BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

Home , ATC code B, Nitrovasodilator

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] BMJ Open

Oral anticoagulation therapy use in patients with atrial fibrillation after the introduction of direct oral anticoagulants: findings from the French healthcare databases, 2011-2016.

ForJournal: peerBMJ Open review only

Manuscript ID bmjopen-2018-026645

Article Type: Research

Date Submitted by the Author: 12-Sep-2018

Complete List of Authors: MAURA, Géric; French National Health Insurance (Caisse Nationale de l’Assurance Maladie/Cnam), Department of Studies in Public Health; University of Bordeaux, Inserm, Bordeaux Population Health Research Center, team Pharmacoepidemiology - UMR 1219, Billionnet, Cécile; French National Health Insurance (Caisse Nationale de l’Assurance Maladie/Cnam), Department of Studies in Public Health Drouin, Jérôme; French National Health Insurance (Caisse Nationale de l’Assurance Maladie/Cnam), Department of Studies in Public Health Weill, Alain; French National Health Insurance (Caisse Nationale de l’Assurance Maladie/Cnam), Department of Studies in Public Health Neumann, Anke; French National Health Insurance (Caisse Nationale de l’Assurance Maladie/Cnam), Department of Studies in Public Health Pariente, Antoine; University of Bordeaux, Inserm, Bordeaux Population Health Research Center, team Pharmacoepidemiology - UMR 1219; CHU Bordeaux

Anticoagulation < HAEMATOLOGY, Adult cardiology < CARDIOLOGY, Keywords: EPIDEMIOLOGY, PUBLIC HEALTH, Thromboembolism < CARDIOLOGY

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 32 BMJ Open

1 2 3 4 Title page 5 6 7 Oral anticoagulation therapy use in patients with atrial fibrillation after the introduction of direct 8 oral anticoagulants: findings from the French healthcare databases, 2011-2016. 9 10 11 Running title: Patterns of OAC use 12 13 14 Géric Maura PharmD1,2, Cécile Billionnet MSc, PhD1, Jérôme Drouin, MSc1, Alain Weill MD1, Anke 15 16 For1 peer review2,3 only 17 Neumann MSc, PhD , Antoine Pariente MD, PhD 18 1 19 Department of Studies in Public Health, French National Health Insurance (Caisse Nationale de 20 21 l’Assurance Maladie/Cnam), 75 986 Paris Cedex 20, France. 22 23 2 University of Bordeaux, Inserm, Bordeaux Population Health Research Center, team 24 25 Pharmacoepidemiology - UMR 1219, F-33000 Bordeaux, France 26 27 3 CHU de Bordeaux, Pharmacologie, F-33000 Bordeaux, France 28 29 Address for correspondence: Géric MAURA, Cnam/DSES/DESP, National Health Insurance, 50 avenue 30 31 du Pr. André Lemierre, 75986 Paris cedex 20; E-mail: [email protected]; Telephone: 32 33 +33(0)172602338; Fax: +33(0)172601724 34 35

36 37 38 39 Word count of text: (excluding the title page, abstract, tables, acknowledgements, contributions and 40 references): 3,990/4,000 41 42 Word count of abstract: 300/300 43 Number of references: 63 44 45 Number of tables: 2 46 47 Number of figures: 3 48 Number of supplementary tables: 2 49 50 51 52 53 54 55 56 57 58 1 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 32

1 2 3 Abstract 4

5 6 Objectives: To describe i) the trend in oral anticoagulant (OAC) use following the introduction of 7 direct oral anticoagulant (DOAC) therapy for stroke prevention in atrial fibrillation (AF) patients; ii) 8 9 the current patterns of use of DOAC therapy in new users with AF in France. 10 Design: i) Repeated cross-sectional study; ii) Population-based cohort study. 11 12 Setting: French National Healthcare databases (Régime général, 50 million beneficiaries). 13 Participants: i) Patients with identified AF in 2011, 2013, and 2016; ii) Patients with AF initiating OAC 14 15 therapy in 2015-2016. 16 For peer review only 17 Primary and secondary outcome measures: i) Trend in oral anticoagulant therapy use in patients 18 with AF; ii) Patterns of use of DOAC therapy in new users with AF. 19 20 Results: Between 2011 and 2016, use of OAC therapy moderately increased (+16%), while use of 21 antiplatelet therapy decreased (-22%) among all patients with identified AF. In 2016, among the 1.1 22 23 million AF patients, 66% used OAC therapy and were more likely to be treated by VKA than DOAC 24 therapy, including patients at higher risk of stroke (63.5%), while 33% used antiplatelet therapy. 25 26 Among 192,851 new users of OAC therapy in 2015-2016 with identified AF, DOAC therapy (66.3%) 27 28 was initiated more frequently than VKA therapy, including in patients at higher risk of stroke (57.8%). 29 Reduced doses were prescribed in 40% of DOAC new users. Several situations of inappropriate use at 30 31 DOAC initiation were identified, including concomitant use of drugs increasing the risk of bleeding 32 (one in three new users) and potential DOAC underdosing. 33 34 Conclusions: OAC therapy use in AF patients remains suboptimal 4 years after the introduction of 35 36 DOACs for stroke prevention in France and improvement in appropriate prescribing regarding DOAC 37 initiation is needed. However, DOAC therapy is now the preferred drug class for initiation of OAC 38 39 therapy in patients with AF, including in patients at higher risk of stroke. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 2 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 32 BMJ Open

1 2 3 4 Strengths and limitations of this study 5 • This study is the first to report both the 2011-2016 trend in oral anticoagulant (OAC) 6 7 coverage in patients with atrial fibrillation (AF) following the introduction of direct oral 8 anticoagulant (DOAC) therapy in France and the current patterns of use of DOACs in new 9 10 users including assessment of potential inappropriate use; 11 • This study is based on reimbursement data for 50 million beneficiaries with access to all OAC 12 13 prescriptions filled in the ambulatory setting; 14 15 • As indications for treatment are not available in the databases, AF was mostly identified on 16 the basis For of discharge peer and long-term review disease diagnosis only codes and an algorithm previously 17 18 validated in the French healthcare databases that helped to further identify AF in 19 outpatients. 20 21 22

23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 3 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 32

1 2 3 Introduction 4 5 6 Direct Oral Anticoagulants (DOACs) have been gradually introduced over the past decade as a more 7 convenient, fixed-dose alternative to vitamin K antagonists (VKAs), the only oral anticoagulant 8 9 therapy available up until now for long-term stroke prevention in patients with non-valvular atrial 10 fibrillation (AF) [1]. Compared to VKA, DOAC therapy avoids the need for regular laboratory 11 12 monitoring of patients by INR testing due to a wider therapeutic window, allows once (rivaroxaban, 13 14 edoxaban) or twice (dabigatran, apixaban) daily intake and is associated with fewer drug-drug 15 interactions to date [2–6]. DOACs have been demonstrated to have similar or superior efficacy to 16 For peer review only 17 warfarin for stroke prevention in nv-AF patients, and these findings have recently been implemented 18 in the 2016 European Society of Cardiology (ESC) guidelines that recommended DOAC over VKA 19 20 therapy in this indication [7]. Moreover, antiplatelet therapy is no longer recommended in these 21 patients [7,8]. The use of DOAC therapy, massively adopted worldwide including in France [9–15], is 22 23 expected to overcome the suboptimal use of oral anticoagulant (OAC) therapy extensively reported 24 25 with VKA therapy, including underprescribing and high discontinuation rates [16,17]. 26 Despite their improved ease of use, DOAC prescribed dose needs to be adjusted to the patient’s 27 28 clinical profile with regards to age, renal function, weight, and risks of bleeding and drug-drug 29 interactions. Two dose regimens are therefore proposed for each DOAC: standard dose regimen (i.e. 30 31 dabigatran 150mg/12h, rivaroxaban 20mg/24h; apixaban 5mg/12h), and reduced-dose regimen (i.e. 32 dabigatran 75mg or 110mg/12h, rivaroxaban 10mg or 15mg/24h; apixaban 2.5mg/12h). The reasons 33 34 for prescribing a reduced-dose regimen are listed in the summary of product characteristics (SmPCs) 35 36 of each DOAC with differences across DOACs as well as in the ESC [7,8] and European Heart Rhythm 37 Association (EHRA) guidelines [18]. Recent publications have suggested that the frequency of use of 38 39 reduced-dose DOACs in clinical practice could largely exceed that expected on the basis of the 40 conditions summarized in these guidelines [11,19–21]. However, national data are lacking concerning 41 42 the current French patterns of DOAC use, including the potential issue of DOAC underdosing. 43 A steady increase in the initiation of DOAC therapy has already been reported in AF patients in 44 45 France [14], but trends in OAC coverage of AF patients following the introduction of DOACs and a 46 47 description of the current national patterns of DOAC use have not yet been reported. This study, 48 based on the French nationwide healthcare databases, therefore had a twofold objective: 1) to 49 50 describe the trends in OAC use following the introduction of DOAC for stroke prevention in AF 51 patients during the 2011-2016 period; 2) to describe the current patterns of use of OAC therapies 52 53 with particular focus on DOAC use in new users with AF. 54 55 56 57 58 4 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 32 BMJ Open

1 2 3 4 Methods 5 6 Data source 7 French national health insurance (Assurance Maladie) covers the entire French population by means 8 9 of several specific schemes according to the beneficiary’s occupational sector, the largest scheme 10 being the ‘Régime général’ (around 50 million beneficiaries). 11 12 This study was conducted using data from the French health insurance system database (SNIIRAM) 13 linked to the French hospital discharge database (PMSI) [22,23]. The SNIIRAM database contains 14 15 individualized, anonymous, and comprehensive data on health spending reimbursements. 16 Demographic dataFor include datepeer of birth, gender, review and vital status. only Drugs are coded according to the 17 18 Anatomical Therapeutic Chemical (ATC) classification. Each packaging of each product is identified by 19 20 means of a national specific pack identifier code providing information on the name of the product, 21 active ingredient and dose per unit, number of units, and route of administration; but, the exact 22 23 dosage prescribed and the indication are not available. However, for most drugs, particularly OACs, 24 each dispensing of the prescribed drug cannot exceed the quantity necessary for one month of 25 26 treatment. The PMSI database provides detailed information about discharge diagnoses and medical 27 procedures related to all hospitalizations in France. This information regarding morbidities is 28 29 completed by diagnoses corresponding to patient eligibility for 100% reimbursement of severe and 30 31 costly long-term diseases (LTD) and disability, such as AF, coronary heart disease, certain debilitating 32 diseases (such as multiple sclerosis or rheumatoid arthritis), HIV infection, cancer, etc. All information 33 34 concerning medical diagnosis is encoded according to the International Classification of Diseases, 35 10th revision (ICD-10). Finally, the SNIIRAM-PMSI databases also indicate medical procedures 36 37 performed in the ambulatory setting, including information about the type and date of all laboratory 38 39 tests performed, but not including their results. 40 The French healthcare databases have been previously described and used in epidemiological and 41 42 pharmacoepidemiological studies [22,24–26]. 43 44 Study populations and study designs 45 A study population was defined for each objective. First, in a repeated cross-sectional study, patients 46 47 with AF were identified for each of the following calendar years: 2011 (no DOAC available for stroke 48 49 prevention in France), 2013 (first calendar year with both rivaroxaban and dabigatran reimbursed for 50 stroke prevention in July 2012) and 2016 (the most recent data available at the time of writing the 51 52 study protocol; dabigatran, rivaroxaban and apixaban were all reimbursed for stroke prevention in 53 France, as apixaban was reimbursed from January 2014 onwards). For each of these calendar years, a 54 55 patient was considered to have AF when at least one diagnosis of AF (ICD-10 code I48) was identified 56 57 58 5 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 32

1 2 3 from discharge and LTD diagnoses in the SNIIRAM-PMSI database in the calendar year considered or 4 during the previous 5 years. Patients with no continuous ‘Régime général’ health insurance coverage 5 6 for at least six years before the calendar year considered were excluded. 7 Second, a retrospective population-based cohort study was performed in patients with AF among 8 9 those initiating DOAC therapy in 2015-2016. To be included in this cohort, patients with continuous 10 11 ‘Régime général’ health insurance coverage had to be identified as DOAC new users: at least one 12 reimbursement for DOAC therapy in 2015-2016 and no reimbursement for any OAC (VKA or DOAC) in 13 14 the previous 24 months. The patient’s index date was the date of first DOAC reimbursement 15 identified during the 2015-2016 period. After exclusion of patients treated for other OAC indications 16 For peer review only 17 i.e. patients treated for deep vein thrombosis/pulmonary embolism (DVT/PE) or with lower limb 18 orthopaedic procedures, DOAC new users treated for AF were identified from the resulting cohort as: 19 20 (a) “DOAC new users with confirmed AF” for those with a diagnosis of AF (ICD-10 code I48) or specific 21 22 AF management procedures identified from LTD or hospitalization discharge data during a six-year 23 pre-index period; and (b) “DOAC new users with probable AF” for outpatients identified using an 24 25 algorithm discriminating AF from DVT/PE with 95% specificity [27]. The remaining patients were not 26 classified as probable FA patients and were excluded. Codes used for identification of AF and all of 27 28 the patient characteristics considered, including comorbidities, are displayed in Supplementary Table 29 1. 30 31 32 Exposure 33 DOAC (dabigatran, rivaroxaban, apixaban) and VKA therapies (fluindione, warfarin and 34 35 acenocoumarol) were identified using ATC codes; edoxaban was not available in France during the 36 period considered. 37 38 39 Outcomes 40 41 Trends in oral anticoagulant therapy use in patients with AF 42 The proportion of AF patients treated by OAC therapy was assessed before and after approval of 43 44 DOAC therapies for stroke prevention in France. Trends in the use of antiplatelet agents were also 45 assessed in AF patients over the same timeframe. 46 47 48 Patterns of use of DOAC therapy in new users with AF 49 The description of patterns of DOAC use in new users treated for AF in 2015-2016 included 50 51 comparison of the baseline characteristics among DOAC new users and compared to those of VKA 52 new users and potential inappropriate use of DOAC therapy was then investigated by identifying: 53 54 (i) DOAC off-label use or non-approved indication/dose: contraindications to DOAC therapy according 55 56 to SmPCs [concomitant coagulopathy, purpura and other haemorrhagic conditions, liver fibrosis and 57 58 6 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 32 BMJ Open

1 2 3 cirrhosis, recent gastrointestinal ulceration or intracranial haemorrhage], valvular atrial fibrillation 4 [DOAC are only approved for non-valvular AF], prosthetic heart valve [contraindicated for 5 6 dabigatran], cancer [DOAC are not approved for prevention of thromboembolism in patients with 7 cancer] and prescription of DOAC doses not approved for stroke prevention in Europe [dabigatran 75 8 9 mg and rivaroxaban 10 mg are not approved for stroke prevention in Europe and are therefore off- 10 11 label doses in AF patients]; (ii) non-compliance with guidelines with respect to follow-up and clinical 12 work-up of patients during the first year following DOAC initiation: no monitoring of patients’ renal 13 14 function [renal function should be assessed at initiation and annually during DOAC therapy [28]], 15 discontinuation of DOAC therapy [OAC therapy is recommended as lifetime treatment in most 16 For peer review only 17 patients with AF] (iii) clinically relevant drug-drug interactions at initiation increasing the bleeding 18 risk at initiation; (iv) potential inappropriate underdosing, i.e. patients in whom the DOAC dose 19 20 prescribed at initiation was inappropriately reduced in view of their individual stroke and bleeding 21 22 risks. 23 24 Data analysis 25 Descriptive analyses are expressed as mean and standard deviation (SD) for continuous variables, 26 27 and numbers and percentages for categorical variables. 28 29 Trends in oral anticoagulant therapy use in patients with AF 30 31 For each calendar year, the proportion of patients treated by a drug was defined by the number of 32 patients with at least one reimbursement for this drug in the calendar year considered over the total 33 34 number of patients identified as having AF in the same year. Proportions are reported according to 35 36 the type of OAC first reimbursed in the year considered. Antiplatelet drugs and OAC therapies were 37 considered to be coprescribed when they were reimbursed at least once on the same day during the 38 39 calendar year studied. Analyses were replicated in subgroups of AF patients: (i) aged 75 years and 40 over; (ii) female and male, separately; (iii) with a history of hospitalization for arterial 41 42 thromboembolic events (ATE); (iv) with concomitant ischaemic heart disease (IHD) or prosthetic 43 heart valve. 44 45 46 Patterns of use of DOAC therapy in new users with AF 47 Baseline characteristics of DOAC new users with AF included sociodemographic data, including 48 49 deprivation index of the patient’s municipality of residence [29], type of initial prescriber, clinical 50 scores predicting the risk of stroke (CHA DS -VASc score) or bleeding (HAS-BLED score) [30,31], 51 2 2 52 adapted to claims data and the other main comorbidities and comedications, including proxies of 53 54 frailty. 55 56 57 58 7 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 32

1 2 3 Compliance with guidelines regarding renal function monitoring and treatment persistence patterns 4 were assessed in new users for whom data for at least one year of follow-up were available, i.e. 5 6 patients included in 2015 and who had not died and had not been hospitalized for 3 months or 7 longer. Compliance with renal function monitoring was assessed at DOAC initiation (no 8 9 reimbursement for renal function monitoring during the three months before and the three months 10 11 after DOAC initiation) and during the first year following treatment initiation. OAC non-persistence 12 patterns were assessed over the one-year period following the index date by calculating proxies of 13 14 OAC discontinuation: number of patients with only one reimbursement and those with five or less 15 reimbursements. 16 For peer review only 17 Drugs increasing the risk of bleeding were those responsible for clinically relevant pharmacodynamic 18 interactions with OAC therapy, i.e. concomitant reimbursement of other parenteral and oral 19 20 antithrombotic drugs, non-steroidal anti-inflammatory drugs (NSAIDs); selective serotonin reuptake 21 22 inhibitors and selective serotonin-norepinephrine reuptake inhibitors (SSRIs and SSNRIs) [7,18,32]. 23 Patients taking concomitant drugs with OAC therapy were defined as those with a reimbursement for 24 25 the drug of interest during the period corresponding to the index date and the following 45 days. 26 Analyses were replicated in VKA new users for descriptive purposes. 27 28 Finally, potential inappropriate underdosing with DOACs was defined as initiation of DOAC therapy in 29 patients at risk of stroke in whom reduced doses of DOAC were prescribed with no identified 30 31 justification. As this study was based on claims data and as, up until 2016, ESC guidelines 32 33 recommended prescribing reduced-dose DOAC in patients with HAS-BLED≥3 [8], the proportion of AF 34 patients initiating reduced-dose DOAC with an HAS-BLED score<3 among all DOAC new users with a 35 36 CHA2DS2-VASc score ≥2 was used to quantify potential inappropriate underdosing in DOAC new 37 users. Analyses were replicated in patients i) with CHA DS -VASc score ≥4, ii) aged 75 and over with a 38 2 2 39 history of ATE. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 8 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 32 BMJ Open

1 2 3 4 Results 5 6 Trends in oral anticoagulant therapy use in patients with AF 7 The number of patients identified from French healthcare databases as having AF increased between 8 9 2011 (N=853,440) and 2016 (N=1,098,657). A high proportion of these patients were identified 10 exclusively by hospitalization discharge diagnosis and this proportion decreased only slightly over this 11 12 time interval from 90.2% (N=770,002) to 86.2% (N=946,657). 13 Between 2011 and 2016, the proportion of patients with at least one reimbursement for OAC 14 15 therapy among all AF patients moderately increased (+16%) from 56.7% to 65.8%, corresponding to a 16 steady decrease inFor VKA use (frompeer 56.6% to review40.8% of all AF patients) only associated with the introduction 17 18 of DOACs (from 0.6% to 27.7%). In 2016, among patients with identified AF, VKA therapy remained 19 20 the preferred OAC therapy (62.0%), including in patients aged 75 years and over (67.0%) and in those 21 with a history of ATE (63.5%). 22 23 Between 2011 and 2016, the use of antiplatelet therapy decreased in AF patients (-22%), but the 24 proportion of patients with concomitant OAC and antiplatelet therapy remained stable (9.7% in 25 26 2016). In 2016, 32.5% of AF patients had at least one reimbursement for antiplatelet therapy during 27 the year. 28 29 Similar trends were observed in the subgroup analyses. A slightly higher rate of OAC therapy was 30 31 observed in patients with a history of ATE (68.4% vs 65.8% in the total cohort in 2016). However, OAC 32 coverage was lower in women than in men with AF (64.7% vs 66.8% in 2016) (Figure 1). 33 34 35 Patterns of use of DOAC therapy in new users with AF 36 37 Baseline characteristics of OAC new users 38 Among 540,914 patients initiating OAC therapy in 2015-2016, a total of 192,851 (35.7%) patients 39 40 were included in the study population, corresponding to 127,841 DOAC new users and 65,010 VKA 41 new users with AF. The mains reasons for ineligibility were other indications or uncertain 42 43 identification of the indication for DOAC (Figure 2). 44 The mean age of the DOAC new users cohort was 74.1±11.6 years; patients over the age of 80 45 46 represented 37.3% of the total cohort. One-half of the DOAC new users were women, mean 47 CHA DS -VASc and HAS-BLED scores were 3.7±1.6 and 2.0±0.9, respectively, and 40.0% received a 48 2 2 49 reduced dose at initiation (62.2% for dabigatran new users). DOAC were mostly initiated by hospital 50 51 practitioners and private cardiologists. Apixaban was the DOAC most commonly initiated; apixaban 52 new users were older and had more comorbidities than the other two groups of DOAC new users. 53 54 VKA new users were older with much more severe disease than each of the three DOAC new users. 55 56 57 58 9 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 32

1 2 3 DOAC therapy was more likely to be initiated than VKA therapy among patients aged 75 years and 4 older (59.4%) and in patients with a history of ATE (57.8%) (Table 1). 5 6 7 Potential inappropriate use of DOAC therapy 8 About 15% of DOAC new users with AF were considered to be using DOAC off-label or for a non- 9 10 approved indication. In particular, 8.5% of DOAC new users with AF had valvular heart disease (8.5%), 11 including prosthetic heart valve (1.5%) and 4.6% had a recently or currently treated cancer (Table 2). 12 13 About 15% and 9% of DOAC new users had no reimbursement for renal function tests at initiation 14 and during the one-year period post-initiation, respectively. Discontinuation during the one-year 15 16 period following For initiation peer was frequent, review as more than 20%only of patients had five or less 17 18 reimbursements (Table 2). 19 Nearly 30% of DOAC new users were using at least one concomitant drug increasing the risk of 20 21 bleeding (52% in VKA new users). The most common concomitant drugs concerned at initiation were 22 antiplatelet agents or parenteral anticoagulants (Table 2). 23 24 Among the 116,391 DOAC new users with AF with a CHA2DS2-VASc score ≥2, 42.9% (N=49,935) of 25 patients were prescribed a reduced dose at initiation, and 29.1% (N=33,845) also had an HAS-BLED 26 27 score <3 meaning that nearly 1 in 3 DOAC new users with AF and at risk of stroke were therefore 28 29 potentially prescribed an inappropriately reduced dose of DOAC at initiation. The proportion of 30 patients potentially underdosed was 33% (N=24,281) in patients with a CHA2DS2-VASc score ≥4 and 31 32 14.5% in patients aged 75 and over with a history of ATE (Figure 3). 33 Among the patients with no criterion justifying dose reduction, i.e. patients with HAS-BLED<3 only, 34 35 these proportions were 39.3%, 51.9% and 58.4%, respectively. Differences in baseline characteristics 36 were observed in patients with HAS-BLED<3 according to the type of DOAC dose prescribed, e.g. 37 38 patients with reduced-dose DOAC were frailer than those with standard-dose DOAC (Supplementary 39 40 Table 2). 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 10 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 32 BMJ Open

1 2 3 4 Discussion 5 6 Main findings 7 OAC therapy use among patients with AF improved in France between 2011 and 2016, but remained 8 9 suboptimal with about 66% of the 1.1 million patients with identified AF treated by OAC therapy in 10 2016. Use of antiplatelet therapy in AF patients decreased over the same period, but still concerned 11 12 33% of all AF patients in 2016. Patients with AF were more likely to be treated by VKA than DOAC 13 therapy, including older patients and those at higher risk of stroke. 14 15 Nearly 193,000 patients with AF were identified as OAC new users in 2015-2016: patients were more 16 likely to be treatedFor by DOAC peer than VKA therapy, review including older onlypatients and those at higher risk of 17 18 stroke. Results on current patterns of use of DOAC therapy in new users with AF suggest several 19 20 situations of inappropriate use, including frequent concomitant use of drugs increasing the risk of 21 bleeding and potential inappropriate underdosing. 22 23 24 Comparison with postmarketing literature and clinical implications 25 The overall improvement of management of AF patients observed with regards to OAC therapy after 26 27 the introduction of DOACs has been reported in many countries [33–36]. The same applies to the 28 steady decrease in VKA use in favour of DOAC therapies [12,37,38], and the gaps remaining in 29 30 optimal OAC coverage and high antiplatelet drug use [39–41]. 31 This study demonstrated channelling of DOAC therapy towards patients at lower risk of stroke and 32 33 bleeding when considering all patients with AF, in line with what has become a common feature 34 reported worldwide in clinical practice by many observational studies on the current patterns of use 35 36 of DOACs [15,42–45]. However, DOAC therapy is now the preferred OAC therapy at initiation in the 37 38 oldest patients and those at higher risk of stroke. This French pattern of OAC use has never been 39 previously reported and contrasts with published results concerning earlier periods [46]. This 40 41 emerging pattern is encouraging for AF management, as older and high-risk patients are those who 42 should derive most benefit from DOAC versus VKA therapy [47]. 43 44 Reduced doses were often prescribed in OAC new users, including dabigatran 75mg and rivaroxaban 45 10mg, which are not approved for stroke prevention in Europe [48]. In addition to the overall 46 47 channelling mentioned above, these findings may reflect a “bleeding avoidance” strategy of 48 49 prescribers (i.e. overestimation of the potential bleeding risk versus the likely benefit of stroke 50 reduction) and the differential perception of the comparative safety of DOACs versus VKA and 51 52 between DOACs. The early safety alert on bleeding in dabigatran-treated patients, followed by the 53 contraindication of this DOAC in patients with prosthetic heart valves, may have reinforced the fears 54 55 of prescribers in relation to the safety of dabigatran, which would explain the difference in reduced- 56 57 58 11 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 32

