efficacy and safety of the novel oral anticoagulants in...

Efficacy and Safety of the Novel Oral Anticoagulants inAtrial Fibrillation

A Systematic Review and Meta-Analysis of the Literature

Francesco Dentali, MD; Nicoletta Riva, MD; Mark Crowther, MD; Alexander G.G. Turpie, MD;Gregory Y.H. Lip, MD; Walter Ageno, MD

Background—Novel oral anticoagulants (NOACs) have been proposed as alternatives to vitamin K antagonists for theprevention of stroke and systemic embolism in patients with atrial fibrillation. Individually, NOACs were at leastnoninferior to vitamin K antagonists, but a clear superiority in overall and vascular mortality was notconsistently proven.

Methods and Results—We performed a meta-analysis of phase II and phase III randomized, controlled trials comparingNOACs with vitamin K antagonists in patients with atrial fibrillation. The MEDLINE and EMBASE databases,supplemented with conference abstract books and www.clinicaltrials.gov, were searched up to the first week of July2012 with no language restriction. Two reviewers performed independent article review and study quality assessment.Data on overall and cardiovascular mortality, stroke or systemic embolism, ischemic stroke, major and intracranialbleeding, and myocardial infarction were collected. NOACs were pooled to perform a comparison with vitamin Kantagonists, calculating pooled relative risks (RRs) and associated 95% confidence intervals (CIs). We retrieved 12studies (3 administering dabigatran, 4 administering rivaroxaban, 2 administering apixaban, and 3 administeringedoxaban) enrolling a total of 54 875 patients. NOACs significantly reduced total mortality (5.61% versus 6.02%; RR,0.89; 95% CI, 0.83–0.96), cardiovascular mortality (3.45% versus 3.65%; RR, 0.89; 95% CI, 0.82–0.98), andstroke/systemic embolism (2.40% versus 3.13%; RR, 0.77; 95% CI, 0.70–0.86). There was a trend toward reducedmajor bleeding (RR, 0.86; 95% CI, 0.72–1.02) with a significant reduction of intracranial hemorrhage (RR, 0.46; 95%CI, 0.39–0.56). No difference in myocardial infarction was observed.

Conclusions—NOACs are associated with an overall clinical benefit compared with vitamin K antagonists. Additional researchis required to confirm these findings outside the context of randomized trials. (Circulation. 2012;126:2381-2391.)

Key Words: anticoagulants � atrial fibrillation � meta-analysis � mortality � stroke

Atrial fibrillation (AF) is the most common cardiacarrhythmia, with an overall prevalence of 5.5% that

increases up to 17.8% in individuals �85 years of age.1 AF isa major risk factor for stroke,2 with a 30-day mortality rate of24% in the absence of treatment.3

Clinical Perspective on p 2391

Vitamin K antagonists (VKAs) are highly effective for theprevention of stroke, mainly of ischemic origin, in patientswith AF, resulting in a 64% risk reduction compared withplacebo and a 37% risk reduction compared with antiplatelettherapy.4 For this reason, VKAs are currently recommendedin all AF patients at moderate to high risk for stroke orsystemic embolism (SE).5 However, VKAs have significant

limitations, particularly their unpredictable anticoagulant re-sponse and numerous food and drug interactions, mandatingregular laboratory monitoring.6 These limitations make treat-ment with VKAs problematic for many patients; as a result,only about half of all potentially eligible AF patients aretreated with VKAs.6

Over the last several years, novel oral anticoagulant drugs(NOACs), including direct thrombin inhibitors and factor Xainhibitors, have been developed. These drugs have the poten-tial to address some of the limitations of VKAs.7 These agentshave fewer food and drug interactions and a more predictableanticoagulant effect, thus allowing fixed dosing without theneed for laboratory monitoring. Furthermore, their shorterhalf-life may produce additional advantages, eg, if temporary

Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.Received April 30, 2012; accepted September 14, 2012.From the Department of Clinical and Experimental Medicine, Insubria University, Varese, Italy (F.D., N.R., W.A.); University of Birmingham Centre

for Cardiovascular Sciences, Birmingham, UK (N.R., G.Y.H.L.); and Division of Hematology (M.C.) and Department of Medicine (A.G.G.T.), McMasterUniversity, Hamilton, Ontario, Canada.

The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.112.115410/-/DC1.

Correspondence to Francesco Dentali, UO Medicina Interna, Ospedale di Circolo, Viale Borri 57, 21100 Varese, Italy. E-mail [email protected]© 2012 American Heart Association, Inc.

Circulation is available at http://circ.ahajournals.org DOI: 10.1161/CIRCULATIONAHA.112.115410

2381

by guest on May 25, 2018

http://circ.ahajournals.org/D

ownloaded from

by guest on M

ay 25, 2018http://circ.ahajournals.org/

Dow

nloaded from



ownloaded from

by guest on M


Dow

nloaded from



ownloaded from

by guest on M


Dow

nloaded from



ownloaded from

by guest on M


Dow

nloaded from



ownloaded from

by guest on M


Dow

nloaded from



ownloaded from

by guest on M


Dow

nloaded from



ownloaded from

by guest on M


Dow

nloaded from



ownloaded from

by guest on M


Dow

nloaded from

www.clinicaltrials.gov

http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.112.115410/-/DC1

http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.112.115410/-/DC1

http://circ.ahajournals.org/
















interruption is required for a surgical procedure or in the caseof an hemorrhagic complication.

The NOACs have been compared with warfarin for theprevention of stroke and SE in patients with AF.8–10 Thesetrials have been favorable for the NOACs but have notconsistently demonstrated superiority over warfarin, particu-larly in terms of overall and vascular mortality. To betterassess the clinical benefit, we carried out a systematicreview of the literature and a meta-analysis of phase II andphase III randomized, clinical trials (RCTs) of these agentscompared with VKAs for the prevention of stroke or SE inpatients with AF.

MethodsA protocol for this review was prospectively developed that detailedthe specific objectives, criteria for study selection, approach to assessstudy quality, outcomes, and statistical methods. This protocol isavailable for review through the investigators.

Data Sources and SearchesWe identified all published studies that compared the risk ofthromboembolic and/or major bleeding (MB) events in AF patientsrandomized to VKAs (warfarin, phenprocoumon, acenocumarol,fluindione, and tecarfarin) or NOACs (dabigatran, AZD0837, sofi-gatran, rivaroxaban, apixaban, edoxaban, betrixaban, eribaxaban,LY517717, YM150, TAK442, and TTP889) using the MEDLINE(1966 to week 1 of July 2012) and the EMBASE (1980 to week 1 ofJuly 2012) electronic databases. The term ximelagatran was ex-cluded from the search because this drug has been withdrawn fromthe market. The search strategy was developed without any languagerestriction and used the medical subject headings and text wordspresented in Table 1 in the online-only Data Supplement. Wesupplemented our search by reviewing abstracts books from thecongresses of the International Society on Thrombosis and Haemo-stasis (2003–2011), European Society of Cardiology (2005–2011),American Society of Hematology (2004–2011), and AmericanCollege of Cardiology (2008–2011) and by manually reviewing thereference lists of all retrieved articles. We also searched on thewww.clinicaltrials.gov Web site to identify unpublished trials.

Study SelectionTwo reviewers (N.R. and F.D.) performed study selection indepen-dently, with disagreements solved through discussion and by theopinion of a third reviewer (W.A.) if necessary. Studies wereconsidered potentially eligible for this systematic review if they metthe following criteria: They were phase III RCTs or phase II RCTsincluding at least one of the evaluated dosages subsequently used inphase III trials; NOACs were compared with therapeutic doses ofVKAs in patients with AF; and thromboembolic and bleeding eventswere objectively assessed in both groups.

For trials that were reported in �1 publication, we extracted datafrom the most complete publication and used other publications toclarify data.

To assess the agreement between reviewers for study selection, weused the � statistic, which measures agreement beyond chance.11

Data Extraction and Quality AssessmentWe extracted and presented data according to the Providing Innova-tive Service Models and Assessment (PRISMA) criteria.12 Tworeviewers (N.R. and F.D.) independently extracted data on study(year of publication, design), population characteristics (number ofpatients, mean age, sex), and treatment (therapeutic indication, typeof drug, dose, and duration).

