- Email: [email protected]
Efficacy and Safety of Rivaroxaban Versus Warfarin in Patients Taking Nondihydropyridine Calcium Channel Blockers for Atrial Fibrillation (from the ROCKET AF Trial)
Efficacy and Safety of Rivaroxaban Versus Warfarin in Patients Taking Nondihydropyridine Calcium Channel Blockers for Atrial Fibrillation (from the ROCKET AF Trial) Jeffrey B. Washam, PharmDa,*, Anne S. Hellkamp, MSa, Yuliya Lokhnygina, PhDa, Jonathan P. Piccini, MD, MHSa, Scott D. Berkowitz, MDb, Christopher C. Nessel, MDc, Richard C. Becker, MDd, Günter Breithardt, MDe, Keith A.A. Fox, MBChBf, Jonathan L. Halperin, MDg, Graeme J. Hankey, MDh, Kenneth W. Mahaffey, MDi, Daniel E. Singer, MDj, and Manesh R. Patel, MDa, on behalf of the ROCKET AF Steering Committee and Investigators Non-dihydropyridine calcium channel blockers (non-DHP CCBs) possess combined P-glycoprotein and moderate CYP3A4 inhibition, which may lead to increased exposure of medications that are substrates for these metabolic pathways, such as rivaroxaban. We evaluated the use and outcomes of non-DHP CCBs in patients with atrial fibrillation (AF) in 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). We assessed clinical outcomes in patients who received non-DHP CCBs and the impact on the efficacy and safety of rivaroxaban compared with warfarin. Stroke or noncentral nervous system (CNS) systemic embolism (SE), major or nonmajor clinically relevant (NMCR) bleeding, all-cause death, and major bleeding were compared according to nonDHP CCB use. At randomization, 1,308 patients (9.2%) were taking a non-DHP CCB. They were more likely to be women, have diabetes and COPD, and less likely to have heart failure and had a lower mean CHADS2 score (3.3 vs 3.5). Non-DHP CCB use was not associated with an increased risk of stroke/non-CNS SE (p [ 0.11) or the composite outcome of NMCR or major bleeding (p [ 0.087). Non-DHP CCB use was associated with an increased risk of major bleeding (adjusted hazard ratio 1.50, 95% CI 1.11 to 2.04) and intracranial hemorrhage (adjusted hazard ratio 2.84, 95% CI 1.53 to 5.29). No significant difference was observed in the primary efficacy (stroke or non-CNS SE; adjusted interaction p value [ 0.38) or safety outcome (NMCR or major bleeding; adjusted interaction p value [ 0.14) between rivaroxaban and warfarin with non-DHP CCB use. In conclusion, although the overall use of non-DHP CCBs was associated with an increased risk of major bleeding and intracranial hemorrhage, the use was not associated with a significant change in the safety or efficacy of rivaroxaban compared with warfarin observed in ROCKET AF. Ó 2017 Elsevier Inc. All rights reserved. (Am J Cardiol 2017;120:588e594) The nondihydropyridine calcium channel blockers (non-DHP CCB), verapamil, and diltiazem, in addition to L-type CCBs, exhibit P-glycoprotein (P-gp) and moderate
a Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina; bBayer HealthCare Pharmaceuticals, Parsippany, New Jersey; cJanssen Research & Development, Raritan, New Jersey; d University of Cincinnati College of Medicine, Cincinnati, Ohio; eUniversity Hospital of Münster, Münster, Germany; fUniversity of Edinburgh, Royal Infirmary of Edinburgh, Edinburgh, United Kingdom; gMount Sinai School of Medicine, New York, New York; hSchool of Medicine and Pharmacology, University of Western Australia, Crawley, Australia; i Stanford Center for Clinical Research, Stanford University School of Medicine, Stanford, California; and jMassachusetts General Hospital, Harvard Medical School, Boston, Massachusetts. Manuscript received February 16, 2017; revised manuscript received and accepted May 5, 2017. The ROCKET AF trial was supported by Janssen Research & Development and Bayer HealthCare Pharmaceuticals. See page 593 for disclosure information. *Corresponding author: Tel: (919) 681-3286; fax: (919) 681-2960. E-mail address: [email protected] (J.B. Washam).
0002-9149/17/$ - see front matter Ó 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2017.05.026
hepatic cytochrome P450 3A4 (CYP3A4) enzyme inhibition. Rivaroxaban is an orally administered direct-acting factor Xa inhibitor that was compared with warfarin for the prevention of stroke or noncentral nervous system (CNS) systemic embolism (SE) in patients with nonvalvular atrial fibrillation in the Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation study (ROCKET AF).1 Rivaroxaban is a substrate for both the P-gp transport and CYP3A4 isoenzyme pathways. Therefore, there is a theoretical concern that concomitant use of rivaroxaban with diltiazem or verapamil could lead to an increase in rivaroxaban drug exposure and anticoagulant effect. To date, no clinical data have been reported on the association of concomitant use of these therapies on clinical outcomes in patients treated with rivaroxaban. The purpose of this analysis was to assess clinical outcomes in patients with atrial fibrillation on oral anticoagulation enrolled in ROCKET AF trial who received non-DHP CCB therapy and to assess the impact of this concomitant therapy on the www.ajconline.org
Arrhythmias and Conduction Disturbances/Non-DHP CCB Use in ROCKET AF
efficacy and safety of rivaroxaban compared with warfarin. Methods The design of the ROCKET AF study has previously been described.2 ROCKET AF was a double-blind, doubledummy trial that randomized 14,264 patients with atrial fibrillation to rivaroxaban (20 mg once daily or 15 mg once daily if creatinine clearance was 30 to 49 ml/min) or adjusted dose warfarin (target international normalized ratio 2.0 to 3.0). The primary efficacy end point was stroke or non-CNS SE and the primary safety end point was the composite of nonmajor clinically relevant (NMCR) and major bleeding.3 Major bleeding was defined as clinically overt bleeding that was associated with one or more of the following: (1) fall in hemoglobin of 2 g/dl or more; (2) a transfusion of 2 or more units of packed red blood cells or whole blood; (3) bleeding at a critical site (intracranial, intraaspinal, intraocular, pericardial, intraarticular, intramuscular with compartment syndrome, and retroperitoneal); or (4) a fatal outcome. NMCR bleeding was defined as overt bleeding not meeting the criteria for major bleeding but associated with medical intervention, unscheduled medical contact, cessation of study treatment, or associated with discomfort for the patient such as pain or impairment of activities of daily life. The intention-to-treat population was used for all efficacy analyses and safety analyses were based on the safety population (randomized patients who received at least 1 dose of study medication). Ninety-three patients from 1 enrolling site were excluded from efficacy analyses due to violations of Good Clinical Practice Guidelines. The overall clinical outcomes from ROCKET AF have previously been reported.1 The present study is a post hoc analysis of ROCKET AF including all randomized patients and grouping patients according to baseline and postrandomization use of a nonDHP CCB. Non-DHP CCB therapy was at the discretion of the treating physician and was captured as a concomitant medication on the case report form at baseline and during study follow-up (every 4 weeks until study completion). An independent clinical end point committee, unaware of treatment assignment, adjudicated clinical end points including all suspected stroke and SE, myocardial infarction, death, and bleeding events. Death events were subclassified into cardiovascular death (including sudden death) and noncardiovascular death. Summary statistics are presented for patterns of nonDHP CCB use, including the number of patients, agent, dose, and exposure time. Patients who took more than 1 non-DHP CCB were counted for the agent taken for the largest proportion of time; in instances when more than 1 dose is reported, the last dose was used. Baseline characteristics are presented for each group with categorical variables presented as counts (percentages) and comparisons were made using Pearson chi-square tests. Continuous variables are presented as medians (25th, 75th percentiles) and comparisons were made using Wilcoxon rank-sum tests. Cox proportional hazards models were used to assess the relation of non-DHP CCB use with subsequent outcomes,
