Article Text
PDF
Abstract
Purpose Combined oral anticoagulant (OAC) and antiplatelet (AP) therapy is generally discouraged in atrial fibrillation (AF) outside of acute coronary syndromes or stenting because of increased bleeding. We evaluated its frequency and possible reasons in a contemporary European AF population.
Methods The PREvention oF thromboembolic events–European Registry in Atrial Fibrillation (PREFER in AF) prospectively enrolled AF patients in France, Germany, Austria, Switzerland, Italy, Spain and the UK from January 2012 to January 2013. We evaluated patterns of combined VKA-AP therapy in this population.
Results Out of 7243 patients enrolled, 5170 (71.4%) were treated with OAC alone, 808 (11.2%) with AP alone and 791 (10.9%) with a combination of OAC and one (dual) or two AP (triple combination therapy). Compared with patients only prescribed OAC, patients on combination treatment had similar Body Mass Index, but more frequently diabetes (p<0.05), dyslipidaemia (p<0.01), coronary heart disease (54.2 vs 18.6%; p<0.01) or peripheral arterial disease (10.2 vs 3.7%; p<0.01). Accordingly, they had a higher mean CHA2DS2VASc (3.7 vs 3.4), and HAS-BLED (2.7 vs 1.9) scores (for both, p<0.01).
Of the 660 patients on dual AP+OAC combination therapy, 629 (95.3%) did not have an accepted indication. Out of the 105 patients receiving triple combination therapy, 67 (63.8%) did not have an accepted indication.
Conclusions The combined use of OAC and AP therapy is not uncommon in AF, largely inappropriate, explained by the coexistence of coronary or peripheral arterial disease, and not influenced by considerations on the risk of bleeding.
Statistics from Altmetric.com
Introduction
Atrial fibrillation (AF) and coronary heart disease (CHD) are common and important causes of morbidity and mortality,1 ,2 sharing many risk factors,1 and frequently coexisting. It has indeed been estimated that 20% of the AF population has concomitant CHD.3 Atrial fibrillation also frequently coexists with cerebrovascular disease4 and peripheral arterial disease (PAD),4 other manifestations of systemic atherosclerosis.
Most patients with AF, with or without coexisting CHD are in need of oral anticoagulant (OAC) therapy with either a vitamin K antagonist (VKA) or one of the non-vitamin K antagonist oral anticoagulants (NOAC) to prevent AF-related strokes.5–7 On the other hand, the main antithrombotic prophylaxis in patients with evidence of vascular disease is antiplatelet therapy. This is carried on, in the vast majority of cases, with aspirin alone in stable patients; or a combination of aspirin with an ADP P2Y12 inhibitor—mostly clopidogrel—in the first months after an acute event or a stent implantation.8 ,9 However, the optimal antithrombotic prophylaxis when AF and vascular disease coexist is still a matter of discussion. Dual antiplatelet therapy is considered appropriate to prevent stent thrombosis after a recent stent implantation (usually for 1 month after implantation of a bare-metal stent (BMS), and 3–6 months after a drug-eluting stent).10 Furthermore, dual antiplatelet therapy can reduce recurrent ischaemic events after an acute coronary syndrome (ACS)11 ,12 and prevent stent thrombosis after stenting better than VKAs.13 Conversely, VKAs are superior to dual antiplatelet therapy to prevent stroke in AF patients.14 In case of coexisting AF and stable vascular disease (CHD, cerebrovascular disease and PAD) in the absence of a recent stenting and a recent ACS, the current ESC Guidelines in AF1 ,7 recommend anticoagulant therapy only, as this proved effective in assuring secondary prophylaxis from vascular events,15 and as the combination of an anticoagulant (VKA in most cases) with aspirin considerably increases the risk of bleeding with uncertain efficacy/safety advantages.16
Here, we analysed the baseline data from a large, contemporary European registry of AF patients (a) to describe the patterns of antithrombotic treatments in patients with AF and (b) to assess suspected reasons for the pattern of combined OAC and antiplatelet drug treatment, and its appropriateness according to ESC guidelines in five representative European regions in such settings.
