If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Quebec, CanadaDepartment of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Quebec, CanadaInstitute of Clinical Pharmacology and Toxicology, Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Berlin, Germany
Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Quebec, CanadaDepartment of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Quebec, CanadaDepartment of Neurology and Neurosurgery, McGill University, Montreal, Quebec, Canada
Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Quebec, CanadaDepartment of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Quebec, CanadaDepartment of Medicine, McGill University, Montreal, Quebec, Canada
Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Quebec, CanadaDepartment of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Quebec, Canada
Requests for reprints should be addressed to Laurent Azoulay, PhD, Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, 3755 Cote Sainte-Catherine Road, H425.1, Montreal, QC H3T 1E2, Canada.
Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Quebec, CanadaDepartment of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Quebec, CanadaGerald Bronfman Department of Oncology, McGill University, Montreal, Quebec, Canada
Patients with nonvalvular atrial fibrillation commonly have comorbidities requiring concurrent use of oral anticoagulants and antiplatelets. There are no real-world data on the comparative safety of concomitant antithrombotic treatments in the era of direct oral anticoagulant (DOACs). Thus, we compared the incidence of intracranial hemorrhage, gastrointestinal bleeding, and other major bleeding between concomitant DOAC-antiplatelet use and concomitant vitamin K antagonist (VKA)-antiplatelet use in patients with nonvalvular atrial fibrillation.
Methods
Using computerized health care databases from Québec, we conducted a cohort study among patients newly diagnosed with nonvalvular atrial fibrillation between January 2011 and March 2014. Cox proportional hazards models yielded hazard ratios (HRs) and 95% confidence intervals (CIs), adjusted for disease risk score, of the study outcomes comparing current concomitant use of DOACs with ≥1 antiplatelet vs current concomitant use of VKAs with ≥1 antiplatelet.
Results
A total of 5301 patients initiated concomitant DOAC-antiplatelet use, while 9106 patients initiated concomitant VKA-antiplatelet use. During a median follow-up of 1.6 months, concomitant DOAC-antiplatelet use was associated with a similar risk of gastrointestinal bleeding (HR 1.08; 95% CI, 0.81-1.45), but with a decreased risk of intracranial hemorrhage (HR 0.46; 95% CI, 0.24-0.91) and other major bleeding (HR 0.68; 95% CI, 0.51-0.91) compared with concomitant VKA-antiplatelet use.
Conclusions
Concomitant DOAC-antiplatelet use was associated with a similar risk of gastrointestinal bleeding, and a lower risk of intracranial hemorrhage and other major bleeding than concomitant VKA-antiplatelet use. These findings could inform physician decision-making in patients requiring concomitant treatment with oral anticoagulants and antiplatelets.
Direct oral anticoagulants with antiplatelets are associated with a similar risk of gastrointestinal bleeding compared with vitamin K antagonists and antiplatelets.
•
Direct oral anticoagulants with antiplatelets are associated with a decreased risk of intracranial hemorrhage and other major bleeding compared with vitamin K antagonists and antiplatelets.
•
There were no duration–response relations for the 3 bleeding outcomes.
•
Age, sex, or baseline bleeding risk did not modify the associations.
Background
Direct oral anticoagulants (DOACs) are increasingly being used for ischemic stroke prevention among patients with nonvalvular atrial fibrillation,
Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials.
However, the safety of DOACs in patients with nonvalvular atrial fibrillation requiring additional use of antiplatelet agents remains uncertain. This is important, as up to 30% of patients with nonvalvular atrial fibrillation may receive concomitant treatments of oral anticoagulants with antiplatelets due to comorbid cardiovascular conditions.
Use and associated risks of concomitant aspirin therapy with oral anticoagulation in patients with atrial fibrillation: insights from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) Registry.
High rates of concomitant antiplatelet use in patients with atrial fibrillation treated with oral anticoagulation: insights from the stroke prevention and rhythm interventions in atrial fibrillation (SPRINT-AF) registry.
Efficacy and safety of the use of non-vitamin K antagonist oral anticoagulants in patients with nonvalvular atrial fibrillation and concomitant aspirin therapy: a meta-analysis of randomized trials.
data are lacking on the comparative safety of concomitant use of antithrombotic drugs in the era of DOACs.
In a meta-analysis of randomized controlled trials (RCTs) assessing the efficacy of DOACs in nonvalvular atrial fibrillation, use of acetylsalicylic acid (ASA) with DOACs was associated with a similar risk of major bleeding (hazard ratio [HR] 0.83; 95% confidence interval [CI], 0.69-1.01), but a decreased risk of intracranial hemorrhage (HR 0.38; 95% CI, 0.26-0.56), compared with use of ASA with VKAs.
Efficacy and safety of the use of non-vitamin K antagonist oral anticoagulants in patients with nonvalvular atrial fibrillation and concomitant aspirin therapy: a meta-analysis of randomized trials.
