Stroke prevention in atrial fibrillation: Past, present and future

Transcription

1 th Anniversary Stroke prevention in atrial fibrillation: Past, present and future Comparing the guidelines and practical decision-making Gregory Y. H. Lip 1,2,3 ; Ben Freedman 4,5 ; Raffaele De Caterina 6 ; Tatjana S. Potpara 3,7 1 Institute of Cardiovascular Science, University of Birmingham, Birmingham, UK; 2 Aalborg Thrombosis Research Unit, Department of Clinical Medicine, Aalborg University, Aalborg, Denmark; 3 School of Medicine, Belgrade University, Belgrade, Serbia; 4 Heart Research Institute, Charles Perkins Centre, Sydney Medical School, University of Sydney, Sydney, Australia; 5 Department of Cardiology and Anzac Research Institute, Concord Hospital, University of Sydney, Australia; 6 G. d Annunzio University and Center of Excellence on Aging CeSI-Met, Chieti, Italy; 7 Cardiology Clinic, Clinical Centre of Serbia, Belgrade, Serbia Summary Concepts and our approaches to stroke prevention in atrial fibrillation (AF) have changed markedly over the last decade. There has been an evolution over the approach to stroke and bleeding risk assessment, as well as new treatment options. An increasing awareness of AF has led to calls to improve the detection of and population screening for AF. Stroke and bleeding risk assessment continues to evolve, and the ongoing debate on balance between simplicity and practicality, against precision medicine will continue. In this review article, we provide an overview of past, present and the (likely) future concepts and approaches to stroke prevention in AF. We propose three simple steps (the Birmingham 3-step ) that offers a practical management pathway to help streamline and simplify decision-making for stroke prevention in patients with AF. Keywords Atrial fibrillation, stroke prevention, non-vitamin K antagonist oral anticoagulants, warfarin Correspondence to: Prof. G. Y. H. Lip University of Birmingham Institute of Cardiovascular Sciences City Hospital, Birmingham B18 7QH, United Kingdom Tel.: , Fax: g.y.h.lip@bham.ac.uk Received: November 22, 2016 Accepted after minor revision: January 2, 2017 Epub ahead of print: June 9, Thromb Haemost 2017; 117: Note: The review process for this paper was fully handled by Christian Weber, Editor in Chief. Concepts and our approaches to stroke prevention in atrial fibrillation (AF) have changed markedly over the last decade. There has been an evolution over the approach to stroke and bleeding risk assessment, as well as new treatment options. An increasing awareness of AF has led to calls to improve the detection of and population screening for AF (1, 2). As the public health impact of AF increases, there also is an urgent need to simplify and streamline patient management pathways, with attention to improving patient adherence and persistence with stroke prevention therapies (3), as this will optimise outcomes and reduce the burden of stroke associated with AF (4). The past The historical randomised controlled trials (RCTs) comparing warfarin to placebo or control, conducted two decades ago, clearly showed that warfarin reduced stroke/systemic embolism by 64 % and all-cause mortality by 26 % (5, 6). Of note, these trials included less than 8000 patients in comparison with either nothing or antiplatelet agents mostly aspirin (5, 6). Clearly, warfarin (and the Vitamin K antagonists, VKA) became the standard of care when stroke prevention with OAC was needed. Nonetheless, the use of VKA was problematic, requiring regular anticoagulation monitoring to keep within a relatively narrow therapeutic range, and furthermore, VKAs were subject to much inter- and intra-patient variability, as well as diet and drug interactions (7). Furthermore, oral anticoagulation (OAC) with the VKAs was associated with increased bleeding risk, the most serious being intracranial bleeding (ICH). Hence, initial stroke prevention efforts were particularly directed towards identifying the high risk AF patients in whom stroke reduction benefits of OAC outweighed the bleeding risk. Clinical characteristics of patients in the non-vka treated or placebo arms of the historical RCTs were used to help identify clinical and echocardiographic risk factors for an increased risk of stroke, and the clinical features were used to formulate stroke risk stratification schemes, such as the CHADS 2 score (8). Whilst the latter was simple, it had only modest predictive value, as reflected by a c-index of approximately 0.60 (a statistical measure of the predictive value, where a c-index of 1.0 is perfect prediction, and 0.5 is 50:50 prediction) (9), similar to most risk scores based on clinical features. Thrombosis and Haemostasis 7/2017 Schattauer 2017

2 Lip et al. Stroke prevention in AF 1231 Attention was therefore directed towards echocardiographic parameters to refine stroke risk assessment, and the presence of moderate-severe left ventricular systolic impairment on two-dimensional echocardiography was independently predictive of stroke in AF (10). The availability of transoesphageal echocardiography allowed additional refinement, where the presence of spontaneous echo contrast, low left atrial appendage velocities and complex aortic plaque were independent predictors of thromboembolism (11), as was the presence of left atrial thrombus (12). The latter could be present asymptomatically in approximately 10 % of patients, even whilst on OAC (12). Nonetheless, the historical RCTs have been debated as they only randomised <10 % of patients screened, and many common risk factors associated with thromboembolism in AF were neither systematically recorded nor consistently defined. A systematic review by the Stroke in AF Working Group identified previous stroke (relative risk [RR] 2.5, 95 % confidence interval [CI] ), increasing age (RR 1.5 per decade, 95 %CI ) diabetes (1.7, ), and history of hypertension (2.0, ) as risk factors, whilst female sex was inconsistently associated with stroke but heart failure and coronary artery disease were inconclusive as risk factors (13). Other systematic reviews by the National institute for Health and Care Excellence (NICE) and Pisters et