Trial Outcomes & Findings for Sequential Expansion of Comparative Effectiveness of Anticoagulants (NCT NCT02081807)

NCT ID: NCT02081807

Last Updated: 2019-10-02

Results Overview

The rate of overall stroke (hemorrhagic, ischemic or stroke of uncertain classification ) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: As primary International Classification of Diseases, Ninth Revision (ICD-9) discharge diagnosis (Dx): 431.x Intracerebral hemorrhage (ICH), 433.x1 Occlusion and stenosis of precerebral arteries with cerebral infarction, 434.x1 Occlusion and stenosis of cerebral arteries with cerebral infarction, 436.x Acute, but ill-defined cerebrovascular events.

• Recruitment status: COMPLETED
• Target enrollment: 221228 participants
• Primary outcome timeframe: From October 2010 to September 2015 (the study period)
• Results posted on: 2019-10-02

Participant Flow

Data for the patients included in this cohort study arose from two de-identified research databases, US MarketScan and Optum Research Database

In this non-interventional study based on existing data, no patients were screened.

Participant milestones

Measure	Optum Clinformatics - Dabigatran Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation	Optum Clinformatics - Warfarin Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation	Optum Clinformatics - Rivaroxaban Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation	Optum Clinformatics - Apixaban Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation	MarketScan - Dabigatran Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation	MarketScan - Warfarin Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation	MarketScan - Rivaroxaban Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation	MarketScan - Apixaban Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation
Overall Study STARTED	7668	26149	14852	9112	27456	76055	40207	19729
Overall Study COMPLETED	7668	26149	14852	9112	27456	76055	40207	19729
Overall Study NOT COMPLETED	0	0	0	0	0	0	0	0

Reasons for withdrawal

Withdrawal data not reported

Baseline Characteristics

Race and Ethnicity were not collected from any participant.

Baseline characteristics by cohort

Measure	Optum Clinformatics - Dabigatran n=7668 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation	Optum Clinformatics - Warfarin n=26149 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation	Optum Clinformatics - Rivaroxaban n=14852 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation	Optum Clinformatics - Apixaban n=9112 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation	MarketScan - Dabigatran n=27456 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation	MarketScan - Warfarin n=76055 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation	MarketScan - Rivaroxaban n=40207 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation	MarketScan - Apixaban n=19729 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation	Total n=221228 Participants Total of all reporting groups
Age, Continuous	63.9 Years STANDARD_DEVIATION 11.3 • n=7668 Participants	69.9 Years STANDARD_DEVIATION 11.8 • n=26149 Participants	68.0 Years STANDARD_DEVIATION 11.9 • n=14852 Participants	71.9 Years STANDARD_DEVIATION 11.2 • n=9112 Participants	67.1 Years STANDARD_DEVIATION 12.1 • n=27456 Participants	71.0 Years STANDARD_DEVIATION 12.1 • n=76055 Participants	67.6 Years STANDARD_DEVIATION 12.2 • n=40207 Participants	69.4 Years STANDARD_DEVIATION 12.3 • n=19729 Participants	69.2 Years STANDARD_DEVIATION 12.2 • n=221228 Participants
Sex: Female, Male Female	2315 Participants n=7668 Participants	10143 Participants n=26149 Participants	5483 Participants n=14852 Participants	3911 Participants n=9112 Participants	9774 Participants n=27456 Participants	29693 Participants n=76055 Participants	15014 Participants n=40207 Participants	7747 Participants n=19729 Participants	84080 Participants n=221228 Participants
Sex: Female, Male Male	5353 Participants n=7668 Participants	16006 Participants n=26149 Participants	9369 Participants n=14852 Participants	5201 Participants n=9112 Participants	17682 Participants n=27456 Participants	46362 Participants n=76055 Participants	25193 Participants n=40207 Participants	11982 Participants n=19729 Participants	137148 Participants n=221228 Participants
Race and Ethnicity Not Collected	—	—	—	—	—	—	—	—	0 Participants Race and Ethnicity were not collected from any participant.

PRIMARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of overall stroke (hemorrhagic, ischemic or stroke of uncertain classification ) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: As primary International Classification of Diseases, Ninth Revision (ICD-9) discharge diagnosis (Dx): 431.x Intracerebral hemorrhage (ICH), 433.x1 Occlusion and stenosis of precerebral arteries with cerebral infarction, 434.x1 Occlusion and stenosis of cerebral arteries with cerebral infarction, 436.x Acute, but ill-defined cerebrovascular events.

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=6125 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=6125 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=23323 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=23323 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=10373 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=10373 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) n=25147 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) n=25147 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) n=7065 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) n=7065 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) n=12523 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) n=12523 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Stroke (Hemorrhagic, Ischemic, or Stroke of Uncertain Classification)	0.90 Incidence rate per 100 person-years Interval 0.6 to 1.3	1.01 Incidence rate per 100 person-years Interval 0.66 to 1.5	0.64 Incidence rate per 100 person-years Interval 0.51 to 0.8	0.99 Incidence rate per 100 person-years Interval 0.8 to 1.2	0.80 Incidence rate per 100 person-years Interval 0.56 to 1.11	1.02 Incidence rate per 100 person-years Interval 0.72 to 1.39	0.78 Incidence rate per 100 person-years Interval 0.64 to 0.94	1.09 Incidence rate per 100 person-years Interval 0.9 to 1.31	1.08 Incidence rate per 100 person-years Interval 0.74 to 1.53	1.26 Incidence rate per 100 person-years Interval 0.86 to 1.8	0.62 Incidence rate per 100 person-years Interval 0.44 to 0.84	1.16 Incidence rate per 100 person-years Interval 0.87 to 1.51

PRIMARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of major bleeding (Major intracranial bleeding and major extracranial bleed ) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. For outcome definitions please refer the descriptions section of outcome 8 (major intracranial bleeding) and outcome 9 (major extracranial bleeding).

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=6125 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=6125 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=23323 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=23323 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=10373 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=10373 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) n=25147 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) n=25147 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) n=7065 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) n=7065 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) n=12523 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) n=12523 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Major Bleeding	2.43 Incidence rate per 100 person-years Interval 1.91 to 3.05	5.23 Incidence rate per 100 person-years Interval 4.35 to 6.24	4.36 Incidence rate per 100 person-years Interval 4.0 to 4.74	6.06 Incidence rate per 100 person-years Interval 5.58 to 6.56	6.2 Incidence rate per 100 person-years Interval 5.5 to 7.0	6.3 Incidence rate per 100 person-years Interval 5.5 to 7.2	6.9 Incidence rate per 100 person-years Interval 6.5 to 7.4	7.5 Incidence rate per 100 person-years Interval 7.0 to 8.0	3.33 Incidence rate per 100 person-years Interval 2.69 to 4.07	6.83 Incidence rate per 100 person-years Interval 5.8 to 7.99	4.25 Incidence rate per 100 person-years Interval 3.74 to 4.8	7.96 Incidence rate per 100 person-years Interval 7.15 to 8.83

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of stroke (hemorrhagic, ischemic or stroke of uncertain classification ) or systemic embolism in patients matched on propensity scores and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. For outcome definitions please refer the descriptions section of outcome 4 (Systemic embolism), outcome 5 (Ischemic stroke), outcome 6 (Hemorrhagic stroke) and outcome 7 (Stroke uncertain classification).

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Stroke or Systemic Embolism	0.92 Incidence rate per 100 person-years Interval 0.77 to 1.08	1.13 Incidence rate per 100 person-years Interval 0.95 to 1.33	1.00 Incidence rate per 100 person-years Interval 0.85 to 1.15	1.29 Incidence rate per 100 person-years Interval 1.11 to 1.5	0.81 Incidence rate per 100 person-years Interval 0.63 to 1.02	1.36 Incidence rate per 100 person-years Interval 1.1 to 1.66	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of systemic embolism in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: ICD-9 Diagnoses: 444.x Arterial embolism.