Methods
Source of data
The Prevention of thromboembolic events–European Registry in Atrial Fibrillation (PREFER in AF) registry was designed as a prospective observational study with a baseline visit at the time of patient enrolment (cross-sectional part) and a 1 year follow-up visit (prospective part). This manuscript refers to the results of the baseline data, which were collected from patients in seven representative European countries (Austria, France, Germany, Italy, Spain, Switzerland, UK—for regional comparisons, not presented here—Austria, Switzerland and Germany were combined into one prespecified region), close to ‘real-life’. Furthermore, consecutive patients were included at each site in order to reduce selection bias. All data were captured through an electronic case report form including a wide range of plausibility checks for the entered variables. Additionally, on-site source data verification was done in approximately 5% of the sites. The study management was executed by Daiichi Sankyo Europe GmbH, Munich, Germany, as sponsor via a contract research organisation (SSS International Clinical Research GmbH, Munich). The study management was overseen by a scientific Steering Committee. Bleeding data collected were all ‘historical data’, that is, collecting documented events that occurred within 12 months prior to baseline. More details of the registry procedures and of the baseline patient characteristics have been described in the online supplementary material and elsewhere.17
Data processing
Patients were grouped into those treated with OAC therapy alone (either a VKA or one of the NOAC), OAC plus AP monotherapy, and OAC plus dual antiplatelet therapy. Clinical correlates for such categories were described.
Patients receiving combination therapy were classified as ‘likely appropriate’ and as ‘likely inappropriate’ (in short: ‘inappropriate’) by the following definitions: for patients on OAC therapy plus one single AP treatment, inappropriateness was defined as the use of combination treatment in the absence of an ACS or any percutaneous coronary intervention (PCI) in the year prior to enrolment (see online supplementary figure 1).
For patients treated with an OAC plus dual antiplatelet therapy, the criterion for inappropriateness included the absence of a BMS implantation ≤1 month, OR the absence of a drug eluting stent (DES) implantation ≤1 year, OR the absence of an ACS ≤1 year (see online supplementary figure 2).
Since most recent advice for the optimal duration of dual antiplatelet therapy after DES implantation or an ACS in the presence of therapy with a VKA now restrict the combination of an OAC and dual antiplatelet therapy to no more than 6 months,18 a sensitivity analysis was also conducted with such more stringent criteria. The aim of this analysis was not to have a clear-cut definition of the appropriate use, but rather to have a clear idea of the certainly inappropriate use, with the awareness of a possible underestimation of the problem by such criteria. As the number of patients receiving NOAC therapy was small, we performed another sensitivity analysis with patients receiving VKA, excluding NOAC-treated patients. Since there is now an overall trend to drop the aspirin component of the triple therapy when, in principle, triple therapy is recommended from Guidelines, we deliberately did not want to address the problem of possibly inappropriate undertreatment, susceptible to changes in recommendations as soon as more data will be made available in the subsequent months or years. Our estimate of inappropriateness is therefore to be intended as confined with the excessive use of antiplatelet drugs in combination with OAC.
Statistical analysis
Variables collected in the electronic case report form (eCRF) at baseline and derived parameters were used. For the analysis of the baseline data, only patients fulfilling the inclusion criteria were taken into account. Binary, categorical and ordinal parameters were summarised by means of absolute and percentage numbers within the various categories. Numerical data were summarised by means of standard statistics (ie, number of available data, number of missing data, mean, SD, minimum, median, maximum, lower and upper quartiles).
Comparisons between groups were performed using Student t test for quantitative data, and using the χ2 test for qualitative and ordinal data. Analysis of variance (ANOVA) was used for comparisons of >2 variables. All p values were interpreted in an exploratory way and with caution.
Statistical analyses were performed using the SAS V.9.2 software. For all analyses, the term, ‘Germany’, includes data from Austria and Switzerland.
Results
Characteristics of the whole study population enrolled in PREFER for this analysis
Out of 7243 patients enrolled, 5170 (71.4%) were treated with an OAC alone, 808 (11.2%) with an AP alone and 791 (10.9%) with a combination of an OAC and one or two AP. Six hundred and sixty (9.1%) patients received OAC and a single AP (ASA or clopidogrel), whereas 105 (1.4%) patients received OAC and a dual AP (ASA and clopidogrel) (figure 1).
