; OD = once daily; PE = pulmonary embolism; PICOS = patients, intervention, comparator, outcomes, study design; QW = once per week; RCT = randomized controlled trial; SLR = systematic literature review; UFH = unfractionated heparin; VKA = vitamin K antagonist; VTE = venous thromboembolism

Table 2. List of Included DOACs and Doses Licensed Dose

Unlicensed Dose

SC RI PT

Treatment Acute Treatment Setting Apixaban

10 mg BID then 5 mg BID

LMWH-Dabigatran

150 mg BID

LMWH-Edoxaban

60 mg or 30 mg OD

Rivaroxaban

15 mg BID then 20 mg OD

Fondaparinux-VKA

NR

UFH-LMWH

NR

LMWH-VKA

NR

10 mg BID 20 mg BID None None

NR

U

NR NR

150 mg BID

Rivaroxaban

20 mg OD

Idraparinux

NA

A

Dabigatran

M

2.5 mg BID 5 mg BID

D

Apixaban

N

Extended Treatment Setting

None None 10 mg OD 30 mg OD 40 mg OD 2.5 mg QW

TE

Abbreviations: BID: twice daily; LMWH: low molecular weight heparin; NA: not applicable; NR: not reported; OD: once daily; QW: once weekly; VKA: vitamin K antagonist

EP

Both licensed and unlicensed treatments were evaluated by the included NMAs (Table 1). The presentation of NMA results focuses on the licensed doses whereas the results for the unlicensed

CC

doses can be found in the S1 Appendix.

A

More than half the NMAs (n = 5; 56%) used a Bayesian methodology. 23-25,28,30 Other less commonly used approaches included the Bucher method (n = 1)27 and frequentist method (n = 1).29 The approach used was unclear in two NMAs.22,26 Three NMAs used a fixed-effects model,25,26,30 two NMAs used a random-effects model,28,29 and two NMAs used both the fixed-

15

and random-effects models.23,24 Two NMAs did not specify the type of statistical model that was used (Table 1).22,27 3.2

GRADE Quality Assessment

SC RI PT

A summary of the GRADE assessment with rationale for down-rating of the evidence is

presented in S1 Tables 3–5. The GRADE assessment was conducted for the direct estimates from RCT data included in the NMAs (when available), the indirect evidence and the NMA

evidence for the outcomes and comparisons of interest. A summary of the GRADE assessment

U

results for apixaban vs. dabigatran, edoxaban and rivaroxaban is summarized in Table 3and

A

CC

EP

TE

D

M

A

N

outlined in Section 3.4.1 and Section 3.4.2.

16

Table 3. Summary of GRADE Ratings of the NMA Evidence for Treatment with Licensed Doses of Apixaban versus Dabigatran, Rivaroxaban, and Edoxaban for All Outcomes in the Acute Treatment of VTE and Extended Treatment Setting

CRB

CRNMB

Recurrent VTE

-

-

-

-

Low-Very low

Acute Apixaban 10 mg then 5 mg BID vs.



 Moderate

 Moderate

ND

Rivaroxaban 15 mg BID then 20 mg OD

 Very low

 Very low

ND

 Very low

LMWH-Edoxaban 60 mg or 30 mg OD

 ModerateHigh

 Moderate

ND

 ModerateLow

N

Extended

Rivaroxaban 20 mg OD Apixaban 5 mg BID vs.

CC

Rivaroxaban 20 mg OD

A



M

Very low  Low-Very low

-

 Moderate



-



Very low

EP

Dabigatran 150 mg BID

-

 Low

D

Dabigatran 150 mg BID

-

TE

Apixaban 2.5 mg BID vs.



Low-Very low

-

 Low  Low

Very low  Low

-



Moderate Very low  Low-Very low

 Very low  Very low

-

 Moderate Very low  Low-Very low

Recurrent PE

All-cause mortality



-



-



Low-Very low

Low-Very low

Low-Very low

 Very low

 Very low

 Very low

 ModerateLow

 ModerateLow

 ModerateLow

-

-

-

U

LMWH-Dabigatran 150 mg BID

Recurrent DVT

SC RI PT

Major bleeding







Very low

Very low

Very low







Very low

Low-Very low

Low-Very low

-

-

-







Very low

Very low

Very low

 Low-Very low

 Low-Very low

 Very low

Abbreviations: BID = twice per day; CRB = clinically relevant bleeding; CRNMB = clinically relevant non-major bleeding; DVT