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Systemic Embolism	0.23 Incidence rate per 100 person-years Interval 0.17 to 0.32	0.14 Incidence rate per 100 person-years Interval 0.08 to 0.22	0.21 Incidence rate per 100 person-years Interval 0.15 to 0.29	0.22 Incidence rate per 100 person-years Interval 0.15 to 0.32	0.05 Incidence rate per 100 person-years Interval 0.02 to 0.11	0.17 Incidence rate per 100 person-years Interval 0.09 to 0.29	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of ischemic stroke in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: As primary ICD-9 discharge diagnosis (Dx): 433.x1 Occlusion and stenosis of precerebral arteries with cerebral infarction, ICD-9 Dx 434.x1 Occlusion and stenosis of cerebral arteries with cerebral infarction

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Ischemic Stroke	0.63 Incidence rate per 100 person-years Interval 0.51 to 0.76	0.80 Incidence rate per 100 person-years Interval 0.65 to 0.97	0.63 Incidence rate per 100 person-years Interval 0.52 to 0.75	0.90 Incidence rate per 100 person-years Interval 0.75 to 1.07	0.61 Incidence rate per 100 person-years Interval 0.46 to 0.79	1.03 Incidence rate per 100 person-years Interval 0.8 to 1.29	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of hemorrhagic in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: As primary ICD-9 discharge diagnosis (Dx): 431.x Intracerebral hemorrhage (ICH)

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Hemorrhagic Stroke	0.07 Incidence rate per 100 person-years Interval 0.03 to 0.12	0.19 Incidence rate per 100 person-years Interval 0.12 to 0.28	0.18 Incidence rate per 100 person-years Interval 0.12 to 0.25	0.18 Incidence rate per 100 person-years Interval 0.12 to 0.26	0.15 Incidence rate per 100 person-years Interval 0.09 to 0.25	0.18 Incidence rate per 100 person-years Interval 0.1 to 0.31	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of stroke uncertain classification in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: As primary ICD-9 discharge diagnosis (Dx): ICD-9 Dx code 436.x (acute, but ill-defined cerebrovascular disease).

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Stroke Uncertain Classification	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.02	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.02	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.01	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.02	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.03	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.04	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of major intracranial bleeding in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: ICD-9 diagnosis: 430.x Subarachnoid hemorrhage (SAH), 431.x Intracerebral hemorrhage (ICH), 432.x other and unspecified intracranial hemorrhage including 432.1x - subdural hemorrhage

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Major Intracranial Bleeding	0.21 Incidence rate per 100 person-years Interval 0.15 to 0.3	0.58 Incidence rate per 100 person-years Interval 0.46 to 0.73	0.44 Incidence rate per 100 person-years Interval 0.35 to 0.55	0.66 Incidence rate per 100 person-years Interval 0.53 to 0.81	0.45 Incidence rate per 100 person-years Interval 0.32 to 0.6	0.69 Incidence rate per 100 person-years Interval 0.51 to 0.92	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of major extracranial bleeding (Major upper GI bleed, major lower and unspecified GI bleed, major urogenital bleed, major other bleed) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. For outcome definitions please refer the descriptions section of outcome 11 (Major upper gastrointestinal bleeding), outcome 12 (Major lower gastrointestinal bleeding), outcome 13 (Major urogenital bleeding) and outcome 14 (Other major bleeding).

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Major Extra-cranial Bleeding	3.78 Incidence rate per 100 person-years Interval 3.48 to 4.1	5.37 Incidence rate per 100 person-years Interval 4.96 to 5.79	6.31 Incidence rate per 100 person-years Interval 5.94 to 6.69	6.54 Incidence rate per 100 person-years Interval 6.11 to 6.99	3.51 Incidence rate per 100 person-years Interval 3.13 to 3.92	6.92 Incidence rate per 100 person-years Interval 6.31 to 7.58	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of major gastrointestinal (GI) bleeding (Major upper GI bleeding, major lower/unspecified GI bleeding) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. For outcome definitions please refer the descriptions section of outcome 11 (Major upper gastrointestinal bleeding) and outcome 12 (Major lower gastrointestinal bleeding).

