Article Text

Download PDFPDF

Original research article
Use of oral anticoagulants in combination with antiplatelet(s) in atrial fibrillation
  1. Caroline Sindet-Pedersen1,2,3,
  2. Laila Staerk1,2,
  3. Morten Lamberts1,
  4. Thomas Alexander Gerds4,5,
  5. Jeffrey S Berger3,
  6. Anders Nissen Bonde1,2,
  7. Jannik Langtved Pallisgaard1,2,
  8. Morten Lock Hansen1,
  9. Christian Torp-Pedersen1,6,
  10. Gunnar H Gislason1,2,4,7,
  11. Jonas Bjerring Olesen1
  1. 1 Department of Cardiology, Copenhagen University Hospital Herlev and Gentofte, Hellerup, Denmark
  2. 2 Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark
  3. 3 School of Medicine, New York University, New York, USA
  4. 4 Department of Research, Danish Heart Foundation, Copenhagen, Denmark
  5. 5 Department of Public Health, Section of Biostatistics, University of Copenhagen, Copenhagen, Denmark
  6. 6 Department of Health, Science and Technology, Department of Epidemiology/Biostatistics and Cardiology, Aalborg University and Hospital, Aalborg, Denmark
  7. 7 National Institute of Public Health, University of Southern Denmark, Odense, Denmark
  1. Correspondence to Caroline Sindet-Pedersen, Department of Cardiology, Copenhagen University Hospital Herlev and Gentofte, Hellerup, Denmark; carolinesindet{at}gmail.com

Abstract

Objectives To investigate temporal trends in the use of non-vitamin K oral anticoagulants (NOACs) and vitamin K antagonists (VKAs) in combination with aspirin and/or clopidogrel in patients with atrial fibrillation (AF) following acute myocardial infarction (MI) and/or percutaneous coronary intervention (PCI).

Methods Using Danish nationwide registries, all patients with AF who survived 30 days after discharge from MI and/or PCI between 22 August 2011 and 30 September 2016 were identified.

Results A total of 2946 patients were included in the study population, of whom 1967 (66.8%) patients were treated with VKA in combination with antiplatelet(s) (VKA+aspirin n=477, VKA+clopidogrel n=439, VKA+aspirin+clopidogrel n=1051) and 979 (33.2%) patients were treated with NOAC in combination with antiplatelet(s) (NOAC+aspirin n=252, NOAC+clopidogrel n=218, NOAC+aspirin+clopidogrel n=509). The overall study population had a median age of 76 years [IQR: 69–82] and consisted of 1995 (67.7%) men. Patients with MI as inclusion event accounted for 1613 patients (54.8%). Patients with high CHA2DS2-VASc score(congestive heart failure, hypertension, age ≥75 years (2 points), diabetes mellitus, history of stroke/transient ischemic attack/systemic thromboembolism (2 points), vascular disease, age 65-75 years, and female sex) accounted for 132 2814 (95.5%) of patients, and patients with high HAS-BLED score (hypertension, abnormal renal/liver function, history of stroke, history of bleeding, age >65 years, non-steroidal anti-inflammatory drug usages, or alcohol abuse, leaving out labile international normalized ratio (not available), and use of antiplatelets (exposure variable)) accounted for 934 (31.7%) of patients. There was an increase from 10% in 2011 to 52% in 2016 in the use of NOACs in combination with antiplatelet(s).

Conclusion From 2011 to 2016, the use of NOAC in combination with antiplatelet(s) increased in patients with AF following MI/PCI and exceeded the use of VKA in combination with antiplatelet(s) by 2016.

  • atrial fibrillation
  • acute myocardial infarction
  • percutaneous coronary intervention
  • drug monitoring
View Full Text