Assessment of appropriateness/inappropriateness of the use of dual therapy (OAC and ASA or clopidogrel), and of triple therapy (OAC plus ASA plus clopidogrel) in AF patients in the PREFER in AF Registry (criteria as defined in online supplementary figures 1 and 2). OAC: oral anticoagulants; AP: antiplatelet agent; ASA: acetylsalicylic acid (aspirin).
A VKA (warfarin, phenprocoumon, acenocoumarol, fluindione) as monotherapy or in-combination therapy was by far the most common anticoagulant used (5519, 92.6% of anticoagulated patients). Only 442 (7.4% of anticoagulated patients) received a NOAC.
Dual therapy with a VKA and a single AP (ASA or clopidogrel) was given in 589 patients (8.1%), and triple therapy (VKA+ASA+clopidogrel) in 94 patients (1.3%). Dual therapy with a NOAC and a single AP (ASA or clopidogrel) was given in 34 patients (0.5%), and triple therapy (a NOAC+ASA+clopidogrel) in 3 patients (0.04%). Antithrombotic therapy was given by cardiologists in the largest percentage (75%), followed by general medicine/family doctors (3%) and coagulation specialists (0.2%) for dual therapy, and in even larger percentages by cardiologists for triple therapy, overall reflecting the enrolment criteria of the PREFER in AF registry. We found no differences in the percentage of appropriateness among these categories.
Details of the categories described above, according to the use of various drugs, are presented in table 1. Data are presented here for the OAC category as a whole.
Descriptive characteristics of the study population
Since data with VKA are, by far, prevailing, additional data on VKA alone or in combination with AP are presented in the online supplementary tables 1–6.
Since the current analysis covers patient data from the time period of January 2012 to January 2013, a rather low number of patients were treated with NOACs (442 patients on monotherapy or any combination therapy, 34 patients on dual, 3 patients on triple therapy) due to a limited market penetration in the recruitment window. No analysis of the appropriate/inappropriate use of the combination treatments was done on this, as such numbers are too low to draw reasonable conclusions.
Comparative characteristics of the populations treated with OAC, with AP and with OAC+AP
Compared with patients only prescribed an OAC, patients on combination treatment had similar Body Mass Index and prevalence of hypertension, but were slightly younger, and more often male, previous smokers and dyslipidemic. Additionally, they more often had valvular heart disease, CHD, coronary stenting, PAD and a history of a previous ischaemic event. Accordingly, patients on combination therapy had slightly higher CHA2DS2VASc (3.7 vs 3.4), and HAS-BLED scores (2.7 vs 1.9) (tables 2 and 3).
Differential (comparative) characteristics of AF patients with OAC alone or in combination with AP—Part I
Differential (comparative) characteristics of AF patients with OAC alone or in combination with AP—Part II
Patients on combination therapy more often had paroxysmal or persistent AF and, conversely, less frequently permanent AF. Similar statistically significant differences were also present in the comparison of patients on triple therapy versus OAC therapy alone, as well as in the comparison of patients on triple therapy versus dual (OAC+single AP) therapy (tables 2 and 3).
Assessment of the appropriateness of combined OAC+AP use
Dual therapy
For a dual therapy with OAC and AP treatment (aspirin or clopidogrel), we estimated appropriateness as in the presence of either an ACS or any PCI within the previous year (see Methods and online supplementary figure 1). Following this definition, only 4.7% of the patients treated with a combination of AP and OAC received this appropriately (see online supplementary figure 1). Diabetes and previous coronary stenting (>1 year before enrolment) were commonly found in patients receiving ‘likely inappropriate’ combination therapy (tables 4 and 5). We found 403 out of 629 patients classified as ‘likely inappropriate’ combination therapy (64.1%) as having evidence of vascular disease (CHD, cerebrovascular disease or PAD). By subtraction, about 36% of patients judged as ‘inappropriately treated’ had no detectable understandable medical reason for the addition of aspirin to VKAs. We also investigated in such patients whether the presence of a mechanical heart valve (a condition where the ACC/AHA Guidelines,19 but not the ESC Guidelines,20 suggest adding aspirin along with VKAs. Only 18 such patients out of 629 had such a condition, and in only 5 cases without the presence of concomitant vascular disease.