Efficacy and safety of the use of non-vitamin K antagonist oral anticoagulants in patients with nonvalvular atrial fibrillation and concomitant aspirin therapy: a meta-analysis of randomized trials.
Moreover, 2 RCTs evaluating concomitant use of DOACs with antiplatelets restricted inclusion to patients with nonvalvular atrial fibrillation undergoing coronary intervention.
Finally, a recently published cohort study addressing the safety of such regimes in real-world clinical practice had methodological limitations such as inclusion of prevalent users and did not report results on specific types of major bleeding.
Thus, to address this important safety issue, we conducted a population-based study to compare the incidence of intracranial hemorrhage, gastrointestinal bleeding, and other major bleeding between concomitant DOAC-antiplatelet use and concomitant VKA-antiplatelet use in patients with nonvalvular atrial fibrillation.
Methods
Data Sources
This study was conducted by linking 3 computerized health care databases from the Canadian province of Québec: Régie de l'assurance maladie du Québec (RAMQ), Maintenance et exploitation des données pour l’étude de la clientèle hospitalière (MEDÉCHO), and the Institut de la statistique du Québec (ISQ).
The RAMQ databases collect information on demographics, medical services (coded using the International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] or enhanced version of ICD-10 for Canada ICD-10-CA), and dispensed outpatient prescriptions. Under the universal health care program, medical services are covered for all Québec residents, while coverage in the Public Prescription Drug Insurance Plan is limited to residents (and their children) with no private drug insurance plans, those at least 65 years of age, and recipients of financial assistance.
The use of prescription claims databases in pharmacoepidemiological research: the accuracy and comprehensiveness of the prescription claims database in Quebec.
MEDÉCHO contains records of all hospitalizations occurring in Québec and includes date and type of admission and discharge, as well as primary and secondary diagnoses (coded using ICD-10-CA). Finally, ISQ contains vital statistics including date and cause of death. The study protocol was approved by the Research Ethics Committee of the Jewish General Hospital, Montreal, Canada. No informed consent was required.
Base Cohort of Patients with Incident Nonvalvular Atrial Fibrillation
Using the databases above, we identified all patients at least 18 years of age with a first inpatient or outpatient diagnosis of atrial fibrillation between January 1, 2011 (when the first DOAC, dabigatran, was approved for stroke prevention in nonvalvular atrial fibrillation in Québec) and March 31, 2014. All patients were required to have been covered by the RAMQ Public Prescription Drug Insurance Plan for at least 1 year prior to the diagnosis of atrial fibrillation. We restricted inclusion to patients with nonvalvular atrial fibrillation by excluding patients with a history of mitral or aortic stenosis or valvular repair, and those with a history of hyperthyroidism, at any time prior to the diagnosis of atrial fibrillation. Finally, we excluded patients dispensed VKAs or DOACs in the year prior to the nonvalvular atrial fibrillation diagnosis to maximize the inclusion of new users during the study period.
Study Cohort of Concomitant Oral Anticoagulant-Antiplatelet Users
Using the base cohort defined above, we identified all patients initiating concomitant use of an oral anticoagulant (DOAC [dabigatran, rivaroxaban, or apixaban] or VKA) with at least one antiplatelet agent (ASA, dipyridamole, clopidogrel, prasugrel, or ticagrelor) during the study period. Thus, concomitant users comprised 3 types of patients: 1) those on antiplatelets with a prescription overlapping a new prescription for an oral anticoagulant; 2) those on an oral anticoagulant with a prescription overlapping a new prescription for at least one antiplatelet; and 3) patients initiating treatment with an oral anticoagulant and at least one antiplatelet on the same day (Supplementary Figure 1, available online). Cohort entry was defined as the first day of concomitant use during the study period. We excluded patients with any bleeding-related hospitalization in the 3 months prior to cohort entry, to exclude events related to precohort entry antithrombotic exposure. Patients were followed until an event (described in detail below), discontinuation of concomitant use (described in detail below), end of registration with the Public Prescription Drug Insurance Plan, death, or end of study period (December 31, 2014), whichever occurred first.
Exposure Groups
We compared patients initiating concomitant use of a DOAC with at least one antiplatelet with patients initiating concomitant use of a VKA with at least one antiplatelet. The latter group served as reference as this represented a clinically relevant comparator, while minimizing potential confounding by indication.
An as-treated exposure definition was used where patients were considered continuously exposed to concomitant use if the prescription durations of the drugs of interest were overlapping each other. We allowed for a 30-day grace period in the event of nonoverlapping prescriptions.
Bleeding Outcomes
We conducted 3 analyses; for intracranial hemorrhage, gastrointestinal bleeding, and other major bleeding separately (ICD-10 codes in Supplementary Table 1, available online). The bleeding outcomes were defined by inpatient diagnoses (captured in MEDÉCHO in primary or secondary position), or bleeding-related deaths (captured in MEDÉCHO or ISQ).
