Accepted Manuscript Antithrombotic treatment in patients with heart failure and associated atrial fibrillation and vascular disease: A nationwide cohort study Morton Lamberts, Gregory Y.H. Lip, Martin H. Ruwald, Morten Lock Hansen, Cengiz Özcan, Soren L. Kristensen, Lars Køber, Christian Torp-Pedersen, Gunnar H. Gislason PII:

S0735-1097(14)02048-8

DOI:

10.1016/j.jacc.2014.03.039

Reference:

JAC 20072

To appear in:

Journal of the American College of Cardiology

Received Date: 14 October 2013 Revised Date:

22 February 2014

Accepted Date: 4 March 2014

Please cite this article as: Lamberts M, Lip GYH, Ruwald MH, Hansen ML, Özcan C, Kristensen SL, Køber L, Torp-Pedersen C, Gislason GH, Antithrombotic treatment in patients with heart failure and associated atrial fibrillation and vascular disease: A nationwide cohort study, Journal of the American College of Cardiology (2014), doi: 10.1016/j.jacc.2014.03.039. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Antithrombotic treatment in patients with heart failure and associated atrial fibrillation and vascular disease: A nationwide cohort study Running title: HF and vascular disease plus AF Morton Lamberts†; Gregory Y.H. Lip§ (+); Martin H. Ruwald†; Morten Lock Hansen†; Cengiz Özcan†; Soren L. Kristensen †; Lars Køber*; Christian Torp-Pedersen#; Gunnar H. Gislason†‡ (+)

joint senior authors

RI PT

(+)

†

Department of Cardiology, Gentofte University Hospital, Hellerup, Copenhagen, Denmark; University of Birmingham, Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK *The Heart Centre, University Hospital of Copenhagen, Rigshospitalet, Denmark # Institute of Health, Science and Technology, Aalborg University, Denmark ‡ National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark.

SC

§

M AN U

Conflicts of interest: The authors report no conflicts of interest.

Funding: Prof. Gislason is supported by an independent research scholarship from the Novo Nordisk Foundation.

AC C

EP

TE D

Address for correspondence: Morten Lamberts, MD Department of Cardiology, Gentofte University Hospital, Post 635 Niels Andersens Vej 65, 2900 Hellerup, DK Tel /FAX: +45 2243 4186 / +45 7020 1283 E-mail: [email protected]

1

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

Abstract: Objectives: To investigate the impact of atrial fibrillation (AF) and antithrombotic treatment on prognosis in patients with heart failure (HF) as well as vascular disease. Background: HF, vascular disease and AF are pathophysiological related and understanding antithrombotic treatment in these conditions is crucial. Methods: In hospitalized patients with HF and co-existing vascular disease (coronary artery disease or peripheral arterial disease) followed from 1997-2009, AF status was categorized as prevalent incident and no AF. Risk of thromboembolism (TE), myocardial infarction (MI) and serious bleeding was assessed by Cox regression models (hazard ratio (HR) with [95% confidence interval (CI)]) with antithrombotic therapy and AF as time-dependent variables. Results: A total of 37,464 patients were included (mean age 74.5 [SD 10.7] years, 36.3% females) with a mean follow-up of 3.0 years during which 20.7% were categorized as prevalent AF and 17.2% as incident AF. In comparison to VKA in prevalent AF, VKA plus antiplatelet was not associated with decreased risk of TE (HR 0.91 [0.73-1.12] or MI 1.11 [0.96-1.28] while bleeding risk was significantly increased (HR 1.31 [1.09-1.57]). Corresponding estimates for incident AF were HR 0.77 [0.56-1.06], 1.07 [0.89-1.28] and 2.71 [1.33-2.21], respectively. In no AF patients, no statistical differences were seen between antithrombotic therapies on TE or MI risk while bleeding risk was significantly raised for VKA with or without single antiplatelet therapy. Conclusions: In AF patients with co-existing HF and vascular disease, adding single antiplatelet on top of VKA therapy is not associated with additional benefit on thromboembolic or coronary risk, but notably increased bleeding risk. Key words: Heart failure, atrial fibrillation, vascular disease, coronary artery disease, antithrombotic treatment.

AC C

EP

TE D

Abbreviations AF, Atrial fibrillation HF, Heart Failure VKA, Vitamin K antagonist ATC system, Anatomical Therapeutic Chemical Classification system ICD, International Classification of Diseases CHA2DS2-VASc, Congestive HF, Hypertension, Age >75 years, Diabetes, Stroke/thromboembolism, Vascular disease, Age 65-74 years, female sex HAS-BLED, Hypertension, Abnormal liver/renal function, Stroke, Bleeding, Labile INR [International Normalized Ratio], Elderly, Drugs WARCEF trial, Warfarin versus Aspirin in Reduced Cardiac Ejection Fraction trial HR, Hazard ratios CI, Confidence interval

2

ACCEPTED MANUSCRIPT Introduction While systolic heart failure (HF) is associated with increased risk of thromboembolism and death, no firm evidence exists on the benefit of antithrombotic treatment in uncomplicated HF in sinus rhythm(1-3). For example, a recent Cochrane review found no convincing

RI PT

evidence that oral anticoagulant therapy modified mortality or vascular events in patients with HF in sinus rhythm(4).

Two conditions commonly related with HF are vascular disease and atrial fibrillation (AF),

SC

and both frequently requiring the use of antithrombotic therapy with antiplatelet drugs and oral anticoagulation, respectively. In patients with coronary or peripheral artery disease,

M AN U

antiplatelet therapy is recommended(5-7), although the benefits of antiplatelet therapy in patients with concomitant HF are less well defined in relation to mortality and vascular events(4). In HF patients with AF, oral anticoagulation is clearly indicated(8,9). The use of antithrombotic therapy has to balance a reduction in thromboembolism against the

TE D

potential increase in risk of bleeding(10). Bleeding whilst on antithrombotic therapy may have implications for subsequent adverse outcomes(11-15). Patients with HF may also be predisposed to more bleeding, due to difficulties with warfarin and liver congestion(16) and

EP

in the recent WARCEF (Warfarin versus Aspirin in Reduced Cardiac Ejection Fraction) trial

AC C

conducted in HF patients in sinus rhythm, the beneficial effects in reducing ischaemic stroke were offset by an increase in major bleeding with warfarin therapy(17). If patients with HF have both vascular disease and AF, a common practice is to concomitantly prescribe oral anticoagulation and antiplatelet therapy; as such patients are considered ‘high risk’. Indeed incident and prevalent AF may confer different risks. In general population studies, there is little evidence for a beneficial effect of such combination antithrombotic therapy on thromboembolism given an increase in serious bleeding(11,12). Limited data are available for HF patients who have both vascular disease and AF.

