ORIGINAL ARTICLE

Treatment Discontinuations With New Oral

Agents for Long-term Anticoagulation: Insights
From a Meta-analysis of 18 Randomized Trials
Including 101,801 Patients
Saurav Chatterjee, MD; Partha Sardar, MD; Jay S. Giri, MD, MPH;
Joydeep Ghosh, MD; and Debabrata Mukherjee, MD, MS

Abstract

Objective: To systematically examine discontinuation rates with new US Food and Drug Admin-
istrationeapproved oral anticoagulants (NOACs) in patients with various indications for long-term
anticoagulation.
Patients and Methods: Poor adherence to medications is considered a potential and frequent cause of
treatment failure. We searched the PubMed, Cochrane Central Register of Controlled Trials, EMBASE,
EBSCO, Web of Science, and CINAHL databases for articles published from January 1, 2001, through
September 15, 2013. The following Medical Subject Heading terms and/or keywords were used for our
database searches: rivaroxaban, dabigatran, apixaban, new oral anticoagulants, oral thrombin inhibitors, and
oral factor Xa inhibitors. Articles in English that focused on randomized controlled trials (RCTs) comparing
NOACs (apixaban, dabigatran, and rivaroxaban) with conventional therapy or placebo were abstracted.
Independent extraction of relevant data was performed by 2 authors. The primary end point of interest
was discontinuation due to all causes. Other end points of interest were discontinuation due to adverse
events, consent withdrawal, and nonadherence.
Results: Eighteen RCTs including a total of 101,801 patients were included for analysis. Total study
drug discontinuation rates were not statistically different with NOACs in comparison to pharmaco-
logically active comparators for treatment of venous thromboembolism/pulmonary embolism (risk ratio
[RR], 0.91; 95% CI, 0.74-1.13; P¼.40) and for NOACs in comparison to warfarin and aspirin for
prevention of stroke in patients with atrial fibrillation (RR, 1.01; 95% CI, 0.87-1.17; P¼.92). In contrast,
in acute coronary syndromes, total study drug discontinuation with NOACs was significantly higher
than with placebo (RR, 1.40; 95% CI, 1.07-1.83; P¼.01). Overall discontinuations were comparable to
those with active comparators.
Conclusion: Study drug discontinuations with NOACs were not significantly different from those with
conventional drugs in treatment of venous thromboembolism/pulmonary embolism and prevention of
stroke in patients with atrial fibrillation but were worse in acute coronary syndromes as noted in evidence
from contemporary RCTs.
ª 2014 Mayo Foundation for Medical Education and Research n Mayo Clin Proc. 2014;89(7):896-907

From St Luke’s-Roosevelt
Hospital Center of the Mount
Sinai Health System, New
York, NY (S.C.); Division of
Cardiology, Texas Tech Uni-
versity Health Sciences Cen-
ter, El Paso, TX (P.S., D.M.);
Division of Cardiology,
Affiliations continued at
the end of this article.
P
oor adherence to medications is consid-
ered to be a potential cause of treatment
failure.1 Previous studies have reported
that adherence to effective therapy translated
into positive health outcomes and lower mor-
tality.1 Nonadherence is more common in
long-term therapy for chronic diseases than
in short-term therapy for acute conditions.2
Nonadherence to conventional anticoagulants
such as vitamin K antagonists (VKAs) has been
associated with inadequate anticoagulation and
high risk of embolic events.3 The new oral anti-
coagulants (NOACs) dabigatran, rivaroxaban,
and apixaban have been evaluated in several in-
dications for long-term anticoagulation such as
stroke prevention in atrial fibrillation (AF), treat-
ment of venous thromboembolism (VTE), and
acute coronary syndrome (ACS).4-6
The NOACs have efficacy and safety com-
parable to that of VKAs for stroke prevention

896 Mayo Clin Proc. n July 2014;89(7):896-907 n http://dx.doi.org/10.1016/j.mayocp.2014.01.030

www.mayoclinicproceedings.org n ª 2014 Mayo Foundation for Medical Education and Research
DISCONTINUATION RATES WITH NEW ORAL ANTICOAGULANTS
in AF and for treatment of VTE.4,6 They have
certain potential advantages over VKA, such
as rapid onset of action, obviation of need for
bridging therapy with heparin, predictable ef-
fects with fixed dosages, and no need for moni-
toring.4,6 However, there is concern regarding
poor adherence with these newer anticoagulant
agents in recent publications.4,7,8 Several ran-
domized controlled trials (RCTs) have reported
high discontinuation rates with these new oral
agents.9-11 Previous reports suggest that higher
doses may be required to achieve similar levels
of anticoagulation efficacy for the same drug in
patients with poor adherence to new anticoag-
ulants.7 The NOACs with twice-daily doses,
dabigatran and apixaban, may have higher rates
of nonadherence because of the increased
frequency of medication administration.7,8
Adherence is essential for these drugs to be
effective without causing major complica-
tions.12-14 However, data related to discontinu-
ation of NOACs is sparse and heterogeneous in
published studies, and various causes of drug
discontinuation have not been examined for a
comprehensive evaluation. We conducted a
meta-analysis to examine the discontinuation
rates of NOACs for various indications for
long-term anticoagulation.
PATIENTS AND METHODS
Data Sources and Searches
We searched the PubMed, Cochrane Central
Register of Controlled Trials, EMBASE, EBSCO,
Web of Science, and CINAHL databases from
January 1, 2001, through September 15, 2013,
for English-language, peer-reviewed publica-
tions. We identified RCTs with the NOACs dabi-
gatran, rivaroxaban, and apixaban for long-term
anticoagulation (more than 12 weeks because
that remains the minimum duration of treatment
recommended for thrombotic events) for various
indications. The following Medical Subject Head-
ing terms and/or keywords were used for our
database searches: rivaroxaban, dabigatran, apixa-
ban, new oral anticoagulants, oral thrombin inhibi-
tors, and oral factor Xa inhibitors. We also
searched related reviews, clinical trials databases,
and the reference lists of all retrieved articles.
Study Selection
We included RCTs that compared NOACs with
conventional anticoagulants or placebo for the
Mayo Clin Proc. n July 2014;89(7):896-907 n http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
treatment of VTE/pulmonary embolism (PE),
ACS, and stroke prevention in patients with
AF. The included studies had to have at least
12 weeks of follow-up. We did not include
studies of orthopedic operations because of
the considerably shorter durations of treatment
and follow-up, the greater variability in base-
line characteristics, and surgery-specific con-
founding factors. Both double-blind and
open-label trial designs were eligible for inclu-
sion. The PRISMA (Preferred Reporting Items
for Systematic Reviews and Meta-analyses)
statement for reporting of systematic reviews
and meta-analyses of RCTs15 was followed for
the protocol of our meta-analysis (Figure 1).
Data Extraction and Quality Assessment
Two physician-reviewers (S.C., P.S.) indepen-
dently extracted data from relevant published
articles after determining the eligibility for in-
clusion. Disagreements regarding data incorpo-
ration were resolved by consensus among all
authors. Methods specified in the Cochrane
Handbook for Systematic Reviews of Interven-
tions16 were followed for objective assessment
of the included trials. We extracted data from
published sources regarding total number of
treated patients, duration of follow-up, and
drugs for the intervention and control groups.
The occurrence of the following 4 end points
was abstracted according to the intention-to-
treat population for individual trials and sepa-
rately for the study drug and control drug:
discontinuation due to all causes, discontinua-
tion due to adverse events, discontinuation due
to consent withdrawal, and discontinuation
due to nonadherence. The definition for each
end point was that specified by the individual
trial (details of the definitions are provided in
the Supplemental Appendix, available online
at http://www.mayoclinicproceedings.org).
Data Synthesis and Analysis
To combine the data from each study, random-
effects (DerSimonian and Laird) models (as
appropriate for using data from published liter-
ature) were used to calculate a summary esti-
mate of discontinuation across all included
studies. When more than one dose of the study
drug was used in a single trial, we added the
data related to particular end points for all
doses, ie, we added the discontinuation events
for different doses of the NOACs and used the
897
 MAYO CLINIC PROCEEDINGS

