Transcript Ruff_Meta-analysis - Clinical Trial Results

Comprehensive Meta-Analysis Comparing the
Efficacy and Safety of NOACs with
Warfarin in AF
An Analysis Including 71,683 Patients from
Four Large Randomized Clinical Trials
Christian T. Ruff, MD, MPH
TIMI Study Group
Brigham and Women’s Hospital
Harvard Medical School
Boston, MA
1
Pivotal Warfarin-Controlled Trials
Stroke Prevention in AF
Warfarin vs. Placebo
2,900 Patients
NOACs vs. Warfarin
71,683 Patients
ROCKET AF
(Rivaroxaban)
2010
6 Trial of Warfarin vs. Placebo
1989-1993
RE-LY
(Dabigatran)
2009
ENGAGE AF-TIMI 48
(Edoxaban)
2013
ARISTOTLE
(Apixaban)
2011
2
Comparative PK/PD of NOACs
Dabigatran
Rivaroxaban
Apixaban
Edoxaban
IIa (thrombin)
Xa
Xa
Xa
Hours to Cmax
1-3
2-4
3-4
1-2
Half-life, hours
12-17
5-13
12
10-14
80
33*
27
50
Transporters
P-gp
P-gp
P-gp
P-gp
CYP Metabolism, %
None
32
<32
<4
Target
Renal Clearance, %
CYP = cytochrome P450; P-gp = P-glycoprotein
*33% renally cleared; 33% excreted unchanged in urine
Pradaxa [package insert]. Ridgefield, CT: Boehringer Ingelheim Pharmaceuticals, Inc. 2013
Xarelto [package insert]. Titusville, NJ: Janssen Pharmaceuticals, Inc. 2011
Weinz et al. Drug Dispos Metab 2009;37:1056–1064
ELIQUIS Summary of Product Characteristics. Bristol Myers Squibb/Pfizer EEIG, UK
Matsushima et al. Am Assoc Pharm Sci 2011; abstract
Ogata, et al. J Clin Pharmacol 2010;50:743–753
Mendell, et al. Am J Cardiovasc Drugs 2013;13:331–342
Bathala, et al. Drug Metab Dispos 2012;40:2250–2255
3
NOAC SPAF Trials
Drug
# Randomized
Dose (mg)
Frequency
Dose Adjustment
At Baseline
After Randomization
Target INR (Warfarin)
Design
RE-LY
ROCKET-AF
ARISTOTLE
ENGAGE AF
Dabigatran
Rivaroxaban
Apixaban
Edoxaban
18,113
14,266
18,201
21,105
150, 110
20
5
60, 30
Twice Daily
Once Daily
Twice Daily
Once Daily
No
20 → 15
5 → 2.5
60 → 30
30 → 15
0
21
5
25
No
No
No
>9%
2.0-3.0
2.0-3.0
2.0-3.0
2.0-3.0
PROBE*
2x blind
2x blind
2x blind
*PROBE = prospective, randomized, open-label, blinded end point evaluation
Connolly SJ, et al. N Engl J Med 2009;361:1139-1151
Patel MR, et al. N Engl J Med 2011;365:883-891
Granger CB, et al. N Engl J Med 2011;365:981-992
Giugliano RP, et al. N Engl J Med 2013; e-pub ahead of print DOI:10.1056/NEJMoa1310907
4
Baseline Characteristics
RE-LY
(Dabigatran)
ROCKET-AF
(Rivaroxaban)
ARISTOTLE
(Apixaban)
ENGAGE AF
(Edoxaban)
# Randomized
18,113
14,264
18,201
21,105
Age, years
72 ± 9
73 [65-78]
70 [63-76]
72 [64-78]
Female, %
37
40
35
38
Paroxysmal AF
32
18
15
25
VKA naive
50
38
43
41
Aspirin Use
40
