OUP user menu

★ Editor's choice ★

Ischaemic cardiac outcomes in patients with atrial fibrillation treated with vitamin K antagonism or factor Xa inhibition: results from the ROCKET AF trial

Kenneth W. Mahaffey , Susanna R. Stevens , Harvey D. White , Christopher C. Nessel , Shaun G. Goodman , Jonathan P. Piccini , Manesh R. Patel , Richard C. Becker , Jonathan L. Halperin , Werner Hacke , Daniel E. Singer , Graeme J. Hankey , Robert M. Califf , Keith A.A. Fox , Günter Breithardt , for the ROCKET AF Investigators
DOI: http://dx.doi.org/10.1093/eurheartj/eht428 233-241 First published online: 16 October 2013

Abstract

Aims We investigated the prevalence of prior myocardial infarction (MI) and incidence of ischaemic cardiovascular (CV) events among atrial fibrillation (AF) patients.

Methods and results In ROCKET AF, 14 264 patients with nonvalvular AF were randomized to rivaroxaban or warfarin. The key efficacy outcome for these analyses was CV death, MI, and unstable angina (UA). This pre-specified analysis was performed on patients while on treatment. Rates are per 100 patient-years. Overall, 2468 (17%) patients had prior MI at enrollment. Compared with patients without prior MI, these patients were more likely to be male (75 vs. 57%), on aspirin at baseline (47 vs. 34%), have prior congestive heart failure (78 vs. 59%), diabetes (47 vs. 39%), hypertension (94 vs. 90%), higher mean CHADS2 score (3.64 vs. 3.43), and fewer prior strokes or transient ischaemic attacks (46 vs. 54%). CV death, MI, or UA rates tended to be lower in patients assigned rivaroxaban compared with warfarin [2.70 vs. 3.15; hazard ratio (HR) 0.86, 95% confidence interval (CI) 0.73–1.00; P = 0.0509]. CV death, MI, or UA rates were higher in those with prior MI compared with no prior MI (6.68 vs. 2.19; HR 3.04, 95% CI 2.59–3.56) with consistent results for CV death, MI, or UA for rivaroxaban compared with warfarin in prior MI compared with no prior MI (P interaction = 0.10).

Conclusion Prior MI was common and associated with substantial risk for subsequent cardiac events. Patients with prior MI assigned rivaroxaban compared with warfarin had a non-significant 14% reduction of ischaemic cardiac events.

  • Atrial fibrillation
  • Myocardial infarction
  • Coronary artery disease
  • Outcomes
  • Factor Xa
  • Rivaroxaban
  • Warfarin

See page 207 for the editorial comment on this article (doi:10.1093/eurheartj/eht480)

Coronary artery disease (CAD) is common in patients with atrial fibrillation (AF).1 Patients with CAD may develop AF and patients with AF may present with acute coronary syndromes (ACS). Prognostic relevance of AF in patients with acute myocardial infarction (MI) is not certain, with some studies showing an independent increased risk and others not.214

Management of patients with AF, ACS, and chronic CAD includes antiplatelet and anticoagulant therapy.1519 Treatment with multiple antiplatelet and anticoagulant agents may be necessary to prevent subsequent thrombotic events, but dual or triple antiplatelet/anticoagulant therapy increases bleeding risk.2022 The bleeding risk may be enhanced in specific patient populations, including the elderly, patients with diabetes, and those with renal insufficiency.

ROCKET AF (Rivaroxaban Once-daily, oral, direct Factor Xa inhibition Compared with vitamin K antagonism for prevention of stroke and Embolism Trial in Atrial Fibrillation) demonstrated that rivaroxaban, a direct factor Xa inhibitor, was noninferior to warfarin in the prevention of stroke and systemic embolism in high-risk patients.23 In this report, we investigate the prevalence of established CAD, defined as prior MI, the incidence of cardiovascular (CV) events in a large population of patients with AF receiving anticoagulant therapy, and the risk of CV events with rivaroxaban compared with warfarin.