1 2 3 dose prescription rates between the three DOACs in this study, despite the intermediate stroke and 4 bleeding risk profile of dabigatran compared to that of rivaroxaban and apixaban in new users. 5 6 Similarly, among DOAC new users, apixaban was prescribed to the oldest and most severe patients. 7 Apixaban was the only DOAC found to be superior to warfarin for all types of bleeding outcome and 8 9 all-cause mortality, which would also illustrate the tendency of physicians to prescribe OAC therapies 10 11 according to bleeding risk [5]. This may also explain the potential inappropriate underdosing 12 observed in DOAC users in this study. This pattern of DOAC use has been previously reported, but 13 14 mostly in field and registry studies based on small sample sizes. The reported inappropriate 15 underdosing rate varies according to studies and the definition used. DOAC underdosing concerned 16 For peer review only 17 30.4% of Turkish patients in the RAMSES study (N=2,086) [49], 18.4% of Japanese patients of the KiCS 18 AF registry (N=1,284) [50,51], between 19.7% and 27.6% of patients in the SAKURA AF registry (N= 19 20 3,266) [52], and 9.4% to 16% of patients in the ORBIT-AF II registry (N=7925) [21,53]. In the subgroup 21 22 of Dutch patients (N=899) enrolled in the XANTUS registry, 33% of patients were also treated with 23 reduced-dose rivaroxaban despite presenting normal renal function [54]. Using a large U.S. 24 25 administrative database, Yao et al found that 13.3% of the 13,392 DOAC new users with no renal 26 indication for dose reduction were potentially underdosed [55]. Taken together with our results, 27 28 these data suggest that inappropriate underdosing might be a common issue in DOAC new users that 29 should be systematically assessed when studying DOAC patterns of use. This is of particular concern, 30 31 as recent data have suggested a relationship between DOAC dose and clinical outcomes [56]. In 32 33 particular, DOAC underdosing has been shown to be associated with increased risk for adverse 34 outcomes [21,55]. 35 36 These patterns of DOAC use contrast with the other patterns concomitantly observed in this study, 37 such as the high level of concomitant prescription of antiplatelet agents and parenteral 38 39 anticoagulants or, to a lesser extent, DOAC use in non-approved indications such as prosthetic heart 40 41 valves that are both associated with an increased risk of bleeding [57–59]. 42 43 Strengths and limitations 44 This study is the first to report the improved trend in OAC coverage in French patients with AF over 45 46 the last five years as well as the recent patterns of use of OAC therapy in new users, particularly 47 including a nationwide assessment of the growing issue of DOAC underdosing, based on health data 48 49 for more than 50 million beneficiaries. Moreover, all OAC prescriptions filled in the ambulatory 50 51 setting are captured in the databases and are reimbursed with no restriction of coverage: selection 52 bias related to the access of patients to more expensive DOAC therapy is therefore not an issue with 53 54 the use of French healthcare databases [22,60]. 55 56 57 58 12 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 32 BMJ Open

1 2 3 However, several limitations related to the nature of the data used should be underlined. First of all, 4 it cannot be verified whether patients actually took the drugs for which they were reimbursed. 5 6 Secondly, as the indication for treatment is not available in the databases, and despite the use of an 7 algorithm to identify AF among outpatients in the French healthcare databases, identification of AF 8 9 was mostly based on non-validated discharge and LTD diagnoses recorded in the databases. 10 11 Moreover, it cannot be excluded that the increase in the identified number of patients with AF over 12 the 2011-2016 period could be partially explained by changes in LTD legislation in 2011 (e.g. 13 14 hypertension was removed from the list of LTD, while access to LTD was facilitated for patients with 15 severe arrhythmia and valvular heart diseases) which could have helped identify patients with AF. 16 For peer review only 17 Thirdly, identification of inappropriate underdosing at DOAC initiation was also indirectly assessed by 18 using stroke and bleeding risk scores computed from claims data. Important medical data such as 19 20 patient’s weight, renal function assessment and exact alcohol consumption are not available in the 21 22 French healthcare databases. Furthermore, the agreement between these empirical scores in 23 patients with AF and the prescriber-assessed stroke and bleeding risk is a subject of discussion [61]. 24 25 However, DOAC misuse and underdosing have also been reported in a French prospective field study 26 based on patients’ medical charts [62]. Of note, as INR values were not available in the databases, 27 28 underdosing with VKA therapy was not assessed in this study, but has been frequently reported and 29 must not be overlooked [52,63]. In addition, the results for DOAC and VKA new users are difficult to 30 31 compare, as they were not adjusted for significant differences in baseline characteristics and this 32 33 comparison was not the purpose of this study. 34 35 36 37 Conclusion 38 OAC therapy use has modestly increased after the introduction of the DOACs for stroke prevention in 39 40 patients with AF in France and DOAC therapy is now the preferred OAC therapy at initiation in older 41 patients at higher risk of stroke. However, results from this nationwide drug utilization study suggest 42 43 the need for improvement in appropriate prescription of OAC therapy in these patients, especially 44 regarding the use of concomitant interacting drugs and the choice of initial DOAC dose. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 13 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 32

1 2 3 Contributor ship statement 4 GM, AP, CB and AN designed the study. All authors have contributed substantially to the 5 6 interpretation of results. In addition: GM, AP and AN drafted the article; CB, JD and AN conducted 7 the statistical analysis; GM, AP, AN, CB, JD, AW provided critical revision of the manuscript for 8 9 important intellectual content. All authors had full access to all of the data (including statistical 10 reports and tables) in the study and can take responsibility for the integrity of the data and the 11 12 accuracy of the data analysis. All authors approved the version to be published. 13 14 15 16 Competing interestsFor peer review only 17 All authors have no conflict of interest. 18 19 GM, CB, JD, AN and AW are employees of the French National Health Insurance (Cnam, Assurance 20 21 Maladie); AP belongs to the French National Institute of Health and Medical Research (Inserm). 22 All authors have no conflicts of interest with the Pharmaceutical Industry. 23 24 25 26 Funding 27 28 The authors received no funding. 29 30 31 Data sharing statement 32 Data are available from the French National Health Insurance (Caisse National de l’Assurance 33 34 Maladie, Cnam) for academic research and the data access permission policy prohibits making the 35 data set publicly available. All databases used in this study only contained anonymous patient 36 37 records. 38 39 40 41 Ethics approval 42 The study was approved by the French data protection agency (Commission Nationale de 43 44 l'Informatique et des Libertés, Cnil). 45

46 47 48 Acknowledgments 49 The authors thank Dr Saul, medical translator, for assistance in writing the manuscript. 50 51 52

53 54 55 56 57 58 14 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 32 BMJ Open

1 2 3 References 4 1 Steinberg BA, Piccini JP. Anticoagulation in atrial fibrillation. BMJ 2014;348:g2116–g2116. 5 doi:10.1136/bmj.g2116 6 7 2 Verheugt FWA, Granger CB. Oral anticoagulants for stroke prevention in atrial fibrillation: current 8 status, special situations, and unmet needs. The Lancet 2015;386:303–10. doi:10.1016/S0140- 9 6736(15)60245-8 10 11 3 Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus Warfarin in Patients with Atrial 12 Fibrillation. N Engl J Med 2009;361:1139–51. doi:10.1056/NEJMoa0905561 13 14 4 Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus Warfarin in Nonvalvular Atrial 15 Fibrillation. N Engl J Med 2011;365:883–91. doi:10.1056/NEJMoa1009638 16 For peer review only 17 5 Granger CB, Alexander JH, McMurray JJV, et al. Apixaban versus Warfarin in Patients with Atrial 18 Fibrillation. N Engl J Med 2011;365:981–92. doi:10.1056/NEJMoa1107039 19 20 6 Giugliano RP, Ruff CT, Braunwald E, et al. Edoxaban versus Warfarin in Patients with Atrial 21 Fibrillation. N Engl J Med 2013;369:2093–104. doi:10.1056/NEJMoa1310907 22 23 7 Kirchhof P, Benussi S, Kotecha D, et al. 2016 ESC Guidelines for the management of atrial 24 fibrillation developed in collaboration with EACTS. Eur Heart J 2016;37:2893–962. 25 doi:10.1093/eurheartj/ehw210 26 27 8 Authors/Task Force Members, Camm AJ, Lip GYH, et al. 2012 focused update of the ESC 28 29 Guidelines for the management of atrial fibrillation: An update of the 2010 ESC Guidelines for 30 the management of atrial fibrillation * Developed with the special contribution of the European 31 Heart Rhythm Association. Eur Heart J 2012;33:2719–47. doi:10.1093/eurheartj/ehs253 32 33 9 Desai NR, Krumme AA, Schneeweiss S, et al. Patterns of Initiation of Oral Anticoagulants in 34 Patients with Atrial Fibrillation— Quality and Cost Implications. Am J Med 2014;127:1075– 35 1082.e1. doi:10.1016/j.amjmed.2014.05.013 36 37 10 Barnes GD, Lucas E, Alexander GC, et al. National Trends in Ambulatory Oral Anticoagulant Use. 38 Am J Med 2015;128:1300–1305.e2. doi:10.1016/j.amjmed.2015.05.044 39 40 11 Haastrup S, Hellfritzsch M, Rasmussen L, et al. Use of Non-Vitamin K Antagonist Oral 41 Anticoagulants 2008-2016: A Danish Nationwide Cohort Study. Basic Clin Pharmacol Toxicol 42 Published Online First: 17 April 2018. doi:10.1111/bcpt.13024 43 44 12 Kjerpeseth LJ, Ellekjær H, Selmer R, et al. Trends in use of warfarin and direct oral anticoagulants 45 in atrial fibrillation in Norway, 2010 to 2015. Eur J Clin Pharmacol Published Online First: 22 July 46 2017. doi:10.1007/s00228-017-2296-1 47 48 13 Staerk L, Fosbøl EL, Gadsbøll K, et al. Non-vitamin K antagonist oral anticoagulation usage 49 according to age among patients with atrial fibrillation: Temporal trends 2011–2015 in Denmark. 50 Sci Rep 2016;6. doi:10.1038/srep31477 51 52 14 Huiart L, Ferdynus C, Renoux C, et al. Trends in initiation of direct oral anticoagulant therapies 53 for atrial fibrillation in a national population-based cross-sectional study in the French health 54 insurance databases. BMJ Open 2018;8:e018180. doi:10.1136/bmjopen-2017-018180 55 56 57 58 15 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 32

1 2 3 15 Urbaniak AM, Strøm BO, Krontveit R, et al. Prescription Patterns of Non-Vitamin K Oral 4 Anticoagulants Across Indications and Factors Associated with Their Increased Prescribing in 5 Atrial Fibrillation Between 2012–2015: A Study from the Norwegian Prescription Database. Drugs 6 Aging 2017;34:635–45. doi:10.1007/s40266-017-0476-4 7 8 16 Broderick JP, Bonomo JB, Kissela BM, et al. Withdrawal of Antithrombotic Agents and Its Impact 9 on Ischemic Stroke Occurrence. Stroke 2011;42:2509–14. doi:10.1161/STROKEAHA.110.611905 10 11 17 Fang MC, Go AS, Chang Y, et al. Warfarin Discontinuation After Starting Warfarin for Atrial 12 Fibrillation. Circ Cardiovasc Qual Outcomes 2010;3:624–31. 13 doi:10.1161/CIRCOUTCOMES.110.937680 14 15 18 Heidbuchel H, Verhamme P, Alings M, et al. Updated European Heart Rhythm Association 16 practical guideFor on the use peer of non-vitamin-K review antagonist anticoagulants only in patients with non- 17 valvular atrial fibrillation: Executive summary. Eur Heart J 2016;:ehw058. 18 doi:10.1093/eurheartj/ehw058 19 20 19 Dillinger J-G, Aleil B, Cheggour S, et al. Dosing issues with non-vitamin K antagonist oral 21 anticoagulants for the treatment of non-valvular atrial fibrillation: Why we should not underdose 22 our patients. Arch Cardiovasc Dis 2018;111:85–94. doi:10.1016/j.acvd.2017.04.008 23 24 20 Marzec LN, Wang J, Shah ND, et al. Influence of Direct Oral Anticoagulants on Rates of Oral 25 Anticoagulation for Atrial Fibrillation. J Am Coll Cardiol 2017;69:2475–84. 26 doi:10.1016/j.jacc.2017.03.540 27 28 21 Steinberg BA, Shrader P, Thomas L, et al. Off-Label Dosing of Non-Vitamin K Antagonist Oral 29 Anticoagulants and Adverse Outcomes. J Am Coll Cardiol 2016;68:2597–604. 30 doi:10.1016/j.jacc.2016.09.966 31 32 22 Tuppin P, Rudant J, Constantinou P, et al. Value of a national administrative database to guide 33 public decisions: From the système national d’information interrégimes de l’Assurance Maladie 34 (SNIIRAM) to the système national des données de santé (SNDS) in France. Rev DÉpidémiologie 35 36 Santé Publique 2017;65:S149–67. doi:10.1016/j.respe.2017.05.004 37 38 23 Bezin J, Duong M, Lassalle R, et al. The national healthcare system claims databases in France, 39 SNIIRAM and EGB: Powerful tools for pharmacoepidemiology. Pharmacoepidemiol Drug Saf 40 Published Online First: 24 May 2017. doi:10.1002/pds.4233 41 42 24 Maura G, Blotière P-O, Bouillon K, et al. Comparison of the Short-Term Risk of Bleeding and 43 Arterial Thromboembolic Events in Nonvalvular Atrial Fibrillation Patients Newly Treated With 44 Dabigatran or Rivaroxaban Versus Vitamin K Antagonists: A French Nationwide Propensity- 45 Matched Cohort Study. Circulation 2015;132:1252–60. 46 doi:10.1161/CIRCULATIONAHA.115.015710 47 48 25 Weill A, Dalichampt M, Raguideau F, et al. Low dose oestrogen combined oral contraception and 49 risk of pulmonary embolism, stroke, and myocardial infarction in five million French women: 50 cohort study. BMJ 2016;353:i2002. 51 52 26 Neumann A, Maura G, Weill A, et al. Clinical Events After Discontinuation of β-Blockers in 53 Patients Without Heart Failure Optimally Treated After Acute Myocardial Infarction: A Cohort 54 Study on the French Healthcare Databases. Circ Cardiovasc Qual Outcomes 2018;11:e004356. 55 doi:10.1161/CIRCOUTCOMES.117.004356 56 57 58 16 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 32 BMJ Open

1 2 3 27 Billionnet C, Alla F, Bérigaud É, et al. Identifying atrial fibrillation in outpatients initiating oral 4 anticoagulants based on medico-administrative data: results from the French national healthcare 5 databases: Identification of Atrial Fibrillation in Claims. Pharmacoepidemiol Drug Saf 6 2017;26:535–43. doi:10.1002/pds.4192 7 8 28 Haute autorité de santé. Guide parcours de soins Fibrillation atriale. 2014.https://www.has- 9 sante.fr/portail/jcms/c_1741768/fr/guide-parcours-de-soins-fibrillation-atriale (accessed 18 Jul 10 2018). 11 12 29 Rey G, Jougla E, Fouillet A, et al. Ecological association between a deprivation index and 13 mortality in France over the period 1997 – 2001: variations with spatial scale, degree of 14 urbanicity, age, gender and cause of death. BMC Public Health 2009;9:33. doi:10.1186/1471- 15 2458-9-33 16 For peer review only 17 30 Olesen JB, Lip GYH, Hansen ML, et al. Validation of risk stratification schemes for predicting 18 stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study. BMJ 19 2011;342:d124–d124. doi:10.1136/bmj.d124 20 21 31 Pisters R, Lane DA, Nieuwlaat R, et al. A novel user-friendly score (HAS-BLED) to assess 1-year risk 22 of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest 2010;138:1093– 23 100. doi:10.1378/chest.10-0134 24 25 32 Renoux C, Vahey S, Dell’Aniello S, et al. Association of Selective Serotonin Reuptake Inhibitors 26 With the Risk for Spontaneous Intracranial Hemorrhage. JAMA Neurol 2017;74:173. 27 doi:10.1001/jamaneurol.2016.4529 28 29 33 Cowan JC, Wu J, Hall M, et al. A 10 year study of hospitalized atrial fibrillation-related stroke in 30 England and its association with uptake of oral anticoagulation. Eur Heart J Published Online 31 First: 5 July 2018. doi:10.1093/eurheartj/ehy411 32 33 34 Steinberg BA, Gao H, Shrader P, et al. International trends in clinical characteristics and oral 34 anticoagulation treatment for patients with atrial fibrillation: Results from the GARFIELD-AF, 35 36 ORBIT-AF I, and ORBIT-AF II registries. Am Heart J 2017;194:132–40. 37 doi:10.1016/j.ahj.2017.08.011 38 39 35 Brown JD, Shewale AR, Dherange P, et al. A Comparison of Oral Anticoagulant Use for Atrial 40 Fibrillation in the Pre- and Post-DOAC Eras. Drugs Aging 2016;33:427–36. doi:10.1007/s40266- 41 016-0369-y 42 43 36 Apenteng PN, Gao H, Hobbs FR, et al. Temporal trends in antithrombotic treatment of real-world 44 UK patients with newly diagnosed atrial fibrillation: findings from the GARFIELD-AF registry. BMJ 45 Open 2018;8:e018905. doi:10.1136/bmjopen-2017-018905 46 47 37 Alalwan AA, Voils SA, Hartzema AG. Trends in utilization of warfarin and direct oral 48 anticoagulants in older adult patients with atrial fibrillation. Am J Health Syst Pharm 49 2017;74:1237–44. doi:10.2146/ajhp160756 50 51 38 Olesen JB, Sorensen R, Hansen ML, et al. Non-vitamin K antagonist oral anticoagulation agents in 52 anticoagulant naive atrial fibrillation patients: Danish nationwide descriptive data 2011-2013. 53 Europace 2015;17:187–93. doi:10.1093/europace/euu225 54 55 56 57 58 17 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 32

1 2 3 39 Lacoin L, Lumley M, Ridha E, et al. Evolving landscape of stroke prevention in atrial fibrillation 4 within the UK between 2012 and 2016: a cross-sectional analysis study using CPRD. BMJ Open 5 2017;7:e015363. doi:10.1136/bmjopen-2016-015363 6 7 40 Averlant L, Ficheur G, Ferret L, et al. Underuse of Oral Anticoagulants and Inappropriate 8 Prescription of Antiplatelet Therapy in Older Inpatients with Atrial Fibrillation. Drugs Aging 9 2017;34:701–10. doi:10.1007/s40266-017-0477-3 10 11 41 Alamneh EA, Chalmers L, Bereznicki LR. Suboptimal Use of Oral Anticoagulants in Atrial 12 Fibrillation: Has the Introduction of Direct Oral Anticoagulants Improved Prescribing Practices? 13 Am J Cardiovasc Drugs 2016;16:183–200. doi:10.1007/s40256-016-0161-8 14 15 42 Douros A, Renoux C, Coulombe J, et al. Patterns of long-term use of non-vitamin K antagonist 16 oral anticoagulantsFor for non-valvularpeer atrial review fibrillation: Quebec only observational study. 17 Pharmacoepidemiol Drug Saf 2017;26:1546–54. doi:10.1002/pds.4333 18 19 43 Komen J, Forslund T, Hjemdahl P, et al. Factors associated with antithrombotic treatment 20 decisions for stroke prevention in atrial fibrillation in the Stockholm region after the introduction 21 of NOACs. Eur J Clin Pharmacol 2017;73:1315–22. doi:10.1007/s00228-017-2289-0 22 23 44 Patel PA, Zhao X, Fonarow GC, et al. Novel Oral Anticoagulant Use Among Patients With Atrial 24 Fibrillation Hospitalized With Ischemic Stroke or Transient Ischemic Attack. Circ Cardiovasc Qual 25 Outcomes 2015;8:383–92. doi:10.1161/CIRCOUTCOMES.114.000907 26 27 45 Lauffenburger JC, Farley JF, Gehi AK, et al. Factors Driving Anticoagulant Selection in Patients 28 With Atrial Fibrillation in the United States. Am J Cardiol 2015;115:1095–101. 29 doi:10.1016/j.amjcard.2015.01.539 30 31 46 Maura G, Billionnet C, Alla F, et al. Comparison of Treatment Persistence with Dabigatran or 32 Rivaroxaban versus Vitamin K Antagonist Oral Anticoagulants in Atrial Fibrillation Patients: A 33 Competing Risk Analysis in the French National Health Care Databases. Pharmacother J Hum 34 Pharmacol Drug Ther 2018;38:6–18. doi:10.1002/phar.2046 35 36 47 Ruff CT, Giugliano RP, Braunwald E, et al. Comparison of the efficacy and safety of new oral 37 38 anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised 39 trials. The Lancet 2014;383:955–62. doi:10.1016/S0140-6736(13)62343-0 40 41 48 Sorensen HT, Riis AH, Lash TL, et al. Statin Use and Risk of Amyotrophic Lateral Sclerosis and 42 Other Motor Neuron Disorders. Circ Cardiovasc Qual Outcomes 2010;3:413–7. 43 doi:10.1161/CIRCOUTCOMES.110.936278 44 45 49 Başaran Ö, Dogan V, Beton O, et al. Suboptimal use of non-vitamin K antagonist oral 46 anticoagulants: Results from the RAMSES study. Medicine (Baltimore) 2016;95:e4672. 47 doi:10.1097/MD.0000000000004672 48 49 50 Ono T, Kohsaka S, Takatsuki S, et al. Inconsistent Dosing of Non–Vitamin K Oral Anticoagulants. J 50 Am Coll Cardiol 2017;70:118. doi:10.1016/j.jacc.2017.03.609 51 52 51 Hsu JC, Akao M, Abe M, et al. International Collaborative Partnership for the Study of Atrial 53 Fibrillation (INTERAF): Rationale, Design, and Initial Descriptives. J Am Heart Assoc 54 2016;5:e004037. doi:10.1161/JAHA.116.004037 55 56 57 58 18 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 32 BMJ Open

1 2 3 52 Okumura Y, Yokoyama K, Matsumoto N, et al. Current use of direct oral anticoagulants for atrial 4 fibrillation in Japan: Findings from the SAKURA AF Registry. J Arrhythmia 2017;33:289–96. 5 doi:10.1016/j.joa.2016.11.003 6 7 53 Steinberg BA, Shrader P, Pieper K, et al. Frequency and Outcomes of Reduced Dose Non–Vitamin 8 K Antagonist Anticoagulants: Results From ORBIT-AF II (The Outcomes Registry for Better 9 Informed Treatment of Atrial Fibrillation II). J Am Heart Assoc 2018;7:e007633. 10 doi:10.1161/JAHA.117.007633 11 12 54 Pisters R, van Vugt SPG, Brouwer MA, et al. Real-life use of Rivaroxaban in the Netherlands: data 13 from the Xarelto for Prevention of Stroke in Patients with Atrial Fibrillation (XANTUS) registry. 14 Neth Heart J 2017;25:551–8. doi:10.1007/s12471-017-1009-9 15 16 55 Yao X, Shah ND,For Sangaralingham peer LR, et al.review Non–Vitamin K Antagonist only Oral Anticoagulant Dosing 17 in Patients With Atrial Fibrillation and Renal Dysfunction. J Am Coll Cardiol 2017;69:2779–90. 18 doi:10.1016/j.jacc.2017.03.600 19 20 56 Reilly PA, Lehr T, Haertter S, et al. The Effect of Dabigatran Plasma Concentrations and Patient 21 Characteristics on the Frequency of Ischemic Stroke and Major Bleeding in Atrial Fibrillation 22 Patients. J Am Coll Cardiol 2014;63:321–8. doi:10.1016/j.jacc.2013.07.104 23 24 57 Bouillon K, Bertrand M, Boudali L, et al. Short-Term Risk of Bleeding During Heparin Bridging at 25 Initiation of Vitamin K Antagonist Therapy in More Than 90 000 Patients With Nonvalvular Atrial 26 Fibrillation Managed in Outpatient Care. J Am Heart Assoc 2016;5:e004065. 27 doi:10.1161/JAHA.116.004065 28 29 58 Chang S-H, Chou I-J, Yeh Y-H, et al. Association Between Use of Non–Vitamin K Oral 30 Anticoagulants With and Without Concurrent Medications and Risk of Major Bleeding in 31 Nonvalvular Atrial Fibrillation. JAMA 2017;318:1250. doi:10.1001/jama.2017.13883 32 33 59 Van de Werf F, Brueckmann M, Connolly SJ, et al. A comparison of dabigatran etexilate with 34 warfarin in patients with mechanical heart valves: The Randomized, phase II study to Evaluate 35 36 the sAfety and pharmacokinetics of oraL dabIGatran etexilate in patients after heart valve 37 replacemeNt (RE-ALIGN). Am Heart J 2012;163:931–937.e1. doi:10.1016/j.ahj.2012.03.011 38 39 60 Steffen M. Universalism, Responsiveness, Sustainability — Regulating the French Health Care 40 System. N Engl J Med 2016;374:401–5. doi:10.1056/NEJMp1504547 41 42 61 Steinberg BA, Kim S, Thomas L, et al. Lack of Concordance Between Empirical Scores and 43 Physician Assessments of Stroke and Bleeding Risk in Atrial Fibrillation: Results From the 44 Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) Registry. 45 Circulation 2014;129:2005–12. doi:10.1161/CIRCULATIONAHA.114.008643 46 47 62 Lafon T, Vallejo C, Hadj M, et al. Mésusage et iatrogénie des anticoagulants oraux directs (AOD) : 48 étude observationnelle dans le service des urgences du CHU de Limoges. Thérapie Published 49 Online First: July 2017. doi:10.1016/j.therap.2017.05.004 50 51 63 Pokorney SD, Simon DN, Thomas L, et al. Stability of International Normalized Ratios in Patients 52 Taking Long-term Warfarin Therapy. JAMA 2016;316:661–3. doi:10.1001/jama.2016.9356 53 54 55 56 57 58 19 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 32

1 2 3 Figure titles and legends 4 5 6 Figure 1. Time trends in the use of oral antithrombotic therapy between 2011 and 2016 in patients 7 with atrial fibrillation in France 8 9 Abbreviations: AF: atrial fibrillation; OAC: oral anticoagulant; VKA: vitamin K antagonist; DOAC: direct 10 oral anticoagulant 11 12 Legend: For Figures 1.b., 1.d and 1.e., estimates from all AF patients already presented in figure 1.a. 13 14 are indicated by dashed lines for the purposes of comparison. 15 16 For peer review only 17 Figure 2. Patient flow chart. 18 Abbreviations: OAC: oral anticoagulant; DOAC: direct oral anticoagulant; VKA: vitamin K antagonist; 19 20 DVT/PE: deep vein thrombosis/pulmonary embolism; AF: atrial fibrillation 21

22 23 Figure 3. Potential DOAC underdosing in new users with AF 24 25 Abbreviations: DOAC: direct oral anticoagulant; AF: atrial fibrillation 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 20 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 21

)

(11.3)

NA VKA .0 680 (8.7) 680 (23.9) 539 (91 139 73 (0.1) , (14.7) 587 (12.9) 399 , , 4.5 (1.6) (1.6) 4.5 309 (0.5) 309 , , 263 (15.8) 263 (19.7) 811 (22.6) 699 N= 65,010 N= 65,010 78 5 1,607 (2.5) 1,607 (2.5) , , , 9 8 7,762 (11.9) 7,762 (11.9) 7,035 (10.8) 15 (97.0) 63,094 59 33,865 (52.1) (52.1) 33,865 (50.3) 32,727 (28.9) 18,806 (16.4) 10,638 (44.4) 28,894 10 12 14 11,884 (18.3) 14,272 (22.0) 1,081 (1.7)

)

) .

