One in Three !! : Atrial Fibrillation and Stroke  

AUTHORS: Professor Rónán Collins, consultant physician in geriatric and stroke medicine and Associate Clinical Professor in Gerontology Trinity College Dublin.
Professor Collins is Clinical lead National Stroke Programme, Member of the British and Irish Stroke Physicians (BIASP), Fellow of European Stroke Organsiation (FESO), Steering group member for EHRA-PATHS and member of the EHRA writing group for Practical guide to anticoagulation with NOACS in non-vavlulat atrial fribrillation.

Atrial Fibrillation (AF) is the commonest cardiac arrhythmia affecting at least 3% of Irish adults over 60 and 6 % over 70 [1]. The worldwide prevalence has been rising and is approximately 600 in men and 375 in women per 100,000 population [2].  The lifetime risk of developing atrial fibrillation is almost 1 in 3 and it is estimated  there may be as many as 18 million people with AF in the EU by 2060.

Despite this very significant burden of disease we are still  in a relative infancy in our understanding of the pathophysiology of AF as a truly systemic condition and  its association with risk of stroke and other illnesses. AF is associated with  a 2-3.5 fold increase in mortality, impaired left ventricular dysfunction in 20-30% of cases, a hazard ratio of 1.6 for the development of dementia independent of previous stroke and associated with an impaired  quality of life in almost 60% cases and a hospitalisation rate of 10-40% annually [3].

While much progress has been made in the last 10 years in particular with collaborative interdisciplinary working and research between cardiologists and stroke physicians, nevertheless there are many unanswered questions and dilemmas in AF.

Much attention has focused on the potential to screen for AF given it fulfils many of the original Wilson and Jungner criteria adopted by the WHO for a disease condition which would be amenable to a screening programme (see table 1.) and more specifically that the condition is common, detectable, associated with serious disease (stroke) which is preventable (with anticoagulation ) [4].iven those facts the national clinical programmes for stroke and chronic disease management in  Ireland have introduced case finding as part of the primary care contract to opportunistically pulse check for AF in an older population. The advent of personal digital devices with the potential to detect irregularities of heart rhythm has expanded this potential to detect AF .

A proof of concept  study using apples i-phone and watch  for example found a positive predictive value 0.84 when a notification was received of an irregular pulse [5]. The recently published ‘Smartphone-based screening for atrial fibrillation: a pragmatic randomised clinical trial (eBrave)’ showed a doubling of detection of clinically relevant atrial fibrillation in a cohort of older people who owned a smartphone in both phases of the crossover design [6]

In addition many more people now have implantable pacemaker and ICD devices capable of detecting AF if interrogated for same and there is increasing use of implantable recorders in patients with a high index of suspicion for AF. Many such devices pick up brief runs of AF or Atrial High Rate Episodes (AHREs)that may suggest a risk of developing AF though it is not proven that  anticoagulation in such cases would prevent more stroke  .

Screening for AF has huge potential benefits not least in preventing stroke, development of AF associated symptoms or preventing irreversible remodelling of the atrium and associated atrial or ventricular tachycardia associated cardiomyopathy.  It can also however create unnecessary anxiety, lead to over diagnosis and treatment with anticoagulation or even potentially lead to further invasive investigations that may have associated risks and be unnecessary .

At present despite a growing body of evidence for use of digital devices, there is no evidence to suggest more strokes would be prevented adopting such strategies over opportunistic screening by pulse checking in high- risk groups. An overview of reported specificities and sensitivities of various detection approaches is outlined below in table 2 .  All suspected AF  must be confirmed by a 12 lead ECG or a 30 second single lead strip read by a physician experienced in AF diagnosis.

Table 2 Reported reliabilities of methods of AF detection

	sensitivity	specificity
Pulse taking	87-97%	70-81%
Automated BP monitors	93-100%	86-92%
Single lead ECG	94-98%	76-95%
Smartphone devices /watches	91-99%	84-100%

• Adapted from Hindricks G et al. European Heart Journal 2021;42(5):373–498

Stroke and AF

In recent years international guidelines have correctly focused more on stroke prevention as the priority in the management of AF [7]. In their “Atrial Fibrillation Better Care (ABC)” approach to AF the latest ESC guidelines on the management of AF stress the ‘A’ -to Avoid stroke / Anticoagulation  as the prime issue of concern before addressing ‘B’ – better symptom control and ‘C’ optimisation of cardiovascular comorbidity and other risk factors ‘ .