Methods

ROCKET AF was a multicentre, randomized, double-blind, double-dummy, event-driven trial comparing fixed-dose rivaroxaban (20 mg daily; 15 mg daily in patients with creatinine clearance 30–49 mL/min) with adjusted-dose warfarin [target international normalized ratio (INR) 2.0–3.0] for prevention of stroke or systemic embolism, as described previously.23,24 The study was supported by Johnson & Johnson Pharmaceutical Research and Development (Raritan, NJ) and Bayer HealthCare AG (Leverkusen, Germany). An international executive committee designed the study and oversaw the trial conduct. The protocol was approved by national regulatory authorities and ethics committees at participating centres. All patients provided written consent.

Study participants and follow-up

Patients with nonvalvular AF at moderate-to-high risk for stroke were recruited at 1178 participating sites in 45 countries. Patients were included if they had a history of stroke, transient ischaemic attack, or systemic embolism or at least two of the following: heart failure or left ventricular ejection fraction ≤35%, hypertension, age ≥75 years, or diabetes mellitus (CHADS2 score ≥2). Patients with prior MI were not excluded. Prior MI was reported on the standard electronic case report form modules by the site investigator. Patients were followed monthly for assessment of clinical outcomes, adverse events, and monitoring of INR values while on therapy. If the study drug was discontinued before the end of the study, then patients were to be followed every 3 months by telephone for ascertainment of events.

Cardiovascular outcomes

CV outcomes were defined in the study analysis plan as CV death, MI, and unstable angina (UA). The principal safety endpoint was the composite of major and non-major clinically relevant (NMCR) bleeding events. Investigators were to report deaths and all suspected strokes, MIs, and bleeding events, and these were adjudicated by an independent clinical events committee that was blinded to treatment assignment. The endpoint definitions have been published.23,24

Statistical analysis

These prespecified analyses evaluated the CV outcomes and efficacy and safety of the randomized treatment comparison in patients with and without prior MI in the on-treatment population. By design, ROCKET AF was a noninferiority trial, and the primary analysis was performed in the safety population while on treatment.23 Outcomes between the patients with and without prior MI were compared using the Cox proportional hazards method and reported as events per 100 patient-years (%/year) and total number of events. The homogeneity of treatment effects on the occurrence of the key efficacy and safety endpoints across subgroups with and without prior MI was tested with a Cox proportional hazards regression model with terms for both main effects and the treatment by prior MI interaction. Models that were adjusted for prior stroke or transient ischaemic attack, creatinine clearance, diastolic blood pressure, history of vascular disease, paroxysmal AF, heart rate, sex, diabetes, hypertension, age, and heart failure were fit for the CV death, MI, or UA composite endpoint.

Estimates and two-sided 95% confidence intervals (CIs) for the hazard ratio (HR) (rivaroxaban vs. warfarin) by prior MI are presented in the safety, on-treatment population for both the efficacy and safety endpoints. Landmark analyses were performed to determine whether the effect of rivaroxaban on outcomes differed in patients who were and were not taking aspirin 30 days following randomization. These analyses included patients who were event-free and being followed 30 days post-randomization; time to event was measured from Day 30. Aspirin and clopidogrel use at baseline, after study entry, and around the time of endpoint MI and post-MI treatment among the 227 patients with an on-treatment MI are summarized with counts and percentages. The Kaplan–Meier plots are presented for the CV death, MI, and UA composite endpoint and for the principal safety endpoint with two stratifications—treatment/prior MI or treatment/baseline antiplatelet use.

All statistical analyses were performed with SAS software (version 9.2, SAS Institute Inc., Cary, NC).

Results

Of the 14 264 patients enrolled in ROCKET AF, 2468 (17.3%) had prior MI at baseline. Table 1 shows the baseline demographics of patients with and without prior MI. Patients with prior MI tended to be older, male, and more often had risk factors for atherosclerosis and used antiplatelet agents. Among warfarin-treated patients with prior MI and those without prior MI, the median (25th, 75th percentiles) time the INR was in the therapeutic range (2.0–3.0, inclusive) was 59.1 (44.9, 71.5%) and 57.6 (42.6, 70.4%), respectively.