.0 0 .0

391 (9391 (83.2) 403 084 (40084 (12.4) 885 (16.7) 299 (6706 (19.4) 797 (1.8) 324 , , , , , , , 3.7 (1.6) 3.7 (1.6) 211 (0.2) 211 925 (17.9) 925 (19.8) 280 (21.2) 110 74.1 (11.6) 74.1 7 2 9,126 (7.1) (7.1) 9,126 (7.4) 9,503 , , , N= 127,841 N= 127,841 Total DOAC DOAC Total 51 15 21 24 64,068 (50.1) (50.1) 64,068 (34.6) 44,202 (21.4) 27,325 (11.4) 14,585 (27.3) 34,932 (17.2) 21,980 116, 106 22 25 27 (19.4) 24,760 (20.2) 25,853 1,913 (1.5)

) )

.0 .0

003 (40.4) 003 (16.6) 692 (7.2) 654 (18 590 (84.9) 596 95 (0.1) , (11.1) 117 , , , , 3.9 (1.6) (1.6) 3.9 847 (1.3) , Apixaban Apixaban 266 (17.5) 266 (19.9) 799 (21.5) 825 N= 64,300 64,300 N= 75.3 (11.3)75.3 4 3,637 (5.7) 3,637 (5.7) 5,608 (8.7) , , , 7 8,448 (13.1) 8,448 (13.1) 9,396 (14.6) 26 10 11 59,816 (93 54 33,375 (51.9) 33,375 (51.9) 23,548 (36.6) 14,087 (21.9) 18,766 (29.2)

11 12 13 12,496 (19.4) 13,142 (20.4) (1.2) 772 BMJ Open DOAC

)

429 (35.7) 429 (4.7) 559 (20.5) 145 (2.4) 294 (81.3) 260 , (14651 (16.6) 053 , , , , 3.5 (1.6) 3.5 (1.6) 100 (0.2) 100 , , 265 (18.9) 265 (19.7) 701 (20.5) 172 N= 54,456 N= 54,456 72.8 (11.9) 72.8 2 1 4,854 (8.9) (8.9) 4,854 (9.1) 4,930 (5.7) 3,126 , , , 7 9 Rivaroxaban Rivaroxaban 19 11 (88.7) 48,308 44 26,147 (48.0) (48.0) 26,147 (32.7) 17,805 (20.7) 11,279 (25.6) 13,924 (20.3) 11,074 10 10 11 (19.6) 10,678 (19.8) 10,794 (1.6) 846

) .0

16 (0.2) 652 (62.2) 652 (12.3) 117 (17.1) 554 (22.7) 062 (83.1) 547 3.7 (1.6) (1.6) 3.7 548 (6.0) 548 (6.0) (7.4) 4,007 3,172 (4.9) (6.0) 7,727 2,361 (3.6) 493 (5.4) 183 (2 635 (7.0) 635 (7.0) 769 (8.5) 168 (1.8) 168 (1.8) (2.1) 1,149 1,276 (2.0) (2.0) 2,593 1,583 (2.4) N= 9,085 9,085 N= , , , , , 74.1 (11.6)74.1 Dabigatran Dabigatran 2,514 (27.7) 2,514 (27.7) 3,352 (36.9) (29.5) 16,057 16,645 (25.9) (32.5) 17,688 (27.5) 35,216 26,674 (41.5) (19.9) 12,969 (37.3) 47,714 (52.9) 34,413 5 1 1 2 8,267 (91.0) 7 4,546 (50.0) 4,546 (50.0) 2,849 (31.4) 1,959 (21.6) 1,207 (13.3) 2,242 (24.7) 1,510 (16.6) 3,093 (34.0) 3,093 (34.0) (34.9) 18,978 21,843 (34.0) (34.4) 43,914 8,511 (13.1) 3,720 (40.9) 3,720 (40.9) (42.1) 22,905 29,316 (45.6) (43.8) 55,941 (60.1) 39,083 394 (15.3) 394 (19.6) 780 (23.3) 113 , , , 1 1 2 1,586 (17.5) 1,917 (21.1) (3.2) 295

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml †

For peer review only †

† deprived)

Characteristics (N; %*) Characteristics %*) (N; mean (SD)mean practitioner

orthopaedic surgeon orthopaedic ,

VASc score VASc - 2

64 64 79

- - DS prescriber’s specialty prescriber’s 2 18-54 18-54 65-74 ≥80 Hospital practitioner Private cardiologist specialist Other private 55 75 ≥90 1 (leastQuintile 3 Quintile 5 (most deprived)) Quintile General Private 0 2 ≥ drugs) H (antihypertensive Quintile 2 Quintile 4 Quintile Overseas departments (SD) score Mean 1 (Heart C failure) D(iabetes) ATE) S(troke: V(ascular diseases) DOAC: reduced doses DOAC:reduced Age (years) First CHA Female sex sex Female index Deprivation events thromboembolic arterial Age≥75 and events thromboembolic arterial no Age<65 and Table 1. Baseline characteristics of anticoagulant-naive patients with atrial fibrillation initiating oral anticoagulants in 2015-2016 2015-2016 in anticoagulants oral initiating fibrillation atrial with patients ofanticoagulant-naive characteristics Baseline 1. Table

Page 21 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 22 of 32 22

VKA 923 (4.5) 923 (36.4) 657 (54.8) 600 (69.5) 153 348 (14.4) 348 , (15.1) 824 , , (13.8) 967 , 212 (0.3) 212 , , , N= 65,010 N= 65,010 2 9 9 8 7,437 (11.4) 7,437 (11.4) 23 35 45 24,175 (37.2) (37.2) 24,175 (25.5) 16,598 (17.6) 11,434

)

756 (0.6) 756 (0.6) 781 9,555 (7.5) 3,700 (2.9) 7,853 (6.1) 7,853 (6.1) N= 127,841 N= 127,841 Total DOAC DOAC Total 28,581 (22.4) 28,581 (64.8) 82,788 (12.6) 16,148 (4253,671 23,367 (18.3) (18.3) 23,367 (19.7) 25,159 (11.7) 14,947

5,134 (8) 361 (0.6) 344 (0.5) Apixaban Apixaban N= 64,300 64,300 N= 1,730 (2.7) 4,125 (6.4) 4,125 (6.4) 8,079 (12.6) 7,442 (11.6) 7,442 (11.6) 15,439 (24.0) 41,031 (63.8) 28,196 (43.9) 12,730 (19.8) 12,730 (19.8) 12,844 (20.0) BMJ Open DOAC DOAC

321 (20.8) 321 (65.7) 761 (39.9) 725 , , (12.8) 964 , 331 (0.6) 331 (0.7) 365 , N= 54,456 N= 54,456 3,729 (6.8) 3,729 (3.1) 1,698 3,204 (5.9) (5.9) 3,204 6 8,971 (16.5) (16.5) 8,971 (11.9) 6,481 Rivaroxaban Rivaroxaban 11 35 21 10,593 (19.5) (19.5) 10,593

64 (0.7) 72 (0.8) 2.0 (0.9) (0.9) 2.0 1.9 (0.9) (0.9) 2.1 2.0 (0.9) (1.0) 2.7 692 (7.6) 272 (3.0) 524 (5.8) 524 (5.8) 826 (9.1) 826 (9.1) (9.6) 5,225 6,951 (10.8) (10.2) 13,002 (23.7) 15,433 169 (1.9) 169 (1.9) (1.8) 967 1,106 (1.7) 2,242 (1.8) 2,372 (3.6) 738 (8.1) 738 (8.1) (8.5) 4,653 6,246 (9.7) (9.1) 11,637 (23.0) 14,947 182 (2.0) 182 (2.0) (2.1) 1,161 1,605 (2.5) 2,948 (2.3) 5,873 (9.0) 345 (3.8) 345 (3.8) (4.5) 2,426 3,822 (5.9) 6,593 (5.2) (21.9) 14,260 N= 9,085 9,085 N= Dabigatran Dabigatran 2,169 (23.9) 2,169 (23.9) (20.9) 11,388 16,834 (26.2) (23.8) 30,391 (56.0) 36,417 1,821 (20.0) 5,996 (66.0) 1,105 (12.2) 3,750 (41.3) 1,666 (18.3) 1,666 (18.3) 1,722 (19.0) 1,024 (11.3) 4,295 (47.3) 4,295 (47.3) (45.6) 24,842 31,469(48.9) (47.4) 60,606 (61.3) 39,842 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only

)

ATC classes

‡

≥5 †

†

§ Characteristics (N; %) Characteristics %) (N;

‡ Parenteral anticoagulant (heparin) anticoagulant Parenteral NSAIDs Alcohol abuse Alcohol Major bleeding B(leeding) B(leeding) D(rug/alcohol) D(rug/alcohol) interactions Drug-drug Antiplatelet drugs (10.4) 947 5,838 (10.7) agents Lipid-lowering 7,570 (11.8) Antiulcer agents (11.2) 14,355 ATC (≥10 classes) Polymedication (36.1) 23,451 3,913 (43.1) (40.9) 22,250 28,812 (44.8) (43.0) 54,975 (49.1) 31,903 disease Ischaemic heart glycosides or cardiac Antiarrhythmics Oral corticosteroids ( Polymedication Frailty (proxies) Frailty (proxies) disease Dementia or Parkinson's disorders Psychiatric Smoking Predisposition Liver function A(bnormal) A(bnormal) function Renal Unless otherwise stated stated Unlessotherwise

HAS-BLED score (SD) score Mean 3 ≥ Other comorbidities

Comedications

† Comorbidities were defined using a rolling 1-year period following the initiation of OAC therapy therapy OAC of initiation the following period 1-year rolling a using defined were Comorbidities † consumption/diseases alcohol smoking or to related hospitalizations or therapy specific for suchas reimbursements proxies by using measured data: alcohol or Smoking ‡ Table 1. Baseline characteristics of oral anticoagulant-naive patients with atrial fibrillation initiating oral anticoagulants in 2015-2016inanticoagulants oral initiating fibrillation withatrial patients ofanticoagulant-naive oral characteristics Baseline 1. Table non- NSAIDs: attack); ischaemic transient or embolism systemic arterial stroke, (ischaemic events thromboembolic arterial ATE: deviation; SD:standard antagonist; K vitamin VKA: anticoagulant; oral direct DOAC: drugs anti-inflammatory steroidal * § Comorbidities were defined using a rolling 4-month period preceding the initiation of OAC therapy therapy OAC of initiation the preceding period 4-month rolling a using defined were Comorbidities § (continued) (continued)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 23

VKA 350 (0.5) 350 N= 65,010 N= 65,010 2,984 (9.4) 2,984 (9.4) 2,426 (7.6) 1,030 (1.6) 5401 (17.0) (17.0) 5401 7,317 (11.6) 7,456 (23.5) 7,456 (23.5) 6,726 (10.3) 6,726 (10.3) 28,112 (44.5) (44.5) 28,112 (19.1) 12,078 (31.2) 19,710 (28.1) 17,770 33,025 (52.3) (52.3) 33,025

408 (0.3) 408 4,863 (9.1) (9.1) 4,863 (9.4) 5,042 3,384 (2.7) 4,503 (3.6) 8083 (15.1) (15.1) 8083 N= 127,841 N= 127,841 Total DOAC DOAC Total 11,568 (9.1) 11,568 12,301 (23.0) (23.0) 12,301 (19.4) 24,496 (17.3) 21,839 (15.5) 19,577 36,992 (29.3) (29.3) 36,992

248 (0.4) 248 Apixaban Apixaban N= 64,300 N= 64,300 5,880 (9.2) 2,138 (8.5) 2,138 (8.5) 1,771 (7.1) 1,520 (2.4) 2,017 (3.2) 3642 (14.5) (14.5) 3642 4,515 (18.0) 4,515 (18.0) 12,382 (19.5) 12,382 (19.5) 11,170 (17.6) 10,026 (15.8)

18,556 (29.2) (29.2) 18,556 DOAC

(unless otherwise stated): rolling 1-year period following the initiation of OAC therapy (index date date (index therapy OAC of initiation the following period 1-year rolling stated): (unlessotherwise

120 (0.2) 120 N= 54,456 N= 54,456 4,852 (9.0) 4,852 2,347 (9.6) (9.6) 2,347 (2.9) 1,577 (3.9) 2,111 3804 (15.5) (15.5) 3804 2,666 (10.9) 2,666 (10.9) 6,557 (26.8) 9,179 (17.0) 8,215 (15.2) Rivaroxaban Rivaroxaban 10,386 (19.3) (19.3) 10,386 15,797 (29.3) 15,797 BMJ Open

40 (0.4) 90 (1.0) 90 (1.0) (0.9) 479 (0.9) 596 1,165 (0.9) 2,142 (3.3) 50 (0.6) 50 (0.6) (0.5) 282 (0.6) 367 (0.5) 699 1,174 (1.8) 357 (3.9) 357 (3.9) (3.4) 1,844 NA 2,201 (3.5) NA 649 (7.1) 649 (7.1) 106 (1.2) (7.6) 4,146 (1.2) 665 6,122 (9.5) 1,096 (1.7) (8.5) 10,917 (1.5) 1,867 (25.3) 16,461 836 (9.3) 287 (3.2) 287 (3.2) 375 (4.2) N= 9,085 9,085 N= 637 (16.8) (16.8) 637 (10.0) 378 (15.9) 605 Dabigatran Dabigatran 1,229 (32.4) 1,229 (32.4) 1,728 (19.2) 1,490 (16.6) 1,336 (14.8) 1,457 (16.0) (16.0) 1,457 (15.8) 8,614 (14.8) 9,542 (15.3) 19,613 NA 2,639 (29.3) (29.3) 2,639

§ emorrhage† a For peer review only For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

‡ up - follow Characteristics (N; %) Characteristics %) (N;

during during Aspirin Aspirin One reimbursement only reimbursement only One reimbursements ≤5 anticoagulants Parenteral Antiplatelet agents ecently ecently hospitalized for coagulopathy, purpura and other Reduced-dose Reduced-dose DOAC not approved for stroke Europe patients in prevention in AF Any valvular heart disease disease heartAny valvular cancer* treated or currently Recently 417 (4.6) (4.6) 2,531 2,898 (4.5) 5,846 (4.6) 4,252 (6.5) Recent gastrointestinal ulceration or intracranial h ulceration gastrointestinal Recent SSNRIs SSRIs and No monitoring of renal function at renal function initiation of monitoring No the year in post-initiation renal function of monitoring No N patterns, (%) Non-persistence anticoagulants or parenteral Antiplatelet agents NSAIDs haemorrhagic conditions* conditions* haemorrhagic Liver fibrosis and cirrhosis* cirrhosis* Liver and fibrosis R

use Inappropriate Contraindications or non-approved indication/dose indication/dose non-approved or Contraindications bleeding of therisk drug increasing Concomitant of use valve Prosthetic heart

‡ Data on paents with at least a one year of follow-up i.e. patients initiating OAC in 2015 after excluding patients who died and those hospitalized for 3 months or longer (N= 3,796; 24,483; 25,118; 53,397 and and 53,397 25,118; 3,796;24,483; (N= longer or months 3 for hospitalized those and died who patients excluding after 2015 in OAC initiating i.e. patients follow-up of year one a least at with paents on Data ‡ considered period users,respectively); new VKA total and DOAC- total apixaban, rivaroxaban, dabigatran-, 31,777 for included) included) users ,new VKA total and DOAC- total apixaban, rivaroxaban, dabigatran-, for 63,180 and 126,464 63,578; 53,885; 9,001; (N= date theindex following period a after 45-day alive still newusers for Data § included) date (index therapy OAC of the initiation following period 6-week rolling considered: period respectively); Table 2. Potential inappropriate use of DOAC therapy in oral anticoagulant-naive patients with atrial fibrillation in 2015-2016 in fibrillation atrial patients with anticoagulant-naive oral therapyDOACin of use inappropriate Potential2. Table selective SSNRIs: and inhibitors reuptake serotonin selective SSRIs: drugs; anti-inflammatory NSAIDs:non-steroidal applicable; not NA: fibrillation; atrial AF: antagonist; K vitamin VKA: anticoagulant; oral direct DOAC: inhibitors reuptake serotonin-norepinephrine therapy OAC of initiation the preceding period 1-year rolling a during procedures specific and/or data, LTD and/or hospitalization using identified Comorbidities * therapy OAC of initiation the preceding period 6-week duringrolling a data hospitalizaon using idenﬁed Comorbidies †

Page 23 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 24 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Figure 1 32 33 190x142mm (300 x 300 DPI) 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 32 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Figure 2 46 47 254x338mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Figure 3 32 33 190x142mm (300 x 300 DPI) 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 1 BMJ Open

For peer review only For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Supplementary Materials Materials Supplementary databases. healthcare French inthe comedications and conditions identify comorbid to used Definitions 1. Table Supplementary Supplementary Table 2. Comparison of selected baseline characteristics of DOAC new users with AF at risk of stroke according to level of HAS-BLED score dose. DOAC of type and

Page 27 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

alpha alpha

blockers blockers Page 28 of 32 - 2 naltrexone

blockers, agents acting on the the on acting agents blockers,

and products with with products and cardioversion nalmefene nalmefene surgery or procedures procedures or surgery

Specific procedures or drug reimbursements reimbursements drug or procedures Specific orthopaedic Lower limb venous ultrasonography, pulmonary/lower limb limb pulmonary/lower ultrasonography, venous limb Lower Nicotine replacement therapy therapy replacement Nicotine Heart valve surgery surgery valve Heart ablation, Radiofrequency limb Lower beta including failure heart for approved medications Specific channel calcium blockers, beta Diuretics, drugs lowering glucose blood Insulins and [peg]interferon (ribavirin, HCV use systemic agents for Antiviral NnRTIs). and alpha ([peg]interferon HCB against and DAA), and acamprosate, disulfiram, agents Nitrovasodilator stay nursing home of level high bed, wheelchair, hospital Home and anticholinergic antagonists: receptor NMDA or Anticholinesterases (bisoprolol, carvedilol, metoprolol), eplerenone, bumetanide, bumetanide, eplerenone, metoprolol), carvedilol, (bisoprolol, only failure heart for licensed when furosemide agents antiadrenergic other and system renin-angiotensin dependence alcohol with patients adult in licensed NSAIDs heparins, agents dopaminergic stabilizers mood conventional antipsychotics, Antidepressants, angiography, ventilation/perfusion scan ventilation/perfusion angiography,

†

I25 -

, I20 ,

LTDdiagnoses G83 E14 F03, G30, G20 G30, F03, - I73 I25 - - - -

J43-J44 J43-J44 I26, I80-I82 I26,I80-I82

), acid acetylsalicylic low-dose (including inhibitors aggregation Platelet I48 I50 E10 I70 I12 N18, C22 K74,B18, K70,K73, I20 G81 F00 I63, I74, G45 G45 I63,I74, F30-39 F20-29,

† K74 BMJ Open , K72 ,

Hospital discharge diagnoses discharge Hospital I13.1, I13.2, E10.2, E11.2, E13.2, E14.2, E14.2, E13.2, E11.2, E10.2, I13.2, I13.1, D50, D51, D52, D53, D55, D56, D57, D58, D57,D58, D56, D55, D53, D51,D52, D50,

G30, G20 G30, :

G83 D69 E14 F03, F03, - - I25 - - -

J43-J44 Z72.0; Z71.6, F17, I48 I50 to related J81 K76.1, I13.9, I13.2, I13.0, I11.0, or I50 E10 E10.5, I73, I71,I72, I70, I25, I24, I23, I22, I20,I21, I12.0, N18, K73, K72.1, K71.7, K71.5, K71.4, K71.3, K70, I85, R18, C78.7 C22, B18, K74,K76.1, Anemia K29.2, I42.6, G72.1, G62.1, G31.2, K6, T51, K70, F10, I20 G81 F00 D65 K71.7 K70.4, K70.3, K70.2, K70.0, I63 (except I63.6), G46 related to I63 or I69.3; I74, G45 G45 I69.3;I74, or I63 to related G46 I63.6), (except I63 G45.4) (except E14.5 E13.5, E12.5, E11.5, N08.3 D62, K92, K62.5, K25-K29, I85, S06.3-S06.6, I60-I62, H92.2, H45.0, H43.1, H35.6, H11.3, R58, R31, N02, N95.0,R04.0 N93.9, N93.8, N92, M25.0, K66.1, J94.2, defects, coagulation or D64 D63, D60,D62, D59, D67,D68, D66, D65, hemorrhagic: other and purpura D69 Z50.2. Z72.1, Z71.4, F30-39 F20-29, I26, I80 (except I80.0), I81, I82 I82 I81, I80.0), (except I26,I80 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

For peer review only other d

disease disease

BLED scores BLED procedures

- Covariates* Covariates*

Parkinson’s

orthopaedic VASc and HAS and VASc

- 2

DS 2 Drug-drug interaction interaction Drug-drug Smoking Deep vein thrombosis/pulmonary embolism embolism thrombosis/pulmonary vein Deep AF definition (Patterns of use of DOAC therapy in new in therapy DOAC of use of (Patterns definition AF fibrillation atrial Nonvalvular limb Lower CHA failure Heart drugs Antihypertensive Diabetes disease vascular Peripheral function renal Abnormal function liver Abnormal bleeding for Predisposition abuse Alcohol comorbidities Other infarction) myocardial (including disease heart Ischemic Frailty or Dementia an purpura coagulopathy, for hospitalized Recently cirrhosis and fibrosis Liver AF) with users valve heart disease/Prosthetic heart valvular of History Z95.4 ou Z95.3 Z95.2, I33-I39/ I05-I09, embolism systemic arterial stroke, (ischemic ATE of History attack) ischemic transient or bleeding Major disorders Psychiatric conditions haemorrhagic Supplementary Table 1. Definitions used to identify comorbid conditions and comedications in the SNIIRAM-PMSI databases. databases. SNIIRAM-PMSI the in comedications and conditions comorbid to identify used 1. Definitions Table Supplementary 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

dose acetylsalicylic acid acetylsalicylic dose - low

receptor antagonists antagonists receptor -

corticoids - and gluco and

- procedures Radiotherapy glycosides digitalis III, verapamil, and I class Antiarrhythmics: ezetimibe fibrates, inhibitors, reductase CoA HMG including inhibitors aggregation Platelet Mineralo H2 and inhibitors pump proton Mainly

D48 - D09, D37 D09, - C00 BMJ Open

, I61, I62, S06.4, S06.5, S06.5, S06.4, I61, ,I62, I60 D48, Z510, Z511 Z510, D48, - D09,D37 - K25, K26, K27, K28, K29.0, K29.0, K28, K27, K25,K26, C00 S06.6 S06.6 For peer review only For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

or intracranial intracranial or

and cardiac glycosides cardiac and

orticosteroids lowering agents lowering - during the 4-month period preceding the index date. Influenza vaccination was determined during the first ‘flu vaccination campaign preceding the index date. date. index the preceding campaign vaccination ‘flu the first during determined was vaccination Influenza date. index the preceding period 4-month the during codes †ICD-10 Recent gastrointestinal ulceration gastrointestinal Recent cancer treated currently or Recently comedications Baseline Antiarrhythmics Lipid drugs Antiplatelet c Oral agents Antiulcer haemorrhage haemorrhage LTD: long-term diseases; ATE: arterial thromboembolic events (mainly stroke); DVT/PE: deep vein embolism; thrombosis/pulmonary NMDA: N-methyl-D-aspartate; HIV: human immunodeficiency virus; PI: Protease * Comorbidities were identified using ICD-10 diagnosis codes for hospital discharge/LTD, or specific procedures, or drug Concomitant reimbursements. medications were identified as those dispensed at least once inhibitor; NNRTIs: Non-nucleoside reverse transcriptase inhibitors; NnRTI: Nucleoside and nucleotide reverse transcriptase inhibitors; DAA: direct-acting antiviral; HMG CoA: 3-hydroxy-3-methyl-glutaryl-CoA; NSAIDs: CoA: HMG 3-hydroxy-3-methyl-glutaryl-CoA; antiviral; direct-acting DAA: inhibitors; transcriptase reverse and nucleotide Nucleoside NnRTI: inhibitors; transcriptase reverse Non-nucleoside NNRTIs: inhibitor; drugs. anti-inflammatory Non-steroidal

Page 29 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 30 of 32

4 70.9 70.9 (N=981) (N=981) Standard doses* doses* Standard

HAS-BLED<3 (N=2,360) HAS-BLED<3 (N=2,360) 81.3 81.3 19 (1.4) 19 (1.4) 11 (0.8) (1.4) 14 (0.7) 7 104 (7.5) (7.5) 104 (6.8) 67 (N=1,379) (N=1,379) 322 (23.4) (23.4) 322 (42.7) 589 (18.8) 259 (16.0) 157 (10.4) 144 (32.2) 316 (11.3) 111 (24.4) 336 (9.7) 95 (50.5) 697 (18.9) 185 (62.3) 611 1013 (73.5) (73.5) 1013 (81.3) 798 1,188 (86.1) 1,188 (86.1) (57.8) 567 Reduced doses* Reduced doses*

81.1 81.1 (N=2,965) (N=2,965) Standard doses* Standard Patients aged ≥75 years old and history of ATE (N=9,503) ATE (N=9,503) history of and old years aged ≥75 Patients

HAS-BLED≥3 (N=7,143) HAS-BLED≥3 (N=7,143) 84,8 84,8 (N=4,178) (N=4,178) Reduced doses* Reduced

70.9 70.9 (N=52,273) Standard doses* Standard

81.3 81.3 HAS-BLED<3 (N=86 118) 118) HAS-BLED<3 (N=86 BMJ Open 425 (1.3) 425 (1.3) 620 (1.8) (0.9) 474 1,264 (2.4) (2.2) 94 (20.9) 874 52 (1.8) (17.2) 510 3169 (9.4) (9.4) 3169 (3.0) 1555 (18.9) 789 (12.0) 355 2,740 (8.1) 2,740 (8.1) 3,342 (6.4) (13.5) 566 2,150 (6.4) (11.1) 330 2,798 (5.4) (18.9) 788 (16.0) 473 (N=33,845) (N=33,845) 7836 (23.2) (23.2) 7836 (17.8) 9302 (20.6) 860 (17.0) 503 6,932 (20.5) 6,932 (20.5) 4,779 (9.1) (62.2) 2,599 6,827 (20.2) 1,642 (55.4) 9,142 (17.5) (32.0) 1,338 (25.6) 758 12836 (37.9) 12836 (41.7) 21820 (62.4) 2606 (71.9) 2132 22,962 (67.8) 22,962 (67.8) 9,984 (19.1) 25,221 (74.5) (83.7) 3,496 41,728 (79.8) 1,720 (58.0) (77.0) 3216 (80.2) 2965 Reduced doses*

74.8 74.8 (N=14,183) (N=14,183) Standard doses* Standard Patients with CHA2DS2-VASc ≥ 2 (N=116,391) (N=116,391) 2 CHA2DS2-VASc with ≥ Patients

81.8 81.8 HAS-BLED≥3 (N=30 273) 273) HAS-BLED≥3 (N=30 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 544 (3.4) 544 (3.4) 384 (2.7) (N=16,090) (N=16,090) 9252 (57.5) (57.5) 9252 (65.0) 9213 3321 (20.6) (20.6) 3321 (17.1) 2426 2121 (13.2) (13.2) 2121 935 (6.6) 7,232 (44.9) 7,232 (44.9) 5,274 (37.2) 6,689 (41.6) 6,689 (41.6) 4,600 (32.4) 2,640 (16.4) 1,809 (12.8) 6,035 (37.5) 4,891 (34.5) 3,627 (22.5) 2,946 (20.8) 12,315 (76.6) (76.6) 12,315 11,364 (80.1) 10,786 (67.0) (67.0) 10,786 3,982 (28.1)

Reduced doses* Reduced doses* For peer review only

N (%) N (%)

cardiologist Hospital practitioner or private or practitioner Hospital

First prescriber’s specialty specialty First prescriber’s Ischaemic heart disease Ischaemic heart disease agents Lipid-lowering practitioner General Age (years), mean Age (years), Frailty (proxies) (proxies) Frailty disease Dementia Parkinson's or Any diseases† valvular valve† Prosthetic heart Antiplatelet initiation at agents ATC (≥10 classes) Polymedication Age ≥ 80 years Age years 80 ≥

* Reduced-dose DOAC: dabigatran 150mg, Rivaroxaban 20 and apixaban 5mg; reduced-dose DOAC: dabigatran 75mg and 110mg, rivaroxaban 10mg and 15mg and apixaban 2.5mg apixaban 15mg 10mg and and rivaroxaban 110mg, DOAC: 75mgdabigatran and 5mg; reduced-dose apixaban 20 and 150mg, DOAC: Rivaroxaban dabigatran Reduced-dose * data hospitalization by only Covariates† defined and type of DOAC dose dose DOAC oftype and Supplementary Table 2. Comparison of selected baseline characteristics of DOAC new users with AF at risk of stroke according to level of HAS-BLED score of with at risk DOAC AF to users score according new level of stroke of of baseline characteristics HAS-BLED selected Table 2. Comparison Supplementary 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 31 of 32 BMJ Open

1 2 STROBE Statement—checklist of items that should be included in reports of observational studies 3 4 Item Page 5 No Recommendation No 6 Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the title or the 1 7 8 abstract 9 (b) Provide in the abstract an informative and balanced summary of what 2 10 was done and what was found 11 Introduction 12 13 Background/rationale 2 Explain the scientific background and rationale for the investigation being 4 14 reported 15 Objectives 3 State specific objectives, including any prespecified hypotheses 4 16 Methods For peer review only 17 18 Study design 4 Present key elements of study design early in the paper 5,6 19 Setting 5 Describe the setting, locations, and relevant dates, including periods of 5 20 recruitment, exposure, follow-up, and data collection 21 Participants 6 (a) Cohort study—Give the eligibility criteria, and the sources and methods 5,6 22 of selection of participants. Describe methods of follow-up 23 24 Case-control study—Give the eligibility criteria, and the sources and 25 methods of case ascertainment and control selection. Give the rationale for 26 the choice of cases and controls 27 Cross-sectional study—Give the eligibility criteria, and the sources and 28 methods of selection of participants 29 30 (b) Cohort study—For matched studies, give matching criteria and number NA 31 of exposed and unexposed 32 Case-control study—For matched studies, give matching criteria and the 33 number of controls per case 34 Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, 6 35 and effect modifiers. Give diagnostic criteria, if applicable 36 37 Data sources/ 8* For each variable of interest, give sources of data and details of methods of 6,7,8 38 measurement assessment (measurement). Describe comparability of assessment methods 39 if there is more than one group 40 Bias 9 Describe any efforts to address potential sources of bias 7,8 41 Study size 10 Explain how the study size was arrived at 9, Fig 42 43 2 44 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If 7 45 applicable, describe which groupings were chosen and why 46 Statistical methods 12 (a) Describe all statistical methods, including those used to control for 7,8 47 confounding 48 49 (b) Describe any methods used to examine subgroups and interactions 7,8 50 (c) Explain how missing data were addressed NA 51 (d) Cohort study—If applicable, explain how loss to follow-up was NA 52 addressed 53 Case-control study—If applicable, explain how matching of cases and 54 55 controls was addressed 56 Cross-sectional study—If applicable, describe analytical methods taking 57 account of sampling strategy 58 (e) Describe any sensitivity analyses 7,8 59 1 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 32

1 2 Continued on next page 3 Results 4 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers potentially 9 5 eligible, examined for eligibility, confirmed eligible, included in the study, Fig 2 6 7 completing follow-up, and analysed 8 (b) Give reasons for non-participation at each stage 9 9 (c) Consider use of a flow diagram Fig 2 10 Descriptive 14* (a) Give characteristics of study participants (eg demographic, clinical, social) and Table 11 12 data information on exposures and potential confounders 1 13 (b) Indicate number of participants with missing data for each variable of interest NA 14 (c) Cohort study—Summarise follow-up time (eg, average and total amount) NA 15 Outcome data 15* Cohort study—Report numbers of outcome events or summary measures over time NA 16 ForCase-control peer study—Report review numbers in each exposure only category, or summary NA 17 18 measures of exposure 19 Cross-sectional study—Report numbers of outcome events or summary measures Fig2,3 20 Table2 21 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and Table 22 their precision (eg, 95% confidence interval). Make clear which confounders were 2 23 24 adjusted for and why they were included 25 (b) Report category boundaries when continuous variables were categorized Table 26 2 27 (c) If relevant, consider translating estimates of relative risk into absolute risk for a NA 28 meaningful time period 29 30 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and 9, 10 31 sensitivity analyses 32 Discussion 33 Key results 18 Summarise key results with reference to study objectives 11 34 35 Limitations 19 Discuss limitations of the study, taking into account sources of potential bias or 13 36 imprecision. Discuss both direction and magnitude of any potential bias 37 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, 11, 12 38 multiplicity of analyses, results from similar studies, and other relevant evidence 39 Generalisability 21 Discuss the generalisability (external validity) of the study results 12 40 41 Other information 42 Funding 22 Give the source of funding and the role of the funders for the present study and, if 14 43 applicable, for the original study on which the present article is based 44 45 *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and 46 47 unexposed groups in cohort and cross-sectional studies. 48 49 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and 50 published examples of transparent reporting. The STROBE checklist is best used in conjunction with this article (freely 51 available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 52 53 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is 54 available at www.strobe-statement.org. 55 56 57 58 59 2 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open

Oral anticoagulation therapy use in patients with atrial fibrillation after the introduction of non-vitamin K antagonist oral anticoagulants: findings from the French healthcare databases, 2011-2016

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2018-026645.R1

Article Type: Research

Date Submitted by the 03-Jan-2019 Author:

Primary Subject Cardiovascular medicine Heading:

Cardiovascular medicine, Epidemiology, Pharmacology and therapeutics, Secondary Subject Heading: Public health

Anticoagulation < HAEMATOLOGY, EPIDEMIOLOGY, PUBLIC HEALTH, Keywords: dabigatran, rivaroxaban, apixaban

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 34 BMJ Open

1 2 3 4 Title page 5 6 7 Oral anticoagulation therapy use in patients with atrial fibrillation after the introduction of non- 8 9 vitamin K antagonist oral anticoagulants: findings from the French healthcare databases, 2011- 10 11 2016. 12 13 14 Running title: Patterns of OAC use 15 16 17 Géric Maura PharmD, PhD1,2, Cécile Billionnet MSc, PhD1, Jérôme Drouin, MSc1, Alain Weill MD1, 18 For peer review only 19 1 2,3 20 Anke Neumann MSc, PhD , Antoine Pariente MD, PhD 21 22 1 Department of Studies in Public Health, French National Health Insurance (Caisse Nationale de 23 24 l’Assurance Maladie/Cnam), 75 986 Paris Cedex 20, France. 25 26 2 University of Bordeaux, Inserm, Bordeaux Population Health Research Center, team 27 28 29 Pharmacoepidemiology - UMR 1219, F-33000 Bordeaux, France 30 31 3 CHU de Bordeaux, Pharmacologie, F-33000 Bordeaux, France 32 33 Address for correspondence: Géric MAURA, Cnam/DSES/DESP, National Health Insurance, 50 avenue 34 35 du Pr. André Lemierre, 75986 Paris cedex 20; E-mail: [email protected]; Telephone: 36 37 38 +33(0)172602338; Fax: +33(0)172601724 39 40 41 42 43 Word count of text: (excluding the title page, abstract, tables, acknowledgements, contributions and 44 45 references): 4,186 46 Word count of abstract: 300/300 47 48 Number of references: 64 49 50 Number of tables: 2 51 52 Number of figures: 3 53 Number of supplementary tables: 3 54 55 56 57 58 59 60

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 34

1 2 3 Abstract 4 5 6 Objectives: To describe i) the trend in oral anticoagulant (OAC) use following the introduction of 7 8 Non-vitamin K antagonist oral anticoagulant (NOAC) therapy for stroke prevention in atrial fibrillation 9 10 (AF) patients; ii) the current patterns of use of NOAC therapy in new users with AF in France. 11 Design: i) Repeated cross-sectional study; ii) Population-based cohort study. 12 13 Setting: French National Healthcare databases (50 million beneficiaries). 14 15 Participants: i) Patients with identified AF in 2011, 2013, and 2016; ii) Patients with AF initiating OAC 16 therapy in 2015-2016. 17 18 Primary and secondaryFor outcome peer measures review: i) Trend in oral anticoagulantonly therapy use in patients 19 20 with AF; ii) Patterns of use of NOAC therapy in new users with AF. 21 Results: Between 2011 and 2016, use of OAC therapy moderately increased (+16%), while use of 22 23 antiplatelet therapy decreased (-22%) among all patients with identified AF. In 2016, among the 1.1 24 25 million AF patients, 66% used OAC therapy and were more likely to be treated by VKA than NOAC 26 therapy, including patients at higher risk of stroke (63.5%), while 33% used antiplatelet therapy. 27 28 Among 192,851 new users of OAC therapy in 2015-2016 with identified AF, NOAC therapy (66.3%) 29 30 was initiated more frequently than VKA therapy, including in patients at higher risk of stroke (57.8%). 31 Reduced doses were prescribed in 40% of NOAC new users. Several situations of inappropriate use at 32 33 NOAC initiation were identified, including concomitant use of drugs increasing the risk of bleeding 34 35 (one in three new users) and potential NOAC underdosing. 36 Conclusions: OAC therapy use in AF patients remains suboptimal 4 years after the introduction of 37 38 NOACs for stroke prevention in France and improvement in appropriate prescribing regarding NOAC 39 40 initiation is needed. However, NOAC therapy is now the preferred drug class for initiation of OAC 41 therapy in patients with AF, including in patients at higher risk of stroke. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 34 BMJ Open

1 2 3 4 Strengths and limitations of this study 5 6  This study is the first to report both the 2011-2016 trend in oral anticoagulant (OAC) 7 coverage in patients with atrial fibrillation (AF) following the introduction of Non-vitamin K 8 9 antagonist oral anticoagulant (NOAC) therapy in France and the current patterns of use of 10 11 NOACs in new users including assessment of potential inappropriate use; 12  This study is based on reimbursement data for 50 million beneficiaries with access to all OAC 13 14 prescriptions filled in the ambulatory setting; 15 16  As indications for treatment are not available in the databases, AF was mostly identified on 17 the basis of discharge and long-term disease diagnosis codes and an algorithm previously 18 For peer review only 19 validated in the French healthcare databases that helped to further identify AF in 20 21 outpatients. 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 34

1 2 3 Introduction 4 5 6 Non-vitamin K antagonist Oral AntiCoagulants (NOACs) have been gradually introduced over the past 7 8 decade as a more convenient, fixed-dose alternative to vitamin K antagonists (VKAs), the only oral 9 10 anticoagulant therapy available up until now for long-term stroke prevention in patients with non- 11 valvular atrial fibrillation (AF) [1]. Compared to VKA, NOAC therapy avoids the need for regular 12 13 laboratory monitoring of patients by INR testing due to a wider therapeutic window, allows once 14 15 (rivaroxaban, edoxaban) or twice (dabigatran, apixaban) daily intake and is associated with fewer 16 drug-drug interactions to date [2–6]. NOACs have been demonstrated to have similar or superior 17 18 efficacy to warfarin forFor stroke peerprevention inreview nv-AF patients, andonly these findings have recently been 19 20 implemented in the 2016 European Society of Cardiology (ESC) guidelines that recommended NOAC 21 over VKA therapy in this indication [7]. Moreover, antiplatelet therapy is no longer recommended in 22 23 these patients [7,8]. The use of NOAC therapy, massively adopted worldwide including in France [9– 24 25 15], is expected to overcome the suboptimal use of oral anticoagulant (OAC) therapy extensively 26 reported with VKA therapy, including underprescribing and high discontinuation rates [16,17]. 27 28 Despite their improved ease of use, NOAC prescribed dose needs to be adjusted to the patient’s 29 30 clinical profile with regards to age, renal function, weight, and risks of bleeding and drug-drug 31 interactions. Two dose regimens are therefore proposed for each NOAC: standard dose regimen (i.e. 32 33 dabigatran 150mg/12h, rivaroxaban 20mg/24h; apixaban 5mg/12h), and reduced-dose regimen (i.e. 34 35 dabigatran 75mg or 110mg/12h, rivaroxaban 10mg or 15mg/24h; apixaban 2.5mg/12h). The reasons 36 37 for prescribing a reduced-dose regimen are listed in the summary of product characteristics (SmPCs) 38 of each NOAC with differences across NOACs as well as in the ESC [7,8] and European Heart Rhythm 39 40 Association (EHRA) guidelines [18]. Recent publications have suggested that the frequency of use of 41 42 reduced-dose NOACs in clinical practice could largely exceed that expected on the basis of the 43 conditions summarized in these guidelines [11,19–21]. However, national data are lacking concerning 44 45 the current French patterns of NOAC use, including the potential issue of NOAC underdosing. 46 47 A steady increase in the initiation of NOAC therapy has already been reported in AF patients in 48 France [14], but trends in OAC coverage of AF patients following the introduction of NOACs and a 49 50 description of the current national patterns of NOAC use have not yet been reported. This study, 51 52 based on the French nationwide healthcare databases, therefore had a twofold objective: 1) to 53 describe the trends in OAC use following the introduction of NOAC for stroke prevention in AF 54 55 patients during the 2011-2016 period; 2) to describe the current patterns of use of OAC therapies 56 57 with particular focus on NOAC use in new users with AF. 58 59 60

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 34 BMJ Open

1 2 3 4 Methods 5 6 Data source 7 8 French national health insurance (Assurance Maladie) covers the entire French population by means 9 of several specific schemes according to the beneficiary’s occupational sector, the largest scheme 10 11 being the ‘Régime général’ (around 50 million beneficiaries). 12 13 This study was conducted using data from the French health insurance system database (Système 14 national d'information inter-régimes de l'Assurance maladie, SNIIRAM) linked to the French hospital 15 16 discharge database (Programme de médicalisation des systèmes d'information, PMSI) [22,23]. The 17 18 SNIIRAM database containsFor individualized, peer anonymous,review and comprehensive only data on health spending 19 20 reimbursements. Demographic data include date of birth, gender, and vital status. Drugs are coded 21 according to the Anatomical Therapeutic Chemical (ATC) classification. Each packaging of each 22 23 product is identified by means of a national specific pack identifier code providing information on the 24 25 name of the product, active ingredient and dose per unit, number of units, and route of 26 administration; but, the exact dosage prescribed and the indication are not available. However, for 27 28 most drugs, particularly OACs, each dispensing of the prescribed drug cannot exceed the quantity 29 30 necessary for one month of treatment. The PMSI database provides detailed information about 31 discharge diagnoses and medical procedures related to all hospitalizations in France. This information 32 33 regarding morbidities is completed by diagnoses corresponding to patient eligibility for 100% 34 35 reimbursement of severe and costly long-term diseases (LTD) and disability, such as AF, coronary 36 heart disease, certain debilitating diseases (such as multiple sclerosis or rheumatoid arthritis), HIV 37 38 infection, cancer, etc. All information concerning medical diagnosis is encoded according to the 39 40 International Classification of Diseases, 10th revision (ICD-10). Finally, the SNIIRAM-PMSI databases 41 also indicate medical procedures performed in the ambulatory setting, including information about 42 43 the type and date of all laboratory tests performed, but not including their results. 44 45 The French healthcare databases have been previously described and used in epidemiological and 46 pharmacoepidemiological studies [22,24–26]. 47 48 49 Study populations and study designs 50 Two study populations were defined; one for each objective. 51 52 To answer the first objective, a repeated cross-sectional study was performed to describe the trends 53 54 in OAC use following the introduction of NOAC in AF patients. Patients with AF were identified in 55 2011 (as none of the NOACs was available for stroke prevention in France) and 2016 (the most recent 56 57 data available at the time of writing the study protocol; dabigatran, rivaroxaban and apixaban were 58 59 all reimbursed for stroke prevention in France, as apixaban was reimbursed from January 2014 60

5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 34

1 2 3 onwards). OAC coverage was also calculated for year 2013 as this year represented the first calendar 4 5 year for which the first two NOACs were available in France, i.e. a pivotal year for the 6 7 pharmacological management of AF by oral anticoagulants. For each of these calendar years, a 8 patient was considered to have AF when at least one diagnosis of AF (ICD-10 code I48) was identified 9 10 from discharge and LTD diagnoses in the SNIIRAM-PMSI database in the calendar year considered or 11 12 during the previous 5 years. Patients with no continuous ‘Régime général’ health insurance coverage 13 for at least six years before the calendar year considered were excluded. 14 15 To answer the second objective, a population-based cohort study was performed including patients 16 17 with AF initiating OAC therapy in 2015-2016. First, OAC new users were identified among patients 18 with continuous ‘RégimeFor général peer’ health review insurance coverage only as those with at least one 19 20 reimbursement for OAC therapy in 2015-2016 and no reimbursement for any OAC (VKA or NOAC) in 21 22 the previous 24 months. The patient’s index date was the date of first OAC reimbursement identified 23 during the 2015-2016 period. Second, the cohort of NOAC news users was restricted to those treated 24 25 for AF: (i) patients treated for other OAC indications i.e. patients treated for deep vein 26 27 thrombosis/pulmonary embolism (DVT/PE) or with lower limb orthopaedic procedures were 28 excluded; (ii) OAC new users treated for AF were identified from the resulting cohort as the sum of 29 30 “OAC new users with confirmed AF” for those with a diagnosis of AF (ICD-10 code I48) or specific AF 31 32 management procedures identified from LTD or hospitalization discharge data during a six-year pre- 33 index period, and “OAC new users with probable AF” for outpatients identified using an algorithm 34 35 discriminating AF from DVT/PE with 95% specificity [27]. The remaining patients were not classified 36 37 as probable FA patients and were excluded. Codes used for identification of AF and all of the patient 38 characteristics considered, including comorbidities, are displayed in Supplementary Table 1. 39 40 41 Patient and public involvement 42 Patients and or public were not involved. 43 44 45 Exposure 46 47 NOAC (dabigatran, rivaroxaban, apixaban) and VKA therapies (fluindione, warfarin and 48 acenocoumarol) were identified using ATC codes; edoxaban was not available in France during the 49 50 period considered. 51 52 53 Outcomes 54 55 Trends in oral anticoagulant therapy use in patients with AF 56 The proportion of AF patients treated by OAC therapy was assessed before and after approval of 57 58 NOAC therapies for stroke prevention in France. Trends in the use of antiplatelet agents were also 59 assessed in AF patients over the same timeframe. 60

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 34 BMJ Open

1 2 3 Patterns of use of NOAC therapy in new users with AF 4 5 The description of patterns of NOAC use in new users treated for AF in 2015-2016 included 6 comparison of the baseline characteristics among NOAC new users and compared to those of VKA 7 8 new users and potential inappropriate use of NOAC therapy was then investigated by identifying: 9 10 (i) NOAC off-label use or non-approved indication/dose: contraindications to NOAC therapy 11 according to SmPCs [concomitant coagulopathy, purpura and other haemorrhagic conditions, liver 12 13 fibrosis and cirrhosis, recent gastrointestinal ulceration or intracranial haemorrhage], valvular atrial 14 15 fibrillation [NOAC are only approved for non-valvular AF], prosthetic heart valve [contraindicated for 16 dabigatran], cancer [NOAC are not approved for prevention of thromboembolism in patients with 17 18 cancer] and prescriptionFor of NOAC peer doses not reviewapproved for stroke only prevention in Europe [dabigatran 75 19 20 mg and rivaroxaban 10 mg are not approved for stroke prevention in Europe and are therefore off- 21 label doses in AF patients]; (ii) non-compliance with guidelines with respect to follow-up and clinical 22 23 work-up of patients during the first year following NOAC initiation: no monitoring of patients’ renal 24 25 function [renal function should be assessed at initiation and annually during NOAC therapy [28]], 26 discontinuation of NOAC therapy [OAC therapy is recommended as lifetime treatment in most 27 28 patients with AF] (iii) clinically relevant drug-drug interactions at initiation increasing the bleeding 29 30 risk at initiation; (iv) potential inappropriate underdosing, i.e. patients in whom the NOAC dose 31 prescribed at initiation was inappropriately reduced in view of their individual stroke and bleeding 32 33 risks. 34 35 36 Data analysis 37 Descriptive analyses are expressed as mean and standard deviation (SD) for continuous variables, 38 39 and numbers and percentages for categorical variables. 40 41 Trends in oral anticoagulant therapy use in patients with AF 42 43 For each calendar year, the proportion of patients treated by a drug was defined by the number of 44 45 patients with at least one reimbursement for this drug in the calendar year considered over the total 46 number of patients identified as having AF in the same year. Proportions are reported according to 47 48 the type of OAC first reimbursed in the year considered. Antiplatelet drugs and OAC therapies were 49 50 considered to be coprescribed when they were reimbursed at least once on the same day during the 51 calendar year studied. Analyses were replicated in subgroups of AF patients: (i) aged 75 years and 52 53 over; (ii) female and male, separately; (iii) with a history of hospitalization for arterial 54 55 thromboembolic events (ATE); (iv) with concomitant ischaemic heart disease (IHD) or prosthetic 56 heart valve. 57 58 59 60

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 34

1 2 3 Patterns of use of NOAC therapy in new users with AF 4 5 Baseline characteristics of NOAC new users with AF included sociodemographic data, including 6 deprivation index of the patient’s municipality of residence [29], type of initial prescriber, clinical 7 8 scores predicting the risk of stroke (CHA2DS2-VASc score) or bleeding (HAS-BLED score) [30,31], 9 10 adapted to claims data and the other main comorbidities and comedications, including proxies of 11 frailty. A negative binomial regression analysis for each NOAC therapy and each baseline 12 13 characteristic was performed to assess the association between these characteristics and the choice 14 15 of NOAC therapy versus VKA therapy, while adjusting for age and sex. 16 Compliance with guidelines regarding renal function monitoring and treatment persistence patterns 17 18 were assessed in newFor users forpeer whom data review for at least one only year of follow-up were available, i.e. 19 20 patients included in 2015 and who had not died and had not been hospitalized for 3 months or 21 longer. Compliance with renal function monitoring was assessed at NOAC initiation (no 22 23 reimbursement for renal function monitoring during the three months before and the three months 24 25 after NOAC initiation) and during the first year following treatment initiation. OAC non-persistence 26 patterns were assessed over the one-year period following the index date by calculating proxies of 27 28 OAC discontinuation: number of patients with only one reimbursement and one-year crude 29 30 discontinuation rates. 31 Drugs increasing the risk of bleeding were those responsible for clinically relevant pharmacodynamic 32 33 interactions with OAC therapy, i.e. concomitant reimbursement of other parenteral and oral 34 35 antithrombotic drugs, non-steroidal anti-inflammatory drugs (NSAIDs); selective serotonin reuptake 36 inhibitors and selective serotonin-norepinephrine reuptake inhibitors (SSRIs and SSNRIs) [7,18,32]. 37 38 Patients taking concomitant drugs with OAC therapy were defined as those with a reimbursement for 39 40 the drug of interest during the period corresponding to the index date and the following 45 days. 41 Analyses were replicated in VKA new users for descriptive purposes. 42 43 Finally, potential inappropriate underdosing with NOACs was defined as initiation of NOAC therapy in 44 45 patients at risk of stroke in whom reduced doses of NOAC were prescribed with no identified 46 47 justification. As this study was based on claims data and as, up until 2016, ESC guidelines 48 recommended prescribing reduced-dose NOAC in patients with HAS-BLED≥3 [8], the proportion of AF 49 50 patients initiating reduced-dose NOAC with an HAS-BLED score<3 among all NOAC new users with a 51 52 CHA2DS2-VASc score ≥2 was used to quantify potential inappropriate underdosing in NOAC new 53 users. Analyses were replicated in patients i) with CHA2DS2-VASc score ≥4, ii) aged 75 and over with a 54 55 history of ATE. 56 57 58 59 60

8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 34 BMJ Open

1 2 3 4 Results 5 6 Trends in oral anticoagulant therapy use in patients with AF 7 8 The number of patients identified from French healthcare databases as having AF increased between 9 2011 (N=853,440) and 2016 (N=1,098,657). A high proportion of these patients were identified 10 11 exclusively by hospitalization discharge diagnosis and this proportion decreased only slightly over this 12 13 time interval from 90.2% (N=770,002) to 86.2% (N=946,657). 14 Between 2011 and 2016, the proportion of patients with at least one reimbursement for OAC 15 16 therapy among all AF patients moderately increased (+16%) from 56.7% to 65.8%, corresponding to a 17 18 steady decrease in VKAFor use (from peer 56.6% to 40.8%review of all AF patients) only associated with the introduction 19 20 of NOACs (from 0.6% to 27.7%). In 2016, among patients with identified AF, VKA therapy remained 21 the preferred OAC therapy (62.0%), including in patients aged 75 years and over (67.0%) and in those 22 23 with a history of ATE (63.5%). 24 25 Between 2011 and 2016, the use of antiplatelet therapy decreased in AF patients (-22%), but the 26 proportion of patients with concomitant OAC and antiplatelet therapy remained stable (9.7% in 27 28 2016). In 2016, 32.5% of AF patients had at least one reimbursement for antiplatelet therapy during 29 30 the year. 31 Similar trends were observed in the subgroup analyses. A slightly higher rate of OAC therapy was 32 33 observed in patients with a history of ATE (68.4% vs 65.8% in the total cohort in 2016). However, OAC 34 35 coverage was lower in women than in men with AF (64.7% vs 66.8% in 2016) (Figure 1). 36 37 Patterns of use of NOAC therapy in new users with AF 38 39 Baseline characteristics of OAC new users 40 41 Among 540,914 patients initiating OAC therapy in 2015-2016, a total of 192,851 (35.7%) patients 42 were included in the study population, corresponding to 127,841 NOAC new users and 65,010 VKA 43 44 new users with AF. The mains reasons for ineligibility were other indications or uncertain 45 46 identification of the indication for NOAC (Figure 2). 47 The mean age of the NOAC new users cohort was 74.1±11.6 years; patients over the age of 80 48 49 represented 37.3% of the total cohort. One-half of the NOAC new users were women, and 40.0% 50 51 received a reduced dose at initiation (62.2% for dabigatran new users). Apixaban was the NOAC most 52 53 commonly initiated; apixaban new users were older and had more comorbidities than the other two 54 groups of NOAC new users. NOAC therapy was more likely to be initiated than VKA therapy among 55 56 patients aged 75 years and older (59.4%) and in patients with a history of ATE (57.8%) (Table 1). 57 58 Characteristics associated with bleeding risk, such as older age, renal impairment, history of bleeding 59 or bleeding predisposition, and treatment with a concomitant drug increasing the risk of bleeding at 60

9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 34

1 2 3 OAC initiation, were strong predictors of being treated with VKA therapy versus NOAC therapies 4 5 (Supplementary Table 2). 6 7 8 Potential inappropriate use of NOAC therapy 9 About 15% of NOAC new users with AF were considered to be using NOAC off-label or for a non- 10 11 approved indication. In particular, 8.5% of NOAC new users with AF had valvular heart disease (8.5%), 12 including prosthetic heart valve (1.5%) and 4.6% had a recently or currently treated cancer (Table 2). 13 14 About 15% and 9% of NOAC new users had no reimbursement for renal function tests at initiation 15 16 and during the one-year period post-initiation, respectively. Discontinuation during the one-year 17 period following initiation was frequent, as more than 20% of patients had five or less 18 For peer review only 19 reimbursements (Table 2). 20 21 Nearly 30% of NOAC new users were using at least one concomitant drug increasing the risk of 22 bleeding (52% in VKA new users). The most common concomitant drugs concerned at initiation were 23 24 antiplatelet agents or parenteral anticoagulants (Table 2). 25 26 Among the 116,391 NOAC new users with AF with a CHA2DS2-VASc score ≥2, 29.1% (N=33,845) were 27 prescribed a reduced dose although they had an HAS-BLED score <3. This meant that nearly 1 in 3 28 29 NOAC new users with AF and at risk of stroke were therefore potentially prescribed an 30 31 inappropriately reduced dose of NOAC at initiation. This proportion was 33% (N=24,281) and 14.5% 32 when defining patients at risk of stroke as patients with a CHA DS -VASc score ≥4 and aged 75 and 33 2 2 34 over with a history of ATE, respectively (Figure 3). 35 36 Differences in baseline characteristics were observed in patients with HAS-BLED<3 according to the 37 type of NOAC dose prescribed, e.g. patients with reduced-dose NOAC were older and frailer than 38 39 those with standard-dose NOAC (Supplementary Table 3). 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 34 BMJ Open

1 2 3 4 Discussion 5 6 Main findings 7 8 OAC therapy use among patients with AF improved in France between 2011 and 2016, but remained 9 suboptimal with about 66% of the 1.1 million patients with identified AF treated by OAC therapy in 10 11 2016. Use of antiplatelet therapy in AF patients decreased over the same period, but still concerned 12 13 33% of all AF patients in 2016. Patients with AF were more likely to be treated by VKA than NOAC 14 therapy, including older patients and those at higher risk of stroke. 15 16 Nearly 193,000 patients with AF were identified as OAC new users in 2015-2016: patients were more 17 18 likely to be treated byFor NOAC thanpeer VKA therapy, review including older only patients and those at higher risk of 19 20 stroke. Results on current patterns of use of NOAC therapy in new users with AF suggest several 21 situations of inappropriate use, including frequent concomitant use of drugs increasing the risk of 22 23 bleeding and potential inappropriate underdosing. 24 25 26 Comparison with postmarketing literature and clinical implications 27 The overall improvement of management of AF patients observed with regards to OAC therapy after 28 29 the introduction of NOACs has been reported in many countries [33–36]. The same applies to the 30 steady decrease in VKA use in favour of NOAC therapies [12,37,38], and the gaps remaining in 31 32 optimal OAC coverage and high antiplatelet drug use [39–41]. 33 34 This study demonstrated channelling of NOAC therapy towards patients at lower risk of stroke and 35 bleeding when considering all patients with AF, in line with what has become a common feature 36 37 reported worldwide in clinical practice by many observational studies on the current patterns of use 38 39 of NOACs [15,42–45]. In particular, data from EORP-AF registry showed that younger age and non- 40 valvular heart diseases were also found to be clinical predictors for being treated with NOAC in other 41 42 South countries (Greece, Italy, Portugal) [46]. 43 44 However, NOAC therapy is now the preferred OAC therapy at initiation in the oldest patients and 45 those at higher risk of stroke. This French pattern of OAC use has never been previously reported and 46 47 contrasts with published results concerning earlier periods [47]. This emerging pattern is encouraging 48 49 for AF management, as older and high-risk patients are those who should derive most benefit from 50 NOAC versus VKA therapy [48]. 51 52 Reduced doses were often prescribed in OAC new users, including dabigatran 75mg and rivaroxaban 53 54 10mg, which are not approved for stroke prevention in Europe [49]. In addition to the overall 55 channelling mentioned above, these findings may reflect a “bleeding avoidance” strategy of 56 57 prescribers (i.e. overestimation of the potential bleeding risk versus the likely benefit of stroke 58 59 reduction) and the differential perception of the comparative safety of NOACs versus VKA and 60

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 34

1 2 3 between NOACs. The early safety alert on bleeding in dabigatran-treated patients, followed by the 4 5 contraindication of this NOAC in patients with prosthetic heart valves, may have reinforced the fears 6 7 of prescribers in relation to the safety of dabigatran, which would explain the difference in reduced- 8 dose prescription rates between the three NOACs in this study, despite the intermediate stroke and 9 10 bleeding risk profile of dabigatran compared to that of rivaroxaban and apixaban in new users. 11 12 Similarly, among NOAC new users, apixaban was prescribed to the oldest and most severe patients. 13 Apixaban was the only NOAC found to be superior to warfarin for all types of bleeding outcome and 14 15 all-cause mortality, which would also illustrate the tendency of physicians to prescribe OAC therapies 16 17 according to bleeding risk [5]. This may also explain the potential inappropriate underdosing 18 observed in NOAC usersFor in this peer study. This reviewpattern of NOAC useonly has been previously reported, but 19 20 mostly in field and registry studies based on small sample sizes. The reported inappropriate 21 22 underdosing rate varies according to studies and the definition used. NOAC underdosing concerned 23 30.4% of Turkish patients in the RAMSES study (N=2,086) [50], 18.4% of Japanese patients of the KiCS 24 25 AF registry (N=1,284) [51,52], between 19.7% and 27.6% of patients in the SAKURA AF registry (N= 26 27 3,266) [53], and 9.4% to 16% of patients in the ORBIT-AF II registry (N=7925) [21,54]. In the subgroup 28 of Dutch patients (N=899) enrolled in the XANTUS registry, 33% of patients were also treated with 29 30 reduced-dose rivaroxaban despite presenting normal renal function [55]. Using a large U.S. 31 32 administrative database, Yao et al found that 13.3% of the 13,392 NOAC new users with no renal 33 indication for dose reduction were potentially underdosed [56]. Taken together with our results, 34 35 these data suggest that inappropriate underdosing might be a common issue in NOAC new users that 36 37 should be systematically assessed when studying NOAC patterns of use. This is of particular concern, 38 as recent data have suggested a relationship between NOAC dose and clinical outcomes [57]. In 39 40 particular, NOAC underdosing has been shown to be associated with increased risk for adverse 41 42 outcomes [21,56]. 43 These patterns of NOAC use contrast with the other patterns concomitantly observed in this study, 44 45 such as the high level of concomitant prescription of antiplatelet agents and parenteral 46 47 anticoagulants or, to a lesser extent, NOAC use in non-approved indications such as prosthetic heart 48 valves that are both associated with an increased risk of bleeding [58–60]. 49 50 51 Strengths and limitations 52 53 This study is the first to report the improved trend in OAC coverage in French patients with AF over 54 the last five years as well as the recent patterns of use of OAC therapy in new users, particularly 55 56 including a nationwide assessment of the growing issue of NOAC underdosing, based on health data 57 58 for more than 50 million beneficiaries. Moreover, all OAC prescriptions filled in the ambulatory 59 setting are captured in the databases and are reimbursed with no restriction of coverage: selection 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 34 BMJ Open

1 2 3 bias related to the access of patients to more expensive NOAC therapy is therefore not an issue with 4 5 the use of French healthcare databases [22,61]. 6 7 However, several limitations related to the nature of the data used should be underlined. First of all, 8 it cannot be verified whether patients actually took the drugs for which they were reimbursed. 9 10 Secondly, as the indication for treatment is not available in the databases, and despite the use of an 11 12 algorithm to identify AF among outpatients in the French healthcare databases, identification of AF 13 was mostly based on non-validated discharge and LTD diagnoses recorded in the databases. 14 15 Moreover, it cannot be excluded that the increase in the identified number of patients with AF over 16 17 the 2011-2016 period could be partially explained by changes in LTD legislation in 2011 (e.g. 18 hypertension was removedFor from peer the list of LTD,review while access toonly LTD was facilitated for patients with 19 20 severe arrhythmia and valvular heart diseases) which could have helped identify patients with AF. 21 22 Thirdly, identification of inappropriate underdosing at NOAC initiation was also indirectly assessed by 23 using stroke and bleeding risk scores computed from claims data. Important medical data such as 24 25 patient’s weight, glomerular filtration rate and exact alcohol consumption are not available in the 26 27 French healthcare databases, which may have led to underestimation of the HAS-BLED score and 28 therefore to overestimation of the proportion of patients potentially underdosed at initiation. These 29 30 missing clinical data may also explain the prescription of reduced dose NOAC at treatment initiation 31 32 in clinical practice. Furthermore, the agreement between these empirical scores in patients with AF 33 and the prescriber-assessed stroke and bleeding risk is a subject of discussion [62]. Consequently, the 34 35 rate of inappropriate underdosing should be interpreted with caution and must be confirmed by 36 37 further studies. However, NOAC misuse and underdosing have also been reported in a French 38 prospective field study based on patients’ medical charts [63]. Of note, as INR values were not 39 40 available in the databases, underdosing with VKA therapy was not assessed in this study, but has 41 42 been frequently reported and must not be overlooked [53,64]. In addition, the results for NOAC and 43 VKA new users are difficult to compare, as they were not adjusted for significant differences in 44 45 baseline characteristics and this comparison was not the purpose of this study. 46 47 48 49 50 Conclusion 51 OAC therapy use has modestly increased after the introduction of the NOACs for stroke prevention in 52 53 patients with AF in France and NOAC therapy is now the preferred OAC therapy at initiation in older 54 patients at higher risk of stroke. However, results from this nationwide drug utilization study suggest 55 56 the need for improvement in appropriate prescription of OAC therapy in these patients, especially 57 58 regarding the use of concomitant interacting drugs and the choice of initial NOAC dose. 59 60

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 34

1 2 3 Contributor ship statement 4 5 GM, AP, CB and AN designed the study. All authors have contributed substantially to the 6 interpretation of results. In addition: GM, AP and AN drafted the article; CB, JD and AN conducted 7 8 the statistical analysis; GM, AP, AN, CB, JD, AW provided critical revision of the manuscript for 9 10 important intellectual content. All authors had full access to all of the data (including statistical 11 reports and tables) in the study and can take responsibility for the integrity of the data and the 12 13 accuracy of the data analysis. All authors approved the version to be published. 14 15 16 Competing interests 17 All authors have no conflict of interest. 18 For peer review only 19 20 Funding 21 The authors received no funding. 22 23 24 Data sharing statement 25 No additional data are available directly from the authors. Permanent access to the French 26 27 healthcare databases is automatically granted to certain government agencies, public institutions 28 and public service authorities. Temporary access for studies and research is possible upon request 29 30 from the national health data institute (INDS). All databases used in this study only contained 31 32 anonymous patient records. 33 34 Ethics approval 35 36 This observational study based on the French healthcare databases was approved by the French Data 37 Protection Agency (Commission Nationale de l'Informatique et des Libertés, Cnil) and did not require 38 39 patient consents or ethics committee approval. 40 41 42 Acknowledgments 43 The authors thank Dr Saul, medical translator, for assistance in writing the manuscript. 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 34 BMJ Open

1 2 3 References 4 5 1 Steinberg BA, Piccini JP. Anticoagulation in atrial fibrillation. BMJ 2014;348:g2116–g2116. 6 doi:10.1136/bmj.g2116 7 8 2 Verheugt FWA, Granger CB. Oral anticoagulants for stroke prevention in atrial fibrillation: current 9 status, special situations, and unmet needs. The Lancet 2015;386:303–10. doi:10.1016/S0140- 10 6736(15)60245-8 11 12 3 Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus Warfarin in Patients with Atrial 13 Fibrillation. N Engl J Med 2009;361:1139–51. doi:10.1056/NEJMoa0905561 14 15 4 Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus Warfarin in Nonvalvular Atrial 16 17 Fibrillation. N Engl J Med 2011;365:883–91. doi:10.1056/NEJMoa1009638 18 For peer review only 19 5 Granger CB, Alexander JH, McMurray JJV, et al. Apixaban versus Warfarin in Patients with Atrial 20 Fibrillation. N Engl J Med 2011;365:981–92. doi:10.1056/NEJMoa1107039 21 22 6 Giugliano RP, Ruff CT, Braunwald E, et al. Edoxaban versus Warfarin in Patients with Atrial 23 Fibrillation. N Engl J Med 2013;369:2093–104. doi:10.1056/NEJMoa1310907 24 25 7 Kirchhof P, Benussi S, Kotecha D, et al. 2016 ESC Guidelines for the management of atrial 26 fibrillation developed in collaboration with EACTS. Eur Heart J 2016;37:2893–962. 27 doi:10.1093/eurheartj/ehw210 28 29 30 8 Authors/Task Force Members, Camm AJ, Lip GYH, et al. 2012 focused update of the ESC 31 Guidelines for the management of atrial fibrillation: An update of the 2010 ESC Guidelines for 32 the management of atrial fibrillation * Developed with the special contribution of the European 33 Heart Rhythm Association. Eur Heart J 2012;33:2719–47. doi:10.1093/eurheartj/ehs253 34 35 9 Desai NR, Krumme AA, Schneeweiss S, et al. Patterns of Initiation of Oral Anticoagulants in 36 Patients with Atrial Fibrillation— Quality and Cost Implications. Am J Med 2014;127:1075- 37 1082.e1. doi:10.1016/j.amjmed.2014.05.013 38 39 10 Barnes GD, Lucas E, Alexander GC, et al. National Trends in Ambulatory Oral Anticoagulant Use. 40 Am J Med 2015;128:1300-1305.e2. doi:10.1016/j.amjmed.2015.05.044 41 42 43 11 Haastrup S, Hellfritzsch M, Rasmussen L, et al. Use of Non-Vitamin K Antagonist Oral 44 Anticoagulants 2008-2016: A Danish Nationwide Cohort Study. Basic Clin Pharmacol Toxicol 45 Published Online First: 17 April 2018. doi:10.1111/bcpt.13024 46 47 12 Kjerpeseth LJ, Ellekjær H, Selmer R, et al. Trends in use of warfarin and direct oral anticoagulants 48 in atrial fibrillation in Norway, 2010 to 2015. Eur J Clin Pharmacol Published Online First: 22 July 49 2017. doi:10.1007/s00228-017-2296-1 50 51 13 Staerk L, Fosbøl EL, Gadsbøll K, et al. Non-vitamin K antagonist oral anticoagulation usage 52 according to age among patients with atrial fibrillation: Temporal trends 2011–2015 in Denmark. 53 54 Sci Rep 2016;6. doi:10.1038/srep31477 55 56 14 Huiart L, Ferdynus C, Renoux C, et al. Trends in initiation of direct oral anticoagulant therapies 57 for atrial fibrillation in a national population-based cross-sectional study in the French health 58 insurance databases. BMJ Open 2018;8:e018180. doi:10.1136/bmjopen-2017-018180 59 60

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 34

1 2 3 15 Urbaniak AM, Strøm BO, Krontveit R, et al. Prescription Patterns of Non-Vitamin K Oral 4 Anticoagulants Across Indications and Factors Associated with Their Increased Prescribing in 5 Atrial Fibrillation Between 2012–2015: A Study from the Norwegian Prescription Database. Drugs 6 7 Aging 2017;34:635–45. doi:10.1007/s40266-017-0476-4 8 9 16 Broderick JP, Bonomo JB, Kissela BM, et al. Withdrawal of Antithrombotic Agents and Its Impact 10 on Ischemic Stroke Occurrence. Stroke 2011;42:2509–14. doi:10.1161/STROKEAHA.110.611905 11 12 17 Fang MC, Go AS, Chang Y, et al. Warfarin Discontinuation After Starting Warfarin for Atrial 13 Fibrillation. Circ Cardiovasc Qual Outcomes 2010;3:624–31. 14 doi:10.1161/CIRCOUTCOMES.110.937680 15 16 18 Heidbuchel H, Verhamme P, Alings M, et al. Updated European Heart Rhythm Association 17 practical guide on the use of non-vitamin-K antagonist anticoagulants in patients with non- 18 valvular atrial fibrillation:For Executivepeer summary. review Eur Heart J 2016;:ehw058.only 19 20 doi:10.1093/eurheartj/ehw058 21 22 19 Dillinger J-G, Aleil B, Cheggour S, et al. Dosing issues with non-vitamin K antagonist oral 23 anticoagulants for the treatment of non-valvular atrial fibrillation: Why we should not underdose 24 our patients. Arch Cardiovasc Dis 2018;111:85–94. doi:10.1016/j.acvd.2017.04.008 25 26 20 Marzec LN, Wang J, Shah ND, et al. Influence of Direct Oral Anticoagulants on Rates of Oral 27 Anticoagulation for Atrial Fibrillation. J Am Coll Cardiol 2017;69:2475–84. 28 doi:10.1016/j.jacc.2017.03.540 29 30 21 Steinberg BA, Shrader P, Thomas L, et al. Off-Label Dosing of Non-Vitamin K Antagonist Oral 31 32 Anticoagulants and Adverse Outcomes. J Am Coll Cardiol 2016;68:2597–604. 33 doi:10.1016/j.jacc.2016.09.966 34 35 22 Tuppin P, Rudant J, Constantinou P, et al. Value of a national administrative database to guide 36 public decisions: From the système national d’information interrégimes de l’Assurance Maladie 37 (SNIIRAM) to the système national des données de santé (SNDS) in France. Rev Épidémiol Santé 38 Publique 2017;65:S149–67. doi:10.1016/j.respe.2017.05.004 39 40 23 Bezin J, Duong M, Lassalle R, et al. The national healthcare system claims databases in France, 41 SNIIRAM and EGB: Powerful tools for pharmacoepidemiology. Pharmacoepidemiol Drug Saf 2017 42 43 Aug;26(8):954-962. doi:10.1002/pds.4233 44 45 24 Maura G, Blotière P-O, Bouillon K, et al. Comparison of the Short-Term Risk of Bleeding and 46 Arterial Thromboembolic Events in Nonvalvular Atrial Fibrillation Patients Newly Treated With 47 Dabigatran or Rivaroxaban Versus Vitamin K Antagonists: A French Nationwide Propensity- 48 Matched Cohort Study. Circulation 2015;132:1252–60. 49 doi:10.1161/CIRCULATIONAHA.115.015710 50 51 25 Weill A, Dalichampt M, Raguideau F, et al. Low dose oestrogen combined oral contraception and 52 risk of pulmonary embolism, stroke, and myocardial infarction in five million French women: 53 cohort study. BMJ 2016;353:i2002. 54 55 56 26 Neumann A, Maura G, Weill A, et al. Clinical Events After Discontinuation of β-Blockers in 57 Patients Without Heart Failure Optimally Treated After Acute Myocardial Infarction: A Cohort 58 Study on the French Healthcare Databases. Circ Cardiovasc Qual Outcomes 2018;11:e004356. 59 doi:10.1161/CIRCOUTCOMES.117.004356 60

16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 34 BMJ Open

1 2 3 27 Billionnet C, Alla F, Bérigaud É, et al. Identifying atrial fibrillation in outpatients initiating oral 4 anticoagulants based on medico-administrative data: results from the French national healthcare 5 databases. Pharmacoepidemiol Drug Saf 2017;26:535–43. doi:10.1002/pds.4192 6 7 8 28 Haute autorité de santé. Guide parcours de soins Fibrillation atriale. 2014.https://www.has- 9 sante.fr/portail/jcms/c_1741768/fr/guide-parcours-de-soins-fibrillation-atriale (accessed 18 Jul 10 2018). 11 12 29 Rey G, Jougla E, Fouillet A, et al. Ecological association between a deprivation index and 13 mortality in France over the period 1997 – 2001: variations with spatial scale, degree of 14 urbanicity, age, gender and cause of death. BMC Public Health 2009;9:33. doi:10.1186/1471- 15 2458-9-33 16 17 30 Olesen JB, Lip GYH, Hansen ML, et al. Validation of risk stratification schemes for predicting 18 stroke and thromboembolismFor peer in patients review with atrial fibrillation: only nationwide cohort study. BMJ 19 20 2011;342:d124–d124. doi:10.1136/bmj.d124 21 22 31 Pisters R, Lane DA, Nieuwlaat R, et al. A novel user-friendly score (HAS-BLED) to assess 1-year risk 23 of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest 2010;138:1093– 24 100. doi:10.1378/chest.10-0134 25 26 32 Renoux C, Vahey S, Dell’Aniello S, et al. Association of Selective Serotonin Reuptake Inhibitors 27 With the Risk for Spontaneous Intracranial Hemorrhage. JAMA Neurol 2017;74:173. 28 doi:10.1001/jamaneurol.2016.4529 29 30 33 Cowan JC, Wu J, Hall M, et al. A 10 year study of hospitalized atrial fibrillation-related stroke in 31 32 England and its association with uptake of oral anticoagulation. Eur Heart J Published Online 33 First: 5 July 2018. doi:10.1093/eurheartj/ehy411 34 35 34 Steinberg BA, Gao H, Shrader P, et al. International trends in clinical characteristics and oral 36 anticoagulation treatment for patients with atrial fibrillation: Results from the GARFIELD-AF, 37 ORBIT-AF I, and ORBIT-AF II registries. Am Heart J 2017;194:132–40. 38 doi:10.1016/j.ahj.2017.08.011 39 40 35 Brown JD, Shewale AR, Dherange P, et al. A Comparison of Oral Anticoagulant Use for Atrial 41 Fibrillation in the Pre- and Post-DOAC Eras. Drugs Aging 2016;33:427–36. doi:10.1007/s40266- 42 43 016-0369-y 44 45 36 Apenteng PN, Gao H, Hobbs FR, et al. Temporal trends in antithrombotic treatment of real-world 46 UK patients with newly diagnosed atrial fibrillation: findings from the GARFIELD-AF registry. BMJ 47 Open 2018;8:e018905. doi:10.1136/bmjopen-2017-018905 48 49 37 Alalwan AA, Voils SA, Hartzema AG. Trends in utilization of warfarin and direct oral 50 anticoagulants in older adult patients with atrial fibrillation. Am J Health Syst Pharm 51 2017;74:1237–44. doi:10.2146/ajhp160756 52 53 38 Olesen JB, Sorensen R, Hansen ML, et al. Non-vitamin K antagonist oral anticoagulation agents in 54 anticoagulant naive atrial fibrillation patients: Danish nationwide descriptive data 2011-2013. 55 56 Europace 2015;17:187–93. doi:10.1093/europace/euu225 57 58 39 Lacoin L, Lumley M, Ridha E, et al. Evolving landscape of stroke prevention in atrial fibrillation 59 within the UK between 2012 and 2016: a cross-sectional analysis study using CPRD. BMJ Open 60 2017;7:e015363. doi:10.1136/bmjopen-2016-015363

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 34

1 2 3 40 Averlant L, Ficheur G, Ferret L, et al. Underuse of Oral Anticoagulants and Inappropriate 4 Prescription of Antiplatelet Therapy in Older Inpatients with Atrial Fibrillation. Drugs Aging 5 2017;34:701–10. doi:10.1007/s40266-017-0477-3 6 7 8 41 Alamneh EA, Chalmers L, Bereznicki LR. Suboptimal Use of Oral Anticoagulants in Atrial 9 Fibrillation: Has the Introduction of Direct Oral Anticoagulants Improved Prescribing Practices? 10 Am J Cardiovasc Drugs 2016;16:183–200. doi:10.1007/s40256-016-0161-8 11 12 42 Douros A, Renoux C, Coulombe J, et al. Patterns of long-term use of non-vitamin K antagonist 13 oral anticoagulants for non-valvular atrial fibrillation: Quebec observational study. 14 Pharmacoepidemiol Drug Saf 2017;26:1546–54. doi:10.1002/pds.4333 15 16 43 Komen J, Forslund T, Hjemdahl P, et al. Factors associated with antithrombotic treatment 17 decisions for stroke prevention in atrial fibrillation in the Stockholm region after the introduction 18 of NOACs. Eur J ClinFor Pharmacol peer 2017;73 :1315–22.review doi:10.1007/s00228-017-2289-0 only 19 20 21 44 Patel PA, Zhao X, Fonarow GC, et al. Novel Oral Anticoagulant Use Among Patients With Atrial 22 Fibrillation Hospitalized With Ischemic Stroke or Transient Ischemic Attack. Circ Cardiovasc Qual 23 Outcomes 2015;8:383–92. doi:10.1161/CIRCOUTCOMES.114.000907 24 25 45 Lauffenburger JC, Farley JF, Gehi AK, et al. Factors Driving Anticoagulant Selection in Patients 26 With Atrial Fibrillation in the United States. Am J Cardiol 2015;115:1095–101. 27 doi:10.1016/j.amjcard.2015.01.539 28 29 46 Lip GYH, Laroche C, Boriani G, et al. Regional differences in presentation and treatment of 30 patients with atrial fibrillation in Europe: a report from the EURObservational Research 31 32 Programme Atrial Fibrillation (EORP-AF) Pilot General Registry. Eur Eur Pacing Arrhythm Card 33 Electrophysiol J Work Groups Card Pacing Arrhythm Card Cell Electrophysiol Eur Soc Cardiol 34 2015;17:194–206. doi:10.1093/europace/euu201 35 36 47 Maura G, Billionnet C, Alla F, et al. Comparison of Treatment Persistence with Dabigatran or 37 Rivaroxaban versus Vitamin K Antagonist Oral Anticoagulants in Atrial Fibrillation Patients: A 38 Competing Risk Analysis in the French National Health Care Databases. Pharmacother J Hum 39 Pharmacol Drug Ther 2018;38:6–18. doi:10.1002/phar.2046 40 41 48 Ruff CT, Giugliano RP, Braunwald E, et al. Comparison of the efficacy and safety of new oral 42 43 anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised 44 trials. The Lancet 2014;383:955–62. doi:10.1016/S0140-6736(13)62343-0 45 46 49 Sorensen HT, Riis AH, Lash TL, et al. Statin Use and Risk of Amyotrophic Lateral Sclerosis and 47 Other Motor Neuron Disorders. Circ Cardiovasc Qual Outcomes 2010;3:413–7. 48 doi:10.1161/CIRCOUTCOMES.110.936278 49 50 50 Başaran Ö, Dogan V, Beton O, et al. Suboptimal use of non-vitamin K antagonist oral 51 anticoagulants: Results from the RAMSES study. Medicine (Baltimore) 2016;95:e4672. 52 doi:10.1097/MD.0000000000004672 53 54 51 Ono T, Kohsaka S, Takatsuki S, et al. Inconsistent Dosing of Non–Vitamin K Oral Anticoagulants. J 55 56 Am Coll Cardiol 2017;70:118. doi:10.1016/j.jacc.2017.03.609 57 58 52 Hsu JC, Akao M, Abe M, et al. International Collaborative Partnership for the Study of Atrial 59 Fibrillation (INTERAF): Rationale, Design, and Initial Descriptives. J Am Heart Assoc 60 2016;5:e004037. doi:10.1161/JAHA.116.004037

