Real-World Evidence of XARELTO: Stroke Risk Reduction in Nonvalvular Atrial Fibrillation

SUMMARY

• Randomized controlled trials represent the highest level of rigor and evidence as compared to real-world studies, which could have certain biases that may not be controlled for with statistical techniques. It is important to consider study methodology strengths and weaknesses when interpreting real-world studies. The majority of published real-world data demonstrates a consistent balance relative to the safety and efficacy of XARELTO when used in clinical practice.^1,2
  • Some limitations of real-world data studies include:
    • Residual confounding due to unmeasured factors that are either not available in the database or were not accounted for in the statistical analysis.
    • Potential misclassification of exposure and outcomes.
    • Composition of the population may limit the generalizability of the findings.
• The international XANTUS study reported an incidence of major bleeding (MB) of 2.1 per 100 patient-years (PY), and an incidence of stroke and systemic embolism (SE) of 0.8 per 100 PY in patients with nonvalvular atrial fibrillation (NVAF) in routine clinical practice.³
• A United States (US) postmarketing safety surveillance study evaluating the use of XARELTO in patients with NVAF reported an incidence of MB of 2.89 per 100 PY.⁴
• Bradley et al (2026)⁵ conducted a cohort study assessing the comparative safety and effectiveness of most commonly prescribed non-vitamin K oral anticoagulants (NOACs) in US patients with NVAF aged <65 years. The study utilized data from the Food and Drug Administration (FDA) Sentinel System, comprising 4 national commercial insurance plans and 1 state Medicaid partner. Patients who initiated standard-dose NOAC therapy between October 19, 2010, and February 28, 2022, and were aged 21 to 64 years were included in the study. Eligibility criteria included a qualifying NOAC dispensing (index date), at least 6 months of continuous medical and drug coverage before the index date, age 21 to 64 years, and a recorded inpatient or outpatient diagnosis of NVAF or nonvalvular atrial flutter within the previous 6 months based on International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) or International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes.
  • In the comparison of XARELTO vs apixaban, the weighted incidence rate per 1000 person-years was 11.6 vs 6.4 for major extracranial bleeding (hazard ratio [HR], 1.91; 95% confidence interval [CI], 1.56-2.34), 10 vs 5.4 for gastrointestinal (GI) bleeding (HR, 1.92; 95% CI, 1.54-2.39), 1.8 vs 1.1 for intracranial hemorrhage (ICH; HR, 1.63; 95% CI, 0.99-2.7), and 4.06 vs 3.88 for thromboembolic stroke (HR, 1.05; 95% CI, 0.77-1.44).
  • In the comparison of XARELTO vs dabigatran, the weighted incidence rate per 1000 person-years was 9.9 vs 7.3 for major extracranial bleeding (HR, 1.42; 95% CI, 0.98-2.07), 8.3 vs 6.6 for GI bleeding (HR, 1.32; 95% CI, 0.89-1.96), 1.5 vs 1.4 for ICH (HR, 1.18; 95% CI, 0.52-2.67), and 3.38 vs 6.46 for thromboembolic stroke (HR, 0.61; 95% CI, 0.39-0.94).
• Fang et al (2025)⁶ conducted a nationwide, retrospective, population-based cohort study using the Taiwan National Health Insurance Research Database (NHIRD) to compare the effectiveness and safety of 4 direct oral anticoagulants (DOACs) in patients with atrial fibrillation (AF) and dementia.
  • In the comparison of XARELTO vs apixaban, HR (95% CI) was 1.05 (0.95-1.17) for the composite outcome, 1.11 (0.86-1.43) for ischemic stroke, 0.97 (0.57-1.66) for acute myocardial infarction, 0.7 (0.46-1.06) for ICH, 1.05 (0.87-1.26) for MB, and 1.04 (0.91-1.18) for all-cause mortality.
  • In the comparison of XARELTO vs edoxaban, HR (95% CI) was 1.02 (0.9-1.1) for the composite outcome, 1.04 (0.76-1.43) for ischemic stroke, 1.31 (0.65-2.65) for acute myocardial infarction, 1.26 (0.82-1.96) for ICH, 1.15 (0.93-1.42) for MB, and 1.03 (0.89-1.19) for all-cause mortality.
  • In the comparison of XARELTO vs dabigatran, HR (95% CI) was 0.82 (0.73-0.91) for the composite outcome, 0.95 (0.74-1.22) for ischemic stroke, 1.16 (0.65-2.06) for acute myocardial infarction, 0.93 (0.6-1.45) for ICH, 0.99 (0.82-1.19) for MB, and 0.77 (0.67-0.88) for all-cause mortality.
• Additional claims database analyses are summarized below.^7-20
• Additional citations, including studies evaluating healthcare costs/utilization,^21-23 identified during a literature search are included in the REFERENCES section for your review.^24-33

PROSPECTIVE OBSERVATIONAL STUDY

The XANTUS (XARELTO for Prevention of Stroke in Patients with Atrial Fibrillation)^3,34 program includes 4 separate phase 4 studies with the same study protocol, designed to cover 4 regions. Each trial in the XANTUS program is a prospective, international, observational, post-authorization, non-interventional, single-arm cohort study designed to assess the safety and efficacy of XARELTO for stroke prevention in NVAF in routine clinical practice.

Study Design: The study screened 10,934 patients from 311 sites. The type, duration, and dose of drug was determined by the treating physician. Data was collected at the start of therapy, at hospital discharge, and every 3 months thereafter for 1 year. Patients who discontinued therapy early were followed up for 30 days after the last dose of XARELTO.