Sum Zero game : The Low  CHA₂DS₂-VASC patient

While the CHA₂DS₂-VASC score has remained the standard tool for risk assessment , there are concerns that it performs modestly and has relatively poor validity ( C-statistic of 0.64 at best) and that the emphasis should perhaps be on a positive bias for anticoagulation  than not . This author still regularly sees cases of stroke due to atrial fibrillation in people with seemingly low risk as indicated by a CHA₂DS₂-VASC score of 0 or 1 (for female sex)  who are on aspirin or nothing rather than an anticoagulant and this approach is not necessarily at odds with current  guidelines.

While at higher scores the CHA₂DS₂-VASC performs modestly it does seem to perform more consistently when the a score is low and the incidence of stroke seems < 1% consistently in those with a score of 0 or 1 (for female sex) . However it is increasingly recognised that some populations with low CHA₂DS₂-VASC scores of 0 or 1 may have appreciably higher risk of stroke than previously thought, (e.g up to 2% risk per annum) and that this risk seems greater in studies of Asian populations and perhaps in other groups too [8,9].

A limitation of CHA₂DS₂-VASC is that it does not include other variables such as chronic kidney disease, obstructive sleep apnoea etc. though these variables  are often closely related to the CHA₂DS₂-VASC risk factors themselves and their inclusion  does not seem to improve score predictability. It is unknown as to whether including factors such co-existent malignancy, smoking or obesity would improve reliability. Then as we age we must recognise that the risk of stroke with AF is a continuum and likely to increase with the duration of AF as the atrium becomes increasingly dilated and dysfunctional  with chronic fibrillation.

For all these reasons it is good practice that people with a low CHA₂DS₂-VASC score where there may be doubt about initiating anticoagulation , should be referred for a specialist opinion and a thorough assessment of risk. This may  include assessment of biomarkers such as D-Dimer, Troponin, NT ProBNP, IL-6 and CRP among others, many of which have been shown to independently improve upon the risk stratification  of CHA₂DS₂-VASC alone. Echocardiographic evaluation of the left atrium to look for  dilatation, spontaneous contrast or thrombus in the left atrial appendage,  low atrial appendage exit velocity and measurement of  Total atrial conduction time (PA-TDI) may also be useful in individual risk stratification [10]. MR imaging is another growing area of interest in assessing left atrial function.   

Both biomarkers and echocardiographic variable have been shown to independently predict risk of stroke and may well improve upon the accuracy of the CHA₂DS₂-VASC score. A number of more complex risk scores such as Garfield-AF, Anticoagulation and Risk factors in Atrial Fibrillation (ATRIA) and ABC-stroke have been shown to modestly improve upon CHA₂DS₂-VASC though a pragmatic balance must also be struck between the time involved to use such tools in clinical practice and added expenses of additional measurements  versus accruing benefit . At least one biomarker based risk stratification tools is currently undergoing evaluation in a RCT (ABC-AF Study NCT 03753490).

While acknowledging the ESC guidelines it is important to remember that, as pointed out in the preamble ,  that a guideline does not robustly cover every situation where there is a paucity of evidence and is not a substitute for physician based judgement of their individual patient. In discussing the uncertainty of apparent low risk of stroke  with patients it is my personal practice to advise anticoagulation in patients with low CHA₂DS₂-VASC scores (i.e. 0 or 1 for female sex alone) with elevated biomarkers or possibly ‘risky’ echocardiographic measurements. All other patients need to understand that the risk of stroke is not Zero, that the risk of stroke  in AF is a continuum and unless there is a high risk of bleeding , that anticoagulation may be the best strategy especially where the risk of bleeding is low. Where a decision is made not to anti-coagulate in the CHA₂DS₂-VASC =0 or 1 (for female sex)  patient , this should be reviewed regularly with reassessment of  risk factors  . Almost 15% of new AF cases will have one new non sex CHA₂DS₂-VASC risk factor detected at one year the majority of new comorbidities detected within 5 months after diagnosis [11]  

There is no role for aspirin or clopidogrel therapy in the prevention of stroke in AF  and indeed it may just increase risk of bleeding without  benefit.