18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 34 BMJ Open

1 2 3 53 Okumura Y, Yokoyama K, Matsumoto N, et al. Current use of direct oral anticoagulants for atrial 4 fibrillation in Japan: Findings from the SAKURA AF Registry. J Arrhythmia 2017;33:289–96. 5 doi:10.1016/j.joa.2016.11.003 6 7 8 54 Steinberg BA, Shrader P, Pieper K, et al. Frequency and Outcomes of Reduced Dose Non–Vitamin 9 K Antagonist Anticoagulants: Results From ORBIT-AF II (The Outcomes Registry for Better 10 Informed Treatment of Atrial Fibrillation II). J Am Heart Assoc 2018;7:e007633. 11 doi:10.1161/JAHA.117.007633 12 13 55 Pisters R, van Vugt SPG, Brouwer MA, et al. Real-life use of Rivaroxaban in the Netherlands: data 14 from the Xarelto for Prevention of Stroke in Patients with Atrial Fibrillation (XANTUS) registry. 15 Neth Heart J 2017;25:551–8. doi:10.1007/s12471-017-1009-9 16 17 56 Yao X, Shah ND, Sangaralingham LR, et al. Non–Vitamin K Antagonist Oral Anticoagulant Dosing 18 in Patients With AtrialFor Fibrillation peer and Renal review Dysfunction. J Amonly Coll Cardiol 2017;69:2779–90. 19 20 doi:10.1016/j.jacc.2017.03.600 21 22 57 Reilly PA, Lehr T, Haertter S, et al. The Effect of Dabigatran Plasma Concentrations and Patient 23 Characteristics on the Frequency of Ischemic Stroke and Major Bleeding in Atrial Fibrillation 24 Patients. J Am Coll Cardiol 2014;63:321–8. doi:10.1016/j.jacc.2013.07.104 25 26 58 Bouillon K, Bertrand M, Boudali L, et al. Short-Term Risk of Bleeding During Heparin Bridging at 27 Initiation of Vitamin K Antagonist Therapy in More Than 90 000 Patients With Nonvalvular Atrial 28 Fibrillation Managed in Outpatient Care. J Am Heart Assoc 2016;5:e004065. 29 doi:10.1161/JAHA.116.004065 30 31 32 59 Chang S-H, Chou I-J, Yeh Y-H, et al. Association Between Use of Non–Vitamin K Oral 33 Anticoagulants With and Without Concurrent Medications and Risk of Major Bleeding in 34 Nonvalvular Atrial Fibrillation. JAMA 2017;318:1250. doi:10.1001/jama.2017.13883 35 36 60 Van de Werf F, Brueckmann M, Connolly SJ, et al. A comparison of dabigatran etexilate with 37 warfarin in patients with mechanical heart valves: The Randomized, phase II study to Evaluate 38 the sAfety and pharmacokinetics of oraL dabIGatran etexilate in patients after heart valve 39 replacemeNt (RE-ALIGN). Am Heart J 2012;163:931-937.e1. doi:10.1016/j.ahj.2012.03.011 40 41 61 Steffen M. Universalism, Responsiveness, Sustainability — Regulating the French Health Care 42 43 System. N Engl J Med 2016;374:401–5. doi:10.1056/NEJMp1504547 44 45 62 Steinberg BA, Kim S, Thomas L, et al. Lack of Concordance Between Empirical Scores and 46 Physician Assessments of Stroke and Bleeding Risk in Atrial Fibrillation: Results From the 47 Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) Registry. 48 Circulation 2014;129:2005–12. doi:10.1161/CIRCULATIONAHA.114.008643 49 50 63 Lafon T, Vallejo C, Hadj M, et al. Mésusage et iatrogénie des anticoagulants oraux directs (AOD) : 51 étude observationnelle dans le service des urgences du CHU de Limoges. Thérapie 2018 May - 52 Jun;73(3):209-215. doi: 10.1016/j.therap.2017.05.004. 53 54 64 Pokorney SD, Simon DN, Thomas L, et al. Stability of International Normalized Ratios in Patients 55 56 Taking Long-term Warfarin Therapy. JAMA 2016;316:661–3. doi:10.1001/jama.2016.9356 57 58 59 60

19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 34

1 2 3 Figure titles and legends 4 5 6 Figure 1. Time trends in the use of oral antithrombotic therapy between 2011 and 2016 in patients 7 8 with atrial fibrillation in France 9 10 Abbreviations: AF: atrial fibrillation; OAC: oral anticoagulant; VKA: vitamin K antagonist; NOAC: Non 11 vitamin K antagonist oral anticoagulant 12 13 Legend: For Figures 1.b., 1.d and 1.e., estimates from all AF patients already presented in figure 1.a. 14 15 are indicated by dashed lines for the purposes of comparison. 16 17 18 Figure 2. Patient flowFor chart. peer review only 19 20 Abbreviations: OAC: oral anticoagulant; NOAC: Non-vitamin K antagonist oral anticoagulant; VKA: 21 vitamin K antagonist; DVT/PE: deep vein thrombosis/pulmonary embolism; AF: atrial fibrillation 22 23 24 25 Figure 3. Potential NOAC underdosing in new users with AF 26 Abbreviations: NOAC: Non-vitamin K antagonist oral anticoagulant; AF: atrial fibrillation 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

20 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 34 BMJ Open

1 2 3 4 Table 1. Baseline characteristics of anticoagulant-naive patients with atrial fibrillation initiating oral anticoagulants in 2015-2016 5 NOAC 6 VKA Characteristics (N; %*) Dabigatran Rivaroxaban Apixaban Total NOAC 7 N= 65,010 N= 9,085 N= 54,456 N= 64,300 N= 127,841 8 NOAC: reduced doses 5,652 (62.2) 19,429 (35.7) 26,003 (40.4) 51,084 (40.0) NA 9 Female sex 4,546 (50.0) 26,147 (48.0) 33,375 (51.9) 64,068 (50.1) 33,865 (52.1) 10 Age (years), mean (SD) 74.1 (11.6) 72.8 (11.9) 75.3 (11.3) 74.1 (11.6) 78.0 (11.3) 11 18-54 548 (6.0) 4,007 (7.4) 3,172 (4.9) 7,727 (6.0) 2,361 (3.6) 12 55-64 For1,117 (12.3)peer7,651 (14.0) review7,117 (11.1) 15,885 only (12.4) 5,680 (8.7) 13 65-74 2,514 (27.7) 16,057 (29.5) 16,645 (25.9) 35,216 (27.5) 12,969 (19.9) 14 75-79 1,554 (17.1) 9,053 (16.6) 10,692 (16.6) 21,299 (16.7) 9,587 (14.7) 15 ≥80 3,352 (36.9) 17,688 (32.5) 26,674 (41.5) 47,714 (37.3) 34,413 (52.9) 16 ≥90 493 (5.4) 2,559 (4.7) 4,654 (7.2) 7,706 (6.0) 8,399 (12.9) 17 Deprivation index 18 Quintile 1 (least deprived) 1,394 (15.3) 10,265 (18.9) 11,266 (17.5) 22,925 (17.9) 10,263 (15.8) Quintile 2 1,586 (17.5) 10,678 (19.6) 12,496 (19.4) 24,760 (19.4) 11,884 (18.3) 19 Quintile 3 1,780 (19.6) 10,701 (19.7) 12,799 (19.9) 25,280 (19.8) 12,811 (19.7) 20 Quintile 4 1,917 (21.1) 10,794 (19.8) 13,142 (20.4) 25,853 (20.2) 14,272 (22.0) 21 Quintile 5 (most deprived)) 2,113 (23.3) 11,172 (20.5) 13,825 (21.5) 27,110 (21.2) 14,699 (22.6) 22 Overseas departments 295 (3.2) 846 (1.6) 772 (1.2) 1,913 (1.5) 1,081 (1.7) 23 First prescriber’s specialty 24 Hospital practitioner 3,720 (40.9) 22,905 (42.1) 29,316 (45.6) 55,941 (43.8) 39,083 (60.1) 25 General practitioner 2,062 (22.7) 11,145 (20.5) 11,590 (18.0) 24,797 (19.4) 15,539 (23.9) 26 Private cardiologist 3,093 (34.0) 18,978 (34.9) 21,843 (34.0) 43,914 (34.4) 8,511 (13.1) 27 Private orthopaedic surgeon 16 (0.2) 100 (0.2) 95 (0.1) 211 (0.2) 73 (0.1) 28 Other private specialist 168 (1.8) 1,149 (2.1) 1,276 (2.0) 2,593 (2.0) 1,583 (2.4) CHA DS -VASc score† 29 2 2 Mean score (SD) 3.7 (1.6) 3.5 (1.6) 3.9 (1.6) 3.7 (1.6) 4.5 (1.6) 30 0 183 (2.0) 1,294 (2.4) 847 (1.3) 2,324 (1.8) 309 (0.5) 31 1 635 (7.0) 4,854 (8.9) 3,637 (5.7) 9,126 (7.1) 1,607 (2.5) 32 ≥ 2 8,267 (91.0) 48,308 (88.7) 59,816 (93.0) 116,391 (90.1) 63,094 (97.0) 33 C (Heart failure) 2,849 (31.4) 17,805 (32.7) 23,548 (36.6) 44,202 (34.6) 32,727 (50.3) 34 H (antihypertensive drugs) 7,547 (83.1) 44,260 (81.3) 54,596 (84.9) 106,403 (83.2) 59,139 (91.0) 35 D(iabetes) 1,959 (21.6) 11,279 (20.7) 14,087 (21.9) 27,325 (21.4) 18,806 (28.9) 36 S(troke: ATE) 1,207 (13.3) 4,930 (9.1) 8,448 (13.1) 14,585 (11.4) 10,638 (16.4) 37 V(ascular diseases) 2,242 (24.7) 13,924 (25.6) 18,766 (29.2) 34,932 (27.3) 28,894 (44.4) † 38 Age≥75 and arterial thromboembolic events 769 (8.5) 3,126 (5.7) 5,608 (8.7) 9,503 (7.4) 7,762 (11.9) † 39 Age<65 and no arterial thromboembolic events 1,510 (16.6) 11,074 (20.3) 9,396 (14.6) 21,980 (17.2) 7,035 (10.8) 40 41 42 21 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 22 of 34

1 2 3 Table 1. (continued) 4 NOAC 5 VKA Characteristics (N; %) Dabigatran Rivaroxaban Apixaban Total NOAC 6 N= 65,010 7 N= 9,085 N= 54,456 N= 64,300 N= 127,841 † 8 HAS-BLED score 9 Mean score (SD) 2.0 (0.9) 1.9 (0.9) 2.1 (0.9) 2.0 (0.9) 2.7 (1.0) ≥ 3 2,169 (23.9) 11,388 (20.9) 16,834 (26.2) 30,391 (23.8) 36,417 (56.0) 10 A(bnormal) 11 Renal function 345 (3.8) 2,426 (4.5) 3,822 (5.9) 6,593 (5.2) 14,260 (21.9) 12 Liver function For169 (1.9)peer967 (1.8) review1,106 (1.7) 2,242 only (1.8) 2,372 (3.6) 13 B(leeding) 14 Predisposition 182 (2.0) 1,161 (2.1) 1,605 (2.5) 2,948 (2.3) 5,873 (9.0) 15 Major bleeding 692 (7.6) 3,729 (6.8) 5,134 (8) 9,555 (7.5) 9,348 (14.4) 16 D(rug/alcohol) 17 Alcohol abuse‡ 272 (3.0) 1,698 (3.1) 1,730 (2.7) 3,700 (2.9) 2,923 (4.5) 18 Drug-drug interactions 947 (10.4) 5,838 (10.7) 7,570 (11.8) 14,355 (11.2) 23,451 (36.1) 19 Parenteral anticoagulant (heparin) 64 (0.7) 331 (0.6) 361 (0.6) 756 (0.6) 9,824 (15.1) 20 Antiplatelet drugs 826 (9.1) 5,225 (9.6) 6,951 (10.8) 13,002 (10.2) 15,433 (23.7) NSAIDs 72 (0.8) 365 (0.7) 344 (0.5) 781 (0.6) 212 (0.3) 21 Other comorbidities† 22 Ischaemic heart disease 1,821 (20.0) 11,321 (20.8) 15,439 (24.0) 28,581 (22.4) 23,657 (36.4) 23 Frailty (proxies) 1,666 (18.3) 8,971 (16.5) 12,730 (19.8) 23,367 (18.3) 24,175 (37.2) 24 Dementia or Parkinson’s disease 524 (5.8) 3,204 (5.9) 4,125 (6.4) 7,853 (6.1) 7,437 (11.4) 25 Psychiatric disorders 1,722 (19.0) 10,593 (19.5) 12,844 (20.0) 25,159 (19.7) 16,598 (25.5) 26 Smoking‡ 1,024 (11.3) 6,481 (11.9) 7,442 (11.6) 14,947 (11.7) 11,434 (17.6) 27 Comedications§ 28 Antiarrhythmics or cardiac glycosides 5,996 (66.0) 35,761 (65.7) 41,031 (63.8) 82,788 (64.8) 35,600 (54.8) 29 Lipid-lowering agents 3,913 (43.1) 22,250 (40.9) 28,812 (44.8) 54,975 (43.0) 31,903 (49.1) 30 Oral corticosteroids 1,105 (12.2) 6,964 (12.8) 8,079 (12.6) 16,148 (12.6) 8,967 (13.8) 31 Antiulcer agents 4,295 (47.3) 24,842 (45.6) 31,469 (48.9) 60,606 (47.4) 39,842 (61.3) 32 Polymedication (≥5 ATC classes) 3,750 (41.3) 21,725 (39.9) 28,196 (43.9) 53,671 (42.0) 45,153 (69.5) 33 Polymedication (≥10 ATC classes) 738 (8.1) 4,653 (8.5) 6,246 (9.7) 11,637 (9.1) 14,947 (23.0) 34 NOAC: Non-vitamin K antagonist oral anticoagulant; VKA: vitamin K antagonist; SD: standard deviation; ATE: arterial thromboembolic events (ischaemic stroke, arterial systemic embolism or transient ischaemic 35 attack); NSAIDs: non-steroidal anti-inflammatory drugs 36 * Unless otherwise stated † Comorbidities were defined using a rolling 1-year period following the initiation of OAC therapy 37 ‡ Smoking or alcohol data: measured by using proxies such as reimbursements for specific therapy or hospitalizations related to smoking or alcohol consumption/diseases 38 § Comorbidities were defined using a rolling 4-month period preceding the initiation of OAC therapy 39 40 41 42 22 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 23 of 34 BMJ Open

1 2 3 4 Table 2. Potential inappropriate use of NOAC therapy in oral anticoagulant-naïve patients with atrial fibrillation in 2015-2016 5 6 NOAC VKA 7 Characteristics (N; %) Dabigatran Rivaroxaban Apixaban Total NOAC N= 65,010 8 N= 9,085 N= 54,456 N= 64,300 N= 127,841 9 Contraindications or non-approved indication/dose 1,457 (16.0) 8,614 (15.8) 9,542 (14.8) 19,613 (15.3) NA 10 Any valvular heart disease 649 (7.1) 4,146 (7.6) 6,122 (9.5) 10,917 (8.5) 16,461 (25.3) 11 Prosthetic heart valve (mechanical or bioprosthetic valves) 106 (1.2) 665 (1.2) 1,096 (1.7) 1,867 (1.5) 6,726 (10.3) Recently hospitalized for coagulopathy, purpura and other 12 90 (1.0) 479 (0.9) 596 (0.9) 1,165 (0.9) 2,142 (3.3) haemorrhagic conditions* For peer review only 13 Liver fibrosis and cirrhosis* 50 (0.6) 282 (0.5) 367 (0.6) 699 (0.5) 1,174 (1.8) 14 Recent gastrointestinal ulceration or intracranial haemorrhage† 40 (0.4) 120 (0.2) 248 (0.4) 408 (0.3) 350 (0.5) 15 Recently or currently treated cancer* 417 (4.6) 2,531 (4.6) 2,898 (4.5) 5,846 (4.6) 4,252 (6.5) 16 Reduced-dose NOAC not approved for stroke prevention in AF 357 (3.9) 1,844 (3.4) NA 2,201 (3.5) NA 17 patients in Europe 18 Inappropriate use during follow-up‡ 19 No monitoring of renal function at initiation 637 (16.8) 3804 (15.5) 3642 (14.5) 8083 (15.1) 5401 (17.0) 20 No monitoring of renal function in the year post-initiation 378 (10.0) 2,347 (9.6) 2,138 (8.5) 4,863 (9.1) 2,984 (9.4) 21 Non-persistence patterns, N (%) 22 One reimbursement only 605 (15.9) 2,666 (10.9) 1,771 (7.1) 5,042 (9.4) 2,426 (7.6) § 23 One-year treatment discontinuation rates 984 (25.9) 6210 (25.4) 4524 (18,0) 11,718 (21,9) 8399 (26.4) Concomitant use of drug increasing the risk of bleeding¶ 2,639 (29.3) 15,797 (29.3) 18,556 (29.2) 36,992 (29.3) 33,025 (52.3) 24 Antiplatelet agents or parenteral anticoagulants 1,728 (19.2) 10,386 (19.3) 12,382 (19.5) 24,496 (19.4) 28,112 (44.5) 25 Parenteral anticoagulants 287 (3.2) 1,577 (2.9) 1,520 (2.4) 3,384 (2.7) 12,078 (19.1) 26 Antiplatelet agents 1,490 (16.6) 9,179 (17.0) 11,170 (17.6) 21,839 (17.3) 19,710 (31.2) 27 Aspirin 1,336 (14.8) 8,215 (15.2) 10,026 (15.8) 19,577 (15.5) 17,770 (28.1) 28 NSAIDs 375 (4.2) 2,111 (3.9) 2,017 (3.2) 4,503 (3.6) 1,030 (1.6) 29 SSRIs and SSNRIs 836 (9.3) 4,852 (9.0) 5,880 (9.2) 11,568 (9.1) 7,317 (11.6) 30 31 NOAC: Non-vitamin K antagonist oral anticoagulant; VKA: vitamin K antagonist; AF: atrial fibrillation; NA: not applicable; NSAIDs: non-steroidal anti-inflammatory drugs; SSRIs: selective serotonin reuptake inhibitors and SSNRIs: selective serotonin-norepinephrine reuptake inhibitors 32 * Comorbidities identified using hospitalization and/or LTD data, and/or specific procedures during a rolling 1-year period preceding the initiation of OAC therapy 33 † Comorbidities identified using hospitalization data during a rolling 6-week period preceding the initiation of OAC therapy 34 ‡ Data on patients with at least a one year of follow-up i.e. patients initiating OAC in 2015 after excluding patients who died and those hospitalized for 3 months or longer (N= 3,796; 24,483; 25,118; 53,397 and 35 31,777 for dabigatran-, rivaroxaban, apixaban, total NOAC- and total VKA new users, respectively); period considered (unless otherwise stated): rolling 1-year period following the initiation of OAC therapy (index date included) 36 § One-year crude discontinuation rate for patients initiating OAC in 2015 who died and those hospitalized for 3 months or longer, defined as prolonged treatment discontinuation i.e. 90-day gap with no medication 37 coverage after the 30-day coverage period of a refill 38 ¶ Data for new users still alive after a 45-day period following the index date (N= 9,001; 53,885; 63,578; 126,464 and 63,180 for dabigatran-, rivaroxaban, apixaban, total NOAC- and total VKA new users , 39 respectively); period considered: rolling 6-week period following the initiation of OAC therapy (index date included) 40 41 42 23 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 24 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Figure 1 32 33 253x190mm (300 x 300 DPI) 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 34 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Figure 2 46 47 190x253mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Figure 3 32 33 253x190mm (300 x 300 DPI) 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 34 BMJ Open

1 2 3 4 Supplementary Materials 5 6 7 Supplementary Table 1. Definitions used to identify comorbid conditions and comedications in the French healthcare databases. 8 9 Supplementary Table 2. Age- and sex-adjusted association between baseline characteristics and the choice of NOAC therapy compared to VKA therapy in 10 11 OAC new users. 12 Supplementary Table 3. Comparison ofFor selected baseline peer characteristics review of DOAC new users with only AF at risk of stroke according to level of HAS-BLED score 13 14 and type of DOAC dose. 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 1

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 28 of 34

1 2 3 Supplementary Table 1. Definitions used to identify comorbid conditions and comedications in the SNIIRAM-PMSI databases. 4 5 Covariates* Hospital discharge diagnoses† LTD diagnoses† Specific procedures or drug reimbursements 6 AF definition (Patterns of use of DOAC therapy in new 7 users with AF) Nonvalvular atrial fibrillation I48 I48 Radiofrequency ablation, cardioversion 8 Lower limb orthopaedic procedures Lower limb orthopaedic surgery or procedures 9 History of valvular heart disease/Prosthetic heart valve I05-I09, I33-I39/ Z95.2, Z95.3 ou Z95.4 Heart valve surgery 10 Deep vein thrombosis/pulmonary embolism I26, I80 (except I80.0), I81, I82 I26, I80-I82 Lower limb venous ultrasonography, pulmonary/lower limb 11 angiography, ventilation/perfusion scan CHA2DS2-VASc and HAS-BLED scores 12 Heart failure ForI50 or I11.0, peer I13.0, I13.2, I13.9, K76.1, review J81 related to I50 I50 onlySpecific medications approved for heart failure including beta-blockers 13 (bisoprolol, carvedilol, metoprolol), eplerenone, bumetanide, 14 furosemide when licensed for heart failure only 15 Antihypertensive drugs Diuretics, beta blockers, calcium channel blockers, agents acting on the renin-angiotensin system and other antiadrenergic agents 16 Diabetes E10-E14 E10-E14 Insulins and blood glucose lowering drugs 17 History of ATE (ischemic stroke, arterial systemic embolism I63 (except I63.6), G46 related to I63 or I69.3; I74, G45 I63, I74, G45 18 or transient ischemic attack) (except G45.4) 19 Peripheral vascular disease I20, I21, I22, I23, I24, I25, I70, I71, I72, I73, E10.5, I70-I73 , I20-I25 20 E11.5, E12.5, E13.5, E14.5 Abnormal renal function N18, I12.0, I13.1, I13.2, E10.2, E11.2, E13.2, E14.2, N18, I12 21 N08.3 22 Abnormal liver function R18, I85, K70, K71.3, K71.4, K71.5, K71.7, K72.1, K73, K70, K73, K74, B18, C22 Antiviral agents for systemic use HCV (ribavirin, [peg]interferon alpha 23 K74, K76.1, B18, C22, C78.7 and DAA), and against HCB ([peg]interferon alpha and NnRTIs). 24 Major bleeding I60-I62, S06.3-S06.6, I85, K25-K29, K62.5, K92, D62, N02, R31, R58, H11.3, H35.6, H43.1, H45.0, H92.2, 25 J94.2, K66.1, M25.0, N92, N93.8, N93.9, N95.0, R04.0 26 Predisposition for bleeding Anemias : D50, D51, D52, D53, D55, D56, D57, D58, 27 D59, D60, D62, D63, D64 or coagulation defects, 28 purpura and other hemorrhagic: D65, D66, D67, D68, 29 D69 Alcohol abuse F10, K70, T51, K6, G31.2, G62.1, G72.1, I42.6, K29.2, disulfiram, acamprosate, nalmefene and products with naltrexone 30 Z71.4, Z72.1, Z50.2. licensed in adult patients with alcohol dependence 31 Drug-drug interaction Platelet aggregation inhibitors (including low-dose acetylsalicylic acid ), 32 heparins, NSAIDs 33 Other comorbidities Ischemic heart disease (including myocardial infarction) I20-I25 I20-I25 Nitrovasodilator agents 34 Frailty G81-G83 G81-G83 Home hospital bed, wheelchair, high level of nursing home stay 35 Dementia or Parkinson’s disease F00-F03, G30, G20 F00-F03, G30, G20 Anticholinesterases or NMDA receptor antagonists: anticholinergic and 36 dopaminergic agents 37 Psychiatric disorders F20-29, F30-39 F20-29, F30-39 Antidepressants, antipsychotics, conventional mood stabilizers Smoking F17, Z71.6, Z72.0; J43-J44 J43-J44 Nicotine replacement therapy 38 Recently hospitalized for coagulopathy, purpura and other D65-D69 39 haemorrhagic conditions 40 Liver fibrosis and cirrhosis K70.0, K70.2, K70.3, K70.4, K71.7, K72, K74 41 42 2

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 29 of 34 BMJ Open

1 2 3 Recent gastrointestinal ulceration or intracranial K25, K26, K27, K28, K29.0, I60, I61, I62, S06.4, S06.5, 4 haemorrhage S06.6 5 Recently or currently treated cancer C00-D09, D37-D48, Z510, Z511 C00-D09, D37-D48 Radiotherapy procedures Baseline comedications 6 Antiarrhythmics and cardiac glycosides Antiarrhythmics: class I and III, verapamil, digitalis glycosides 7 Lipid-lowering agents HMG CoA reductase inhibitors, fibrates, ezetimibe 8 Antiplatelet drugs Platelet aggregation inhibitors including low-dose acetylsalicylic acid 9 Oral corticosteroids Mineralo- and gluco-corticoids 10 Antiulcer agents Mainly proton pump inhibitors and H2-receptor antagonists

11 LTD: long-term diseases; ATE: arterial thromboembolic events (mainly stroke); DVT/PE: deep vein thrombosis/pulmonary embolism; NMDA: N-methyl-D-aspartate; HIV: human immunodeficiency virus; PI: Protease 12 inhibitor; NNRTIs: Non-nucleoside reverse transcriptaseFor inhibitors; NnRTI: peer Nucleoside and nucleotide review reverse transcriptase inhibitors; DAA:only direct-acting antiviral; HMG CoA: 3-hydroxy-3-methyl-glutaryl-CoA; NSAIDs: 13 Non-steroidal anti-inflammatory drugs. 14 15 * Comorbidities were identified using ICD-10 diagnosis codes for hospital discharge/LTD, or specific procedures, or drug reimbursements. Concomitant medications were identified as those dispensed at least once during the 4-month period preceding the index date. Influenza vaccination was determined during the first ‘flu vaccination campaign preceding the index date. 16 †ICD-10 codes 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 3

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 30 of 34

1 2 3 Supplementary Table 2. Age- and sex-adjusted association between baseline characteristics and the choice of NOAC therapy compared to VKA therapy in 4 OAC new users. 5 Dabigatran vs VKA Rivaroxaban vs VKA Apixaban vs VKA 6 Baseline characteristics 7 RR* 95% CI p-value† RR* 95% CI p-value† RR* 95% CI p-value† 8 Female sex 1.11 1.05 1.17 *** 1.03 1.00 1.07 * 1.07 1.04 1.09 *** 9 Age (years) 10 18-54 (Ref) 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 11 55-64 0.87 0.78 0.98 * 0.92 0.87 0.97 ** 0.97 0.92 1.02 12 For peer review only 13 65-74 0.86 0.77 0.95 ** 0.89 0.84 0.94 *** 0.98 0.93 1.03

14 75-79 0.73 0.65 0.81 *** 0.77 0.73 0.82 *** 0.91 0.87 0.96 *** 15 80-84 0.58 0.52 0.65 *** 0.65 0.61 0.68 *** 0.83 0.79 0.87 *** 16 17 85-89 0.44 0.39 0.49 *** 0.50 0.48 0.54 *** 0.74 0.70 0.77 *** 18 >=90 0.28 0.25 0.32 *** 0.37 0.35 0.40 *** 0.61 0.57 0.65 *** 19 Deprivation index 20 quintile 1 (least deprived) (Ref) 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 21 quintile 2 0.97 0.91 1.05 0.94 0.92 0.97 *** 0.97 0.95 1.00 * 22 quintile 3 0.99 0.93 1.07 0.90 0.87 0.92 *** 0.94 0.92 0.97 *** 23 24 quintile 4 0.96 0.89 1.03 0.85 0.83 0.87 *** 0.90 0.88 0.93 ***