View this table:
Table 1

Baseline demographics and clinical characteristics according to treatment allocation and history of prior myocardial infarction

Prior MINo Prior MI
Rivaroxaban (n = 1182)Warfarin (n = 1286)Total (n = 2468)Rivaroxaban (n = 5949)Warfarin (n = 5847)Total (n = 11 796)
Age73 (66, 79)74 (67, 79)73 (67, 79)73 (65, 78)72 (65, 78)73 (65, 78)
Female sex, no. (%)290 (24.5)338 (26.3)628 (25.4)2540 (42.7)2492 (42.6)5032 (42.7)
BMI, kg/m229 (25, 32)28 (26, 32)28 (26, 32)28 (25, 32)28 (25, 32)28 (25, 32)
Systolic BP, mm Hg130 (120, 140)130 (120, 140)130 (120, 140)130 (120, 140)130 (120, 140)130 (120, 140)
Diastolic BP, mm Hg80 (70, 85)80 (70, 84)80 (70, 85)80 (70, 85)80 (70, 86)80 (70, 86)
Creatinine clearance65 (50, 85)64 (50, 82)65 (50, 83)68 (53, 88)68 (53, 87)68 (53, 88)
AF type, no. (%)
 Persistent949 (80.3)1005 (78.1)1954 (79.2)4837 (81.3)4757 (81.4)9594 (81.3)
 Paroxysmal219 (18.5)259 (20.1)478 (19.4)1026 (17.2)1010 (17.3)2036 (17.3)
 Newly diagnosed14 (1.2)22 (1.7)36 (1.5)86 (1.4)80 (1.4)166 (1.4)
CHADS2 score3 (3, 4)3 (3, 4)3 (3, 4)3 (3, 4)3 (3, 4)3 (3, 4)
Baseline medications, no. (%)
 Aspirin569 (48.1)602 (46.8)1171 (47.4)2017 (33.9)2017 (34.5)4034 (34.2)
 VKA782 (66.2)843 (65.6)1625 (65.8)3661 (61.5)3618 (61.9)7279 (61.7)
 Thienopyridine32 (2.7)51 (4.0)83 (3.4)119 (2.0)130 (2.2)249 (2.1)
Clinical risk factors, no. (%)
 Hypertension1102 (93.2)1214 (94.4)2316 (93.8)5334 (89.7)5260 (90.0)10 594 (89.8)
 Congestive HF927 (78.4)994 (77.3)1921 (77.8)3540 (59.5)3447 (59.0)6987 (59.2)
 Diabetes547 (46.3)602 (46.8)1149 (46.6)2331 (39.2)2215 (37.9)4546 (38.5)
 Stroke or TIA536 (45.3)595 (46.3)1131 (45.8)3218 (54.1)3119 (53.3)6337 (53.7)
 PAD136 (11.5)153 (11.9)289 (11.7)265 (4.5)285 (4.9)550 (4.7)
 COPD189 (16.0)174 (13.5)363 (14.7)565 (9.5)569 (9.7)1134 (9.6)
  • Values presented as median (25th, 75th percentiles), unless otherwise indicated.

  • AF, atrial fibrillation; BMI, body mass index; BP, blood pressure; COPD, chronic obstructive pulmonary disease; HF, heart failure; MI, myocardial infarction; PAD, peripheral arterial disease; TIA, transient ischaemic attack.

Stroke, cardiac, and bleeding outcomes are shown in Table 2 and Figure 1 for patients with and without prior MI. Patients with and without prior MI had similar rates of stroke and non-central nervous system embolism. Patients with prior MI had more recurrent CV events and higher mortality. These patients also had higher rates of major bleeding. After adjustment for differences in baseline characteristics, participants with prior MI remained at elevated risk for recurrent CV events (HR 1.64, 95% CI 1.21–2.24; P = 0.002) (Supplementary material online, Table S1).