Disease Risk Score
To control for confounding, we calculated disease risk scores (DRS) using a historical cohort from the same data source.
This historical cohort comprised patients newly diagnosed with nonvalvular atrial fibrillation between January 1, 2007 and December 31, 2010 who were on concomitant VKA-antiplatelet use. The DRS included the following variables measured at cohort entry: age, sex, alcohol-related disorders, hypertension, ischemic stroke or transient ischemic attack, congestive heart failure, coronary artery disease, peripheral vascular disease, diabetes mellitus, renal or liver disease, all diagnosed at any time prior to cohort entry. We also considered time from nonvalvular atrial fibrillation diagnosis to initiation of oral anticoagulation, cancer (other than nonmelanoma skin cancer) diagnosed in the year prior to cohort entry, and history of major bleeding 3 to 15 months prior to cohort entry. Moreover, the model included use of nonsteroidal anti-inflammatory drugs, selective serotonin reuptake inhibitors, proton pump inhibitors, and H2 blockers in the year prior to cohort entry, because these drugs have been associated with bleeding. Finally, we measured the number of nonantithrombotic drugs in the year prior to cohort entry as a proxy for overall health. For each bleeding outcome, we fitted a Cox proportional hazards model to the historical cohort, which was then applied to the study cohort to estimate the relative hazard of bleeding for each cohort member.
Crude incidence rates and rate differences for the bleeding outcomes with 95% CIs based on the Poisson distribution were calculated for each exposure group. We used Cox proportional hazards models to estimate HRs and 95% CIs for each bleeding outcome (intracranial hemorrhage, gastrointestinal bleeding, other major bleeding) associated with concomitant DOAC-antiplatelet use compared with concomitant VKA-antiplatelet use. The models were adjusted for DRS by including an interaction term between DRS modeled as a categorical variable (quintiles for intracranial hemorrhage and deciles for gastrointestinal bleeding and other major bleeding) and DRS as a continuous variable, as well as year of cohort entry.
Secondary Analyses
We conducted 7 secondary exploratory analyses. First, we repeated the primary analyses for the individual DOACs (dabigatran, rivaroxaban, and apixaban). Second, to assess a possible duration–response relation between concomitant DOAC-antiplatelet use and the incidence of the bleeding outcomes, we estimated HRs for 3 prespecified duration categories (<3, 3-6, and >6 months). Third, we assessed whether the risk of the bleeding outcomes associated with concomitant DOAC-antiplatelet use varies among patient subgroups (age ≤75 vs >75 years, males vs females, or a modified HAS-BLED [uncontrolled hypertension, abnormal renal or liver function, previous stroke, bleeding history or predisposition, age >65 years, alcohol-related disorders] score
of 0-2 vs 3-7). Fourth, we assessed the association between concomitant DOAC-antiplatelet use and risk of genitourinary bleeding, a common complication of antithrombotic treatment.
Fifth, we assessed the risk of the bleeding outcomes associated with concomitant DOAC-ASA use vs concomitant VKA-ASA use. This analysis was limited to patients using ASA as the only antiplatelet at cohort entry and censored upon addition or switch to another antiplatelet during follow-up. Sixth, we assessed the risk of the bleeding outcomes in patients on DOACs with at least 2 antiplatelets compared with patients on VKAs with at least 2 antiplatelets. Finally, we assessed whether the risk of the bleeding outcomes differed between patients entering the study cohort upon addition of an oral anticoagulant, an antiplatelet, or both.
Sensitivity Analyses
We conducted 6 sensitivity analyses to assess the robustness of our findings. First, to assess potential exposure misclassification, we repeated the primary analyses using a 15-day grace period between successive prescriptions. Second, we used a stricter outcome definition considering only primary diagnoses for hospitalized bleeding or fatal bleeding events. Third, to assess potential informative censoring, as discontinuation of antithrombotic treatment could be related to the outcome, we repeated the primary analyses using an intention-to-treat approach, where exposure is defined based on the drug used at cohort entry, and limiting follow-up to 6 months. Fourth, we excluded patients with a history of both VKA and DOAC use between nonvalvular atrial fibrillation diagnosis and cohort entry. Fifth, we repeated the primary analyses after accounting for competing risk due to death.
Finally, we censored on the first major bleeding event during follow-up, thus not allowing patients to contribute events to multiple types of bleeds. All analyses were conducted with SAS version 9.4 (SAS Institute, Cary, NC) and R (R Foundation for Statistical Computing, Vienna, Austria).
Results
A total of 5301 patients initiated concomitant DOAC-antiplatelet use and 9106 patients initiated concomitant VKA-antiplatelet use during the study period (Figure 1). The median duration of follow-up was 1.4 months for concomitant DOAC-antiplatelet use and 1.7 months for concomitant VKA-antiplatelet use (1.6 months overall), generating a total of 14,751 person-years. This follow-up period was primarily driven by discontinuation of antiplatelets (60%) (Supplementary Table 2, available online). During the follow-up period, 65 patients had an intracranial hemorrhage, 253 had a gastrointestinal bleeding, and 308 had other major bleeding episodes.