Outcome Measures: The primary outcome was to assess the safety of XARELTO in routine practice, recorded as adverse events (AEs) or serious AEs (SAEs), including MB (as defined by the International Society on Thrombosis and Haemostasis [ISTH] criteria), all-cause death, and any other AEs and SAEs. Secondary outcome measures included symptomatic thromboembolic events (stroke, noncentral nervous system SE, transient ischemic attack [TIA], and myocardial infarction [MI], and non-MB events). Other outcomes included treatment persistence, patient satisfaction, healthcare resource utilization (HCRU), and details of treatment interruption and interventions.

Results: There were 6784 patients included in the safety population; 5336 (78.7%) patients received XARELTO 20 mg once daily, 1410 (20.8%) received XARELTO 15 mg once daily, and 35 (0.5%) patients received a XARELTO dose besides the 15 or 20 mg once daily dose. The mean observation period was 329 days.

• The mean patient age was 71.5 years and mean body mass index was 28.3 kg/m². The mean and median CHADS₂ (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, and stroke [double weight]) score was 2 each. The mean CHA₂DS₂-VASc (congestive heart failure, hypertension, age ≥75 [doubled], diabetes, stroke [doubled], vascular disease, age 65 to 74 and sex category [female]) score was 3.4; the median CHA₂DS₂-VASc score was 3.
• The rate of MB, thromboembolic events, and all-cause death was low and increased over time. A total of 6522 patients (96.1%) did not experience any outcomes of treatment-emergent MB, all-cause death, or stroke/SE.
• MB was mostly treated using conservative methods while nonspecific reversal agents were rarely used.
• See Table: Incidence of Treatment-Emergent Thromboembolic and Bleeding Events and All-Cause Death.

• The adjudicated cause of death was due primarily to cardiovascular causes (41.5%), followed by cancer (19.5%).
• The incidence of MB events and symptomatic thromboembolic events was found to increase with age.
• See Table: Safety and Efficacy Outcomes Based on Age.

• Treatment persistence remained high over the 1-year study period, with a discontinuation rate at the end of the observation period of 20.1%.

A pre-planned pooled analysis of the XANTUS, XANAP (Xarelto for Prevention of Stroke in Patients With Atrial Fibrillation in Asia), and XANTUS-EL (Xarelto for Prevention of Stroke in Patients With Atrial Fibrillation in Latin America and EMEA Region) registries was performed to assess the global safety profile of XARELTO in routine clinical practice.

• Primary outcomes were treatment-emergent MB, AEs/SAEs, and all-cause death.
• A total of 11,122 patients were enrolled from 554 centers worldwide. The mean patient age was 70.5 years and 57.1% of patients were male.
• There were 190 treatment-emergent MB events in 172 patients. Overall, 179 patients experienced 196 symptomatic thromboembolic events (1.8 events/100 PY).
• A total of 3796 patients (34.1%) had a treatment-emergent AE, and 1647 patients (14.6%) had a treatment-emergent SAE. The incidence of all-cause death was 187 (1.7%).
• Event rates (events/100 PY) were as follows: MB, 1.7 (95% CI, 1.5-2; lowest: Latin America, 0.7; highest: Western Europe, Canada, and Israel, 2.3); all-cause death, 1.9 (95% CI, 1.6-2.2; lowest: Eastern Europe, 1.5; highest: Latin America, Middle East, and Africa, 2.7); and stroke or SE, 1 (95% CI, 0.8-1.2; lowest: Latin America, 0; highest: East Asia, 1.8).
• Key limitations in these studies included:
  • The studies in the program were single-arm, open-label designed.
  • Interference with patient management was not allowed because of the non-interventional nature of the program.

POSTMARKETING SAFETY SURVEILLANCE

A 5-year observational, postmarketing safety surveillance study is being conducted to provide longitudinal safety data of XARELTO in patients with NVAF.⁴ Electronic medical records from the US Department of Defense database were reviewed to assess patients for MB hospitalizations among XARELTO users. MB events were identified using primary hospital discharge diagnoses and included GI bleeding, hemorrhagic strokes and other intracranial bleeds, genitourinary bleeding, and bleeding at other sites, occurring during and up to 7 days post discontinuation of XARELTO use. Data has been presented for the first 2.5 years (January 1, 2013 - June 30, 2015) of the ongoing 5-year study.

• A total of 44,793 XARELTO users were identified during the 2.5-year period (1352 MB events in 1293 patients). The overall MB incidence rate, based on a patient’s first MB event, was 2.84 (95% CI, 2.69-3) per 100 PY.
• A total of 35 patients in the MB group died during hospitalization, for a fatal bleeding rate of 0.10 per 100 PY (95% CI, 0.07-0.15). Of the 35 patients who died, 26 (74.3%) had an ICH, and 9 (25.7%) had GI bleeding.⁴ The mean age at death was 80.3 years.³⁵
• In the NVAF cohort older patients (78.7 years vs 76.2 years), those with hypertension (87.4% vs 66.8%), coronary heart disease (51.4% vs 30.9%), heart failure (37.7% vs 20.2%), renal disease (24.1% vs 15.2%), and diabetes mellitus (36.5% vs 26.6%) along with greater mean CHA₂DS₂-VASc scores at baseline (4.5 vs 3.5), had higher rates of MB compared to no MB events, respectively.⁴

MB results are presented in Table: MB Event Rates, Locations, and Hospitalization Data in the NVAF Cohort.

MB rates in the pivotal phase 3 ROCKET AF study, which compared XARELTO to warfarin in patients with NVAF, reported an event rate of 3.6 per 100 PY.³⁷ Per the algorithm, bleeding events determined to be trauma-related were excluded.