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Major Gastrointestinal (GI) Bleeding	2.39 Incidence rate per 100 person-years Interval 2.15 to 2.65	2.82 Incidence rate per 100 person-years Interval 2.53 to 3.13	3.82 Incidence rate per 100 person-years Interval 3.53 to 4.12	3.43 Incidence rate per 100 person-years Interval 3.13 to 3.76	1.99 Incidence rate per 100 person-years Interval 1.7 to 2.3	3.59 Incidence rate per 100 person-years Interval 3.16 to 4.07	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of Major upper gastrointestinal bleeding in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: ICD-9 diagnoses: 531.0x, 531.2x, 531.4x, 531.6x, 532.0x, 532.2x, 532.4x, 532.6x, 533.0x, 533.2x, 533.4x, 533.6x, 534.0x, 534.2x, 534.4x, 534.6x, 578.0 OR ICD-9 procedure code 44.43 (endoscopic control of gastric or duodenal bleeding) OR Current Procedural Terminology (CPT) code 43255 (upper gastrointestinal endoscopy including esophagus, stomach, and either the duodenum and/or jejunum as appropriate with control of bleeding, any method).

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Major Upper Gastrointestinal Bleeding	0.51 Incidence rate per 100 person-years Interval 0.41 to 0.64	0.90 Incidence rate per 100 person-years Interval 0.74 to 1.08	0.99 Incidence rate per 100 person-years Interval 0.85 to 1.15	0.88 Incidence rate per 100 person-years Interval 0.73 to 1.05	0.64 Incidence rate per 100 person-years Interval 0.49 to 0.83	0.89 Incidence rate per 100 person-years Interval 0.68 to 1.14	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of lower gastrointestinal bleeding in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: Lower GI/unspecified GI site bleeds :Diverticulosis of small intestine with hemorrhage: 562.02, Diverticulitis of small intestine with hemorrhage: 562.03, Diverticulosis of colon with hemorrhage: 562.12, Diverticulitis of colon with hemorrhage: 562.13, Hemorrhage of rectum and anus: 569.3x, Angiodysplasia of intestine with hemorrhage: 569.85, Blood in stool: 578.1x, Hemorrhage of GI tract, unspecified: 578.9

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Major Lower Gastrointestinal Bleeding	2.18 Incidence rate per 100 person-years Interval 1.96 to 2.43	2.48 Incidence rate per 100 person-years Interval 2.21 to 2.77	3.40 Incidence rate per 100 person-years Interval 3.13 to 3.69	3.09 Incidence rate per 100 person-years Interval 2.8 to 3.39	1.61 Incidence rate per 100 person-years Interval 1.36 to 1.9	3.17 Incidence rate per 100 person-years Interval 2.76 to 3.62	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of major urogenital bleeding in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: ICD-9 diagnoses: Hematuria: ICD-9 Dx: 599.7, Excessive/frequent menstruation: ICD-9 Dx 626.2x and secondary diagnosis indicating acute bleeding: anemia (280.0, 285.1, 285.9). Across databases, only one event for dabigatran versus no event among warfarin initiators observed. Across database, four events for rivaroxaban versus no event among warfarin initiators observed. Across database, no events for apixaban and warfarin observed. Therefore HR estimate is not possible.

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Major Urogenital Bleeding	0.01 Incidence rate per 100 person-years Interval 0.0 to 0.03	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.02	0.02 Incidence rate per 100 person-years Interval 0.01 to 0.06	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.02	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.03	0.00 Incidence rate per 100 person-years Interval 0.0 to 0.04	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of Other major bleeding in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: Other major bleeds: Hemathrosis: 719.1x, Hemopericardium: 423.0x, Hemoptysis: 786.3x, Epistaxis: 784.7x, Hemorrhage not specified 459.0x, Acute posthemorrhagic anemia 285.1x