Differential (comparative) characteristics of AF patients treated appropriately/inappropriately with dual therapy of OAC and AP—Part I
Differential (comparative) characteristics of AF patients treated appropriately/inappropriately with dual therapy of OAC and AP—Part II
Triple therapy
For patients treated with an OAC plus dual AP therapy, the criterion for inappropriateness included the absence of a BMS implantation ≤1 month, OR the absence of a DES implantation ≤1 year, OR the absence of an ACS ≤1 year. Relative figures for these categories of appropriate/inappropriate use are shown in online supplementary figure 2.
Patients with PCIs performed within 12 months prior to baseline in those treated with dual therapy were 23, of whom 8 were treated with BMS, 14 treated with DES, and 1 patient with missing information. In those treated with triple therapy, patients with PCIs performed within 12 months prior to baseline were 33, of whom 13 were treated with BMS, 19 treated with DES and 1 patient with missing information.
Of the entire population, 36.2% received triple combination treatment appropriately according to the ESC definitions (implantation of a BMS ≤1 month OR a DES implantation ≤1 year OR ACS ≤1 year) (see online supplementary figure 2), suggesting that ‘triple therapy’ was more often discontinued than maintained in such conditions, in accordance with current recommendations.
Likely inappropriate ‘triple therapy’ was associated with the presence of a previous valve replacement, CHD, previous coronary stenting (>1 year remote), a previous myocardial infarction, and a previous ‘other ischemic-thromboembolic event (table 6 and 7). In 10 out of 67 patients labelled as ‘likely inappropriate’ triple therapy (15%) we could find no reasonable explanation for such a pattern in the presence of any such conditions.
Differential (comparative) characteristics of AF patients treated appropriately/inappropriately with triple therapy of OAC and AP—Part I
Differential (comparative) characteristics of AF patients treated appropriately/inappropriately with triple therapy of OAC and AP—Part II
Sensitivity analyses
Current indications for patients with AF receiving a coronary stent for an ACS are shortening the time of triple therapy (OAC+ASA+clopidogrel) to 3–6 months rather than a year18 because of the prevailing risk of bleeding over the risk of stent thrombosis. Such a recommendation is likely to be reinforced in the light of lesser risk of stent thrombosis with the newer generations of stents. Because of this, we also conducted a sensitivity analysis for the appropriate versus inappropriate use of triple therapy by restricting the timing for appropriate use after an ACS to 6 months rather than a year. As expected, the inappropriate use of combination therapy with stricter definition of the need for combination therapy was higher, with >70% of patients treated with triple therapy here judged as inappropriately treated (see online supplementary figure 3).
Further sensitivity analyses were done regarding the appropriate use of dual and triple therapy when replacing OAC by VKA. The results were comparable with those of the main analysis. Of the 589 patients on dual combination therapy, 563 (95.6%) patients did not have an accepted indication for this according to the 2012 ESC guidelines, while 26 (4.4%) did. In the 94 patients receiving a triple combination therapy, 63 (67.0%) patients did not meet the criteria for a combination therapy, while 31 (33.0%) did (see online supplementary figures 4 and 5).
Discussion
Main findings
Our analysis of the PREFER in AF registry data suggests that the combination of a single AP agent (often aspirin) and OAC is commonly used without a clinical indication, while ‘triple therapy’ (OAC, ASA and clopidogrel) is more often used in line with clinical guidance. To the best of our knowledge, this is the largest dataset so far published on this subject.
Patterns of combination therapy in PREFER in AF in 2012
A considerable proportion (9.9%) of our consecutive AF patients received combination therapy with a VKA and one or more AP, mostly (8.1% of the total) in the form of a VKA+ aspirin or clopidogrel (dual therapy). Our analysis suggests that this ‘dual therapy’ is often used without a compelling indication, even though the evidence for an increased risk of bleeding is clear.21 ,22 This possibly reflects the long-held belief that arterial vascular events (ACS, peripheral emboli) can only be prevented by AP therapy. Indeed, vascular disease was much more common in patients receiving combination therapy. Additionally, risk factors for recurrent vascular events (diabetes, previous myocardial infarction, history of smoking, dyslipoproteinemia, chronic kidney disease) were more common in patients on combination therapy. Furthermore, combination therapy patients had, more often, a remote history of stenting, suggestive of a pattern where discontinuing the AP agent was ‘forgotten’ or indefinitely postponed. Interestingly, a history of previous ischaemic stroke was no more prevalent in patients receiving combination therapy, probably indicating a concern of doctors about a higher risk of cerebral bleeding. The inference from the observation of such prescription pattern is that vascular disease or a prosthetic valve implantation are likely to be important driving forces in the prescription of AP therapy in addition to anticoagulants.