25 quintile 5 (most deprived) 1.00 0.94 1.07 0.85 0.83 0.87 *** 0.91 0.89 0.93 *** 26 Overseas departments 1.60 1.41 1.82 *** 0.83 0.77 0.89 *** 0.77 0.72 0.83 *** 27 28 First prescriber’s specialty 29 Hospital practitioner (Ref) 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

30 General practitioner 1.37 1.29 1.46 *** 1.16 1.10 1.23 *** 1.00 0.96 1.04

31 Private cardiologist 2.92 2.76 3.08 *** 1.80 1.70 1.90 *** 1.65 1.59 1.71 *** 32 Private orthopedic surgeon 1.96 1.20 3.21 ** 1.48 1.20 1.81 *** 1.32 1.07 1.61 ** 33 Other private specialist 1.07 0.93 1.24 1.12 1.04 1.21 ** 1.03 0.97 1.10 34 35 From CHA2DS2-VASc score 36 Heart Failure 0.52 0.49 0.54 *** 0.70 0.68 0.71 *** 0.77 0.75 0.78 *** 37 Antihypertensive drugs 0.60 0.57 0.63 *** 0.74 0.72 0.76 *** 0.80 0.78 0.82 *** 38 Diabetes 0.67 0.63 0.71 *** 0.76 0.74 0.77 *** 0.80 0.79 0.82 *** 39 40 ATE 0.85 0.79 0.92 *** 0.71 0.68 0.74 *** 0.90 0.88 0.92 *** 41 42 4

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 31 of 34 BMJ Open

1 2 3 Vascular diseases 0.47 0.42 0.53 *** 0.63 0.61 0.66 *** 0.71 0.67 0.74 *** 4 From HAS-BLED score 5 6 Abnormal renal function 0.17 0.15 0.19 *** 0.31 0.30 0.33 *** 0.40 0.38 0.42 *** 7 Abnormal liver function 0.46 0.39 0.54 *** 0.58 0.54 0.62 *** 0.60 0.57 0.64 *** 8 Bleeding predisposition 0.53 0.47 0.59 *** 0.63 0.60 0.65 *** 0.70 0.67 0.72 *** 9 Major bleeding 0.23 0.20 0.26 *** 0.36 0.33 0.39 *** 0.41 0.38 0.45 *** 10 11 Alcohol abuse 0.55 0.48 0.62 *** 0.69 0.65 0.72 *** 0.68 0.65 0.72 *** 12 Drug-drug interactions 0.22For 0.20 peer 0.24 *** review 0.36 0.34 0.37 ***only 0.40 0.37 0.43 *** 13 Parenteral anticoagulants 0.04 0.03 0.06 *** 0.07 0.06 0.07 *** 0.07 0.06 0.07 *** 14 Antiplatelets drugs 0.34 0.31 0.37 *** 0.49 0.47 0.51 *** 0.56 0.53 0.60 *** 15 16 NSAIDs 1.80 1.43 2.28 *** 1.24 1.12 1.38 *** 1.18 1.06 1.32 ** 17 Other comorbidities 18 Valvular heart diseases 0,25 0,20 0,31 *** 0,40 0,36 0,45 *** 0,47 0,41 0,54 *** 19 Ischemic heart diseases 0.51 0.46 0.56 *** 0.65 0.63 0.67 *** 0.74 0.71 0.77 *** 20 21 Frailty (proxies) 0.48 0.45 0.51 *** 0.58 0.55 0.61 *** 0.64 0.62 0.66 *** 22 Dementia or Parkinson’s disease 0.66 0.60 0.72 *** 0.82 0.78 0.85 *** 0.77 0.75 0.80 *** 23 Psychiatric disorders 0.74 0.70 0.78 *** 0.86 0.84 0.88 *** 0.86 0.84 0.88 *** 24 Smoking 0.56 0.52 0.60 *** 0.71 0.69 0.74 *** 0.74 0.71 0.76 *** 25 26 Comedications 27 Antiarrhythmics or cardiac glycosides 1.42 1.35 1.50 *** 1.22 1.19 1.25 *** 1.18 1.16 1.20 *** 28 Lipid-lowering agents 0.76 0.71 0.81 *** 0.81 0.79 0.84 *** 0.90 0.87 0.92 *** 29 Oral corticosteroids 0.85 0.80 0.91 *** 0.94 0.91 0.97 *** 0.94 0.91 0.96 *** 30 31 Antiulcer agents 0.61 0.56 0.66 *** 0.73 0.71 0.75 *** 0.78 0.75 0.81 *** 32 Polymedication (≥5 ATC classes) 0.37 0.35 0.40 *** 0.55 0.54 0.57 *** 0.60 0.59 0.62 *** 33 Polymedication (≥10 ATC classes) 0.33 0.30 0.36 *** 0.49 0.46 0.52 *** 0.54 0.51 0.57 *** 34 35 NOAC: Non vitamin K antagonist oral anticoagulant; VKA: vitamin K antagonist; OAC: oral anticoagulant; RR: Relative risk; IC; Confidence interval; ATE: arterial thromboembolic events 36 (ischaemic stroke, arterial systemic embolism or transient ischaemic attack); NSAIDs: non-steroidal anti-inflammatory drugs 37 * RR determined using negative binomial regression analysis adjusting for age and sex(except for age and sex covariates, adjusted for sex and for age only, respectively). 38 † p-values : *** p < 0,001 ; ** p < 0,01 ; * p < 0,05 Reading example: Age and sex being equal, frailty reduced the probability for a prescription of dabigatran (instead of VKA) by 52% (1 minus the estimated RR of 0.48). For rivaroxaban and 39 apixaban, this reduction was 42% and 36%, respectively. 40 41 42 5

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 32 of 34

1 2 3 4 Supplementary Table 3. Comparison of selected baseline characteristics of DOAC new users with AF at risk of stroke according to level of HAS-BLED score 5 and type of DOAC dose. 6 7 8 Patients with CHA2DS2-VASc ≥ 2 (N=116,391) Patients aged ≥75 years old and history of ATE (N=9,503) N (%) 9 HAS-BLED≥3 (N=30 273) HAS-BLED<3 (N=86 118) HAS-BLED≥3 (N=7,143) HAS-BLED<3 (N=2,360) 10 Reduced doses* Standard doses* Reduced doses* Standard doses* Reduced doses* Standard doses* Reduced doses* Standard doses* 11 (N=16,090) (N=14,183) (N=33,845) (N=52,273) (N=4,178) (N=2,965) (N=1,379) (N=981) 12 For peer review only 13 Age (years), mean 81.8 74.8 81.3 70.9 84,8 81.1 81.3 70.9 14 Age ≥ 80 years 10,786 (67.0) 3,982 (28.1) 22,962 (67.8) 9,984 (19.1) 3,496 (83.7) 1,720 (58.0) 1,188 (86.1) 567 (57.8) 15 First prescriber’s specialty 16 Hospital practitioner or private 12,315 (76.6) 11,364 (80.1) 25,221 (74.5) 41,728 (79.8) 3216 (77.0) 2965 (80.2) 1013 (73.5) 798 (81.3) 17 cardiologist 18 General practitioner 3321 (20.6) 2426 (17.1) 7836 (23.2) 9302 (17.8) 860 (20.6) 503 (17.0) 322 (23.4) 157 (16.0) 19 20 Frailty (proxies) 6,689 (41.6) 4,600 (32.4) 6,932 (20.5) 4,779 (9.1) 2,599 (62.2) 1,642 (55.4) 589 (42.7) 316 (32.2) 21 Dementia or Parkinson's disease 2121 (13.2) 935 (6.6) 3169 (9.4) 1555 (3.0) 789 (18.9) 355 (12.0) 259 (18.8) 111 (11.3) 22 Any valvular diseases† 2,640 (16.4) 1,809 (12.8) 2,740 (8.1) 3,342 (6.4) 566 (13.5) 330 (11.1) 144 (10.4) 95 (9.7) 23 Prosthetic heart valve† 544 (3.4) 384 (2.7) 425 (1.3) 474 (0.9) 94 (2.2) 52 (1.8) 19 (1.4) 14 (1.4) 24 25 Ischaemic heart disease 7,232 (44.9) 5,274 (37.2) 6,827 (20.2) 9,142 (17.5) 1,338 (32.0) 758 (25.6) 336 (24.4) 185 (18.9) 26 Antiplatelet agents at initiation 6,035 (37.5) 4,891 (34.5) 620 (1.8) 1,264 (2.4) 874 (20.9) 510 (17.2) 11 (0.8) 7 (0.7) 27 Lipid-lowering agents 9252 (57.5) 9213 (65.0) 12836 (37.9) 21820 (41.7) 2606 (62.4) 2132 (71.9) 697 (50.5) 611 (62.3) 28 Polymedication (≥10 ATC classes) 3,627 (22.5) 2,946 (20.8) 2,150 (6.4) 2,798 (5.4) 788 (18.9) 473 (16.0) 104 (7.5) 67 (6.8) 29

30 * Reduced-dose DOAC: dabigatran 150mg, Rivaroxaban 20 and apixaban 5mg; reduced-dose DOAC: dabigatran 75mg and 110mg, rivaroxaban 10mg and 15mg and apixaban 2.5mg 31 † Covariates defined by hospitalization data only 32 33 34 35 36 37 38 39 40 41 42 6

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 33 of 34 BMJ Open

Oral anticoagulation therapy use in patients with atrial fibrillation after the introduction of non-vitamin K antagonist oral anticoagulants: findings from the French healthcare databases, 2011-2016

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2018-026645.R2

Article Type: Research

Date Submitted by the 04-Feb-2019 Author:

Primary Subject Cardiovascular medicine Heading:

Cardiovascular medicine, Epidemiology, Pharmacology and therapeutics, Secondary Subject Heading: Public health

Anticoagulation < HAEMATOLOGY, dabigatran, rivaroxaban, apixaban, Keywords: claims database, France

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 36 BMJ Open

1 2 3 4 Title page 5 6 7 Oral anticoagulation therapy use in patients with atrial fibrillation after the introduction of non- 8 9 vitamin K antagonist oral anticoagulants: findings from the French healthcare databases, 2011- 10 11 2016. 12 13 14 Running title: Patterns of OAC use 15 16 17 Géric Maura PharmD, PhD1,2, Cécile Billionnet MSc, PhD1, Jérôme Drouin, MSc1, Alain Weill MD1, 18 For peer review only 19 1 2,3 20 Anke Neumann MSc, PhD , Antoine Pariente MD, PhD 21 22 1 Department of Studies in Public Health, French National Health Insurance (Caisse Nationale de 23 24 l’Assurance Maladie/Cnam), 75 986 Paris Cedex 20, France. 25 26 2 University of Bordeaux, Inserm, Bordeaux Population Health Research Center, team 27 28 29 Pharmacoepidemiology - UMR 1219, F-33000 Bordeaux, France 30 31 3 CHU de Bordeaux, Pharmacologie, F-33000 Bordeaux, France 32 33 Address for correspondence: Géric MAURA, Cnam/DSES/DESP, National Health Insurance, 50 avenue 34 35 du Pr. André Lemierre, 75986 Paris cedex 20; E-mail: [email protected]; Telephone: 36 37 38 +33(0)172602338; Fax: +33(0)172601724 39 40 41 42 43 Word count of text: (excluding the title page, abstract, tables, acknowledgements, contributions and 44 45 references): 4,231 46 Word count of abstract: 300/300 47 48 Number of references: 64 49 50 Number of tables: 2 51 52 Number of figures: 3 53 Number of supplementary tables: 3 54 55 56 57 58 59 60

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 36

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 36 BMJ Open

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 36

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 36 BMJ Open

5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 36

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 36 BMJ Open

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 36

8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 36 BMJ Open

9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 36

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 36 BMJ Open

1 2 3 4 Discussion 5 6 Main findings 7 8 OAC therapy use among patients with AF improved in France between 2011 and 2016, but remained 9 suboptimal with about 66% of the 1.1 million patients with identified AF treated by OAC therapy in 10 11 2016. Use of antiplatelet therapy in AF patients decreased over the same period, but still concerned 12 13 33% of all AF patients in 2016. Patients with AF were more likely to be treated by VKA than NOAC 14 therapy, including older patients and those at higher risk of stroke. 15 16 Nearly 193,000 patients with AF were identified as OAC new users in 2015-2016: patients were more 17 18 likely to be treated byFor NOAC thanpeer VKA therapy, review including older only patients and those at higher risk of 19 20 stroke. Results on current patterns of use of NOAC therapy in new users with AF suggest several 21 situations of inappropriate use, including frequent concomitant use of drugs increasing the risk of 22 23 bleeding and potential inappropriate underdosing. 24 25 26 Comparison with postmarketing literature and clinical implications 27 The overall improvement of management of AF patients observed with regards to OAC therapy after 28 29 the introduction of NOACs has been reported in many countries [33–36]. The same applies to the 30 steady decrease in VKA use in favour of NOAC therapies [12,37,38], and the gaps remaining in 31 32 optimal OAC coverage and high antiplatelet drug use [39–41]. 33 34 This study demonstrated channelling of NOAC therapy towards patients at lower risk of stroke and 35 bleeding when considering all patients with AF, in line with what has become a common feature 36 37 reported worldwide in clinical practice by many observational studies on the current patterns of use 38 39 of NOACs [15,42–45]. In particular, data from the ESC-sponsored ‘EURObservational Research 40 Programme on AF’ (EORP-AF) General Long-Term Registry showed that younger age, having fewer 41 42 risk factors or a history of non-valvular heart diseases were also found to be clinical predictors for 43 44 being treated with NOACs vs. VKAs [46]. 45 However, NOAC therapy is now the preferred OAC therapy at initiation in the oldest patients and 46 47 those at higher risk of stroke. This French pattern of OAC use has never been previously reported and 48 49 contrasts with published results concerning earlier periods [47]. This emerging pattern is encouraging 50 for AF management, as older and high-risk patients are those who should derive most benefit from 51 52 NOAC versus VKA therapy [48]. 53 54 Reduced doses were often prescribed in OAC new users, including dabigatran 75mg and rivaroxaban 55 10mg, which are not approved for stroke prevention in Europe [49]. In addition to the overall 56 57 channelling mentioned above, these findings may reflect a “bleeding avoidance” strategy of 58 59 prescribers (i.e. overestimation of the potential bleeding risk versus the likely benefit of stroke 60

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 36

1 2 3 reduction) and the differential perception of the comparative safety of NOACs versus VKA and 4 5 between NOACs. The early safety alert on bleeding in dabigatran-treated patients, followed by the 6 7 contraindication of this NOAC in patients with prosthetic heart valves, may have reinforced the fears 8 of prescribers in relation to the safety of dabigatran, which would explain the difference in reduced- 9 10 dose prescription rates between the three NOACs in this study, despite the intermediate stroke and 11 12 bleeding risk profile of dabigatran compared to that of rivaroxaban and apixaban in new users. 13 Similarly, among NOAC new users, apixaban was prescribed to the oldest and most severe patients. 14 15 Apixaban was the only NOAC found to be superior to warfarin for all types of bleeding outcome and 16 17 all-cause mortality, which would also illustrate the tendency of physicians to prescribe OAC therapies 18 according to bleedingFor risk [5]. peer This may review also explain the only potential inappropriate underdosing 19 20 observed in NOAC users in this study. This pattern of NOAC use has been previously reported, but 21 22 mostly in field and registry studies based on small sample sizes. The reported inappropriate 23 underdosing rate varies according to studies and the definition used. NOAC underdosing concerned 24 25 30.4% of Turkish patients in the RAMSES study (N=2,086) [50], 18.4% of Japanese patients of the KiCS 26 27 AF registry (N=1,284) [51,52], between 19.7% and 27.6% of patients in the SAKURA AF registry (N= 28 3,266) [53], and 9.4% to 16% of patients in the ORBIT-AF II registry (N=7925) [21,54]. In the subgroup 29 30 of Dutch patients (N=899) enrolled in the XANTUS registry, 33% of patients were also treated with 31 32 reduced-dose rivaroxaban despite presenting normal renal function [55]. Using a large U.S. 33 administrative database, Yao et al found that 13.3% of the 13,392 NOAC new users with no renal 34 35 indication for dose reduction were potentially underdosed [56]. Taken together with our results, 36 37 these data suggest that inappropriate underdosing might be a common issue in NOAC new users that 38 should be systematically assessed when studying NOAC patterns of use. This is of particular concern, 39 40 as recent data have suggested a relationship between NOAC dose and clinical outcomes [57]. In 41 42 particular, NOAC underdosing has been shown to be associated with increased risk for adverse 43 outcomes [21,56]. 44 45 These patterns of NOAC use contrast with the other patterns concomitantly observed in this study, 46 47 such as the high level of concomitant prescription of antiplatelet agents and parenteral 48 anticoagulants or, to a lesser extent, NOAC use in non-approved indications such as prosthetic heart 49 50 valves that are both associated with an increased risk of bleeding [58–60]. 51 52 53 Strengths and limitations 54 This study is the first to report the improved trend in OAC coverage in French patients with AF over 55 56 the last five years as well as the recent patterns of use of OAC therapy in new users, particularly 57 58 including a nationwide assessment of the growing issue of NOAC underdosing, based on health data 59 for more than 50 million beneficiaries. Moreover, all OAC prescriptions filled in the ambulatory 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 36 BMJ Open

1 2 3 setting are captured in the databases and are reimbursed with no restriction of coverage: selection 4 5 bias related to the access of patients to more expensive NOAC therapy is therefore not an issue with 6 7 the use of French healthcare databases [22,61]. 8 However, several limitations related to the nature of the data used should be underlined. First of all, 9 10 it cannot be verified whether patients actually took the drugs for which they were reimbursed. 11 12 Secondly, as the indication for treatment is not available in the databases, and despite the use of an 13 algorithm to identify AF among outpatients in the French healthcare databases, identification of AF 14 15 was mostly based on non-validated discharge and LTD diagnoses recorded in the databases. 16 17 Moreover, it cannot be excluded that the increase in the identified number of patients with AF over 18 the 2011-2016 periodFor could peer be partially explainedreview by changes only in LTD legislation in 2011 (e.g. 19 20 hypertension was removed from the list of LTD, while access to LTD was facilitated for patients with 21 22 severe arrhythmia and valvular heart diseases) which could have helped identify patients with AF. 23 Thirdly, identification of inappropriate underdosing at NOAC initiation was also indirectly assessed by 24 25 using stroke and bleeding risk scores computed from claims data. Important medical data such as 26 27 patient’s weight, glomerular filtration rate and exact alcohol consumption are not available in the 28 French healthcare databases, which may have led to underestimation of the HAS-BLED score and 29 30 therefore to overestimation of the proportion of patients potentially underdosed at initiation. These 31 32 missing clinical data may also explain the prescription of reduced dose NOAC at treatment initiation 33 in clinical practice. Furthermore, the agreement between these empirical scores in patients with AF 34 35 and the prescriber-assessed stroke and bleeding risk is a subject of discussion [62]. Consequently, the 36 37 rate of inappropriate underdosing should be interpreted with caution and must be confirmed by 38 further studies. However, NOAC misuse and underdosing have also been reported in a French 39 40 prospective field study based on patients’ medical charts [63]. Of note, as INR values were not 41 42 available in the databases, underdosing with VKA therapy was not assessed in this study, but has 43 been frequently reported and must not be overlooked [53,64]. In addition, as stated in the 2016 ESC 44 45 guidelines [7], HAS-BLED score is not designed to evaluate prescription of NOAC type and dosage and 46 47 no longer must be used for this purpose in clinical practice. 48 Finally, the results for NOAC and VKA new users are difficult to compare, as they were not adjusted 49 50 for significant differences in baseline characteristics and this comparison was not the purpose of this 51 52 study. 53 54 55 56 Conclusion 57 58 OAC therapy use has modestly increased after the introduction of the NOACs for stroke prevention in 59 patients with AF in France and NOAC therapy is now the preferred OAC therapy at initiation in older 60

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 36

1 2 3 patients at higher risk of stroke. However, results from this nationwide drug utilization study suggest 4 5 the need for improvement in appropriate prescription of OAC therapy in these patients, especially 6 7 regarding the use of concomitant interacting drugs and the choice of initial NOAC dose. 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 36 BMJ Open

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 36

1 2 3 References 4 5 1 Steinberg BA, Piccini JP. Anticoagulation in atrial fibrillation. BMJ 2014;348:g2116–g2116. 6 doi:10.1136/bmj.g2116 7 8 2 Verheugt FWA, Granger CB. Oral anticoagulants for stroke prevention in atrial fibrillation: current 9 status, special situations, and unmet needs. The Lancet 2015;386:303–10. doi:10.1016/S0140- 10 6736(15)60245-8 11 12 3 Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus Warfarin in Patients with Atrial 13 Fibrillation. N Engl J Med 2009;361:1139–51. doi:10.1056/NEJMoa0905561 14 15 4 Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus Warfarin in Nonvalvular Atrial 16 17 Fibrillation. N Engl J Med 2011;365:883–91. doi:10.1056/NEJMoa1009638 18 For peer review only 19 5 Granger CB, Alexander JH, McMurray JJV, et al. Apixaban versus Warfarin in Patients with Atrial 20 Fibrillation. N Engl J Med 2011;365:981–92. doi:10.1056/NEJMoa1107039 21 22 6 Giugliano RP, Ruff CT, Braunwald E, et al. Edoxaban versus Warfarin in Patients with Atrial 23 Fibrillation. N Engl J Med 2013;369:2093–104. doi:10.1056/NEJMoa1310907 24 25 7 Kirchhof P, Benussi S, Kotecha D, et al. 2016 ESC Guidelines for the management of atrial 26 fibrillation developed in collaboration with EACTS. Eur Heart J 2016;37:2893–962. 27 doi:10.1093/eurheartj/ehw210 28 29 30 8 Authors/Task Force Members, Camm AJ, Lip GYH, et al. 2012 focused update of the ESC 31 Guidelines for the management of atrial fibrillation: An update of the 2010 ESC Guidelines for 32 the management of atrial fibrillation * Developed with the special contribution of the European 33 Heart Rhythm Association. Eur Heart J 2012;33:2719–47. doi:10.1093/eurheartj/ehs253 34 35 9 Desai NR, Krumme AA, Schneeweiss S, et al. Patterns of Initiation of Oral Anticoagulants in 36 Patients with Atrial Fibrillation— Quality and Cost Implications. Am J Med 2014;127:1075- 37 1082.e1. doi:10.1016/j.amjmed.2014.05.013 38 39 10 Barnes GD, Lucas E, Alexander GC, et al. National Trends in Ambulatory Oral Anticoagulant Use. 40 Am J Med 2015;128:1300-1305.e2. doi:10.1016/j.amjmed.2015.05.044 41 42 43 11 Haastrup S, Hellfritzsch M, Rasmussen L, et al. Use of Non-Vitamin K Antagonist Oral 44 Anticoagulants 2008-2016: A Danish Nationwide Cohort Study. Basic Clin Pharmacol Toxicol 45 Published Online First: 17 April 2018. doi:10.1111/bcpt.13024 46 47 12 Kjerpeseth LJ, Ellekjær H, Selmer R, et al. Trends in use of warfarin and direct oral anticoagulants 48 in atrial fibrillation in Norway, 2010 to 2015. Eur J Clin Pharmacol 2017;73:1417–25. 49 doi:10.1007/s00228-017-2296-1 50 51 13 Staerk L, Fosbøl EL, Gadsbøll K, et al. Non-vitamin K antagonist oral anticoagulation usage 52 according to age among patients with atrial fibrillation: Temporal trends 2011–2015 in Denmark. 53 54 Sci Rep 2016;6. doi:10.1038/srep31477 55 56 14 Huiart L, Ferdynus C, Renoux C, et al. Trends in initiation of direct oral anticoagulant therapies 57 for atrial fibrillation in a national population-based cross-sectional study in the French health 58 insurance databases. BMJ Open 2018;8:e018180. doi:10.1136/bmjopen-2017-018180 59 60

16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 36 BMJ Open

1 2 3 15 Urbaniak AM, Strøm BO, Krontveit R, et al. Prescription Patterns of Non-Vitamin K Oral 4 Anticoagulants Across Indications and Factors Associated with Their Increased Prescribing in 5 Atrial Fibrillation Between 2012–2015: A Study from the Norwegian Prescription Database. Drugs 6 7 Aging 2017;34:635–45. doi:10.1007/s40266-017-0476-4 8 9 16 Broderick JP, Bonomo JB, Kissela BM, et al. Withdrawal of Antithrombotic Agents and Its Impact 10 on Ischemic Stroke Occurrence. Stroke 2011;42:2509–14. doi:10.1161/STROKEAHA.110.611905 11 12 17 Fang MC, Go AS, Chang Y, et al. Warfarin Discontinuation After Starting Warfarin for Atrial 13 Fibrillation. Circ Cardiovasc Qual Outcomes 2010;3:624–31. 14 doi:10.1161/CIRCOUTCOMES.110.937680 15 16 18 Heidbuchel H, Verhamme P, Alings M, et al. Updated European Heart Rhythm Association 17 practical guide on the use of non-vitamin-K antagonist anticoagulants in patients with non- 18 valvular atrial fibrillation:For Executivepeer summary. review Eur Heart J 2016;:ehw058.only 19 20 doi:10.1093/eurheartj/ehw058 21 22 19 Dillinger J-G, Aleil B, Cheggour S, et al. Dosing issues with non-vitamin K antagonist oral 23 anticoagulants for the treatment of non-valvular atrial fibrillation: Why we should not underdose 24 our patients. Arch Cardiovasc Dis 2018;111:85–94. doi:10.1016/j.acvd.2017.04.008 25 26 20 Marzec LN, Wang J, Shah ND, et al. Influence of Direct Oral Anticoagulants on Rates of Oral 27 Anticoagulation for Atrial Fibrillation. J Am Coll Cardiol 2017;69:2475–84. 28 doi:10.1016/j.jacc.2017.03.540 29 30 21 Steinberg BA, Shrader P, Thomas L, et al. Off-Label Dosing of Non-Vitamin K Antagonist Oral 31 32 Anticoagulants and Adverse Outcomes. J Am Coll Cardiol 2016;68:2597–604. 33 doi:10.1016/j.jacc.2016.09.966 34 35 22 Tuppin P, Rudant J, Constantinou P, et al. Value of a national administrative database to guide 36 public decisions: From the système national d’information interrégimes de l’Assurance Maladie 37 (SNIIRAM) to the système national des données de santé (SNDS) in France. Rev DÉpidémiologie 38 Santé Publique 2017;65:S149–67. doi:10.1016/j.respe.2017.05.004 39 40 23 Bezin J, Duong M, Lassalle R, et al. The national healthcare system claims databases in France, 41 SNIIRAM and EGB: Powerful tools for pharmacoepidemiology. Pharmacoepidemiol Drug Saf 42 43 Published Online First: 24 May 2017. doi:10.1002/pds.4233 44 45 24 Maura G, Blotière P-O, Bouillon K, et al. Comparison of the Short-Term Risk of Bleeding and 46 Arterial Thromboembolic Events in Nonvalvular Atrial Fibrillation Patients Newly Treated With 47 Dabigatran or Rivaroxaban Versus Vitamin K AntagonistsCLINICAL PERSPECTIVES: A French 48 Nationwide Propensity-Matched Cohort Study. Circulation 2015;132:1252–60. 49 doi:10.1161/CIRCULATIONAHA.115.015710 50 51 25 Weill A, Dalichampt M, Raguideau F, et al. Low dose oestrogen combined oral contraception and 52 risk of pulmonary embolism, stroke, and myocardial infarction in five million French women: 53 cohort study. BMJ 2016;353:i2002. 54 55 56 26 Neumann A, Maura G, Weill A, et al. Clinical Events After Discontinuation of β-Blockers in 57 Patients Without Heart Failure Optimally Treated After Acute Myocardial Infarction: A Cohort 58 Study on the French Healthcare Databases. Circ Cardiovasc Qual Outcomes 2018;11:e004356. 59 doi:10.1161/CIRCOUTCOMES.117.004356 60