Figure 1Flowchart describing the construction of base and study cohorts.
Table 1 presents the characteristics of the nonvalvular atrial fibrillation patients on concomitant DOAC-antiplatelet use vs concomitant VKA-antiplatelet use at cohort entry. Patients on concomitant DOAC-antiplatelet use were less likely to have renal or vascular disease or congestive heart failure than patients on concomitant VKA-antiplatelet use, and they also had slightly lower HAS-BLED scores. In both groups, most patients were on single antiplatelet treatment (92% in the DOAC group vs 86% in the VKA group) and the most common antiplatelet was ASA (94% in the DOAC group vs 93% in the VKA group).
Table 1Baseline Demographics and Clinical Characteristics of the Cohort Stratified by Concomitant Use of Antithrombotic Drugs at Cohort Entry
Characteristic
Concomitant DOAC-Antiplatelet Use
Concomitant VKA-Antiplatelet Use
Total
5301
9106
Age, y (mean, SD)
75.5 (9.1)
78.0 (9.0)
Male, n (%)
2819 (53.2)
4709 (51.7)
Time from NVAF diagnosis to VKA/DOAC initiation, d (mean, SD)
77.0 (180.2)
38.6 (116.3)
History of alcohol-related disorders, n (%)
210 (4.0)
427 (4.7)
History of arterial hypertension, n (%)
4544 (85.7)
8337 (91.6)
History of cancer, n (%)
787 (14.9)
1399 (15.4)
History of congestive heart failure, n (%)
1043 (19.7)
2804 (30.8)
History of coronary artery disease, n (%)
2577 (48.6)
5588 (61.4)
History of diabetes mellitus, n (%)
1685 (31.8)
3577 (39.3)
History of liver disease, n (%)
269 (5.1)
676 (7.4)
History of peripheral vascular disease, n (%)
556 (10.5)
1718 (18.9)
Prior ischemic stroke or transient ischemic attack, n (%)
We calculated a CHA2DS2-VASc21 score and a modified HAS-BLED18 score for each patient, defining uncontrolled arterial hypertension as intake of two or more antihypertensive drugs in the year prior to cohort entry.
We calculated a CHA2DS2-VASc21 score and a modified HAS-BLED18 score for each patient, defining uncontrolled arterial hypertension as intake of two or more antihypertensive drugs in the year prior to cohort entry.
1.4 (0.8)
1.7 (1.0)
Number of antiplatelets at cohort entry, mean (SD)
1.1 (0.3)
1.1 (0.4)
One, n (%)
4886 (92.2)
7862 (86.3)
Two or more, n (%)
415 (7.8)
1244 (13.7)
Class of antiplatelets at cohort entry, n (%)
Acetylsalicylic acid
4972 (93.8)
8487 (93.2)
P2Y12 inhibitors
744 (14.0)
1863 (20.5)
H2 blockers, n (%)
85 (1.6)
185 (2.0)
Nonsteroidal anti-inflammatory drugs, n (%)
947 (17.9)
1527 (16.8)
Proton pump inhibitors, n (%)
2752 (51.9)
5577 (61.3)
Selective serotonin reuptake inhibitors, n (%)
457 (8.6)
901 (9.9)
Number of non-antithrombotic drugs, mean (SD)
11.5 (6.0)
13.7 (6.6)
CHA2DS2-VASc = congestive heart failure, arterial hypertension, age ≥75 years, diabetes mellitus, prior stroke or transient ischemic attack, vascular disease, age 65-74 years, female sex; DOAC = direct oral anticoagulant; HAS-BLED = abnormal renal or liver function, previous stroke or transient ischemic attack, bleeding history or predisposition, age >65 years, alcohol-related disorders or drugs; NVAF = nonvalvular atrial fibrillation; SD = standard deviation; VKA = vitamin K antagonist.
Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
Table 2 shows the results related to concomitant DOAC-antiplatelet use and the risk of the bleeding outcomes (18,21). Concomitant DOAC-antiplatelet use was associated with a similar risk of gastrointestinal bleeding (HR 1.08; 95% CI, 0.81-1.45), but with a decreased risk of intracranial hemorrhage (HR 0.46; 95% CI, 0.24-0.91) and other major bleeding (HR 0.68; 95% CI, 0.51-0.91), compared with concomitant VKA-antiplatelet use. Cumulative incidence curves for the bleeding outcomes are presented in Supplementary Figures 2-4, available online.