al. found evidence for structural heart disease (including prior myocardial infarction or cardiac dysfunction) and female sex as risk factors (14, 15). Nonetheless, additional information on stroke risk factors has been obtained from large observational cohorts. The present Stroke risk stratification and treatment decisions Currently, the CHA 2 DS 2 -VASc score (16) is commonly advocated in most guidelines (see Table 1), and extends the older CHADS 2 score by recognising additional risk factors. For example, age is a powerful driver of stroke risk, so that age gets one point, and age 75 gets 2 points; in Asian cohorts, where stroke risk seems higher than in non-asian cohorts, stroke risk seems to rise from age of 50 upwards, such that a recalibration of CHA 2 DS 2 -VASc has been proposed with 1 point for age 50 74, with a significant improvement in stroke prediction over the conventional CHA 2 DS 2 -VASc score (17). The V in CHA 2 DS 2 -VASc refers to complicated vascular disease, usually defined as myocardial infarction or revascularisation, peripheral artery disease or the presence of complex aortic plaque (18). The Sc (Sex category) criterion refers to female sex, which older literature suggested increased stroke risk (19), but it became apparent there was an agedependency to this risk, such that females aged <65 in the absence of any additional stroke risk factors were at low risk (with stroke rates of <1 %/year) (20). Also, the C criterion in CHA 2 DS 2 -VASc is broader than that in CHADS 2, and also refers to recent decompensated heart failure (usually indicated by a hospitalisation) irrespective of the left ventricular ejection fraction (LVEF) (21), thus including heart failure with reduced ejection fraction (HFrEF) or preserved ejection fraction (HFpEF), cardiomyopathy, etc., as well as asymptomatic or symptomatic moderate-severe LV systolic impairment on cardiac imaging. Even a single stroke risk factor confers an accentuated risk of stroke and mortality, and the net clinical benefit (NCB) balancing stroke reduction against serious bleeding with OAC use is clearly positive in such patients, compared to aspirin or no treatment (22 24). In contrast, aspirin has a neutral or negative NCB, indicating no advantage even in patients with a CHA 2 DS 2 -VASc score of 1 (or 2 in females) (25). Whilst CHA 2 DS 2 -VASc includes the common stroke risk factors seen in everyday clinical practice, it does not include every possible stroke risk factor especially since some have less supportive evidence as an independent predictor. For example, AF patients with chronic kidney disease (CKD) are at high risk for stroke, death, myocardial infraction and bleeding but CKD does not appear to independently improve the predictive value of CHA 2 DS 2 -VASc, which is perhaps unsurprising since CKD is associated with the components of CHA 2 DS 2 -VASc such as age, heart failure, diabetes mellitus, vascular disease etc (26). As would be expected, the CHA 2 DS 2 -VASc score also predicts stroke and mortality in non-af populations, given it is essentially a cluster of the common risk factors for stroke and mortality (27, 28). The modest predictive value of CHA 2 DS 2 -VASc for identifying high risk patients can clearly be improved (at least statistically) by including substantially more clinical variables (QStroke, GAR- FIELD score), or the addition of biomarkers (29-31). With complicated clinical risk scores and multiple parameters, some refinement is possible, with a c-index statistically improved to approximately 0.7, especially with machine learning methodology of large cohorts. Biomarkers can clearly improve stroke prediction (32), and addition of any biomarker (whether blood, urine or imaging based) will always improve on clinical risk prediction, at least statistically and changing the c-index to This concept is not new, and even a decade ago (33), biomarkers were already shown to refine stroke risk stratification. The same biomarkers are also predictors of stroke, death, myocardial infarction and bleeding, and to weigh or balance these outcomes can only lead to treatment uncertainties and confusion in decision-making. Importantly, the predictive value of biomarkers should ideally be shown in nonanticoagulated general population cohorts, which is difficult, while recent studies proposing biomarkers have been conducted in highly selected anticoagulated RCT cohorts (31, 34). Also, complicated scores and multiple biomarkers and/or imaging simply add additional expense and reduce simplicity or practicality of widespread use in everyday clinical practice and busy clinics, especially since the default should be to offer stroke prevention (which is OAC) unless the patient is shown to be low risk. Moreover, most of the increase in risk prediction in these studies is by improving prediction in mid to high risk range, where all will require anticoagulation regardless of actual risk. The real clinical unmet need is to improve separation of the low risk subset who do not need anticoagulation from the intermediate risk group who would benefit, not provided by addition of current bio - Schattauer 2017 Thrombosis and Haemostasis 7/2017

3 1232 Lip et al. Stroke prevention in AF markers, and may be unethical to test prospectively in non-anticoagulated patients. The 2016 European Society of Cardiology (ESC) AF guidelines (35) use the CHA 2 DS 2 -VASc score in a categorised approach, Table 1: A comparison of the most recent ( ) guidelines published in English on antithrombotic prophylaxis in atrial fibrillation. Reference Scoring system recommended Recommendation for anticoagulation in general Recommendation for NOACs Recommendation for intermediaterisk patients Role of aspirin monotherapy JCS 2013 Circ J 2014; 78: (72) CHA 2 DS 2 Anticoagulation therapy based on risk assessment for cerebral infarction and bleeding is recommended. Class I LOE A NOACs should be considered, whenever indicated, as the first line therapy in patients with a CHADS 2 score of 2. Class I LOE A Anticoagulation therapy with rivaroxaban, edoxaban or warfarin should be considered for intermediate-risk patients with a CHA 2 DS 2 score of 1. LOE B Antiplatelet