Statistics from Altmetric.com

Introduction

Patients with atrial fibrillation (AF) presenting with acute myocardial infarction (MI) and/or undergoing percutaneous coronary intervention (PCI) poses a great challenge with regard to the management of antithrombotic therapy.1 2 These patients are at a high risk of ischaemic events, such as stroke, recurrent MI and stent thrombosis.1 2 Antithrombotic treatment with oral anticoagulants (OACs), such as vitamin K antagonists (VKAs) or one of the non-vitamin K antagonist oral anticoagulants (NOACs), is a cornerstone in the management of patients with AF at risk of stroke.1 3 Substantial changes in stroke prophylaxis have been observed in patients with AF since the introduction of NOACs.1 4–7 When patients with AF present with MI and/or undergo PCI, they have indication for dual or triple therapy with aspirin and/or clopidogrel on top of their OACs.1 2 The introduction of the NOACs has added to the complexity on treatment decision in these patients, as a multiplicity of treatment combinations are now present.4 The clinical decision making on the use of NOACs in combination with aspirin, clopidogrel or both in the setting of AF and MI/PCI is based on either consensus documents or expert opinions (level C evidence).1 8 9 Considerable knowledge gaps exist about how NOACs in combination with antiplatelets are being used in the management of patients with AF presenting with MI or undergoing PCI. Thus the aims of the study were (1) to investigate initiation patterns of dual or triple therapy with NOACs compared with VKA in patients with AF following MI, PCI or both; (2) to investigate patient characteristics associated with initiation of NOACs compared with VKAs; and (3) to investigate to which degree standard and reduced dosage NOACs are used.

Methods

Databases

Data from the Danish nationwide administrative registries were used, which included the Danish National Prescription Registry, the Danish National Patient Register and the Danish Civil Registration System.10–12 From these registries information on all dispensed prescription medicines from pharmacies, hospital contacts and date of birth can be obtained.11–14 A personal unique number is provided to every Danish resident, at birth or immigration, and this number enables the linkage of the individual registries.13 The Danish National Patient Registry holds information on all hospital contacts and diagnoses from 1978 (inpatient and outpatient), including emergency department visits registered since 1995.12 All diagnoses are registered according to the International Classification of Diseases (ICD) and are classified as a primary diagnosis, and if needed also as secondary diagnoses.12 From 1995 information on all dispensed prescription drugs has been available through the Danish National Prescription Registry.11 Information on date of filled prescription, strength and package size can be found in this registry.11 15 Information on date of birth, date of death and gender is available through the Danish Civil Registration System.

Study population

From the Danish National Patient Register, we identified patients with AF with an MI or PCI between 22 August 2011 and 30 September 2016. Patients were eligible for inclusion if they had not had an MI or undergone PCI within 1 year before the index date, defined as the date of admission with an MI, PCI or both. We identified patients treated with available NOACs (rivaroxaban approved 6 February 2012, dabigatran approved 22 August 2011, apixaban approved 10 December 2012) or VKAs, and concomitantly treated with aspirin, clopidogrel or both. To allow patients to claim a prescription from the pharmacies after a discharge with MI, PCI or both, we included patients who were in treatment 30 days after discharge. This means that day 30 after discharge was considered as baseline. Patients were excluded if they died within 30 days after an MI, PCI or both, were aged below 30 or above 100 years, had valvular disease (mechanical heart valves or rheumatic heart disease), or were not on oral anticoagulation in combination with aspirin and/or clopidogrel.

Exposure to treatment regimens

The exposure to any of the treatment regimens was identified through the National Prescription Registry, where treatment regimens were classified into six categories based on filled prescription from the pharmacies. These included VKA+aspirin, NOAC+aspirin, VKA+clopidogrel, NOAC+clopidogrel, VKA+aspirin+clopidogrel, and NOAC+aspirin+clopidogrel. The exposure to any of the treatment regimens was calculated using an algorithm, which has been described in detail elsewhere.8 15 Briefly, based on filled prescriptions, we estimated the exposure to any of the treatment regimens at any given time point.8 15 The algorithm uses the date of filled prescriptions, number of tablets and strength to estimate the exposure.8 15 VKAs have individual dosing, whereas NOACs are fixed dose regimens. Standard dosages of NOACs and antiplatelets were defined as rivaroxaban 20 mg once daily, dabigatran 150 mg twice daily, apixaban 5 mg twice daily, aspirin 75 mg once daily and clopidogrel 75 mg once daily. Reduced dosages of NOACs were defined as rivaroxaban 15 mg once daily, dabigatran 110 mg twice daily and apixaban 2.5 mg twice daily. Treatment groups represented are the treatment group that the patients were in exactly 30 days after the index MI, PCI or both.