A

= deep venous thrombosis; OD = once daily; PE = pulmonary embolism; VTE = venous thromboembolism

3.2.1 Acute Treatment of VTE In the RCTs evaluating DOACs in the acute treatment of VTE, across bleeding, recurrent VTE and all-cause mortality outcomes, the direct evidence ranged from high quality to very low quality, but all of the very low quality direct evidence was from trials evaluating unlicensed 17

doses. When the direct evidence was down rated, the main reasons were due to risk of bias mostly due to lack of blinding of participants/personnel and for imprecision of effect estimates. The outcomes and comparisons with NMA estimates of the highest rated quality were apixaban

SC RI PT

10 mg then 5 mg BID vs. edoxaban 60 mg or 30 mg OD and apixaban 10 mg then 5 mg BID vs. dabigatran 150 mg BID. These comparisons for major bleeding and CRB were rated of moderate or high quality (Table 3). For recurrent VTE/DVT/PE and all-cause mortality, the quality of the NMA estimates ranged from moderate to very low across DOAC comparisons, with the

comparison of rivaroxaban 15 mg BID then 20 mg OD rated as very low quality across all

N

U

outcomes (Table 3).

A

3.2.2 Extended Treatment of VTE

M

In the RCTs evaluating DOACs in the extended treatment setting, the quality of the direct evidence ranges from moderate to low quality for major bleeding, from high to moderate for

D

recurrent VTE, from moderate to low for recurrent DVT, from high to low for recurrent PE, and

TE

from high to very low for all-cause mortality.

EP

In the evidence in the extended treatment setting, the only comparison and outcome with a moderate quality rating was apixaban 2.5 mg BID vs. dabigatran 150 mg BID for risk of CRB.

CC

The comparison of apixaban 2.5 mg and 5 mg BID vs. dabigatran 150 mg BID for risk of

A

recurrent VTE ranged from moderate to very low. For all other comparisons and outcomes in the extended treatment setting, the quality ranged from low to very low (Table 3).

18

3.3

Safety Outcomes

3.3.1 Major Bleeding A summary of the evidence on major bleeding for acute VTE treatment and extended treatment

SC RI PT

settings are provided in Figure 2 and Figure 3, respectively. Comparative evidence on major

bleeding between apixaban 10 mg then 5 mg twice per day (BID) and other anticoagulants for acute VTE treatment were reported across four NMAs24,26,27,30 (Figure 2). Overall, the risk of

major bleeding was significantly lower with apixaban 10 mg BID then 5 mg BID compared with

U

LMWH-dabigatran 150 mg BID, LMWH-edoxaban 60 mg or 30 mg OD, and fondaparinux–

N

VKA combination in the treatment of acute VTE (range of effect estimates: 0.30-0.43; Figure 2).

A

The difference in risk of major bleeding was not significant between apixaban 10 mg then 5 mg

M

BID compared to rivaroxaban 15 mg BID then 20 mg OD for treatment of acute VTE (Figure 2). Overall, the risk of major bleeding was significantly lower with apixaban 5 mg and 2.5 mg

D

compared with rivaroxaban 15 mg BID then 20 mg OD in the extended treatment setting (range

A

CC

EP

TE

of effect estimates: 0.02-0.03; Figure 3).

19

M

A

N

U

SC RI PT

Figure 2. Major Bleeding—Treatment of Acute VTE

D

Abbreviations: BID = twice per day; CI = confidence interval; FE = fixed effect; HR = hazard ratio; LMWH = low-molecular

TE

weight heparin; NMA = network meta-analysis; OD = once daily; OR = odds ratio; RE = random effect; RR = relative risk; UFH

A

CC

EP

= unfractionated heparin; VKA = vitamin K antagonist; VTE = venous thromboembolism

20

D

M

A

N

U

SC RI PT

Figure 3. Major Bleeding—Extended Treatment

TE

Abbreviations: BID = twice per day; CI = confidence interval; FE = fixed effect; NMA = network meta-analysis; OD = once

EP

daily; OR = odds ratio; RE = random effect; RR = relative risk

None of the NMAs evaluated the risk of CRNMB in the acute treatment of VTE for the

CC

comparisons of interest. Evidence on CRNMB in the extended treatment setting was reported by three NMAs22,25,28 (S1 Figure 4). The risk of CRNMB was significantly lower with apixaban 2.5