CLAIMS DATABASE ANALYSES

Bradley et al (2026)⁵ conducted a cohort study assessing the comparative safety and effectiveness of most commonly prescribed NOACs in US patients with NVAF aged <65 years. The study utilized data from the FDA Sentinel System, comprising 4 national commercial insurance plans and 1 state Medicaid partner. Patients who initiated standard-dose NOAC therapy between October 19, 2010, and February 28, 2022, and were aged 21 to 64 years were included in the study. Eligibility criteria included a qualifying NOAC dispensing (index date), at least 6 months of continuous medical and drug coverage before the index date, age 21 to 64 years, and a recorded inpatient or outpatient diagnosis of NVAF or nonvalvular atrial flutter within the previous 6 months based on ICD-9-CM or ICD-10-CM codes. Data analysis was performed between December 2022 and August 2023.

• New-user cohorts were identified for each of the following pairwise NOAC comparisons: XARELTO (n=57,932) vs apixaban (n=96,057), XARELTO (n=57,399) vs dabigatran (n=20,188), and dabigatran (n=20,163) vs apixaban (n=96,668).
• Across all cohorts, the median follow-up duration was 62 days (interquartile range [IQR], 32-126).
• Incidence rates of bleeding events were numerically higher among XARELTO users than among apixaban and dabigatran users. See Table: Outcome Events, Weighted Incidence Rates, and HRs for Bleeding Events and Stroke in XARELTO Pairwise Comparisons.

• Key limitations of the study included the following:
  • Despite the use of a new-user active-comparator approach with extensive propensity score adjustment, the possibility of residual confounding could not be excluded.
  • Outcome classification was based on diagnosis codes captured in administrative claims data rather than adjudicated clinical events, introducing the potential for misclassification.
  • Administrative claims databases did not capture detailed clinical characteristics, lifestyle factors, or use of over-the-counter medications such as aspirin, which may have affected treatment selection and outcomes.

Fang et al (2025)⁶ conducted a nationwide, retrospective, population-based cohort study to compare the effectiveness and safety of 4 DOACs in patients with AF and dementia. Data were obtained from Taiwan NHIRD for the period from 2009 to 2020. The index date was defined as the date of initiation of oral anticoagulant therapy following an AF diagnosis.

• The final study population included 9922 patients.
• Of these, 3873 received XARELTO, 2302 received dabigatran, 2181 received apixaban, and 1566 received edoxaban.
• Propensity score matching was performed to create 6 matched cohorts for the main analysis as follows:
  • Dabigatran vs XARELTO, n=4596
  • Edoxaban vs XARELTO, n=3132
  • XARELTO vs apixaban, n=4360
  • Dabigatran vs apixaban, n=4334
  • Dabigatran vs edoxaban, n=3132
  • Edoxaban vs apixaban, n=3132
• See Table: Outcomes in Patients With AF and Dementia: XARELTO Pairwise Comparisons.

• In subgroup analyses, no significant interaction was observed for the composite outcome when dabigatran was compared with apixaban, edoxaban, or XARELTO across subgroups stratified by age, sex, history of stroke, and prior antiplatelet use.
• Key limitations of the study included the following:
  • Dementia was identified using International Classification of Diseases (ICD) codes; therefore, dementia severity and its potential impact on DOAC effectiveness and safety could not be assessed.
  • The sample size was insufficient to evaluate outcomes according to standard-dose vs low-dose regimens for individual DOACs.
  • The study population consisted almost exclusively of Asian patients, which may limit generalizability to other populations.
  • Unmeasured confounders, including clinician prescribing preferences, patient frailty, and perceived bleeding risk, may have influenced DOAC selection.
  • Renal function data, including creatinine clearance (CrCl) and glomerular filtration rate (GFR), were unavailable. Although chronic kidney disease (CKD) was included in the propensity score matching to balance baseline characteristics between DOAC pairs, the absence of renal function data may have introduced confounding by contraindication.
    • The absence of renal function measures is particularly relevant for dabigatran because of its predominant renal elimination and the association between impaired renal function and adverse outcomes.

Mitchell et al (2025)⁷ conducted a retrospective cohort study with a new-user design in patients with AF aged ≥75 years who were newly initiated on anticoagulant therapy to compare safety (MB, non-MB, myocardial infarction, and death) and effectiveness (stroke) outcomes between periods of exposure and no exposure to anticoagulant treatment at the time of the event and across different anticoagulant exposures. Data were obtained from the UK Clinical Practice Research Datalink (CPRD) for the period from January 1, 2013, to December 27, 2017. Eligible patients were aged ≥75 years on the index date and had research-standard data for at least 1 year before the index date and for 1 month of follow-up after the index date.

• The index anticoagulant was warfarin for 10,149 patients and a DOAC for 10,018 patients.
• The median duration of follow-up was 1.49 years (IQR, 0.71-2.55).
• During the study period, 3940 (19.5%) patients experienced at least 1 period without anticoagulant exposure.
• A total of 5603 (27.8%) patients experienced a change in anticoagulant exposure, including 3772 (18.7%) who switched once, 1154 (5.7%) who switched twice, and 677 (3.3%) who switched more than twice.
• See Tables: Incidence Rates of Outcomes per 1000 Person-Years and Unadjusted HRs for Patients With AF Aged ≥75 Years: Exposed vs Unexposed and XARELTO vs Warfarin Exposure Comparisons and HRs for Effectiveness and Safety Outcomes Associated With XARELTO, Apixaban, and All DOACs Compared With Warfarin: Cox Proportional Hazards Models Adjusted for Age, Sex, and Relevant Comorbidities.
• The incidence of MB and GI bleeding was higher with DOACs as a group than with warfarin. When analyzed individually, the incidence of MB, non-MB, and GI bleeding was lower with apixaban and higher with XARELTO than with warfarin.
• Compared with warfarin, XARELTO was associated with
  • A 30% higher risk of MB (HR, 1.33; 95% CI, 1.15-1.55)
  • A 30% higher risk of non-MB (HR, 1.29; 95% CI, 1.16-1.44)
  • A 46% higher risk of GI bleeding (HR, 1.46; 95% CI, 1.23-1.73)
• Neither apixaban (HR, 1.01; 95% CI, 0.84-1.21) nor XARELTO (HR, 1.14; 95% CI, 0.96-1.35) was associated with a significant change in the risk of death compared with warfarin.