Comorbidities and concomitant medication

Comorbidities registered within 10 years of baseline were identified using the ICD-10 diagnosis codes (online supplementary table 1). Concomitant medication was identified from claimed prescriptions up to 180 days prior to baseline and it was identified using ATC codes (online supplementary table 1). The CHA2DS2-VASc score was used to examine the risk of stroke (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, history of stroke/transient ischaemic attack/systemic thromboembolism, vascular disease, age 65–75 years and female sex), and the HAS-BLED score was used to examine the risk of bleeding (hypertension, abnormal renal/liver function, history of stroke, history of bleeding, >65 years of age, non-steroidal anti-inflammatory drugs or alcohol abuse, not including labile internationally normalised ratio (INR) (not available) and use of antiplatelets (explanatory variable)).

Supplementary file 1

Statistical analyses

Baseline characteristics according to the six different treatment regimens are presented as median with IQR for continuous variables and count with percentages for categorical variables. The distribution of the included patients across the six treatment regimens was calculated separately for each calendar year and further within subgroups defined by CHA2DS2-VASc score (low-intermediate (0–2), high (>2)), HAS-BLED score (low-intermediate (0–2), high (>2)), age groups (<65, 65–64, >75), gender, previous stroke and previous bleeding. Calendar time trends of treatment preference were tested using multinomial logistic regression (outcome: six treatment groups) and reported as p values corresponding to likelihood ratio tests for a model adjusted for calendar year against a model without covariates. Differences in calendar time trends of treatment preference were tested between subgroups using multinomial logistic regression adjusted for calendar year, and subgroups, and was reported as p values corresponding to likelihood ratio test (with vs without interaction term). Using only data from 2016, the χ2 test was used to investigate if treatment preference depended on any of the subgroups.

Results

In the period between 22 August 2011 and 30 September 2016, a total of 2946 patients with AF with MI, PCI or both were included (figure 1).

Figure 1

Selection process. Flow chart over selection of study population between August 2011 and September 2016. Valvular disease was defined as mechanical heart valves or rheumatic heart disease. No patients were treated with a combination of a NOAC and ticagrelor or prasugrel. MI, myocardial infarction; NOAC, non-vitamin K oral anticoagulant; PCI, percutaneous coronary intervention; VKA, vitamin K antagonist.

Table 1 shows a summary of the baseline characteristics of the overall study population and according to the different treatment regimens (online supplementary figures S1 and S2).

The study population had a median age of 76 years [IQR: 69–82) and consisted of 1995 (67.7%) men. Patients with MI as the inclusion event accounted for 1613 patients (54.8%), of whom 228 (14.1%) had a PCI within 1 day following hospital admission. The PCI (and no MI) group consisted of 1333 (45.2%) patients, with 1000 (75.0%) patients having a stent implanted. The mean CHA2DS2-VASc score was 4.23 (SD: 1.65), and the mean HAS-BLED score was 2.15 (SD: 0.94).

Table 1

Baseline characteristics

Among patients continuing with NOAC treatment following MI and/or PCI (n=528), we assessed the distribution of NOAC dosages in combination with antiplatelet(s). We found that the majority of patients were on standard-dosage NOAC before MI and/or PCI (60.0%), whereas the majority (64.1%) was on a reduced dosage when antiplatelet(s) was added.

Temporal trends

Figure 2 shows the temporal trends in the initiation of dual therapy with an OAC+single antiplatelet versus triple therapy with an OAC+dual antiplatelet therapy of included patients. The use of triple therapy increased during the study period (43% in 2011 to 60% in 2016).

Figure 2

Treatment initiation of dual versus triple therapy. Treatment initiation according to treatment OAC+single antiplatelet therapy and OAC+dual antiplatelet therapy between August 2011 and September 2016. OAC, oral anticoagulant.

Figure 3 shows the temporal trends in initiation according to the six different treatment regimens in patients with MI, PCI or both of included patients. During the study period from 2011 to 2016, an increase in the initiation of NOACs in combination with antiplatelet(s) was observed (10% in 2011 to 52% in 2016), which by 2016 was higher than the initiation of VKA in combination with antiplatelet(s). The increase in the use of NOACs in combination with antiplatelet(s) was mainly driven by anincrease in the use of NOAC+clopidogrel+ aspirin, which increased from 4% in 2011 to 32% in 2016, and a more modest increase in the use of NOAC+clopidogrel (1% in 2011 to 11% in 2016).

Figure 3

Treatment initiation according to treatment groups between August 2011 and September 2016. NOAC, non-vitamin K oral anticoagulant; VKA, vitamin K antagonist.