View this table:
Table 2

History of myocardial infarction vs. no history of myocardial infarction (on-treatment, safety population)

Prior MINo Prior MI
EndpointTotalEvents per 100 patient-yearsTotalEvents per 100 patient-yearsHR (95% CI)P value
Stroke outcomes
 Stroke or non-CNS systemic embolism711.913611.930.99 (0.77, 1.28)0.9529
 Any stroke611.643441.840.89 (0.68, 1.17)0.4239
  Hemorrhagic stroke90.24700.370.65 (0.32, 1.30)0.2191
  Ischaemic stroke501.352601.390.97 (0.72, 1.31)0.8451
  Unknown stroke20.05160.090.64 (0.15, 2.76)0.5455
 Stroke disability
  Fatal stroke200.54940.501.07 (0.66, 1.74)0.7761
  Disabling stroke130.35870.460.75 (0.42, 1.34)0.3342
  Non-disabling stroke240.651510.810.80 (0.52, 1.24)0.3199
  Stroke unknown disability50.13140.071.80 (0.65, 4.99)0.2599
 Non-CNS systemic embolism100.27170.092.96 (1.35, 6.45)0.0065
CV outcomes
 MI1022.781250.674.16 (3.20, 5.40)<.0001
 All-cause mortality1393.743191.702.19 (1.80, 2.68)<.0001
  CV death1213.252421.292.51 (2.02, 3.13)<.0001
  Non-CV death80.22470.250.86 (0.41, 1.83)0.7016
  Death unknown cause100.27300.161.68 (0.82, 3.43)0.1559
 UA381.03600.323.21 (2.14, 4.82)<.0001
 CV death or MI2095.703551.902.99 (2.52, 3.55)<.0001
 CV death, MI, or UA2426.684082.193.04 (2.59, 3.56)<.0001
Bleeding outcomes
 Principal safety endpoint55517.07236914.261.19 (1.08, 1.30)0.0003
 Major bleeding1524.146293.401.21 (1.02, 1.45)0.0325
 Hemoglobin/hematocrit drop1213.294382.361.39 (1.13, 1.70)0.0014
 Transfusion701.892621.401.34 (1.03, 1.74)0.0302
 Critical organ bleeding310.831931.030.81 (0.55, 1.18)0.2676
 Fatal major bleeding90.24730.390.62 (0.31, 1.24)0.1772
 Intracranial bleeding180.481210.640.75 (0.46, 1.23)0.2541
 Non-major clinically relevant bleeding44213.38189411.241.18 (1.06, 1.31)0.0018
  • CI, confidence interval; CNS, central nervous system; CV, cardiovascular; HR, hazard ratio; MI, myocardial infarction; UA, unstable angina.

Figure 1

Clinical outcomes in patients with and without prior myocardial infarction.

Major bleeding rates in patients with and without baseline use of aspirin and/or a thienopyridine were 4.59 and 3.73 for patients with prior MI vs. 4.27 and 2.97 for patients without prior MI (interaction P value = 0.408).

Table 3 shows CV outcomes by randomly assigned treatment in the overall trial. Patients assigned to rivaroxaban tended to have lower rates of CV events with a non-significant 14% reduction in the hazard of CV death, MI, or UA [HR 0.86, 95% CI 0.73–1.00; P = 0.051 (adjusted HR 0.88, 95% CI 0.75–1.03; adjusted P = 0.103)].

View this table:
Table 3

Overall rivaroxaban treatment effect on cardiovascular events (safety population, on treatment)

RivaroxabanWarfarin
TotalEvents per 100 patient-yearsTotalEvents per 100 patient-yearsHR (95% CI)P value
CV death, MI, or UA2982.703523.150.86 (0.73, 1.00)0.0509
CV death or MI2602.353042.700.87 (0.74, 1.02)0.0934
All-cause mortality2081.872502.210.85 (0.70, 1.02)0.0735
CV death1701.531931.710.89 (0.73, 1.10)0.2891
Non-CV death210.19340.300.63 (0.36, 1.08)0.0936
Death unknown cause170.15230.200.75 (0.40, 1.41)0.3700
MI1010.911261.120.81 (0.63, 1.06)0.1211
UA410.37570.510.73 (0.49, 1.09)0.1225
  • CI, confidence interval; CV, cardiovascular; HR, hazard ratio; MI, myocardial infarction; UA, unstable angina.

Table 4 shows cardiac, stroke, and bleeding outcomes in patients with and without prior MI by treatment assignment and the MI by treatment interaction P values. Several interaction P values are <0.05, suggesting a differential treatment effect of rivaroxaban compared with warfarin in patients with prior MI. A reduction of all-cause mortality with rivaroxaban was observed in patients without prior MI, while this was not seen among patients with prior MI (interaction P = 0.0296). For major or NMCR bleeding and NMCR bleeding, excess bleeding with rivaroxaban was observed only among patients with prior MI (interaction P = 0.0352 and 0.0427, respectively). Consistent reductions in intracranial haemorrhage and fatal bleeding were observed with rivaroxaban in patients with and without prior MI.