A

mg BID compared with rivaroxaban 20 mg OD;22,25 one study reported a difference compared to dabigatran 150 mg BID,22 while the other study found no difference25 (S1 Figure 4). In contrast, the risk of CRNMB with apixaban 5 mg BID compared to rivaroxaban 20 mg OD varied between NMAs—one reported a significant difference in favor of apixaban,22 but another NMA found no difference.28 The results for apixaban 5 mg BID compared to dabigatran 150 mg BID 21

were also varied between NMAs. One study reported a significantly lower risk with apixaban 5 mg BID compared with dabigatran 150 mg BID, but two NMAs found no differences22,28 (S1 Figure 4).

SC RI PT

Results of the analyses of CRB (the composite outcome of major bleeding and CRNMB) were reported by two NMAs in the treatment of acute VTE26,30 and by three NMAs in the extended treatment setting.25,29,30 In the treatment of acute VTE, apixaban 10 mg then 5 mg BID was

associated with significant reduction in risk of CRB compared with dabigatran 150 mg BID, edoxaban 60 mg or 30 mg OD, and rivaroxaban 15 mg BID then 20 mg OD (range of effect

U

estimates: 0.47-0.72; Figure 4). In the extended treatment setting, the risk of CRB was

N

significantly lower with apixaban at both 2.5 mg BID and 5 mg BID doses than rivaroxaban 20

A

mg OD (range of effect estimates: 0.01-0.24; Figure 5). The risk of CRB was significantly lower

M

with apixaban 2.5 mg BID compared with dabigatran 150 mg BID (RR [95%CI] 0.42[0.18-

D

0.98]), but for apixaban 5 mg BID, the results conflicted across NMAs. One NMA showed a

TE

significant risk reduction with apixaban 5 mg BID compared with dabigatran 150 mg BID, while

A

CC

EP

a second NMA showed no difference.

22

U

SC RI PT

Figure 4. Clinically Relevant Bleeding—Acute Treatment for VTE

N

Abbreviations: BID = twice per day; CI = confidence interval; FE = fixed effect; NMA = network meta-analysis; OD = once daily;

A

CC

EP

TE

D

M

A

RR = relative risk; VTE = venous thromboembolism

23

M

A

N

U

SC RI PT

Figure 5. Clinically Relevant Bleeding—Extended Treatment

Abbreviations: BID = twice per day; CI = confidence interval; CRNMB = clinically relevant non-major bleeding; FE = fixed

D

effect; NMA = network meta-analysis; QW = once per week; OD = once daily; RE = random effect; RR = relative risk; VTE =

TE

venous thromboembolism

Evidence on the risk of intracranial bleeding in the acute treatment setting was reported in only

EP

one study and indicated there were no differences between apixaban and comparators (dabigatran 150 mg BID, edoxaban 60 mg or 30 mg OD, and rivaroxaban 15 mg BID then 20 mg OD) in the

CC

acute treatment setting26 (S1 Figure 5). Data for intracranial bleeding were not reported in the

A

extended treatment setting. None of the NMAs reported comparative evidence on gastrointestinal bleeding.

24

3.4

Efficacy Outcomes

3.4.1 Recurrent VTE/DVT/PE Comparative indirect evidence on the risk of recurrent VTE/DVT/PE between apixaban and

SC RI PT

other DOACs and non-DOACs in the acute and extended treatment settings were reported in four NMAs24,26,27,30 and five NMAs,22,23,25,29,30 respectively. There were no significant differences in the risk of recurrent VTE (DVT or PE) between apixaban and the evaluated comparators in

either treatment setting (the treatment of acute VTE or the extended treatment) (Figure 6 and

U

Figure 7). Additionally, there were no differences in risk of recurrent DVT or recurrent PE in

A

CC

EP

TE

D

M

A

N

either treatment setting (data not shown).