Table 2Crude and Adjusted Hazard Ratios for the Association Between the DOAC-Antiplatelet Concomitant Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation
Adjusted for disease risk score and calendar year. References: Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138:1093-1100. 18 Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137:263-272. 21
Intracranial hemorrhage
Current concomitant VKA-antiplatelet use
9106
54
3508
1.5 (1.2-2.0)
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
5301
11
1461
0.8 (0.4-1.4)
−0.3 (-1.6-1.1)
0.47 (0.25-0.91)
0.46 (0.24-0.91)
Gastrointestinal bleeding
Current concomitant VKA-antiplatelet use
9106
181
3462
5.2 (4.5-6.1)
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
5301
72
1448
5.0 (3.9-6.3)
−0.8 (-1.4–0.2)
0.87 (0.66-1.14)
1.08 (0.81-1.45)
Other major bleeding
Current concomitant VKA-antiplatelet use
9106
246
3421
7.2 (6.3-8.2)
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
5301
62
1450
4.3 (3.3-5.5)
−2.9 (-4.3–1.5)
0.56 (0.43-0.74)
0.68 (0.51-0.91)
CI = confidence interval; DOAC = direct oral anticoagulant; HR = hazard ratio; VKA = vitamin K antagonist.
Per 100 person-years.
† Adjusted for disease risk score and calendar year.References:Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138:1093-1100.
Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137:263-272.
Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
Chronic duodenal ulcer with hemorrhage/perforation
K270
Acute peptic ulcer with hemorrhage
K272
Acute peptic ulcer with hemorrhage/perforation
K274
Chronic peptic ulcer with hemorrhage
K276
Chronic peptic ulcer with hemorrhage/perforation
K280
Acute gastrojejunal ulcer with hemorrhage
K282
Acute gastrojejunal ulcer with hemorrhage/perforation
K284
Chronic gastrojejunal ulcer with hemorrhage
K286
Chronic gastrojejunal ulcer with hemorrhage/perforation
K290
Acute hemorrhagic gastritis
K2921
Alcoholic gastritis (with hemorrhage)
K2961
Other gastritis with bleeding
K2971
Gastritis (unspecified with bleeding)
K2991
Gastroduodenitis (unspecified with bleeding)
K2981
Duodenitis (with hemorrhage)
K3181
Angiodysplasia of stomach and duodenum with hemorrhage
K5521
Angiodysplasia of colon with hemorrhage
K5711
Diverticulosis of small intestine without perforation or abscess with bleeding
K5713
Diverticulitis of small intestine without perforation or abscess with bleeding
K5731
Diverticulosis of large intestine without perforation or abscess with bleeding
K5733
Diverticulitis of large intestine without perforation or abscess with bleeding
K625
Hemorrhage of anus and rectum
K661
Hemoperitoneum
Other
D683
Hemorrhagic disorder due to circulating anticoagulants
D699
Unspecified hemorrhagic conditions
H0523
Orbital hemorrhage
H113
Conjunctival hemorrhage
H313
Choroidal hemorrhage and rupture
H356
Retinal hemorrhage
H431
Vitreous hemorrhage
H44819
Hemophthalmos, unspecified eye
H450
Vitreous hemorrhage in diseases classified elsewhere
H922
Otorrhagia
I230
Hemopericardium as current complication following acute myocardial infarction
I312
Hemopericardium NEC
J942
Hemothorax
M250
Hemarthrosis
M7981
Nontraumatic hematoma of soft tissue
N02
Recurrent and persistent hematuria
N421
Congestion and hemorrhage of prostate
N836
Hematosalpinx
N837
Hematoma of broad ligament
N857
Hematometra
N897
Hematocolpos
N92
Excessive, frequent and irregular menstruation
N93
Other abnormal uterine and vaginal bleeding
N950
Postmenopausal bleeding
R04
Hemorrhage from respiratory passages
R31
Unspecified hematuria
R233
Spontaneous ecchymoses
R58
Hemorrhage NEC
T810
Hemorrhage and hematoma complicating a procedure NEC
T792
Traumatic secondary and recurrent hemorrhage and seroma, initial encounter
GIB = gastrointestinal bleeding; ICD = International Classification of Diseases; ICH = intracranial hemorrhage; NEC = not elsewhere classified NOS = not otherwise specified.
Drug-specific analyses revealed similar associations for the 3 DOACs, although these were based on fewer events (Supplementary Table 3, available online). We observed no duration–response relations for the bleeding outcomes (Supplementary Table 4, available online), and age, sex, or HAS-BLED score did not modify the associations (Supplementary Tables 5-7, available online). Concomitant DOAC-antiplatelet use was associated with a trend toward an increased risk of genitourinary bleeding compared with concomitant VKA-antiplatelet use (HR 1.42; 95% CI, 0.88-2.27), with the difference being driven by rivaroxaban and apixaban (Supplementary Table 8, available online). Comparing concomitant DOAC-ASA use vs concomitant VKA-ASA use, or concomitant use of a DOAC with at least 2 antiplatelet agents vs concomitant use of a VKA with at least 2 antiplatelet agents yielded results similar with those of the primary analysis (Supplementary Tables 9, 10, available online). Finally, we observed no effect modification for the bleeding outcomes after stratifying patients based on how they entered the cohort (Supplementary Table 11, available online). The results of the sensitivity analyses were consistent with those of the primary analysis (summarized in Figure 2 and presented in detail in Supplementary Tables 12-17, available online).