Identification
3220 Records identified 510 Additional records
through identified through
database searching other sources

1070 Duplicates removed

Screening

2593 Records excluded

- 2251 Not a randomized controlled
trial
- 229 No control arm
2660 Records screened - 113 Study related to surgical
and other indications

48 Full-text articles excluded

- 21 Not a randomized controlled
67 Full-text articles trial
assessed for eligibility
Eligibility

- 19 Subgroup analysis
- 8 Study related to other
indications

1 Study excluded
19 Studies included in
qualitative synthesis (DVT rather than VTE/PE, AF,
or ACS)
Included

18 Studies included in
quantitative synthesis
(meta-analysis)

FIGURE 1. Search strategy and study selection per PRISMA (Preferred Reporting Items for Systematic
Reviews and Meta-analyses) checklist. ACS ¼ acute coronary syndrome; AF ¼ atrial fibrillation; DVT ¼
deep venous thrombosis; VTE/PE ¼ venous thromboembolism/pulmonary embolism.

sum of the number of events with different
doses for the purpose of analysis. We calculated
the risk ratio (RR) estimates and associated
95% CIs for each of the oral anticoagulants
separately and for each indication of use. The
Cochran Q test and the Higgins I2 test were
used for heterogeneity testing. A Cochran Q
test P<.10 and a Higgins I2>50% were consid-
ered indicative of significant heterogeneity. We
also created funnel plots graphically showing
the logarithm of the SE and the effect size to
evaluate publication bias and performed the
Egger regression test as recommended in the
Cochrane Handbook.16 Analyses were done us-
ing the Review Manager Version 5.2 (Nordic
Cochrane Centre, Cochrane Collaboration)
and STATA/SE 11 (StataCorp). We carried
out sensitivity analysis using data from only
n n
898 Mayo Clin Proc.
the phase 3 trials, for the “double-blind, sham
INR [international normalized ratio]” trials,
and for the individual medications under
consideration. Having noted different periods
of follow-up for the individual trials, we calcu-
lated the rate ratios of events from the absolute
number of events and person-years of follow-
up (obtained via multiplication of the sample
size of the trial population by the mean dura-
tion of follow-up), assuming a constant rate
of events for the individual trial. The natural
logarithms of the rate ratios were combined in
a Poisson regression model using a generic
inverse variance method.16 We also performed
meta-regression analyses in the presence of
extensive heterogeneity noted with a particular
outcome in an attempt to identify, if possible,
the covariates responsible for the heterogeneity.
July 2014;89(7):896-907 http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
DISCONTINUATION RATES WITH NEW ORAL ANTICOAGULANTS
Objective evaluation of trial quality and risk of
bias in reporting data for individual studies
was carried out using Cochrane metrics.16
Another sensitivity analysis was performed
with only studies with low risk of bias, low
bias risk being determined by the presence
of high risk in any one of the criteria assessed
in the risk of bias table. Noting the varied na-
ture of the trials of NOACs in ACS (in which
placebo was the comparator rather than phar-
macologically active medications), a separate
analysis was done comparing all-cause discon-
tinuations with NOACs with discontinuations
with all active comparators regardless of the
indication.
RESULTS
Study Characteristics
Database searches identified 3220 publications,
and an additional 510 reports were identified
from other sources. After assessing for eligibility
and exclusion as shown in Figure 1, 67 articles
were reviewed in full text. One notable study
of apixaban for treatment of deep venous throm-
bosis was excluded after qualitative synthesis
because of inconsistency in the reported data,
unavailability of adequate data for analysis, and
incompatibility with our inclusion criteria, ie, it
did not involve VTE/PE, AF, or ACS.17 Ulti-
mately, 18 RCTs involving 101,801 patients
were included in our meta-analysis.9-11,18-32 Of
the included trials, 6 were studies involving
treatment of VTE/PE, 7 focused on patients
with AF, and 5 involved treatment of ACS.
Four trials evaluated dabigatran, 8 evaluated
rivaroxaban, and 6 evaluated apixaban. The
characteristics of the included trials are provided
in Supplemental Table 1 (available online at
http://www.mayoclinicproceedings.org). Treat-
ment duration ranged from 12 weeks to 12
months in VTE/PE trials, from 12 weeks to 2
years in AF studies, and from 6 months to 13
months in ACS studies. Comparator groups in
all VTE/PE studies except one24 received VKA af-
ter initial bridging with low-molecular-weight
heparin therapy. In AF trials, all comparator
groups received warfarin except in the AVER-
ROES (Apixaban Versus Acetylsalicylic Acid
[ASA] to Prevent Stroke in Atrial Fibrillation Pa-
tients Who Have Failed or Are Unsuitable for
Vitamin K Antagonist Treatment) trial,18 which
used aspirin. In ACS trials, comparator groups
Mayo Clin Proc. n July 2014;89(7):896-907 n http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
received placebo along with other indicated ther-
apy (Supplemental Table 1, available online at
http://www.mayoclinicproceedings.org).
Meta-analysis
Treatment Discontinuation Due to All Cau-
ses. When analysis was done with all causes of
discontinuation together, in treatment of VTE/
PE, discontinuation rates were not statistically
different with NOACs in comparison to VKA
(RR, 0.91; 95% CI. 0.74-1.13; P¼.40). In pa-
tients with AF, no statistically significant differ-
ence was found in total drug discontinuation
between NOACs and warfarin (RR, 1.01; 95%
CI, 0.87-1.17; P¼.92). In patients with ACS, to-
tal study drug discontinuation with NOACs was
significantly higher than with placebo (RR, 1.40;
95% CI, 1.07-1.83; P¼.01). However, all of
these results were associated with marked het-
erogeneity (Figure 2). Sensitivity analyses of
phase 3 trials only revealed all-cause discontin-
uation rates for VTE/PE (RR, 0.84; 95% CI, 0.68-
1.03; P¼.10), AF (RR, 1.01; 95% CI, 0.87-1.17;
P¼.71), and ACS (RR, 1.11; 95% CI, 1.03-1.19;
P¼.004) to be similar to the overall discontinu-
ation rates. When only the direct thrombin in-
hibitor dabigatran was assessed for VTE/PE (RR,
1.11; 95% CI, 0.92-1.33; P¼.27), AF (RR, 1.38;
95% CI, 0.94-2.04; P¼.10), and ACS (only 1
trial31) (RR 1.26; 95% CI, 0.96-1.66; P¼.09), the
results again were consistent with the overall
findings. In the case of the pooled factor Xa in-
hibitors for VTE/PE (RR, 0.86; 95% CI, 0.71-
1.04; P¼.16), AF (RR, 0.94; 95% CI, 0.85-1.05;
P¼.26), and ACS (RR, 1.43; 95% CI, 1.06-1.94;
P¼.02), the discontinuation rates again were
congruent with the overall analysis (Figure 2).
Meta-regression analyses did not reveal the
reason for the heterogeneity noted with the dis-
continuations of NOACs in AF and ACS, while
for VTE/PE, the differences in baseline covariates
accounted for the heterogeneity (Supplemental
Figure 1, available online at http://www.mayo
clinicproceedings.org). A sensitivity analysis
was carried out using only double-blind, sham
INR trials. The outcomes were congruent with
the primary analyses for VTE/PE (RR, 0.88; 95%
CI, 0.56-1.39; P¼.14) and AF (RR, 0.97; 95%
CI, 0.85-1.10; P¼.65).
The follow-up adjusted analyses in the
Poisson model were congruent with the unad-
justed results, with rates of all-cause discontinua-
tions being significantly improved in VTE/PE (RR,
899
 MAYO CLINIC PROCEEDINGS