• Key limitations of the study included the following:
  • Anticoagulant exposure was estimated; therefore, temporality may have been affected if an outcome event resulted in oral anticoagulant discontinuation but was recorded during a period of nonexposure.

Natale et al (2025)⁸ conducted a retrospective cohort study to evaluate the effectiveness, safety, HCRU, and costs among patients with NVAF and obstructive sleep apnea (OSA) who were initiated treatment with XARELTO compared with warfarin in a US healthcare claims database. The study used data from the IQVIA PharMetrics^® Plus adjudicated health plan claims database and was conducted between November 1, 2010, and December 23, 2022. Eligible patients were aged ≥18 years on index date and had ≥1 pharmacy claim for XARELTO or warfarin between November 1, 2011, and December 23, 2022, with the date of the first claim designated as the index date.

• A total of 14,765 patients were included, of whom 9133 received XARELTO and 5632 received warfarin.
• The mean (standard deviation) follow-up duration was 2.30 (2.15) years overall, 2.31 (2.12) years among XARELTO users, and 2.29 (2.19) years among warfarin users.
• See Table: Clinical Outcomes With XARELTO vs Warfarin in Patients With NVAF and Sleep Apnea (Propensity Score Overlap Weighted).
• In the intent-to-treat (ITT) analysis, XARELTO was consistently associated with a reduced incidence rate of stroke or SE compared with warfarin across patient subgroups, with significant reductions observed among patients aged <65 years, males, nonobese patients, patients with diabetes, patients without heart failure, patients with a history of prior stroke, and patients with metabolic syndrome.
• Owing to the small proportion of patients who were confirmed to have received positive airway pressure (PAP) therapy, subgroup analyses according to the PAP therapy status were not performed.
• Results of the on-treatment analysis were similar, except that a significant reduction in stroke/SE was observed among patients without diabetes, whereas the reduction among patients with diabetes was not significant.
• Consistent with the overall population, the incidence rate of MB with XARELTO vs warfarin was not significantly different across patient subgroups in both the ITT and on-treatment analyses.
• Patients with NVAF and OSA receiving XARELTO demonstrated higher adherence vs those receiving warfarin (49.6% vs 43.8%; odds ratio, 1.13; 95% CI, 1.07-1.2; P<0.0001).

• Key limitations of the study included the following:
  • Administrative claims data may be subject to coding errors and inconsistencies.
  • Residual confounding could not be excluded despite cohort-balancing efforts.
  • Findings may not be generalizable to the overall US population because the study primarily included younger patients with commercial or self-insurance coverage.
  • Claims data did not permit assessment of OSA severity and its potential impact on outcomes.
  • Prescription claims do not confirm whether medication was taken as prescribed, and over-the-counter medications such as aspirin may be underreported.
  • Limited laboratory data prevented assessment of time in therapeutic international normalized ratio (INR) range among warfarin users.

Chun et al (2024)⁹ conducted a study to evaluate outcomes and severity of ischemic stroke among patients with AF receiving vitamin K antagonists and DOACs. In addition, patients receiving DOACs were divided into 4 treatment groups, and outcomes of each patient were compared. Data were obtained from the Korean National Health Insurance database for the period from 2002 to 2019. Patients who received oral anticoagulants through at least 2 prescriptions were identified.

• The study included 9578 patients who experienced an ischemic stroke during anticoagulation therapy between 2014 and 2019.
• Of the included patients, 3343 (34.9%) received warfarin and 6235 (65.1%) received DOACs.
• Among DOAC users, 1702 received XARELTO, 965 received dabigatran, 2320 received apixaban, and 1248 received edoxaban.
• See Table: Hazards for Mortality, Intracranial Hemorrhage, and Cerebral Disability Associated With XARELTO.

• Key limitations of the study included the following:
  • Clinical information was unavailable because the analysis was based on claims data.
  • National Institutes of Health Stroke Scale (NIHSS) scores were unavailable; therefore, stroke severity could not be assessed using this measure.
  • INR information was unavailable, preventing assessment of the impact of inappropriate anticoagulation treatment on stroke severity among warfarin users.
  • CrCl data were unavailable, limiting evaluation of its effect on anticoagulant effectiveness.
  • The study did not analyze the development of new stroke after ischemic stroke as the primary outcome variable. Failure to obtain this information had little impact on the study results. In cases where stroke occurred despite anticoagulant therapy, causes other than cardioembolism should be considered. However, due to the lack of clinical information, the effect of these causes relative to other causes could not be analyzed.
  • The study was conducted in a single country and a specific racial group, which may limit generalizability of the findings.

Perreault et al (2022)¹³ conducted a cohort analysis using administrative data from Med-Echo databases (a Quebec administrative database) to evaluate the effectiveness and safety of low-dose vs standard-dose XARELTO and apixaban. Hospitalized patients ≥18 years of age subsequently discharged with AF as a primary or secondary diagnosis, (based on ICD-9 or ICD-10 codes) from January 2011 to December 2017, were included. The primary effectiveness outcome was ischemic stroke/SE. The primary safety outcome was MB, including ICH, GI bleeding, and MB from other sites.

• Inclusion criteria consisted of patients who filled a new prescription for XARELTO (low-dose 15 mg or standard-dose 20 mg once daily) or apixaban (low-dose 2.5 mg or standard-dose 5 mg twice daily) within a year of hospital discharge, had no exposure to oral anticoagulants one year prior to the claim index date, and maintained pharmacy coverage for at least 12 months with continuous enrollment in an insurance drug plan for at least 1 year prior to the claim index date.
• Patients were categorized as under treatment (UT) if their prescription was filled within 30 days after the end of the last treatment period. An ITT analysis was also performed. An IPTW approach was utilized to balance the distribution of patient characteristics between groups and create 2 IPTW populations using 1:1 propensity score matching (low-dose and standard-dose XARELTO and low-dose and standard-dose apixaban).
• A total of 16,967 qualifying AF patient claims were identified for XARELTO or apixaban as follows: 1722 for low-dose XARELTO, 4639 for standard-dose XARELTO, 3833 for low-dose apixaban, and 6773 for standard-dose apixaban.
• No significant difference for low-dose vs standard-dose XARELTO was observed for the primary effectiveness outcome of ischemic stroke/SE in the UT or ITT (HR, 1.16; 95% CI, 0.70-1.93 and HR, 1.28; 95% CI, 0.83-1.99). No significant difference was observed for the safety composite in the UT or ITT (HR, 0.98; 95% CI, 0.69-1.41 and HR, 1.01; 95% CI, 0.72-1.41) between low-dose and standard-dose XARELTO. An increased risk of acute MI associated with low-dose XARELTO was observed in the UT analysis (HR, 2.07; 95% CI, 1.21-3.52).
• Low-dose apixaban was associated with an increased risk for ischemic stroke/SE, all-cause mortality, and effectiveness composite events (ischemic stroke/SE, all-cause mortality, and acute MI) compared to standard-dose apixaban (HR, 1.95; 95% CI, 1.38-2.76; HR, 1.99; 95% CI, 1.46-2.70; HR, 1.74; 95% CI, 1.41-2.13, respectively). No significant decrease in the incidence of safety composite events was observed in the UT and ITT analysis (HR, 0.76; 95% CI, 0.56-1.02 and HR, 0.82; 95% CI, 0.62-1.08).
• Key limitations of the study included:
  • The study may not be generalizable because the patient population was mostly representative of older White adults.
  • Due to the observational nature of this study, further evaluation is needed to determine the influence of residual or confounding bias by unadjusted factors.
  • Exact weight and estimated glomerular filtration rate (eGFR) data were not provided to confirm the appropriateness of each prescription.

Milentijevic et al (2021)¹⁰ conducted a retrospective study using the IBM MarketScan Commercial and Medicare databases from 2011 to 2019 to provide further insight in terms of stroke overall and by severity among NVAF patients receiving XARELTO vs warfarin.

• A total of 13,599 XARELTO and 39,861 warfarin-treated patients were included in the study.
• The mean observation period from index date to either stroke, or end of eligibility or end of data was 28 months.
• A total of 272 (2%) patients on XARELTO and 1303 (3.3%) patients on warfarin developed TIA or ischemic or hemorrhagic stroke.
• XARELTO patients had lower risk for stroke overall (HR, 0.82; 95% CI, 0.76-0.88) and for minor (NIHSS score 1 to <5; HR, 0.83; 95% CI, 0.74-0.93), moderate (NIHSS score 5 to <16; HR, 0.88; 95% CI, 0.78-0.99), and severe stroke (NIHSS score 16 to 42; HR, 0.44; 95% CI, 0.22-0.91).
• Key limitations in the study included:
  • The database lacked information on anticoagulant adherence and control, and patients may have received other stroke prevention treatments.
  • Information on bleeding risk was not assessed in the study as it focused specifically on stroke outcomes.

Alberts et al (2020)¹¹ conducted a retrospective cohort study to compare XARELTO with warfarin for stroke and all-cause mortality risk reduction in a real-world setting. Claims from the Optum^® Clinformatics database between July 2011 and December 2017 were used to identify patients who started treatment with XARELTO or warfarin within 30 days following initial diagnosis of NVAF. Prior to NVAF diagnosis, patients had 6 months of continuous health plan enrollment and a CHA₂DS₂-VASc score of ≥2. Stroke severity was determined by the NIHSS, imputed based on machine learning algorithms.

• Stroke and all-cause mortality risks were compared by treatment using Cox proportional hazard regression, with inverse probability of treatment weighting (IPTW) used to balance cohorts for baseline risk factors.
• During a mean follow-up of 27 months, 175 XARELTO-treated patients and 536 warfarin-treated patients developed a stroke.
• XARELTO reduced stroke risk by 19% compared to warfarin (HR, 0.81; 95% CI: 0.73-0.91).
• Analysis by stroke severity showed significant risk reductions with XARELTO for severe stroke (48% reduction; NIHSS score, 16-42; HR, 0.52; 95% CI, 0.33-0.82) and for minor stroke (19% reduction; NIHSS score, 1 to <5; HR, 0.81; 95% CI, 0.68-0.96).
• No significant difference between XARELTO and warfarin was found for moderate stroke (NIHSS score, 5 to <16; HR, 0.93; 95% CI, 0.78-1.1).
• A total of 41 XARELTO-treated patients and 147 warfarin-treated patients died poststroke, of which 12 and 67, respectively, died within 30 days, representing a 24% mortality risk reduction with XARELTO (HR, 0.76; 95% CI, 0.61-0.95) poststroke and 59% reduction (HR, 0.41; 95% CI, 0.28-0.6) within 30 days.
• Key limitations in the study included:
  • The study did not provide a measure of adherence and persistence to the anticoagulant therapy, anticoagulant control, or other cardiovascular event prevent strategies.
  • The study did not assess complications associated with organ-related bleeding as its focus was on stroke outcomes, it did include ICH as a stroke outcome.