3

ACCEPTED MANUSCRIPT In a real-life cohort of HF patients with vascular disease, our objective was to assess the relation of incident or prevalent AF to thromboembolism and serious bleeding. Second, we also assessed the effectiveness and safety of ongoing antithrombotic treatment in such

AC C

EP

TE D

M AN U

SC

RI PT

patients.

4

ACCEPTED MANUSCRIPT Methods Registries We linked information on the individual level from several nationwide databases. The National Patient Registry classifies all hospital contacts according to the International

RI PT

Classification of Diseases (ICD) since 1977 (with the 8th revision until 1994, and then the 10th revision.). Coding is performed for the primary diagnosis of contact, and if appropriate one or more secondary diagnoses, and when identifying diagnoses in the registries either was

SC

allowed (18). Procedures performed are also coded according to Nordic Medical Statistics Committee of Surgical Procedures. From the national prescription registry, we collected

M AN U

information on strength, number of tablets and date of dispensing for each individual according to the Anatomical Therapeutic Chemical Classification system (ATC) endorsed by the WHO(19). Vital status and cause of death according to the ICD 10th revision were obtained from the Danish Personal Registration System and the National Causes of Death

TE D

register, respectively(20). Using a unique number we retrospectively linked this information for each individual. All ICD and ATC codes used are available as supplementary material. Study population

EP

All Danish residents with a first-time HF hospitalization were identified between January 1,

AC C

1997 and December 31, 2009. We included patients with prior diagnosis of myocardial infarction (MI), aortic plaque and peripheral artery disease, and procedures on coronary arteries (coronary artery bypass and coronary intervention) as markers of vascular disease. The date of study inclusion with HF was the date of discharge in patients alive. The presence of no AF was defined for patients without an AF diagnosis (since 1977) prior to HF hospitalization, while prevalent AF patients had a diagnosis of AF prior to HF hospitalization. During the study period no AF patients were continuously screened for an AF diagnosis and included as incident AF at the date of a first-time AF admission. Hence, the

5

ACCEPTED MANUSCRIPT study population initially comprised of patients with a HF hospitalization and co-existing vascular disease with status of prevalent (known) or no AF. During follow-up, no AF patients could subsequently change status to incident AF (Figure 1). The categorizing of AF patients was predefined as occurrence of AF (either prevalent or incident) from a first-time HF

RI PT

hospitalization might pose different risks (e.g. duration of AF disease burden, influencing antithrombotic treatment strategy, progression of HF). HF

SC

The administrative discharge coding for HF classified HF as hypertensive (ICD-10 DI11.0), cardiomyopathy (ICD-10 DI42, including dilated, alcoholic and obstructive cardiomyopathy),

M AN U

acute pulmonary oedema (ICD-10 DJ81.9), and unspecified HF (ICD-10 DI50 including decompensated heart failure [ICD-10 I50.9]). To assess the severity of HF, we calculated the daily dosage of loop diuretics prior to and following HF hospitalization: Group 1 (0 to 39 mg), Group 2 (40 mg to 79 mg), Group 3 (80 mg to 159 mg), and Group 4 (>=160 mg), as

Antithrombotic treatment

TE D

previously done(21).

For each individual, all prescriptions of aspirin, clopidogrel and vitamin K antagonists (VKA

EP

i.e. warfarin and phenprocoumon) were identified and the following commonly used

AC C

treatment regimens were classified; single antiplatelet therapy (aspirin or clopidogrel), VKA, and VKA plus single antiplatelet therapy. In no AF patients dual antiplatelet therapy (aspirin and clopidogrel) was also assessed for the primary outcomes. Ongoing exposure to antithrombotic treatment was determined from claimed prescriptions as previously done (11,22). Shortly explained, from the number of tablets dispensed and the strength of tablets, an average day-to-day dose was defined. Patients were allowed to change group but could only be exposed to one treatment group at any given time, and was only considered at risk when having tablets available for consumption. Subsequent antithrombotic treatment at

6

ACCEPTED MANUSCRIPT baseline was defined as any claimed prescriptions of VKA, antiplatelet drugs or both up to 30 days following HF discharge(11). Outcomes Thromboembolism was defined as hospitalization or death from ischemic stroke, transient

RI PT

ischemic attack, and arterial embolism. Serious bleeding was defined as hospitalization or death from intracranial, gastrointestinal, respiratory, urogenital bleeding and anaemia caused by bleeding. As secondary outcomes, recurrent HF hospitalization and myocardial infarction

SC

(MI) including hospitalization and coronary death were used. Due to study design of

continuous inclusion of incident AF, death was included as an outcome for prevalent and no

M AN U

AF patients only. The outcome definitions have previously been used (22-24). For overall thrombosis risk, an outcome including both thromboembolism and MI was also defined. Other pharmacotherapy and comorbidity

Any prescriptions 180 days prior to inclusion of the following drugs defined other

TE D

pharmacotherapy: Renin-angiotensin inhibitors, beta-blockers, spironolactone, thiazides, loop-diuretics, non-steroidal anti-inflammatory drugs, and statins. For risk factors for thromboembolism and bleeding, we calculated scores of CHA2DS2-VASc (Congestive HF,

EP

Hypertension, Age >75 years, Diabetes, Stroke/thromboembolism, Vascular disease, Age 65-

AC C

74 years, female sex) and HAS-BLED (Hypertension, Abnormal liver/renal function, Stroke, Bleeding, Labile INR [International Normalized Ratio], Elderly, Drugs [NSAID]) from recorded comorbidities and pharmacotherapy as previously used and validated(23,25). All patients were scored at least 2 for CHA2DS2-VASc according to HF and vascular disease status. INR values were not available in the registries, and use of aspirin was not incorporated in the (modified) HAS-BLED score, as this was an explanatory variable. For incident AF patients, all characteristics were determined at the date of the first-time AF diagnosis. Statistical analyses