Information on the following outcomes was collected in the 2groups: number of total and ischemic strokes (ISs), SE, total andcardiovascular mortality, MB and intracranial bleeding, and myocar-dial infarctions (MIs). No attempt was made to reclassify bleeding

events. However, in the included studies, MBs were classifiedmostly according to International Society on Thrombosis and Hae-mostasis criteria13 as bleeding causing a fall in hemoglobin levels of�2 g/dL, bleeding leading to transfusion of �2 U whole blood or redcells, symptomatic bleeding in a critical area (intracranial, intraspi-nal, intraocular, retroperitoneal, intrarticular, pericardial, or intra-muscular with compartment syndrome), or bleeding events leadingto death.

If outcome data for extraction could not be identified, wecontacted the study authors by e-mail, with a reminder after 15 days.For unpublished trials, we also contacted the pharmaceuticalcompanies.

Two reviewers (N.R. and F.D.) independently assessed studyquality using a validated scale14 based on the following criteria:methods used to generate the randomization sequence, method ofdouble blinding, and description of patient withdrawals and drop-outs. A score of 1 point was given for each criterion satisfied, and 1additional point was given for high quality of randomization anddouble blinding, for a maximum of 5 points. Studies with a score �2were considered high quality, and studies with a score �2 wereconsidered low quality. Although concealed treatment allocation isnot part of this rating scale, it was included in our study qualityassessment.

We resolved disagreements about study data extraction by con-sensus or by discussion with a third reviewer (W.A.).

Statistical Analyses

Primary AnalysesWe determined pooled relative risks (RRs) and corresponding 95%confidence intervals (CIs) for all-cause and cardiovascular mortalityin AF patients who received VKAs or treatment with a NOAC.Furthermore, the pooled RR of any cardioembolic event (whichincluded stroke or SE), IS, MB and intracranial bleeding, and MI inthe 2 arms of treatment was calculated. Because transient ischemicattacks are frequently subjective, seldom consistently reported, andnot usually considered a primary outcome in AF trials, we decidednot to include them in our analysis.

Although we acknowledge that direct thrombin inhibitors andfactor Xa inhibitors have different pharmacodynamic and pharma-cokinetic properties, they all act on the final phase of the coagulationcascade, the common pathway, and represent as a whole the currentalternative to VKAs. For this reason, we considered it appropriate topool the NOACs together to perform a comparison with VKAs.

From phase II dose-ranging RCTs, we summed all groups whosetotal daily dose was equal to the regimens used in phase III RCTs.For all the treatment effects that were statistically significant, wedetermined the absolute risk reduction (ARR) or the absolute riskincrease and the corresponding number needed to treat (NNT) ornumber needed to harm. Data were pooled by use of a fixed-effectsmodel (Mantel-Haenszel method),15 and results were comparedwith the results obtained with a random-effects (RE) model(DerSimonian-Laird method).16 A value of P�0.05 was consideredstatistically significant. All analyses were performed with Review-Manager software (RevMan, version 5.1.6 for Windows; TheCochrane Collaboration, Oxford, UK; 2008). Because combiningtrials with extremely low or zero event rates can yield biased results,we repeated the analyses using Comprehensive Meta-Analysis soft-ware, version 2 (Biostat Software Corp, Englewood, NJ), whichprovides exact fixed-effect point and interval estimates for the oddsratio.17 The appropriateness of pooling data across studies wasassessed with the use of the Cochran Q and the I2 test forheterogeneity, which measures the inconsistency across study resultsand describes the proportion of total variation in study estimates thatis due to heterogeneity rather than sampling error.18 For the prepa-ration of the forest plots, we also used the Meta Data Viewersoftware version 1.02 (National Toxicology Program, ResearchTriangle Park, NC).19 The presence of publication bias was investi-gated by the use of funnel plots of effect size versus standard error.20

2382 Circulation November 13, 2012



ownloaded from



Sensitivity AnalysesWe repeated sensitivity analyses by using only studies that satisfiedeach item of our prespecified quality evaluation.21 Furthermore,separate analyses of studies published in peer-reviewed journalswere provided.

Subgroup AnalysesWe planned to perform separate analyses of studies assessingshort-term outcomes (1–3 months) and long-term outcomes (�1year).

We also performed separate analyses including, from phase IItrials, only the exact same dose regimen of the NOAC that wassubsequently used in phase III trials and, in a separate analysis,excluding the lower dosage of dabigatran (110 mg twice daily),which is not licensed by the US Food and Drug Administration.

ResultsStudy Identification and SelectionWe identified 1454 potentially relevant studies: 364 fromMEDLINE and 1090 from EMBASE. A total of 317 studieswere duplicated, and 1106 studies were excluded after titleand abstract screening. The remaining 31 studies were re-trieved in full for detailed evaluation. A list of the 24excluded studies and reasons for exclusion is available onrequest. Two additional studies were identified from a per-sonal library.22,23 We identified 3 more studies by searchingunpublished trials on www.clinicaltrials.gov.24–26 A review ofthe reference lists of included studies did not provide anyadditional references. Twelve studies were therefore includedin this systematic review.8–10,22–30 Interobserver agreementfor study selection was almost perfect (��0.89). The studyidentification and selection progression are summarized inFigure I in the online-only Data Supplement.

Data from 2 studies were supplemented with informationextracted from more recent publications.31,32 Supplementarydata for 7 trials were provided by the investigators involvedin the trial or the pharmaceutical companies.22–27,29 Data ondabigatran and rivaroxaban were also supplemented withinformation from their Food and Drug Administrationreviews.33,34

Study Characteristics and Study QualityBaseline characteristics of patients included in the studies aresummarized in the Table. For 4 NOACs (apixaban, dabigatran,edoxaban, and rivaroxaban), a phase III RCT was published orwas ongoing. The comparator VKA was warfarin in all studies.All studies were published in English. Study size ranged from100 patients26 to 18 201 patients,10 for a total of 54 875 includedpatients. Eight studies were phase II RCTs22,24–30 and 4 studieswere phase III RCTs.8–10,23 Three studies involved dabiga-tran,8,24,30 4 involved rivaroxaban,9,23,25,26 2 involved apixa-ban,10,27 and 3 involved edoxaban.22,28,29

Quality assessment items are summarized in Table II in theonline-only Data Supplement. With the use of the Jadadscore, 6 studies were classified as low quality22,24–27,30 and 6as high quality.8–10,23,28,29 All the trials were randomized, butthe method to generate the randomization sequence wasadequately reported in 4 studies. Three studies were double-blind and double-dummy. All studies provided a descriptionof patient withdrawals, and 6 reported concealed treatmentallocation.

OutcomesTable III in the online-only Data Supplement summarizes theoutcomes assessed during oral anticoagulant treatment. Whenpossible, we performed analyses using intention-to-treat pop-ulations. The intention-to-treat populations were not availablefor all outcomes for 2 studies9,23; therefore, we used thesafety-on-treatment populations.

Primary Analyses

Total MortalityDeath occurred in 1715 of 30 584 patients (5.61%) treatedwith NOACs and in 1416 of 23 531 patients (6.02%) treatedwith VKAs (Figure 1A). Use of NOACs was associated witha significant reduction in total death (RR, 0.89; 95% CI,0.83–0.96; I2�0%; ARR, 0.41%; NNT, 244). Reanalysiswith an RE model did not change these results.

Cardiovascular MortalityCardiovascular death occurred in 1054 of 30 584 patients(3.45%) treated with NOACs and in 858 of 23 531 patients(3.65%) treated with VKAs (Figure 1B). Use of NOACs wasassociated with a significant reduction in cardiovasculardeath (RR, 0.89; 95% CI, 0.82–0.98; I2�0%; ARR, 0.20%;NNT, 500). Reanalysis with an RE model did not changethese results.

Stroke or SE and ISStroke or SE occurred in 733 of 30 604 patients (2.40%)treated with NOACs and in 736 of 23 539 patients (3.13%)treated with VKAs (Figure 2A). Use of NOACs was associ-ated with a significant reduction in the composite outcome ofstroke or SE (RR, 0.77; 95% CI, 0.70–0.86; I2�0%; ARR,0.73%; NNT, 137). Reanalysis with an RE model did notchange these results.