Supplementary Table 3Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Drug-Specific Analysis)
Supplementary Table 4Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Duration-Response Analysis)
Exposure
< 3 Months HR (95% CI)
3-5.9 Months HR (95% CI)
≥6 Months HR (95% CI)
P for Interaction
Intracranial hemorrhage
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.37 (0.15-0.92)
0.35 (0.04-2.82)
0.63 (0.22-1.87)
.73
Gastrointestinal bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
1.09 (0.75-1.57)
1.18 (0.61-2.29)
0.90 (0.48-1.70)
.83
Other major bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.67 (0.46-0.97)
0.84 (0.43-1.64)
0.62 (0.34-1.12)
.77
CI = confidence interval; DOAC = direct oral anticoagulant; HR = hazard ratio; VKA = vitamin K antagonist.
Supplementary Table 5Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Interaction with Age)
Exposure
≤75 Years of Age HR (95% CI)
>75 Years of Age HR (95% CI)
P for Interaction
Intracranial hemorrhage
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.23 (0.05-0.99)
0.63 (0.30-1.33)
.22
Gastrointestinal bleeding
Current concomitant VKA-antiplatelet use
1.00 [Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
1.32 (0.86-2.04)
0.94 (0.64-1.37)
.22
Other major bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.78 (0.52-1.18)
0.61 (0.41-0.91)
.38
CI = confidence interval; DOAC = direct oral anticoagulant; HR = hazard ratio; VKA = vitamin K antagonist.
Supplementary Table 6Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Interaction with Sex)
Exposure
Female HR (95% CI)
Male HR (95% CI)
P for Interaction
Intracranial hemorrhage
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.66 (0.24-1.80)
0.37 (0.15-0.88)
.38
Gastrointestinal bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
1.34 (0.88-2.03)
0.93 (0.63-1.36)
.19
Other major bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.53 (0.32-0.89)
0.77 (0.54-1.09)
.24
CI = confidence interval; DOAC = direct oral anticoagulant; HR = hazard ratio; VKA = vitamin K antagonist.
Supplementary Table 7Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Interaction with HAS-BLED
We used a modified HAS-BLED score including uncontrolled arterial hypertension, abnormal renal or liver function, previous stroke or transient ischemic attack, bleeding history or predisposition, age >65 years, and alcohol-related disorders, because all patients were exposed to antiplatelets, and international normalized ratio values were not available. We defined uncontrolled arterial hypertension as intake of ≥2 antihypertensive drugs in the year prior to cohort entry.
Score)
Exposure
HAS-BLED Score 0-2 HR (95% CI)
HAS-BLED Score 3-7 HR (95% CI)
P for Interaction
Intracranial hemorrhage
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.62 (0.31-1.24)
–
.98
Gastrointestinal bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
1.25 (0.91-1.73)
0.72 (0.36-1.46)
.16
Other major bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.68 (0.49-0.94)
0.79 (0.42-1.49)
0.69
CI = confidence interval; DOAC = direct oral anticoagulant; HAS-BLED = uncontrolled arterial hypertension, abnormal renal function, abnormal liver function, previous stroke, bleeding history or predisposition, labile international normalized ratio, age >65 years old, alcohol-related disorders or drugs; HR = hazard ratio; VKA = vitamin K antagonist.
We used a modified HAS-BLED score including uncontrolled arterial hypertension, abnormal renal or liver function, previous stroke or transient ischemic attack, bleeding history or predisposition, age >65 years, and alcohol-related disorders, because all patients were exposed to antiplatelets, and international normalized ratio values were not available. We defined uncontrolled arterial hypertension as intake of ≥2 antihypertensive drugs in the year prior to cohort entry.
Supplementary Table 8Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Genitourinary Bleeding Among Patients with Nonvalvular Atrial Fibrillation
Supplementary Table 9Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-ASA Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation
Supplementary Table 10Crude and Adjusted Hazard Ratios for the Association Between Concomitant Use of a DOAC with Multiple Antiplatelets and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation
Supplementary Table 11Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Interaction with the Way Patients Entered the Cohort)
Exposure
Addition of Antiplatelets HR (95% CI)
Addition of an Oral Anticoagulant HR (95% CI)
Concomitant Initiation of an Oral Anticoagulant and Antiplatelets HR (95% CI)
P for Interaction
Intracranial hemorrhage
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
0.61 (0.06-5.85)
0.46 (0.23-0.92)
–
.97
Gastrointestinal bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
1.66 (0.44-6.22)
1.03 (0.76-1.40)
1.65 (0.54-5.06)
.58
Other major bleeding
Current concomitant VKA-antiplatelet use
[Reference]
[Reference]
[Reference]
Current concomitant DOAC-antiplatelet use
1.96 (0.65-5.87)
0.63 (0.46-0.86)
0.78 (0.23-2.62)
.14
CI = confidence interval; DOAC = direct oral anticoagulant; HR = hazard ratio; VKA = vitamin K antagonist.