7

ACCEPTED MANUSCRIPT Continuous variables are presented as mean with standard deviation (SD), and categorical as numbers with percentage. All rates are crude incidence rates calculated as events per 100 person-years with 95% confidence intervals (CI). Hazard ratios (HR) estimates with 95% CI for outcomes were calculated in a Cox proportional-hazard model with antithrombotic

RI PT

treatment and AF status as time-varying variables. These models were adjusted for age, sex, inclusion year, HF severity group (daily dosage of loop diuretics at inclusion), and

CHA2DS2-VASc score for events of thromboembolism and HAS-BLED for events of serious

SC

bleeding. For the secondary outcomes, adjustment included evidence-based HF medication (beta-blockers, renin-angiotensin receptor inhibitors and spironolactone) for HF

M AN U

hospitalization and coronary risk factors/medication (beta-blockers, renin-angiotensin receptor inhibitors, statins, diabetes, hypertension, and renal failure) for events of MI. In models not assessing antithrombotic treatment, ongoing antithrombotic treatment was also used for adjustment. As mentioned previously, characteristics were updated in patients

TE D

changing from no AF to incident AF status. To illustrate overall prognosis in the population, we calculated Kaplan-Meier survival estimates for patients with prevalent and no AF at inclusion (the no AF group comprised subsequently incident AF patients) (Figure 2). For

EP

sensitivity and the potential reduction of unmeasured confounders, we performed matching

AC C

analyses using a propensity score model. We defined controls as no AF patients and cases as the presence of AF (whether prevalent and incident) and used risk set matching by date of potential AF. This allowed a patient without AF to be defined as a control, and subsequently as a case if an AF hospitalization occurred. Propensity score was calculated by Cox regression model conditional on baseline variables of age, inclusion year and risk factors included in the CHA2DS2-VASc and HAS-BLED score. Matching was performed using the Greedy matching macro (at http://mayoresearch.mayo.edu/mayo/research/biostat/upload/gmatch.sas, 16 August 2013,

8

ACCEPTED MANUSCRIPT last date accessed). Patients were followed to death or the end of the study period (December 31, 2009). For model control, assumptions were not violated (linearity of continuous variables, proportional hazard assumptions and lack of clinical relevant interaction). Statistical software packages SAS 9.2 (SAS Inc., NJ) and Stata11.0 (StataCorp, TX) were

RI PT

used. Ethics

The study was approved by the Danish Data Protection agency (ref 2007-41-1667), and data

based study, Danish law does not require ethical approval.

M AN U

Results

SC

were made available to us so no individuals could be identified. As a retrospective registry-

A total of 37,464 patients with HF and vascular disease were included (mean age 74.5 years [SD 10.7], 36.3% females). Of these 7,804 (20.7%) had prevalent AF while a further 6,432 (17.2%) developed incident AF (Figure 1). The characteristics of the study population

TE D

according to AF status (no, prevalent or incident AF) are shown in Table 1. In patients with no AF, potential indications for antithrombotic therapy are provided in Supplementary Table 2. Mean time between first date of hospitalization of vascular disease and inclusion (with HF)

EP

was 6.5 years (SD 6.5) and median 4.5 years (IQR 0.8-10.5). Mean CHA2DS2-VASc and

AC C

HAS-BLED scores were 5.0 [SD 1.5] and 2.1 [SD 1.0], respectively. During a mean followup of 3.0 years (median 3.3 years [interquartile range 0.9-7.9], 23,154 (61.8%) died. Prevalent AF patients were more likely to die compared to no AF patients (Figure 2). A total of 4,272 (11.4%), 4,383 (11.7%), 17,889 (47.7%), and 13,003 (34.7%) events of thromboembolism, serious bleeding, recurrent HF hospitalization and MI occurred, respectively. Relation of AF status and outcomes

9

ACCEPTED MANUSCRIPT Total person-years accumulated for prevalent AF was 20,691 person-years and for incident AF 15,758 person-years. The ‘no AF’ group accumulated 77,317 person-years. Mean time to incidence AF was 473 days (SD 787) and median 56 days [IQR 0-632]. Crude rates of thromboembolism (events per 100 person years [95% CI]) were 5.8 [5.5-6.2], 4.6 [4.2-5.0],

RI PT

and 4.1 [3.9-4.2] for prevalent, incident and no AF, respectively. For serious bleeding, the corresponding crude rates were 5.6 [5.3-6.0], 4.6 [4.3-5.0], and 3.7 [3.5-3.8]. Figures 3A-B show that incident and prevalent AF had similar HRs of thromboembolism and serious

SC

bleeding and the risk was higher than in patients without AF. For the secondary outcomes of HF hospitalization and MI, crude rates and HRs are shown in Figure 3C-D. Among patients

M AN U

with either prevalent or incident AF, no marked difference was apparent for risk of recurrent HF hospitalization. Concerning risk of MI, an increased risk was seen for incident AF compared to no AF or prevalent AF. No clinical relevant effect modification was present for the use of evidence-based HF medication (beta-blockers, renin-angiotensin receptor

TE D

inhibitors and spironolactone) among AF patients compared to no AF patients. In the propensity score matched model, risk of thromboembolism (HR 1.29 [1.20-1.38]) and bleeding (HR 1.48 [1.38-1.60]) among patients with AF compared to patients without AF

EP

resembled the main analyses.