drugs may be considered for patients who cannot use oral anticoagulants. Class IIb LOE C NICE 2014 nice.org.uk/guidance/cg180 (73) CHA 2 DS 2 VASc For most people the benefit of anticoagulation outweighs the bleeding risk No specific preference for NOACs (apixaban, dabigatran etexilate, rivaroxaban) over warfarin in patients in whom anticoagulation is indicated Consider anticoagulation for men with a CHA 2 DS 2 VASc score of 1. Take the bleeding risk into account. Do not offer aspirin mono - therapy solely for stroke prevention to people with AF ACC/AHA 2014 J Am Coll Cardiol 2014; 64: e1 76 (36) CHA 2 DS 2 VASc With prior stroke, TIA, or CHA 2 DS 2 VASc score 2, Class I, LOE A Warfarin (Class I, LOE A) or a NOAC (dabigatran, rivaroxaban, or apixaban). Class I, LOE B With nonvalvular AF and a CHA 2 DS 2 VASc score of 1, no antithrombotic therapy or treatment with oral anticoagulant or aspirin may be considered. Class Iib LOE C With nonvalvular AF and a CHA 2 DS 2 VASc score of 1, no antithrombotic therapy or treatment with oral anticoagulant or aspirin may be considered. Class IIb LOE C where low risk is defined as a CHA 2 DS 2 -VASc score 0 (or 1 in females), moderate risk is a CHA 2 DS 2 -VASc score 1 in males (or 2 in females), and high risk is 2 stroke risk factors based on CHA 2 DS 2 -VASc; in these guidelines, high risk patients are recom- CCS 2016 Can J Cardiol 2016; 32: (74) CHADS-65 In general, oral anticoagulation (OAC) therapy is recommended for all patients with AF except those younger than 65 years of age with a Congestive Heart Failure, Hypertension, Age, Diabetes, Stroke/TIA (CHADS 2 ) score of 0. When OAC is indicated, a NOAC is recommended in preference to a VKA for nonvalvular AF (NVAF) (also with concomitant CAD) In CHADS 2 0 patients "aspirin (acetylsalicylic acid; ASA) is recommended in patients with CAD/vascular disease, and there is no indication for antithrombotic treatment in the absence of CAD/vascular disease". In CHADS 2 0 patients "aspirin (acetylsalicylic acid; ASA) is recommended in patients with CAD/vascular disease, and there is no indication for antithrombotic treatment in the absence of CAD/vascular disease". ESC 2016 Eur Heart J 2016; PMID: DOI: /eurheartj/ehw210 (35) CHA 2 DS 2 VASc Oral anticoagulation therapy to prevent thromboembolism is recommended for all male AF patients with a CHA 2 DS 2 -VASc score of 2 or more. Class I LOE A Oral anticoagulation therapy to prevent thromboembolism is recommended in all female AF patients with a CHA 2 DS 2 -VASc score of 3 or more. Class I LOE A When oral anticoagulation is initiated in a patient with AF who is eligible for a NOAC (apixaban, dabigatran, edoxaban, or rivaroxaban), a NOAC is recommended in preference to a Vitamin K antagonist. Class I LOE A Oral anticoagulation therapy to prevent thromboembolism should be considered in male AF patients with a CHA 2 DS 2 VASc score of 1, considering individual characteristics and patient preferences. Class IIa LOE B Oral anticoagulation therapy to prevent hromboembolism should be considered in female AF patients with a CHA 2 DS 2 VASc score of 2, considering individual characteristics and patient preferences. Class IIa LOE B. Antiplatelet monotherapy is not recommended for stroke prevention in AF patients, regardless of stroke risk. Class III LOE B Thrombosis and Haemostasis 7/2017 Schattauer 2017

4 Lip et al. Stroke prevention in AF 1233 Table 1: Continued Combination therapy JCS 2013 A combination of antiplatelet and anticoagulant drugs may be considered for patients with CAD who undergo percutaneous coronary intervention (PCI) or surgical revascularisation. Iib LOE C NICE 2014 No specific recommendations given ACC/AHA 2014 After coronary re - vascularisation in patients with CHA 2 DS 2 VASc score 2, it may be reasonable to use clopidogrel concurrently with oral anticoagulants but without aspirin. Class IIb LOE B CCS 2016 It is recommended that patients who refuse warranted OAC therapy should receive the combination of ASA and clopidogrel. When CAD is present, some expert clinicians prefer a combination of a NOAC and ASA rather than a NOAC alone in preference to warfarin alone for patients perceived to be at higher risk of coronary events and low risk of major bleeding and might choose a NOAC alone as a reasonable option in those with average to lower risk of coronary events and higher risk of bleeding. ESC 2016 Combinations of oral anticoagulants and platelet inhibitors increase bleeding risk and should be avoided in AF patients without another indication for platelet inhibition. Class III LOE B Abbreviations: ACC/AHA = American College of Cardiology/American Heart Association; CAD = coronary artery disease; CCS = Canadian Cardiovascular Society; ESC = European Society of Cardiology; JCS = Japanese Cardiovascular Society; LOE = Level of Evidence; NICE = National Institute for Health and Care Excellence (UK); NOAC = non-vitamin K antagonist oral anticoagulant; OAC = oral anticoagulant; TIA = transient ischaemic attack. mended OAC (Class I recommendation), moderate risk patients should be considered for OAC (Class IIa recommendation) and low risk patients are recommended no antithrombotic therapy. The ESC guidelines do not recommend the use of aspirin solely for stroke prevention in AF. The 2014 AHA/ACC/ HRS (American Heart Association/American College of Cardiology/Heart Rhythm Society) guidelines (36) categorise risk as 0, 1 or 2 where those with score 0 are recommended no antithrombotic therapy, whilst those with score 2 are recommended OAC; however, those with score =1 are recommended nothing, aspirin or OAC depending on patient values and preferences and individualisation of therapy. Other guidelines are discussed further below. The continued mention of aspirin in guidelines may be one reason for its use as an option by some prescribers (37). To simplify and improve clinical management, we should perhaps start by avoiding the obsession with trying to identity a particular event rate for a particular point on the CHA 2 DS 2 -VASc score. Unsurprisingly, there would be a wide variation in reported event rates for particular scores points in different cohorts (38, 39), as stroke rates are dependent on study population, clinical setting, ethnicity, etc. Indeed, hospitalised cohorts have higher rates than community based AF population cohorts, but AF patients have high rates of hospitalisation whilst individual AF patients may be in the community and stable today, they may well be hospitalised next week, and even non-cardiovascular hospitalisations (e. g. from sepsis) increase the risk markedly. The electrophysiological characteristics of AF (AF pattern, atrial flutter, AF with a reversible cause, AF efficiently treated with drugs or ablation, brady-tachy syndrome, etc) are not relevant predictors for a higher or a lower risk of stroke, and are not sufficiently robust to decide on antithrombotic strategies. Also, with a single stroke risk factor, not all risk factors carry equal weight (40, 41). Of the CHA 2 DS 2 -VASc score factors, 1 point from age is a powerful driver of stroke risk but even within this category, the age of 65 is clearly lower risk than the age of 74 years. Also, well-controlled hypertension clearly has lower stoke risk compared to uncontrolled blood pressure, but what may be controlled hypertension today is not necessarily the case if patients do not take their antihypertensive medication. Notwithstanding the fact that even one stroke risk factor confers an excess risk of stroke and death, and that OAC has a positive NCB with one or more stroke risk factors (25), the CHA 2 DS 2 -VASc score should be used in a stepwise approach to initially identify low risk patients (Step 1), defined as a CHA 2 DS 2 -VASc score of 0 in males or 1 in females; such patients do not need any antithrombotic therapy. Step 2 is then to offer stroke prevention (i. e. OAC) to those with 1 stroke risk factors at this step it does not matter if the CHA 2 DS 2 -VASc score is 2, 3 or higher, or whether we have 1 or 10 biomarkers added, the patient merits stroke prevention to reduce the risk of a potentially fatal or severely disabling AF-related stroke. At Step 2, the decision with OAC these days is either a VKA with good quality anticoagulation control (as reflected by a good time in therapeutic range [TTR] >65 70 %), or a non-vka oral anticoagulant (NOAC), now tested against warfarin in over 80,000 patients. Despite the introduction of the NOACs, the VKAs are still widely used worldwide. However if VKAs are used, the challenge is how to identify those patients who are going to achieve a good TTR, as well as persistence with treatment, given that treatment discontinuations leads to more adverse outcomes (4). Many common clinical and genetic factors have been associated with labile INRs (42). The more common clinical factors have been used to formulate the SAMe-TT 2 R 2 score (43), which is a simple score that has been validated to identify those patients likely to do well on VKA with a good TTR (score 0 2) or those less likely to achieve a good TTR (score >2) that can be flagged up for more regular review with INR checks or education/counselling Schattauer 2017 Thrombosis and Haemostasis 7/2017

5 1234 Lip et al. Stroke prevention in AF (that has been shown to improve TTR) or these days, to be started on NOAC instead. The SAMe-TT 2 R 2 score has now been validated in multiple independent cohorts (44, 45), and not only helps to identify those patients less likely to do well on VKA but is predictive of labile INRs and the adverse outcomes associated with labile INRs, e. g. thromboembolism, death and bleeding. The SAMe- TT 2 R 2 score can therefore be incorporated into the patient pathway (Step 3). These three simple steps (the Birmingham 3-step ) offers a management pathway to help streamline decision-making for stroke prevention in patients with AF, and is illustrated in Figure 1. The NOACs have changed how we approach stroke prevention in AF, as they offer relative efficacy, safety and convenience compared to VKAs, in patients with AF, with an overall clear net clinical benefit (46, 47). The efficacy and safety of NOACs is even greater in Asians compared to non-asians. With specific antidotes or reversal agents already licensed and available for dabigatran (specifically, idarucizumab [48, 49]), or in development for the Factor Xa inhibitors (eg. andexanet alpha [50]), we are also in the era where we can essentially switch off any anticoagulation effect with the NOACs should the need arise, for example, following severe bleeding or the need for urgent surgery in a patient whilst on a NOAC (51, 52). The large Phase 3 NOAC clinical trials have been complemented by large post-marketing observational real world cohorts that augment and support the RCT data, showing effectiveness and safety with the NOACs versus warfarin these perhaps contrast the broad use of the new drugs in the real world against the artificial setting of a highly controlled protocol-based RCT cohort (53). The volume of these papers has largely reflected the sequence the individual NOACs were approved and marketed (54 57). Concerns about some adverse outcomes, such as myocardial infarction with some of the NOACs also seem unfounded (58). In the Danish registries, Larsen et al. (59) reported that all NOACs seemed to be safe and effective alternatives to warfarin in a routine care setting, and no significant differences were found between usual dose NOACs (dabigatran 150 mg bid, rivaroxaban 20 mg and apixaban 5 mg bid) and warfarin for ischaemic stroke. The risks of death, any bleeding, or major bleeding were significantly lower for apixaban and dabigatran compared with warfarin. Comparative effectiveness or safety real world cohorts are also been published, and subtle differences are emerging and providing differentiation between the different NOAC agents (60). For example, the independent FDA Medicare analysis by Graham et al. (61) reported that treatment with rivaroxaban 20 mg once daily was associated with similar risks of thromboembolic stroke but statistically significant increases in ICH and major extracranial bleeding, including major gastrointestinal bleeding, compared with dabigatran 150 mg twice daily. Thus, apart from VKA (with good TTR), we have four NOACs licensed for AF stroke prevention compared to five years ago when prescribers did not have a choice (VKA only, nothing else), we are now spoilt for choice, and can fit the NOAC drug to particular patient characteristics (62, 63). After all, AF patients are not homogeneous, and to assume one drug fits all is simply unrealistic and misleading. Bleeding risk assessment The other side of the equation of thromboprophylaxis is bleeding risk assessment, which has been subject to many misconceptions and misuse (64). Unlike stroke risk factors, many bleeding risk factors can be modified and bleeding risk assessment is a dynamic process that needs evaluation at regular review and follow-up. Figure 1: The Birmingham 3-step. Management pathway to streamline decision-making for stroke prevention in patients with atrial fibrillation. Thrombosis and Haemostasis 7/2017 Schattauer 2017