Figure 2Forest plot summarizing the results of the primary analysis and sensitivity analyses, showing adjusted hazard ratios and 95% confidence intervals for the association between DOAC-antiplatelet concomitant use and risk of the three bleeding outcomes.
Supplementary Figure 1Diagram depicting cohort entry with different patterns of initiation of concomitant use of antithrombotic drugs. The black line represents person-time prior to cohort entry spanning from the diagnosis of nonvalvular atrial fibrillation to the initiation of the first antithrombotic drug (either an antiplatelet or an oral anticoagulant). The blue line represents person-time prior to cohort entry exposed to an oral anticoagulant (either a vitamin K antagonist [VKA] or a direct oral anticoagulant [DOAC]). Red line represents person-time prior to cohort entry exposed to at least one antiplatelet. Cohort entry corresponded to the initiation of DOAC-antiplatelet or VKA-antiplatelet concomitant use. Red-and-blue line represents person-time after cohort entry. Scenario A corresponds to cohort entry upon addition of a DOAC or a VKA to at least one antiplatelet, scenario B corresponds to cohort entry upon addition of at least one antiplatelet to a DOAC or a VKA, and scenario C corresponds to cohort entry upon concomitant initiation of a DOAC or a VKA and at least one antiplatelet.
Supplementary Figure 2Cumulative incidence of intracranial hemorrhage among concomitant DOAC-antiplatelet and VKA-antiplatelet users. DOAC = direct oral anticoagulant; VKA = vitamin K antagonist.
Supplementary Figure 3Cumulative incidence of gastrointestinal bleeding among concomitant DOAC-antiplatelet and VKA-antiplatelet users. DOAC = direct oral anticoagulant; VKA = vitamin K antagonist.
Supplementary Figure 4Cumulative incidence of other major bleeding among concomitant DOAC-antiplatelet and VKA-antiplatelet users. DOAC = direct oral anticoagulant; VKA = vitamin K antagonist.
Supplementary Table 12Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (15-Day Grace Period)
Supplementary Table 13Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Stricter Outcome Definition)
Supplementary Table 14Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Intention-to-Treat)
Supplementary Table 15Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Excluding Patients with a Switch Between a VKA and a DOAC Prior to Cohort Entry)
Supplementary Table 16Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Competing Risk)
Supplementary Table 17Crude and Adjusted Hazard Ratios for the Association Between Concomitant DOAC-Antiplatelet Use and the Risk of Bleeding Outcomes Among Patients with Nonvalvular Atrial Fibrillation (Censoring on First Bleeding Event)
Our study assessed the risk of intracranial hemorrhage, gastrointestinal bleeding, and other major bleeding associated with concomitant DOAC-antiplatelet use in patients with nonvalvular atrial fibrillation. Compared with concomitant VKA-antiplatelet use, concomitant DOAC-antiplatelet use was associated with a similar risk of gastrointestinal bleeding (HR 1.08; 95% CI, 0.81-1.45), but with a decreased risk of intracranial hemorrhage (HR 0.46; 95% CI, 0.24-0.91) and other major bleeding (HR 0.68; 95% CI, 0.51-0.91). The findings of the primary analyses remained consistent in several sensitivity analyses.
Our results on intracranial hemorrhage are congruent with a recent meta-analysis of RCTs, which showed that in patients using ASA at baseline, DOACs were associated with a decreased risk compared with VKAs (HR 0.38; 95% CI, 0.26-0.56). Importantly, the incidence rates of intracranial hemorrhage in our study were approximately twofold higher than the ones reported in the trials comparing different DOACs with VKAs in monotherapy.
Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials.
This increase in the absolute risk of intracranial hemorrhage combined with the strong reduction in the relative risk associated with concomitant DOAC-antiplatelet use highlight the importance of choosing the appropriate oral anticoagulant in patients with nonvalvular atrial fibrillation requiring additional use of antiplatelets.
Our results on gastrointestinal bleeding showing a similar risk among concomitant DOAC-antiplatelet use and concomitant VKA-antiplatelet use are opposed to the previously reported increased risk with DOACs in monotherapy compared with VKAs in monotherapy.
Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials.
A possible explanation could be the preferential concomitant prescribing of VKAs and antiplatelets instead of DOACs and antiplatelets in high-risk patients, as the higher prevalence of proton-pump inhibitors in the VKA group alludes. However, the pharmacologic properties of the different antithrombotic medications could also account for this phenomenon.
Indeed, while bleeding episodes related to VKAs and antiplatelets usually occurs in the upper gastrointestinal tract, bleeding episodes related to dabigatran and rivaroxaban, the 2 most prevalent DOACs in our cohort, originate more often in the lower gastrointestinal tract.