589
with patients classified according to their baseline use; patients were included in models for as long as they remained in their baseline group. Patients who were taking non-DHP CCBs at baseline and subsequently stopped, or patients who were not taking one at baseline and subsequently started, were censored at those time points. Because either new start or cessation of non-DHP CCBs can be influenced by patient characteristics or intervening events that might also be related to the outcomes, patients were weighted by the inverse probability of continuing in their therapy group. Weights were applied to the Cox models with a robust sandwich variance estimator.4 Unadjusted summary event rates (events per 100 patient-years) were also weighted. For the recurrent event of hospitalizations, we used the method of Wei et al5 for multiple failure times with a robust sandwich variance estimator. These models incorporated the weighting described previously. Because only a small number of patients had multiple hospitalizations, these models included the first and second hospitalizations only. A single weighted parameter estimate was used to generate a significance test (z score) as well as a hazard ratio (HR) and CI. Event rates (events per 100 patient-years) and total number of events are presented for all clinical outcomes. Group comparisons were made using Cox models, adjusted for previously identified predictors of each end point. Efficacy outcomes models included the following covariates: age, gender, body mass index, region, diabetes, previous stroke/transient ischemic attack, vascular disease, congestive heart failure, hypertension, chronic obstructive pulmonary disease (COPD), paroxysmal atrial fibrillation, diastolic blood pressure, creatinine clearance (calculated using the Cockroft-Gault equation), heart rate, and abstinence from alcohol. Safety outcomes models included the following covariates: age, gender, region, previous stroke/transient ischemic attack, anemia, previous gastrointestinal bleed, COPD, diastolic blood pressure, creatinine clearance, platelets, albumin, previous aspirin use, previous vitamin K antagonist use, and previous thienopyridine use. Rates of missing data were quite low; when missing, covariates were imputed using the median for continuous variables and the mode for categorical variables within groups of patients on or not on a non-DHP CCB at baseline. All models contained the randomized treatment (rivaroxaban vs dose-adjusted warfarin). Three types of models were generated: comparison of non-DHP CCB use with no use; comparison of rivaroxaban with warfarin within groups defined by non-DHP CCB use; and comparison of rivaroxaban with warfarin within groups defined by baseline renal function (creatinine clearance of 30 to 49, 50 to 80, and >80 ml/min) among patients taking a non-DHP CCB. The latter 2 model types included terms for the interaction of group with randomized treatment. HRs with 95% CIs and p values are presented. For models that include interaction terms, subgroup HRs are presented only where the interaction is significant (p <0.05). Results Of the 14,264 randomized patients in ROCKET AF, 1,308 (9.2%) were taking a non-DHP CCB: 800 (61.1%)
590
The American Journal of Cardiology (www.ajconline.org)
received diltiazem, 506 (38.7%) received verapamil, and 2 (0.2%) received gallopamil. The median (25th, 75th) daily dose(s) of the non-DHP CCBs were 180 mg (120, 240) for diltiazem and 180 mg (120, 240) for verapamil. The 2 patients taking gallopamil were receiving daily doses of 50 and 100 mg, respectively. The median duration of non-DHP CCB exposure during the study was 21.4 (14.5, 28.2) months. Baseline characteristics of the patients according to nonDHP CCB use at randomization are listed in Table 1. NonDHP CCB use was balanced between randomized treatment assignments: 677 patients assigned to rivaroxaban and 631 assigned to warfarin. Patients receiving non-DHP CCBs were more often women, had a longer median time since AF diagnosis, and a slightly lower mean CHADS2 score. Several co-morbid medical conditions were more frequent in the non-DHP CCB group, including diabetes and COPD. However, previous stroke, transient ischemic attack, or nonCNS SE, previous myocardial infarction, and heart failure were less frequent in those receiving non-DHP CCBs. The event rates and adjusted HRs for the efficacy and safety outcomes according to non-DHP CCB use are listed in Table 2. The use of a non-DHP CCB was not associated with a significant increase in the risk for the primary efficacy outcome of stroke or non-CNS SE (adjusted HR 1.39, 95% CI 0.93 to 2.09, p ¼ 0.11) or the primary safety composite outcome of major clinically relevant or NMCR bleeding (adjusted HR 1.17, 95% CI 0.98 to 1.39, p ¼ 0.087). NonDHP CCB use was associated with an increase in the risk for major bleeding (adjusted HR 1.50, 95% CI 1.11e2.04, p ¼ 0.0091) and intracranial hemorrhage (adjusted HR 2.84, 95% CI 1.53e5.29, p ¼ 0.001). The occurrence of other clinical outcomes including all-cause death, cardiovascular death, myocardial infarction, and all-cause hospitalization was not significantly different in patients receiving versus those not receiving a non-DHP CCB. The associations of non-DHP CCB use on the treatment effects of rivaroxaban versus warfarin are illustrated in Figure 1. No significant interactions between non-DHP CCB use and randomized treatment effects were observed, including the primary efficacy outcome of stroke or nonCNS SE (interaction p ¼ 0.38), all-cause death (interaction p ¼ 0.41), and stroke (interaction p ¼ 0.50). Similarly, no significant interactions were observed between non-DHP CCB use and the treatment effects of rivaroxaban versus warfarin for the primary safety composite outcome of major clinically relevant or NMCR bleeding (interaction p ¼ 0.14) as well as major bleeding (interaction p ¼ 0.99) and intracranial hemorrhage (interaction p ¼ 0.51). An additional analysis was conducted to assess the association of non-DHP CCB use across subgroups of renal function for the primary safety composite of major or NMCR bleeding events. The unadjusted rates of bleeding events in patients taking a non-DHP CCB according to treatment assignment across renal function subgroups are listed in Table 3. Renal function classifications were no chronic kidney disease (creatinine clearance >80 ml/min), mild chronic kidney disease (50 to 80 ml/min), and moderate chronic kidney disease (30 to 49 ml/min). In patients taking a non-DHP CCB, no difference was observed in the treatment effect of rivaroxaban versus