6 Lip et al. Stroke prevention in AF 1235 Bleeding risk assessment is to flag up the patients potentially at high risk of bleeding for more careful follow-up and most importantly, to address the reversible bleeding risk factors. A high bleeding risk score is not an excuse to withhold OAC, as the NCB in such patients is often greater given that the stroke and bleeding risk track each other and the magnitude of gain from stroke reduction will far outweigh the small risk of serious bleeding. A history of falls is not a contraindication to OAC, and in a modelling analysis, the anticoagulated patient would need to fall 295 times per year, for the benefit of stroke reduction with OAC to be outweighed by the potential harm of serious bleeding (65). This is perhaps a simplification, as clearly a traumatic fall or uncontrolled epileptic seizures confers greater risk than less dangerous reasons for falls (e.g. syncope). Many bleeding risk factors have been identified, and the more common ones have been used to formulate bleeding risk scores, of which the HEMORR2HAGES, HAS-BLED, ATRIA, ORBIT and ABC-Bleeding scores have been derived for use in the AF population. The HAS-BLED score (66) has been popular, since it draws attention to the common reversible bleeding risk factors (uncontrolled blood pressure, labile INRs if on warfarin, excessive alcohol intake or concomitant use of aspirin or NSAIDs in an anticoagulated patient). The HAS-BLED has been validated in patients on no antithrombotic therapy, aspirin or anticoagulation (whether VKA or non-vka anticoagulants), and is therefore applicable at all stages in the patient management pathway. HAS-BLED has also been validated in AF and non-af populations, and is predictive of bleeding during bridging, acute coronary syndrome or percutaneous coronary interventions. Unlike other scores, HAS-BLED is also predictive of intracranial haemorrhage. Last, the simple HAS-BLED score performs as well as, and often superior, to other bleeding risk scores (67). Indeed, newer simple bleeding risk scores would significantly underperform in predicting bleeding whilst on VKA, as they do not consider labile INR or TTR as a criterion, and TTR is a powerful determinant of bleeding risks (67, 68). Figure 2: Thromboprophylaxis in AF a summary of the principal steps recommended by selected international guidelines. VKA: Vitamin K Antagonists; OAC: Oral Anticoagulant; CAD: Coronary Artery Disease; ASA: Acetyl Salicylic Acid; NOAC: Non-VKA Oral Anticoagulants; TTR: Time in Therapeutic Range; SAMe-TT 2 R 2 : female sex, age of <60 years, history of two or more co-morbidities (hypertension, diabetes mellitus, coronary artery disease/myocardial infarction, peripheral arterial disease, congestive heart failure, previous stroke, pulmonary disease, and hepatic or renal disease) or treatment with drugs interacting with VKAs (e. g. amiodarone) 1 point each, tobacco use and non-caucasian ethnicity 2 points each. Schattauer 2017 Thrombosis and Haemostasis 7/2017

7 1236 Lip et al. Stroke prevention in AF Adherence, CVD risk management, and multi - disciplinary clinics One of the major issues with current anticoagulant treatment is that it is not always offered to patients at increased stroke risk according to guideline recommendation, often because of an overestimate of falls or bleeding risk, and an underestimate of risk from paroxysmal AF (69). Once prescribed, reduced adherence to anticoagulant therapy results in subtherapeutic levels, and discontinuation of anticoagulant: even after just one year following initiation, only about 65 % prescribed VKA, and 80 % for NOACs are still taking the drugs (4). Multidisciplinary nurse-led clinics have been proposed as a way to increase guideline and medication adherence, and improve outcomes, including addressing cardiovascular risk factors. The results appear favourable and cost effective (70, 71). What do the guidelines say? Table 1 summarises the principal international guidelines from the Japanese Cardiology Society (JCS) (72), National Institute for Health and Care Excellence (NICE) (73), AHA/ACC/ HRS (36), Canadian Cardiology Society (CCS) (74) and ESC (35). The principal decision steps have been summarised into Figure 2, where the Birmingham 3-step pathway is also shown to put decision making into context. While the ESC, AHA/ACC/HRS and NICE guidelines recommend the CHA 2 DS 2 -VASc score as the stroke assessment tool, the JCS guidelines opt for the use of CHADS 2 score with consideration of age 65 74, vascular disease and cardiomyopathy (i.e. other non- CHADS 2 components of the CHA 2 DS 2 -VASc score) and the CCS guidelines recommend a modified CHADS65 score ( Table 1). Both JCS and CCS well-acknowledge the role of additional stroke risk factors (i. e. age years, prior myocardial infarction, complex aortic plaque and peripheral arterial disease) which are included in the CHA 2 DS 2 -VASc score (excluding only female sex). As mentioned above, female sex-related risk of stroke is increased usually with the presence of 1 additional CHA 2 DS 2 -VASc stroke risk factors (thus, a female AF patient with no additional stroke risk factors would have a CHA 2 DS 2 -VASc score of 1 and should be classified into the truly low stroke risk stratum). The use of OAC in high-risk patients (that is, those with 2 stroke