A. Risk ratio for discontinuation of NOACs due to all causes for VTE/PE, unadjusted event rates
NOACS Comparator
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weight M-H, random [95% Cl] M-H, random, 95% Cl

AMPLIFY-EXT,24 2013 243 1653 188 829 20.8% 0.65 [0.55-0.77]

EINSTEIN,19 2010 196 1731 244 1718 20.7% 0.80 [0.67-0.95]
Einstein-DVT Dose,22 2008 48 406 15 137 9.2% 1.08 [0.63-1.86]
EINSTEIN-PE,20 2012 258 2420 297 2413 21.3% 0.87 [0.74-1.01]
ODIXa-DVT,21 2007 73 487 10 126 7.6% 1.89 [1.00-3.55]
RE-COVER,23 2009 204 1273 183 1266 20.4% 1.11 [0.92-1.33]
Total pts 7970 6489 100.0% 0.91 [0.74-1.13]
Total events 1022 937

Heterogeneity: τ2=0.05; c2=25.18; df=5 (P<.001); I2=80% 0.01 0.1 1 10 100

Test for overall effect: Z=0.84 (P=.40) Favors NOACs Favors comparator

B. Risk ratio for discontinuation of NOACs due to all causes for stroke prevention in atrial fibrillation, unadjusted event rates
NOACS Comparator
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weight M-H, random [95% Cl] M-H, random, 95% Cl

ARISTOTLE,10 2011 2310 9120 2493 9081 21.4% 0.92 [0.88-0.97]

ARISTOTLE J,26 2011 12 148 9 74 3.0% 0.67 [0.29-1.51]
AVERROES,18 2011 503 2808 572 2791 19.7% 0.87 [0.78-0.97]
PETRO,28 2007 36 445 2 70 1.1% 2.83 [0.70-11.50]
RE-LY,27 2009 2372 12,091 902 6022 20.9% 1.31 [1.22-1.40]
J-ROCKET AF,25 2012 83 640 96 640 12.8% 0.86 [0.66-1.14]
ROCKET AF,9 2011 1691 7081 1584 7090 21.1% 1.07 [1.01-1.14]
Total pts 32,333 25,768 100.0% 1.01 [0.87-1.17]
Total events 7007 5658

Heterogeneity: τ2=80.05; c2=80.05; df=6 (P<.001); I2=93% 0.01 0.1 1 10 100

Test for overall effect: Z=0.10 (P=.92) Favors NOACs Favors comparator

C. Risk ratio for discontinuation of NOACs due to all causes for ACS, unadjusted event rates
NOACS Comparator
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weight M-H, random [95% Cl] M-H, random, 95% Cl

APPRAISE-,30 2009 480 1104 84 611 19.2% 3.16 [2.56-3.90]

APPRAISE-2,31 2011 863 3705 748 3687 21.4% 1.15 [1.06-1.26]
ATLAS ACS-TIMI 46,29 2009 347 1823 149 907 20.0% 1.16 [0.97-1.38]
ATLAS ACS 2–TIMI 51,11 2012 2914 10,350 1355 5176 21.6% 1.07 [1.01-1.13]
RE-DEEM,32 2011 270 1505 53 373 17.8% 1.26 [0.96-1.66]
Total pts 18,487 10,754 100.0% 1.40 [1.07-1.83]
Total events 4874 2398

Heterogeneity: τ2=0.08; c2=98.15; df=4 (P<.001); I2=95% 0.5 0.7 1 1.5 2

Test for overall effect: Z=2.48 (P=.01) Favors NOACs Favors comparator

FIGURE 2. A, Risk ratio for discontinuation of new oral anticoagulants (NOACs) due to all causes for venous thromboembolism/
pulmonary embolism (VTE/PE), unadjusted event rates. B, Risk ratio for discontinuation of NOACs due to all causes for stroke
prevention in atrial fibrillation, unadjusted event rates. C, Risk ratio for discontinuation of NOACs due to all causes for acute coronary
syndrome (ACS), unadjusted event rates. AMPLIFg-EXT ¼ Apixaban after the Initial Management of Pulmonary Embolism and Deep
Vein Thrombosis with First-Line Therapy -Extended Treatment; APPRAISE ¼ Apixaban for Prevention of Acute Ischemic Events;
APPRAISE-2 ¼ Apixaban for Prevention of Acute Ischemic Events 2; ARISTOTLE ¼ Apixaban for Reduction in Stroke and Other
Thromboembolic Events in Atrial Fibrillation; ARISTOTLE-J ¼ Apixaban for Reduction in Stroke and Other Thromboembolic Events
(continued on next page)

n n
900 Mayo Clin Proc. July 2014;89(7):896-907 http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
DISCONTINUATION RATES WITH NEW ORAL ANTICOAGULANTS