Separate data for IS were available from 11 studies.8–10,22–28,30

IS occurred in 560 of 29 871 patients (1.87%) treated withNOACs and in 470 of 23 281 patients (2.02%) treated withVKAs (Figure 2B). Use of NOACs was associated with anonsignificant reduction in the number of ISs (RR, 0.92; 95%CI, 0.81–1.04; I2�22%). Reanalysis with an RE model didnot change these results.

MB and Intracranial BleedingThe definition of MB and the proportion of patients takingVKAs before inclusion and/or taking concomitant aspirinduring the trial were not declared in many studies and werevariable in the others (the Table).

MB, as defined by the study, occurred in 1498 of 30 599patients (4.90%) treated with NOACs and in 1304 of 23 548patients (5.54%) treated with VKAs (Figure 3A). Use ofNOACs was associated with a significant reduction in the rateof MB events (RR, 0.86; 95% CI, 0.80–0.93; ARR, 0.64%;I2�57%; NNT, 157). Reanalysis using an RE model demon-strated a trend toward a reduction in MB (RR, 0.86; 95% CI,0.72–1.02). Given the significant heterogeneity (P�0.05), weperformed a subgroup analysis examining each NOAC (Fig-ure II in the online-only Data Supplement). In this analysis,apixaban and dabigatran reduced MB events, whereas neitherrivaroxaban nor edoxaban reduced bleeding.

Dentali et al Novel Oral Anticoagulants in Atrial Fibrillation 2383



ownloaded from



Table. Characteristics of Included Studies

Study (Year) Study Design PopulationPatients

Randomized, n

Patients Characteristics*

Age, ySex,

M/F, % CHADS2 ScoreVKA at

Inclusion, %Concomitant

ASA, %

NCT0113640824

(2007)RCT phase II, Udabigatran and

warfarin

NVAF, �1 risk factor (HTN, DM,HF, previous stroke or TIA, age�75 y, CAD), Japanese patients

174 (53 and 59each dabigatran

group, 62 warfarin)

68.4�8.6 (mean�SD) 88.0/12.0 NA NA NA

PETRO30 (2007) RCT phase II, Bdabigatran, Uwarfarin and

aspirin

NVAF, �1 risk factor (HTN, DM,HF, previous stroke or TIA, age

�75 y)

502 (105, 166, 161each dabigatran

group, 70 warfarin)

70�8.3 (mean�SD) 81.9/18.1 3 (median) 100 36.1

RE-LY8 (2009) RCT phase III,B dabigatran, U

warfarin

NVAF, �1 risk factor (previousstroke or TIA, HF, age �75 y,age 65–74 y�DM or HTN or

CAD)

18 113 (6015 and6076 each

dabigatran group,6022 warfarin)

71 (mean) 63.6/36.4 2.1 (mean) 61.8 20.5

Weitz et al29

(2010)RCT phase II, B

edoxaban, Uwarfarin

NVAF, CHADS2 score �2 1146 (235, 245,235, 180 each

edoxaban group,251 warfarin)

65�8.7 (mean�SD) 62.1/37.9 2 (63.3% patients) 35.6 NA

Chung et al28


edoxaban, Uwarfarin

NVAF, CHADS2 score �1, Asianpatients

235 (79 and 80each edoxaban

group, 76 warfarin)

64.9�9.1, 65.9�7.7,64.5�9.5 (mean�SD

in each group)

65.4/34.6 2.0�1.10, 1.9�1.03,1.8�1.10 (mean�SD

in each group)

51.7 39.7

Yamashita etal22 (2012)

RCT phase II, Bedoxaban, U

warfarin

NVAF, CHADS2 score �1,Japanese patients

536 (135, 135, 132each edoxaban

group, 134warfarin)

69.4, 69.5, 68.4, 68.8(mean in each group)

82.5/17.5 1.9, 2.1, 2.1, 2.2(mean in each group)

84.8 25.1

ARISTOTLE-J27


apixaban, Uwarfarin

NVAF, �1 risk factor (age �75 y,HF, HTN, DM, previous stroke

or TIA), Japanese patients

222 (74 and 74each apixaban

group, 74 warfarin)

69.3, 70.7, 71.7(mean in each group)

82.9/17.1 1.8, 2.1, 1.9 (meanin each group)

85.3 24.8

ARISTOTLE10

(2011)RCT phase III,

B apixaban andwarfarin

NVAF or flutter, �1 risk factor(age �75 y, previous stroke or

TIA or SE, HF, DM, HTN)

18 201 (9120apixaban, 9081

warfarin)

70 (median) 64.7/35.3 2.1 (mean) 57.1 NA

NCT0097324525

(2008)RCT phase II, Urivaroxaban and

warfarin

NVAF age �60 y or �1 riskfactor (HTN, DM, CAD, HF)

Japanese patients

102 (26, 25, 24each rivaroxaban

group, 27 warfarin)

65.9�8.8, 65.7�8.2,67.4�7.2, 68.7�8.4(mean�SD in each

group)

78.4/21.6 NA NA NA

NCT0097332326

(2008)RCT phase II, Urivaroxaban and

warfarin

NVAF, age �60 y or �1 riskfactor (HTN, DM, CAD, HF),

Japanese patients

100 (24, 26, 24each rivaroxaban

group, 26 warfarin)

67.7�9.3, 70.5�9.5,66.3�11.2, 67.6�10.4(mean�SD in each

group)

80.0/20.0 NA NA NA

ROCKET-AF9

(2011)RCT phase III,B rivaroxabanand warfarin

NVAF, CHADS2 score �2 14 264 (7131rivaroxaban, 7133

warfarin)

73 (median) 60.3/39.7 3.5 (mean) 62.4 35.6

J-ROCKET-AF23

(2012)RCT phase III,B rivaroxabanand warfarin

NVAF; prior stroke, TIA, or SEor �2 risk factors (HF, HTN,age �75 y, DM), Japanese

patients

1280 (640rivaroxaban, 640

warfarin)

71.1�8.1 (mean�SD) 80.6/19.4 3.25 90.0 NA

ARISTOTLE indicates Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation; ASA, acetylsalicylic acid (aspirin); B, blinded; bid, twice daily; CAD,coronary artery disease; DM, diabetes mellitus; HF, heart failure; HTN, hypertension; INR, international normalized ratio; ISTH, International Society on Thrombosis andHaemostasis; J-ARISTOTLE, Japanese Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation; J-ROCKET-AF, An Efficacy and Safety Study ofRivaroxaban With Warfarin for the Prevention of Stroke and Non-Central Nervous System Systemic Embolism in Patients With Non-Valvular Atrial Fibrillation in Japan;NVAF, nonvalvular atrial fibrillation; od, once daily; RCT, randomized, controlled trial; RE-LY, Randomized Evaluation of Long-Term Anticoagulant Therapy; ROCKET-AF,An Efficacy and Safety Study of Rivaroxaban With Warfarin for the Prevention of Stroke and Non-Central Nervous System Systemic Embolism in Patients WithNon-Valvular Atrial Fibrillation; SE, systemic embolism; TIA, transient ischemic attack; TTR, time within therapeutic-range INR; U, unblinded; and VKA, vitamin Kantagonist. For the CHADS2 score, 1 point is given for cardiac heart failure, hypertension, age �75 years, and diabetes mellitus; 2 points are given for previous stroke,transient ischemic attack, and systemic embolism.

*When the characteristics of randomized patients were not available, we reported the characteristics of patients actually treated (studies in References 22–24,28–29).

†TTR was calculated in all studies according to the Rosendaal method but using different time frames.‡See the text for the ISTH definition of major bleeding.




ownloaded from


Intracranial bleeding occurred in 180 of 30 599 patients(0.59%) treated with NOACs and in 307 of 23 548 patients(1.30%) treated with VKAs (Figure 3B). Use of NOACs wasassociated with a significant reduction in the risk of intracra-nial bleeding events (RR, 0.46; 95% CI, 0.39–0.56; I2�34%;ARR, 0.71%; NNT, 141). Reanalysis with an RE model didnot change these results.

Myocardial InfarctionMI occurred in 394 of 30 584 patients (1.29%) treated withNOACs and in 304 of 23 531 patients (1.29%) treated withVKAs (Figure 4). There was no difference in the risk ofdeveloping MI between NOACs and VKAs (RR, 0.99;95% CI, 0.85–1.15; I2�55%). Reanalysis using an REmodel did not show significant changes in the results (RR,1.00; 95% CI, 0.75–1.33). Given the high level of hetero-

geneity, we performed a subgroup analysis separately toassess each novel drug, but we did not find any statisticallysignificant difference between the NOACs and VKAs(Figure III in the online-only Data Supplement).