warfarin across renal function subgroups (p value for interaction ¼ 0.76). Discussion In this analysis of ROCKET AF, we report the outcomes associated with non-DHP CCB use. There are 3 major findings in this analysis. First, overall non-DHP CCB use was not associated with an increased risk of stroke or nonCNS SE or the composite of major clinically relevant or NMCR bleeding. However, an association was observed between non-DHP CCB use and an increased risk for major bleeding and intracranial hemorrhage. Second, there was no evidence of a differential treatment effect (safety or efficacy) in patients receiving non-DHP CCBs compared with the overall trial population. Finally, we did not observe any evidence of differential efficacy or safety of rivaroxaban versus warfarin in patients with non-DHP CCB use according to renal function. Non-DHP CCB use was observed in approximately 10% of patients in ROCKET AF. Consistent with guideline recommendations,6 patients taking a non-DHP CCB were more likely to have COPD and less likely to have heart failure. Overall, the use of non-DHP CCBs was associated with an increase in the risk for major bleeding. However, no significant differences were observed in the trial’s primary efficacy or safety outcomes between rivaroxaban and warfarin with the use of non-DHP CCBs, and no renal function interaction was observed on the treatment effect of rivaroxaban versus warfarin among those on non-DHP CCB therapy. Non-DHP L-type CCBs such as diltiazem and verapamil exhibit inhibition of both CYP3A4 and P-gp. The US Food and Drug Administration classifies agents including verapamil, diltiazem, erythromycin, and dronedarone as combined P-gp and moderate CYP3A4 inhibitors.7 To date, no pharmacokinetic or pharmacodynamics data have been published to assess the impact of non-DHP CCBs on rivaroxaban drug exposure. However, pharmacokinetic studies have been conducted with the combined P-gp and moderate CYP3A4 inhibitor erythromycin.8,9 These studies were designed to assess the combined impact of erythromycin and renal insufficiency on rivaroxaban exposure. Results showed that rivaroxaban exposure was increased in patients with renal insufficiency on concomitant erythromycin compared with those with normal renal function who were not receiving erythromycin. Although the use of a non-DHP CCB was associated with an increased risk of major bleeding events including intracranial hemorrhage, we did not observe a significant difference in the treatment effect of rivaroxaban versus warfarin in non-DHP CCB users versus nonusers. Based on this lack of impact on the treatment effect of rivaroxaban versus warfarin, factors other than an isolated increase in rivaroxaban exposure are likely to have contributed to the observed increased major bleeding with non-DHP CCB use. Most notably, non-DHP CCB users versus nonusers were more likely to have a history of COPD (21% vs 9%; p <0.0001). A history of COPD has been identified as an independent predictor of major bleeding events in ROCKET AF.10 An analysis from the Apixaban for Reduction in
Arrhythmias and Conduction Disturbances/Non-DHP CCB Use in ROCKET AF
591
Table 1 Baseline characteristics for patients classified by non-DHP CCB use Variable Randomized to rivaroxaban Age (years) Women Race White Black Asian Other Type of Atrial Fibrillation Persistent Paroxysmal New onset Time since Atrial Fibrillation diagnosis, years CHADS2 score, mean (SD) CHADS2 score 1 2 3 4 5 6 At Presentation Body Mass Index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate (beats/min) Creatinine clearance (mL/min) Atrial Fibrillation on baseline electrocardiogram New York Heart Association class I II III IV Ejection fraction <40% Baseline comorbidities Prior stroke, TIA, or non-CNS embolism Hypertension Diabetes Coronary artery disease Prior myocardial infarction Congestive heart failure Valvular disease Carotid occlusive disease Peripheral arterial disease Chronic Obstructive Pulmonary Disease Liver disease Sleep apnea Medications Prior VKA use Prior chronic ASA use ACE inhibitor/ARB at baseline Beta blocker at baseline Digitalis at baseline Diuretic at baseline Class I or III antiarrhythmic
Non-DHP CCB Used (N¼1308) 677 (52%) 74 (67, 79) 595 (45%)
Non-DHP CCB Not Used (N¼12,956) 6454 (50%) 73 (65, 78) 5065 (39%)
1110 10 162 26
(85%) (1%) (12%) (2%)
10,769 170 1624 393
(83%) (1%) (13%) (3%)
1068 223 17 4.1 3.3
(82%) (17%) (1%) (1.3, 8.2) (1.0)
10,480 2291 185 3.1 3.5
(81%) (18%) (1%) (0.9, 7.0) (0.9)
P 0.18 <.0001 <.0001 0.053
0.78
1 251 549 339 138 30 28.4 130 80 77 67.0 1193
(<1%) (19%) (42%) (26%) (11%) (2%) (25.2, 32.9) (120, 140) (70, 84) (68, 86) (52.0, 88.0) (92%)
2 1608 5667 3752 1675 252 28.1 130 80 76 67.0 11,637
<.0001 <.0001 <.0001
(<1%) (12%) (44%) (29%) (13%) (2%) (25.1, 31.9) (120, 140) (70, 86) (67, 86) (52.0, 87.0) (90%)
0.014 0.041 <.0001 0.016 0.59 0.15 <.0001
814 340 143 10 103
(62%) (26%) (11%) (1%) (10%)
5734 4694 2400 126 2422
(44%) (36%) (19%) (1%) (24%)
<.0001
666 1183 565 298 188 644 234 58 67 270 55 118
(51%) (90%) (43%) (23%) (14%) (49%) (18%) (4%) (5%) (21%) (4%) (9%)
7145 11,727 5130 3063 2280 8264 1827 535 772 1227 692 527
(55%) (91%) (40%) (24%) (18%) (64%) (14%) (4%) (6%) (9%) (5%) (4%)
0.0034 0.93 0.011 0.49 0.0033 <.0001 0.0002 0.60 0.22 <.0001 0.079 <.0001
980 405 830 417 567 771 99
(75%) (31%) (63%) (32%) (43%) (59%) (8%)
7924 4800 9753 8833 4901 7719 1582
(61%) (37%) (75%) (68%) (38%) (60%) (12%)
<.0001 <.0001 <.0001 <.0001 <.0001 0.66 0.079
Continuous variables are shown as median (25th, 75th percentiles), except where noted; categorical variables shown as n (%). ACE ¼ angiotensin-converting enzyme; AF ¼ atrial fibrillation; ARB ¼ angiotensin receptor blocker; ASA ¼ acetylsalicylic acid; BMI ¼ body mass index; CAD ¼ coronary artery disease; CNS ¼ central nervous system; COPD ¼ chronic obstructive pulmonary disease; ECG ¼ electrocardiogram; non-DHP CCB ¼ nondihydropyridine calcium channel blocker; NYHA ¼ New York Heart Association; SD ¼ standard deviation; TIA ¼ transient ischemic attack; VKA ¼ vitamin K antagonist.
592
The American Journal of Cardiology (www.ajconline.org)
Table 2 Efficacy and safety outcomes according to non-DHP CCB use Outcomes
Efficacy Outcomes N Stroke/non-CNS embolism All-cause death Cardiovascular death Stroke Myocardial infarction Hospitalization Safety Outcomes N Major/NMCR bleeding Major bleeding ICH
Non-DHP CCB Used Events/100 pt-yrs (total events)
Non-DHP CCB Not Used Events/100 pt-yrs (total events)
Non-DHP CCB Used vs. Not Used
1301 2.96 (54) 5.30 (107) 2.73 (64) 2.88 (51) 0.57 (17) 12.81 (315)
12,870 2.25 (508) 4.68 (1065) 3.00 (689) 2.11 (475) 1.09 (249) 10.24 (2345)
1.39 1.03 0.91 1.45 0.61 1.15
1303 18.96 (321) 5.85 (100) 1.65 (23)
12,933 14.58 (2537) 3.42 (659) 0.57 (113)
1.17 (0.98, 1.39) 1.50 (1.11, 2.04) 2.84 (1.53, 5.29)
Adjusted HR (CI)
(0.93, (0.74, (0.60, (0.96, (0.29, (0.95,
2.09) 1.44) 1.37) 2.20) 1.25) 1.40)
P
0.11 0.87 0.65 0.078 0.17 0.15
0.087 0.0091 0.0010
Hazard ratios and p values are from Cox models that use inverse propensity weighting to account for differences between patients on and off a non-DHP CCB and for different probabilities of discontinuation. All models include other established predictors of endpoints. Event rates incorporate weighting but are not adjusted for covariate. CI ¼ confidence interval; CNS ¼ central nervous system; HR ¼ hazard ratio; ICH ¼ intracranial hemorrhage; NMCR ¼ nonmajor clinically relevant; nonDHP CCB ¼ nondihydropyridine calcium channel blocker.
Figure 1. Treatment comparisons for efficacy and safety outcomes within CCB used and not used group. ICH ¼ intracranial hemorrhage; MI ¼ myocardial infarction.