AC C

Relation of antithrombotic therapy according to AF status on risk of thromboembolism and serious bleeding

Amongst HF patients with co-existing vascular disease and prevalent AF, thromboembolism rates were highest amongst those on single antiplatelet therapy and lowest for VKA plus single antiplatelet therapy (Figure 3A). No statistical difference on risk of thromboembolism was found for VKA plus single antiplatelet therapy compared to VKA (HR 0.91 [0.73-1.12]). Bleeding risk was significantly increased for VKA plus single antiplatelet therapy compared to VKA alone (HR 1.31 [1.09-1.57])(Figure 3B). In HF patients with incident AF,

10

ACCEPTED MANUSCRIPT thromboembolism rates were higher amongst those on antiplatelet therapy and lowest in those with combined VKA and antiplatelet therapy. Bleeding risk was greater in patients with VKA plus single antiplatelet compared to VKA only users. Amongst HF patients with no AF, the risk of thromboembolism was similar between single antiplatelet therapy, VKA and

RI PT

VKA plus single antiplatelet therapy. Bleeding risk was lowest in single antiplatelet therapy and highest in VKA plus single antiplatelet therapy. For fatal bleedings only, no differences were seen between the antithrombotic therapies (data not shown), although increased crude

SC

rates for prevalent (0.8 events per 100 person-years) and incident (0.7 events per 100 personyears) AF were seen compared to no AF patients (0.4 events per 100 person-years). Dual

M AN U

antiplatelet therapy (aspirin and clopidogrel) was frequently used in no AF patients (4,608 person-years accumulated), and crude rates were 3.5 (95% CI 3.0-4.1) and 5.5 events per 100 person-years (95% CI 4.8-6.2) of thromboembolism and bleeding, respectively. Relative to single antiplatelet therapy, the risk of thromboembolism was significantly reduced (HR 0.82

TE D

[0.69-0.97]) while the risk of bleeding was significantly increased (HR 1.53 [1.33-1.76]) when on dual antiplatelet therapy.

Relation of antithrombotic therapy according to AF status on secondary outcomes

EP

In patients without AF, VKA and VKA plus single antiplatelet therapy were associated with

AC C

increased risk of HF hospitalization compared to single antiplatelet therapy. No difference was found between VKA plus single antiplatelet and VKA monotherapy (Figure 3C). Among AF patients, no significant differences were found between antithrombotic treatment regimens, although adding single antiplatelet to VKA was associated with a HR of 1.11 [1.00-1.23] for risk of HF hospitalization in prevalent AF patients. Regardless of AF status, no statistical significant difference was found between VKA plus single antiplatelet and VKA monotherapy on the risk of MI (Figure 3D). Among AF patients, single antiplatelet therapy was associated with increased risk of MI. For the combined outcome of thromboembolism

11

ACCEPTED MANUSCRIPT and MI, no statistical difference was found for VKA plus single antiplatelet compared to VKA only for prevalent (HR 1.00 [0.89-1.14]) or incident AF (HR 0.97 [0.82-1.14]) (supplementary Table 3). Discussion

RI PT

In this study, we have shown that amongst patients with HF and vascular disease, the

presence of incident and prevalent AF conferred similar hazard ratios of thromboembolism, which was greater than those with no AF. However, the risk of serious bleeding was for

SC

incident and prevalent AF particular high when single antiplatelet were added on top of

VKA. Second, we have shown that amongst patients with no AF, there was no difference

M AN U

between antiplatelet therapy and VKA for thromboembolism, but serious bleeding was increased with VKA therapy.

AF contributes to a high risk of stroke and thromboembolic events in HF, and our data support previous studies(26,27). For either prevalent or incident AF, we found similar risks

TE D

of thromboembolism, serious bleeding and HF hospitalization suggesting that regardless of the first appearance of AF the prognosis is worsened for these specific outcomes. Incident AF was associated with increased risk of coronary events while prevalent AF was not compared

EP

to no AF patients. It has been suggested that AF could be a marker of disease progression and

AC C

our findings support previous studies that found new-onset AF was independently associated with cardiovascular events and death in both healthy individuals and in HF patients (28,29). HF, whether due to reduced ejection fraction or preserved ejection fraction has been associated with thromboembolism especially when AF is present. Indeed, the ‘C’ in the CHA2DS2-VASc score refers to recent acute decompensated HF or the presence of moderatesevere left ventricular dysfunction(30). Nonetheless, reliance on diagnostic coding of ‘any HF’ may be less reliable, as only approximately 50% of such patients actually have confirmed HF in the primary care setting (8,31). Although a considerable higher predictive

12

ACCEPTED MANUSCRIPT value of HF has been found in hospitalized patients, the risk and mechanisms of thrombosis related to type and degree of HF is still unresolved(32). Unsurprisingly, ‘any HF’ did not emerge as an independent stroke risk factor in the large Swedish AF cohort study(33), but has been associated with thromboembolism in other AF populations(23). AF carries a particularly

RI PT

poor prognosis in HF patients, with mortality being significantly greater with AF compared to no AF shown in our nationwide cohort study. Importantly, we show a beneficial impact of VKA therapy in these patients concerning thromboembolism protection compared to single

SC

antiplatelet, and even in the historical trials, VKA therapy did significantly reduce all-cause mortality by 26%(34).

M AN U

The use of antithrombotic therapy has to balance the reduction in thromboembolism against an increase in bleeding risk. Our data show that amongst HF patients with no AF, there is no difference between antiplatelet therapy and VKA for thromboembolism, but serious bleeding is significantly less. These findings are supportive of the recent European Society of

TE D

Cardiology guidelines on HF, which give a class III recommendation on the use of VKA in HF patients without AF concerning thrombosis protection(8). This is also consistent with observations in WARCEF, which showed a significant reduction in ischaemic stroke with

EP

VKA but at the cost of greater major bleeding risk(17). As expected, the combination of

AC C

VKA and antiplatelet therapy is associated with even more bleeding risk. In the presence of AF (whether prevalent or incident), VKA treated patients with HF had lower thromboembolism compared to antiplatelet therapy, with a higher bleeding risk. When the analysis was confined to fatal bleeds only, the difference between VKA and single antiplatelet therapy was non-significant. Patients with prior vascular disease are at increased risk of coronary event, and we also assessed if combination therapy of VKA and a single antiplatelet might provide further protection from MI. In both AF and no AF patients, we did not find statistical difference between VKA with or without a single antiplatelet agent

13

ACCEPTED MANUSCRIPT suggesting adequate coronary prophylaxis with VKA monotherapy. Our data support findings from a previous controlled trial of favourable efficacy of VKA (against aspirin and VKA plus aspirin) during 26 months of follow-up post-MI (35). A meta-analysis of trials prior to year 2000 concluded that oral anticoagulant therapy (moderate or high intensity therapy) with or

RI PT

without aspirin was beneficial for secondary coronary artery disease protection whilst degree of bleeding hazard was uncertain (36). Both studies allowed for effects at different INR levels but were limited from small sample size, did not specifically include HF patients nor

SC

investigated contemporary real-life patients with currently used treatment regimens.