View this table:
Table 4

Treatment by myocardial infarction history interaction (on-treatment, safety population)

Prior MINo Prior MI
EndpointEvents per 100 patient-years (Rivaroxaban)Events per 100 patient-years (Warfarin)HR (95% CI)Events per 100 patient-years (Rivaroxaban)Events per 100 patient-years (Warfarin)HR (95% CI)Treatment by prior MI P value
Stroke outcomes
 Stroke or non-CNS systemic embolism1.422.350.61 (0.37, 0.99)1.752.110.83 (0.67, 1.02)0.2522
 Any stroke1.371.890.73 (0.43, 1.21)1.711.970.87 (0.70, 1.07)0.5355
  Haemorrhagic stroke0.230.250.90 (0.24, 3.34)0.270.480.55 (0.34, 0.90)0.5007
  Ischaemic or unknown stroke type1.141.630.70 (0.40, 1.22)1.441.500.96 (0.76, 1.22)0.3069
 Non-CNS systemic embolism0.060.460.12 (0.02, 0.98)0.040.140.31 (0.10, 0.94)0.4529
CV outcomes
 MI2.772.800.99 (0.67, 1.46)0.570.770.73 (0.51, 1.04)0.2626
 All-cause mortality4.043.461.17 (0.84, 1.63)1.461.950.75 (0.60, 0.94)0.0296
  CV death3.592.951.22 (0.85, 1.74)1.141.440.79 (0.61, 1.02)0.0533
  Non-CV death0.280.151.86 (0.45, 7.79)0.170.330.52 (0.28, 0.94)0.1046
  Death unknown cause0.170.360.48 (0.12, 1.87)0.150.170.87 (0.42, 1.78)0.4498
 UA0.801.240.65 (0.34, 1.26)0.290.350.81 (0.49, 1.35)0.5961
 CV death or MI5.945.491.08 (0.83, 1.42)1.682.130.79 (0.64, 0.97)0.0711
 CV death, MI, or UA6.806.571.04 (0.81, 1.33)1.942.450.79 (0.65, 0.96)0.1008
Bleeding outcomes
 Principal safety endpoint18.8415.511.21 (1.03, 1.43)14.2014.310.99 (0.92, 1.08)0.0352
 Major bleeding4.753.611.32 (0.96, 1.81)3.383.420.99 (0.84, 1.15)0.1119
 Haemoglobin/hematocrit drop3.872.781.39 (0.97, 1.99)2.562.161.19 (0.98, 1.43)0.4395
 Transfusion2.411.431.68 (1.04, 2.70)1.511.301.16 (0.91, 1.48)0.1817
 Critical organ bleeding0.680.970.71 (0.34, 1.46)0.841.220.69 (0.52, 0.92)0.9455
 Fatal major bleeding0.170.300.56 (0.14, 2.25)0.250.520.49 (0.30, 0.79)0.8497
 Intracranial bleeding0.340.610.56 (0.21, 1.50)0.520.770.68 (0.47, 0.97)0.7240
 Non-major clinically relevant bleeding14.9412.011.24 (1.03, 1.50)11.2411.231.00 (0.92, 1.10)0.0427
  • CI, confidence interval; CNS, central nervous system; CV, cardiovascular; HR, hazard ratio; MI, myocardial infarction; UA, unstable angina.

The Kaplan–Meier curves for CV death, MI, and UA in patients with and without prior MI by treatment assignment are shown in Figure 2. A Kaplan–Meier plot that was adjusted for selected baseline characteristics is provided as Supplementary material online, Figure S1. After adjustment, there was some evidence that rivaroxaban vs. warfarin was more beneficial in preventing CV events in those without prior MI than in those with prior MI, but this was non-significant (HR among those without prior MI 0.79, 95% CI 0.65–0.96; HR among those with prior MI 1.06, 95% CI 0.82–1.36; interaction P = 0.075). The Kaplan–Meier curves for CV death, MI, and UA by treatment and by antiplatelet use at baseline are shown in Supplementary material online, Figure S2. Event rates tended to be higher for patients with aspirin or clopidogrel use at baseline and lowest in patients on neither at baseline, although the number of patients on clopidogrel at baseline was low. Supplementary material online, Figure S3, shows the principal safety outcome in the same subgroups. Similar trends are observed with the highest event rates in patients on both aspirin and clopidogrel at baseline.