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Other Major Bleeding	2.32 Incidence rate per 100 person-years Interval 2.08 to 2.57	3.65 Incidence rate per 100 person-years Interval 3.33 to 4.01	4.33 Incidence rate per 100 person-years Interval 4.02 to 4.65	4.59 Incidence rate per 100 person-years Interval 4.24 to 4.97	2.54 Incidence rate per 100 person-years Interval 2.22 to 2.9	4.86 Incidence rate per 100 person-years Interval 4.35 to 5.41	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of Transient Ischemic Attack (TIA) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: ICD-9 Dx code 435.xx (transient cerebral ischemia) as the principal (primary) discharge diagnosis

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Transient Ischemic Attack (TIA)	0.21 Incidence rate per 100 person-years Interval 0.15 to 0.3	0.43 Incidence rate per 100 person-years Interval 0.32 to 0.56	0.26 Incidence rate per 100 person-years Interval 0.19 to 0.34	0.31 Incidence rate per 100 person-years Interval 0.23 to 0.42	0.22 Incidence rate per 100 person-years Interval 0.14 to 0.34	0.32 Incidence rate per 100 person-years Interval 0.2 to 0.48	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of Myocardial infarction (MI) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: ICD-9 Dx 410.X (acute myocardial infarction) excluding 410.x2 (subsequent episode of care), as the principal (primary) or the next (secondary) diagnosis AND a length of stay (LOS) between 3-180 days, or death if LOS is < 3 days

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Myocardial Infarction (MI)	0.48 Incidence rate per 100 person-years Interval 0.38 to 0.6	0.60 Incidence rate per 100 person-years Interval 0.48 to 0.75	0.53 Incidence rate per 100 person-years Interval 0.43 to 0.64	0.65 Incidence rate per 100 person-years Interval 0.52 to 0.8	0.57 Incidence rate per 100 person-years Interval 0.43 to 0.75	0.74 Incidence rate per 100 person-years Interval 0.55 to 0.97	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of Venous Thromboembolism (VTE) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. For outcome definitions please refer the descriptions section of outcome 18 (Deep vein thrombosis (DVT)) and outcome 19 (Pulmonary Embolism (PE)).

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Venous Thromboembolism (VTE)	0.67 Incidence rate per 100 person-years Interval 0.55 to 0.81	1.14 Incidence rate per 100 person-years Interval 0.96 to 1.34	1.20 Incidence rate per 100 person-years Interval 1.04 to 1.37	1.65 Incidence rate per 100 person-years Interval 1.45 to 1.88	0.57 Incidence rate per 100 person-years Interval 0.43 to 0.75	1.21 Incidence rate per 100 person-years Interval 0.96 to 1.5	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of DVT in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: Validated algorithm: ICD-9 451.1x, ICD-9 451.2x,ICD-9 451.81, ICD-9 451.9x, ICD-9 453.1x, ICD-9 453.2x, ICD-9 453.8x, ICD-9 453.9x ; Not in the validated algorithm but will be included following Mini-Sentinel recommendation for VTE outcome: ICD-9 453.40 (Venous embolism and thrombosis of unspecified deep vessels of lower extremity (includes DVT), ICD-9 453.41 (Venous embolism and thrombosis of deep vessels of proximal lower extremity (includes femoral, iliac, popliteal, thigh, and upper leg), ICD-9 453.42 (Venous embolism and thrombosis of deep vessels of distal lower extremity (includes calf, lower leg, peroneal, and tibia), ICD-9 453.0 (Hepatic vein thrombosis)

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Deep Vein Thrombosis (DVT)	0.46 Incidence rate per 100 person-years Interval 0.36 to 0.58	0.78 Incidence rate per 100 person-years Interval 0.64 to 0.95	0.78 Incidence rate per 100 person-years Interval 0.66 to 0.93	1.09 Incidence rate per 100 person-years Interval 0.93 to 1.28	0.41 Incidence rate per 100 person-years Interval 0.29 to 0.56	0.75 Incidence rate per 100 person-years Interval 0.57 to 0.99	—	—	—	—	—	—

SECONDARY outcome

Timeframe: From October 2010 to September 2015 (the study period)

Population: Patients with non-valvular atrial fibrillation (NVAF) and new initiators of dabigatran and warfarin during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database or MarketScan before treatment initiation and matched on propensity scores (PS). Primary analysis represented an 'as treated' approach.