Sensitivity AnalysesSensitivity analyses including only high-quality8–10,23,28.29

and published8–10,22.23,27–30 studies confirmed the results ofthe primary analyses (Figures IV–VII in the online-only DataSupplement).

Similarly, repeating our analyses with ComprehensiveMeta Analysis software including trials with extremely lowor zero event rates did not change the results of the primaryanalyses (results available on request).

Table. Continued

Intervention Drug Comparator Drug† Duration of Treatment Major Bleeding (Definition)‡

Dabigatran 110 mg bid, 150 mg bid Warfarin (INR range, 2.0–3.0 or 1.6–2.6 if age �70 y); TTR, NA 12 wk ISTH criteria

Dabigatran 50 mg bid, 150 mg bid,300 mg bid alone or combined with

ASA 81 or 325 mg

Warfarin (INR range, 2.0–3.0); TTR, 57.2% 12 wk ISTH criteria�bleeding eventsrequiring surgery

Dabigatran 110 mg bid, 150 mg bid Warfarin (INR range, 2.0–3.0 or 2.0–2.6 if Japanese age �70 y); TTR(mean), 64%

2.0 y (median) ISTH criteria�bleeding eventsrequiring inotropic agents or

surgery

Edoxaban 30 mg od, 30 mg bid,60 mg od, 60 mg bid

Warfarin (INR range, 2.0–3.0); TTR, 49.7% 12 wk ISTH criteria

Edoxaban 30 mg od, 60 mg od Warfarin (INR range, 2.0–3.0); TTR, 45.1% 3 mo Modified ISTH criteria (transfusion�800 mL packed red blood cells

or whole blood)

Edoxaban 30 mg od, 45 mg od,60 mg od

Warfarin (INR range, 2.0–3.0 if age �70 y or 1.6–2.6 if age �70 y);TTR, 73% for �70 y and 83% for age �70 y

12 wk Modified ISTH criteria (bloodtransfusion �4 U)

Apixaban 2.5 mg bid, 5 mg bid Warfarin (INR range, 2.0–3.0 if age �70 y or 2.0–2.6 if age �70 y);TTR, overall �60% of patients had INR within the 2.0–3.0 range for

60% of the treatment period regardless of age

12 wk ISTH criteria

Apixaban 5 mg bid or 2.5 mg bid(if �2 criteria: age �80 y, bodyweight �60 kg, serum creatinine

�1.5 mg/dL)

Warfarin (INR range, 2.0–3.0); TTR (mean), 62.2% 1.8 y (median) ISTH criteria (hemoglobin dropwas censored over a 24-h period)

Rivaroxaban 10 mg od, 15 mg od,20 mg od

Warfarin (INR range, 2.0–3.0 or 1.6–2.6 if age �70 y); TTR, NA 28 d ISTH criteria

Rivaroxaban 2.5 mg bid, 5 mg bid,10 mg bid

Warfarin (INR range, 2.0–3.0 or 1.6–2.6 if age �70 y); TTR, NA 28 d ISTH criteria

Rivaroxaban 20 mg od or 15 mg od(if creatinine clearance 30–49

mL/min)

Warfarin (INR range, 2.0–3.0); TTR (mean), 55% 590 d (median) ISTH criteria�bleeding eventsassociated with permanent

disability

Rivaroxaban 15 mg od or 10 mg od(if creatinine clearance 30–49

mL/min)

Warfarin (INR range, 2.0–3.0 if age �70 y or 1.6–2.6 if age �70 y);TTR, 65%

71 and 69 wk (median forrivaroxaban and warfarin,

respectively)

ISTH criteria




ownloaded from


Subgroup AnalysesIn the subgroup analysis that included only studies withshort-term follow-up,22,24–30 only the rate of stroke and SEappears to be significantly reduced in patients randomized toNOACs, whereas the rates of cardiovascular and total death,IS, MB, intracranial bleeding, and MI were similar in the 2groups. In contrast, the subgroup analyses that included onlystudies with long-term follow-up8–10,23 and only the doseregimens used in phase III studies provided the same resultsas the primary analyses.

The analysis excluding the lower dosage of dabigatranshowed a reduction of IS with NOACs compared with VKAs(RR, 0.86; 95% CI, 0.76–0.99; I2�30%) using a fixed-effects model. This result was not statistically significant withan RE model (RR, 0.85; 95% CI, 0.70–1.02; Figures IV–VIIin the online-only Data Supplement).

Publication BiasFunnel plots of effect size versus standard error are summa-rized in Figures VIII–XI in the online-only Data Supplement.The funnel plot for stroke or SE was asymmetrical with theabsence of studies on the right side of the plot, whereas thefunnel plot for MI was lacking studies on the left side ofthe plot. The funnel plots for total mortality and MB appearedsymmetrical.

DiscussionThis is, to the best of our knowledge, the first systematicreview and meta-analysis of phase II and phase III RCTs thatcompared the NOACs with warfarin for the prevention ofstroke and SE in patients with AF. Our analysis, incorporat-ing �50 000 patients, found a statistically significant 11%RR reduction in the incidence of both total mortality and

Figure 1. Total (A) and cardiovascular (B) mortality during oral anticoagulant treatment. NOAC indicates novel oral anticoagulant; VKA,vitamin K antagonists; M-H, Mantel-Haenszel; CI, confidence interval; RE-LY, Randomized Evaluation of Long-Term AnticoagulantTherapy; ARISTOTLE, Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation; J-ARISTOTLE, Japanese Apixabanfor the Prevention of Stroke in Subjects With Atrial Fibrillation; ROCKET-AF, An Efficacy and Safety Study of Rivaroxaban WithWarfarin for the Prevention of Stroke and Non-Central Nervous System Systemic Embolism in Patients With Non-Valvular AtrialFibrillation; and J-ROCKET-AF, An Efficacy and Safety Study of Rivaroxaban With Warfarin for the Prevention of Stroke and Non-Central Nervous System Systemic Embolism in Patients With Non-Valvular Atrial Fibrillation in Japan.




ownloaded from


cardiovascular mortality, which corresponds to an NNT of244 patients to prevent 1 death and to an NNT of 500 patientsto prevent 1 cardiovascular death. The observed advantage ofthe NOACs is consistent for all outcomes, including strokeand SE reduction (RR reduction, 23%; NNT, 137) and MBreduction (RR reduction, 14%; NNT, 157).

Following the favorable results of the individual clinicaltrials, dabigatran and rivaroxaban have received approvalfrom the regulatory agencies and apixaban is expected to belicensed in the near future. However, the cost-effectiveness ofthese compounds remains unclear. This lack of clarity exists,in part, as a result of the observation that single studies havereported small and often nonstatistically significant differ-ences between the NOACs and warfarin for hard end pointssuch as overall mortality and vascular mortality.

We believe that our study could provide more accurateestimates of the expected clinical benefits of the NOACs.Taken together, our results suggest that the use of the NOACs

not only provides practical advantages over the VKAs butalso is associated with an overall clinical benefit, suggestingtheir cost-effectiveness.

More important, the NOACs both reduce clinical eventsand offer the possibility of increasing the use of adequateprophylactic strategies in patients with AF. It is well knownthat AF remains a major cause of stroke2 and that the severityof stroke is greater in patients with AF than in othersubgroups.35 Despite this evidence, the use of VKAs in thereal world of patients with AF remains unacceptably low,with an overall prevalence of treated high-risk patients notexceeding 70%.36 The availability of NOACs has the poten-tial to reduce the incidence of AF-related strokes and SEbecause of both their superior efficacy and their potential tobe more widely used compared with the VKAs.