FIGURE 2. (Continued)
in Atrial Fibrillation-Japan; ATLAS ACS-2-TIMI 46 ¼ Anti-Xa Therapy to Lower Cardiovascular Events in Addition to Standard
Therapy in Subjects with Acute Coronary Syndrome -Thrombolysis in Myocardial Infarction 46; ATLAS ACS-2-TIMI 51 ¼ Anti-Xa
Therapy to Lower Cardiovascular Events in Addition to Standard Therapy in Subjects with Acute Coronary Syndrome -Thrombolysis
in Myocardial Infarction 51; AVERROES ¼ Apixaban Versus Acetylsalicylic Acid to Prevent Stroke in Atrial Fibrillation Patients Who
Have Failed or Are Unsuitable for Vitamin K Antagonist Treatment; EINSTEIN ¼ Oral Direct Factor Xa Inhibitor Rivaroxaban in
Patients With Acute Symptomatic Deep Vein Thrombosis; EINSTEIN-DVT Dose ¼ Once-daily Oral Direct Factor Xa Inhibitor
BAY59-7939 in Patients With Acute Symptomatic Deep-vein Thrombosis; EINSTEIN-PE ¼ Oral Direct Factor Xa Inhibitor Rivar-
oxaban in Patients With Acute Symptomatic Pulmonary Embolism; J-ROCKET AF ¼ Japan-Rivaroxaban Once Daily Oral Direct
Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation; ODIXa-
DVT ¼ Oral Direct Factor Xa Inhibitor BAY 59-7939 in Patients With Acute Symptomatic Deep-Vein Thrombosis; RE-DEEM ¼
Dose Finding Study for Dabigatran Etexilate in Patients With Acute Coronary Syndrome; RE-LY ¼ Randomized Evaluation of Long-
Term Anticoagulation Therapy; ROCKET AF ¼ Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K
Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation.

0.80; 95% CI, 0.68-0.94; P¼.006), comparable in
AF (RR, 1.01; 95% CI, 0.87-1.17; P¼.92), and
worse in ACS (RR, 1.40; 95% CI, 1.07-1.82;
P¼.01) (Supplemental Figure 2, available online
at http://www.mayoclinicproceedings.org).
Treatment Discontinuation Due to Adverse
Events. Discontinuation due to adverse events
(per individual trial definition) was not different
with NOACs in comparison to VKA for treatment
of VTE/PE (RR, 1.12; 95% CI, 0.70-1.80; P¼.64).
Similar results were found in patients with AF
(RR, 1.18; 95% CI, 0.87-1.60; P¼.30). Again,
these results showed significant heterogeneity.
Discontinuation due to adverse events was signif-
icantly higher with NOACs in patients with ACS
(RR, 1.24; 95% CI, 1.08-1.43; P¼.003), without
any significant heterogeneity (Figure 3, Table).
Treatment Discontinuation Due to Consent
Withdrawal. Discontinuation due to consent
withdrawal was significantly lower for NOACs
compared with VKA in patients with VTE/PE
(RR, 0.68; 95% CI, 0.47-0.99; P¼.05),
although associated with significant heteroge-
neity (I2¼60%; P¼.03). In patients with AF, the
consent withdrawal rate was not significantly
different between NOACs and comparators
(RR, 1.11; 95% CI, 0.92-1.32; P¼.28). In pa-
tients with ACS, consent withdrawal was signif-
icantly higher with NOACs than with placebo
(RR, 1.12; 95% CI, 1.01-1.25; P¼.03), without
significant heterogeneity (Figure 4, Table).
Treatment Discontinuation Due to Non-
adherence. Nonadherence rates were numer-
ically lower with NOACs than comparators in
patients with VTE/PE, although the difference
was not statistically significant (RR, 0.61; 95%
CI, 0.36-1.04; P¼.07 for the overall effect).
However, in patients with ACS, NOACs had
statistically similar nonadherence rates (RR,
1.19; 95% CI, 0.43-3.30; P¼.74) compared
with placebo. No significant heterogeneity
was present. In AF trials, rates of discontinua-
tion for nonadherence were not reported
(Figure 5, Table).
There was also no significant publication
bias detected with examination of funnel plots
for the primary outcomes, as well with the
Egger regression test (P>.20 for all outcomes)
(Supplemental Figure 3, available online at
http://www.mayoclinicproceedings.org).
Risk of bias for individual trials was
assessed (Supplemental Table 2, available
online at http://www.mayoclinicproceedings.
org), and a sensitivity analysis was carried out
using only low bias risk trials for the outcome
of discontinuation due to all causes. The out-
comes were congruent with the primary analyses
for VTE/PE (RR, 0.88; 95% CI, 0.56-1.39;
P¼.14), AF (RR, 0.97; 95% CI, 0.85-1.10;
P¼.65), and ACS (RR, 1.16; 95% CI, 1.08-
1.25; P<0.001). The absolute percentages of
discontinuations for various indications for
the individual NOACs were also computed
(Supplemental Table 3, available online at
http://www.mayoclinicproceedings.org).
Treatment Discontinuation Due to All Causes
With NOACs vs Active Comparators (Excluding
ACS Trials). Overall discontinuations were com-
parable between NOACs and pharmacologically
active comparators when evaluated in the setting

Mayo Clin Proc. n July 2014;89(7):896-907 n http://dx.doi.org/10.1016/j.mayocp.2014.01.030 901

www.mayoclinicproceedings.org
 MAYO CLINIC PROCEEDINGS

A. Risk ratio for discontinuation of NOACs due to adverse events for VTE/PE, unadjusted event rates

NOACS Comparators Risk ratio

Risk ratio
Study or subgroup Events Pts Events Pts Weight M-H, random [95% Cl] M-H, random, 95% Cl

AMPLIFY-EXT,24 2013 123 1653 126 829 21.3% 0.49 [0.39-0.62]

EINSTEIN,19 2010 74 1731 67 1718 20.4% 1.10 [0.79-1.52]
Einstein-DVT Dose,22 2008 21 406 5 137 11.7% 1.42 [0.54-3.69]
EINSTEIN-PE,20 2012 111 2420 92 2413 21.0% 1.20 [0.92-1.58]
ODIXa-DVT,21 2007 41 487 1 126 4.6% 10.61 [1.47-76.37]
RE-COVER,23 2009 115 1273 86 1266 21.0% 1.33 [1.02-1.74]
Total pts 7970 6489 100.0% 1.12 [0.70-1.80]
Total events 485 377

Heterogeneity: τ2=0.26; c2=47.28; df=5 (P<.001); I2=89% 0.01 0.1 1 10 100

Test for overall effect: Z=0.47 (P=.64) Favors NOACs Favors comparator

B. Risk ratio for discontinuation of NOACs due to adverse events for stroke prevention in atrial fibrillation, unadjusted event rates

NOACS Comparator Risk ratio

Risk ratio
Study or subgroup Events Pts Events Pts Weighta M-H, random [95% Cl] M-H, random, 95% Cl

ARISTOTLE,10 2011 679 9120 738 9081 24.5% 0.92 [0.83-1.01]

ARISTOTLE-J,26 2011 8 148 4 74 5.4% 1.00 [0.31-3.21]
PETRO,28 2007 29 445 0 70 1.2% 9.39 [0.58-152.00]
RE-LY,27 2009 731 12091 197 6022 23.7% 1.85 [1.58-2.18]
J-ROCKET AF,25 2012 83 640 96 640 21.0% 0.86 [0.66-1.14]
ROCKET AF,9 2011 694 7081 498 7090 24.3% 1.19 [1.07-1.34]
Total pts 29525 22977 100.0% 1.18 [0.87-1.60]
Total events 2124 1533

Heterogeneity: τ2=0.10; c2=63.22; df=5 (P<.001); I2=92% 0.01 0.1 1 10 100

Test for overall effect: Z=1.04 (P=.30) Favors NOACs Favors comparator

C. Risk ratio for discontinuation of NOACs due to adverse events for ACS, unadjusted event rates

NOACS Comparator
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weighta M-H, random [95% Cl] M-H, random, 95% Cl

APPRAISE-1,30 2009 94 1104 53 611 16.9% 0.99 [0.71-1.35]

APPRAISE-2,31 2011 313 3705 241 3687 47.9% 1.29 [1.10-1.52]
ATLAS ACS-TIMI 46,29 2009 190 1823 66 907 22.8% 1.43 [1.10-1.87]
RE-DEEM,32 2011 135 1505 30 373 12.6% 1.12 [0.76-1.63]
Total pts 8137 5578 100.0% 1.24 [1.08-1.43]
Total events 732 390

Heterogeneity: τ2=0.00; c2=3.67; df=3 (P=.30); I2=18% 0.01 0.1 1 10 100

Test for overall effect: Z=2.98 (P=.003) Favors NOACs Favors comparators

FIGURE 3. A, Risk ratio for discontinuation of new oral anticoagulants (NOACs) due to adverse events for venous thromboembolism/
pulmonary embolism (VTE/PE), unadjusted event rates. B, Risk ratio for discontinuation of NOACs due to adverse events for stroke
prevention in atrial fibrillation, unadjusted event rates. C, Risk ratio for discontinuation of NOACs due to adverse events for acute
coronary syndrome (ACS), unadjusted event rates. For expansion of abbreviations and acronyms, see legend for Figure 2. aSum is greater
than 100 due to rounding.

n n
902 Mayo Clin Proc. July 2014;89(7):896-907 http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
DISCONTINUATION RATES WITH NEW ORAL ANTICOAGULANTS
of AF and VTE/PE (RR, 0.96; 95% CI, 0.85-1.08;
P¼.48).
DISCUSSION
To the best of our knowledge, the present meta-
analysis provides the first extensive summary of
the discontinuation rates of NOACs in clinical
trials for long-term anticoagulation. Our analysis
showed that total discontinuation rates with
NOACs are not different compared with conven-
tional drugs in patients with VTE/PE (12.8% vs
14.4% for VKA) and AF (21.67% vs 21.96%
for warfarin and aspirin) but were significantly
higher in patients with ACS (26.4% vs 22.3%
for placebo; P¼.01). There may be possible ex-
planations for the consistently higher discontin-
uation rates with NOACs in patients with ACS
seen in our analysis. Patients with ACS are
already receiving dual antiplatelet therapies,
and risk of bleeding is likely to be higher with tri-
ple therapy.5 Also, the fact remains that the
comparator group in ACS trials was placebo
and not a VKA or other pharmacologically active
medication, ie, there was no “double-dummy”
comparator in the ACS trials.11 However, the
possibility remains that there was a higher fre-
quency of discontinuation due to adverse events
in the NOAC arms of the ACS studies because
patients in the other arm did not receive any
kind of active medication. Consent withdrawal
rates with use of NOACs were lower compared
with VKA in the treatment of VTE/PE.
Considerable concern has been expressed
in recent articles regarding drug discontinua-
tion with NOACs, which have been attributed
to intolerance due to minor adverse effects
such as gastrointestinal discomfort on initia-
tion, major prohibitive factors such as severe
bleeding with sustained use, natural wariness
of using a new medication, highlighting of
adverse events by the media, or simple patient
preference.4,7 Our meta-analysis on this issue
provides reassurance that the drug discontinu-
ation rate with NOACs is similar to that of com-
parators in patients with AF and VTE/PE and
that NOACs have been appropriately approved
for use in these indications per current guide-
lines. In patients with ACS, consistently high
rates of discontinuation due to most causes
may limit the use of NOACs, even if they may
be of some benefit in efficacy end points.
Another concern is that although our anal-
ysis provides reassuring evidence of comparable
Mayo Clin Proc. n July 2014;89(7):896-907 n http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
TABLE. Risk Ratios for NOAC Discontinuation vs Comparators for Various
Causes and Indications for Long-term Anticoagulation
Variable Risk ratio (95% CI)
Discontinuation due to adverse events
For treatment of VTE/PE 1.12 (0.70-1.80)
For patients with AF 1.18 (0.87-1.60)
For treatment of ACS 1.24 (1.08-1.43)
Discontinuation due to consent withdrawal
For treatment of VTE/PE 0.68 (0.47-0.99)
For patients with AF 1.11 (0.92-1.32)
For treatment of ACS 1.12 (1.01-1.25)
Discontinuation due to nonadherence
For treatment of VTE/PE 0.61 (0.36-1.04)
For treatment of ACS 1.19 (0.43-3.30)
ACS ¼ acute coronary syndrome; AF ¼ atrial fibrillation; NOAC ¼ new oral anticoagulant; VTE/
PE ¼ venous thromboembolism/pulmonary embolism.
discontinuation rates with use of NOACs in the
setting of clinical trials, it may not be reflective
of “real-world” clinical practice. However, on
perusal of recent data, it appears that NOACs
have largely been well tolerated in the general
population, and the rates of serious adverse
events in day-to-day clinical practice leading
to discontinuation of the drugs have been
few.32-35 These early data from clinical practice
provide some measure of reassurance of similar
adverse event rates and resultant discontinua-
tion of NOACs (the data currently being mainly
available for dabigatran, eg, real-world major
bleeding rates of 1.3%34 and 3.11% in the RE-
LY trial27) in routine practice settings as in the
well-controlled environment of clinical trials.
Many factors associated with poor compli-
ance with prescribed medicines have been identi-
fied in several studies, including frequent dosing,
social problems, poor patient-physician commu-
nication, educational level, number of doses,
employment status, and disability.7,8 In our anal-
ysis, we found that discontinuation due to
adverse events is still a considerable problem
with NOACs, even in the largely well-regulated
world of RCTs, although discontinuation rates
are at least comparable to those of conventional
drugs for their approved indications of stroke
prevention and VTE/PE.
Significantly less consent withdrawal with
NOACs in only VTE/PE studies was an interesting
finding. It is a matter of debate whether these find-
ings are due to actual greater tolerability with
NOACs than with VKA or if they simply reflect a
Berksonian bias in which the NOAC arms were
903
 MAYO CLINIC PROCEEDINGS

A. Risk ratio for discontinuation of NOACs due to consent withdrawal for VTE/PE unadjusted event rates
NOACS Comparator
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weighta M-H, random [95% Cl] M-H, random, 95% Cl

AMPLIFY-EXT,24 2013 19 1653 6 829 11.0% 1.59 [0.64-3.96]

EINSTEIN,19 2010 34 1731 67 1718 23.4% 0.50 [0.34-0.76]
Einstein-DVT Dose,22 2008 6 406 3 137 6.0% 0.67 [0.17-2.66]
EINSTEIN-PE,20 2012 66 2420 118 2413 27.0% 0.56 [0.41-0.75]
ODIXa-DVT,21 2007 10 487 7 126 10.5% 0.37 [0.14-0.95]
RE-COVER,23 2009 39 1273 36 1266 22.2% 1.08 [0.69-1.68]
Total pts 7970 6489 100.0% 0.68 [0.47-0.99]
Total events 174 237

Heterogeneity: τ2=0.11; c2=12.49; df= 5 (P=.03); I2=60% 0.01 0.1 1 10 100

Test for overall effect: Z= 2.00 (P=.05) Favors NOACs Favors comparators

B. Risk ratio for discontinuation of NOACs due to consent withdrawal for stroke prevention in atrial fibrillation, unadjusted event rates
NOACS Comparator
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weight M-H, random [95% Cl] M-H, random, 95% Cl

RE-LY,27 2009 914 12,091 375 6022 49.2% 1.21 [1.08-1.36]

ROCKET-AF,9 2011 618 7081 613 7090 50.8% 1.01 [0.91-1.12]
Total pts 19,172 13112 100.0% 1.11 [0.92-1.32]
Total events 1532 988

Heterogeneity: τ2=0.01; c2=5.26; df=1 (P=.02); I2=81% 0.01 0.1 1 10 100

Test for overall effect: Z=1.09 (P=.28) Favors NOACs Favors comparator

C. Risk ratio for discontinuation of NOACs due to consent withdrawal for ACS, unadjusted event rates
NOACS Comparator
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weighta M-H, random [95% Cl] M-H, random, 95% Cl

APPRAISE-2,31 2011 196 3705 155 3687 24.7% 1.26 [1.02-1.55]

ATLAS ACS-TIMI 46,29 2009 79 1823 42 907 8.4% 0.94 [0.65-1.35]
ATLAS ACS-2–TIMI 51,11 2012 890 10,350 403 5176 67.0% 1.10 [0.99-1.24]
Total pts 15,878 9770 100.0% 1.12 [1.01-1.25]
Total events 1165 600

Heterogeneity: τ2=0.00; c2=2.21; df=2 (P=.33); I2=10% 0.01 0.1 1 10 100

Test for overall effect: Z=2.15 (P=.03) Favors NOACs Favors comparators

FIGURE 4. A, Risk ratio for discontinuation of new oral anticoagulants (NOACs) due to consent withdrawal for venous thromboem-
bolism/pulmonary embolism (VTE/PE), unadjusted event rates. B, Risk ratio for discontinuation of NOACs due to consent withdrawal for
stroke prevention in atrial fibrillation, unadjusted event rates. C, Risk ratio for discontinuation of NOACs due to consent withdrawal for
acute coronary syndrome (ACS), unadjusted event rates. For expansion of abbreviations and acronyms, see legend for Figure 2. aSum is
greater than 100 due to rounding.

unintentionally subjected to more intensive coun-
seling and vigilance, causing greater adherence.
Our meta-analysis sheds insight on the
important issue of nonadherence with NOACs.
Because of the recent introduction of these
NOACs in clinical practice, real-world toler-
ability and adherence data from these
n n
904 Mayo Clin Proc.
medications remain pending, especially for
a reasonable period of follow-up. Strategies
and interventions to improve adherence to
NOACs should be developed to obtain
optimal benefits from these newer agents,
such as dedicated anticoagulation clinics.36
Also, rigid definitions should be developed
July 2014;89(7):896-907 http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
DISCONTINUATION RATES WITH NEW ORAL ANTICOAGULANTS

A. Risk ratio for discontinuation of NOACs due to nonadherence for VTE/PE, unadjusted event rates
NOACS Comparators
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weight M-H, random [95% Cl] M-H, random, 95% Cl

ODIXa-DVT,21 2007 2 487 0 126 3.0% 1.30 [0.06-26.93]

RE-COVER,23 2009 21 1273 35 1266 97.0% 0.60 [0.35-1.02]
Total pts 1760 1392 100.0% 0.61 [0.36-1.04]
Total events 23 35

Heterogeneity: τ2=0.00; c2=.25; df=1 (P=.62); I2=0% 0.01 0.1 1 10 100

Test for overall effect: Z=1.83 (P=.07) Favors NOACs Favors comparator

B. Risk ratio for discontinuation of NOACs due to nonadherence for ACS, unadjusted event rates
NOACS Comparator
Risk ratio Risk ratio
Study or subgroup Events Pts Events Pts Weight M-H, random [95% Cl] M-H, random, 95% Cl

APPRAISE,30 2009 7 1104 6 611 41.7% 0.85 [0.22-1.91]

RE-DEEM,32 2011 67 1505 9 373 58.3% 1.85 [0.93-3.67]
Total pts 2609 984 100.0% 1.19 [0.43-3.30]
Total events 74 15

Heterogeneity: τ2=0.34; c2=2.59; df=1 (P=.11); I2=61% 0.01 0.1 1 10 100

Test for overall effect: Z=0.34 (P=.74) Favors NOACs Favors comparator

FIGURE 5. A, Risk ratio for discontinuation of new oral anticoagulants (NOACs) due to nonadherence for venous thromboem-
bolism/pulmonary embolism (VTE/PE), unadjusted event rates. B, Risk ratio for discontinuation of NOACs due to nonadherence for
acute coronary syndrome (ACS), unadjusted event rates. For expansion of abbreviations and acronyms, see legend for Figure 2.

for identifying and addressing various rea-
sons for drug discontinuations.
STUDY LIMITATIONS
The studies included in our analysis have
inherent differences in designs and baseline pa-
rameters. Data on drug discontinuation was not
uniformly documented in most of the studies.
Discontinuation data in a few studies was based
on a pill counting method, which is susceptible
to misinterpretations and confounders.7 Defini-
tions of the data on the individual components
of discontinuation were not uniform, and inter-
pretation of our results was limited by extensive
heterogeneity on occasion. Generalizability re-
mains an issue because this study was limited
by its assessment of why patients stopped using
these drugs in a trial setting. The application of
our findings to clinical practice remains conjec-
tural at present because of the paucity of the
data and the relatively recent approval of NOACs.
Rates of discontinuation for the various drugs
may differ in trials and in routine clinical practice.
We also did not perform a mixed-treatment
comparison analysis or network meta-analysis
because of the heterogeneity noted with the pri-
mary pair-wise analyses, indicating variability in
the included trials, as well as the variations in the
different treatment arms, eg, NOACs, warfarin,
aspirin, low-molecular-weight heparin, and pla-
cebo, which have potential for major biases and
errors, especially when used in the context of
Markov chain Monte Carlo simulations.37,38
CONCLUSION
Study drug discontinuations with NOACs are
comparable with those of conventional drugs
in the treatment of VTE/PE and the prevention
of stroke in patients with AF in contemporary
RCTs. Considerably worse rates of discontinu-
ation in ACS may potentially limit the use of
NOACs in this population.
ACKNOWLEDGMENTS
Dr Chatterjee had full access to all of the data in
the study and takes responsibility for the integ-
rity of the data and the accuracy of the data

Mayo Clin Proc. n July 2014;89(7):896-907 n http://dx.doi.org/10.1016/j.mayocp.2014.01.030 905

www.mayoclinicproceedings.org

analysis. Technical appendix, statistical code,
and data set are available from the correspond-
ing author (sauravchatterjeemd@gmail.com).
SUPPLEMENTAL ONLINE MATERIAL
Supplemental material can be found online at
http://www.mayoclinicproceedings.org.
Abbreviations and Acronyms: ACS = acute coronary
syndrome; AF = atrial fibrillation; NOAC = new oral anti-
coagulant; PE = pulmonary embolism; RCT = randomized
controlled trial; RR = risk ratio; VKA = vitamin K antagonist;
VTE = venous thromboembolism
Affiliations (Continued from the first page of this
article.): University of Pennsylvania, Philadelphia (J.S.G.);
and Division of Cardiology, Columbia University College
of Physicians and Surgeons, New York, NY (J.G.).
Correspondence: Address to Saurav Chatterjee, MD, Divi-
sion of Cardiology, St Luke’s-Roosevelt Hospital Center,
Clark Bldg, 3rd floor, 1111 Amsterdam Ave, New York,
NY 10025 (sauravchatterjeemd@gmail.com).
REFERENCES
1. Simpson SH, Eurich DT, Majumdar SR, et al. A meta-analysis of
the association between adherence to drug therapy and mor-
tality. BMJ. 2006;333(7557):15.
2. Osterberg L, Blaschke T. Adherence to medication. N Engl J
Med. 2005;353(5):487-497.
3. Lavitola Pde L, Sampaio RO, Oliveira WA, et al. Warfarin or
aspirin in embolism prevention in patients with mitral valvul-
opathy and atrial fibrillation. Arq Bras Cardiol. 2010;95(6):749-
755.
4. Adam SS, McDuffie JR, Ortel TL, Williams JW Jr. Comparative
effectiveness of warfarin and new oral anticoagulants for the
management of atrial fibrillation and venous thromboembolism:
a systematic review. Ann Intern Med. 2012;157(11):796-807.
5. Komócsi A, Vorobcsuk A, Kehl D, Aradi D. Use of new-
generation oral anticoagulant agents in patients receiving anti-
platelet therapy after an acute coronary syndrome: systematic
review and meta-analysis of randomized controlled trials. Arch
Intern Med. 2012;172(20):1537-1545.
6. De Caterina R, Husted S, Wallentin L, et al; Coordinating Com-
mittee. New oral anticoagulants in atrial fibrillation and acute
coronary syndromes: ESC Working Group on Thrombosis-
Task Force on Anticoagulants in Heart Disease position paper.
J Am Coll Cardiol. 2012;59(16):1413-1425.
7. Rodriguez RA, Carrier M, Wells PS. Non-adherence to new
oral anticoagulants: a reason for concern during long-term anti-
coagulation? J Thromb Haemost. 2013;11(2):390-394.
8. Bellamy L, Rosencher N, Eriksson B. Adherence to a new oral
anticoagulant treatment prescription: dabigatran etexilate. Pa-
tient Prefer Adherence. 2009;3:173-177.
9. Patel MR, Mahaffey KW, Garg J, et al; ROCKET AF Investigators.
Rivaroxaban versus warfarin in nonvalvular atrial fibrillation.
N Engl J Med. 2011;365(10):883-891.
10. Granger CB, Alexander JH, McMurray JJ, et al; ARISTOTLE
Committees and Investigators. Apixaban versus warfarin in
patients with atrial fibrillation. N Engl J Med. 2011;365(11):
981-992.
11. Mega JL, Braunwald E, Wiviott SD, et al; ATLAS ACS 2eTIMI
51 Investigators. Rivaroxaban in patients with a recent acute
coronary syndrome. N Engl J Med. 2012;366(1):9-19.
n n
906 Mayo Clin Proc. July 2014;89(7):896-907
MAYO CLINIC PROCEEDINGS
12. Schulman S, Crowther MA. How I treat with anticoagulants in
2012: new and old anticoagulants, and when and how to
switch. Blood. 2012;119(13):3016-3023.
13. Kazmi RS, Lwaleed BA. New anticoagulants: how to deal with
treatment failure and bleeding complications. Br J Clin Pharma-
col. 2011;72(4):593-603.
14. Eikelboom JW, Wallentin L, Connolly SJ, et al. Risk of bleeding
with 2 doses of dabigatran compared with warfarin in older and
younger patients with atrial fibrillation: an analysis of the Ran-
domized Evaluation of Long-Term Anticoagulant Therapy
(RE-LY) trial. Circulation. 2011;123(21):2363-2372.
15. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement
for reporting systematic reviews and meta-analyses of studies
that evaluate health care interventions: explanation and elabo-
ration. Ann Intern Med. 2009;151(4):W65-W94.
16. Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Re-
views of Interventions. Version 5.1.0. http://www.cochrane-
handbook.org. Updated March 2011. Accessed July 1, 2013.
17. Goldhaber SZ, Leizorovicz A, Kakkar AK, et al; ADOPT Trial
Investigators. Apixaban versus enoxaparin for thrombopro-
phylaxis in medically ill patients. N Engl J Med. 2011;365(23):
2167-2177.
18. Connolly SJ, Eikelboom J, Joyner C, et al; AVERROES Steering
Committee and Investigators. Apixaban in patients with atrial
fibrillation. N Engl J Med. 2011;364(9):806-817.
19. EINSTEIN Investigators, Bauersachs R, Berkowitz SD,
Brenner B, et al. Oral rivaroxaban for symptomatic venous
thromboembolism. N Engl J Med. 2010;363(26):2499-2510.
20. EINSTEINePE Investigators, Büller HR, Prins MH, Lensin AW,
et al. Oral rivaroxaban for the treatment of symptomatic pul-
monary embolism. N Engl J Med. 2012;366(14):1287-1297.
21. Agnelli G, Gallus A, Goldhaber SZ, et al; ODIXa-DVT Study In-
vestigators. Treatment of proximal deep-vein thrombosis with
the oral direct factor Xa inhibitor rivaroxaban (BAY 59-
7939): the ODIXa-DVT (Oral Direct Factor Xa Inhibitor
BAY 59-7939 in Patients With Acute Symptomatic Deep-
Vein Thrombosis) study. Circulation. 2007;116(2):180-187.
22. Buller HR, Lensing AW, Prins MH, et al; Einstein-DVT Dose-
Ranging Study Investigators. A dose-ranging study evaluating
once-daily oral administration of the factor Xa inhibitor rivarox-
aban in the treatment of patients with acute symptomatic deep
vein thrombosis: the Einstein-DVT Dose-Ranging Study. Blood.
2008;112(6):2242-2247.
23. Schulman S, Kearon C, Kakkar AK, et al; RE-COVER Study Group.
Dabigatran versus warfarin in the treatment of acute venous
thromboembolism. N Engl J Med. 2009;361(24):2342-2352.
24. Agnelli G, Buller HR, Cohen A, et al; AMPLIFY-EXT Investiga-
tors. Apixaban for extended treatment of venous thromboem-
bolism. N Engl J Med. 2013;368(8):699-708.
25. Hori M, Matsumoto M, Tanahashi N, et al; J-ROCKET AF Study
Investigators. Rivaroxaban vs. warfarin in Japanese patients with
atrial fibrillation: the J-ROCKET AF study. Circ J. 2012;76(9):
2104-2111.
26. Ogawa S, Shinohara Y, Kanmuri K. Safety and efficacy of the
oral direct factor Xa inhibitor apixaban in Japanese patients
with non-valvular atrial fibrillation: the ARISTOTLE-J study.
Circ J. 2011;75(8):1852-1859.
27. Connolly SJ, Ezekowitz MD, Yusuf S, et al; RE-LY Steering
Committee and Investigators. Dabigatran versus warfarin in pa-
tients with atrial fibrillation [published correction appears in N
Engl J Med. 2010;363(19):1877]. N Engl J Med. 2009;361(12):
1139-1151.
28. Ezekowitz MD, Reilly PA, Nehmiz G, et al. Dabigatran with or
without concomitant aspirin compared with warfarin alone in
patients with nonvalvular atrial fibrillation (PETRO Study). Am
J Cardiol. 2007;100(9):1419-1426.
29. Mega JL, Braunwald E, Mohanavelu S, et al; ATLAS ACS-TIMI 46
Study Group. Rivaroxaban versus placebo in patients with acute
coronary syndromes (ATLAS ACS-TIMI 46): a randomised,
double-blind, phase II trial. Lancet. 2009;374(9683):29-38.
http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
DISCONTINUATION RATES WITH NEW ORAL ANTICOAGULANTS
30. APPRAISE Steering Committee and Investigators, Alexander JH,
Becker RC, Bhatt DL, et al. Apixaban, an oral, direct, selective fac-
tor Xa inhibitor, in combination with antiplatelet therapy after
acute coronary syndrome: results of the Apixaban for Preven-
tion of Acute Ischemic and Safety Events (APPRAISE) trial. Circu-
lation. 2009;119(22):2877-2885.
31. Alexander JH, Lopes RD, James S, et al; APPRAISE-2 Investiga-
tors. Apixaban with antiplatelet therapy after acute coronary
syndrome. N Engl J Med. 2011;365(8):699-708.
32. Oldgren J, Budaj A, Granger CB, et al; RE-DEEM Investigators.
Dabigatran vs. placebo in patients with acute coronary syn-
dromes on dual antiplatelet therapy: a randomized, double-
blind, phase II trial. Eur Heart J. 2011;32(22):2781-2789.
33. Rowley CP, Bernard ML, Brabham WW, et al. Safety of contin-
uous anticoagulation with dabigatran during implantation of car-
diac rhythm devices. Am J Cardiol. 2013;111(8):1165-1168.
Mayo Clin Proc. n July 2014;89(7):896-907 n http://dx.doi.org/10.1016/j.mayocp.2014.01.030
www.mayoclinicproceedings.org
34. Sørensen R, Gislason G, Torp-Pedersen C, et al. Dabigatran use
in Danish atrial fibrillation patients in 2011: a nationwide study.
BMJ Open. 2013;3(5).
35. Southworth MR, Reichman ME, Unger EF. Dabigatran and post-
marketing reports of bleeding. N Engl J Med. 2013;368(14):
1272-1274.
36. Testa S, Paoletti O, Zimmermann A, Bassi L, Zambelli S,
Cancellieri E. The role of anticoagulation clinics in the era of
new oral anticoagulants. Thrombosis. 2012;2012:835356.
37. Salanti G, Higgins JP, Ades AE, Ioannidis JP. Evaluation of net-
works of randomized trials. Stat Methods Med Res. 2008;
17(3):279-301.
38. Song F, Loke YK, Walsh T, Glenny AM, Eastwood AJ,
Altman DG. Methodological problems in the use of indirect
comparisons for evaluating healthcare interventions: survey of
published systematic reviews. BMJ. 2009;338:b1147.
907