25

D

M

A

N

U

SC RI PT

Figure 6. Recurrent VTE—Treatment of Acute VTE

Abbreviations: BID = twice per day; CI = confidence interval; FE = fixed effect; HR = hazard ratio; LMWH = low-molecular

TE

weight heparin; NMA = network meta-analysis; OD = once daily; OR = odds ratio; RE = random effect; RR = relative risk; UFH

A

CC

EP

= unfractionated heparin; VKA = vitamin K antagonist; VTE = venous thromboembolism

26

SC RI PT

Figure 7. Recurrent VTE—Extended Treatment

*

*

D

M

A

N

U

*

TE

Abbreviations: BID = twice per day; CI = confidence interval; FE = fixed effect; NMA = network meta-analysis; QW = once per week; OD = once daily; OR = odds ratio; RE = random effect; RR = relative risk; VTE = venous thromboembolism

EP

*Results for extended follow-up period

CC

3.4.2 All-cause Mortality Comparative indirect evidence on the risk of all-cause mortality between apixaban and other

A

DOACs and non-DOACs in the acute26,27,30 and extended22,25,30 treatment settings were reported in three NMAs each. There were no significant differences in the risk of all-cause mortality between apixaban and the evaluated comparators in both treatment settings (S1 Figure 6 and S1 Figure 7).

27

4

Discussion

Across nine NMAs evaluating the safety and efficacy of DOACs for the treatment of VTE, the indirect evidence consistently points to a significant reduction in risk of major bleeding and

SC RI PT

clinically relevant bleeding with apixaban compared to other DOACs. Evidence of moderate to high quality confirmed consistently that treatment with apixaban significantly reduced the risk of major bleeding compared to dabigatran, edoxaban, and also fondaparinux-VKA combination, in all published indirect comparisons. Clinically relevant bleeding was also significantly reduced

with apixaban compared with dabigatran, edoxaban, and rivaroxaban. Results relating to major

U

bleeding and clinically relevant bleeding events were consistent across the NMAs, independently

N

of any differences in the design and execution of these analyses. There were no differences in

A

efficacy outcomes between the DOACs, but uncertainty for some of these results was high, given

M

the wide credible intervals, especially those for all-cause mortality. However, for some

D

comparisons, the NMAs reported inconsistent results. This is likely to be due to differences in

TE

the number of RCTs contributing to the evidence network for each NMA, with larger networks having been included in more recent NMAs compared to the earliest example. Additionally,

EP

some NMAs had different inclusion/exclusion criteria and different methodological approaches (random vs. fixed effects) although not enough information were available in the NMA

CC

publications to further explore the impact of such differences in the observed results. There were generally no significant differences between apixaban and other anticoagulants for the following

A

outcomes: intracranial bleeding, recurrent/symptomatic VTE/DVT/PE, non-fatal DVT, non-fatal PE, or all-cause mortality. This SLR of NMAs and indirect treatment comparisons of DOACs provides a new insight into the relative impact of apixaban compared to other anticoagulants for the most clinically relevant outcomes in patients with VTE. Additionally, this SLR considered

28

the quality of the evidence contributed to each of the quantitative results by applying a wellestablished NMA quality tool which further reinforces the validity of its conclusions. Providing therapy with a good safety profile is especially important in patients with VTE who

SC RI PT

require extended or indefinite anticoagulation as the majority of the burden of bleeding occurs in this setting.

Bleeding incidence is highest in the acute and standard treatment phases and bleeding may lead to the interruption or even the early discontinuation of anticoagulation in the phase where this

U

therapy is most critical. In the first three months of anticoagulant therapy for VTE, bleeding and

N

VTE recurrences have similar case fatality but after the first three months of treatment, the case

A

fatality rates for bleeding are three times higher than those of recurrences.1,31-33 Despite the

M

extensive data on the effectiveness of appropriately managed anticoagulants, fear of bleeding events leads to premature discontinuation resulting in a loss of benefit.34

D

In the present study, for patients requiring extended treatment, the collective body of indirect

TE

evidence from the NMAs points to a reduced risk of major bleeding with apixaban compared to

EP

rivaroxaban. Significant reduction in the risk of clinically relevant bleeding was seen with apixaban 2.5 mg BID compared to rivaroxaban and dabigatran and apixaban 5 mg BID

CC

compared with rivaroxaban across indirect comparisons. Results of this review provide insight into the relative impact of DOACs and other anticoagulants in the overall VTE population.

A

Results of this review align with previous evidence of a net clinical benefit of apixaban compared with other DOACs.35 Similar benefits in the risk of major bleeding have been reported with apixaban compared to VKAs and DOACs in non-valvular atrial fibrillation and in the primary prevention of VTE.36-39 A recent observational study in non-valvular atrial fibrillation examined the safety and efficacy of different DOACs in four large, US commercial claims 29

databases, and showed that apixaban was associated with a lower risk of stroke/systemic embolism and major bleeding compared to dabigatran and rivaroxaban.40 Evidence on the comparative efficacy and safety of DOACs and other anticoagulants on the

SC RI PT

components of major bleeding, such as gastrointestinal bleeding, is unclear. Such data were not reported in any of the NMAs included in this review although collected by some of the included RCTs.

