Dabigatran and Warfarin in Vitamin K Antagonist Naive and Experienced Cohorts With Atrial Fibrillation

Transcription

1 Dabigatran and Warfarin in Vitamin K Antagonist Naive and Experienced Cohorts With Atrial Fibrillation Michael D. Ezekowitz, MBChB, DPhil, FRCP; Lars Wallentin, MD, PhD; Stuart J. Connolly, MD; Amit Parekh, MD; Michael R. Chernick, PhD; Janice Pogue, MSc; Timothy H. Aikens, BA; Sean Yang, MSc; Paul A. Reilly, PhD; Gregory Y.H. Lip, MD; Salim Yusuf, FRCPC, DPhil; and the RE-LY Steering Committee and Investigators Background The comparison of anticoagulants dabigatran and warfarin might be most equitable in vitamin K antagonist (VKA) naive patients. Methods and Results Warfarin and 2 doses of dabigatran 110 mg BID (D110) and 150 mg BID (D150) were compared in a balanced population of VKA-naive ( 62 days of lifetime VKA exposure, with 33% never prescribed a VKA) and VKA-experienced patients with atrial fibrillation (n ). For VKA-naive and -experienced patients assigned warfarin, the time in therapeutic range (international normalized ratio 2.0 to 3.0) was 62% and 67%, respectively, and 61% and 66% for those never and ever prescribed a VKA. In VKA-naive patients, stroke and systemic embolism rates were 1.57%, 1.07%, and 1.69% per year for D110, D150, and warfarin, respectively. D110 was similar to warfarin (P 0.65); D150 was superior (P 0.005). Major bleeding rates were 3.11%, 3.34%, and 3.57% per year, respectively. D110 and D150 were similar to warfarin (P 0.19 and P 0.55). Intracranial bleeding rates were 0.19%, 0.33%, and 0.73% per year, respectively. D110 and D150 were lower than warfarin (P and P 0.005). In VKA-experienced patients, stroke and systemic embolism rates were 1.51%, 1.15%, and 1.74% per year for D110, D150, and warfarin, respectively. D110 was similar to warfarin (P 0.32); D150 was superior (P 0.007). Major bleeding rates were 2.66%, 3.30%, and 3.57% per year, respectively. D110 was lower than warfarin (P 0.003); D150 was similar (P 0.41). Intracranial bleeding rates were 0.26%, 0.32%, and 0.79% per year, respectively. D110 and D150 were lower than warfarin (P for both). Results were similar for patients never on a VKA. Conclusions Previous VKA exposure does not influence the benefits of dabigatran at either dose compared with warfarin. Clinical Trial Registration Unique identifier: NCT (Circulation. 2010;122: ) Key Words: anticoagulants arrhythmia atrial fibrillation stroke prevention The Randomized Evaluation of Long-term anticoagulation therapy (RE-LY) trial was a global trial that evaluated 2 doses of the novel anticoagulant dabigatran (110 mg BID [D110] and 150 mg BID [D150]) against usual therapy with warfarin (international normalized ratio [INR] 2.0 to 3.0) for the prevention of stroke and systemic embolism in patients with nonvalvular atrial fibrillation. 1 The design and main results have been published. 1,2 This analysis is a prespecified substudy of the RE-LY trial comparing 2 subpopulations: patients naive to and experienced with vitamin K antagonists (VKAs). 2 All patients in the RE-LY trial fell into 1 of these 2 distinct subgroups. The primary study outcome of RE-LY was stroke or systemic embolism. The primary safety outcome was major hemorrhage. Other outcomes included myocardial infarction, pulmonary embolism, transient ischemic attack, hospitalization, and death. An international team of adjudicators reviewed documents in local languages after blinding. Each primary and secondary outcome was adjudicated by 2 independent investigators who were blinded to assigned treatment. All transient ischemic attacks were reviewed to ensure that strokes had not been missed. To detect possible unreported events, symptom questionnaires were regularly administered to patients, and adverse event and hospitalization reports were scrutinized to detect unreported primary or secondary outcomes. 1,2 Clinical Perspective on p 2253 Post hoc analyses of clinical trials and reviews of registry cohorts suggest that patients with atrial fibrillation who are naive to VKAs have a different response when first treated Received June 22, 2010; accepted September 24, From the Lankenau Institute for Medical Research and The Heart Center, Wynnewood, Pa (M.D.E., A.P., M.R.C., T.H.A.); Uppsala Clinical Research Centre and Department of Medical Sciences, Uppsala University, Uppsala, Sweden (L.W.); Population Health Research Institute, McMaster University and Hamilton Health Sciences, Hamilton, Ontario, Canada (S.J.C., J.P., S. Yang, S. Yusuf); Boehringer Ingelheim Pharmaceuticals, Ridgefield, Conn (P.A.R.); and University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK (G.Y.H.L.). Correspondence to Michael D. Ezekowitz, MBChB, DPhil, FRCP, Lankenau Institute for Medical Research, 100 Lancaster Ave, Suite G36, Wynnewood, PA ezekowitzm@mlhs.org 2010 American Heart Association, Inc. Circulation is available at DOI: /CIRCULATIONAHA

2 Ezekowitz et al VKA-Naive Analysis of RE-LY 2247 with warfarin compared with those who are VKA experienced It is believed that patients who are VKA experienced have demonstrated the ability to comply with an anticoagulation regimen, have a determined, personalized dose of the VKA that achieves a therapeutic INR value between 2.0 and 3.0, and have passed the VKA stress test, thereby reducing the chance that sources of major bleeding are yet to be uncovered. Dabigatran is a direct thrombin inhibitor. After oral administration, patients are therapeutically anticoagulated within 2 hours. It is 80% renally excreted. 11 In the primary analysis, D110 was not inferior and D150 was superior to warfarin for prevention of stroke and systemic embolism, whereas major bleeding rates were lower for D110 and similar for D150 compared with warfarin, and intracranial bleeding rates were lower for both dabigatran doses. 1 Given the uncertainty relative to the evaluation of a novel anticoagulant in VKAnaive and -experienced patients, RE-LY was designed to have a balanced enrollment of VKA-naive and -experienced patients. 1,2 It is the first trial to prospectively evaluate a novel anticoagulant in VKA-naive and -experienced populations. Hypothesis It is expected that there will be an interaction between treatment groups and prior VKA experience on the outcomes of the study. Methods Study Participants Patients (18 113) were recruited from 951 clinical centers in 44 countries. Patients were eligible if they had atrial fibrillation documented on ECG at screening or within 6 months before and at least 1 of the following characteristics: previous stroke or transient ischemic attack, a left ventricular ejection fraction of 40%, New York Heart Association class II or higher heart failure symptoms within 6 months before screening, an age of at least 75 years, or an age of 65 to 74 years with diabetes mellitus, hypertension, or coronary artery disease. Reasons for exclusion were presence of a severe heart valve disorder, stroke within 14 days or severe stroke within 6 months before screening, a condition that increased the risk of hemorrhage, a creatinine clearance of 30 ml/min, active liver disease, and pregnancy or chance of becoming pregnant. 2 In the RE-LY trial, VKA-naive was defined as having a total of 62 days of lifetime VKA exposure. 1 This time frame was chosen to allow for identification of new-onset atrial fibrillation in the hospital and recruitment as an outpatient. Initially, the trial did not mandate but encouraged sites to voluntarily recruit equal proportions of VKA-naive and -experienced patients. After 8 months of enrollment, it became clear that this goal would not be met. To achieve a balanced cohort, the RE-LY operations committee implemented a quota system that required every site to enroll 50% VKA-naive patients from then on. Because there is no standard definition of VKA naive, secondary analyses of the relative impacts of the 2 therapies were performed with 2 alternative definitions of VKA naive: patients not on a VKA at randomization 3 6 and patients who had never been on a VKA. These secondary analyses were performed to evaluate the sensitivity of the results to the choice of definition. Administration The study was funded by Boehringer Ingelheim and coordinated by the Population Health Research Institute (Hamilton, Ontario, Canada). An operations committee composed of the 2 co principal investigators (M.D.E., S.J.C.), 2 cochairmen (L.W., S.Y.), and 2 pharmaceutical company representatives (P.A.R., M. Haehl), with assistance from an international steering committee, was responsible for the design, conduct, data analysis, and reporting of the study. The study was approved by all appropriate national regulatory authorities and ethics committees. All patients provided written, informed consent. All authors vouch for the accuracy and completeness of the data and this analysis. Statistical Analysis The baseline characteristics of the VKA-naive and -experienced cohorts were compared by Fisher exact test, as were the reasons for permanent study drug discontinuation between treatment groups. To examine study outcomes, hazard ratios (equivalent to relative risk [RR]) were calculated comparing D110 and D150 with warfarin in VKA-naive and -experienced patients. Two-sided 95% confidence intervals (CIs) were presented in each case, and P values for interactions were used to determine whether the relative effects of treatment on the major outcomes varied by the subgroups of interest. 12 Within each subgroup, P values comparing dabigatran with warfarin are reported with the use of a test for superiority. This analysis was repeated for the 2 additional definitions of VKA-naive patients: patients not on a VKA at randomization and patients who have never been on a VKA. We planned to enroll patients, which we estimated would provide 84% power to evaluate the noninferiority of each dose of dabigatran. A protocol change was made by the operations committee during the enrollment period without knowledge of emerging treatment effects to increase the sample size to patients to maintain the statistical power in case of a low event rate. No formal power calculations were made for the VKA-naive and -experienced cohorts. Results Characteristics of the Study Participants With the use of the RE-LY definition, approximately half of the patients were VKA naive (50.4%) (Table 1). 13 Compared with the VKA-experienced cohort, the VKA-naive cohort had fewer men (58.9% versus 68.4%; P 0.001), fewer patients with diabetes mellitus (22.5% versus 24.2%; P 0.007), more patients with hypertension (80.4% versus 77.3%; P 0.001), and a higher use of aspirin (54.5% versus 24.8%; P 0.001). There were fewer patients with prior myocardial infarction (14.9% versus 18.3%; P 0.001), coronary artery disease (24.7% versus 30.9%; P 0.001), stroke (11.0% versus 14.1%; P 0.001), and transient ischemic attack (7.9% versus 10.4%; P 0.001). There was no significant difference in congestive heart failure rates between cohorts. The mean CHADS 2 score was 2.11 in the VKA-naive cohort and 2.18 in the VKA-experienced cohort (P 0.001; CHADS 2 is an acronym for cardiac failure, hypertension, age, diabetes mellitus, and prior stroke or TIA). With the use of the not-on-a-vka-at-randomization and never-been-on-a-vka definitions of VKA naive, VKA-naive patients were more likely to have congestive heart failure than VKA-experienced patients but were otherwise similar to the patient population according to the RE-LY definition (Table 1). Mean time in therapeutic INR range (TTR) for patients on warfarin was 64% for the study overall and 62% and 67% for VKA-naive (RE-LY definition) and -experienced cohorts, respectively, with the use of the Rosendaal method. 14 The differences occurred primarily in the early months of the study; the TTR in VKA-naive patients was 43%, 62%, and 67% in months 1, 6, and 12 after randomization, whereas that

3 2248 Circulation November 30, 2010 Table 1. Treatment Groups and Baseline Characteristics of the Study Participants by VKA Cohort and VKA Status Definition* VKA Naive VKA Experienced P Never Been on a VKA Ever Been on a VKA P No VKA at Randomization VKA at Randomization P Total randomized 9123 (50.4) 8989 (49.6) 5748 (32.7) (67.3) 6924 (38.2) (61.8) Dabigatran, 110 mg BID 3004 (49.9) 3011 (50.1) 1868 (31.1) 3952 (65.7) 2264 (37.6) 3751 (62.4) Dabigatran, 150 mg BID 3026 (49.8) 3049 (50.2) 1909 (31.4) 3979 (65.5) 2317 (38.1) 3759 (61.9) Warfarin 3093 (51.4) 2929 (48.6) 1972 (32.7) 3899 (64.7) 2344 (38.9) 3678 (61.1) Mean age, y Age 75 y 4755 (52.1) 4675 (52.0) (53.6) 6275 (53.0) (53.5) 5955 (53.2) 0.71 Male 5369 (58.9) 6145 (68.4) (57.7) 7867 (66.5) (59.1) 7419 (66.3) CHADS 2 score Congestive heart failure 2890 (31.7) 2903 (32.3) (33.4) 3736 (31.6) (33.9) 3444 (30.8) Hypertension 7331 (80.4) 6951 (77.3) (80.4) 9227 (78.0) (79.9) 8753 (78.2) Diabetes mellitus 2049 (22.5) 2172 (24.2) (21.9) 2843 (24.0) (22.4) 2673 (23.9) 0.02 Coronary artery disease 2257 (24.7) 2777 (30.9) (22.8) 3559 (30.1) (24.7) 3323 (29.7) Prior myocardial infarction 1363 (14.9) 1642 (18.3) (14.0) 2109 (17.8) (15.0) 1966 (17.6) Prior stroke 1002 (11.0) 1271 (14.1) (10.7) 1603 (13.6) (11.3) 1492 (13.3) Prior transient ischemic attack 725 (7.9) 938 (10.4) (7.6) 1184 (10.0) (7.8) 1121 (10.0) Baseline aspirin use 4971 (54.5) 2226 (24.8) (62.4) 3350 (28.3) (62.3) 2887 (25.8) *All data are presented as number (percentage) unless otherwise noted. Percentages in the top portion were calculated over total population and total randomized to each drug group, respectively. Percentages in the bottom portion were calculated over total VKA-naive and -experienced populations. VKA naive is defined as a total lifetime use of a VKA of 62 days. The CHADS 2 score is a measure of the risk of stroke in which congestive heart failure, hypertension, an age of 75 years or older, and diabetes mellitus are each assigned 1 point and previous stroke or transient ischemic attack is assigned 2 points; the score is calculated by summing all the points for a given patient. 13 for VKA-experienced patients was 55%, 68%, and 69% in months 1, 6, and 12. Permanent Study Drug Discontinuation Rates In each of the 3 treatment groups, VKA-naive patients had higher rates of permanent study drug discontinuation compared with VKA-experienced patients. In the warfarin and D110 groups, this was statistically significant (P and P 0.02, respectively). Rates of discontinuation because of gastrointestinal symptoms in both the VKA-naive and -experienced populations were higher for dabigatran compared with warfarin (D110 versus warfarin, VKA naive, 2.3% versus 0.9%, P 0.001; VKA experienced, 2.1% versus 0.4%, P 0.001; D150 versus warfarin, VKA naive, 2.1% versus 0.9%, P 0.001; VKA experienced, 2.1% versus 0.4%, P 0.001) (Table 2). Discontinuation due to patient decision was lower in the VKA-experienced cohort than in the VKA-naive cohort for both D150 and warfarin (P 0.03 and P 0.001, respectively) and was similar in each for D110 (Table 3). With the use of the never-been-on-a-vka and not-on-a-vkaat-randomization definitions of VKA naive, comparisons of discontinuation rates were similar to those with the use of the RE-LY definition. Primary Outcome Combined stroke and systemic embolism rates were similar in the D110 group for both the VKA-naive and -experienced cohorts compared with warfarin (RR 0.93; 95% CI, 0.70 to 1.25; P 0.65 and RR 0.87; 95% CI, 0.66 to 1.15; P 0.32, respectively; P for interaction 0.72) (Table 4 and Figure 1). In the D150 group, both VKA-naive (RR 0.63; 95% CI, 0.46 to 0.87; P 0.005) and -experienced patients (RR 0.66; 95% CI, 0.49 to 0.89; P 0.007) had a statistically significantly lower risk of stroke or systemic embolism compared with warfarin (P for interaction 0.84). With the use of the 2 alternative definitions of VKA naive, comparisons of primary outcome rates were similar, and interactions were nonsignificant. Major Bleeding Major bleeding rates were lower in VKA-experienced patients when D110 was compared with warfarin (RR 0.74; 95% CI, 0.60 to 0.90; P 0.003) (Table 4 and Figure 2). VKA-naive patients in the D110 group (RR 0.87; 95% CI, 0.72 to 1.07; P 0.19) and VKA-naive (RR 0.94; 95% CI, 0.77 to 1.15; P 0.55) and -experienced patients (RR 0.92; 95% CI, 0.76 to 1.12; P 0.41) in the D150 group were similar compared with warfarin. The P values for the interaction of treatment group and VKA status are 0.25 for D110 versus warfarin and 0.90 for D150 versus warfarin. With the use of the 2 alternative definitions of VKA naive, comparisons of major bleeding rates were similar, and interactions were nonsignificant. Intracranial Bleeding Intracranial bleeding rates were lower in the D110 VKAnaive and -experienced cohorts (RR 0.27; 95% CI, 0.14 to 0.52; P 0.001; RR 0.32; 95% CI, 0.18 to 0.56; P 0.001, respectively; P for interaction 0.66) and the D150 VKAnaive and -experienced cohorts (RR 0.46; 95% CI, 0.27 to 0.78; P 0.005; RR 0.40; 95% CI, 0.24 to 0.67; P 0.001, respectively; P for interaction 0.71) compared with warfarin (Table 4 and Figure 3).

4 Ezekowitz et al VKA-Naive Analysis of RE-LY 2249 Table 2. Permanent Study Drug Discontinuation by Treatment Group, VKA Cohort, and VKA Status Definition* Total Permanent Discontinuation Gastrointestinal Symptoms With the use of both the never-been-on-a-vka and the not-on-a-vka-at-randomization definitions of VKA naive, intracranial bleeding rates for D150 VKA-naive patients were similar to those for warfarin. Rates for D150 VKAexperienced and D110 VKA-naive and -experienced patients were lower compared with warfarin. Interactions were nonsignificant (Table 4). Other Outcomes Life-Threatening Bleeding, Disabling Stroke, and Death Rates of life-threatening bleeding, disabling stroke, and death (when combined) were statistically significantly lower in the VKA-experienced cohort in both the D110 (RR 0.82; 95% CI, 0.70 to 0.96; P 0.01) and D150 groups (RR 0.80; 95% CI, 0.68 to 0.93; P 0.004) compared with warfarin but similar for VKA-naive patients (Table 4). Myocardial Infarction There were no significant differences in the rates of myocardial infarction. Patient Decision Serious Adverse Event Outcome Event Gastro- Intestinal Bleeding VKA naive, n (%) Total 1880 (20.6) 162 (1.8) 713 (7.8) 217 (2.4) 238 (2.6) 111 (1.2) D (21.3) (2.3) (7.5) (2.8) (3.1) (1.1) 0.90 D (21.7) (2.1) (8.3) (2.8) (2.5) (1.6) 0.20 Warfarin 582 (18.8) 27 (0.9) 237 (7.7) 49 (1.6) 68 (2.2) 31 (1.0) VKA experienced, n (%) Total 1679 (18.7) 140 (1.6) 576 (6.4) 217 (2.4) 248 (2.8) 81 (0.9) D (20.1) (2.1) (7.2) (2.6) (3.3) (0.8) 0.37 D (21.6) (2.1) (7.3) (2.7) (2.9) (1.1) 0.78 Warfarin 416 (14.2) 11 (0.4) 138 (4.7) 56 (1.9) 62 (2.1) 23 (0.8) Never been on a VKA, n (%) Total 1242 (21.6) 111 (1.9) 510 (8.9) 128 (2.2) 149 (2.6) 67 (1.2) D (21.5) (2.5) (8.1) (2.7) (2.8) (0.9) 0.50 D (23.3) (2.5) (9.7) (2.5) (2.8) (1.6) 0.25 Warfarin 396 (20.1) 17 (0.9) 172 (8.7) 31 (1.6) 44 (2.2) 20 (1.0) Ever been on a VKA, n (%) Total 2187 (18.5) 186 (1.6) 738 (6.2) 290 (2.5) 324 (2.7) 117 (1.0) D (20.2) (2.2) (7.0) (2.7) (3.4) (1.0) 0.71 D (20.7) (2.0) (6.9) (2.9) (2.6) (1.1) 0.81 Warfarin 567 (14.5) 21 (0.5) 189 (4.8) 70 (1.8) 83 (2.1) 32 (0.8) No VKA at randomization, n (%) Total 1506 (21.8) 132 (1.9) 601 (8.7) 162 (2.3) 187 (2.7) 79 (1.1) D (21.8) (2.5) (8.1) (3.0) (3.0) (0.9) 0.76 D (23.3) (2.5) (9.4) (2.5) (2.8) (1.5) 0.28 Warfarin 474 (20.2) 19 (0.8) 200 (8.5) 36 (1.5) 54 (2.3) 23 (1.0) VKA at randomization, n (%) Total 2053 (18.3) 170 (1.5) 688 (6.1) 272 (2.4) 299 (2.7) 113 (1.0) D (20.0) (2.1) (6.9) (2.6) (3.3) (1.0) 0.45 D (20.7) (1.9) (6.8) (2.8) (2.6) (1.2) 1.00 Warfarin 524 (14.2) 19 (0.5) 175 (4.8) 69 (1.9) 76 (2.1) 31 (0.8) *All percentages were calculated over total patients in each treatment group by cohort. Total permanent discontinuation P values were calculated with total patients in each treatment group by cohort. All other P values were calculated with total permanent discontinuations in each treatment group by cohort. VKA naive is defined as a total lifetime use of a VKA of 62 days. Gastrointestinal symptoms included pain, vomiting, and diarrhea. Gastrointestinal Bleeding Gastrointestinal bleeding rates were similar for D110 compared with warfarin but significantly higher in both the D150 VKA-naive (RR 1.56; 95% CI, 1.15 to 2.10; P 0.004) and -experienced cohorts (RR 1.42; 95% CI, 1.06 to 1.89; P 0.02) compared with warfarin. With the use of the 2 alternative definitions of VKA naive, rates of these outcomes were similar except for a significant interaction between treatment group and VKA status for D150 versus warfarin on the life-threatening bleeding, disabling stroke, and death outcome with the not-on-a-vka-atrandomization definition (P for interaction 0.04) (Table 4). Comparison of VKA-Naive and -Experienced Patients Within Treatment Groups With the use of the RE-LY definition of VKA naive, in the D110 group the life-threatening bleeding, disabling stroke, and death (combined) rate was lower in VKA-experienced patients than in VKA-naive patients (RR 0.83; 95% CI, 0.71 to 0.98; P 0.03), as was the cardiovascular death rate

5 2250 Circulation November 30, 2010 Table 3. Comparison of VKA-Naive and -Experienced Cohorts (RE-LY Definition) Within Assigned Treatment Groups* D110 Experienced vs Naive D150 Experienced vs Naive Warfarin Experienced vs Naive RR (95% CI) P RR (95% CI) P RR (95% CI) P Permanent study drug discontinuation Gastrointestinal symptoms 0.97 ( ) ( ) ( ) 0.31 Serious adverse event 0.94 ( ) ( ) ( ) 0.33 Patient decision 0.92 ( ) ( ) ( ) Outcome event 1.06 ( ) ( ) ( ) 0.83 Gastrointestinal bleeding 0.80 ( ) ( ) ( ) 0.88 Efficacy and safety outcome events Stroke or systemic embolism 0.93 ( ) ( ) ( ) 0.83 Major bleeding 0.87 ( ) ( ) ( ) 0.82 Intracranial bleeding 1.22 ( ) ( ) ( ) 0.72 Life-threatening bleeding, disabling 0.83 ( ) ( ) ( ) 0.92 stroke, and death Myocardial infarction 0.80 ( ) ( ) ( ) 0.69 Gastrointestinal bleeding 0.86 ( ) ( ) ( ) 0.66 Cardiovascular death 0.73 ( ) ( ) ( ) 0.41 *The RE-LY definition of VKA naive is a total lifetime use of a VKA of 62 days. (RR 0.73; 95% CI, 0.58 to 0.92; P 0.007) (Table 3). In the D150 group, the life-threatening bleeding, disabling stroke, and death and cardiovascular death rates trended lower in VKA-experienced patients compared with VKA-naive patients. There were no other significant differences in RR for any outcomes between VKA-naive and -experienced patients within treatment groups. Discussion RE-LY was the first prospective study evaluating a novel oral anticoagulant to include a VKA-naive cohort for comparison with a VKA-experienced cohort as a prespecified analysis. 