risk factors) and no therapy in AF patients without additional stroke risk factors is consistently recommended by all guidelines, whilst the recommendations for those at intermediate stroke risk (i.e. single stroke risk factor) are heterogeneous and include no therapy, aspirin or (preferably) OAC ( Table 1). When recommend OAC, the ESC, CCS and JCS guidelines give a preference to NOACs, whilst the NICE and AHA/ACC/HRS guidelines do not specify any preference to NOACs or VKAs. Also, the JCS guidelines specify that the target INR range with VKAs in patients aged 70 years should be (instead of in all other AF patients or in all other AF guidelines). Importantly, all guidelines make a distinction between so-called valvular versus non-valvular AF and recommend the use of NOAC only in patients with non-valvular AF, based on the results of the Phase 2 RE-ALIGN trial (75) which showed higher rates of stroke, valve thrombosis and major bleeding with dabigatran compared to warfarin. Valvular AF, which is currently not a satisfying terminology, has been defined in recent years as AF in patients with mechanical prosthetic heart valves or moderate-to severe mitral stenosis. All such patients should be given a VKA, whilst AF patients with any other heart valve abnormality may be given either NOACs or VKAs (76, 77). Stroke risk in such patients can also be assessed using the CHA 2 DS 2 -VASc score (78). All guidelines highlight the importance of bleeding risk management by addressing the modifiable bleeding risk factors and scheduling AF patients at increased risk of bleeding for a more intense clinical follow-up during OAC treatment. The NICE and JCS guidelines explicitly recommend the HAS-BLED score for bleeding risk assessment, whilst the AHA/ACC/HRS guidelines mention several bleeding risk scores without any formal recommendation, but with the notion that the HAS-BLED score better discriminates bleeding risk in comparison to the HEMORR2HAGES or ATRIA score. As discussed above, the ESC guidelines provide a long list of modifiable (uncontrolled hypertension, labile INR or TTR<60 % with VKAs, concomitant aspirin or non-steroidal antiinflammatory drug use, excessive alcohol use), potentially modifiable (anaemia, impaired renal or liver function, reduced platelet count or function) most of which are included in the HAS-BLED score and non-modifiable (age >65 years, previous bleeding, prior stroke, renal replacement therapy, liver cirrhosis, malignancy, genetic factors) or biomarker-based bleeding risk factors (e.g. high-sensitivity troponin). Whilst mentioning several bleeding risk scores (including the HAS-BLED, ORBIT and ABC scores), the ESC guidelines also do not provide any specific recommendation on the bleeding risk assessment tool or the definition of high-bleeding risk category, perhaps because a high-bleeding risk would not change the decision to initiate OAC, although this may be relevant with combination OAC and antiplatelet drugs in the setting of acute coronary syndrome or percutaneous coronary interventions. All guidelines provide guidance on the optimal management of special AF populations (e.g. elderly, patients with chronic kidney disease, etc.) or those presenting with an acute coronary syndrome, major bleeding or undergoing surgery or invasive procedures (79, 80). The ESC guidelines particularly extend on the management of acute stroke, major bleeding and the use of OAC post ICH, emphasising the role of integrated approach to AF management. Patient education and their values and preferences with respect to the management of stroke and bleeding risk are increasingly acknowledged in the formal AF guidelines. Finally, the ESC, AHA/ACC/HRS and NICE guidelines emphasise that documentation of AF using an electrocardiographic (ECG) or monitor/device recording is necessary for the diagnosis of AF, which is required before initiating stroke prevention therapy. Along with recommending opportunistic screening for AF by manual pulse checking (and subsequent ECG recording in those Thrombosis and Haemostasis 7/2017 Schattauer 2017

8 Lip et al. Stroke prevention in AF 1237 with palpated pulse irregularity) in all individuals aged >65 years (the ESC) or in individuals with symptoms suggestive of AF (the NICE), the ESC, NICE and AHA/ACC/HRS guidelines recommend prolonged screening for AF in post-stroke/tia patients, using short-term ECG recording followed by continuous ECGmonitoring for at least 72 hours. All three guidelines mention that long-term monitoring using non-invasive ECG monitors or implanted loop recorders should be considered for the detection of asymptomatic AF in post-stroke patients or to document AF in those with symptoms suggestive of paroxysmal AF. The ESC guidelines further emphasise the significance of detecting asymptomatic (or silent) AF by suggesting that systematic ECG screening may be considered in all individuals aged >75 years or at high risk of stroke (81), and added an ECG rhythm strip to pulse palpation as a recommended way to undertake opportunistic screening. The ESC guidelines also recommended regular device interrogation in patients implanted with pacemakers or implantable cardioverter defibrillators (ICD) for AF or atrial high rate episodes (AHRE). At present, the detection of AHRE should trigger further ECG monitoring to document AF before initiating stroke prevention therapy. Stroke and bleeding risk assessment continues to evolve, and the ongoing debate on balance between simplicity and practicality, against precision medicine will continue. As the NOACs are neither new nor novel anymore, new oral anticoagulants drugs targeting Factor IX, XI and other novel mechanisms are in development (90) as the quest for the holy grail of stroke prevention with no excess of bleeding continues. Conflicts of interest GYHL: Consultant for Bayer/Janssen, BMS/Pfizer, Biotronik, Medtronic, Boehringer-Ingelheim, Microlife and Daiichi-Sankyo. Speaker for Bayer, BMS/Pfizer, Medtronic, Boehringer Ingelheim, Microlife, Roche and Daiichi-Sankyo. No fees are received personally. BF: Grants, personal fees, non-commercial support from Bayer; Personal fees from