The present study evaluated the quality of estimates from the NMAs using the GRADE

U

framework. The evidence for major bleeding and clinically relevant bleeding was of the highest

N

quality thus we are the most confident in results of NMAs comparing DOACs with respect to

A

these outcomes. The risk of major bleeding was significantly reduced with apixaban compared

M

to dabigatran, edoxaban, and also fondaparinux-VKA combination in all published indirect comparisons. The risk of clinically relevant bleeding was also significantly reduced with

D

apixaban compared with dabigatran, edoxaban, and rivaroxaban. The evidence for non-major

TE

clinically relevant bleeding, recurrent VTE/DVT/PE and all-cause mortality were rated of low to very low quality and should be interpreted cautiously. The most common reason for down-rating

EP

the quality of this evidence was for imprecision, which was likely driven by a low number of

CC

trials available for each DOAC comparison and low frequency of occurrence of these outcomes. Limitations and Strengths

A

4.1

To best of our knowledge, this is the first SLR of NMAs in acute and extended treatment setting for VTE. This SLR adds to the gains seen by prospectively planned living NMA, which can facilitate timely treatment recommendations and avoid unnecessary research waste by providing strong evidence for the effectiveness of different treatments in a disease area.41 Although this

30

SLR aimed to identify the most robust direct and indirect evidence from NMAs comparing apixaban versus other DOACs or non-DOACs, there were some limitations that could restrict the applicability of its findings. Firstly, the search was restricted to English peer reviewed

SC RI PT

publications so relevant NMA publications in other languages or published in non-indexed sources (such as HTA websites) may have been missed. This SLR did not conduct a separate

search of RCTs comparing apixaban versus other treatment for the treatment of VTE in either

acute or extended setting, so there may be a possibility the retrieved NMAs have missed relevant publications in this area. Another important consideration when evaluating the results of this

U

SLR of NMAs is that there was significant overlap in the RCTs included across the NMAs and

N

each SLR was searched at a different date. Therefore, older NMAs will not have as much

A

evidence contributing to the network as more recently published NMAs. This may have resulted

M

in different findings for the same outcome and same comparisons published at different dates. Furthermore, the lack of information available regarding the primary RCTs included in the

D

NMAs has restricted in some instances the presentation and quality assessment of effect

TE

estimates. Also, it is difficult to make any assertions on the optimal anticoagulants in specific

EP

subpopulations, such as patients with reduced renal function, cancer or other comorbidities because such evidence was lacking in the included NMAs. A possible reason for this could be

CC

the lack of the reporting of subgroup data in primary trial publications. It is likely that some of

A

these data are reported in secondary publications that were not identified by NMAs. Furthermore, NMA as a statistical technique inherits all the methodological challenges present in a standard meta-analysis (assessment of bias, heterogeneity, precision, and reliability of estimates) but with increased complexity due to the number of comparisons involved. Strong assumptions required for NMAs, such as homogeneity (equivalence of treatment effects across

31

trials within each pairwise comparison), transitivity (the validity of making indirect comparisons), and consistency (the equivalence of direct and indirect evidence) must be met.42-44 In our example, NMA data were the only way to estimate comparative treatment effects between

SC RI PT

DOACs for VTE, given that there were no RCTs with direct comparisons among these treatments of interest. This review was restricted to the information at the NMA level, including inclusion and exclusion criteria and outcome definitions. Based on the authors’ clinical expertise, the trials had similar inclusion/exclusion criteria of adult patients with VTE (PE and/or DVT)

who were not taking medication or had conditions that would interfere with DOAC treatment.