1,2 The hypothesis tested was supported by the lower TTR in VKA-naive patients assigned to warfarin and the higher withdrawal rates in the VKA-naive cohort irrespective of treatment assignment. VKA-naive patients had a lower TTR than VKA-experienced in the early months of therapy. Despite higher permanent assigned therapy discontinuation rates in VKA-naive patients (the analysis was on an intentionto-treat basis) and a higher TTR in VKA-experienced patients, there was no significant interaction between treatment and prior VKA use on major outcomes. The only exception was for the combined secondary outcome measure of lifethreatening bleeding, disabling stroke, and death in the D150 group compared with warfarin and only for the not-on-a- VKA-at-randomization definition (P 0.04) (Table 4 and Figure 1). The alternative definitions were not part of the prespecified analysis, and the importance of this interaction remains unclear. There were differences in baseline characteristics (particularly prior stroke, myocardial infarction, and transient ischemic attack, gender, coronary artery disease, and baseline aspirin use) between the VKA-naive and -experienced cohorts. Some, but not all, of these differences might be explained by multiple testing and/or the large sample size that enables small, clinically nonsignificant effects to be detected. 15 Despite demographic differences between VKA-naive and -experienced patients, the outcome of the study was not affected. There has been inconsistency in the definition of VKAnaive in previous studies. Several early studies of warfarin versus placebo or aspirin recruited patients who were VKA naive, but the definition of VKA naive varied. For example, the Copenhagen Atrial Fibrillation, Aspirin and Anticoagulation (AFASAK) trial excluded patients with 6 months of previous anticoagulation therapy. 7 The Stroke Prevention in Nonrheumatic Atrial Fibrillation (SPINAF) trial had patients stop anticoagulation for 6 months before randomization to warfarin or placebo. 8 More recent trials, such as the Stroke Prevention using an ORal Thrombin Inhibitor in atrial Fibrillation (SPORTIF) III and V trials, which compared ximelagatran with warfarin for stroke risk reduction in atrial fibrillation, included VKA-naive patients defined as those not on a VKA at study entry. These patients constituted 27% and 15% of the total number in the warfarin arms of SPORTIF III and V, respectively. 3,4 With the use of the SPORTIF definition, a post hoc analysis found higher INR variability for VKA-naive patients versus experienced patients (SD, 0.85 versus 0.61; P 0.001), 9 a finding similar to that reported here. Greater INR variability and lower TTR values are strong predictors of high event rates on VKAs. 16 The Atrial fibrillation Clopidogrel Trial with Irbesartan for prevention of Vascular Events (ACTIVE W) trial compared a combination of clopidogrel and aspirin with warfarin for prevention of vascular events in patients with atrial fibrillation. Like the SPORTIF trials, ACTIVE W did not prospectively recruit a cohort of VKA-naive patients. In a post hoc analysis of the trial, 22% of patients in the warfarin arm were VKA naive (not on a VKA at study entry). With the use of this definition,

6 Ezekowitz et al VKA-Naive Analysis of RE-LY 2251 Table 4. Efficacy and Safety Outcomes by Treatment Group, VKA Cohort, and VKA Status Definition* D110 D150 Warfarin D110 vs Warfarin D150 vs Warfarin Event n %/y n %/y n %/y RR (95% CI) P Interaction P RR (95% CI) P Interaction P Stroke or systemic embolism VKA naive ( ) ( ) VKA experienced ( ) ( ) Never been on a VKA ( ) ( ) Ever been on a VKA ( ) ( ) No VKA at randomization ( ) ( ) VKA at randomization ( ) ( ) Major bleeding VKA naive ( ) ( ) 0.55 VKA experienced ( ) ( ) Never been on a VKA ( ) ( ) 0.75 Ever been on a VKA ( ) ( ) No VKA at randomization ( ) ( ) 0.88 VKA at randomization ( ) ( ) Intracranial bleeding VKA naive ( ) ( ) VKA experienced ( ) ( ) Never been on a VKA ( ) ( ) 0.16 Ever been on a VKA ( ) ( ) No VKA at randomization ( ) ( ) 0.07 VKA at randomization ( ) ( ) Life-threatening bleeding, disabling stroke, and death VKA naive ( ) ( ) 0.38 VKA experienced ( ) ( ) Never been on a VKA ( ) ( ) 0.74 Ever been on a VKA ( ) ( ) No VKA at randomization ( ) ( ) 0.90 VKA at randomization ( ) ( ) Myocardial infarction VKA naive ( ) ( ) 0.40 VKA experienced ( ) ( ) Never been on a VKA ( ) ( ) 0.41 Ever been on a VKA ( ) ( ) No VKA at randomization ( ) ( ) 0.13 VKA at randomization ( ) ( ) Gastrointestinal bleeding VKA naive ( ) ( ) VKA experienced ( ) ( ) Never been on a VKA ( ) ( ) 0.01 Ever been on a VKA ( ) ( ) No VKA at randomization ( ) ( ) VKA at randomization ( ) ( ) *VKA naive is defined as a total lifetime use of a VKA of 62 days. Interactive P values are for the interaction between treatment group and VKA status by definition of VKA status. unlike the findings from RE-LY, VKA-naive patients randomized to warfarin had higher rates of stroke and systemic embolism (4.71% versus 3.72% per year) and major hemorrhage (2.92% versus 2.02% per year) compared with VKAexperienced patients, but, like RE-LY, they had a higher rate of warfarin discontinuation (P 0.008) and worse INR control (P 0.001). 5 The Birmingham Atrial Fibrillation Treatment of the Aged (BAFTA) trial compared warfarin with

7 2252 Circulation November 30, 2010 Figure 1. Comparisons of primary end points by treatment group for the overall population and VKA-naive and -experienced cohorts. VKA naive is defined as a total lifetime use of a VKA of 62 days. Interaction P values are for the interaction between treatment group and VKA status. aspirin for stroke risk reduction in elderly patients with atrial fibrillation. Sixty percent of patients randomized to warfarin were VKA naive (not on a VKA at study entry). By this definition, as in RE-LY (although in an older population), stroke and systemic embolism and major hemorrhage rates in the warfarin arm were similar in the VKA-naive and -experienced cohorts (stroke and systemic embolism, 2.0% versus 1.4% per year; major hemorrhage, 2.1% versus 1.6% per year). 6 Because there is no standard definition of VKA naive, the analyses for the current report were done with the use of the prespecified definition from the RE-LY trial ( 62 days of lifetime VKA use) and 2 alternative definitions (not on a VKA at randomization and were never on a VKA). The Figure 2. Cumulative hazard rates for major bleeding by treatment group (VKA-naive and -experienced cohorts). VKA naive is defined as a total lifetime use of a VKA of 62 days. Figure 3. Cumulative hazard rates for intracranial bleeding by treatment group (VKA-naive and -experienced cohorts). VKA naive is defined as a total lifetime use of a VKA of 62 days.

8 Ezekowitz et al VKA-Naive Analysis of RE-LY 2253 results were similar for all 3 definitions, demonstrating that dabigatran is a safe and effective alternative to warfarin in this context regardless of length or timing of prior VKA use. The definitions evaluated represent extremes, and thus one can conclude that the choice of definition did not matter. If differences in outcomes between VKA-experienced and -naive patients are only important for a short time after onset of treatment, the relatively long mean duration of exposure of 2 years in the RE-LY trial may have masked the short-term differences. The differences in INR control for the first 6 months in the warfarin treatment group suggest that this may be the case for VKAs. The clinical implications of this analysis of the RE-LY trial are important. The results of this analysis demonstrate that both patients starting dabigatran without prior VKA experience and those switching to dabigatran from a VKA benefit from dabigatran at either dose compared with warfarin. Acknowledgments We would like to thank the patients who participated in the trial, the Steering Committee of the trial, and the staff at each site. Sources of Funding This study was funded by a grant from Boehringer Ingelheim. Disclosures Dr Ezekowitz, Dr Wallentin, Dr Connolly, Dr Lip, and Dr Yusuf received research grants, consulting fees, and lecture fees from Boehringer Ingelheim Pharmaceuticals Inc. Dr Reilly is an employee of Boehringer Ingelheim. Dr Parekh received grant support from Boehringer Ingelheim. Dr Chernick, J. Pogue, T.H. Aikens, and S. Yang have no disclosures. References 1. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, Pogue J, Reilly PA, Themeles E, Varrone J, Wang S, Alings M, Xavier D, Zhu J, Diaz R, Lewis BS, Darius H, Diener H-C, Joyner CD, Wallentin L; the RE-LY Steering Committee and Investigators. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361: Ezekowitz MD, Connolly S, Parekh A, Reilly PA, Varrone J, Wang S, Oldgren J, Themeles E, Wallentin L, Yusuf S. Rationale and design of RE-LY: randomized evaluation of long-term anticoagulant therapy, warfarin, compared with dabigatran. Am Heart J. 2009;157: ,810.e Albers GW, Diener HC, Grind M, Halperin JL, Horrow J, Olsson SB, Petersen P, Vahanian A, Frison L, Nevinson M, Partridge S; Executive Steering Committee on Behalf of the SPORTIF III Investigators. Stroke prevention with the oral direct thrombin inhibitor ximelagatran compared with warfarin in patients with non-valvular atrial fibrillation (SPORTIF III): randomised controlled trial. Lancet. 2003;362: Albers GW, Diener HC, Frison L, Grind M, Nevinson M, Partridge S, Halperin JL, Horrow J, Olsson SB, Petersen P, Vahanian A; SPORTIF Executive Steering Committee for the SPORTIF V Investigators. Ximelagatran vs warfarin for stroke prevention in patients with nonvalvular atrial fibrillation. JAMA. 2005;293: Connolly S, Pogue J, Hart R, Pfeffer M, Hohnloser S, Chrolavicius S, Yusuf S; the ACTIVE Writing Group on behalf of the ACTIVE Investigators. Clopidogrel plus aspirin versus oral anticoagulation for atrial fibrillation in the Atrial fibrillation Clopidogrel Trial with Irbesartan for prevention of Vascular Events (ACTIVE W): a randomised controlled trial. Lancet. 2006;367: Mant J, Hobbs FDR, Fletcher K, Roalfe A, Fitzmaurice D, Lip GYH, Murray E; on behalf of the BAFTA investigators and the Midland Research Practices Network (MidReC). Warfarin versus aspirin for stroke prevention in an elderly community population with atrial fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): a randomized controlled trial. Lancet. 2007;370: Peterson P, Boysen G, Godtfredsen J, Andersen ED, Andersen B. Placebo-controlled, randomized trial of warfarin and aspirin for prevention of thromboembolic complications in chronic atrial fibrillation: the Copenhagen AFASAK study. Lancet. 1989;1: Ezekowitz MD, Bridgers SL, James KE, Carliner NH, Colling CL, Gornick CC, Krause-Steinrauf H, Kurtzke JF, Nazarian SM, Radford MJ, Rickles FR, Shabetai R, Deykin D; for the Veterans Affairs Stroke Prevention in Nonrheumatic Atrial Fibrillation Investigators. Warfarin in the prevention of stroke associated with nonrheumatic atrial fibrillation. N Engl J Med. 1992;327: Hylek EM, Frison L, Henault LE, Cupples A. Disparate stroke rates on warfarin among contemporaneous cohorts with atrial fibrillation: potential insights into risk from a comparative analysis of SPORTIF III versus SPORTIF V. Stroke. 2008;39: Hylek EM, Evans-Molina C, Shea C, Henault LE, Regan S. Major hemorrhage and tolerability of warfarin in the first year of therapy among elderly patients with atrial fibrillation. Circulation. 2007;115: Stangier J. Clinical pharmacokinetics and pharmacodynamics of the oral direct thrombin inhibitor dabigatran etexilate. Clin Pharmacokinet. 2008; 47: Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in medicine: reporting of subgroup analyses in clinical trials. N Engl J Med. 2007;357: Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA. 2001; 285: Rosendaal FR, Cannegieter SC, van der Meer FJ, Briët E. A method to determine the optimal intensity of oral anticoagulant therapy. Thromb Haemost. 1993;69: Berger VW. Selection Bias and Covariate Imbalances in Randomized Clinical Trials. Chichester, UK: John Wiley & Sons; Connolly SJ, Pogue J, Eikelboom J, Flaker G, Commerford P, Franzosi MG, Healey JS, Yusuf S; on behalf of the ACTIVE W Investigators. Benefit of oral anticoagulant over antiplatelet therapy in atrial fibrillation depends on the quality of international normalized ratio control achieved by centers and countries as measured by time in therapeutic range. Circulation. 2008;118: CLINICAL PERSPECTIVE This report compares the novel anticoagulant dabigatran with the current standard, warfarin, in vitamin K antagonist naive and experienced populations of patients with atrial fibrillation who are at risk of stroke. It was believed that patients who were vitamin K antagonist experienced were likely to do better than patients who are new to anticoagulation because they have demonstrated the ability to comply with an anticoagulation regimen, they have a personalized dose of warfarin that achieves a therapeutic international normalized ratio value, and they have passed the vitamin K antagonist stress test, thereby reducing the chance of uncovering major sources of bleeding. This report found that dabigatran, the novel direct thrombin inhibitor, was better than the comparator warfarin for both vitamin K antagonist naive and experienced patients and therefore could be used clinically in both cases.