Figure 2

Cardiovascular death, myocardial infarction, and unstable angina in patients with and without prior myocardial infarction by treatment assignment.

Because aspirin and prior MI are likely strongly linked and concomitant use of aspirin and anticoagulant therapies are of particular clinical interest, the relationship between aspirin use and outcomes was modelled in two ways (Supplementary material online, Tables S2 and S3). The first model included all patients in the safety population and contained aspirin use at study entry. The second model was a 30-day landmark analysis with an indicator for aspirin use 30 days after randomization because the use of aspirin at baseline could have been altered because of inclusion in the study and initiation of anticoagulant therapy. The time point of 30 days was chosen because aspirin use should be consistent by this time. In both models, aspirin use was associated with higher event rates, but the treatment effect of rivaroxaban compared with warfarin did not differ depending on whether aspirin was used (interaction P values all >0.3).

Overall, a total of 227 MIs occurred while on study drug treatment for a median of 0.8 (0.4, 1.4) years following study drug initiation. The use of fibrinolytic therapy; use of percutaneous coronary intervention (PCI) within 2 h of the time of MI; and the use of antiplatelet, anticoagulant, and other cardiac medications are shown in Supplementary material online, Table S4. In total, 22 patients were treated with lytic therapy and none of these had intracranial hemorrhage within 5 days of lytic use. Few patients were treated with PCI. About half of the patients were treated with aspirin and the use of other cardiac medications was low. Overall, 84 patients had study drug permanently stopped within 5 days of MI, 74 had study drug temporarily stopped, and 69 remained on study drug through the MI event.

Supplementary material online, Table S5, shows aspirin and thienopyridine use at baseline and 30 days. Overall, 36.5% of patients were on aspirin at randomization and ∼60% of those remained on aspirin at 30 days post-enrollment with a similar distribution between treatment groups. Few patients were on thienopyridine at baseline since use within 5 days of randomization was an exclusion criterion. A total of 126 (0.9%) patients were on both aspirin and a thienopyridine at randomization and 58 (0.4%) were on both at 30 days; these were protocol deviations since the use of a thienopyridine during the course of the study was only allowed for those receiving a coronary artery stent.

Supplementary material online, Table S6, shows the proportion of patients on antiplatelet therapy within the 30 days prior to MI or the principal safety endpoint (major or NMCR bleeding). Approximately 30% of patients were on aspirin or thienopyridine therapy prior to these events.

Discussion

In the high-risk AF population enrolled in ROCKET AF, a history of MI was common (∼20% of patients). A 14% reduction in the hazard for CV death, MI, or UA was observed while patients were taking rivaroxaban with comparable reductions in each of the individual components of the composite. Patients with prior MI had worse CV and bleeding outcomes compared with patients without prior MI. The overall trial demonstrated that rivaroxaban was noninferior to warfarin for the primary efficacy endpoint of the composite of stroke and non-central nervous system embolism with similar rates of major or NMCR bleeding. In patients with and without prior MI while taking the study drug, the comparison of rivaroxaban with warfarin for the key efficacy endpoint of CV death, MI, and UA was consistent with the overall trial results. A statistically significant interaction suggests that patients with prior MI had worse major/NMCR bleeding outcomes with rivaroxaban compared with warfarin that likely were associated with higher use of concomitant aspirin, although other factors not measured could have contributed or it could be due to chance finding due to multiple comparisons. However, consistent reductions in intracranial haemorrhage and fatal bleeding with rivaroxaban were observed in both those with and without prior MI.