U

Across the trials, most outcomes, including stroke and major bleeding, had similar definitions,

N

but there is typically more heterogeneity in how clinically relevant non-major bleeding is defined

A

and interpreted. For example, some studies define clinically relevant non-major bleeding as overt

M

bleeding not meeting the criteria for major bleeding but associated with medical intervention, contact with a physician, interruption of the study drug, or discomfort or impairment in carrying

D

out activities of daily life. Other studies classify bleeding into major and anything that is not

TE

major, which is classified as minor or clinically relevant non-major bleeding. Studies can also

EP

define non-major bleeding by their own criteria, such as presence of skin hematoma, nose bleed, rectal bleed, bleeding leading to hospitalization, or anything deemed clinically relevant by the

CC

investigator. Furthermore, we assessed the quality of the evidence (direct, indirect) fitted in the NMAs and the strength of the NMA results by implementing the GRADE assessment, a well-

A

recognized, international tool of assessing the quality of NMAs. Further research is required to confirm clinical similarities across the trials included across NMAs.

32

5

Conclusion

In our SLR of NMAs in trials of patients with VTE, most of the evidence points to a significant reduction in risk of major bleeding and clinically relevant bleeding with apixaban (10 mg then 5

SC RI PT

mg BID dose) compared to dabigatran, edoxaban, and the fondaparinux–VKA combination in the acute VTE treatment setting. In the extended treatment setting, there was a reduction of risk of major bleeding with apixaban (2.5 mg or 5 mg BID doses) compared to dabigatran and rivaroxaban. In the absence of head-to-head trials, which are needed to fully understand

A

CC

EP

TE

D

M

A

N

VTE, well-conducted NMAs provide the best evidence.

U

comparative efficacy and safety of DOACs and other anticoagulants used in the management of

33

6

Acknowledgements

The authors would like to thank Ajibade O. Ashaye, Valentina Ricci, and Juliana Lewis, for their assistance with the manuscript. Disclosures

SC RI PT

7

NR Hill and S Lister are employees of Bristol-Myers Squibb Company.

SE Berger and D Milenković are employed by Evidera Inc., which provides consulting and other research services to pharmaceutical, medical device, and related organizations. In their salaried

U

positions, they work with a variety of companies and organizations, and are precluded from

N

receiving payments or honoraria directly from these organizations for services rendered.

A

AT Cohen reports grants and personal fees from Bristol-Myers Squibb Company and Pfizer Inc.

M

during the conduct of the study; personal fees from Boehringer Ingelheim, grants and personal fees

D

from Bristol-Myers Squibb Company, grants and personal fees from Daiichi-Sankyo Europe,

TE

personal fees from Johnson & Johnson, grants and personal fees from Pfizer, Inc., personal fees from Portola, personal fees from Sanofi, personal fees from XO1, personal fees from Janssen,

EP

personal fees from ONO Pharmaceuticals, and grants and personal fees from Bayer AG, outside

8

CC

the submitted work.

Acknowledgement of Any Grant Support

A

This study was funded by Pfizer, Inc. and Bristol-Myers Squibb Company

34

References 1.

Kearon C, Akl EA, Ornelas J, et al. Antithrombotic Therapy for VTE Disease: CHEST Guideline and Expert Panel Report. Chest. 2016;149(2):315-352 Members ATF, Konstantinides SV, Torbicki A, et al. 2014 ESC Guidelines on the

SC RI PT

2.

diagnosis and management of acute pulmonary embolism: The Task Force for the

Diagnosis and Management of Acute Pulmonary Embolism of the European Society of

Cardiology (ESC) Endorsed by the European Respiratory Society (ERS). European heart journal. 2014;35(43):3033-3073

van Es N, Coppens M, Schulman S, Middeldorp S, Buller HR. Direct oral anticoagulants

U

3.

N

compared with vitamin K antagonists for acute venous thromboembolism: evidence from

4.

M

A

phase 3 trials. Blood. 2014;124(12):1968-1975

Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with

Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus warfarin in nonvalvular atrial

TE

5.

D

atrial fibrillation. N Engl J Med. 2009;361(12):1139-1151

fibrillation. N Engl J Med. 2011;365(10):883-891 Zirlik A, Bode C. Vitamin K antagonists: relative strengths and weaknesses vs. direct oral

EP

6.

CC

anticoagulants for stroke prevention in patients with atrial fibrillation. Journal of thrombosis and thrombolysis. 2017;43(3):365-379

A

7.

8.

Agnelli G, Buller HR, Cohen A, et al. Oral apixaban for the treatment of acute venous thromboembolism. N Engl J Med. 2013;369(9):799-808 Botticelli Investigators WC, Buller H, Deitchman D, Prins M, Segers A. Efficacy and safety of the oral direct factor Xa inhibitor apixaban for symptomatic deep vein

35

thrombosis. The Botticelli DVT dose-ranging study. J Thromb Haemost. 2008;6(8):13131318 9.