In this meta-analysis, we have combined the results ofclinical trials carried out with 4 NOACs. These drugs presentsome important differences in terms of mechanisms of action;

Figure 2. Stroke or systemic embolism (A) and ischemic stroke (B) during oral anticoagulant treatment. NOAC indicates novel oral anti-coagulant; VKA, vitamin K antagonists; M-H, Mantel-Haenszel; CI, confidence interval; RE-LY, Randomized Evaluation of Long-TermAnticoagulant Therapy; ARISTOTLE, Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation; J-ARISTOTLE, JapaneseApixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation; ROCKET-AF, An Efficacy and Safety Study of RivaroxabanWith Warfarin for the Prevention of Stroke and Non-Central Nervous System Systemic Embolism in Patients With Non-Valvular AtrialFibrillation; and J-ROCKET-AF, An Efficacy and Safety Study of Rivaroxaban With Warfarin for the Prevention of Stroke and Non-Central Nervous System Systemic Embolism in Patients With Non-Valvular Atrial Fibrillation in Japan.




ownloaded from


1 drug (dabigatran etexilate) is a direct thrombin inhibitor and3 drugs (rivaroxaban, apixaban, and edoxaban) are directfactor Xa inhibitors. Moreover, there are some differences inthe mechanisms of excretion, in their mean half-lives, and inthe drug-drug interactions, among others.37 These differencesmay suggest that combining the results of these drugs maynot be appropriate. However, there were no signs of hetero-geneity when the outcomes of total and cardiovascularmortality, stroke, or SE were analyzed, thus suggesting thatthe advantages in terms of efficacy are consistent among allthe new agents included in our study. On the other hand,significant heterogeneity was documented when the outcomeof MB was analyzed.

This finding may be due in part to some drug-specific orregimen-specific differences in terms of safety, although noneof the NOACs were less safe than warfarin, and to patient-

specific characteristics. Indeed, some studies enrolled anintermediate-risk population with a mean CHADS2 score of�2,8,10,22,27–29 whereas other studies enrolled a population athigher risk not only for thromboembolic but also for bleedingcomplications, with a mean CHADS2 score �3.9,23

When the analysis was repeated with the exclusion of thelower dose of dabigatran, which is not approved in the UnitedStates and is recommended for more fragile patients in othercountries, the results were fully comparable to those of themain analysis in terms of total mortality and safety, whereasa tendency toward a greater benefit in the reduction of IS wasobserved.

There is great interest in the potential increased risk of MIwith the use of the NOACs, particularly dabigatran. Theresults of a recent meta-analysis of trials involving dabigatranfor the primary and secondary prevention of cardiovascular

Figure 3. Major (A) and intracranial (B) bleeding during oral anticoagulant treatment. NOAC indicates novel oral anticoagulant; VKA,vitamin K antagonists; M-H, Mantel-Haenszel; CI, confidence interval; RE-LY, Randomized Evaluation of Long-Term AnticoagulantTherapy; ARISTOTLE, Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation; J-ARISTOTLE, Japanese Apixabanfor the Prevention of Stroke in Subjects With Atrial Fibrillation; ROCKET-AF, An Efficacy and Safety Study of Rivaroxaban WithWarfarin for the Prevention of Stroke and Non-Central Nervous System Systemic Embolism in Patients With Non-Valvular Atrial Fibril-lation; and J-ROCKET-AF, An Efficacy and Safety Study of Rivaroxaban With Warfarin for the Prevention of Stroke and Non-CentralNervous System Systemic Embolism in Patients With Non-Valvular Atrial Fibrillation in Japan.




ownloaded from


diseases found a 33% RR increase in acute coronary eventswith the novel direct thrombin inhibitors compared withtraditional anticoagulant drugs.38 Another recent analysiscomparing warfarin with other antithrombotic drugs (ximela-gatran, dabigatran, idraparinux, and clopidogrel) in AF clin-ical trials found a 23% RR reduction in the rate of MI withwarfarin, suggesting the protective effect of the VKA.39 Inthis meta-analysis, we failed to detect any difference in theoverall risk of MI, with a 1.29% rate in both NOAC- andVKA-treated patients. Of interest, no statistically significantdifference was detected after subgroup analysis.

The strengths of this study include the rigorous method-ological approach, the selection of all the studies performedwith the 4 NOACs considered, and the consistency of theresults of sensitivity analyses. Furthermore, to the greatestextent possible, we confined our analysis to clinically rele-vant events, and because all the studies were performed as acomponent of product registration, it is likely that all reportedoutcome events were objectively confirmed.

The study has a number of limitations. First, because this wasa study-level meta-analysis, we were unable to confirm theoverall results in specific subgroups of patients according to theirbaseline stroke or bleeding risk. Population characteristics werequite different among the single studies; however, severalsubgroup analyses have already been published suggesting theconsistency of the principal findings across different subgroupssuch as patients with previous stroke31,40.41 or more advancedage.42 Second, we could not compare the patients receivingNOACs with different subgroups of warfarin-treated patientsaccording to the time in therapeutic range, and it has been shownthat the magnitude of the benefit of the NOACs compared withstandard treatment is dependent on the quality of control ofwarfarin.43 Third, the results of our meta-analysis are drivenmainly by 3 large RCTs involving dabigatran,8 rivaroxaban,9

and apixaban,10 whereas fewer data are available on edoxaban

because the phase III RCT is currently ongoing. Fourth, thefunnel plot for stroke or SE was asymmetrical with a lack ofstudies on the right part of the plot, suggesting that unpublishedstudies likely to demonstrate an increased risk of stroke or SEwith NOACs were not included in our meta-analysis. Instead,the funnel plot for MI was lacking studies on the left side of theplot, suggesting that studies demonstrating a reduction in the riskof MI with NOACs were not included. However, because weperformed an extensive research of the literature, includingabstracts presented at congresses of several international societ-ies, and because we contacted pharmaceutical companies askingfor unpublished trials, the existence of other trials not included inour systematic review is extremely unlikely.

ConclusionsNOACs reduced overall and cardiovascular mortality, strokeand SE, and MB and intracranial bleeding compared withwarfarin. These favorable efficacy and safety profiles nowneed to be confirmed in postmarketing studies.

AcknowledgmentsWe are grateful to Drs Jeffrey Weitz, Masatsugu Hori, AndreasClemens, Mariko Kajikawa, and Kazuhiro Kanmuri for kindlyproviding additional data.

DisclosuresDr Crowther discloses having served on advisor boards for LeoPharma, Pfizer, Bayer, Boehringer Ingelheim, Alexion, CSL Beh-ring, and Artisan Pharma. Dr Crowther has prepared educationalmaterials for Pfizer, Octapharm, and CSL Behring; has providedexpert testimony for Bayer; and holds a Career Investigator Awardfrom the Heart and Stroke Foundation of Ontario and the LeoPharma Chair in Thromboembolism Research at McMaster Univer-sity. Dr Crowther’s institution has received funding for research projectsfrom Boehringer Ingelheim, Octapharm, Pfizer, and Leo Pharma. DrLip has served as a consultant for Bayer, Astellas, Merck, AstraZeneca,Sanofi, BMS/Pfizer, Biotronik, Portola, and Boehringer Ingelheim andhas been on the speakers’ bureau for Bayer, BMS/Pfizer, Boehringer

Figure 4. Myocardial infarction during oral anticoagulant treatment. NOAC indicates novel oral anticoagulant; VKA, vitamin K antago-nists; M-H, Mantel-Haenszel; CI, confidence interval; RE-LY, Randomized Evaluation of Long-Term Anticoagulant Therapy; ARISTOTLE,Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation; J-ARISTOTLE, Japanese Apixaban for the Prevention of Strokein Subjects With Atrial Fibrillation; ROCKET-AF, An Efficacy and Safety Study of Rivaroxaban With Warfarin for the Prevention of Strokeand Non-Central Nervous System Systemic Embolism in Patients With Non-Valvular Atrial Fibrillation; and J-ROCKET-AF, An Efficacyand Safety Study of Rivaroxaban With Warfarin for the Prevention of Stroke and Non-Central Nervous System Systemic Embolism inPatients With Non-Valvular Atrial Fibrillation in Japan.




ownloaded from


Ingelheim, and Sanofi-Aventis. Dr Ageno has served on the advisoryboards for Bayer, BMS/Pfizer, and Daiichi Sankyo; has receivedhonoraria for speaking activities from Boehringer Ingelheim, Bayer,BMS, Pfizer, Sanofi, GlaxoSmithKline; and has received funding forresearch projects from Bayer, GlaxoSmithKline, and Alexion. Dr Turpiehas served on advisory boards for Bayer, Astellas, and Takeda and hasreceived honoraria for speaking activities from Boehringer Ingelheim,Bayer, BMS, Pfizer, Sanofi, and GlaxoSmithKline. Drs Dentali andRiva report no conflicts.