Chrischilles et al (2018)¹⁷ used the FDA Sentinel system from Aetna, Humana, Optum, and HealthCore databases from November 1, 2011 through April 30, 2015 to compare the safety of XARELTO initiators vs warfarin initiators on ICH, GI bleeding, and ischemic stroke outcomes.

• A total of 36,173 XARELTO and 79,520 warfarin initiators were variable‐ratio matched within 2 monitoring periods for the GI bleeding outcome, with similar sample sizes in the ICH and ischemic stroke outcomes.
• Patients were ≥21 years of age and had a diagnosis of AF or atrial flutter. Patients were excluded if they received a diagnosis of mitral stenosis, mechanical heart valve, joint replacement, renal dialysis, or a history of renal transplant.
• The follow‐up for each outcome ended at the earliest of any of the following: occurrence of that outcome event, initiation of a different anticoagulant, health plan disenrollment, death, discontinuation of the initiated therapy, or reaching the end of the assessment period.
• XARELTO initiators were approximately 4 years younger and had fewer stroke or bleeding risk factors compared with warfarin initiators. XARELTO and warfarin initiators with a prior ischemic stroke diagnosis were 7.5% and 11.6%, respectively. Prior GI bleeding had occurred in 3.6% and 5.5%, and ICH in 0.6%, and 1.3%, of XARELTO and warfarin initiators, respectively.
• HR for XARELTO vs warfarin initiators for ischemic stroke was 0.61 (0.47-0.79), for GI bleeding was 1.47 (1.29-1.67), and for ICH was 0.71 (0.5-1.01).
• For GI bleeding, HR for XARELTO vs warfarin initiators varied by age: <66 years old: HR, 0.88 (0.6-1.3) (P=0.0002); ≥66 years old: HR, 1.49 (1.3-1.71) (P=0.0002).
• Key limitations in the study included:
  • It is possible some patients could have had an AF diagnosis in previous years and taken warfarin before a long period of non-adherence or non-problematic AF.

Lip et al (2018)¹⁸ conducted a retrospective, observational study of patients initiating treatment with apixaban, dabigatran, XARELTO, or warfarin for NVAF, known as ARISTOPHANES (Anticoagulants for Reduction in Stroke: Observational Pooled Analysis on Health Outcomes and Experience of Patients). The study pooled patient records from 4 US commercial claims databases and Centers for Medicare and Medicaid Services Medicare data to compare stroke/SE and MB among a large NVAF population (n=285,292). Patients were selected for analysis between January 1, 2013, and September 30, 2015.

After propensity-score matching, 6 matched cohorts were created using the patient population. The NOAC vs warfarin group consisted of: apixaban-warfarin (n=57,929), dabigatran-warfarin (n=26,838), and XARELTO-warfarin (n=83,007). The NOAC vs NOAC group consisted of: apixaban-dabigatran (n=27,096), apixaban-XARELTO (n=62,619), dabigatran-XARELTO (n=27,538).

The NOACs, when compared to warfarin, were all associated with lower rates of stroke/SE, apixaban (HR, 0.61; 95% CI, 0.54-0.69), dabigatran (HR, 0.80; 95% CI, 0.68-0.94), and XARELTO (HR, 0.75; 95% CI, 0.69-0.82). In the NOACs compared to warfarin cohort, apixaban and dabigatran cohort were associated with lower MB (HR, 0.58; 95% CI, 0.54-0.62 and HR 0.73; 95% CI, 0.66-0.81 respectively), and XARELTO (HR, 1.07; 95% CI, 1.02-1.13) had a higher rate.

Rates of stroke/SE and MB across the NOACs differed. See Table: Comparison of Stroke/SE and Major Bleeding Between NOACs.

• Key limitations of this study included:
  • NOAC vs NOAC cohorts are for hypothesis generation due to the lack of head-to-head trials.
  • There was no evaluation of the dose reduction criteria for the NOACs, and INR measurements were not available.

Nielsen et al (2017)¹⁶ conducted an observational, cohort study to compare the effectiveness and safety of reduced doses of NOACs (i.e., XARELTO 15 mg once daily, apixaban 2.5 mg twice daily and dabigatran 110 mg twice daily) with warfarin in patients with AF who were naïve to oral anticoagulants. Study data were obtained from 3 nationwide administrative registries in Denmark.

Patients were followed from the onset of treatment until April 30, 2016 for the occurrence of the primary efficacy outcome (i.e. the combined endpoint of ischemic stroke or SE), and safety outcomes were reported as "any bleeding" and included hemorrhagic stroke and MB and GI bleeding.

Among 55,644 patients with AF who met inclusion criteria, the cohort was distributed according to treatment: apixaban n=4400; dabigatran n=8875; XARELTO n=3476; warfarin n=38,893. There were 21,949 patients analyzed according to “indication for dose reduction” (age ≥80 years and/or renal disease).

A total of 1779 patients experienced an ischemic stroke or SE event during the first year of follow-up. During one year of follow-up, apixaban was associated with higher weighted event rate of ischemic stroke/SE (4.8%), while dabigatran, XARELTO, and warfarin had event rates of 3.3%, 3.5%, and 3.7%, respectively.

The weighted event rates for bleeding outcomes were similar for apixaban, XARELTO, and warfarin at 5.1%, 5.6%, and 5.1%, respectively, and lower for dabigatran (4.1%).

For effectiveness and safety outcomes at 1-year follow-up according to the patient’s initiated treatment as compared to warfarin, see Table: Effectiveness and Safety Outcomes at 1-Year Follow-up According to Initiated Treatment in Comparison to Warfarin.

• Key limitations of this study included:
  • Quality of treatment in associations with social status and socioeconomic factors could have contributed to differences in outcomes associated with treatments.

The study did not have access to patient’s eGFR or CrCl in order to identify patients with impaired renal function.

Yao et al (2017)¹⁵ conducted a retrospective analysis from OptumLabs Data of 14,865 subjects with NVAF using administrative claims and laboratory data from OptumLabs Data to investigate patterns of NOAC dosing and the associated risks of stroke and MB in routine clinical practice. The analysis examined the use of a standard dose in patients with a renal indication for dose reduction (potential overdosing), as well as the use and outcomes of a reduced dose in patients with no renal indication for dose reduction (potential underdosing).

Patients who initiated XARELTO, apixaban, or dabigatran between October 1, 2010 and September 30, 2015, and had creatinine test results before treatment initiation were included. Patients with valvular heart disease, eGFR <15 mL/min/1.73 m², or with other indications for NOACs were excluded.

Patients were considered to have a renal indication for dose reduction if they were prescribed dabigatran and had eGFR <30 mL/min/1.73 m², XARELTO and eGFR <50 mL/min/1.73 m², or apixaban and serum creatinine (SCr) level ≥1.5 mg/dL.

The primary effectiveness outcome was ischemic stroke or SE, and the primary safety outcome was MB.

• Among patients who had a renal indication for dose reduction (n=1473):
  • The median age was 79 years, the median GFR was 39 mL/min/1.73 m², the median CHA₂DS₂-VASc score was 5, and the median HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile INR, elderly, drugs/alcohol concomitantly) score was 4. Patients were followed up for a median of 3.6 months.
• Among patients without a renal indication for dose reduction (n=13,392):
  • The median age was 70 years, the median eGFR was 73 mL/min/1.73 m², the median CHA₂DS₂- VASc score was 4, and the median HAS-BLED score was 2. Patients were followed up for a median of 4 months.
• Among patients with a renal indication for dose reduction, 43% received standard doses. Among patients with no renal indication for dose reduction, 13.3% received reduced doses.
• There was no statistically significant difference in the risk of stroke/SE between standard-dose vs reduced-dose NOACs (2.32 vs 1.85 per 100 PY; HR, 1.66; 95% CI, 0.4-6.88). Bleeding risk was significantly higher in patients prescribed standard dose NOACs vs reduced dose (11.29 and 5.06 per 100 PY; HR, 2.19; 95% CI, 1.07-4.46)
• Among patients with no renal indication for dose reduction, use of reduced-dose vs standard-dose apixaban was associated with a nearly 5-fold higher risk of stroke/SE (2.57 and 0.54 per 100 PY; HR, 4.87; 95% CI, 1.3-18.26) but a similar risk of MB (6.01 and 4.64 per 100 PY; HR, 1.29; 95% CI, 0.48-3.42).
• In the dabigatran group, there was no statistically significant relationship between dose reduction and risk of stroke/SE (reduced-dose: 1.64 vs standard-dose: 1.75 per 100 PY; HR, 0.92; 95% CI, 0.3-2.87) or MB (reduced dose: 4.99 vs standard-dose: 5.54 per 100 PY; HR, 0.91; 95% CI, 0.45-1.85).
• In the XARELTO-treated patients’ group, there was no statistically significant relationship between dose reduction and risk of stroke/SE (reduced-dose: 1.23 vs standard-dose: 1.65 per 100 PY; HR, 0.71; 95% CI, 0.24-2.09) or MB (reduced-dose: 5.42 vs standard-dose: 4.90 per 100 PY; HR, 1.09; 95% CI, 0.63-1.87).
• Key limitations of this study included:
  • The average follow-up time was limited to the first six months. Therefore, the findings do not provide physicians with long term risk assessment.
  • Only the most recent SCr result before treatment initiation was abstracted, which may not necessarily reflect patient kidney function during follow-up.
  • Because weight was not available in their database, the main analyses relied on SCr level as the apixaban dose indication, differing from the approved product label.
  • The number of events and event rates were low; therefore, the findings should be viewed as hypothesis-generating and need to be confirmed by future studies.

REVISIT-US (Real-world EVIdence on Stroke prevention In patients with aTrial Fibrillation in the United States)¹⁹ was a retrospective analysis of US Truven MarketScan claims data evaluating the safety and effectiveness of newly initiated XARELTO or apixaban vs warfarin. The primary objective was to assess the real-world effectiveness and safety of these agents individually compared to warfarin, and not to directly compare XARELTO to apixaban. Patients were oral anticoagulant naïve adults with NVAF, a baseline CHA_sDS₂-VASc score ≥2, and ≥180 days of constant prescription and medical coverage. Patients with a history of prior stroke, SE, or ICH were excluded.

The primary study endpoint was the combination of ischemic stroke or ICH, identified by the presence of an International Classification of Diseases 9 Clinical Modification (ICD-9-CM) primary diagnosis code as recommended by US FDA "Mini-Sentinel" postmarketing surveillance system coding schemas.

• A total of 11,411 patients prescribed XARELTO between January 2012 and October 2014 were propensity-score matched to an equal number of warfarin patients to minimize the differences in baseline characteristics between the 2 groups.
• The demographic characteristics between the XARELTO group and the warfarin matched cohort were similar.
• XARELTO was associated with a significant 39% reduction of the combined endpoint of ischemic stroke or ICH vs warfarin. Detailed results are presented in Table: Annual Incidence of Ischemic and Hemorrhagic Stroke.