When dividing patients according to inclusion event (MI or PCI), similar trends were observed (figure 4). It was interesting to note that more patients undergoing PCI were initiated on triple therapy (35% VKA, 39% NOAC in 2016), as compared with patients with MI (23% VKA, 25% NOAC in 2016).

Figure 4

Treatment initiation according to inclusion event. Treatment initiation according to treatment groups between August 2011 and September, divided according to inclusion event. AF, atrial fibrillation; MI, myocardial infarction; NOAC, non-vitamin K oral anticoagulant; PCI, percutaneous coronary intervention; VKA, vitamin K antagonist.

Changes in treatment preference

Treatment preference changed significantly between 2011 and 2016 (p value <0.001). However, calendar time changes in treatment preference were not significantly different in subgroups defined by CHA2DS2-VASc score (p value=0.50) and HAS-BLED score (p value=0.11) (figure 5).

Figure 5

Treatment preference within subgroups of CHA2DS2-VASc (congestive heart failure, hypertension, age ≥75 years (2 points), diabetes mellitus, history of stroke/transient ischemic attack/systemic thromboembolism (2 points), vascular disease, age 65-75 years, and female sex) and HAS-BLED scores (hypertension, abnormal renal/liver function, history of stroke, history of bleeding, age >65 years, non-steroidal anti-inflammatory drug usages, or alcohol abuse, leaving out labile international normalized ratio (due to availability), and use of antiplaelets (exposure variable)). Treatment preference in per cent according to CHA2DS2-VASc and HAS-BLED scores between August 2011 and September 2016. NOAC, non-vitamin K oral anticoagulant; VKA, vitamin K antagonist.

In 2016, the preferred treatment for patients with high CHA2DS2-VASc score was NOAC+aspirin+clopidogrel (34%), followed by VKA+aspirin+clopidogrel (30%). This pattern was similar and not significantly different (p=0.23) in patients with intermediate CHA2DS2-VASc score, where NOAC+aspirin+clopidogrel (34%) and VKA+aspirin+clopidogrel (28%) were the preferred treatments.

In 2016, the preferred treatment for patients with high HAS-BLED score was VKA+aspirin+clopidogrel (33%), followed by NOAC+aspirin+clopidogrel (30%). This was opposite, but not significantly different (p value=0.57), in patients with low to intermediate HAS-BLED score, where the preferred treatment was NOAC+aspirin+clopidogrel (34%), followed by VKA+aspirin+clopidogrel (28%).

It was found that treatment preference depended on sex (p value=0.008) (online supplementary figure S3) in 2016. The preferred treatment in women was NOAC+aspirin+clopidogrel (35%), followed by VKA+aspirin+clopidogrel (21%). The percentage of patients initiating VKA+aspirin+clopidogrel was higher in men (34%) than in women; however, most male patients were still initiated on NOAC+aspirin+clopidogrel (35%).

Treatment preference did not depend significantly on age in 2016 (p value=0.63) (figure 6). Noteworthy, a total of 71% of patients under the age of 65 were initiated on triple therapy, whereas the number was lower for patients over the age of 85, where 56% were initiated on triple therapy in 2016.

Figure 6

Treatment preference within subgroups of age. Treatment preference in per cent according to age groups (<65, 65–74, >85) between August 2011 and September 2016. NOAC, non-vitamin K oral anticoagulant; VKA, vitamin K antagonist.

Results regarding the changes in treatment preference according to previous bleeding and stroke can be found in online supplementary material; the choice of treatment did not depend significantly on whether the patient had a previous stroke (p value=0.26) or bleeding (p value=0.80) (online supplementary figures S4 and S5).

Discussion

This study investigated the temporal trends in the initiation of NOACs and VKA in combination with aspirin, clopidogrel or both. The main findings were that the use of NOACs in combination with antiplatelet(s) increased between 2011 and 2016. The increasing use of NOACs in combination with antiplatelet(s) was mainly driven by an increase in the use of NOAC+aspirin+clopidogrel, followed by a modest increase in the use of NOAC+clopidogrel. In 2016, treatment preference did not depend significantly on predicted stroke or bleeding risk, nor was age found to have a significant influence. However, sex was found to influence the choice of treatment by the end of the study period.