Schulman S, Kearon C, Kakkar AK, et al. Dabigatran versus warfarin in the treatment of

10.

SC RI PT

acute venous thromboembolism. N Engl J Med. 2009;361(24):2342-2352 Schulman S, Kakkar AK, Goldhaber SZ, et al. Treatment of acute venous

thromboembolism with dabigatran or warfarin and pooled analysis. Circulation. 2014;129(7):764-772 11.

Hokusai VTEI, Buller HR, Decousus H, et al. Edoxaban versus warfarin for the treatment

Buller HR, Lensing AWA, Prins MH, et al. A dose-ranging study evaluating once-daily

N

12.

U

of symptomatic venous thromboembolism. N Engl J Med. 2013;369(15):1406-1415

A

oral administration of the factor Xa inhibitor rivaroxaban in the treatment of patients with

M

acute symptomatic deep vein thrombosis: the Einstein–DVT Dose-Ranging Study. Blood.

Investigators E, Bauersachs R, Berkowitz SD, et al. Oral rivaroxaban for symptomatic

TE

13.

D

2008;112(6):2242

venous thromboembolism. N Engl J Med. 2010;363(26):2499-2510 Investigators E-P, Buller HR, Prins MH, et al. Oral rivaroxaban for the treatment of

EP

14.

CC

symptomatic pulmonary embolism. N Engl J Med. 2012;366(14):1287-1297 15.

Agnelli G, Buller HR, Cohen A, et al. Apixaban for extended treatment of venous

A

thromboembolism. N Engl J Med. 2013;368(8):699-708

16.

Schulman S, Kearon C, Kakkar AK, et al. Extended Use of Dabigatran, Warfarin, or Placebo in Venous Thromboembolism. New England Journal of Medicine. 2013;368(8):709-718

36

17.

Buller HR. Once-Daily Oral Rivaroxaban Versus Placebo in the Long-Term Prevention of Recurrent Symptomatic Venous Thromboembolism. the Einstein-Extension Study. Blood. 2009;114(22):LBA-2 Linkins L, O'Donnell M, Julian JA, Kearon C. Intracranial and fatal bleeding according

SC RI PT

18.

to indication for long-term oral anticoagulant therapy. J Thromb Haemost. 2010;8(10):2201-2207 19.

Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for

systematic reviews and meta-analyses: the PRISMA statement. BMJ. 2009;339:b2535.

Higgins JP, Green S. Cochrane Handbook for Systematic Reviews of Interventions. Vol

U

20.

Puhan MA, Schunemann HJ, Murad MH, et al. A GRADE Working Group approach for

A

21.

N

4: John Wiley & Sons; 2011

M

rating the quality of treatment effect estimates from network meta-analysis. BMJ.

Alotaibi G, Alsaleh K, Wu C, McMurtry MS. Dabigatran, rivaroxaban and apixaban for

TE

22.

D

2014;349:g5630

extended venous thromboembolism treatment: network meta-analysis. Int Angiol.

Castellucci LA, Cameron C, Le Gal G, et al. Efficacy and safety outcomes of oral

CC

23.

EP

2014;33(4):301-308

anticoagulants and antiplatelet drugs in the secondary prevention of venous

A

thromboembolism: systematic review and network meta-analysis. BMJ. 2013;347:f5133

24.

Castellucci LA, Cameron C, Le Gal G, et al. Clinical and safety outcomes associated with treatment of acute venous thromboembolism: a systematic review and meta-analysis. JAMA. 2014;312(11):1122-1135

37

25.

Cohen AT, Hamilton M, Bird A, et al. Comparison of the Non-VKA Oral Anticoagulants Apixaban, Dabigatran, and Rivaroxaban in the Extended Treatment and Prevention of Venous Thromboembolism: Systematic Review and Network Meta-Analysis. PLoS One.

26.

SC RI PT

2016;11(8):e0160064 Hirschl M, Kundi M. New oral anticoagulants in the treatment of acute venous thromboembolism - a systematic review with indirect comparisons. Vasa. 2014;43(5):353-364 27.

Kang N, Sobieraj DM. Indirect treatment comparison of new oral anticoagulants for the

Rollins BM, Silva MA, Donovan JL, Kanaan AO. Evaluation of oral anticoagulants for

N

28.