36
31
29
CHADS2
0-1
2
3-6
13
33 32
35
87
30
34
53
47
36
Connolly SJ, et al. N Engl J Med 2009;361:1139-1151
Patel MR, et al. N Engl J Med 2011;365:883-891
Granger CB, et al. N Engl J Med 2011;365:981-992
Giugliano RP, et al. N Engl J Med 2013; e-pub ahead of print DOI:10.1056/NEJMoa1310907
5
Trial Metrics
RE-LY
(Dabigatran)
ROCKET-AF
(Rivaroxaban)
ARISTOTLE
(Apixaban)
ENGAGE AF
(Edoxaban)
Median Follow-Up, years
2.0
1.9
1.8
2.8
Median TTR
66
58
66
68
Lost to Follow-Up, N
20
32
90
1
*TTR, time in therapeutic range
Connolly SJ, et al. N Engl J Med 2009;361:1139-1151
Patel MR, et al. N Engl J Med 2011;365:883-891
Granger CB, et al. N Engl J Med 2011;365:981-992
Giugliano RP, et al. N Engl J Med 2013; e-pub ahead of print DOI:10.1056/NEJMoa1310907
6
All NOACS: Stroke or SEE
Risk Ratio (95% CI)
0.66 (0.53 - 0.82)
RE-LY
[150 mg]
ROCKET AF
0.88 (0.75 - 1.03)
ARISTOTLE
0.80 (0.67 - 0.95)
ENGAGE AF-TIMI 48
0.88 (0.75 - 1.02)
[60 mg]
Combined
0.81 (0.73 - 0.91)
[Random Effects Model]
N=58,541
0.5
Heterogeneity p=0.13
p=<0.0001
Favors NOAC
1
Favors Warfarin
2
Ruff CT, et al. Lancet 2013 [in-press]
7
Secondary Efficacy Outcomes
Risk Ratio (95% CI)
Ischemic Stroke
0.92 (0.83 - 1.02)
p=0.10
Hemorrhagic Stroke
0.49 (0.38 - 0.64)
p<0.0001
MI
0.97 (0.78 - 1.20)
p=0.77
All-Cause Mortality
0.90 (0.85 - 0.95)
p=0.0003
0.2
0.5
Favors NOAC
1
2
Favors Warfarin
Heterogeneity p=NS for all outcomes
Ruff CT, et al. Lancet 2013 [in-press]
8
All NOACS: Major Bleeding
Risk Ratio (95% CI)
0.94 (0.82 - 1.07)
RE-LY
[150 mg]
ROCKET AF
1.03 (0.90 - 1.18)
ARISTOTLE
0.71 (0.61 - 0.81)
ENGAGE AF-TIMI 48
0.80 (0.71 - 0.90)
[60 mg]
Combined
0.86 (0.73 - 1.00)
[Random Effects Model]
p=0.06
N=58,498
0.5
Heterogeneity p=0.001
Favors NOAC
1
Favors Warfarin
2
Ruff CT, et al. Lancet 2013 [in-press]
9
Secondary Safety Outcomes
Risk Ratio (95% CI)
0.48 (0.39 - 0.59)
ICH
p<0.0001
1.25 (1.01 - 1.55)
GI Bleeding
0.2
p=0.043
0.5
Favors NOAC
1
2
Favors Warfarin
Heterogeneity
ICH, p=0.22
GI Bleeding, p=0.009
Ruff CT, et al. Lancet 2013 [in-press]
10
Subgroups: Stroke or SEE
Age
Gender
Diabetes
Prior Stroke or TIA
CrCl
CHADS2 Score
VKA Status
Center-Based TTR
Risk Ratio (95% CI)
P-Interaction
<75
0.85 (0.73 - 0.99)
p=0.38
≥75
0.78 (0.68 - 0.88)
Female
0.78 (0.65 - 0.94)
Male
0.84 (0.75 - 0.94)
No
0.83 (0.74 - 0.93)
Yes
0.80 (0.69 - 0.93)
No
0.78 (0.66 - 0.91)
Yes
0.86 (0.76 - 0.98)
<50
0.79 (0.65 - 0.96)
50-80
0.75 (0.66 - 0.85)
>80
0.98 (0.79 - 1.22)
0-1
0.75 (0.54 - 1.04)
2
0.86 (0.70 - 1.05)
3-6
0.80 (0.72 - 0.89)
Naive