Thus, the common localization of VKA and antiplatelet-related gastrointestinal bleeding could lead to a higher additive risk due to concomitant antiplatelet use with VKAs than with DOACs.
Recently, 2 RCTs compared the risk of major bleeding between concomitant DOAC-antiplatelet use and concomitant VKA-antiplatelet use,
ClinicalTrials.gov. NIH U.S. National Library of Medicine. A study of apixaban in patients with atrial fibrillation, not caused by a heart valve problem, who are at risk for thrombosis (blood clots) due to having had a recent coronary event, such as a heart attack or a procedure to open the vessels of the heart. Available at: https://clinicaltrials.gov/ct2/show/NCT02415400. Accessed April 9, 2018.
Although they have considerably expanded our knowledge about concomitant antithrombotic treatment, a direct comparison with our study is challenging. First, these trials were restricted to nonvalvular atrial fibrillation patients undergoing coronary intervention.
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.
Second, while trial participants randomized to VKAs received dual antiplatelet treatment, patients randomized to DOACs received either single antiplatelet treatment
Nevertheless, the lower risk of major bleeding among concomitant DOAC-antiplatelet use shown in the RCTs is congruent with our findings.
Our study has several strengths. First, the use of an as-treated exposure definition for both oral anticoagulants and antiplatelets allowed us to assess concomitant use with high precision. This is particularly important in drug–drug interaction studies aiming to investigate the effects associated with the concurrent use of 2 or more medications. Second, the population-based setting and the few exclusion criteria applied during study cohort assembly make the results of this study highly generalizable. Finally, we defined bleeding events based on related hospitalization or death likely maximizing the sensitivity and specificity of this outcome definition.
Our study also has some limitations. First, given its observational nature, residual confounding is possible. To mitigate this potential bias, we adjusted for DRS, including 18 clinically important risk factors. Of note, adjustment had only a minor effect on the point estimates of the bleeding outcomes. Second, the median follow-up in our study was short (1.6 months). However, this was expected given current recommendations limiting concomitant use of oral anticoagulants with antiplatelets to few months for most patients.
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.
Finally, our drug-specific analyses should be interpreted cautiously given the low number of exposed events.
In summary, the results of this population-based study indicate that compared with concomitant VKA-antiplatelet use, concomitant DOAC-antiplatelet use is associated with a similar risk of gastrointestinal bleeding, but a lower risk of intracranial hemorrhage and other major bleeding. These findings could help physician decision-making in patients requiring concomitant treatment with oral anticoagulants and antiplatelets.
Acknowledgments
AD is a recipient of a Research Fellowship from the German Research Foundation (Deutsche Forschungsgemeinschaft). SS is the recipient of the James McGill Professorship award. KBF and LA are recipients of a Chercheur-Boursier Award from the Fonds de recherche du Québec – Santé (FRQS) and William Dawson Scholar awards from McGill University.
References
Kirchhof P
Benussi S
Kotecha D
et al.
2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS.
Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials.
Use and associated risks of concomitant aspirin therapy with oral anticoagulation in patients with atrial fibrillation: insights from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) Registry.
High rates of concomitant antiplatelet use in patients with atrial fibrillation treated with oral anticoagulation: insights from the stroke prevention and rhythm interventions in atrial fibrillation (SPRINT-AF) registry.
Efficacy and safety of the use of non-vitamin K antagonist oral anticoagulants in patients with nonvalvular atrial fibrillation and concomitant aspirin therapy: a meta-analysis of randomized trials.
The use of prescription claims databases in pharmacoepidemiological research: the accuracy and comprehensiveness of the prescription claims database in Quebec.
Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
ClinicalTrials.gov. NIH U.S. National Library of Medicine. A study of apixaban in patients with atrial fibrillation, not caused by a heart valve problem, who are at risk for thrombosis (blood clots) due to having had a recent coronary event, such as a heart attack or a procedure to open the vessels of the heart. Available at: https://clinicaltrials.gov/ct2/show/NCT02415400. Accessed April 9, 2018.
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.
Funding: This work was supported, in part, by an infrastructure grant from the Canadian Foundation for Innovation (CFI), and the dataset was acquired thanks to unrestricted funding from Bayer Pharma AG. The sponsors had no influence on the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Conflict of Interest: SS has received research grants from Bayer Pharma AG, Boehringer Ingelheim, and Bristol-Myers-Squibb, and has participated in advisory board meetings or as speaker for AstraZeneca, Boehringer Ingelheim, and Novartis. AD, CR, HY, KBF, and LA have no conflicts of interest to disclose.
Authorship: LA had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Assessed 3-15 months prior to cohort entry.
30964-1&id=gr1.jpg){kind=link}
30964-1&id=gr2.jpg){kind=link}
30964-1&id=gr3.jpg){kind=link}
30964-1&id=gr4.jpg){kind=link}
30964-1&id=gr5.jpg){kind=link}
30964-1&id=gr6.jpg){kind=link}