References1. Heeringa J, van der Kuip DA, Hofman A, Kors JA, van Herpen G,

Stricker BH, Stijnen T, Lip GY, Witteman JC. Prevalence, incidence andlifetime risk of atrial fibrillation: the Rotterdam study. Eur Heart J. 2006;27:949–953.

2. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independentrisk factor for stroke: the Framingham study. Stroke. 1991;22:983–988.

3. Hylek EM, Go AS, Chang Y, Jensvold NG, Henault LE, Selby JV, SingerDE. Effect of intensity of oral anticoagulation on stroke severity andmortality in atrial fibrillation. N Engl J Med. 2003;349:1019–1026.

4. Hart RG, Pearce LA, Aguilar MI. Meta-analysis: antithrombotic therapyto prevent stroke in patients who have nonvalvular atrial fibrillation. AnnIntern Med. 2007;146:857–867.

5. Camm AJ, Kirchhof P, Lip GY, Schotten U, Savelieva I, Ernst S, VanGelder IC, Al-Attar N, Hindricks G, Prendergast B, Heidbuchel H, AlfieriO, Angelini A, Atar D, Colonna P, De Caterina R, De Sutter J, Goette A,Gorenek B, Heldal M, Hohloser SH, Kolh P, Le Heuzey JY, PonikowskiP, Rutten FH. Guidelines for the management of atrial fibrillation: theTask Force for the Management of Atrial Fibrillation of the EuropeanSociety of Cardiology (ESC). Eur Heart J. 2010;31:2369–2429.

6. Ansell J, Hirsh J, Hylek E, Jacobson A, Crowther M, Palareti G. Phar-macology and management of the vitamin K antagonists: AmericanCollege of Chest Physicians Evidence-Based Clinical Practice Guidelines(8th Edition). Chest. 2008;133:160S–198S.

7. Ahrens I, Lip GY, Peter K. New oral anticoagulant drugs in cardiovas-cular disease. Thromb Haemost. 2010;104:49–60.

8. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A,Pogue J, Reilly PA, Themeles E, Varrone J, Wang S, Alings M, XavierD, Zhu J, Diaz R, Lewis BS, Darius H, Diener HC, Joyner CD, WallentinL; RE-LY Steering Committee and Investigators. Dabigatran versuswarfarin in patients with atrial fibrillation [published correction appearsin N Engl J Med. 2010;363:1877]. N Engl J Med. 2009;361:1139–1151.

9. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, BreithardtG, Halperin JL, Hankey GJ, Piccini JP, Becker RC, Nessel CC, PaoliniJF, Berkowitz SD, Fox KA, Califf RM; ROCKET AF Investigators.Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N EnglJ Med. 2011;365:883–891.

10. Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, HannaM, Al-Khalidi HR, Ansell J, Atar D, Avezum A, Bahit MC, Diaz R,Easton JD, Ezekowitz JA, Flaker G, Garcia D, Geraldes M, Gersh BJ,Golitsyn S, Goto S, Hermosillo AG, Hohnloser SH, Horowitz J, MohanP, Jansky P, Lewis BS, Lopez-Sendon JL, Pais P, Parkhomenko A,Verheugt FW, Zhu J, Wallentin L; ARISTOTLE Committees and Inves-tigators. Apixaban versus warfarin in patients with atrial fibrillation.N Engl J Med. 2011;365:981–992.

11. Landis JR, Koch GG. The measurement of observer agreement for cate-gorical data. Biometrics. 1977;33:159–174.

12. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferredreporting items for systematic reviews and meta-analyses: the PRISMAstatement. Ann Intern Med. 2009;151:264–269.

13. Schulman S, Kearon C; Subcommittee on Control of Anticoagulation ofthe Scientific and Standardization Committee of the International Societyon Thrombosis and Haemostasis. Definition of major bleeding in clinicalinvestigations of antihemostatic medicinal products in non-surgicalpatients. J Thromb Haemost. 2005;3:692–694.

14. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, GavaghanDJ, McQuay HJ. Assessing the quality of reports of randomized clinicaltrials: is blinding necessary? Control Clin Trials. 1996;17:1–12.

15. Mantel N, Haenszel W. Statistical aspects of the analysis of data fromretrospective studies of disease. J Natl Cancer Inst. 1959;22:719–748.

16. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control ClinTrials. 1986;7:177–188.

17. Sweeting MJ, Sutton AJ, Lambert PC. What to add to nothing? Use andavoidance of continuity corrections in meta-analysis of sparse data. StatMed. 2004;23:1351–1375.

18. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsis-tency in meta-analyses. BMJ. 2003;327:557–560.

19. Boyles AL, Harris SF, Rooney AA, Thayer KA. Forest Plot Viewer: anew graphing tool. Epidemiology. 2011;22:746–747.

20. Sterne JA, Egger M, Smith GD. Systematic reviews in health care:investigating and dealing with publication and other biases in meta-anal-ysis. BMJ. 2001;323:101–105.

21. Juni P, Witschi A, Bloch R, Egger M. The hazards of scoring the qualityof clinical trials for meta-analysis. JAMA. 1999;282:1054–1060.

22. Yamashita T, Koretsune Y, Yasaka M, Inoue H, Kawai Y, Yamaguchi T,Uchiyama S, Matsumoto M, Ogawa S. Randomized, multicenter,warfarin-controlled phase II study of edoxaban in Japanese patients withnon-valvular atrial fibrillation. Circ J. 2012;76:1840–1847.

23. Hori M, Matsumoto M, Tanahashi N, Momomura SI, Uchiyama S, GotoS, Izumi T, Koretsune Y, Kajikawa M, Kato M, Ueda H, Iwamoto K,Tajiri M; J-ROCKET AF Study Investigators. Rivaroxaban vs. warfarinin Japanese patients with atrial fibrillation. Circ J. 2012;76:2104–2111.

24. A dose response study of dabigatran etexilate (BIBR 1048) in pharma-codynamics and safety in patients with non-valvular atrial fibrillation incomparison to warfarin. http://www.clinicaltrials.gov/ct2/show/NCT01136408. Accessed September 2, 2011.

25. BAY59-7939 in Atrial Fibrillation Once Daily (OD). http://www.clinicaltrials.gov/ct2/show/NCT00973245. Accessed September 2, 2011.

26. BAY59-7939 Japanese in Atrial Fibrillation (2nd). http://www.clinicaltrials.gov/ct2/show/NCT00973323. Accessed September 2, 2011.

27. Ogawa S, Shinohara Y, Kanmuri K. Safety and efficacy of the oral directfactor Xa inhibitor apixaban in Japanese patients with non-valvular atrialfibrillation: the ARISTOTLE-J study. Circ J. 2011;75:1852–1859.

28. Chung N, Jeon HK, Lien LM, Lai WT, Tse HF, Chung WS, Lee TH,Chen SA. Safety of edoxaban, an oral factor Xa inhibitor, in Asianpatients with non-valvular atrial fibrillation. Thromb Haemost. 2011;105:535–544.

29. Weitz JI, Connolly SJ, Patel I, Salazar D, Rohatagi S, Mendell J,Kastrissios H, Jin J, Kunitada S. Randomised, parallel-group, multicentre,multinational phase 2 study comparing edoxaban, an oral factor Xainhibitor, with warfarin for stroke prevention in patients with atrial fibril-lation. Thromb Haemost. 2010;104:633–641.

30. Ezekowitz MD, Reilly PA, Nehmiz G, Simmers TA, Nagarakanti R,Parcham-Azad K, Pedersen KE, Lionetti DA, Stangier J, Wallentin L.Dabigatran with or without concomitant aspirin compared with warfarinalone in patients with nonvalvular atrial fibrillation (PETRO Study). Am JCardiol. 2007;100:1419–1426.

31. Easton JD, Lopes RD, Bahit MC, Wojdyla DM, Granger CB, WallentinL, Alings M, Goto S, Lewis BS, Rosenqvist M, Hanna M, Mohan P,Alexander JH, Diener HC; ARISTOTLE Committees and Investigators.Apixaban compared with warfarin in patients with atrial fibrillation andprevious stroke or transient ischaemic attack: a subgroup analysis of theARISTOTLE trial. Lancet Neurol. 2012;11:503–511.