The endpoint of HF admissions and death is commonly the primary outcome in HF trials. In

M AN U

our analysis, we chose a secondary outcome of HF hospitalization (without death) as prescribed antithrombotic medication would likely be withdrawn in terminally ill patients. We found that this outcome was increased amongst those with AF (whether prevalent or incident) compared to no AF. Concerning different antithrombotic strategies in AF patients

TE D

we did not find profound differences in risk of HF hospitalization. Relative to single antiplatelet therapy in no AF patient, we surprisingly found an increased risk of HF hospitalization with VKA with or without single antiplatelet therapy. This contradicts

EP

findings from the WARCEF trial, which found a non-significant increase in a secondary

AC C

outcome of heart failure hospitalization with aspirin compared to warfarin in HF patients in sinus rhythm. However these findings are readily explained by the fact that no information was available of the specific indication of therapy, and patients receiving VKA could be considered having a greater disease burden e.g. potential ‘non-registered’ AF burden. Of note, use of renin-angiotensin system inhibitors and other HF evidence-based medication did not influence antithrombotic treatment effect in the present population. Limitations

14

ACCEPTED MANUSCRIPT This study is limited by its dependence on retrospective registry data, and inherent in the observational design causal interpretation of treatment effects is not possible. The diagnoses of AF, HF, MI, and ischemic stroke have been validated in the registries with positive predictive values of 97%, 81%, 93% and 97%, respectively (32,37-39). INR measurements

RI PT

were not available in the registries. Although actual ongoing VKA exposure was

continuously updated and these data demonstrates everyday antithrombotic treatment

strategies, we had no information on specific intensities of anticoagulation selected by the

SC

prescribing physician. As efficacy and hazard have been shown to be influenced by target INR levels in especially controlled settings, this limitation should be acknowledged when

M AN U

interpreting the results (36). Selection bias could be present in the ‘no AF’ group, as we did not have data on silent AF before their initial presentation with their arrhythmia diagnosis. Thus, the ‘no AF’ patients may have included a number of such asymptomatic AF patients who have an equally bad prognosis as symptomatic patients (40,41). This is also implied by

TE D

the high risk of HF hospitalizations in this group and many patients treated with VKA despite no registered indication (supplementary Table 2). Nonetheless, we clearly show a mortality difference between prevalent AF when compared to ‘no AF’ subjects. The availability of

EP

prolonged ECG monitoring may enable greater detection of AF episodes. We may also not

AC C

have accounted for patients with an outpatient diagnosis for HF, who have not been hospitalised. Confounding by indication may be present (i.e. patients perceived at higher risk of thrombosis is treated with more intense antithrombotic therapy). We did not have information on the specific indication of antithrombotic therapy and unmeasured confounders could affect the outcomes under investigation, though we controlled for a wide range of known risk factors/prophylactic medication for the specific outcomes. It should be noted that for the outcome of MI, all patients were included as having indication for antiplatelet therapy but with a wide range and timing of previous conditions of vascular disease, which could

15

ACCEPTED MANUSCRIPT have affected antithrombotic treatment prescribed. Consequently, our findings do not support any recommendation following an acute ischemic event or following percutaneous coronary intervention. Finally, compared to other large stroke prevention trials, our outcomes differentiates strokes into ischaemic and haemorrhagic for better discrimination of the

RI PT

effectiveness and safety of antithrombotic treatment. Conclusions

In conclusion, the presence of AF (prevalent or incident) is an adverse feature in HF patients

SC

with vascular disease, and the arrhythmia has an effect on thromboembolism/bleeding and HF hospitalizations. No further beneficial effect on thromboembolism or coronary risk was

M AN U

apparent when adding a single antiplatelet drug to VKA in patients with AF (but with an

AC C

EP

TE D

increase in bleeding risk), whilst antiplatelet therapy only is inadequate.

16

ACCEPTED MANUSCRIPT References 1.

Lip GY, Piotrponikowski P, Andreotti F et al. Thromboembolism and antithrombotic therapy for heart failure in sinus rhythm. An Executive Summary of a joint Consensus Document from the ESC Heart Failure Association and the ESC Working Group on

2.

RI PT

Thrombosis. Thromb Haemost 2012;108:1009-22.

Lip GY, Gibbs CR. Antiplatelet agents versus control or anticoagulation for heart failure in sinus rhythm. Cochrane Database Syst Rev 2001:CD003333.

Massie BM. Aspirin use in chronic heart failure: what should we recommend to the practitioner? J Am Coll Cardiol 2005;46:963-6.

Lip GY, Wrigley BJ, Pisters R. Anticoagulation versus placebo for heart failure in

M AN U

4.

SC

3.

sinus rhythm. Cochrane Database Syst Rev 2012;6:CD003336. 5.

Patrono C, Baigent C, Hirsh J, Roth G, American College of Chest P. Antiplatelet drugs: American College of Chest Physicians Evidence-Based Clinical Practice

6.

TE D

Guidelines (8th Edition). Chest 2008;133:199S-233S.

Patrono C, Bachmann F, Baigent C et al. Expert consensus document on the use of antiplatelet agents. The task force on the use of antiplatelet agents in patients with

EP

atherosclerotic cardiovascular disease of the European society of cardiology. Eur

7.

AC C

Heart J 2004;25:166-81.

Warfarin Antiplatelet Vascular Evaluation Trial I, Anand S, Yusuf S et al. Oral anticoagulant and antiplatelet therapy and peripheral arterial disease. N Engl J Med 2007;357:217-27.

8.

McMurray JJ, Adamopoulos S, Anker SD et al. ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2012: The Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of

17

ACCEPTED MANUSCRIPT Cardiology. Developed in collaboration with the Heart Failure Association (HFA) of the ESC. Eur Heart J 2012;33:1787-847. 9.

Authors/Task Force M, Camm AJ, Lip GY et al. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation: An update of the 2010 ESC

RI PT

Guidelines for the management of atrial fibrillation * Developed with the special contribution of the European Heart Rhythm Association. Eur Heart J 2012;33:271947.

Lip GY, Andreotti F, Fauchier L et al. Bleeding risk assessment and management in

SC

10.

atrial fibrillation patients. Executive Summary of a Position Document from the

M AN U

European Heart Rhythm Association [EHRA], endorsed by the European Society of Cardiology [ESC] Working Group on Thrombosis. Thromb Haemost 2011;106:9971011. 11.