0.75 (0.66 - 0.86)
Experienced
0.85 (0.70 - 1.03)
<66%
0.77 (0.65 - 0.92)
≥66%
0.82 (0.71 - 0.95)
0.5
Ruff CT, et al. Lancet 2013 [in-press]
1
Favors NOAC
p=0.52
p=0.73
p=0.30
p=0.12
p=0.76
p=0.31
p=0.60
2
Favors Warfarin
11
Subgroups: Major Bleeding
Age
Gender
Diabetes
Prior Stroke or TIA
CrCl
CHADS2 Score
VKA Status
Center-Based TTR
Risk Ratio (95% CI)
P-Interaction
<75
0.79 (0.67 - 0.94)
p=0.28
≥75
0.93 (0.74 - 1.17)
Female
0.75 (0.58 - 0.97)
Male
0.90 (0.72 - 1.12)
No
0.71 (0.54 – 0.93)
Yes
0.90 (0.78 - 1.04)
No
0.85 (0.72 - 1.01)
Yes
0.89 (0.77 - 1.02)
<50
0.74 (0.52 - 1.05)
50-80
0.91 (0.76 - 1.08)
>80
0.85 (0.66 - 1.10)
0-1
0.60 (0.45 - 0.80)
2
0.88 (0.65 - 1.20)
3-6
0.86 (0.71 - 1.04)
Naive
0.84 (0.76 - 0.93)
Experienced
0.87 (0.70 - 1.08)
<66%
0.69 (0.59 - 0.81)
≥66%
0.93 (0.76 - 1.13)
0.2
Ruff CT, et al. Lancet 2013 [in-press]
0.5
Favors NOAC
1
2
Favors Warfarin
p=0.29
p=0.12
p=0.70
p=0.57
p=0.09
p=0.78
p=0.022
12
Low Dose Regimens
Efficacy & Safety Outcomes
Dabigatran 110 mg & Edoxaban 30 mg
Risk Ratio (95% CI)
1.03 (0.84 - 1.27)
Stroke or SEE
p=0.74
1.28 (1.02 - 1.60)
Ischemic Stroke
p=0.045
0.33 (0.23 - 0.46)
Hemorrhagic Stroke
p<0.0001
MI
1.25 (1.04 - 1.50)
p=0.019
0.89 (0.83 - 0.96)
All-Cause Mortality
p=0.003
0.65 (0.43 - 1.00)
Major Bleeding
p=0.05
0.31 (0.24 - 0.41)
ICH
p<0.0001
0.89 (0.57 - 1.37)
GI Bleeding
p=0.58
N=26,107
Heterogeneity
P=NS for outcomes except:
Major Bleeding, p=<0.001
GI Bleeding, p=0.01
0.2
0.5
Favors Low Dose NOAC
1
2
Favors Warfarin
Ruff CT, et al. Lancet 2013 [in-press]
13
Comprehensive Meta-Analysis Comparing the
Efficacy and Safety of NOACs with
Warfarin in AF
An Analysis Including 71,683 Patients from
Four Large Randomized Clinical Trials
Christian T. Ruff, MD, MPH
TIMI Study Group
Brigham and Women’s Hospital
Harvard Medical School
Boston, MA
14
Disclosures
Research Support:
Daiichi Sankyo, AstraZeneca
Consultant and Advisory Boards:
Boehringer Ingelheim, Daiichi Sankyo, BristolMeyers Squibb
15
Stroke Prevention in AF
Warfarin vs. Placebo
AFASAK-1 (671)
SPAF (421)
BAATAF (420)
CAFA (378)
SPINAF (571)
EAFT (439)
All Trials (n=6)
64%
100%
50%
Warfarin Better
0%
-50%
-100%
Warfarin Worse
Hart RG, et al. Ann Intern Med 2007;146:857-867.
16
Pivotal Warfarin-Controlled Trials
Stroke Prevention in AF
Warfarin vs. Placebo
2,900 Patients
NOACs vs. Warfarin
71,683 Patients
ROCKET AF
(Rivaroxaban)
2010
6 Trials of Warfarin vs. Placebo
1989-1993
RE-LY
(Dabigatran)
2009
ENGAGE AF-TIMI 48
(Edoxaban)
2013
ARISTOTLE
(Apixaban)
2011
17
Meta-Analysis