Stroke and Other Thromboembolic Events in Atrial Fibrillation trial found COPD to be independently associated with an increased risk of cardiovascular and noncardiovascular
mortality among patients with atrial fibrillation.11 Although a history of COPD was a covariate in the bleeding model in the present analysis, it is possible that the presence of COPD
Arrhythmias and Conduction Disturbances/Non-DHP CCB Use in ROCKET AF
593
Table 3 Composite of major or NMCR bleeding events in patients receiving non-DHP CCB classified by baseline renal function* Renal Function
Rivaroxaban
Warfarin
N
Events/100 pt-yrs (total events)
N
Events/100 pt-yrs (total events)
224 307 139
20.27 (54) 17.63 (87) 33.18 (34)
205 285 140
12.59 (43) 14.75 (64) 28.14 (38)
P-value for interaction of CrCl and treatment
0.76 CrCl > 80 mL/min CrCl 50-80 mL/min CrCl 30-49 mL/min
Event rates incorporate weighting but are not adjusted for covariate. CrCl ¼ creatinine clearance. * Patients missing baseline creatinine clearance (CrCl) data (n ¼ 3) or who did not receive study drug (n ¼ 5) are omitted from the analysis in this table.
may be a marker of co-morbid medical illness that we were not able to completely account for. In patients taking non-DHP CCBs, the presence of mild or moderate renal insufficiency was not associated with an increased risk of the composite outcome of major clinically relevant or NMCR bleeding among patients assigned to rivaroxaban. Although the current US prescribing information for rivaroxaban cautions that “rivaroxaban should not be used in patients with creatinine clearance 15 to <80 ml/ min who are receiving concomitant combined P-gp and moderate CYP3A4 inhibitors unless the potential benefit justifies the potential risk,” we did not observe an interaction with renal function for major clinically relevant or NMCR bleeding among those receiving a non-DHP CCB.12 Although pharmacokinetic studies and simulations have value, these findings highlight the limitations of pharmacokinetic data in predicting clinical outcomes and reinforce the primacy of large, randomized clinical trials. This analysis was observational, and the use of nonDHP CCBs was not randomized. We adjusted for baseline differences between non-DHP CCB users and nonusers with inverse probability weighting approaches and multivariable modeling. Nonetheless, our results may be subject to residual confounding. Moreover, the non-DHP CCB doses were relatively low and therefore may not be generalizable to patients taking the highest doses of nonDHP CCBs. Although a minority of patients enrolled in ROCKET AF were treated with non-DHP CCB therapy, the cohort was sizable at 1,308 patients and represents the largest clinical experience on the concomitant use of nonDHP CCBs and rivaroxaban in an atrial fibrillation population. Acknowledgment: Drs. Washam and Patel had full access to all the data in the study; take responsibility for the integrity of the data and the accuracy of the data analysis; and drafted the manuscript. All authors made substantial contributions to the design of the work and the acquisition, analysis, or interpretation of data; reviewed and revised the work; gave final approval of the version to be submitted; and agree to be accountable for all aspects of the work. The authors certify that the manuscript represents valid work and that neither this one nor one with substantially similar content under their authorship has been published or is being considered for publication elsewhere.
Elizabeth Cook who is an employee of the Duke Clinical Research Institute provided editorial support. Disclosures Dr. Piccini received research grant from Janssen Pharmaceuticals, ARCA Biopharma, Boston Scientific, GE Healthcare, and Johnson & Johnson/Janssen Scientific Affairs and consulting fees from Bristol-Myers Squibb/Pfizer and Johnson & Johnson, Forest Laboratories, Janssen Scientific Affairs, Spectranetics, and Medtronic. Dr. Berkowitz is an employee of Bayer HealthCare Pharmaceuticals. Dr. Nessel is an employee of Janssen Research & Development. Dr. Becker has served as a consultant/advisory board member of Janssen, Portola, Cook, Boehringer Ingelheim. Dr. Breithardt received institutional research grant from BMS/Pfizer, Sanofi Aventis, St. Jude; is a speaker’s bureau of Bayer, BMS/Pfizer, Daiichi Sankyo; has served as a consultant/advisory board member of Bayer, BMS/Pfizer. Dr. Fox received institutional research grant from Bayer, Janssen, AstraZeneca; honoraria from Bayer, AstraZeneca, GlaxoSmithKline, Janssen, Sanofi; has served as a consultant/advisory board member of Bayer, Lilly, AstraZeneca, Sanofi. Dr. Halperin has served as a consultant/advisory board member of Bayer AG HealthCare, Boehringer Ingelheim, Daiichi Sankyo, Johnson & Johnson, OrthoMcNeil-Janssen Pharmaceuticals, Pfizer, Sanofi Aventis, Biotronik, Boston Scientific, Janssen, and Medtronic. Dr. Hankey is a speaker’s bureau of Bayer and as a consultant/ advisory board member of Bayer, Sanofi; Others: AC Immune (chair, Data Monitoring Committee). Mahaffey received institutional research grant from Amgen, Daiichi, Johnson & Johnson, Medtronic, St. Jude, and Tenax; has served as a consultant for American College of Cardiology, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Cubist, Eli Lilly, Elsevier, Epson, Forest, GlaxoSmithKline, Johnson & Johnson, Medtronic, Merck, Mt. Sinai, Myokardia, Omthera, Portola, Purdue Pharma, Spring Publishing, The Medicines Company, Vindico, and WebMD; and received equity from BioPrint Fitness. Dr. Singer received institutional research grant from Johnson & Johnson, Bristol-Myers Squibb, Boehringer Ingelheim, Medtronic and has served as a consultant/advisory board member of Boehringer Ingelheim, Bristol-Myers Squibb, CVS Health, Johnson & Johnson, Merck, Pfizer, and St. Jude Medical. Dr. Patel received institutional research grant
594
The American Journal of Cardiology (www.ajconline.org)
from AstraZeneca, CSL, HeartFlow, Janssen Research & Development, Johnson & Johnson, Maquet, Medtronic, and NHLBI and has served as a consultant for AstraZeneca, Bayer, CSL, Genzyme, Janssen Research & Development, Medtronic, and Merck. 1. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, Breithardt G, Halperin JL, Hankey GJ, Piccini JP, Becker RC, Nessel CC, Paolini JF, Berkowitz SD, Fox KA, Califf RM. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011;365:883e891. 2. 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;159:340e347. 3. Schulman S, Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 2005;3:692e694. 4. Zhang M, Tsiatis AA, Davidian M, Pieper KS, Mahaffey KW. Inference on treatment effects from a randomized clinical trial in the presence of premature treatment discontinuation: the SYNERGY trial. Biostatistics 2011;12:258e269. 5. Wei LJ, Lin DY, Weissfeld L. Regression analysis of multivariate incomplete failure time data by modeling marginal distributions. J Am Stat Assoc 1989;84:1065e1073. 6. January CT, Wann LS, Alpert JS, Calkins H, Cigarroa JE, Cleveland JC Jr, Conti JB, Ellinor PT, Ezekowitz MD, Field ME, Murray KT, Sacco RL, Stevenson WG, Tchou PJ, Tracy CM, Yancy CW. 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. J Am Coll Cardiol 2014;64:e1ee76.