Sorensen R, Hansen ML, Abildstrom SZ et al. Risk of bleeding in patients with acute

TE D

myocardial infarction treated with different combinations of aspirin, clopidogrel, and vitamin K antagonists in Denmark: a retrospective analysis of nationwide registry data. Lancet 2009;374:1967-74.

Hansen ML, Sorensen R, Clausen MT et al. Risk of bleeding with single, dual, or

EP

12.

AC C

triple therapy with warfarin, aspirin, and clopidogrel in patients with atrial fibrillation. Arch Intern Med 2010;170:1433-41. 13.

Gullov AL, Koefoed BG, Petersen P. Bleeding during warfarin and aspirin therapy in patients with atrial fibrillation: the AFASAK 2 study. Atrial Fibrillation Aspirin and Anticoagulation. Arch Intern Med 1999;159:1322-8.

14.

Buresly K, Eisenberg MJ, Zhang X, Pilote L. Bleeding complications associated with combinations of aspirin, thienopyridine derivatives, and warfarin in elderly patients following acute myocardial infarction. Arch Intern Med 2005;165:784-9.

18

ACCEPTED MANUSCRIPT 15.

Moscucci M, Fox KA, Cannon CP et al. Predictors of major bleeding in acute coronary syndromes: the Global Registry of Acute Coronary Events (GRACE). Eur Heart J 2003;24:1815-23.

16.

Hirsh J, Dalen J, Anderson DR et al. Oral anticoagulants: mechanism of action,

17.

RI PT

clinical effectiveness, and optimal therapeutic range. Chest 2001;119:8S-21S.

Homma S, Thompson JL, Pullicino PM et al. Warfarin and aspirin in patients with heart failure and sinus rhythm. N Engl J Med 2012;366:1859-69.

Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J

SC

18.

Public Health 2011;39:30-3.

Kildemoes HW, Sorensen HT, Hallas J. The Danish National Prescription Registry.

M AN U

19.

Scand J Public Health 2011;39:38-41. 20.

Helweg-Larsen K. The Danish Register of Causes of Death. Scand J Public Health 2011;39:26-9.

Gislason GH, Rasmussen JN, Abildstrom SZ et al. Persistent use of evidence-based

TE D

21.

pharmacotherapy in heart failure is associated with improved outcomes. Circulation 2007;116:737-44.

Lamberts M, Olesen JB, Ruwald MH et al. Bleeding after initiation of multiple

EP

22.

AC C

antithrombotic drugs, including triple therapy, in atrial fibrillation patients following myocardial infarction and coronary intervention: a nationwide cohort study. Circulation 2012;126:1185-93. 23.

Olesen JB, Lip GY, Hansen ML et al. Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study. BMJ 2011;342:d124.

24.

Olesen JB, Lip GY, Kamper AL et al. Stroke and bleeding in atrial fibrillation with chronic kidney disease. N Engl J Med 2012;367:625-35.

19

ACCEPTED MANUSCRIPT 25.

Olesen JB, Lip GY, Lindhardsen J et al. Risks of thromboembolism and bleeding with thromboprophylaxis in patients with atrial fibrillation: A net clinical benefit analysis using a 'real world' nationwide cohort study. Thromb Haemost 2011;106:739-49.

26.

Effects of enalapril on mortality in severe congestive heart failure. Results of the

RI PT

Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS). The CONSENSUS Trial Study Group. N Engl J Med 1987;316:1429-35. 27.

Fuster V, Gersh BJ, Giuliani ER, Tajik AJ, Brandenburg RO, Frye RL. The natural

28.

SC

history of idiopathic dilated cardiomyopathy. Am J Cardiol 1981;47:525-31.

Conen D, Chae CU, Glynn RJ et al. Risk of death and cardiovascular events in

29.

M AN U

initially healthy women with new-onset atrial fibrillation. JAMA 2011;305:2080-7. Rivero-Ayerza M, Scholte Op Reimer W, Lenzen M et al. New-onset atrial fibrillation is an independent predictor of in-hospital mortality in hospitalized heart failure patients: results of the EuroHeart Failure Survey. Eur Heart J 2008;29:1618-

30.

TE D

24.

Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a

EP

novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest

31.

AC C

2010;137:263-72.

Davies M, Hobbs F, Davis R et al. Prevalence of left-ventricular systolic dysfunction and heart failure in the Echocardiographic Heart of England Screening study: a population based study. Lancet 2001;358:439-44.

32.

Kumler T, Gislason GH, Kirk V et al. Accuracy of a heart failure diagnosis in administrative registers. Eur J Heart Fail 2008;10:658-60.

20

ACCEPTED MANUSCRIPT 33.

Friberg L, Rosenqvist M, Lip GY. Evaluation of risk stratification schemes for ischaemic stroke and bleeding in 182 678 patients with atrial fibrillation: the Swedish Atrial Fibrillation cohort study. Eur Heart J 2012;33:1500-10.

34.

Hart RG, Pearce LA, Aguilar MI. Adjusted-dose warfarin versus aspirin for

35.

RI PT

preventing stroke in patients with atrial fibrillation. Ann Intern Med 2007;147:590-2. van Es RF, Jonker JJ, Verheugt FW, Deckers JW, Grobbee DE. Aspirin and

coumadin after acute coronary syndromes (the ASPECT-2 study): a randomised

36.

Anand SS, Yusuf S. Oral anticoagulant therapy in patients with coronary artery

M AN U

disease: a meta-analysis. JAMA 1999;282:2058-67. 37.

SC

controlled trial. Lancet 2002;360:109-13.

Krarup LH, Boysen G, Janjua H, Prescott E, Truelsen T. Validity of stroke diagnoses in a National Register of Patients. Neuroepidemiology 2007;28:150-4.

38.

Mukamal KJ, Tolstrup JS, Friberg J, Jensen G, Gronbaek M. Alcohol consumption

TE D

and risk of atrial fibrillation in men and women: the Copenhagen City Heart Study. Circulation 2005;112:1736-42. 39.

Madsen M, Davidsen M, Rasmussen S, Abildstrom SZ, Osler M. The validity of the

EP

diagnosis of acute myocardial infarction in routine statistics: a comparison of

AC C

mortality and hospital discharge data with the Danish MONICA registry. J Clin Epidemiol 2003;56:124-30. 40.

Freudenberger RS, Wilson AC, Kostis JB, Investigators A, Committees. Comparison of rate versus rhythm control for atrial fibrillation in patients with left ventricular dysfunction (from the AFFIRM Study). Am J Cardiol 2007;100:247-52.

41.

Wyse DG, Waldo AL, DiMarco JP et al. A comparison of rate control and rhythm control in patients with atrial fibrillation. N Engl J Med 2002;347:1825-33.

21

ACCEPTED MANUSCRIPT Figure legends Figure 1 – Study population Heart failure patients with co-existing vascular disease were categorized as prevalent or no atrial fibrillation at inclusion. Patients with no atrial fibrillation were continuously screened for incident atrial fibrillation during follow-up, and could change status accordingly.

RI PT

Figure 2 – Prognosis in first-time hospitalized heart failure patients with vascular disease according to presence of atrial fibrillation Kaplan-Meier survival estimates adjusted for age (75 years).

AC C

EP

TE D

M AN U

SC

Figure 3A-D – Risk of outcomes Rates are events per 100 person years with 95% CI. Adjusted for age, gender, inclusion year, HF severity group and CHA2DS2-VASc (for thromboembolism), HAS-BLED (for serious bleeding), evidence-based pharmacotherapy (for HF hospitalization) and coronary risk factors/medication (MI). Single AP denotes either aspirin or clopidogrel. All p values denotes test for difference. Abbreviations: AP, antiplatelet; VKA, vitamin K antagonists; HR, hazard ratio; CI, confidence interval; HF, heart failure; CHA2DS2-VASc and HAS-BLED, please see text.

22

ACCEPTED MANUSCRIPT Table 1– Characteristics of study population Prevalent atrial fibrillation

Incident atrial fibrillation#

Total

29,660

7,804

6,432

Females

36.4%

35.7%

34.9%

Age, years (SD)

73.9 (11.0)

76.8 (9.3)

Diagnosis

93.5%

90.3%

Procedure

27.3%

31.5%

M AN U

Heart failure discharge diagnosis

SC

Vascular disease

RI PT

No atrial fibrillation

76.8 (9.4)

94.0% 23.4%

2.3%

2.1%

2.4%

Cardiomyopathy

3.0%

3.0%

2.7%

Decompensated

32.0%

31.1%

32.9%

Acute pulmonary oedema

3.8%

2.5%

2.9%

61.4%

59.1%

TE D

Hypertensive

Unspecified

58.9%

Severity group (at /after inclusion)

60.8% / 38.1% 55.2% / 38.0%

58.4% / 31.5%

19.1% / 17.9% 18.9% / 15.8%

22.4% / 20.4%

15.3% / 32.4% 18.8% / 31.7%

15.2% / 35.8%

4.9% / 11.6%

7.1% / 14.5%

4.3% / 12.3%

4.9 (1.5)

5.3 (1.4)

5.2 (1.4)

2

4.2%

1.5%

1.7%

3

12.7%

7.9%

9.3%

4

23.6%

21.3%

21.2%

5

26.9%

26.9%

27.9%

Group 2

AC C

Group 3

EP

Group 1

Group 4

CHA2DS2-VASc score, mean (SD)

23

18.7%

21.7%

21.8%

7

9.4%

13.5%

12.0%

>7

4.4%

7.1%

6.1%

2.1 (1.0)

2.3 (1.0)

2.3 (1.0)

Low (score 0-1)

30.3%

20.1%

21.8%

Intermediate (score 2)

38.5%

39.6%

High (score ≥3)

31.2%

40.3%

RAS inhibitors

51.7%

55.5%

59.7%

Beta-blockers

45.7%

55.6%

52.2%

Loop-diuretics

55.9%

66.5%

71.6%

Statins

39.3%

41.7%

42.1%

Spironolactone

14.4%

17.7%

23.2%

Glucose-lowering medication

18.0%

17.8%

17.6%

21.3%

18.6%

19.3%

36.3%

29.2%

30.3%

Vitamin K antagonists

7.3%

20.7%

25.0%

2.6%

6.0%

8.0%

HAS-BLED, mean (SD)

M AN U

TE D

NSAID

SC

Other pharmacotherapy

RI PT

6

AC C

ACCEPTED MANUSCRIPT

39.8%

38.4%

Antiplatelet

Both

EP

Subsequent antithrombotic treatment*

Abbreviations: n (%), number (rounded column percent); SD, standard deviation; VKA, vitamin K antagonist; AF, atrial fibrillation; HAS-BLED, CHA2DS2-VASc, please see text. RAS, renin-angiotensin system; NSAID, non-steroidal anti-inflammatory drug. * At least one prescription claimed within 30 days from discharge, # group is subgroup of the no AF group, and all scores and pharmocotherapy are calculated from date of incident AF

24

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT Supplementary material Supplementary Table 1 Diagnoses, procedures, and pharmacotherapy used for defining the study population, comorbidity, and outcomes

Study population Defined from diagnosis of

ICD8: 425, 4270, 4271

hypertensive, cardiomyopathy

ICD10: I110, I50, I42, J81

RI PT

Heart failure

(including dilated and obstructive oedema, decompensated and unspecified heart failure.

Defined from diagnosis of myocardial ICD8: 410, 440 ICD10: I21, I22, I700, I702-I709 infarction, peripheral arterial disease

M AN U

Vascular disease

and aortic plaque orcoronary procedures Atrial fibrillation

SC

cardiomyopathy), acute pulmonary

Defined from diagnosis

NCSP: KFN

ICD8: 42793, 42794

TE D

ICD10: I48

Comorbidity

AC C

EP

CHA2DS2-VASc and HAS-BLED scores Defined from diagnosis of Previous ischemic stroke, transient ischemic thrombosis* attack, pulmonary embolus, or arterial embolus Hypertension

ICD8: 444, 450, 433, 434, 435, 436, 437, 438 ICD10: I63, I64, I26, I74, G458, G459

Defined from combination

Treatment: Adrenergic α-antagonist,

treatment with a least two classes

non-loop-diuretics, vasodilators, beta-

of antihypertensive drugs. This

blockers, calcium channel blockers, and

definition of hypertension has a

renin-angiotensin system inhibitors.

positive predictive value of 80.0% and a specificity 94.7%1 Diabetes mellitus

Defined from glucose-lowering

ATC: A10

medication Alcohol abuse

Defined from diagnosis and

ICD8: 291, 303

ACCEPTED MANUSCRIPT ICD10: E244, E52, F1, G312, G621,

reported during hospitalization

G721, I426, K292, K70, K860, L278A,

Defined from diagnoses of liver cancer, chronic liver disease, liver

O354, T51, Z714, Z721 ICD8: 070, 155, 571-573 ICD10: B15-B19, C22, D684C, I982B, K70-K77, DQ618A, Z944

surgery, cirrhosis, and hepatitis or

NCSP: KJJC

RI PT

Liver disease

adverse alcohol consumption

liver procedure Defined from diagnosis of chronic

failure

glomerulonephritis, chronic tubulointestinal nephropathy, nonend-stage chronic kidney disease, nephropathy. Previously validated with a specificity of (ref LiseKamper)

Previous bleeding

Defined from diagnosis of

gastrointestinal, intracranial,

TE D

respiratory, and urinary tract

bleedings; and anemia caused by bleeding.

EP

Pharmacotherapy Antiplatelet

ICD10: E102, E112, E132, E142, I120, M200, M313, M319, M321B, N02-N08, N11-N12, N14, N18-N19, N26, N158N160, N162-N164, N168, Q612-Q613, Q615, Q619

M AN U

and diabetic and hypertensive

ICD8: 403-404, 581-584,25002,40039, 59009, 59320, 75310-11, 75319

SC

Chronic renal

Defined from ATC codes of aspirin

ICD8: 430-431, N852-N853 ICD10: I60-I62, I690-I692, J942, K250, K254, K260, K264, K270, K280, K920K922, N02, R04, R31, S064-S066

ATC: B01AC06, B01AC04, N02BA01

or clopidogrel antagonist Renin-

Defined from ATC codes of warfarin

AC C

Vitamin K

ATC: B01AA03, B01AA04

or marcoumar Defined from ATC code

ATC: C09

Beta-blockers

Defined from ATC code

ATC: C07

Loop-diuretics

Defined from ATC code

ATC: C03C

Statins

Defined from ATC code

ATC: C10A

Spironolactone

Defined from ATC code

ATC: C03D

angiotensin

system inhibitors

ACCEPTED MANUSCRIPT Glucose-lowering Defined from ATC code

ATC: A10

medication Defined from ATC code

ATC: M01A

RI PT

NSAID

Outcomes Thromboembolism

Death from or diagnosis of

ICD10: I63-I64, G458-G459, I74

ischemic stroke, transient ischemic attack, or arterial embolism

ICD10: I60-I62, I690-I692, J942,

gastrointestinal, intracranial,

K250, K254, K260, K264, K270,

respiratory, and urinary tract

K280, K920-K922, N02, R04,

bleedings; and anemia caused by bleeding.

Heart failure

SC

Death from or diagnosis of

M AN U

Bleeding

R31, S064-S066,

Hospitalization of heart failure

ICD10: I110, I42, I50, J819

Myocardial

Coronary death or diagnosis of

ICD10: I21-22, I20-I25 (coronary

infarction

myocardial infarction

hospitalization

TE D

(secondary)

(secondary)

8th revision of the International Classification of Diseases system 10th revision of the International Classification of Diseases system The Nordic Medical Statistics Committees Classification of Surgical Procedures

EP

ICD8: ICD10: NCSP:

death).

AC C

* For HAS-BLED only stroke isincluded in the score. Reference 1: Olesen JB, Lip GY, Hansen ML et al. Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: Nationwide cohort study. BMJ. 2011;342:d124

ACCEPTED MANUSCRIPT Supplementary Table 2 – Potential indications for antithrombotic therapy in heart failure patients with co-existing vascular disease without atrial fibrillation.

AC C

EP

TE D

M AN U

SC

RI PT

VKA # (%) No AF patients§ (%) Antiplatelet# (%) Any indications 4,139 (14.0) 3,124 (13.6) 1,087 (32.5) Recent venousthromboembolism 622 (2.1) 421 (67.7) 343 (55.1) Recent arterialembolus 157 (0.5) 113 (72.0) 37 (23.6) Recent muralthrombus 5 (0.0) 3 (60.0) 4 (80.0) Valvulardisease 3,436 (11.6) 2,646 (77.0) 751 (21.8) Previousvalvular procedure 533 (1.8) 410 (76.9) 329 (61.7) Total 29,660 (100.0) 23,061 (77.8) 3,344 (11.3) Recent defines within 1 year prior to inclusion. Antiplatelet denotes aspirin or clopidogrel. Abbreviation: AF, atrial fibrillation; VKA, vitamin K antagonist. § Vertical percentage given of heart failure patients with co-existing vascular disease and no AF with potential antithrombotic indications, # Horizontal percentage given of patients on antithrombotic therapy at inclusion according to indications.

ACCEPTED MANUSCRIPT Supplementary Table 3 -Risk of thromboembolism or MI (total events 15,758) Thromboembolism or MI Treatment regimen

Prevalent AF

HR [95% CI]

VKA plus single antiplatelet VKA Single antiplatelet

13.0 [11.9-14.2]

1.00 [0.89-1.14]

13.4 [12.3-14.6]

reference

20.1 [19.1-21.1]

1.39 [1.26-1.54]

VKA plus single antiplatelet VKA Single antiplatelet

11.2 [11.9-14.2]

0.97 [0.82-1.14]

12.3 [10.9-13.9]

reference

18.3 [17.2-19.5]

1.41 [1.23-1.61]

SC

Incident AF

IR [95% CI]

RI PT

AF status

13.5 [12.0-15.1] 0.95 [0.84-1.07] VKA plus single antiplatelet 13.1 [11.6-14.8] 0.88 [0.78-1.00] VKA 14.6 [14.2-14.9] Single antiplatelet reference Adjusted for age, gender, inclusion year, HF severity group and CHA2DS2-VASc. Abbreviation: IR, crude incidence rate (events per 100 person-years); HR, hazard ratio; CI, confidence interval; CHA2DS2-VASc, please see text

AC C

EP

TE D

M AN U

No AF