U

treatment of acute venous thromboembolism. Thromb Res. 2014;133(6):1145-1151

A

the extended treatment of venous thromboembolism using a mixed-treatment comparison,

Sobieraj DM, Coleman CI, Pasupuleti V, Deshpande A, Kaw R, Hernandez AV.

D

29.

M

meta-analytic approach. Clinical therapeutics. 2014;36(10):1454-1464. e1453

TE

Comparative efficacy and safety of anticoagulants and aspirin for extended treatment of venous thromboembolism: A network meta-analysis. Thrombosis research.

Sterne JA, Bodalia PN, Bryden PA, et al. Oral anticoagulants for primary prevention,

CC

30.

EP

2015;135(5):888-896

treatment and secondary prevention of venous thromboembolic disease, and for

A

prevention of stroke in atrial fibrillation: systematic review, network meta-analysis and

31.

cost-effectiveness analysis. 2017 Linkins LA, Choi PT, Douketis JD. Clinical impact of bleeding in patients taking oral anticoagulant therapy for venous thromboembolism: a meta-analysis. Ann Intern Med. 2003;139(11):893-900 38

32.

Ruiz-Gimenez N, Suarez C, Gonzalez R, et al. Predictive variables for major bleeding events in patients presenting with documented acute venous thromboembolism. Findings from the RIETE Registry. Thromb Haemost. 2008;100(1):26-31 Carrier M, Le Gal G, Wells PS, Rodger MA. Systematic review: case-fatality rates of

SC RI PT

33.

recurrent venous thromboembolism and major bleeding events among patients treated for venous thromboembolism. Ann Intern Med. 2010;152(9):578-589 34.

Sen S, Dahlberg KW. Physician's fear of anticoagulant therapy in nonvalvular atrial fibrillation. Am J Med Sci. 2014;348(6):513-521

Cohen AT, Imfeld S, Rider T. Phase III trials of new oral anticoagulants in the acute

U

35.

N

treatment and secondary prevention of VTE: comparison and critique of study

Lip GY, Larsen TB, Skjøth F, Rasmussen LH. Indirect comparisons of new oral

M

36.

A

methodology and results. Adv Ther. 2014;31(5):473-493

D

anticoagulant drugs for efficacy and safety when used for stroke prevention in atrial

37.

TE

fibrillation. Journal of the American College of Cardiology. 2012;60(8):738-746 Mantha S, Ansell J. An indirect comparison of dabigatran, rivaroxaban and apixaban for

Skjoth F, Larsen TB, Rasmussen LH, Lip GY. Efficacy and safety of edoxaban in

CC

38.

EP

atrial fibrillation. Thromb Haemost. 2012;108(3):476-484

comparison with dabigatran, rivaroxaban and apixaban for stroke prevention in atrial

A

fibrillation. An indirect comparison analysis. Thromb Haemost. 2014;111(5):981-988

39.

Rasmussen LH, Larsen TB, Graungaard T, Skjøth F, Lip GY. Primary and secondary prevention with new oral anticoagulant drugs for stroke prevention in atrial fibrillation: indirect comparison analysis. Bmj. 2012;345:e7097

39

40.

Deitelzweig S, Keshishian A, Li X, et al. Comparison of effectiveness, safety, and the net clinical outcome between different direct oral anticoagulants in 162,707 non-valvular atrial fibrillation patients treated in us clinical practice. Journal of the American College

41.

SC RI PT

of Cardiology. 2018;71(11) Nikolakopoulou A, Mavridis D, Furukawa TA, et al. Living network meta-analysis

compared with pairwise meta-analysis in comparative effectiveness research: empirical study. BMJ. 2018;360:k585 42.

Catalá-López F, Tobías A, Cameron C, Moher D, Hutton B. Network meta-analysis for

U

comparing treatment effects of multiple interventions: an introduction. Rheumatology

Catalá-López F, Hutton B, Moher D. The transitive property across randomized

A

43.

N

international. 2014;34(11):1489-1496

M

controlled trials: if B is better than A, and C is better than B, will C be better than A?

Song F, Xiong T, Parekh-Bhurke S, et al. Inconsistency between direct and indirect

TE

44.

D

Revista espanola de cardiologia. 2014;67(08):597-602

comparisons of competing interventions: meta-epidemiological study. Bmj.

A

CC

EP

2011;343:d4909

40