7. Food and Drug Administration web site. Available at: http://www.fda. gov/Drugs/DevelopmentApprovalProcess/DevelopmentResources/Drug InteractionsLabeling/ucm093664.htm#potency. Accessed on September 15, 2015. 8. Grillo JA, Zhao P, Bullock J, Booth BP, Lu M, Robie-Suh K, Berglund EG, Pang KS, Rahman A, Zhang L, Lesko LJ, Huang SM. Utility of a physiologically-based pharmacokinetic (PBPK) modeling approach to quantitatively predict a complex drug-drug-disease interaction scenario for rivaroxaban during the drug review process: implications for clinical practice. Biopharm Drug Dispos 2012;33: 99e110. 9. Moore KT, Vaidyanathan S, Natarajan J, Ariyawansa J, Haskell L, Turner KC. An open-label study to estimate the effect of steady-state erythromycin on the pharmacokinetics, pharmacodynamics, and safety of a single dose of rivaroxaban in subjects with renal impairment and normal renal function. J Clin Pharmacol 2014;54:1407e1420. 10. Goodman SG, Wojdyla DM, Piccini JP, White HD, Paolini JF, Nessel CC, Berkowitz SD, Mahaffey KW, Patel MR, Sherwood MW, Becker RC, Halperin JL, Hacke W, Singer DE, Hankey GJ, Breithardt G, Fox KA, Califf RM. Factors associated with major bleeding events: insights from the ROCKET AF trial (rivaroxaban once-daily oral direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and embolism trial in atrial fibrillation). J Am Coll Cardiol 2014;63:891e900. 11. Durheim MT, Cyr DD, Lopes RD, Thomas LE, Tsuang WM, Gersh BJ, Held C, Wallentin L, Granger CB, Palmer SM, AlKhatib SM. Chronic obstructive pulmonary disease in patients with atrial fibrillation: insights from the ARISTOTLE trial. Int J Cardiol 2016;202:589e594. 12. Fuhrmann JT, Schmeisser A, Schulze MR, Wunderlich C, Schoen SP, Rauwolf T, Weinbrenner C, Strasser RH. Levosimendan is superior to enoximone in refractory cardiogenic shock complicating acute myocardial infarction. Crit Care Med 2008;36:2257e2266.
a Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina; bBayer HealthCare Pharmaceuticals, Parsippany, New Jersey; cJanssen Research & Development, Raritan, New Jersey; d University of Cincinnati College of Medicine, Cincinnati, Ohio; eUniversity Hospital of Münster, Münster, Germany; fUniversity of Edinburgh, Royal Infirmary of Edinburgh, Edinburgh, United Kingdom; gMount Sinai School of Medicine, New York, New York; hSchool of Medicine and Pharmacology, University of Western Australia, Crawley, Australia; i Stanford Center for Clinical Research, Stanford University School of Medicine, Stanford, California; and jMassachusetts General Hospital, Harvard Medical School, Boston, Massachusetts. Manuscript received February 16, 2017; revised manuscript received and accepted May 5, 2017. The ROCKET AF trial was supported by Janssen Research & Development and Bayer HealthCare Pharmaceuticals. See page 593 for disclosure information. *Corresponding author: Tel: (919) 681-3286; fax: (919) 681-2960. E-mail address: [email protected] (J.B. Washam).
0002-9149/17/$ - see front matter Ó 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2017.05.026
hepatic cytochrome P450 3A4 (CYP3A4) enzyme inhibition. Rivaroxaban is an orally administered direct-acting factor Xa inhibitor that was compared with warfarin for the prevention of stroke or noncentral nervous system (CNS) systemic embolism (SE) in patients with nonvalvular atrial fibrillation in the Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation study (ROCKET AF).1 Rivaroxaban is a substrate for both the P-gp transport and CYP3A4 isoenzyme pathways. Therefore, there is a theoretical concern that concomitant use of rivaroxaban with diltiazem or verapamil could lead to an increase in rivaroxaban drug exposure and anticoagulant effect. To date, no clinical data have been reported on the association of concomitant use of these therapies on clinical outcomes in patients treated with rivaroxaban. The purpose of this analysis was to assess clinical outcomes in patients with atrial fibrillation on oral anticoagulation enrolled in ROCKET AF trial who received non-DHP CCB therapy and to assess the impact of this concomitant therapy on the www.ajconline.org
Arrhythmias and Conduction Disturbances/Non-DHP CCB Use in ROCKET AF
efficacy and safety of rivaroxaban compared with warfarin. Methods The design of the ROCKET AF study has previously been described.2 ROCKET AF was a double-blind, doubledummy trial that randomized 14,264 patients with atrial fibrillation to rivaroxaban (20 mg once daily or 15 mg once daily if creatinine clearance was 30 to 49 ml/min) or adjusted dose warfarin (target international normalized ratio 2.0 to 3.0). The primary efficacy end point was stroke or non-CNS SE and the primary safety end point was the composite of nonmajor clinically relevant (NMCR) and major bleeding.3 Major bleeding was defined as clinically overt bleeding that was associated with one or more of the following: (1) fall in hemoglobin of 2 g/dl or more; (2) a transfusion of 2 or more units of packed red blood cells or whole blood; (3) bleeding at a critical site (intracranial, intraaspinal, intraocular, pericardial, intraarticular, intramuscular with compartment syndrome, and retroperitoneal); or (4) a fatal outcome. NMCR bleeding was defined as overt bleeding not meeting the criteria for major bleeding but associated with medical intervention, unscheduled medical contact, cessation of study treatment, or associated with discomfort for the patient such as pain or impairment of activities of daily life. The intention-to-treat population was used for all efficacy analyses and safety analyses were based on the safety population (randomized patients who received at least 1 dose of study medication). Ninety-three patients from 1 enrolling site were excluded from efficacy analyses due to violations of Good Clinical Practice Guidelines. The overall clinical outcomes from ROCKET AF have previously been reported.1 The present study is a post hoc analysis of ROCKET AF including all randomized patients and grouping patients according to baseline and postrandomization use of a nonDHP CCB. Non-DHP CCB therapy was at the discretion of the treating physician and was captured as a concomitant medication on the case report form at baseline and during study follow-up (every 4 weeks until study completion). An independent clinical end point committee, unaware of treatment assignment, adjudicated clinical end points including all suspected stroke and SE, myocardial infarction, death, and bleeding events. Death events were subclassified into cardiovascular death (including sudden death) and noncardiovascular death. Summary statistics are presented for patterns of nonDHP CCB use, including the number of patients, agent, dose, and exposure time. Patients who took more than 1 non-DHP CCB were counted for the agent taken for the largest proportion of time; in instances when more than 1 dose is reported, the last dose was used. Baseline characteristics are presented for each group with categorical variables presented as counts (percentages) and comparisons were made using Pearson chi-square tests. Continuous variables are presented as medians (25th, 75th percentiles) and comparisons were made using Wilcoxon rank-sum tests. Cox proportional hazards models were used to assess the relation of non-DHP CCB use with subsequent outcomes,
589