We next went on to interpreting what percentage of AF patients on combination treatment did so without a valid reason, according to current recommendations. For this analysis, we used lenient criteria (considering the use of combination therapy for 1 year after a stent placement, or an ACS, or other embolic event acceptable). Even on this lenient definition, 95% of patients prescribed OAC plus single AP therapy, and 64% of patients prescribed OAC plus dual AP therapy were judged to do so for inappropriate reasons based on the timing from a recent ACS with or without stenting. In this latter subset, figures become even higher (77%) by adopting the more recent criteria for inappropriateness of dual antiplatelet therapy after DES, based on more recent evaluations of the trade-off between stent thrombosis and bleeding.18
History of an ACS and stenting (both <1 year before the censored time) are statistically significant correlates of (and possibly important reasons for) appropriate prescription patterns. From our data, we infer that one important reason for largely inappropriately adding a single AP therapy (mostly aspirin) to an OAC is a carry-over effect of a prescription pattern in patients with a history of an ACS or stenting, now at >1 year time distance. This overuse of combination therapy, often given to patients with a relatively high bleeding risk (higher HAS-BLED score), also suggests excessive fear of thrombotic events without too much simultaneous concern for the possibly ominous consequences of bleeding.23
PREFER in AF, started at a time when not all NOACs were available throughout Europe, collected data from only a small proportion (6.1%) of all enrolled patients on NOACs, with 0.5% of patients receiving a NOAC in combination with aspirin or clopidogrel, and only three patients out of 7243 having been prescribed a NOAC plus dual AP therapy. While the numbers are small, the same overtreatment patterns here appear to prevail.
Comparisons with other registry data
Several other contemporary registries24–30 have addressed and described the patterns of combined use of OAC and antiplatelet agents in the setting of AF. While most such studies26 ,29 ,30 have pointed out that CHD is the strongest reason for combination therapy of OAC plus antiplatelet agents, ours is the only registry-based study specifically addressing the issue of the appropriateness/inappropriateness of such combinations.
Limitations
PREFER in AF provides a contemporary snapshot of the management of AF in seven European countries, and illustrates the changes in AF management after publication of the ESC guidelines on AF in 2010. Despite consecutive enrolment and selection of ‘representative sites’ (outpatients and inpatients, cardiologists and other physicians), similar to other registries, we cannot rule out selection bias at the centre or patient levels. While the inappropriateness of adding antiplatelet agents to a VKA is rather well established at the present time, this issue is still debateable with the NOACs (small numbers accrued in the present registry). Therefore, the presentation of data for VKAs as opposed to all OACs in the main section of this report rather than in the online supplementary is debateable. We have preferred this presentation, since at this time no differential behaviour of VKAs and NOACs has been presented in this regard.
Conclusions
Combination therapy with an OAC and one or two AP agents is common in patients with AF in Europe, mostly driven by the concomitance of vascular disease and mostly prescribed for reasons judged inappropriate according to current standards. In a relatively large percentage (35% for dual therapy; 15% of triple therapy) combination therapy was found in the absence of any evidence of concomitant vascular disease. Very few such cases could be explained by the presence of a mechanical valve prosthesis, making the vast majority of such cases totally unjustifiable even by lenient criteria for inappropriateness and taking into account indications different from the ESC guidelines.
Because of the higher risk of bleeding entailed by combination therapy, such prescription pattern should clearly be the target of better education in practicing physicians. Combination therapy should be judged even more dangerous and even more often inappropriate with the expanding use of more potent antiplatelet agents prasugrel and ticagrelor, arguing for the conservative nature of our warning against such use. More educational efforts should be devoted to instruct physicians’ behaviour in patients in whom AF coexists with vascular or prosthetic valve disease. Additionally, more data to inform on the optimal management of these patients would be helpful.