We found that by 2016, the use of NOACs in any combination with antiplatelet(s) was exceeding that of VKA in combination with antiplatelet(s), which shows that physicians are widely using NOACs in combination with antiplatelet(s). Although not entirely comparable, previous studies based on the same registries have shown that the use of NOACs is rapidly increasing in patients with AF.6 7 The increased use of NOACs found in this study and previous studies is likely explained by several factors including guideline recommendations and the convenience of NOACs, for example, no frequent blood monitoring, no food restriction and fixed dose regimen. The finding that treatment preference for any of the treatment groups did not depend on either CHA2DS2-VASc or HAS-BLED score by the end of the study period and only depended on sex suggests that other factors, not included in the registries, may be associated with the choice of treatment.

The 2016 European Society of Cardiology (ESC) guidelines on the management of AF now recommend NOAC over VKA for stroke prophylaxis in patients with AF, unless contraindicated (level A evidence).3 However, current guidelines do not differentiate between NOAC and VKA in patients with AF with acute coronary syndrome (ACS) or PCI (level C evidence).1 3 The guideline recommends initial treatment with triple therapy for at least 4 weeks and for the shortest duration as possible in these patients. We found an increased use of triple therapy during the study period, which is in accordance with guidelines; however, we did not examine the length of triple therapy, as this is out of scope for this paper.

Limited evidence exists to support the clinical decision making on the choice of oral anticoagulation therapy in this setting, and treatment choice is mainly driven by the convenience of the NOACs, despite the fact that good evidence and experience exist with the use of VKA.1 3 16 NOACs have been found to have a favourable safety profile when compared with VKA in patients with AF. Yet it is difficult to assess if this benefit is still present when NOACs are combined with antiplatelet(s) in the setting of MI/PCI and AF.17–19

The two APRAISE-trials (Apixaban for Prevention of Acute Ischemic and Safety Events), where apixaban in combination with aspirin or clopidogrel was examined in patients with ACS, were terminated early due to excess bleeding in the placebo group.20 21 The two ATLAS trials (Anti-Xa Therapy to Lower cardiovascular events in addition to Aspirin with or without thienopyridine therapy in Subjects with Acute Coronary Syndrome), where rivaroxaban in combination with aspirin or clopidogrel was examined in patients with ACS, showed a benefit when rivaroxaban was used in very low dosage (2.5 mg).22 23 Post-hoc analyses from randomised clinical trials showed that the relative safety and efficacy of apixaban, dabigatran or rivaroxaban compared with VKA for the prevention of stroke/systemic embolism in patients with AF were not affected by concomitant use of antiplatelets.16–19 Furthermore, the PIONEER-AF-PCI study (Open-Label, Randomized, Controlled, Multicenter Study Exploring Two Treatment Strategies of Rivaroxaban and a Dose-Adjusted Oral Vitamin K Antagonist Treatment Strategy in Subjects with Atrial Fibrillation who Undergo Percutaneous Coronary Intervention) found that 15 mg rivaroxaban once daily in combination with single antiplatelet therapy, or 2.5 mg rivaroxaban once daily in combination with dual antiplatelet therapy, was associated with a significantly lower risk of clinically significant bleeding compared with standard therapy VKA in combination with dual antiplatelet therapy in patients with AF undergoing PCI.24 However, the study was not powered for efficacy endpoints. Two other studies investigating the safety and efficacy of dabigatran (REDUAL-PCI (Randomized Evaluation of Dual Therapy With Dabigatran versus Triple Therapy With Warfarin in Patients With Nonvalvular Atrial Fibrillation undergoing Percutanous Coronary Intervention)) and apixaban (AUGUSTUS (Apixaban Versus Warfarin in Patients with AF and ACS or PCI)) compared with VKA are ongoing. These studies will be important contributions to the growing evidence on the safety and efficacy of NOACs in combination with antiplatelet(s), which will evidently make the clinical decision making more evidence-based.16 25

To minimise bleeding complications, the 2016 ESC guidelines on the management of AF suggest the lowest effective dose of NOAC for stroke prophylaxis in AF to be used in combination with antiplatelet(s) (level C evidence).1 3 Our analyses showed that the majority of patients were initiated on a reduced dosage NOAC after discharge with MI, PCI or both. In addition, we also found that the majority of patients who were in treatment with a NOAC before admission and continued treatment with a NOAC after discharge were also initiated on a reduced dosage when an antiplatelet was added. However, it is unknown if low-dose NOAC is effective in preventing strokes, and further research is needed to establish this.