with patients classified according to their baseline use; patients were included in models for as long as they remained in their baseline group. Patients who were taking non-DHP CCBs at baseline and subsequently stopped, or patients who were not taking one at baseline and subsequently started, were censored at those time points. Because either new start or cessation of non-DHP CCBs can be influenced by patient characteristics or intervening events that might also be related to the outcomes, patients were weighted by the inverse probability of continuing in their therapy group. Weights were applied to the Cox models with a robust sandwich variance estimator.4 Unadjusted summary event rates (events per 100 patient-years) were also weighted. For the recurrent event of hospitalizations, we used the method of Wei et al5 for multiple failure times with a robust sandwich variance estimator. These models incorporated the weighting described previously. Because only a small number of patients had multiple hospitalizations, these models included the first and second hospitalizations only. A single weighted parameter estimate was used to generate a significance test (z score) as well as a hazard ratio (HR) and CI. Event rates (events per 100 patient-years) and total number of events are presented for all clinical outcomes. Group comparisons were made using Cox models, adjusted for previously identified predictors of each end point. Efficacy outcomes models included the following covariates: age, gender, body mass index, region, diabetes, previous stroke/transient ischemic attack, vascular disease, congestive heart failure, hypertension, chronic obstructive pulmonary disease (COPD), paroxysmal atrial fibrillation, diastolic blood pressure, creatinine clearance (calculated using the Cockroft-Gault equation), heart rate, and abstinence from alcohol. Safety outcomes models included the following covariates: age, gender, region, previous stroke/transient ischemic attack, anemia, previous gastrointestinal bleed, COPD, diastolic blood pressure, creatinine clearance, platelets, albumin, previous aspirin use, previous vitamin K antagonist use, and previous thienopyridine use. Rates of missing data were quite low; when missing, covariates were imputed using the median for continuous variables and the mode for categorical variables within groups of patients on or not on a non-DHP CCB at baseline. All models contained the randomized treatment (rivaroxaban vs dose-adjusted warfarin). Three types of models were generated: comparison of non-DHP CCB use with no use; comparison of rivaroxaban with warfarin within groups defined by non-DHP CCB use; and comparison of rivaroxaban with warfarin within groups defined by baseline renal function (creatinine clearance of 30 to 49, 50 to 80, and >80 ml/min) among patients taking a non-DHP CCB. The latter 2 model types included terms for the interaction of group with randomized treatment. HRs with 95% CIs and p values are presented. For models that include interaction terms, subgroup HRs are presented only where the interaction is significant (p <0.05). Results Of the 14,264 randomized patients in ROCKET AF, 1,308 (9.2%) were taking a non-DHP CCB: 800 (61.1%)
590
The American Journal of Cardiology (www.ajconline.org)
received diltiazem, 506 (38.7%) received verapamil, and 2 (0.2%) received gallopamil. The median (25th, 75th) daily dose(s) of the non-DHP CCBs were 180 mg (120, 240) for diltiazem and 180 mg (120, 240) for verapamil. The 2 patients taking gallopamil were receiving daily doses of 50 and 100 mg, respectively. The median duration of non-DHP CCB exposure during the study was 21.4 (14.5, 28.2) months. Baseline characteristics of the patients according to nonDHP CCB use at randomization are listed in Table 1. NonDHP CCB use was balanced between randomized treatment assignments: 677 patients assigned to rivaroxaban and 631 assigned to warfarin. Patients receiving non-DHP CCBs were more often women, had a longer median time since AF diagnosis, and a slightly lower mean CHADS2 score. Several co-morbid medical conditions were more frequent in the non-DHP CCB group, including diabetes and COPD. However, previous stroke, transient ischemic attack, or nonCNS SE, previous myocardial infarction, and heart failure were less frequent in those receiving non-DHP CCBs. The event rates and adjusted HRs for the efficacy and safety outcomes according to non-DHP CCB use are listed in Table 2. The use of a non-DHP CCB was not associated with a significant increase in the risk for the primary efficacy outcome of stroke or non-CNS SE (adjusted HR 1.39, 95% CI 0.93 to 2.09, p ¼ 0.11) or the primary safety composite outcome of major clinically relevant or NMCR bleeding (adjusted HR 1.17, 95% CI 0.98 to 1.39, p ¼ 0.087). NonDHP CCB use was associated with an increase in the risk for major bleeding (adjusted HR 1.50, 95% CI 1.11e2.04, p ¼ 0.0091) and intracranial hemorrhage (adjusted HR 2.84, 95% CI 1.53e5.29, p ¼ 0.001). The occurrence of other clinical outcomes including all-cause death, cardiovascular death, myocardial infarction, and all-cause hospitalization was not significantly different in patients receiving versus those not receiving a non-DHP CCB. The associations of non-DHP CCB use on the treatment effects of rivaroxaban versus warfarin are illustrated in Figure 1. No significant interactions between non-DHP CCB use and randomized treatment effects were observed, including the primary efficacy outcome of stroke or nonCNS SE (interaction p ¼ 0.38), all-cause death (interaction p ¼ 0.41), and stroke (interaction p ¼ 0.50). Similarly, no significant interactions were observed between non-DHP CCB use and the treatment effects of rivaroxaban versus warfarin for the primary safety composite outcome of major clinically relevant or NMCR bleeding (interaction p ¼ 0.14) as well as major bleeding (interaction p ¼ 0.99) and intracranial hemorrhage (interaction p ¼ 0.51). An additional analysis was conducted to assess the association of non-DHP CCB use across subgroups of renal function for the primary safety composite of major or NMCR bleeding events. The unadjusted rates of bleeding events in patients taking a non-DHP CCB according to treatment assignment across renal function subgroups are listed in Table 3. Renal function classifications were no chronic kidney disease (creatinine clearance >80 ml/min), mild chronic kidney disease (50 to 80 ml/min), and moderate chronic kidney disease (30 to 49 ml/min). In patients taking a non-DHP CCB, no difference was observed in the treatment effect of rivaroxaban versus
warfarin across renal function subgroups (p value for interaction ¼ 0.76). Discussion In this analysis of ROCKET AF, we report the outcomes associated with non-DHP CCB use. There are 3 major findings in this analysis. First, overall non-DHP CCB use was not associated with an increased risk of stroke or nonCNS SE or the composite of major clinically relevant or NMCR bleeding. However, an association was observed between non-DHP CCB use and an increased risk for major bleeding and intracranial hemorrhage. Second, there was no evidence of a differential treatment effect (safety or efficacy) in patients receiving non-DHP CCBs compared with the overall trial population. Finally, we did not observe any evidence of differential efficacy or safety of rivaroxaban versus warfarin in patients with non-DHP CCB use according to renal function. Non-DHP CCB use was observed in approximately 10% of patients in ROCKET AF. Consistent with guideline recommendations,6 patients taking a non-DHP CCB were more likely to have COPD and less likely to have heart failure. Overall, the use of non-DHP CCBs was associated with an increase in the risk for major bleeding. However, no significant differences were observed in the