First to contain data from all 4 phase 3 warfarin-controlled trials

Robust sample size
−
−

Precision in assessing relative benefit of NOACs in key clinical
subgroups
Effects of agents on important secondary outcomes
Pooled data for FXa and thrombin inhibitors
−
−
−
Target key coagulation enzymes
Trials share similar design
Agents used interchangeably clinically and grouped together by Guidelines

Separate meta-analysis of low dose dabigatran and edoxaban

Comprehensive picture of the NOACs as a therapeutic option
18
Comparative PK/PD of NOACs
Dabigatran
Rivaroxaban
Apixaban
Edoxaban
IIa (thrombin)
Xa
Xa
Xa
Hours to Cmax
1-3
2-4
3-4
1-2
Half-life, hours
12-17
5-13
12
10-14
80
33*
27
50
Transporters
P-gp
P-gp
P-gp
P-gp
CYP Metabolism, %
None
32
<32
<4
Target
Renal Clearance, %
CYP = cytochrome P450; P-gp = P-glycoprotein
*33% renally cleared; 33% excreted unchanged in urine
Pradaxa [package insert]. Ridgefield, CT: Boehringer Ingelheim Pharmaceuticals, Inc. 2013
Xarelto [package insert]. Titusville, NJ: Janssen Pharmaceuticals, Inc. 2011
Weinz et al. Drug Dispos Metab 2009;37:1056–1064
ELIQUIS Summary of Product Characteristics. Bristol Myers Squibb/Pfizer EEIG, UK
Matsushima et al. Am Assoc Pharm Sci 2011; abstract
Ogata, et al. J Clin Pharmacol 2010;50:743–753
Mendell, et al. Am J Cardiovasc Drugs 2013;13:331–342
Bathala, et al. Drug Metab Dispos 2012;40:2250–2255
19
NOAC SPAF Trials
Drug
# Randomized
Dose (mg)
Frequency
Dose Adjustment
At Baseline
After Randomization
Target INR (Warfarin)
Design
RE-LY
ROCKET-AF
ARISTOTLE
ENGAGE AF
Dabigatran
Rivaroxaban
Apixaban
Edoxaban
18,113
14,266
18,201
21,105
150, 110
20
5
60, 30
Twice Daily
Once Daily
Twice Daily
Once Daily
No
20 → 15
5 → 2.5
60 → 30
30 → 15
0
21
5
25
No
No
No
>9%
2.0-3.0
2.0-3.0
2.0-3.0
2.0-3.0
PROBE*
2x blind
2x blind
2x blind
*PROBE = prospective, randomized, open-label, blinded end point evaluation
Connolly SJ, et al. N Engl J Med 2009;361:1139-1151
Patel MR, et al. N Engl J Med 2011;365:883-891
Granger CB, et al. N Engl J Med 2011;365:981-992
Giugliano RP, et al. N Engl J Med 2013; e-pub ahead of print DOI:10.1056/NEJMoa1310907
20
Baseline Characteristics
RE-LY
(Dabigatran)
ROCKET-AF
(Rivaroxaban)
ARISTOTLE
(Apixaban)
ENGAGE AF
(Edoxaban)
# Randomized
18,113
14,264
18,201
21,105
Age, years
72 ± 9
73 [65-78]
70 [63-76]
72 [64-78]
Female, %
37
40
35
38
Paroxysmal AF
32
18
15
25
VKA naive
50
38
43
41
Aspirin Use
40
36
31
29
CHADS2
0-1
2
3-6
13
33 32
35
87
30
34
53
47
36
Connolly SJ, et al. N Engl J Med 2009;361:1139-1151
Patel MR, et al. N Engl J Med 2011;365:883-891
Granger CB, et al. N Engl J Med 2011;365:981-992
Giugliano RP, et al. N Engl J Med 2013; e-pub ahead of print DOI:10.1056/NEJMoa1310907
21
Trial Metrics
RE-LY
(Dabigatran)
ROCKET-AF
(Rivaroxaban)
ARISTOTLE
(Apixaban)
ENGAGE AF
(Edoxaban)
Median Follow-Up, years
2.0
1.9
1.8
2.8
Median TTR
66
58
66
68
Lost to Follow-Up, N
20
32
90
1
*TTR, time in therapeutic range
Connolly SJ, et al. N Engl J Med 2009;361:1139-1151
Patel MR, et al. N Engl J Med 2011;365:883-891
Granger CB, et al. N Engl J Med 2011;365:981-992
Giugliano RP, et al. N Engl J Med 2013; e-pub ahead of print DOI:10.1056/NEJMoa1310907
22
All NOACS: Stroke or SEE
Risk Ratio (95% CI)
0.66 (0.53 - 0.82)
RE-LY
[150 mg]
ROCKET AF
0.88 (0.75 - 1.03)
ARISTOTLE
0.80 (0.67 - 0.95)
ENGAGE AF-TIMI 48
0.88 (0.75 - 1.02)
[60 mg]
Combined
0.81 (0.73 - 0.91)
[Random Effects Model]
N=58,541
0.5
Heterogeneity p=0.13
p=<0.0001
Favors NOAC
1
Favors Warfarin
2
Ruff CT, et al. Lancet 2013 [in-press] 23
Secondary Efficacy Outcomes