• After matching 4083 apixaban and 4083 warfarin users, apixaban was found to have similar risk of the combined endpoint vs warfarin (HR, 0.63, 95% CI, 0.35-1.12). A significant 62% reduction in ICH (HR, 0.38, 95% CI, 0.17-0.88) and a similar ischemic stroke risk (HR, 1.13, 95% CI, 0.49-2.63) were found in patients receiving apixaban.
• Key limitations of the study included:
  • Since this is a retrospective database analysis, there is no reporting of laboratory SCr and clinical data which are required to determine if prescribing was consistent with the labeling. Additionally, it was not possible to determine the duration of time warfarin users spent in the therapeutic INR range of 2 to 3.

Graham et al (2016)²⁰ conducted a retrospective study of 118,891 NVAF patients that compared the risks of thromboembolic stroke, mortality, ICH, and major extracranial bleeding in patients initiating treatment with standard doses of dabigatran or XARELTO treatment for stroke prevention. Patients were 65 years or older, enrolled in fee-for-service Medicare, and initiated treatment with dabigatran (150 mg twice daily) or XARELTO (20 mg once daily) from November 4, 2011, through June 10, 2014.

Primary outcomes were thromboembolic stroke, ICH, major extracranial bleeding, including GI bleeding, and mortality. Major extracranial bleeding was defined as a fatal bleeding event, a hospitalized bleeding event requiring transfusion, or hospitalization with hemorrhage into an extracranial critical site.

• A total of 52,240 dabigatran and 66,651 XARELTO patients contributed to 15,524 and 20,199 PY of on-treatment follow-up for a mean duration of 108 and 111 days, respectively.
• There were 306 thromboembolic strokes, 176 ICH events, 1209 major extracranial bleeding events of which 1018 (84.2%) were GI, and 846 deaths.
• Primary outcome results comparing XARELTO to dabigatran are presented in the table: Primary Outcome Results for XARELTO vs Dabigatran.

• For the mortality endpoint, XARELTO risk was significantly increased in patients >75 years and in those with a CHADS₂ score >2.
• The secondary analysis found no differences over time for any of the study outcomes except thromboembolic stroke, for which XARELTO risk was decreased during the first 90 days of use, but not thereafter (HR, 0.71; 95% CI, 0.55-0.93 vs HR, 1.14; 95% CI, 0.74-1.75; P for interaction=0.07).
• Key limitations of the study included:
  • In this study, patients taking a lower dose of the study drug were excluded. It is possible that a higher proportion of XARELTO-treated patients with renal impairment were treated off-label with the standard dose because of the broader CrCl range guiding XARELTO dosing. If this occurred, a greater anticoagulant effect with XARELTO might be observed.

Larsen et al (2016)¹² conducted an observational cohort study to compare the effectiveness and safety of XARELTO, dabigatran, and apixaban compared with warfarin using a nationwide Danish cohort of patients with AF who were naïve to oral anticoagulants. The study analyzed data from three Danish nationwide databases from August 2011-October 2015. Effectiveness outcomes were ischemic stroke; a composite of ischemic stroke or SE; death; and a composite of ischemic stroke, SE, or death (identified via ICD-10 diagnostic codes). Safety outcomes were any bleeding (intracranial, major, GI, traumatic intracranial), intracranial bleeding, and MB.

The study population (n=61,678) was distributed according to treatment type: warfarin (n=35,436, 57%), dabigatran (n=12,701, 21%), XARELTO (n=7192, 12%), and apixaban (n=6349, 10%). During the first year of follow-up, 1702 ischemic stroke or SE events were observed.

• Effectiveness results for these NOACs compared to warfarin are presented in the table: HRs for 1-Year Follow-up (Intention-to-Treat) for DOACs Compared With Warfarin for Stroke and Death Endpoints.

• Safety results for the 3 DOACs compared to warfarin are presented in the table: HRs for 1-Year Follow-up (Intention-to-Treat) for DOACs Compared With Warfarin for Bleeding Endpoints.

• Key limitations of this study included:
  • This study could not capture lab parameters or time in therapeutic range (TTR) for warfarin users.
  • This analysis did not focus on direct comparisons of one DOAC agent against another.

Yao et al (2016)¹⁴ conducted a retrospective analysis using administrative claims data from OptumLabs Data Warehouse evaluating stroke and bleeding outcomes associated with XARELTO, apixaban and dabigatran compared to warfarin. The primary effectiveness outcome was stroke or SE, including ischemic stroke, hemorrhagic stroke, and SE. The primary safety outcome was MB, including GI bleeding, intracranial bleeding, and bleeding from other sites. The outcomes were identified using ICD-9 codes in the primary or secondary diagnosis positions of inpatient claims.

A total of 3 matched cohorts were created using 1:1 propensity score matching: XARELTO vs warfarin (n=32,350), apixaban vs warfarin (n=15,390), and dabigatran vs warfarin (n=28,614).

HRs for DOACs compared with warfarin for effectiveness endpoints are listed in Table: HRs for DOACs Compared With WRF for Effectiveness Endpoints.

HRs for DOACs compared with warfarin for safety endpoints are listed in the table: HRs for DOACs Compared With WRF for Safety Endpoints.

• Key limitations of this study included:
  • Claims databases are unable to capture certain clinical and health behavior parameters such as type of AF or left ventricular ejection fraction.
  • This study was not able to accurately assess mortality.
  • This study was unable to obtain INR values for every patient treated with warfarin, since only ~40 patients had lab values.

LITERATURE SEARCH

A literature search of MEDLINE^®, EMBASE^®, BIOSIS Previews^®, DERWENT^® (and/or other resources, including internal/external databases) was conducted on 21 August 2026.