trial’s primary efficacy or safety outcomes between rivaroxaban and warfarin with the use of non-DHP CCBs, and no renal function interaction was observed on the treatment effect of rivaroxaban versus warfarin among those on non-DHP CCB therapy. Non-DHP L-type CCBs such as diltiazem and verapamil exhibit inhibition of both CYP3A4 and P-gp. The US Food and Drug Administration classifies agents including verapamil, diltiazem, erythromycin, and dronedarone as combined P-gp and moderate CYP3A4 inhibitors.7 To date, no pharmacokinetic or pharmacodynamics data have been published to assess the impact of non-DHP CCBs on rivaroxaban drug exposure. However, pharmacokinetic studies have been conducted with the combined P-gp and moderate CYP3A4 inhibitor erythromycin.8,9 These studies were designed to assess the combined impact of erythromycin and renal insufficiency on rivaroxaban exposure. Results showed that rivaroxaban exposure was increased in patients with renal insufficiency on concomitant erythromycin compared with those with normal renal function who were not receiving erythromycin. Although the use of a non-DHP CCB was associated with an increased risk of major bleeding events including intracranial hemorrhage, we did not observe a significant difference in the treatment effect of rivaroxaban versus warfarin in non-DHP CCB users versus nonusers. Based on this lack of impact on the treatment effect of rivaroxaban versus warfarin, factors other than an isolated increase in rivaroxaban exposure are likely to have contributed to the observed increased major bleeding with non-DHP CCB use. Most notably, non-DHP CCB users versus nonusers were more likely to have a history of COPD (21% vs 9%; p <0.0001). A history of COPD has been identified as an independent predictor of major bleeding events in ROCKET AF.10 An analysis from the Apixaban for Reduction in
Arrhythmias and Conduction Disturbances/Non-DHP CCB Use in ROCKET AF
591
Table 1 Baseline characteristics for patients classified by non-DHP CCB use Variable Randomized to rivaroxaban Age (years) Women Race White Black Asian Other Type of Atrial Fibrillation Persistent Paroxysmal New onset Time since Atrial Fibrillation diagnosis, years CHADS2 score, mean (SD) CHADS2 score 1 2 3 4 5 6 At Presentation Body Mass Index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate (beats/min) Creatinine clearance (mL/min) Atrial Fibrillation on baseline electrocardiogram New York Heart Association class I II III IV Ejection fraction <40% Baseline comorbidities Prior stroke, TIA, or non-CNS embolism Hypertension Diabetes Coronary artery disease Prior myocardial infarction Congestive heart failure Valvular disease Carotid occlusive disease Peripheral arterial disease Chronic Obstructive Pulmonary Disease Liver disease Sleep apnea Medications Prior VKA use Prior chronic ASA use ACE inhibitor/ARB at baseline Beta blocker at baseline Digitalis at baseline Diuretic at baseline Class I or III antiarrhythmic
Non-DHP CCB Used (N¼1308) 677 (52%) 74 (67, 79) 595 (45%)
Non-DHP CCB Not Used (N¼12,956) 6454 (50%) 73 (65, 78) 5065 (39%)
1110 10 162 26
(85%) (1%) (12%) (2%)
10,769 170 1624 393
(83%) (1%) (13%) (3%)
1068 223 17 4.1 3.3
(82%) (17%) (1%) (1.3, 8.2) (1.0)
10,480 2291 185 3.1 3.5
(81%) (18%) (1%) (0.9, 7.0) (0.9)
P 0.18 <.0001 <.0001 0.053
0.78
1 251 549 339 138 30 28.4 130 80 77 67.0 1193
(<1%) (19%) (42%) (26%) (11%) (2%) (25.2, 32.9) (120, 140) (70, 84) (68, 86) (52.0, 88.0) (92%)
2 1608 5667 3752 1675 252 28.1 130 80 76 67.0 11,637
<.0001 <.0001 <.0001
(<1%) (12%) (44%) (29%) (13%) (2%) (25.1, 31.9) (120, 140) (70, 86) (67, 86) (52.0, 87.0) (90%)
0.014 0.041 <.0001 0.016 0.59 0.15 <.0001
814 340 143 10 103
(62%) (26%) (11%) (1%) (10%)
5734 4694 2400 126 2422
(44%) (36%) (19%) (1%) (24%)
<.0001
666 1183 565 298 188 644 234 58 67 270 55 118
(51%) (90%) (43%) (23%) (14%) (49%) (18%) (4%) (5%) (21%) (4%) (9%)
7145 11,727 5130 3063 2280 8264 1827 535 772 1227 692 527
(55%) (91%) (40%) (24%) (18%) (64%) (14%) (4%) (6%) (9%) (5%) (4%)
0.0034 0.93 0.011 0.49 0.0033 <.0001 0.0002 0.60 0.22 <.0001 0.079 <.0001
980 405 830 417 567 771 99
(75%) (31%) (63%) (32%) (43%) (59%) (8%)
7924 4800 9753 8833 4901 7719 1582
(61%) (37%) (75%) (68%) (38%) (60%) (12%)
<.0001 <.0001 <.0001 <.0001 <.0001 0.66 0.079
Continuous variables are shown as median (25th, 75th percentiles), except where noted; categorical variables shown as n (%). ACE ¼ angiotensin-converting enzyme; AF ¼ atrial fibrillation; ARB ¼ angiotensin receptor blocker; ASA ¼ acetylsalicylic acid; BMI ¼ body mass index; CAD ¼ coronary artery disease; CNS ¼ central nervous system; COPD ¼ chronic obstructive pulmonary disease; ECG ¼ electrocardiogram; non-DHP CCB ¼ nondihydropyridine calcium channel blocker; NYHA ¼ New York Heart Association; SD ¼ standard deviation; TIA ¼ transient ischemic attack; VKA ¼ vitamin K antagonist.
592
The American Journal of Cardiology (www.ajconline.org)
Table 2 Efficacy and safety outcomes according to non-DHP CCB use Outcomes
Efficacy Outcomes N Stroke/non-CNS embolism All-cause death Cardiovascular death Stroke Myocardial infarction Hospitalization Safety Outcomes N Major/NMCR bleeding Major bleeding ICH
Non-DHP CCB Used Events/100 pt-yrs (total events)
Non-DHP CCB Not Used Events/100 pt-yrs (total events)
Non-DHP CCB Used vs. Not Used
1301 2.96 (54) 5.30 (107) 2.73 (64) 2.88 (51) 0.57 (17) 12.81 (315)
12,870 2.25 (508) 4.68 (1065) 3.00 (689) 2.11 (475) 1.09 (249) 10.24 (2345)
1.39 1.03 0.91 1.45 0.61 1.15
1303 18.96 (321) 5.85 (100) 1.65 (23)
12,933 14.58 (2537) 3.42 (659) 0.57 (113)
1.17 (0.98, 1.39) 1.50 (1.11, 2.04) 2.84 (1.53, 5.29)
Adjusted HR (CI)
(0.93, (0.74, (0.60, (0.96, (0.29, (0.95,
2.09) 1.44) 1.37) 2.20) 1.25) 1.40)
P
0.11 0.87 0.65 0.078 0.17 0.15
0.087 0.0091 0.0010
Hazard ratios and p values are from Cox models that use inverse propensity weighting to account for differences between patients on and off a non-DHP CCB and for different probabilities of discontinuation. All models include other established predictors of endpoints. Event rates incorporate weighting but are not adjusted for covariate. CI ¼ confidence interval; CNS ¼ central nervous system; HR ¼ hazard ratio; ICH ¼ intracranial hemorrhage; NMCR ¼ nonmajor clinically relevant; nonDHP CCB ¼ nondihydropyridine calcium channel blocker.
Figure 1. Treatment comparisons for efficacy and safety outcomes within CCB used and not used group. ICH ¼ intracranial hemorrhage; MI ¼ myocardial infarction.
Stroke and Other Thromboembolic Events in Atrial Fibrillation trial found COPD to be independently associated with an increased risk of cardiovascular and noncardiovascular
mortality among patients with atrial fibrillation.11 Although a history of COPD was a covariate in the bleeding model in the present analysis, it is possible that the presence of COPD
Arrhythmias and Conduction Disturbances/Non-DHP CCB Use in ROCKET AF
593
Table 3 Composite of major or NMCR bleeding events in patients receiving non-DHP CCB classified by baseline renal function* Renal Function
Rivaroxaban
Warfarin
N
Events/100 pt-yrs (total events)
N
Events/100 pt-yrs (total events)
224 307 139
20.27 (54) 17.63 (87) 33.18 (34)
205 285 140
12.59 (43) 14.75 (64) 28.14 (38)
P-value for interaction of CrCl and treatment
0.76 CrCl > 80 mL/min CrCl 50-80 mL/min CrCl 30-49 mL/min
Event rates incorporate weighting but are not adjusted for covariate. CrCl ¼ creatinine clearance. * Patients missing baseline creatinine clearance (CrCl) data (n ¼ 3) or who did not receive study drug (n ¼ 5) are omitted from the analysis in this table.