The reductions in CV death, MI, and UA observed with rivaroxaban are consistent with results from completed trials with rivaroxaban in patients after ACS.25,26 However, direct comparison of treatment effects with rivaroxaban between the two trials is not proper. The ATLAS 2 trial specifically excluded patients with an indication for anticoagulation therapy. Further, there was frequent use of dual antiplatelet therapy and the doses of rivaroxaban were one-quarter (2.5 mg twice daily) and one-half (5 mg twice daily) the dose used in ROCKET AF (20 mg once daily). Of note, in this post-ACS population, rivaroxaban use resulted in significant increases in major bleeding and intracranial haemorrhage compared with placebo.26 The observations that patients with prior MI had more bleeding than patients without prior MI are consistent with ATLAS 2, likely related to differences in bleeding risk (e.g. older age, more comorbidity) and the higher use of antiplatelet therapy. However, the number of patients with prior MI and outcome events precludes more rigorous analyses to explore the associations of these post-randomization covariates.

Another factor Xa inhibitor, apixaban, has undergone similar evaluations in post-ACS patients with AF. In contrast to the rivaroxaban experience utilizing different dosing regimens, the same dose of apixaban was studied both after ACS27 and in AF.28 The data show significant excess in bleeding and no efficacy post-ACS with apixaban compared with placebo but significant reductions in stroke, death, and bleeding in patients with AF compared with warfarin.

Dabigatran, a factor IIa inhibitor, has been shown to reduce thromboembolic events in patients with AF.29 A significant 38% relative increase in MI was initially reported for the 150 mg dose,29 but after re-evaluation of MI events due to possible underreporting in an open-label trial setting, a nonsignificant 27% excess was reported.30 An excess of MI events with dabigatran was identified in a systematic overview of seven trials including AF, deep vein thrombosis, and ACS populations;31,32 however, these findings must be placed in the context of a simultaneous demonstration of a significant reduction in fatal events with dabigatran, recognizing the challenges of examining non-fatal MI rates as a competing outcome. Until trials are conducted that directly compare factor IIa and Xa inhibitors, it will remain unknown if a potential advantage exists with either class of drug or specific agents to prevent MI in patients with AF.

Management of antiplatelet and anticoagulant therapy in patients with AF who suffer acute MI and in patients with CAD who develop AF is challenging. Indeed, as the population ages, the number of patients with both conditions will increase. The optimal strategy to provide adequate antiplatelet and anticoagulant therapy is unclear. Consistent with our findings, recent observations from the RE-LY trial show an increased risk of bleeding and thromboembolic events associated with antiplatelet therapy compared with no antiplatelet therapy, and consistent treatment effects compared with warfarin regardless of aspirin use.33 Observational data have shown an increased risk of bleeding with antiplatelet therapy together with an anticoagulant—so-called “triple therapy.”20,34 Currently, three ongoing randomized trials in patients with AF undergoing PCI are evaluating different strategies of antiplatelet and anticoagulant therapy.3537 Clinical considerations to decrease bleeding have been suggested,20 including using low-dose antiplatelet therapy; using only single antiplatelet therapy combined with anticoagulant treatment;35 using bare metal stents rather than drug-eluting stents, when appropriate; targeting an optimal INR of 2.0–2.5; proton-pump inhibitor prophylaxis; and shorter duration of triple therapy. The potential to use a new oral anticoagulant instead of warfarin in this setting has not yet been explored. The appropriate long-term management strategy of rivaroxaban in patients with AF who suffer acute MI is not clear because no studies have been done to determine whether a lower dose of rivaroxaban when used with dual antiplatelet therapy (e.g. dosing regimen employed in ATLAS 2) will be sufficient to prevent AF-related thromboembolic events.

For patients suffering MI during treatment for AF with rivaroxaban as occurred in ROCKET AF, too few patients had fibrinolytic therapy or primary PCI to provide clear guidance driven by definitive data. However, the strategy used during the trial of recommending primary PCI, if available, and balancing risks and benefit prior to lysis seems prudent.

Limitations

Our analyses have several limitations. The subgroups defined by MI status at baseline were prespecified, but the trial was not powered to make definitive observations for the randomized treatments within subgroups. Subgroups defined by post-randomization events, such as MI, are potentially confounded so only descriptive analyses were performed on these patients. Because of the relatively small number of MI events, statistical models using landmark techniques or inverse probability weighting analyses to better understand the impact of changes in the treatment of antiplatelet therapy during study drug treatment are not appropriate. Finally, we did not adjust for multiple comparisons.