Key messages
What is already known on this subject?
-
Combined oral anticoagulants (OAC) and antiplatelet (AP) therapy is generally discouraged in atrial fibrillation (AF) outside of acute coronary syndromes or stenting because of increased bleeding. Despite this, several registries have reported frequent combined use in AF.
What might this study add?
-
We evaluated frequency of the combined use and, most important, possible reasons in a largely unselected contemporary European AF population participating in the European PREFER in AF Registry. Out of 7243 patients enrolled, 791 (10.9%) were treated with a combination of OAC and one (dual) or two AP (triple combination therapy). Compared with patients only prescribed OAC, patients on combination treatment had more frequently diabetes, dyslipidaemia, coronary heart disease, or peripheral arterial disease, a higher mean CHA2DS2VASc, and HAS-BLED scores. Of the 660 patients on dual AP+OAC combination therapy, 629 (95.3%) did not have an accepted indication. Out of the 105 patients receiving triple combination therapy, 67 (63.8%) did not have an accepted indication.
How might this impact on clinical practice?
-
Since the combined use of OAC and AP therapy is common in AF, but also largely inappropriate, and largely explained by the coexistence of coronary or peripheral arterial disease, more educational efforts should be directed in instructing doctors of the dangers deriving from the combined use. Efforts at better implementing guidelines in this setting are warranted.
Acknowledgments
We thank all centres which participated in this registry (a full list is available online), and all patients who gave consent to participate. Furthermore, we thank Markus Schwertfeger (employee of Daiichi Sankyo, Europe) for medical advice.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
-
Contributors This manuscript was conceived and written by the first author with the help of all members of the Steering Committee. As to authors affiliated with the Sponsor, specific roles were as follows: Bettina Ammentorp-Schmidt: set-up and management of the registry; co-development of the manuscript; Tessa Schliephacke: co-development of the manuscript; Paul-Egbert Reimitz: statistical advice and programming for the manuscript; Josef Schmitt: responsible statistician for the registry; statistical advice and programming for the manuscript; Christine Schober: co-development of the manuscript. The first author takes responsibility for the overall content of the manuscript as a guarantor.
-
Funding The PREFER in AF registry is funded by Daiichi Sankyo Europe. Apart from the selection of the countries, all design aspects were decided by the scientific Steering Committee and executed by an independent Contract Research Organisation. The members of the Steering Committee received honoraria for their advice in the planning of the registry.
-
Competing interests RDC: lecture fees and honoraria from Daiichi-Sankyo, Boehringer-Ingelheim, Bayer, Bristol-Myers Squibb, Pfizer, and Lilly; J.-YLH: consultant/conferences/advisory board for Sanofi-Aventis, BMS/Pfizer, Meda, Boehringer-Ingelheim, MSD, Bayer, Servier, and Daiichi-Sankyo; HD: Steering Committee member and National Coordinator for Germany RE-LY, APPRAISE-1 and 2, Garfield Registry, and PREFER in AF. Fees, honoraria, and research funding from AstraZeneca, Bayer, Berlin-Chemie, Boehringer-Ingelheim, Boston Scientific, Bristol-Myers Squibb, Daiichi-Sankyo, Lilly, MSD, BMFT, Harvard Med. Res. Inst., and Thrombosis Research Institute; JLZ: speaker honoraria from Sanofi, Servier, and Daiichi-Sankyo; BA and JS: employees of Daiichi-Sankyo Europe; PK: Consulting fees/honoraria from AstraZeneca, Bayer, Boehringer-Ingelheim, Bristol Myer Squibb, Daiichi-Sankyo, 3M Medica, MEDA Pharma, Medtronic, Merck, Otsuka, Pfizer, Sanofi-Aventis, Servier, Siemens, Takeda. Research grants from Cardiovascular Therapeutics, European Union, Fondation LeDucq, German Federal Ministry for Education and Research (BMBF), German Research Foundation (DFG), 3M Medica, MEDA Pharma, Medtronic, OMRON, St Jude Medical.
-
Patient consent Obtained.
-
Ethics approval Ethics Committees at all sites involved.
-
Provenance and peer review Not commissioned; externally peer reviewed.