may be a marker of co-morbid medical illness that we were not able to completely account for. In patients taking non-DHP CCBs, the presence of mild or moderate renal insufficiency was not associated with an increased risk of the composite outcome of major clinically relevant or NMCR bleeding among patients assigned to rivaroxaban. Although the current US prescribing information for rivaroxaban cautions that “rivaroxaban should not be used in patients with creatinine clearance 15 to <80 ml/ min who are receiving concomitant combined P-gp and moderate CYP3A4 inhibitors unless the potential benefit justifies the potential risk,” we did not observe an interaction with renal function for major clinically relevant or NMCR bleeding among those receiving a non-DHP CCB.12 Although pharmacokinetic studies and simulations have value, these findings highlight the limitations of pharmacokinetic data in predicting clinical outcomes and reinforce the primacy of large, randomized clinical trials. This analysis was observational, and the use of nonDHP CCBs was not randomized. We adjusted for baseline differences between non-DHP CCB users and nonusers with inverse probability weighting approaches and multivariable modeling. Nonetheless, our results may be subject to residual confounding. Moreover, the non-DHP CCB doses were relatively low and therefore may not be generalizable to patients taking the highest doses of nonDHP CCBs. Although a minority of patients enrolled in ROCKET AF were treated with non-DHP CCB therapy, the cohort was sizable at 1,308 patients and represents the largest clinical experience on the concomitant use of nonDHP CCBs and rivaroxaban in an atrial fibrillation population. Acknowledgment: Drs. Washam and Patel had full access to all the data in the study; take responsibility for the integrity of the data and the accuracy of the data analysis; and drafted the manuscript. All authors made substantial contributions to the design of the work and the acquisition, analysis, or interpretation of data; reviewed and revised the work; gave final approval of the version to be submitted; and agree to be accountable for all aspects of the work. The authors certify that the manuscript represents valid work and that neither this one nor one with substantially similar content under their authorship has been published or is being considered for publication elsewhere.
Elizabeth Cook who is an employee of the Duke Clinical Research Institute provided editorial support. Disclosures Dr. Piccini received research grant from Janssen Pharmaceuticals, ARCA Biopharma, Boston Scientific, GE Healthcare, and Johnson & Johnson/Janssen Scientific Affairs and consulting fees from Bristol-Myers Squibb/Pfizer and Johnson & Johnson, Forest Laboratories, Janssen Scientific Affairs, Spectranetics, and Medtronic. Dr. Berkowitz is an employee of Bayer HealthCare Pharmaceuticals. Dr. Nessel is an employee of Janssen Research & Development. Dr. Becker has served as a consultant/advisory board member of Janssen, Portola, Cook, Boehringer Ingelheim. Dr. Breithardt received institutional research grant from BMS/Pfizer, Sanofi Aventis, St. Jude; is a speaker’s bureau of Bayer, BMS/Pfizer, Daiichi Sankyo; has served as a consultant/advisory board member of Bayer, BMS/Pfizer. Dr. Fox received institutional research grant from Bayer, Janssen, AstraZeneca; honoraria from Bayer, AstraZeneca, GlaxoSmithKline, Janssen, Sanofi; has served as a consultant/advisory board member of Bayer, Lilly, AstraZeneca, Sanofi. Dr. Halperin has served as a consultant/advisory board member of Bayer AG HealthCare, Boehringer Ingelheim, Daiichi Sankyo, Johnson & Johnson, OrthoMcNeil-Janssen Pharmaceuticals, Pfizer, Sanofi Aventis, Biotronik, Boston Scientific, Janssen, and Medtronic. Dr. Hankey is a speaker’s bureau of Bayer and as a consultant/ advisory board member of Bayer, Sanofi; Others: AC Immune (chair, Data Monitoring Committee). Mahaffey received institutional research grant from Amgen, Daiichi, Johnson & Johnson, Medtronic, St. Jude, and Tenax; has served as a consultant for American College of Cardiology, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Cubist, Eli Lilly, Elsevier, Epson, Forest, GlaxoSmithKline, Johnson & Johnson, Medtronic, Merck, Mt. Sinai, Myokardia, Omthera, Portola, Purdue Pharma, Spring Publishing, The Medicines Company, Vindico, and WebMD; and received equity from BioPrint Fitness. Dr. Singer received institutional research grant from Johnson & Johnson, Bristol-Myers Squibb, Boehringer Ingelheim, Medtronic and has served as a consultant/advisory board member of Boehringer Ingelheim, Bristol-Myers Squibb, CVS Health, Johnson & Johnson, Merck, Pfizer, and St. Jude Medical. Dr. Patel received institutional research grant
594
The American Journal of Cardiology (www.ajconline.org)
from AstraZeneca, CSL, HeartFlow, Janssen Research & Development, Johnson & Johnson, Maquet, Medtronic, and NHLBI and has served as a consultant for AstraZeneca, Bayer, CSL, Genzyme, Janssen Research & Development, Medtronic, and Merck. 1. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, Breithardt G, Halperin JL, Hankey GJ, Piccini JP, Becker RC, Nessel CC, Paolini JF, Berkowitz SD, Fox KA, Califf RM. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011;365:883e891. 2. 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;159:340e347. 3. Schulman S, Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 2005;3:692e694. 4. Zhang M, Tsiatis AA, Davidian M, Pieper KS, Mahaffey KW. Inference on treatment effects from a randomized clinical trial in the presence of premature treatment discontinuation: the SYNERGY trial. Biostatistics 2011;12:258e269. 5. Wei LJ, Lin DY, Weissfeld L. Regression analysis of multivariate incomplete failure time data by modeling marginal distributions. J Am Stat Assoc 1989;84:1065e1073. 6. January CT, Wann LS, Alpert JS, Calkins H, Cigarroa JE, Cleveland JC Jr, Conti JB, Ellinor PT, Ezekowitz MD, Field ME, Murray KT, Sacco RL, Stevenson WG, Tchou PJ, Tracy CM, Yancy CW. 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. J Am Coll Cardiol 2014;64:e1ee76.
7. Food and Drug Administration web site. Available at: http://www.fda. gov/Drugs/DevelopmentApprovalProcess/DevelopmentResources/Drug InteractionsLabeling/ucm093664.htm#potency. Accessed on September 15, 2015. 8. Grillo JA, Zhao P, Bullock J, Booth BP, Lu M, Robie-Suh K, Berglund EG, Pang KS, Rahman A, Zhang L, Lesko LJ, Huang SM. Utility of a physiologically-based pharmacokinetic (PBPK) modeling approach to quantitatively predict a complex drug-drug-disease interaction scenario for rivaroxaban during the drug review process: implications for clinical practice. Biopharm Drug Dispos 2012;33: 99e110. 9. Moore KT, Vaidyanathan S, Natarajan J, Ariyawansa J, Haskell L, Turner KC. An open-label study to estimate the effect of steady-state erythromycin on the pharmacokinetics, pharmacodynamics, and safety of a single dose of rivaroxaban in subjects with renal impairment and normal renal function. J Clin Pharmacol 2014;54:1407e1420. 10. Goodman SG, Wojdyla DM, Piccini JP, White HD, Paolini JF, Nessel CC, Berkowitz SD, Mahaffey KW, Patel MR, Sherwood MW, Becker RC, Halperin JL, Hacke W, Singer DE, Hankey GJ, Breithardt G, Fox KA, Califf RM. Factors associated with major bleeding events: insights from the ROCKET AF trial (rivaroxaban once-daily oral direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and embolism trial in atrial fibrillation). J Am Coll Cardiol 2014;63:891e900. 11. Durheim MT, Cyr DD, Lopes RD, Thomas LE, Tsuang WM, Gersh BJ, Held C, Wallentin L, Granger CB, Palmer SM, AlKhatib SM. Chronic obstructive pulmonary disease in patients with atrial fibrillation: insights from the ARISTOTLE trial. Int J Cardiol 2016;202:589e594. 12. Fuhrmann JT, Schmeisser A, Schulze MR, Wunderlich C, Schoen SP, Rauwolf T, Weinbrenner C, Strasser RH. Levosimendan is superior to enoximone in refractory cardiogenic shock complicating acute myocardial infarction. Crit Care Med 2008;36:2257e2266.