Risk Ratio (95% CI)
Ischemic Stroke
0.92 (0.83 - 1.02)
p=0.10
Hemorrhagic Stroke
0.49 (0.38 - 0.64)
p<0.0001
MI
0.97 (0.78 - 1.20)
p=0.77
All-Cause Mortality
0.90 (0.85 - 0.95)
p=0.0003
0.2
0.5
Favors NOAC
1
2
Favors Warfarin
Heterogeneity p=NS for all outcomes
Ruff CT, et al. Lancet 2013 [in-press] 24
All NOACS: Major Bleeding
Risk Ratio (95% CI)
0.94 (0.82 - 1.07)
RE-LY
[150 mg]
ROCKET AF
1.03 (0.90 - 1.18)
ARISTOTLE
0.71 (0.61 - 0.81)
ENGAGE AF-TIMI 48
0.80 (0.71 - 0.90)
[60 mg]
Combined
0.86 (0.73 - 1.00)
[Random Effects Model]
p=0.06
N=58,498
0.5
Heterogeneity p=0.001
Favors NOAC
1
Favors Warfarin
2
Ruff CT, et al. Lancet 2013 [in-press]
25
Secondary Safety Outcomes
Risk Ratio (95% CI)
0.48 (0.39 - 0.59)
ICH
p<0.0001
1.25 (1.01 - 1.55)
GI Bleeding
0.2
p=0.043
0.5
Favors NOAC
1
2
Favors Warfarin
Heterogeneity
ICH, p=0.22
GI Bleeding, p=0.009
Ruff CT, et al. Lancet 2013 [in-press]
26
Subgroups: Stroke or SEE
Age
Gender
Diabetes
Prior Stroke or TIA
CrCl
CHADS2 Score
VKA Status
Center-Based TTR
Risk Ratio (95% CI)
P-Interaction
<75
0.85 (0.73 - 0.99)
p=0.38
≥75
0.78 (0.68 - 0.88)
Female
0.78 (0.65 - 0.94)
Male
0.84 (0.75 - 0.94)
No
0.83 (0.74 - 0.93)
Yes
0.80 (0.69 - 0.93)
No
0.78 (0.66 - 0.91)
Yes
0.86 (0.76 - 0.98)
<50
0.79 (0.65 - 0.96)
50-80
0.75 (0.66 - 0.85)
>80
0.98 (0.79 - 1.22)
0-1
0.75 (0.54 - 1.04)
2
0.86 (0.70 - 1.05)
3-6
0.80 (0.72 - 0.89)
Naive
0.75 (0.66 - 0.86)
Experienced
0.85 (0.70 - 1.03)
<66%
0.77 (0.65 - 0.92)
≥66%
0.82 (0.71 - 0.95)
0.5
Ruff CT, et al. Lancet 2013 [in-press]
1
Favors NOAC
p=0.52
p=0.73
p=0.30
p=0.12
p=0.76
p=0.31
p=0.60
2
Favors Warfarin
27
Subgroups: Major Bleeding
Age
Gender
Diabetes
Prior Stroke or TIA
CrCl
CHADS2 Score
VKA Status
Center-Based TTR
Risk Ratio (95% CI)
P-Interaction
<75
0.79 (0.67 - 0.94)
p=0.28
≥75
0.93 (0.74 - 1.17)
Female
0.75 (0.58 - 0.97)
Male
0.90 (0.72 - 1.12)
No
0.71 (0.54 – 0.93)
Yes
0.90 (0.78 - 1.04)
No
0.85 (0.72 - 1.01)
Yes
0.89 (0.77 - 1.02)
<50
0.74 (0.52 - 1.05)
50-80
0.91 (0.76 - 1.08)
>80
0.85 (0.66 - 1.10)
0-1
0.60 (0.45 - 0.80)
2
0.88 (0.65 - 1.20)
3-6
0.86 (0.71 - 1.04)
Naive
0.84 (0.76 - 0.93)
Experienced
0.87 (0.70 - 1.08)
<66%
0.69 (0.59 - 0.81)
≥66%
0.93 (0.76 - 1.13)
0.2
Ruff CT, et al. Lancet 2013 [in-press]
0.5
Favors NOAC
1
2
Favors Warfarin
p=0.29
p=0.12
p=0.70
p=0.57
p=0.09
p=0.78
p=0.022
28
ACTIVE-W: Stroke or SEE
TTR ≥ 65%
TTR < 65%
0.10
RR = 1.83
RR = 1.11
P = 0.47
0.08
P < 0.0001
0.06
0.04
Clopi + ASA
C+A
OAC
VKA
0.02
0.02
0.04
0.06
Clopi + ASA
C+A
OAC
0.0
VKA
0.0
Event Rate (%)
0.08
0.10
P-interaction = 0.013
0.0
0.5
1.0
Years
Connolly SJ, et al. Circulation 2008;118:2029-2037
1.5
0.0
0.5
1.0
1.5
Years
29
ACTIVE-W: Major Bleeding
TTR ≥ 65%
TTR < 65%
RR = 0.68
P = 0.08
0.03
0.03
0.04
P = 0.027
OAC
0.02
0.02
C+A
C+A
0.0
0.01
0.01
OAC
0.0
Event Rate (%)
0.05
RR = 1.55
0.04
0.05
P-interaction = 0.0006
0.0
0.5
1.0
Years
Connolly SJ, et al. Circulation 2008;118:2029-2037
1.5
0.0
0.5
1.0
1.5
Years
30
Low Dose Regimens
Efficacy & Safety Outcomes
Dabigatran 110 mg & Edoxaban 30 mg
Risk Ratio (95% CI)
1.03 (0.84 - 1.27)
Stroke or SEE
p=0.74
1.28 (1.02 - 1.60)
Ischemic Stroke
p=0.045
0.33 (0.23 - 0.46)
Hemorrhagic Stroke
p<0.0001
MI
1.25 (1.04 - 1.50)
p=0.019
0.89 (0.83 - 0.96)
All-Cause Mortality
p=0.003
0.65 (0.43 - 1.00)
Major Bleeding
p=0.05
0.31 (0.24 - 0.41)
ICH
p<0.0001
0.89 (0.57 - 1.37)
GI Bleeding
p=0.58
N=26,107
Heterogeneity
P=NS for outcomes except:
Major Bleeding, p=<0.001
GI Bleeding, p=0.01
0.2
0.5
Favors Low Dose NOAC
1
2
Favors Warfarin
Ruff CT, et al. Lancet 2013 [in-press]
31
Conclusions

NOACs significantly reduce stroke (19%)
−
Primarily driven by reduction in hemorrhagic stroke (51%)

NOACs significantly reduce mortality (10%)

Trend toward less bleeding
−
−

Substantial reduction in ICH (52%)
Increased GI bleeding (25%)
The relative efficacy and safety of NOACs consistent across a wide
spectrum of AF patients
−
Even less bleeding when INR not as well controlled

Low dose NOAC regimens reduce mortality and have a very
favorable bleeding profile but more ischemic events

Differences in agents, patients, and trials may not be accounted for
−
Heterogeneity major bleeding and GI bleeding
32
BACK – UP
33
Factor Xa Inhibitors: Stroke or SEE
Risk Ratio (95% CI)
ROCKET AF
0.88 (0.75 - 1.03)
ARISTOTLE
0.80 (0.67 - 0.95)
ENGAGE AF-TIMI 48
0.88 (0.75 - 1.02)
Combined
0.86 (0.78 - 0.94)
[Random Effects Model]
p=0.0011
N=46,443
0.5
1
Favors NOAC
Heterogeneity p=0.65
2
Favors Warfarin
34
Factor Xa Inhibitors: Bleeding
Risk Ratio (95% CI)
1.03 (0.89, 1.18)
ROCKET
0.70 (0.61, 0.81)
ARISTOTLE
0.80 (0.71, 0.90)
ENGAGE AF-TIMI 48
0.83 (0.68, 1.02)
Combined
[Random Effects Model]
N=46,400
Favors NOAC
Heterogeneity p=0.0006
2
1
0.5
Favors Warfarin
35