32. Hart RG, Diener HC, Yang S, Connolly SJ, Wallentin L, Reilly PA,Ezekowitz MD, Yusuf S. Intracranial hemorrhage in atrial fibrillationpatients during anticoagulation with warfarin or dabigatran: the RE-LYtrial. Stroke. 2012;43:1511–1517.

33. Center for Drug Evaluation and Research. Medical review. ApplicationNo. 22-512. http://www.accessdata.fda.gov/drugsatfda_docs/nda/2010/022512Orig1s000MedR.pdf. Accessed May 18, 2011.

34. Center for Drug Evaluation and Research. Medical review. ApplicationNo. 202439. http://www.accessdata.fda.gov/drugsatfda_docs/nda/2011/202439Orig1s000MedR.pdf. Accessed July 19, 2012.

35. Steger C, Pratter A, Martinek-Bregel M, Avanzini M, Valentin A, SlanyJ, Stollberger C. Stroke patients with atrial fibrillation have a worseprognosis than patients without: data from the Austrian Stroke Registry.Eur Heart J. 2004;25:1734–1740.

36. Ogilvie IM, Newton N, Welner SA, Cowell W, Lip GY. Underuse of oralanticoagulants in atrial fibrillation: a systematic review. Am J Med.2010;123:638–645.e4.

37. Ageno W, Gallus AS, Wittkowsky A, Crowther M, Hylek EM, Palareti G.Oral anticoagulant therapy: Antithrombotic Therapy and Prevention ofThrombosis, 9th Edition: American College of Chest PhysiciansEvidence-Based Clinical Practice Guidelines. Chest. 2012;141:e44S–e88S.




ownloaded from

http://www.clinicaltrials.gov/ct2/show/NCT01136408






http://www.accessdata.fda.gov/drugsatfda_docs/nda/2010/022512Orig1s000MedR.pdf





38. Uchino K, Hernandez AV. Dabigatran association with higher risk ofacute coronary events: meta-analysis of noninferiority randomized con-trolled trials. Arch Intern Med. 2012;172:397–402.

39. Lip GY, Lane DA. Does warfarin for stroke thromboprophylaxis protectagainst MI in atrial fibrillation patients? Am J Med. 2010;123:785–789.

40. Diener HC, Connolly SJ, Ezekowitz MD, Wallentin L, Reilly PA, YangS, Xavier D, Di Pasquale G, Yusuf S; RE-LY Study Group. Dabigatrancompared with warfarin in patients with atrial fibrillation and previoustransient ischaemic attack or stroke: a subgroup analysis of the RE-LYtrial. Lancet Neurol. 2010;9:1157–1163.

41. Hankey GJ, Patel MR, Stevens SR, Becker RC, Breithardt G, Carolei A,Diener HC, Donnan GA, Halperin JL, Mahaffey KW, Mas JL, MassaroA, Norrving B, Nessel CC, Paolini JF, Roine RO, Singer DE, Wong L,Califf RM, Fox KA, Hacke W; ROCKET AF Steering Committee Inves-tigators. Rivaroxaban compared with warfarin in patients with atrial

fibrillation and previous stroke or transient ischaemic attack: a subgroupanalysis of ROCKET AF. Lancet Neurol. 2012;11:315–322.

42. Eikelboom JW, Wallentin L, Connolly SJ, Ezekowitz M, Healey JS,Oldgren J, Yang S, Alings M, Kaatz S, Hohnloser SH, Diener HC,Franzosi MG, Huber K, Reilly P, Varrone J, Yusuf S. Risk of bleedingwith 2 doses of dabigatran compared with warfarin in older and youngerpatients with atrial fibrillation: an analysis of the Randomized Evaluationof Long-Term Anticoagulant Therapy (RE-LY) trial. Circulation. 2011;123:2363–2372.

43. Wallentin L, Yusuf S, Ezekowitz MD, Alings M, Flather M, FranzosiMG, Pais P, Dans A, Eikelboom J, Oldgren J, Pogue J, Reilly PA, YangS, Connolly SJ, on behalf of the RE-LY investigators. Efficacy and safetyof dabigatran compared with warfarin at different levels of internationalnormalised ratio control for stroke prevention in atrial fibrillation: ananalysis of the RE-LY trial. Lancet. 2010;376:975–983.

CLINICAL PERSPECTIVENovel oral anticoagulants, including direct thrombin inhibitors and factor Xa inhibitors, have been proposed as alternativesto vitamin K antagonists for the prevention of stroke and systemic embolism in patients with atrial fibrillation. Weconducted a systematic review and a meta-analysis of published and unpublished phase II and III randomized, controlledtrials comparing the novel oral anticoagulants with the vitamin K antagonists in patients with atrial fibrillation. Weretrieved 12 studies (3 administering dabigatran, 4 administering rivaroxaban, 2 administering apixaban, and 3administering edoxaban) enrolling �50 000 patients. We found that the novel oral anticoagulants significantly reducedoverall mortality, cardiovascular mortality, and stroke or systemic embolism. Furthermore, these drugs showed a trendtoward reduced major bleeding, with a significant reduction of intracranial hemorrhage. The novel oral anticoagulants maypotentially address some of the limitations of vitamin K antagonists because they have fewer food and drug interactionsand a more predictable anticoagulant effect, thus allowing fixed-dose regimens without the need for routine laboratorymonitoring. Our meta-analysis also provided robust evidence on the overall clinical benefit of the novel oral anticoagulants,suggesting their cost-effectiveness in the real-life healthcare systems.

Go to http://cme.ahajournals.org to take the CME quiz for this article.




ownloaded from


and Walter AgenoFrancesco Dentali, Nicoletta Riva, Mark Crowther, Alexander G.G. Turpie, Gregory Y.H. Lip

Review and Meta-Analysis of the LiteratureEfficacy and Safety of the Novel Oral Anticoagulants in Atrial Fibrillation: A Systematic

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 2012 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/CIRCULATIONAHA.112.115410

2012;126:2381-2391; originally published online October 15, 2012;Circulation.

http://circ.ahajournals.org/content/126/20/2381World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://circ.ahajournals.org/content/suppl/2012/10/15/CIRCULATIONAHA.112.115410.DC1Data Supplement (unedited) at:

http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer this process is available in the

click Request Permissions in the middle column of the Web page under Services. Further information aboutOffice. Once the online version of the published article for which permission is being requested is located,

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the EditorialCirculationin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:



ownloaded from

http://circ.ahajournals.org/content/126/20/2381

http://circ.ahajournals.org/content/suppl/2012/10/15/CIRCULATIONAHA.112.115410.DC1

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://circ.ahajournals.org//subscriptions/


SUPPLEMENTAL MATERIAL

Supplemental Table 1: search strategy EMBASE from 1980 to 2012 week 27

1. exp heart atrium fibrillation/ (60418) 2. atrial fibrillation.tw. (45976) 3. exp heart atrium flutter/ (7772) 4. atrial flutter.tw. (5194) 5. or/1-4 (70227) 6. (dabigatran or BIBR1048 or BIBR-1048 or "BIBR 1048").mp. (2396) 7. (AZD-0837 or AZD0837 or "AZD 0837" or AR-H067637 or ARH067637 or "AR

H067637").mp. (115) 8. (sofigatran or "MCC 977" or MCC-977 or MCC977).mp. (12) 9. (rivaroxaban or "BAY 59 7939" or "BAY 59-7939" or "BAY 597939" or BAY59-7939 or

BAY597939).mp. (1971) 10. (apixaban or BMS-562247 or BMS562247 or "BMS 562247").mp. (1116) 11. (edoxaban or DU-176b or DU176b or "DU 176b").mp. (390) 12. (betrixaban or PRT054021 or PRT-054021 or "PRT 054021" or MK-4448 or MK4448 or "MK