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 36

1 2 3 27 Billionnet C, Alla F, Bérigaud É, et al. Identifying atrial fibrillation in outpatients initiating oral 4 anticoagulants based on medico-administrative data: results from the French national healthcare 5 databases: Identification of Atrial Fibrillation in Claims. Pharmacoepidemiol Drug Saf 6 7 2017;26:535–43. doi:10.1002/pds.4192 8 9 28 Haute autorité de santé. Guide parcours de soins Fibrillation atriale. 2014.https://www.has- 10 sante.fr/portail/jcms/c_1741768/fr/guide-parcours-de-soins-fibrillation-atriale (accessed 18 Jul 11 2018). 12 13 29 Rey G, Jougla E, Fouillet A, et al. Ecological association between a deprivation index and 14 mortality in France over the period 1997 – 2001: variations with spatial scale, degree of 15 urbanicity, age, gender and cause of death. BMC Public Health 2009;9:33. doi:10.1186/1471- 16 2458-9-33 17 18 30 Olesen JB, Lip GYH,For Hansen peer ML, et al. Validation review of risk stratification only schemes for predicting 19 20 stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study. BMJ 21 2011;342:d124–d124. doi:10.1136/bmj.d124 22 23 31 Pisters R, Lane DA, Nieuwlaat R, et al. A novel user-friendly score (HAS-BLED) to assess 1-year risk 24 of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest 2010;138:1093– 25 100. doi:10.1378/chest.10-0134 26 27 32 Renoux C, Vahey S, Dell’Aniello S, et al. Association of Selective Serotonin Reuptake Inhibitors 28 With the Risk for Spontaneous Intracranial Hemorrhage. JAMA Neurol 2017;74:173. 29 doi:10.1001/jamaneurol.2016.4529 30 31 32 33 Cowan JC, Wu J, Hall M, et al. A 10 year study of hospitalized atrial fibrillation-related stroke in 33 England and its association with uptake of oral anticoagulation. Eur Heart J Published Online 34 First: 5 July 2018. doi:10.1093/eurheartj/ehy411 35 36 34 Steinberg BA, Gao H, Shrader P, et al. International trends in clinical characteristics and oral 37 anticoagulation treatment for patients with atrial fibrillation: Results from the GARFIELD-AF, 38 ORBIT-AF I, and ORBIT-AF II registries. Am Heart J 2017;194:132–40. 39 doi:10.1016/j.ahj.2017.08.011 40 41 35 Brown JD, Shewale AR, Dherange P, et al. A Comparison of Oral Anticoagulant Use for Atrial 42 43 Fibrillation in the Pre- and Post-DOAC Eras. Drugs Aging 2016;33:427–36. doi:10.1007/s40266- 44 016-0369-y 45 46 36 Apenteng PN, Gao H, Hobbs FR, et al. Temporal trends in antithrombotic treatment of real-world 47 UK patients with newly diagnosed atrial fibrillation: findings from the GARFIELD-AF registry. BMJ 48 Open 2018;8:e018905. doi:10.1136/bmjopen-2017-018905 49 50 37 Alalwan AA, Voils SA, Hartzema AG. Trends in utilization of warfarin and direct oral 51 anticoagulants in older adult patients with atrial fibrillation. Am J Health Syst Pharm 52 2017;74:1237–44. doi:10.2146/ajhp160756 53 54 38 Olesen JB, Sorensen R, Hansen ML, et al. Non-vitamin K antagonist oral anticoagulation agents in 55 56 anticoagulant naive atrial fibrillation patients: Danish nationwide descriptive data 2011-2013. 57 Europace 2015;17:187–93. doi:10.1093/europace/euu225 58 59 60

18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 36 BMJ Open

1 2 3 39 Lacoin L, Lumley M, Ridha E, et al. Evolving landscape of stroke prevention in atrial fibrillation 4 within the UK between 2012 and 2016: a cross-sectional analysis study using CPRD. BMJ Open 5 2017;7:e015363. doi:10.1136/bmjopen-2016-015363 6 7 8 40 Averlant L, Ficheur G, Ferret L, et al. Underuse of Oral Anticoagulants and Inappropriate 9 Prescription of Antiplatelet Therapy in Older Inpatients with Atrial Fibrillation. Drugs Aging 10 2017;34:701–10. doi:10.1007/s40266-017-0477-3 11 12 41 Alamneh EA, Chalmers L, Bereznicki LR. Suboptimal Use of Oral Anticoagulants in Atrial 13 Fibrillation: Has the Introduction of Direct Oral Anticoagulants Improved Prescribing Practices? 14 Am J Cardiovasc Drugs 2016;16:183–200. doi:10.1007/s40256-016-0161-8 15 16 42 Douros A, Renoux C, Coulombe J, et al. Patterns of long-term use of non-vitamin K antagonist 17 oral anticoagulants for non-valvular atrial fibrillation: Quebec observational study. 18 PharmacoepidemiolFor Drug Safpeer 2017;26:1546–54. review doi:10.1002/pds.4333 only 19 20 21 43 Komen J, Forslund T, Hjemdahl P, et al. Factors associated with antithrombotic treatment 22 decisions for stroke prevention in atrial fibrillation in the Stockholm region after the introduction 23 of NOACs. Eur J Clin Pharmacol 2017;73:1315–22. doi:10.1007/s00228-017-2289-0 24 25 44 Patel PA, Zhao X, Fonarow GC, et al. Novel Oral Anticoagulant Use Among Patients With Atrial 26 Fibrillation Hospitalized With Ischemic Stroke or Transient Ischemic Attack. Circ Cardiovasc Qual 27 Outcomes 2015;8:383–92. doi:10.1161/CIRCOUTCOMES.114.000907 28 29 45 Lauffenburger JC, Farley JF, Gehi AK, et al. Factors Driving Anticoagulant Selection in Patients 30 With Atrial Fibrillation in the United States. Am J Cardiol 2015;115:1095–101. 31 32 doi:10.1016/j.amjcard.2015.01.539 33 34 46 Boriani G, Proietti M, Laroche C, et al. Contemporary stroke prevention strategies in 11 096 35 European patients with atrial fibrillation: a report from the EURObservational Research 36 Programme on Atrial Fibrillation (EORP-AF) Long-Term General Registry. Eur Eur Pacing Arrhythm 37 Card Electrophysiol J Work Groups Card Pacing Arrhythm Card Cell Electrophysiol Eur Soc Cardiol 38 2018;20:747–57. doi:10.1093/europace/eux301 39 40 47 Maura G, Billionnet C, Alla F, et al. Comparison of Treatment Persistence with Dabigatran or 41 Rivaroxaban versus Vitamin K Antagonist Oral Anticoagulants in Atrial Fibrillation Patients: A 42 43 Competing Risk Analysis in the French National Health Care Databases. Pharmacother J Hum 44 Pharmacol Drug Ther 2018;38:6–18. doi:10.1002/phar.2046 45 46 48 Ruff CT, Giugliano RP, Braunwald E, et al. Comparison of the efficacy and safety of new oral 47 anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised 48 trials. The Lancet 2014;383:955–62. doi:10.1016/S0140-6736(13)62343-0 49 50 49 Sorensen HT, Riis AH, Lash TL, et al. Statin Use and Risk of Amyotrophic Lateral Sclerosis and 51 Other Motor Neuron Disorders. Circ Cardiovasc Qual Outcomes 2010;3:413–7. 52 doi:10.1161/CIRCOUTCOMES.110.936278 53 54 50 Başaran Ö, Dogan V, Beton O, et al. Suboptimal use of non-vitamin K antagonist oral 55 56 anticoagulants: Results from the RAMSES study. Medicine (Baltimore) 2016;95:e4672. 57 doi:10.1097/MD.0000000000004672 58 59 51 Ono T, Kohsaka S, Takatsuki S, et al. Inconsistent Dosing of Non–Vitamin K Oral Anticoagulants. J 60 Am Coll Cardiol 2017;70:118. doi:10.1016/j.jacc.2017.03.609

19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 36

1 2 3 52 Hsu JC, Akao M, Abe M, et al. International Collaborative Partnership for the Study of Atrial 4 Fibrillation (INTERAF): Rationale, Design, and Initial Descriptives. J Am Heart Assoc 5 2016;5:e004037. doi:10.1161/JAHA.116.004037 6 7 8 53 Okumura Y, Yokoyama K, Matsumoto N, et al. Current use of direct oral anticoagulants for atrial 9 fibrillation in Japan: Findings from the SAKURA AF Registry. J Arrhythmia 2017;33:289–96. 10 doi:10.1016/j.joa.2016.11.003 11 12 54 Steinberg BA, Shrader P, Pieper K, et al. Frequency and Outcomes of Reduced Dose Non–Vitamin 13 K Antagonist Anticoagulants: Results From ORBIT-AF II (The Outcomes Registry for Better 14 Informed Treatment of Atrial Fibrillation II). J Am Heart Assoc 2018;7:e007633. 15 doi:10.1161/JAHA.117.007633 16 17 55 Pisters R, van Vugt SPG, Brouwer MA, et al. Real-life use of Rivaroxaban in the Netherlands: data 18 from the Xarelto Forfor Prevention peer of Stroke review in Patients with Atrialonly Fibrillation (XANTUS) registry. 19 20 Neth Heart J 2017;25:551–8. doi:10.1007/s12471-017-1009-9 21 22 56 Yao X, Shah ND, Sangaralingham LR, et al. Non–Vitamin K Antagonist Oral Anticoagulant Dosing 23 in Patients With Atrial Fibrillation and Renal Dysfunction. J Am Coll Cardiol 2017;69:2779–90. 24 doi:10.1016/j.jacc.2017.03.600 25 26 57 Reilly PA, Lehr T, Haertter S, et al. The Effect of Dabigatran Plasma Concentrations and Patient 27 Characteristics on the Frequency of Ischemic Stroke and Major Bleeding in Atrial Fibrillation 28 Patients. J Am Coll Cardiol 2014;63:321–8. doi:10.1016/j.jacc.2013.07.104 29 30 58 Bouillon K, Bertrand M, Boudali L, et al. Short-Term Risk of Bleeding During Heparin Bridging at 31 32 Initiation of Vitamin K Antagonist Therapy in More Than 90 000 Patients With Nonvalvular Atrial 33 Fibrillation Managed in Outpatient Care. J Am Heart Assoc 2016;5:e004065. 34 doi:10.1161/JAHA.116.004065 35 36 59 Chang S-H, Chou I-J, Yeh Y-H, et al. Association Between Use of Non–Vitamin K Oral 37 Anticoagulants With and Without Concurrent Medications and Risk of Major Bleeding in 38 Nonvalvular Atrial Fibrillation. JAMA 2017;318:1250. doi:10.1001/jama.2017.13883 39 40 60 Van de Werf F, Brueckmann M, Connolly SJ, et al. A comparison of dabigatran etexilate with 41 warfarin in patients with mechanical heart valves: The Randomized, phase II study to Evaluate 42 43 the sAfety and pharmacokinetics of oraL dabIGatran etexilate in patients after heart valve 44 replacemeNt (RE-ALIGN). Am Heart J 2012;163:931-937.e1. doi:10.1016/j.ahj.2012.03.011 45 46 61 Steffen M. Universalism, Responsiveness, Sustainability — Regulating the French Health Care 47 System. N Engl J Med 2016;374:401–5. doi:10.1056/NEJMp1504547 48 49 62 Steinberg BA, Kim S, Thomas L, et al. Lack of Concordance Between Empirical Scores and 50 Physician Assessments of Stroke and Bleeding Risk in Atrial Fibrillation: Results From the 51 Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) Registry. 52 Circulation 2014;129:2005–12. doi:10.1161/CIRCULATIONAHA.114.008643 53 54 63 Lafon T, Vallejo C, Hadj M, et al. Mésusage et iatrogénie des anticoagulants oraux directs (AOD) : 55 56 étude observationnelle dans le service des urgences du CHU de Limoges. Thérapie Published 57 Online First: July 2017. doi:10.1016/j.therap.2017.05.004 58 59 64 Pokorney SD, Simon DN, Thomas L, et al. Stability of International Normalized Ratios in Patients 60 Taking Long-term Warfarin Therapy. JAMA 2016;316:661–3. doi:10.1001/jama.2016.9356

20 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 36 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

21 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 36

22 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 36 BMJ Open

1 2 3 4 Table 1. Baseline characteristics of anticoagulant-naive patients with atrial fibrillation initiating oral anticoagulants in 2015-2016 5 NOAC 6 VKA Characteristics (N; %*) Dabigatran Rivaroxaban Apixaban Total NOAC 7 N= 65,010 N= 9,085 N= 54,456 N= 64,300 N= 127,841 8 NOAC: reduced doses 5,652 (62.2) 19,429 (35.7) 26,003 (40.4) 51,084 (40.0) NA 9 Female sex 4,546 (50.0) 26,147 (48.0) 33,375 (51.9) 64,068 (50.1) 33,865 (52.1) 10 Age (years), mean (SD) 74.1 (11.6) 72.8 (11.9) 75.3 (11.3) 74.1 (11.6) 78.0 (11.3) 11 18-54 548 (6.0) 4,007 (7.4) 3,172 (4.9) 7,727 (6.0) 2,361 (3.6) 12 55-64 For1,117 (12.3)peer7,651 (14.0) review7,117 (11.1) 15,885 only (12.4) 5,680 (8.7) 13 65-74 2,514 (27.7) 16,057 (29.5) 16,645 (25.9) 35,216 (27.5) 12,969 (19.9) 14 75-79 1,554 (17.1) 9,053 (16.6) 10,692 (16.6) 21,299 (16.7) 9,587 (14.7) 15 ≥80 3,352 (36.9) 17,688 (32.5) 26,674 (41.5) 47,714 (37.3) 34,413 (52.9) 16 ≥90 493 (5.4) 2,559 (4.7) 4,654 (7.2) 7,706 (6.0) 8,399 (12.9) 17 Deprivation index 18 Quintile 1 (least deprived) 1,394 (15.3) 10,265 (18.9) 11,266 (17.5) 22,925 (17.9) 10,263 (15.8) Quintile 2 1,586 (17.5) 10,678 (19.6) 12,496 (19.4) 24,760 (19.4) 11,884 (18.3) 19 Quintile 3 1,780 (19.6) 10,701 (19.7) 12,799 (19.9) 25,280 (19.8) 12,811 (19.7) 20 Quintile 4 1,917 (21.1) 10,794 (19.8) 13,142 (20.4) 25,853 (20.2) 14,272 (22.0) 21 Quintile 5 (most deprived)) 2,113 (23.3) 11,172 (20.5) 13,825 (21.5) 27,110 (21.2) 14,699 (22.6) 22 Overseas departments 295 (3.2) 846 (1.6) 772 (1.2) 1,913 (1.5) 1,081 (1.7) 23 First prescriber’s specialty 24 Hospital practitioner 3,720 (40.9) 22,905 (42.1) 29,316 (45.6) 55,941 (43.8) 39,083 (60.1) 25 General practitioner 2,062 (22.7) 11,145 (20.5) 11,590 (18.0) 24,797 (19.4) 15,539 (23.9) 26 Private cardiologist 3,093 (34.0) 18,978 (34.9) 21,843 (34.0) 43,914 (34.4) 8,511 (13.1) 27 Private orthopaedic surgeon 16 (0.2) 100 (0.2) 95 (0.1) 211 (0.2) 73 (0.1) 28 Other private specialist 168 (1.8) 1,149 (2.1) 1,276 (2.0) 2,593 (2.0) 1,583 (2.4) CHA DS -VASc score† 29 2 2 Mean score (SD) 3.7 (1.6) 3.5 (1.6) 3.9 (1.6) 3.7 (1.6) 4.5 (1.6) 30 0 183 (2.0) 1,294 (2.4) 847 (1.3) 2,324 (1.8) 309 (0.5) 31 1 635 (7.0) 4,854 (8.9) 3,637 (5.7) 9,126 (7.1) 1,607 (2.5) 32 ≥ 2 8,267 (91.0) 48,308 (88.7) 59,816 (93.0) 116,391 (90.1) 63,094 (97.0) 33 C (Heart failure) 2,849 (31.4) 17,805 (32.7) 23,548 (36.6) 44,202 (34.6) 32,727 (50.3) 34 H (antihypertensive drugs) 7,547 (83.1) 44,260 (81.3) 54,596 (84.9) 106,403 (83.2) 59,139 (91.0) 35 D(iabetes) 1,959 (21.6) 11,279 (20.7) 14,087 (21.9) 27,325 (21.4) 18,806 (28.9) 36 S(troke: ATE) 1,207 (13.3) 4,930 (9.1) 8,448 (13.1) 14,585 (11.4) 10,638 (16.4) 37 V(ascular diseases) 2,242 (24.7) 13,924 (25.6) 18,766 (29.2) 34,932 (27.3) 28,894 (44.4) † 38 Age≥75 and arterial thromboembolic events 769 (8.5) 3,126 (5.7) 5,608 (8.7) 9,503 (7.4) 7,762 (11.9) † 39 Age<65 and no arterial thromboembolic events 1,510 (16.6) 11,074 (20.3) 9,396 (14.6) 21,980 (17.2) 7,035 (10.8) 40 41 42 23 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 24 of 36

1 2 3 Table 1. (continued) 4 NOAC 5 VKA Characteristics (N; %) Dabigatran Rivaroxaban Apixaban Total NOAC 6 N= 65,010 7 N= 9,085 N= 54,456 N= 64,300 N= 127,841 † 8 HAS-BLED score 9 Mean score (SD) 2.0 (0.9) 1.9 (0.9) 2.1 (0.9) 2.0 (0.9) 2.7 (1.0) ≥ 3 2,169 (23.9) 11,388 (20.9) 16,834 (26.2) 30,391 (23.8) 36,417 (56.0) 10 A(bnormal) 11 Renal function 345 (3.8) 2,426 (4.5) 3,822 (5.9) 6,593 (5.2) 14,260 (21.9) 12 Liver function For169 (1.9)peer967 (1.8) review1,106 (1.7) 2,242 only (1.8) 2,372 (3.6) 13 B(leeding) 14 Predisposition 182 (2.0) 1,161 (2.1) 1,605 (2.5) 2,948 (2.3) 5,873 (9.0) 15 Major bleeding 692 (7.6) 3,729 (6.8) 5,134 (8) 9,555 (7.5) 9,348 (14.4) 16 D(rug/alcohol) 17 Alcohol abuse‡ 272 (3.0) 1,698 (3.1) 1,730 (2.7) 3,700 (2.9) 2,923 (4.5) 18 Drug-drug interactions 947 (10.4) 5,838 (10.7) 7,570 (11.8) 14,355 (11.2) 23,451 (36.1) 19 Parenteral anticoagulant (heparin) 64 (0.7) 331 (0.6) 361 (0.6) 756 (0.6) 9,824 (15.1) 20 Antiplatelet drugs 826 (9.1) 5,225 (9.6) 6,951 (10.8) 13,002 (10.2) 15,433 (23.7) NSAIDs 72 (0.8) 365 (0.7) 344 (0.5) 781 (0.6) 212 (0.3) 21 Other comorbidities† 22 Ischaemic heart disease 1,821 (20.0) 11,321 (20.8) 15,439 (24.0) 28,581 (22.4) 23,657 (36.4) 23 Frailty (proxies) 1,666 (18.3) 8,971 (16.5) 12,730 (19.8) 23,367 (18.3) 24,175 (37.2) 24 Dementia or Parkinson’s disease 524 (5.8) 3,204 (5.9) 4,125 (6.4) 7,853 (6.1) 7,437 (11.4) 25 Psychiatric disorders 1,722 (19.0) 10,593 (19.5) 12,844 (20.0) 25,159 (19.7) 16,598 (25.5) 26 Smoking‡ 1,024 (11.3) 6,481 (11.9) 7,442 (11.6) 14,947 (11.7) 11,434 (17.6) 27 Comedications§ 28 Antiarrhythmics or cardiac glycosides 5,996 (66.0) 35,761 (65.7) 41,031 (63.8) 82,788 (64.8) 35,600 (54.8) 29 Lipid-lowering agents 3,913 (43.1) 22,250 (40.9) 28,812 (44.8) 54,975 (43.0) 31,903 (49.1) 30 Oral corticosteroids 1,105 (12.2) 6,964 (12.8) 8,079 (12.6) 16,148 (12.6) 8,967 (13.8) 31 Antiulcer agents 4,295 (47.3) 24,842 (45.6) 31,469 (48.9) 60,606 (47.4) 39,842 (61.3) 32 Polymedication (≥5 ATC classes) 3,750 (41.3) 21,725 (39.9) 28,196 (43.9) 53,671 (42.0) 45,153 (69.5) 33 Polymedication (≥10 ATC classes) 738 (8.1) 4,653 (8.5) 6,246 (9.7) 11,637 (9.1) 14,947 (23.0) 34 NOAC: Non-vitamin K antagonist oral anticoagulant; VKA: vitamin K antagonist; SD: standard deviation; ATE: arterial thromboembolic events (ischaemic stroke, arterial systemic embolism or transient ischaemic 35 attack); NSAIDs: non-steroidal anti-inflammatory drugs 36 * Unless otherwise stated † Comorbidities were defined using a rolling 1-year period following the initiation of OAC therapy 37 ‡ Smoking or alcohol data: measured by using proxies such as reimbursements for specific therapy or hospitalizations related to smoking or alcohol consumption/diseases 38 § Comorbidities were defined using a rolling 4-month period preceding the initiation of OAC therapy 39 40 41 42 24 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 25 of 36 BMJ Open

1 2 3 4 Table 2. Potential inappropriate use of NOAC therapy in oral anticoagulant-naïve patients with atrial fibrillation in 2015-2016 5 6 NOAC VKA 7 Characteristics (N; %) Dabigatran Rivaroxaban Apixaban Total NOAC N= 65,010 8 N= 9,085 N= 54,456 N= 64,300 N= 127,841 9 Contraindications or non-approved indication/dose 1,457 (16.0) 8,614 (15.8) 9,542 (14.8) 19,613 (15.3) NA 10 Any valvular heart disease 649 (7.1) 4,146 (7.6) 6,122 (9.5) 10,917 (8.5) 16,461 (25.3) 11 Prosthetic heart valve (mechanical or bioprosthetic valves) 106 (1.2) 665 (1.2) 1,096 (1.7) 1,867 (1.5) 6,726 (10.3) Recently hospitalized for coagulopathy, purpura and other 12 90 (1.0) 479 (0.9) 596 (0.9) 1,165 (0.9) 2,142 (3.3) haemorrhagic conditions* For peer review only 13 Liver fibrosis and cirrhosis* 50 (0.6) 282 (0.5) 367 (0.6) 699 (0.5) 1,174 (1.8) 14 Recent gastrointestinal ulceration or intracranial haemorrhage† 40 (0.4) 120 (0.2) 248 (0.4) 408 (0.3) 350 (0.5) 15 Recently or currently treated cancer* 417 (4.6) 2,531 (4.6) 2,898 (4.5) 5,846 (4.6) 4,252 (6.5) 16 Reduced-dose NOAC not approved for stroke prevention in AF 357 (3.9) 1,844 (3.4) NA 2,201 (3.5) NA 17 patients in Europe 18 Inappropriate use during follow-up‡ 19 No monitoring of renal function at initiation 637 (16.8) 3804 (15.5) 3642 (14.5) 8083 (15.1) 5401 (17.0) 20 No monitoring of renal function in the year post-initiation 378 (10.0) 2,347 (9.6) 2,138 (8.5) 4,863 (9.1) 2,984 (9.4) 21 Non-persistence patterns, N (%) 22 One reimbursement only 605 (15.9) 2,666 (10.9) 1,771 (7.1) 5,042 (9.4) 2,426 (7.6) § 23 One-year treatment discontinuation rates 984 (25.9) 6210 (25.4) 4524 (18,0) 11,718 (21,9) 8399 (26.4) Concomitant use of drug increasing the risk of bleeding¶ 2,639 (29.3) 15,797 (29.3) 18,556 (29.2) 36,992 (29.3) 33,025 (52.3) 24 Antiplatelet agents or parenteral anticoagulants 1,728 (19.2) 10,386 (19.3) 12,382 (19.5) 24,496 (19.4) 28,112 (44.5) 25 Parenteral anticoagulants 287 (3.2) 1,577 (2.9) 1,520 (2.4) 3,384 (2.7) 12,078 (19.1) 26 Antiplatelet agents 1,490 (16.6) 9,179 (17.0) 11,170 (17.6) 21,839 (17.3) 19,710 (31.2) 27 Aspirin 1,336 (14.8) 8,215 (15.2) 10,026 (15.8) 19,577 (15.5) 17,770 (28.1) 28 NSAIDs 375 (4.2) 2,111 (3.9) 2,017 (3.2) 4,503 (3.6) 1,030 (1.6) 29 SSRIs and SSNRIs 836 (9.3) 4,852 (9.0) 5,880 (9.2) 11,568 (9.1) 7,317 (11.6) 30 31 NOAC: Non-vitamin K antagonist oral anticoagulant; VKA: vitamin K antagonist; AF: atrial fibrillation; NA: not applicable; NSAIDs: non-steroidal anti-inflammatory drugs; SSRIs: selective serotonin reuptake inhibitors and SSNRIs: selective serotonin-norepinephrine reuptake inhibitors 32 * Comorbidities identified using hospitalization and/or LTD data, and/or specific procedures during a rolling 1-year period preceding the initiation of OAC therapy 33 † Comorbidities identified using hospitalization data during a rolling 6-week period preceding the initiation of OAC therapy 34 ‡ Data on patients with at least a one year of follow-up i.e. patients initiating OAC in 2015 after excluding patients who died and those hospitalized for 3 months or longer (N= 3,796; 24,483; 25,118; 53,397 and 35 31,777 for dabigatran-, rivaroxaban, apixaban, total NOAC- and total VKA new users, respectively); period considered (unless otherwise stated): rolling 1-year period following the initiation of OAC therapy (index date included) 36 § One-year crude discontinuation rate for patients initiating OAC in 2015 who died and those hospitalized for 3 months or longer, defined as prolonged treatment discontinuation i.e. 90-day gap with no medication 37 coverage after the 30-day coverage period of a refill 38 ¶ Data for new users still alive after a 45-day period following the index date (N= 9,001; 53,885; 63,578; 126,464 and 63,180 for dabigatran-, rivaroxaban, apixaban, total NOAC- and total VKA new users , 39 respectively); period considered: rolling 6-week period following the initiation of OAC therapy (index date included) 40 41 42 25 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 26 of 36

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 30 of 36

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 31 of 36 BMJ Open

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 32 of 36

30 General practitioner 1.37 1.29 1.46 *** 1.16 1.10 1.23 *** 1.00 0.96 1.04

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 33 of 36 BMJ Open

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 34 of 36

43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 35 of 36 BMJ Open