The rate of Pulmonary Embolism (PE) in patients who are matched on propensity score and calendar quarter of initiation. Event rates were calculated as the total number of patients in each treatment group who had the outcome during follow-up divided by the total person-years at risk in the cohort. The outcome definition includes following: ICD-9 415.1x (pulmonary embolism and infarction)

Outcome measures

Measure	Optum Clinformatics - Dabigatran vs. Warfarin (Dabigatran) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Dabigatran vs. Warfarin (Warfarin) n=29448 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with dabigatran	MarketScan - Dabigatran vs. Warfarin (Dabigatran) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of dabigatran, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Dabigatran vs. Warfarin (Warfarin) n=35520 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during October 2010 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with dabigatran	Optum Clinformatics - Rivaroxaban vs. Warfarin (Rivaroxaban) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Rivaroxaban vs. Warfarin (Warfarin) n=19588 Participants Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with rivaroxaban	MarketScan - Rivaroxaban vs. Warfarin (Rivaroxaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of rivaroxaban, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Rivaroxaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during July 2011 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with rivaroxaban	Optum Clinformatics - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with warfarin	Optum Clinformatics - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in Optum Clinformatics database before treatment initiation and matched on propensity scores with apixaban	MarketScan - Apixaban vs. Warfarin (Apixaban) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of apixaban, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with warfarin	MarketScan - Apixaban vs. Warfarin (Warfarin) Patients with non- valvular atrial fibrillation (NVAF) and new initiators of warfarin, during January 2013 to September 2015 who were enrolled for a minimum of 12 months in MarketScan before treatment initiation and matched on propensity scores with apixaban
Pulmonary Embolism (PE)	0.29 Incidence rate per 100 person-years Interval 0.21 to 0.38	0.46 Incidence rate per 100 person-years Interval 0.35 to 0.59	0.54 Incidence rate per 100 person-years Interval 0.44 to 0.66	0.81 Incidence rate per 100 person-years Interval 0.67 to 0.97	0.22 Incidence rate per 100 person-years Interval 0.14 to 0.34	0.62 Incidence rate per 100 person-years Interval 0.45 to 0.83	—	—	—	—	—	—

Adverse Events

Optum Clinformatics - Dabigatran

Serious events: 0 serious events

Other events: 0 other events

Deaths: 0 deaths

Optum Clinformatics - Warfarin

Serious events: 0 serious events

Other events: 0 other events

Deaths: 0 deaths

Optum Clinformatics - Rivaroxaban

Serious events: 0 serious events

Other events: 0 other events

Deaths: 0 deaths

Optum Clinformatics - Apixaban

Serious events: 0 serious events

Other events: 0 other events

Deaths: 0 deaths

MarketScan - Dabigatran

Serious events: 0 serious events

Other events: 0 other events

Deaths: 0 deaths

MarketScan - Warfarin

Serious events: 0 serious events

Other events: 0 other events

Deaths: 0 deaths

MarketScan - Rivaroxaban

Serious events: 0 serious events

Other events: 0 other events

Deaths: 0 deaths

MarketScan - Apixaban

Serious events: 0 serious events

Other events: 0 other events

Deaths: 0 deaths

Serious adverse events

Adverse event data not reported

Other adverse events

Adverse event data not reported

Additional Information

Boehringer Ingelheim, Call Center

Boehringer Ingelheim

Phone: 1-800-243-0127

Email: clintriage.rdg@boehringer-ingelheim.com

Results disclosure agreements

• Principal investigator is a sponsor employee Boehringer Ingelheim (BI) acknowledges that investigators have the right to publish the study results. Investigators shall provide BI with a copy of any publication or presentation for review prior to any submission. Such review will be done with regard to proprietary information, information related to patentable inventions, medical, scientific, and statistical accuracy within 60 days. BI may request a delay of the publication in order to protect BI's intellectual property rights.
• Publication restrictions are in place

Restriction type: OTHER