4448").mp. (184) 13. (eribaxaban or "PD 0348292" or PD-0348292 or PD0348292).mp. (36) 14. (LY517717 or LY-517717 or "LY 517717").mp. (149) 15. (YM150 or YM-150 or "YM 150").mp. (196) 16. (TAK-442 or TAK442 or "TAK 442").mp. (62) 17. (TTP889 or TTP-889 or "TTP 889").mp. (41) 18. or/6-17 (3528) 19. exp antivitamin K/ (4785) 20. ("vitamin K antagonis$" or "antivitamin K").tw. (2434) 21. exp coumarin derivative/ (77527) 22. warfarin.tw. or warfarin/ (54300) 23. (phenprocoumon or phenprocumon).tw. or phenprocoumon/ (3594) 24. (acenocumarol or acenocoumarol).tw. or acenocoumarol/ (4114) 25. (fluindione or phenindione or anisindione).mp. (1593) 26. (tecarfarin or ATI-5923 or ATI5923 or "ATI 5923").mp. (50) 27. or/19-26 (83443) 28. 5 and 18 and 27 (1290) 29. limit 28 to conference abstract (136) 30. limit 28 to conference paper (64) 31. 28 not 29 not 30 (1090)

Supplemental Table 2. Quality of included studies

Study Trial described as randomized

Adequate description of

method of generate

randomization sequence

Trial described as double

blind

Adequate description

of method of double

blinding

Description of

withdrawal and

dropouts

Sum Jadad’s score

Concealed treatment allocation

NCT 01136408 (24)

Yes No No No Yes 2 Unclearly concealed

PETRO (30)


RE-LY (8)

Yes Yes No No Yes 3 Adequately concealed

Weitz et al. (29)


Chung et al. (28)


Yamashita et al. (22)


ARISTOTLE-J (27)


ARISTOTLE (10)

Yes No Yes Yes Yes 4 Adequately concealed

NCT 00973245 (25)


NCT 00973323 (26)


ROCKET-AF (9)

Yes Yes Yes Yes Yes 5 Adequately concealed

J-ROCKET-AF (23)

Yes No Yes Yes Yes 4 Adequately concealed

Supplemental Table 3. Outcomes examined during oral anticoagulant treatment, comparing new oral anticoagulants with warfarin in atrial fibrillation

Study

Total Death

(n/N)

CV death

(n/N)

Stroke or SE

(n/N)

Ischemic stroke

(n/N)

Major bleeding

(n/N)

Intracranial bleeding

(n/N)

Myocardial infarction

(n/N)

NOAC W NOAC W NOAC W NOAC W NOAC W NOAC W NOAC W

NCT 01136408 (24) *

D 110 bd

0 / 46

0 / 62

0 / 46

0 / 62

0 / 46

1 / 62

0 / 46

1 / 62

0 / 46 1 / 62

***

0 / 46

0 / 62

0 / 46

0 / 62 D 150

bd 0 / 58 0 / 58 0 / 58 0 / 58 1 / 58 0 / 58 0 / 58

PETRO (30) ** D 150

bd 0 / 166 0 / 70 0 / 166 0 / 70 0 / 166 0 / 70 0 / 166 0 / 70 0 / 166 0 / 70 0 / 166 0 / 70 0 / 166 0 / 70

RE-LY (8)

D 110 bd

446 / 6015

487 / 6022

289 / 6015

317 / 6022

183 / 6015

202 / 6022

152 / 6015

134 / 6022

342 / 6015

421 / 6022

27 / 6015

90 / 6022

98 / 6015

75 / 6022

D 150 bd

438 / 6076

274 / 6076

134 / 6076

103 / 6076

399 / 6076

37 / 6076

97 / 6076

Weitz et al. (29) *

E 30 od

6 / 235

3 / 250

2 / 235

2 / 250

1 / 235

3 / 250

NA

NA

0 / 235

1 / 250

0 / 235

0 / 250

2 / 235

0 / 250 E 30 bd

4 / 244 4 / 244 3 / 244 NA 5 / 244 2 / 244 1 / 244

E 60 od

1 / 234 0 / 234 1 / 234 NA 1 / 234 1 / 234 2 / 234

Chung et al. (28)

E 30 od

0 / 79

1 / 75

0 / 79

0 / 75

0 / 79

0 / 75

0 / 79

0 / 75

0 / 79

2 / 75

0 / 79

0 / 75

0 / 79

0 / 75 E 60 od

1 / 80 1 / 80 0 / 80 0 / 80 0 / 80 0 / 80 0 / 80

Yamashita et al.

(22) *

E 30 od

0 / 130

1 / 125

0 / 130

0 / 125

0 / 130

0 / 125

0 / 130

0 / 125

0 / 130

0 / 125

0 / 130

0 / 125

0 / 130

0 / 125 E 60 od

1 / 130 0 / 130 0 / 130 0 / 130 2 / 130 1 / 130 0 / 130

ARISTOTLE-J (27) *

A 2.5 bd

0 / 74 0 / 74

0 / 74 0 / 74

0 / 74 3 / 74

0 / 74 2 / 74

0 / 72 1 / 75

0 / 72 1 / 75

0 / 74 0 / 74

A 5 bd 0 / 74 0 / 74 0 / 74 0 / 74 0 / 71 0 / 71 0 / 74

ARISTOTLE (10) A 2.5

or 5 bd 603 / 9120

669 / 9081

308 / 9120

344 / 9081

212 / 9120

265 / 9081 149 / 9120

155 / 9081

327 / 9088

462 / 9052

52 / 9088

122 / 9052

90 / 9120

102 / 9081

NCT 00973245 (25) *

R 10 od

0 / 26

0 / 27

0 / 26

0 / 27

0 / 26

0 / 27

0 / 26

0 / 27

0 / 26

0 / 27

0 / 26

0 / 27

0 / 26

0 / 27 R 15 od

0 / 25 0 / 25 0 / 25 0 / 25 0 / 25 0 / 25 0 / 25

R 20 od

0 / 24 0 / 24 0 / 24 0 / 24 0 / 24 0 / 24 0 / 24

NCT 00973323 (26) *

R 5 bd 0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26

0 / 26 R 10 bd

0 / 24 0 / 24 0 / 24 0 / 24 0 / 24 0 / 24 0 / 24

ROCKET-AF

(9)

R 15 or 20 od

208 / 7061

****

250 / 7082

****

170 / 7061

****

193 / 7082

****

188 / 7081

240 / 7090

149 / 7061

****

161 / 7082

****

395 / 7111

386 / 7125

55 / 7111

84 / 7125

101 / 7061

****

126 / 7082

****

J-ROCKET-AF (23) *

R 10 or 15 od

7 / 637

****

5 / 637

****

6 / 637

****

2 / 637

****

11 / 637

****

22 / 637

****

7 / 637

****

17 / 637

**** 26 / 639 30 / 639 5 / 639 10 / 639

3 / 637

****

1 / 637

****

Legend:

A = apixaban, CV = cardiovascular, D = dabigatran, E = edoxaban, NOAC = new oral anticoagulant drug, R = rivaroxaban, SE = systemic embolism, W = warfarin

* Additional information were provided by the investigators involved in the trial or the pharmaceutical companies

** Given some discrepancies in the number of patients analyzed for bleeding and thromboembolic events in the original publication, we considered the number of patients randomized to dabigatran 150 mg twice daily in the absence of a reply from the authors

*** We did not consider the ischemic stroke with haemorrhagic transformation, reported in the bulletin of the pharmaceutical company as major bleeding, and we therefore counted only 1 major bleeding in the warfarin group

**** Events actually occurred during treatment were available only for the safety on-treatment population (9) or the per-protocol population on treatment (23)

Supplemental Figure 1. Studies selection progression

Supplemental Figure 2. Subgroup analysis of major bleeding according to different new oral anticoagulants

Supplemental Figure3. Subgroup analysis of myocardial infarction according to different new oral anticoagulants

Supplemental Figure4. Results of sensitivity and subgroup analyses for total and cardiovascular mortality

Supplemental Figure 5. Results of sensitivity and subgroup analyses for stroke or systemic embolism and ischemic stroke

Supplemental Figure 6. Results of sensitivity and subgroup analyses for major and intracranial bleeding *

* For the outcome major bleeding, the random effect model showed RR 0.86 (95% CI, 0.72‐1.02).

Supplemental Figure 7. Results of sensitivity and subgroup analyses for myocardial infarction

Supplemental Figure 8. Funnel plot of studies assessing total mortality

Supplemental Figure 9. Funnel plot of studies assessing stroke or systemic embolism

Supplemental Figure 10. Funnel plot of studies assessing major bleeding

Supplemental Figure 11. Funnel plot of studies assessing myocardial infarction

efficacy and safety of the novel oral anticoagulants in...

Documents