Strengths and limitations

The main strength of this study was the possibility of describing temporal trends of NOACs and VKA in combination with aspirin, clopidogrel or both in real-life patients, regardless of socioeconomic status, geography or participation in a healthcare insurance programme.12 26 Second, the Danish nationwide registries have been associated with high internal validity, including the diagnoses of AF and MI.11 12 26–29 A number of limitations exist, which include the lack of important clinical variables such as labile INR, body mass index, smoking and alcohol consumption. Furthermore, no data exist with regard to the type of stent used (bare-metal stents, drug-eluting stents), or whether the patients underwent multivessel or single-vessel PCI. Aspirin can be purchased both over the counter and on prescription, thus introducing misclassification bias.30 It has been found that the potential for identifying use of aspirin in the registries is high, which can partly be explained from the fact that there is an economic incentive to purchase the drug on prescription because of the reimbursement system on prescription drugs in Denmark.30 Because of a limited number of patients, it was not possible to investigate the temporal trends associated with the individual NOACs. Furthermore, it is unknown if results are generalisable to other countries or ethnicities.

Conclusion

In the setting of patients with AF with MI, PCI or both, the use of NOACs in combination with antiplatelet(s) is increasing, and by 2016 has exceeded that of VKA in combination with antiplatelet(s). It was found that in 2016 the treatment preference did not depend significantly on predicted stroke or bleeding risk and that sex was the most important factor with regard to treatment preference. Future studies are needed to clarify the relative safety and efficacy of NOAC in combination with antiplatelet(s) in comparison to VKA in combination with antiplatelet(s).

Key messages

What is already known on this subject?

Patients with atrial fibrillation (AF) who present with myocardial infarction (MI) and/or undergo percutaneous coronary intervention (PCI) have indication for dual or triple therapy with aspirin and/or clopidogrel on top of their oral anticoagulant with vitamin K antagonists (VKA) or non-vitamin K oral anticoagulants (NOAC).

What might this study add?

The use of NOAC in combination with antiplatelet(s) increased from 2011 to 2016 in patients with AF following MI and/or PCI, and exceeded the use of VKA in combination with antiplatelet(s) by 2016.

How might this impact on clinical practice?

The increased use of NOACs in combination with antiplatelet(s) is noteworthy. More evidence is needed regarding the comparative safety and effectiveness of VKA and NOAC in combination with antiplatelet(s).

References

View Abstract

Footnotes

  • Contributors All authors contributed significantly to the conception and design of the study, acquisition of data, or analysis and interpretation of the data. CS-P drafted the manuscript, and JBO, LS, ML, ANB, JSB, JLP, CT-P, MLH, TAG and GHG interpreted the data and critically revised the manuscript for intellectual content. All authors gave final approval of the version for publication.

  • Funding This work was supported by the Danish Council of Independent Research (grant no: 4183-00008). GHG is supported by an unrestricted clinical research scholarship from the Novo Nordisk Foundation. The sponsor had no influence on the study design, interpretation of results or the decision to submit the manuscript for publication.

  • Competing interests CS-P has nothing to declare. TAG has nothing to declare. LS has received funding for research from Boehringer Ingelheim. JLP has received funding for research from Boehringer Ingelheim and Bayer. CT-P declares research contracts with Bayer and Biotronic and speaker fees for Bayer and BMS. GHG has received funding for research from Boehringer Ingelheim, Pfizer, BMS, AstraZeneca and Bayer, and declares ownership of stocks in Lundbeck A/S, Novo Nordisk A/S and ALK Abello Pharmaceuticals. JBO has received speaker fees from Bristol-Myers Squibb, Boehringer Ingelheim, Bayer and AstraZeneca, previous funding for research from the Lundbeck Foundation, and current funding for research from the Bristol-Myers Squibb and the Capital Region of Denmark, Foundation for Health Research. ML has received speaker fees from Bristol-Meyer Squibb and Bayer. JSB has received research support from AstraZeneca and is a consultant for Merck, AstraZeneca and Janssen. ANB has nothing to declare. MLH has received speaker fees from Bristol-Meyer Squibb.

  • Ethics approval Retrospective registry-based studies do not require approval from the Research Ethics Committee System. The Danish Data Protection Agency had approved use of data for this study (ref no: 2007-58–0015/GEH-2014–012 I-Suite no: 02720.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles