Abstract
Background
Atrial fibrillation (AF) increases risk of stroke 5‐fold. Carotid artery disease (CD) also augments the risk of stroke, yet there are limited data about the interplay of these 2 diseases and clinical outcomes in patients with comorbid AF and CD.
Hypothesis
Among patients with both AF and CD, use of rivaroxaban when compared with warfarin is associated with a lower risk of stroke.
Methods
This post hoc analysis from ROCKET AF aimed to determine absolute rates of stroke/systemic embolism (SE) and bleeding, and the efficacy and safety of rivaroxaban compared with warfarin in patients with AF and CD (defined as history of carotid occlusive disease or carotid revascularization [endarterectomy and/or stenting]).
Results
A total of 593 (4.2%) patients had CD at enrollment. Patients with and without CD had similar rates of stroke or SE (adjusted hazard ratio [HR]: 0.99, 95% confidence interval [CI]: 0.66‐1.48, P = 0.96), and there was no difference in major or nonmajor clinically relevant bleeding (adjusted HR: 1.04, 95% CI: 0.88‐1.24, P = 0.62). The efficacy of rivaroxaban compared with warfarin for the prevention of stroke/SE was not statistically significant in patients with vs those without CD (interaction P = 0.25). The safety of rivaroxaban vs warfarin for major or nonmajor clinically relevant bleeding was similar in patients with and without CD (interaction P = 0.64).
Conclusions
Patients with CD in ROCKET AF had similar risk of stroke/SE compared with patients without CD. Additionally, there was no interaction between CD and the treatment effect of rivaroxaban or warfarin for stroke prevention or safety endpoints.
Keywords: Atrial Fibrillation, Carotid Artery Disease, Rivaroxaban, Warfarin
1. INTRODUCTION
Patients with atrial fibrillation (AF) have a 4‐fold to 5‐fold higher risk of ischemic stroke, and AF‐related strokes are associated with higher morbidity and mortality.1, 2 The mechanism of AF‐related stroke is most commonly cardioembolic. However, an analysis of patients with AF and ischemic stroke from the Stroke Prevention in Atrial Fibrillation (SPAF) I–III trials demonstrated that 24% of strokes were noncardioembolic and 24% were of uncertain cause. Carotid artery disease (CD) is independently associated with an increased risk of stroke and may contribute to the risk of stroke among patients with AF.3 The 2011 guidelines on extracranial and vertebral artery disease provide a class IIa recommendation to administer a vitamin K antagonist for patients with extracranial cerebrovascular atherosclerosis and AF.4 Yet there are limited data describing the frequency and prognostic significance of comorbid CD and AF. Furthermore, the role of novel oral anticoagulants, such as rivaroxaban, in reducing the risk of stroke is presently unclear in patients with AF and CD.
The Rivaroxaban Once Daily, Oral, Direct Factor Xa Inhibition Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET AF) trial demonstrated that rivaroxaban was noninferior to warfarin for the prevention of stroke and systemic embolism (SE) in patients with AF.5 The goal of this subgroup analysis is to evaluate the frequency of comorbid AF and CD, determine the rates of stroke in patients with AF and CD, and assess the safety and efficacy of rivaroxaban vs warfarin in patients with AF with and without CD.
2. METHODS
The ROCKET AF trial design was previously described.6 ROCKET AF was a multicenter, international, double‐blind, event‐driven trial that randomized 14 264 patients with nonvalvular AF at increased risk of stroke (CHADS2 score ≥ 2) to receive rivaroxaban or warfarin.5 Patients in the rivaroxaban group received a fixed dose of 20 mg daily (or 15 mg daily in those with a calculated creatinine clearance (CrCl) of 30–49 mL/min). Warfarin was dose‐adjusted to achieve a target international normalized ratio of 2.0 to 3.0. Exclusion criteria included an indication for anticoagulation for a condition other than AF, CrCl <30 mL/min, prosthetic heart valve, mitral stenosis, or an increased risk of bleeding. The ROCKET AF protocol encouraged investigators to manage patients according to local standards of care. Therefore, the concomitant use of aspirin up to, but not exceeding, 100 mg per day was permitted. Furthermore, thienopyridines were not allowed for 5 days prior to randomization or during the study. However, patients who underwent vascular interventions could receive dual antiplatelet therapy with aspirin and a thienopyridine at the provider's discretion.6 The study protocol was reviewed and approved by the institutional review board or ethics committee at each participating site and by the coordinating center's institutional review board. All patients provided written informed consent prior to randomization.
The primary efficacy outcome of ROCKET AF was stroke or SE. Secondary efficacy outcomes included all‐cause death, cardiovascular death, myocardial infarction (MI), and the individual components of the composite endpoint of stroke or SE. Efficacy endpoints were measured until the time of site notification of study termination. The primary safety endpoint was a combination of major or nonmajor clinically relevant (NMCR) bleeding. Bleeding events involving the central nervous system that met the definition of stroke were designated as hemorrhagic strokes and were included in both the primary efficacy and safety endpoints. Safety endpoints were measured until 2 days after the last study drug dose. All events were adjudicated using predefined endpoint definitions by an independent clinical events committee whose members were unaware of treatment assignment. Stroke was defined as a sudden, focal neurological deficit of presumed cerebrovascular cause that was neither reversible within 24 hours nor due to another readily identified cause. Brain imaging was encouraged but not required to distinguish between hemorrhagic and ischemic stroke.
CD was defined by a history of carotid occlusive disease or a history of carotid revascularization (carotid stenting or carotid endarterectomy) documented in the trial case‐report form. This analysis of patients with CD in ROCKET AF was not prespecified. There were 14 264 patients randomized in ROCKET AF; 1 patient had missing data from medical history and was excluded from the analysis.
2.1. Statistical analysis
The intention‐to‐treat study population was used for all efficacy outcome analyses. Cox proportional hazards models were employed to evaluate outcomes based on patients with and without CD and patients randomized to warfarin or rivaroxaban. The statistical models used included a term for the interaction between randomized treatment and CD, as well as covariates identified as predictive of outcomes by modeling in the full ROCKET AF cohort. The Cox models for efficacy outcomes contained the following covariates: age, sex, body mass index, geographic region, paroxysmal AF, diabetes, prior stroke/transient ischemic attack (TIA), prior MI, peripheral artery disease, congestive heart failure, hypertension, chronic obstructive pulmonary disease, diastolic blood pressure, CrCl (Cockcroft‐Gault), heart rate, and abstinence from alcohol.
Safety endpoints were analyzed using the as‐treated population (randomized patients who received ≥1 dose of study drug). The covariates for the safety endpoints were age, sex, geographic region, prior stroke/TIA, anemia, prior gastrointestinal bleed, chronic obstructive pulmonary disease, diastolic blood pressure, CrCl, platelet count, albumin, and prior aspirin, vitamin K antagonist, or thienopyridine use.
Continuous variables are shown as medians and interquartile ranges, except where noted; categorical variables are shown as counts and percentages. P values are from Wilcoxon rank‐sum tests for continuous variables and Pearson χ2 tests for categorical variables. Outcomes are presented as events per 100 patient‐years. Risk relationships are presented as adjusted hazard ratios (HR) with 95% confidence intervals (CI) derived from the adjusted Cox models. The time to event for each group was assessed using the Kaplan–Meier method. All analyses were performed with SAS version 9.2 (SAS Institute, Inc., Cary, NC).
3. RESULTS
Of the 14 264 patients randomized in ROCKET AF, 593 (4.2%) had CD. Of those with CD, 158 (27%) had prior carotid endarterectomy, 49 (8%) had previous carotid artery stenting, and 9 (2%) had both prior carotid endarterectomy and angioplasty/stent. The baseline demographics of patients with and without CD are presented in Table 1. Patients with CD were slightly older (74 vs 73 years; P < 0.0001), had a higher average CHADS2 score (3.9 vs 3.5; P < 0.0001), and had higher rates of several comorbidities. Specifically, the CD cohort had a higher frequency of prior stroke, TIA, or non–central nervous system embolism (72% vs 54%; P < 0.0001) and a higher proportion of prior MI (32% vs 17%, P < 0.0001; Table 1).