BMS/Pfizer, Boehringer-Ingelheim, Servier, AstraZeneca, Gilead (not in the area of atrial fibrillation or stroke). TSP: Consultant and speaker fees from Bayer, Pfizer, and Boehringer Ingelheim. RDC: Lecture fees and honoraria from Boehringer-Ingelheim, Bayer, BMS/Pfizer, Daiichi-Sankyo, Lilly, and Novartis; Research grants from Boehringer-Ingelheim, Bayer, BMS/Pfizer. The future The last 10 years has been an exciting time in the field of stroke prevention in AF, and in a modelling analysis, our data point to a substantial increase in the human and economic cost burden of AF and so emphasize the need to reduce this burden (82). This may be achieved by the increased use of OACs, particularly with the NOACs. Whilst we focus on stroke prevention in AF, many deaths in AF are actually not stroke related, and the majority are cardiovascular (83). OAC use is independently associated with a reduction in all cause, CV and non-cv mortality. Nonetheless, OAC use in patients with AF significantly reduces stroke to non-af rates, but whilst mortality is reduced, there is still an excess compared to non-af patients (84). Unanswered questions remain, for example, the use of OACs in AF patients with a prior ICH (85), significant valvular heart disease (78), the very elderly (86) and those with severe kidney disease (87), given that such patients were excluded from RCTs. Also, when to re-initiate OAC in high risk patients after a recent acute ischaemic stroke or gastrointestinal bleed (88, 89). In modelling the effect of the increasing use of NOACs in preference to warfarin in a European population with atrial fibrillation (AF), we recently estimated that the introduction of NOACs in 2010 eliminated over 88,000 thromboembolisms and deaths annually, of which over 17,000 were ischaemic strokes (82). At a oneyear cost of Euro 30k per ischaemic stroke, this strategy saved Euro 510 million annually. At a conservative rate of increase in the AF population of 2.2-fold from 2005, in 2050 there will be around 180,000 AF-related ischaemic strokes that, at an inflation-adjusted cost of around Euro 62k per stroke, sums to Euro 11,116 million. References 1. Kirchhof P, et al. A roadmap to improve the quality of atrial fibrillation management: proceedings from the fifth Atrial Fibrillation Network/European Heart Rhythm Association consensus conference. Europace 2016; 18: Freedman SB, Lowres N. Asymptomatic Atrial Fibrillation: The Case for Screening to Prevent Stroke. JAMA : the journal of the American Medical Association 2015; 314: Raparelli V, et al. Adherence to oral anticoagulant therapy in patients with atrial fibrillation. Focus on non-vitamin K antagonist oral anticoagulants. Thromb Haemost 2016; Martinez C, et al. Therapy persistence in newly diagnosed non-valvular atrial fibrillation treated with warfarin or NOAC. A cohort study. Thromb Haemost 2016; 115: Hart RG, et al. Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med 2007; 146: Hart RG, et al. Adjusted-dose warfarin versus aspirin for preventing stroke in patients with atrial fibrillation. Ann Intern Med 2007; 147: De Caterina R, et al. Vitamin K antagonists in heart disease: current status and perspectives (Section III). Position paper of the ESC Working Group on Thrombosis--Task Force on Anticoagulants in Heart Disease. Thromb Haemost 2013; 110: Keogh C, et al. Validation of the CHADS2 clinical prediction rule to predict ischaemic stroke. A systematic review and meta-analysis. Thromb Haemost 2011; 106: Karthikeyan G, Eikelboom JW. The CHADS2 score for stroke risk stratification in atrial fibrillation--friend or foe? Thromb Haemost 2010; 104: Echocardiographic predictors of stroke in patients with atrial fibrillation: a prospective study of 1066 patients from 3 clinical trials. Arch Intern Med 1998; 158: Donal E, et al. EACVI/EHRA Expert Consensus Document on the role of multimodality imaging for the evaluation of patients with atrial fibrillation. Eur Heart joj Cardiovasc Imag 2016; 17: Di Minno MN, et al. Prevalence of left atrial thrombus in patients with non-valvular atrial fibrillation. A systematic review and meta-analysis of the literature. Thromb Haemost 2016; 115: Stroke Risk in Atrial Fibrillation Working G. Independent predictors of stroke in patients with atrial fibrillation: a systematic review. Neurology 2007; 69: Schattauer 2017 Thrombosis and Haemostasis 7/2017

9 1238 Lip et al. Stroke prevention in AF 14. Hughes M, et al. Stroke and thromboembolism in atrial fibrillation: a systematic review of stroke risk factors, risk stratification schema and cost effectiveness data. Thromb Haemost 2008; 99: Pisters R, et al. Stroke and thromboembolism in atrial fibrillation. Circ J 2012; 76: Lip GY, et al. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest 2010; 137: Chao TF, et al. Validation of a Modified CHA2DS2-VASc Score for Stroke Risk Stratification in Asian Patients With Atrial Fibrillation: A Nationwide Cohort Study. Stroke 2016; 47: Violi F, et al. Ankle-Brachial Index and cardiovascular events in atrial fibrillation. The ARAPACIS Study. Thromb Haemost 2016; 115: Wagstaff AJ, et al. Is female sex a risk factor for stroke and thromboembolism in patients with atrial fibrillation? A systematic review and meta-analysis. QJM 2014; 107: Friberg L, et al. Assessment of female sex as a risk factor in atrial fibrillation in Sweden: nationwide retrospective cohort study. Br Med J 2012; 344: e Banerjee A, et al. Ejection fraction and outcomes in patients with atrial fibrillation and heart failure: the Loire Valley Atrial Fibrillation Project. Eur J Heart Fail 2012; 14: Fauchier L, et al. Should Atrial Fibrillation Patients With Only 1 Nongender-Related CHA2DS2-VASc Risk Factor Be Anticoagulated? Stroke 2016; 47: Hung Y, et al. Is an Oral Anticoagulant Necessary for Young Atrial Fibrillation Patients With a CHA2DS2-VASc Score of 1 (Men) or 2 (Women)? J Am Heart Assoc 2016; 5: pii: e Lip GY, Nielsen PB. Should Patients With Atrial Fibrillation and 1 Stroke Risk Factor (CHA2DS2-VASc Score 1 in Men, 2 in Women) Be Anticoagulated? Yes: Even 1 Stroke Risk Factor Confers a Real Risk of Stroke. Circulation 2016; 133: Lip GY, et al. Non-valvular atrial fibrillation patients with none