Conclusions

In a trial population of patients with AF, ∼20% had a history of prior MI. Patients with prior MI are at higher risk for subsequent cardiac and bleeding events compared with patients without prior MI. While patients were taking rivaroxaban, a 14% reduction in the hazard (HR 0.86, 95% CI 0.73–1.00; P = 0.0509) for CV death, MI, or UA was observed. The overall trial results of the effect of rivaroxaban compared with warfarin on stroke and systemic embolism outcomes were consistent in patients with and without prior MI. Patients with prior MI had worse major/NMCR bleeding outcomes with rivaroxaban compared with warfarin that likely were associated with higher use of aspirin therapy in this group. However, consistent reductions in intracranial haemorrhage and fatal bleeding were observed with rivaroxaban compared with warfarin in patients with and without prior MI.

Funding

This work was supported by Johnson & Johnson Pharmaceutical Research & Development, Raritan, NJ and Bayer HealthCare AG, Leverkusen, Germany.

Conflict of interest: K.W.M.: grant support (significant) from AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly, GlaxoSmithKline, Johnson & Johnson, Merck, Momenta Pharmaceuticals, Novartis, Portola, Pozen, Regado Biotechnologies, Sanofi-Aventis, Schering-Plough (now Merck), The Medicines Company; consulting fees (significant) from AstraZeneca and Johnson & Johnson, (modest) Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly, GlaxoSmithKline, Merck, Novartis, Ortho/McNeill, Pfizer, Polymedix, Sanofi-Aventis, Schering-Plough (now Merck). S.R.S.: None. H.D.W.: consulting fees/honoraria and/or research grant support from Sanofi-Aventis, Eli Lilly, The Medicines Company, NIH, Pfizer, Roche, Johnson & Johnson, Schering-Plough, Merck Sharp & Dohme, AstraZeneca, GlaxoSmithKline, Daiichi Sankyo, Bristol Myers Squibb, and Regado. C.C.N.: Employee of Johnson & Johnson Research and Development. S.G.G.: consulting fees/honoraria and/or research grant support from Bayer, Johnson & Johnson, Boehringer Ingelheim, Bristol Myers Squibb, Sanofi-Aventis. J.P.P.: grants for clinical research from Johnson & Johnson and Boston Scientific; and consulting and/or advisory board fees from Medtronic, Forest Laboratories, Sanofi Aventis, and Johnson & Johnson. M.R.P.: honoraria from Johnson & Johnson and Bayer HealthCare for serving on the executive committee of the ROCKET AF trial; consulting fees from Ortho McNeil Janssen, and Bayer HealthCare; and advisory board fees from Genzyme. R.C.B.: research support from Bayer and Johnson & Johnson. J.L.H.: honoraria from Johnson & Johnson and Bayer; and advisory board fees from Boehringer Ingelheim, Bristol Myers-Squibb, and Pfizer. W.H.: honoraria from Johnson & Johnson and Bayer; and advisory board fees from Boehringer Ingelheim. D.E.S.: supported, in part, by the Eliot B. and Edith C. Shoolman fund of the Massachusetts General Hospital (Boston, MA, USA); has received consulting fees from Bayer HealthCare, Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Johnson & Johnson, Pfizer, and Sanofi; and serves as a member of the executive committee of the ROCKET AF trial of rivaroxaban vs. warfarin in patients with AF sponsored by Johnson & Johnson and Bayer HealthCare. G.J.H.: honoraria from Johnson & Johnson, Bayer, and Sanofi-Aventis; and has received fees for serving on trial adjudication committees and an advisory board for Boehringer Ingelheim. R.M.C.: consulting fees and research funding from Johnson & Johnson; all other industry interactions are listed at www.dcri.org. K.A.A.F.: grants and honoraria from Bayer, Lilly, Boehringer Ingelheim, Sanofi-Aventis, and GlaxoSmithKline. G.B.: honoraria from Johnson & Johnson and Bayer; and advisory board fees from Boehringer Ingelheim, Bristol-Myers Squibb, Pfizer, and Sanofi-Aventis.

References

View Abstract