Table 1.
Baseline demographics and clinical features of patients with and without CD
| Variable | All Randomized Patients, N = 14 263a | Patients With CD, n = 593 | Patients Without CD, n = 13 670 | P Valueb |
|---|---|---|---|---|
| Randomized to rivaroxaban | 7131 (50) | 293 (49) | 6838 (50) | 0.77 |
| Age, y | 73 (65–78) | 74 (68–79) | 73 (65–78) | <0.0001 |
| Female sex | 5659 (40) | 165 (28) | 5494 (40) | <0.0001 |
| Type of AF | 0.078 | |||
| Persistent | 11 548 (81) | 459 (77) | 11 089 (81) | |
| Paroxysmal | 2513 (18) | 124 (21) | 2389 (17) | |
| New onset/newly diagnosed | 202 (1.0) | 10 (2.0) | 192 (1.0) | |
| CHADS2 score, mean (SD) | 3.5 (0.9) | 3.9 (1.0) | 3.5 (0.9) | <0.0001 |
| CHADS2 score | <0.0001 | |||
| 1 | 3 (<1.0) | — | 3 (<1.0) | |
| 2 | 1859 (13) | 48 (8.0) | 1811 (13) | |
| 3 | 6216 (44) | 185 (31) | 6031 (44) | |
| 4 | 4091 (29) | 191 (32) | 3900 (29) | |
| 5 | 1812 (13) | 137 (23) | 1675 (12) | |
| 6 | 282 (2.0) | 32 (5.0) | 250 (2.0) | |
| Presenting characteristics | ||||
| BMI, kg/m2 | 28.2 (25.1–32.0) | 27.8 (25.1–30.8) | 28.2 (25.1–32.0) | 0.031 |
| SBP, mm Hg | 130 (120–140) | 130 (120–140) | 130 (120–140) | 0.10 |
| DBP, mm Hg | 80 (70–85) | 80 (70–82) | 80 (70–85) | <0.0001 |
| Heart rate, bpm | 76 (67–86) | 73 (65–84) | 76 (68–86) | 0.0002 |
| CrCl, mL/minc | 67 (52–87) | 62 (48–82) | 68 (52–87) | <0.0001 |
| Baseline comorbidities | ||||
| Prior stroke, TIA, or non‐CNS embolism | 7810 (55) | 426 (72) | 7384 (54) | <0.0001 |
| PAD | 839 (6.0) | 155 (26) | 684 (5.0) | <0.0001 |
| HTN | 12 909 (91) | 554 (93) | 12 355 (90) | 0.013 |
| DM | 5695 (40) | 239 (40) | 5456 (40) | 0.85 |
| Prior MI | 2467 (17) | 188 (32) | 2279 (17) | <0.0001 |
| CHF | 8907 (62) | 349 (59) | 8558 (63) | 0.064 |
| COPD | 1497 (11) | 92 (16) | 1405 (10) | <0.0001 |
| Medications | ||||
| Prior VKA use | 8903 (62) | 453 (76) | 8450 (62) | <0.0001 |
| Prior chronic ASA use | 5205 (36) | 245 (41) | 4960 (36) | 0.013 |
| ACEI/ARB at baseline | 10 583 (74) | 453 (76) | 10 130 (74) | 0.21 |
| β‐Blocker at baseline | 9249 (65) | 426 (72) | 8823 (65) | 0.0003 |
| Digitalis at baseline | 5468 (38) | 190 (32) | 5278 (39) | 0.0013 |
| Diuretic at baseline | 8490 (60) | 398 (67) | 8092 (59) | 0.0001 |
| Prior treatment of CD | 198 (33) | |||
| Carotid endarterectomy | 158 (27) | |||
| Time since last procedure, y | 5.0 (2.0–8.8) | |||
| Carotid angioplasty/stent | 49 (8) | |||
| Time since last procedure, y | 1.9 (1.2–3.7) |
Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin II receptor blocker; ASA, acetylsalicylic acid (aspirin); BMI, body mass index; CD, carotid artery disease; CHADS2, CHF, HTN, age > 75 y, DM, prior stroke/TIA/TE; CHF, congestive heart failure; CNS, central nervous system; DBP, diastolic blood pressure; DM, diabetes mellitus; HTN, hypertension; IQR, interquartile range; MI, myocardial infarction; PAD, peripheral arterial disease; ROCKET AF, Rivaroxaban Once Daily, Oral, Direct Factor Xa Inhibition Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation; SBP, systolic blood pressure; SD, standard deviation; TE, thromboembolism; TIA, transient ischemic attack; VKA, vitamin K antagonist.
Data are presented as n (%) or median (IQR), unless otherwise noted.
There were 14 264 patients randomized in ROCKET AF; 1 patient had missing data from medical history and was excluded from the analysis.
P values are from Wilcoxon rank‐sum tests for continuous variables and from Pearson χ2 tests for categorical variables.
Calculated using the Cockcroft‐Gault equation.
The rate of the primary efficacy endpoint of stroke or SE was higher in patients with CD (2.95 events/100 patient‐years) compared with those without CD (2.24 events/100 patient‐years), but the difference attenuated after adjustment (adjusted HR: 0.99, 95% CI: 0.66‐1.48, P = 0.96; Figure 1, Table 2). Similarly, there were no differences regarding the secondary endpoints, including all‐cause death, vascular death, or MI (Table 2). After adjustment for covariates, there were no statistically significant differences between patients with or without CD for the primary safety endpoint of major/NMCR bleeding (Figure 2). Patients with CD had 18.02 bleeding events/100 patient‐years vs 14.58 events/100 patient‐years for those without CD (adjusted HR: 1.04, 95% CI: 0.88‐1.24, P = 0.62; Table 2). Patients with CD had higher unadjusted rates of both efficacy and safety endpoints, but the neutral findings derived from adjusted models suggest that these differences were driven by a higher comorbidity burden among patients with CD.
Figure 1.
Primary efficacy endpoint (stroke or SE) in patients with and without carotid artery disease. Abbreviations: CI, confidence interval; SE, systemic embolism
Table 2.
Efficacy and safety endpoints in patients with and without CD
| Outcomes | CD, Events/100 PY (Total Events) | No CD, Events/100 PY (Total Events) | CD vs No CD, Adjusted HR (95% CI) | P Value |
|---|---|---|---|---|
| Efficacy outcomesa | ||||
| N | 589 | 13 581 | ||
| Stroke or SE | 2.95 (31) | 2.24 (544) | 0.99 (0.66‐1.48) | 0.96 |
| Ischemic stroke | 2.08 (22) | 1.61 (392) | 0.93 (0.57‐1.49) | 0.76 |
| Hemorrhagic stroke | 0.43 (4) | 0.35 (75) | — | — |
| All‐cause death | 6.89 (74) | 4.62 (1140) | 0.95 (0.73‐1.22) | 0.66 |
| CV death | 4.93 (53) | 2.93 (723) | 0.97 (0.72‐1.31) | 0.85 |
| MI | 2.27 (24) | 1.01 (248) | 0.96 (0.61‐1.50) | 0.86 |
| Safety outcomesb | ||||
| N | 591 | 13 644 | ||
| Major or NMCR bleeding | 18.02 (145) | 14.58 (2779) | 1.04 (0.88‐1.24) | 0.62 |
| Major bleeding | 5.60 (51) | 3.44 (730) | 1.26 (0.95‐1.69) | 0.11 |
| ICH | 0.54 (5) | 0.62 (134) | — | — |
Abbreviations: CD, carotid artery disease; CI, confidence interval; CV, cardiovascular; HR, hazard ratio; ICH, intracranial hemorrhage; MI, myocardial infarction; NMCR, nonmajor clinically relevant; PY, patient‐years; SE, systemic embolism.
Analysis performed in intention‐to‐treat population.
Analysis performed in as‐treated population.
Figure 2.
Primary safety endpoint (major or NMCR bleeding) in patients with and without carotid artery disease. Abbreviations: CI, confidence interval; NMCR, nonmajor clinically relevant
Among patients with CD, the primary efficacy endpoint rate was higher in the rivaroxaban group as compared with the warfarin group (adjusted HR: 1.32, 95% CI: 0.65‐2.69), but the rate was lower in the rivaroxaban vs the warfarin group among those with no CD (adjusted HR: 0.86, 95% CI: 0.73‐1.02; Figure 3). However, the CIs for these estimates were overlapping and the test for interaction was not significant (P = 0.25). Furthermore, there were no statistically significant interactions between CD and rivaroxaban or warfarin with respect to the secondary endpoints of all‐cause death, vascular death, or MI (Figure 3). There was no statistically significant difference (interaction P = 0.64) in the risk relationship between rivaroxaban and warfarin for the primary safety endpoint of major/NMCR bleeding for patients with CD (HR: 1.13, 95% CI: 0.81‐1.56) vs those without CD (HR: 1.04, 95% CI: 0.97‐1.12; Figure 3). The lack of significant interactions for any endpoint suggests that treatment risk relationships are consistent across subgroups stratified by presence or absence of CD.
Figure 3.
Outcomes by presence or absence of carotid artery disease and treatment with rivaroxaban vs warfarin. Abbreviations: CD, carotid artery disease; CI, confidence interval; HR, hazard ratio; MI, myocardial infarction; NMCR, nonmajor clinically relevant; PY, patient‐years; SE, systemic embolism
4. DISCUSSION
This subgroup analysis examining patients with CD and AF in the ROCKET AF trial has 3 major findings. First, a history of CD was present in 4.2% of patients with AF. Second, unadjusted event rates of stroke or SE, all‐cause death, and MI were higher for patients with CD; however, these differences did not remain statistically significant after adjustment for baseline factors. Third, there was no treatment interaction with rivaroxaban vs warfarin for stroke or SE in patients with AF and without CD.
The SPAF II trial randomized patients with AF to receive warfarin or aspirin to assess stroke reduction; a cohort analysis from SPAF II examined patients with AF and CD.7 Among this cohort of 676 patients with AF who were age > 70 years, the frequency of a carotid stenosis ≥50% systematically measured by carotid ultrasound was 12%, compared with 4.2% frequency of CD in ROCKET AF.8 The difference in CD frequency between these 2 trials possibly reflects the absence of dedicated carotid imaging in ROCKET AF. The SPAF II cohort did not demonstrate a difference in the rate of stroke among patients with AF and CD, a result mirrored by our findings in ROCKET AF. The combination of these 2 studies suggests that a diagnosis of CD in patients with AF does not increase the adjusted risk of stroke or SE while on oral anticoagulation. Alternatively, with incomplete data on severity and morphology of carotid lesions, it is possible that we do not have enough statistical power to detect a difference in outcomes among patients with CD with high‐risk lesions.
The results from this subgroup analysis are consistent with findings from the overall trial population demonstrating that rivaroxaban is noninferior to warfarin for the primary efficacy endpoint of stroke and SE and the principal safety endpoint of major and NMCR bleeding events. These results may empower clinicians caring for medically complex patients with AF with comorbid CD to tailor therapies based on patient preferences, such as cost and convenience, rather than a biological difference in stroke or bleeding outcomes. Furthermore, clinicians considering anticoagulation for patients with AF are unlikely to refine patients' risk profiles by obtaining carotid artery imaging.
The 2014 American Heart Association/American College of Cardiology/Heart Rhythm Society and 2016 European Society of Cardiology AF guidelines recommend the CHA2DS2‐VASc score for stroke risk assessment.9, 10 Vascular disease is defined as prior MI, peripheral artery disease, or aortic plaque, and excludes CD.11 The results of our subgroup analysis suggest that there is not a compelling reason to include CD in this risk model when patients are on appropriate oral anticoagulation.
4.1. Study limitations
There are several limitations to our study. First, because we did not employ imaging data to establish the rate of CD, we are likely underreporting the rate of asymptomatic CD. Moreover, without imaging data we cannot evaluate the percent stenosis and morphology of the carotid plaque, both factors that affect the risk of stroke.4 Second, the determination of CD in the ROCKET AF trial was based on patient report in the case‐report form. Therefore, the frequency of 4.2% may be underestimated due to underreporting from patients with recall bias. Third, the total number of patients with CD (n = 593) likely does not provide enough power to detect nuanced differences. Fourth, CD was not stratified; hence, the proportion of patients with CD in each treatment group was not balanced. Finally, this was a post hoc subgroup analysis that was not prespecified, thus requiring us to interpret these results with caution.12
5. CONCLUSION
This subgroup analysis of the ROCKET AF trial examined the clinical implications of CD in an AF population. The frequency of CD in this population was 4.2%. There were no differences after adjustment in the rates of stroke/SE or bleeding in patients with AF with or without CD. Rivaroxaban was noninferior to warfarin for the prevention of stroke and SE in patients with or without CD.
Conflicts of interest
W.S. Jones reports research grants from the American Heart Association, AstraZeneca, Boston Scientific, Bristol‐Myers Squibb, and Patient‐Centered Outcomes Research Institute; has received consulting fees/honoraria from Mondopoint; and reports other fees from Daiichi Sankyo. R.C. Becker reports research grants from AstraZeneca and consulting fees/honoraria from Boehringer Ingelheim, Daiichi‐Sankyo, and Portola. S.D. Berkowitz is an employee of Bayer HealthCare Pharmaceuticals. G. Breithardt reports consulting fees/honoraria from Bayer HealthCare, Johnson & Johnson, and Sanofi‐Aventis. K. Fox has received research grants from Eli Lilly and GlaxoSmithKline and consulting fees/honoraria from AstraZeneca, Bayer, Janssen, Merck & Co., and Sanofi‐Aventis/Regeneron. J.L. Halperin reports consulting fees/honoraria from Bayer HealthCare Pharmaceuticals, Boehringer Ingelheim, Medtronic, Ortho‐McNeil‐Janssen Pharmaceuticals, and Pfizer. G.J. Hankey has received speaker's bureau fees from Bayer HealthCare Pharmaceuticals. K.W. Mahaffey has received research grants from Amgen, Boehringer Ingelheim, Daiichi‐Sankyo, Johnson & Johnson, Medtronic, St. Jude Medical, and Tenax Therapeutics; reports consulting fees/honoraria from Eli Lilly, the American College of Cardiology, AstraZeneca, BAROnova, Bayer, Bio2 Medical, Boehringer Ingelheim, Bristol‐Myers Squibb, Cubist, Elsevier (American Heart Journal), Epson, Forest, GlaxoSmithKline, Johnson & Johnson, Medtronic, Merck, Mt. Sinai, MyoKardia, Omthera, Portola, Purdue Pharma, Spring Publishing, The Medicines Company, Vindico, and WebMD; reports ownership interest/partnership/principal involvement in BioPrint Fitness; and lists other interests at http://med.stanford.edu/profiles/kenneth-mahaffey. C.C. Nessel is an employee of Janssen Research and Development. D.E. Singer has received research grants from Boehringer Ingelheim, Bristol‐Myers Squibb, Johnson & Johnson, and Medtronic, and reports consulting fees/honoraria from Boehringer Ingelheim, Bristol‐Myers Squibb, CVS Health, Johnson & Johnson, and Medtronic. J.P. Piccini reports research grants from ARCA Biopharma, Boston Scientific, Gilead Sciences, Johnson & Johnson, ResMed, and St. Jude Medical; has received consulting fees/honoraria from Johnson & Johnson, Medtronic, and Spectranetics; and reports other support from GlaxoSmithKline. M.R. Patel has received research grants from the Agency for Healthcare Research and Quality, AstraZeneca, Janssen Pharmaceutical Research & Development, Johnson & Johnson, Maquet, the National Heart, Lung, and Blood Institute, and the Patient‐Centered Outcomes Research Institute; and reports consulting fees/honoraria from Bayer HealthCare Pharmaceuticals, Genzyme, Medscape (theheart.org), and Merck. The authors declare no other potential conflicts of interest.