or one additional risk factor of the CHA2DS2-VASc score. A comprehensive net clinical benefit analysis for warfarin, aspirin, or no therapy. Thromb Haemost 2015; 114: Friberg L, et al. Balancing stroke and bleeding risks in patients with atrial fibrillation and renal failure: the Swedish Atrial Fibrillation Cohort study. Eur Heart J 2015; 36: Guo Y, et al. Multiple risk factors and ischaemic stroke in the elderly Asian population with and without atrial fibrillation. An analysis of 425,600 Chinese individuals without prior stroke. Thromb Haemost 2016; 115: Melgaard L, et al. Assessment of the CHA2DS2-VASc Score in Predicting Ischaemic Stroke, Thromboembolism, and Death in Patients With Heart Failure With and Without Atrial Fibrillation. J Am Med Assoc 2015; 314: Hippisley-Cox J, et al. Derivation and validation of QStroke score for predicting risk of ischaemic stroke in primary care and comparison with other risk scores: a prospective open cohort study. Br Med J 2013; 346: f Hijazi Z, et al. The ABC (age, biomarkers, clinical history) stroke risk score: a biomarker-based risk score for predicting stroke in atrial fibrillation. Eur Heart J 2016; 37: Ruff CT, et al. Cardiovascular Biomarker Score and Clinical Outcomes in Patients With Atrial Fibrillation: A Subanalysis of the ENGAGE AF-TIMI 48 Randomized Clinical Trial. J Am Med Assoc Cardiol 2016; 1: Siegbahn A, et al. D-dimer and factor VIIa in atrial fibrillation prognostic values for cardiovascular events and effects of anticoagulation therapy. A RE-LY substudy. Thromb Haemost 2016; 115: Lip GY, et al. Additive role of plasma von Willebrand factor levels to clinical factors for risk stratification of patients with atrial fibrillation. Stroke 2006; 37: Hijazi Z, et al. The novel biomarker-based ABC (age, biomarkers, clinical history)-bleeding risk score for patients with atrial fibrillation: a derivation and validation study. Lancet 2016; 387: Kirchhof P, et al ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS: The Task Force for the management of atrial fibrillation of the European Society of Cardiology (ESC). Developed with the special contribution of the European Heart Rhythm Association (EHRA) of the ESC. Endorsed by the European Stroke Organisation (ESO). Eur Heart J 2016; Epub ahead of print. 36. January CT, et al AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol 2014; 64: e Freedman SB, et al. Misperceptions of aspirin efficacy and safety may perpetuate anticoagulant underutilization in atrial fibrillation. Eur Heart J 2015; 36: Quinn GR, et al. Wide Variation in Reported Rates of Stroke Across Cohorts of Patients with Atrial Fibrillation. Circulation 2017; 135: Nielsen PB, Lip GYH. Adding Rigor to Stroke Rate Investigations in Patients With Atrial Fibrillation. Circulation 2017; Epub ahead of print. 40. Nielsen PB, Chao TF. The risks of risk scores for stroke risk assessment in atrial fibrillation. Thromb Haemost 2015; 113: Olesen JB, Torp-Pedersen C. Stroke risk in atrial fibrillation: Do we anticoagulate CHADS2 or CHA2DS2-VASc >/=1, or higher? Thromb Haemost 2015; 113: Wilson MR, et al. Quality and predictors of anticoagulant control with vitamin K antagonist for stroke prevention in atrial fibrillation. Thromb Haemost 2016; 116: Apostolakis S, et al. Factors affecting quality of anticoagulation control among patients with atrial fibrillation on warfarin: the SAMe-TT (2)R (2) score. Chest 2013; 144: Ruiz-Ortiz M, et al. Validation of the SAMe-TT2R2 score in a nationwide population of nonvalvular atrial fibrillation patients on vitamin K antagonists. Thromb Haemost 2015; 114: Bernaitis N, et al. The Sex, Age, Medical History, Treatment, Tobacco Use, Race Risk (SAMe TT2R2) Score Predicts Warfarin Control in a Singaporean Population. J Stroke Cerebrovasc Dis 2017; 26: Ruff CT, et al. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials. Lancet 2014; 383: Renda G, et al. Net Clinical Benefit of Non-vitamin K Antagonist Oral Anticoagulants Versus Warfarin in Phase III Atrial Fibrillation Trials. Am J Med 2015; 128: e Pollack CV, Jr., et al. Idarucizumab for Dabigatran Reversal. N Engl J Med 2015; 373: Pollack CV, Jr., et al. Design and rationale for RE-VERSE AD: A phase 3 study of idarucizumab, a specific reversal agent for dabigatran. Thromb Haemost 2015; 114: Connolly SJ, et al. Andexanet Alfa for Acute Major Bleeding Associated with Factor Xa Inhibitors. N Engl J Med 2016; 375: Greinacher A, et al. Reversal of anticoagulants: an overview of current developments. Thromb Haemost 2015; 113: Ageno W, et al. Managing reversal of direct oral anticoagulants in emergency situations. Anticoagulation Education Task Force White Paper. Thromb Haemost 2016; 116: Freedman B, Lip GY. Unreal world or real world data in oral anticoagulant treatment of atrial fibrillation. Thromb Haemost 2016; 116: Avgil-Tsadok M, et al. Dabigatran use in elderly patients with atrial fibrillation. Thromb Haemost 2016; 115: Carmo J, et al. Dabigatran in real-world atrial fibrillation. Meta-analysis of observational comparison studies with vitamin K antagonists. Thromb Haemost 2016; 116: Hecker J, et al. Effectiveness and safety of rivaroxaban therapy in daily-care patients with atrial fibrillation. Results from the Dresden NOAC Registry. Thromb Haemost 2016; 115: Beyer-Westendorf J, et al. Effectiveness and safety of dabigatran therapy in daily-care patients with atrial fibrillation. Results from the Dresden NOAC Registry. Thromb Haemost 2015; 113: Darwiche W, et al. Risk of myocardial infarction and death in patients with atrial fibrillation treated with dabigatran or vitamin K antagonists. Meta-analysis of observational analyses. Thromb Haemost 2016; 116: Larsen TB, et al. Comparative effectiveness and safety of non-vitamin K antagonist oral anticoagulants and warfarin in patients with atrial fibrillation: propensity weighted nationwide cohort study. Br Med J 2016; 353: i Lip GY, et al. Real-world comparison of major bleeding risk among non-valvular atrial fibrillation patients initiated on apixaban, dabigatran, rivaroxaban, or warfarin. A propensity score matched analysis. Thromb Haemost 2016; 116: Thrombosis and Haemostasis 7/2017 Schattauer 2017