Kochar A, Hellkamp AS, Lokhnygina Y, et al. Efficacy and safety of rivaroxaban compared with warfarin in patients with carotid artery disease and nonvalvular atrial fibrillation: Insights from the ROCKET AF trial. Clin Cardiol. 2018;41:39–45. 10.1002/clc.22846
Funding information ROCKET AF was funded by Johnson & Johnson Pharmaceutical Research & Development and Bayer HealthCare Pharmaceuticals. The sponsor had no role in the decision to submit the article for publication.
REFERENCES
- 1. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham study. Stroke. 1991;22:983–988. [DOI] [PubMed] [Google Scholar]
- 2. Miller PS, Andersson FL, Kalra L. Are cost benefits of anticoagulation for stroke prevention in atrial fibrillation underestimated? Stroke. 2005;36:360–366. [DOI] [PubMed] [Google Scholar]
- 3. Grau AJ, Weimar C, Buggle F, et al. Risk factors, outcome, and treatment in subtypes of ischemic stroke: the German Stroke Data Bank. Stroke. 2001;32:2559–2566. [DOI] [PubMed] [Google Scholar]
- 4. Brott TG, Halperin JL, Abbara S, et al. 2011. ASA/ACCF/AHA/AANN/AANS/ACR/ASNR/CNS/SAIP/SCAI/SIR/SNIS/SVM/SVS guideline on the management of patients with extracranial carotid and vertebral artery disease: Executive summary [published correction appears in Circulation. 2011;124:e145]. Circulation. 2011;124:489–532.21282505 [Google Scholar]
- 5. Patel MR, Mahaffey KW, Garg J, et al; ROCKET AF Investigators . Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883–891. [DOI] [PubMed] [Google Scholar]
- 6. ROCKET AF Study Investigators . Rivaroxaban Once Daily, Oral, Direct Factor Xa Inhibition Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation: rationale and design of the ROCKET AF study. Am Heart J. 2010;159:340–347. [DOI] [PubMed] [Google Scholar]
- 7. Warfarin versus aspirin for prevention of thromboembolism in atrial fibrillation : Stroke Prevention in Atrial Fibrillation II study. Lancet. 1994;343:687–691. [PubMed] [Google Scholar]
- 8. Kanter MC, Tegeler CH, Pearce LA, et al. Carotid stenosis in patients with atrial fibrillation: prevalence, risk factors, and relationship to stroke in the Stroke Prevention in Atrial Fibrillation study. Arch Intern Med. 1994;154:1372–1377. [PubMed] [Google Scholar]
- 9. January CT, Wann LS, Alpert JS, et al. 2014. AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society [published correction appears in J Am Coll Cardiol. 2014;64:2305–2307]. J Am Coll Cardiol. 2014;64:e1–e76. [DOI] [PubMed] [Google Scholar]
- 10. Kirchhof P, Benussi S, Kotecha D, et al. 2016. ESC guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J. 2016;37:2893–2962. [DOI] [PubMed] [Google Scholar]
- 11. Lip GY, Nieuwlaat R, Pisters R, et al. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor–based approach: the Euro Heart Survey on atrial fibrillation. Chest. 2010;137:263–272. [DOI] [PubMed] [Google Scholar]
- 12. Pocock SJ, Lubsen J. More on subgroup analyses in clinical trials. N Engl J Med. 2008;358:2076–2077. [DOI] [PubMed] [Google Scholar]