The AF patient with stroke

A third of our strokes in Ireland are associated with AF and AF is a strong predictor of recurrent events  Secondary prevention for the AF stroke patient is of prime importance and  applies to both stroke due to ischaemia and probably surprisingly also in cases of stroke due intracerebral haemorrhage .

Post-stroke AF patient with cerebral infarction

Stroke recurrence is an early risk in AF . A significant decision is when to anticoagulate the AF stroke patient after a stroke due to infarction.. This is generally a specialist decision made with the benefit of repeated neuroimaging and assessment of infarct size and the presence or absence of significant haemorrhagic transformation. A traditional role of thumb, largely dictated by concern for haemorrhagic transformation in the era of Warfarin, had been to start anticoagulation in the post-stroke AF patient on the same day -2 days for TIA small infarcts, 4-7 days for moderate sized infarcts and 10-14 days for large volume infarcts. This was always a judgement based on expert consensus rather than expert science.

More recently   the TIMING study published in Circulation [12], showed that it was safe to start anticoagulation with a NOAC in the AF patient with cerebral infarction < 5 days (early) compared to starting > 5 days (delayed) . ‘Early’ initiation was not inferior to ‘Delayed’ initiation with a NOAC and though, perhaps not surprisingly, it was associated with less recurrent events numerically, it was not found to be superior. A caveat to the trial was the incidence of intracerebral haemorrhage was very low (only 3 cases and not symptomatic) compared to anticipated 3-4% incidence from previous studies in a warfarin era, and the median NIHSS was only 4 suggesting mainly milder and perhaps lower volume infractions, thought the latter was not detailed. A previous small study initiating rivaroxaban or warfarin within 5 days of infarction showed no difference in rates of recurrence or bleeding on neuroimaging at 4 weeks and the AREST study comparing early initiation with apixaban within 5 days versus delayed initiation with warfarin showed a 2.1 % incidence of haemorrhagic transformation in the warfarin group versus non with apixaban [13,14].  

Three further RCTs ( OPTIMAS NCT 03759938, ELAN NCT 03148457 and START NCT 03021928 )aim to examine the issue of superiority of an early versus delayed strategy to initiation of anticoagulation with a NOAC in the post-stroke AF patient, but the tide is moving to earlier initiation of anticoagulation for AF patients post stroke due to infarction and with a NOAC as the safest option.

The Post stroke AF patient with intracerebral haemorrhage

It seems counter-intuitive   and perhaps even dangerous to be considering anticoagulation in the AF patient who has suffered an intracerebral haemorrhage (ICH), but good longitudinal data shows that the AF patient who has suffered a stroke due to ICH  does better in terms of recurrent stroke events (both ischemic  and haemorrhagic ) if started on anticoagulation than not {15,16]  . A number of caveats need to be stated here, the initiation needs careful risk assessment by a stroke specialist with repeated neuroimaging to ensure the volume of ICH is stable and resolving, that the patient does not have underlying cerebral amyloid angiopathy and that underlying anatomical risks such as aneurysm or AV malformation have been adequately addressed and treated and that blood pressure control is good.

If the ICH had occurred in the AF patient on an anticoagulant already , then it is important to undertake an assessment of pre-morbid drug adherence; review the  time in therapeutic range (TTR) if on warfarin ; ensure correct dose of NOAC  adjusted for creatinine clearance, age and weight (where indicated); conduct a review of the full prescription for  other drugs that increase haemorrhagic risk (e.g. concomitant antiplatelet use which may no longer be needed ) .

In such cases where the AF patient had been on an oral anticoagulant before the ICH , then consideration should be given to switching from warfarin to a NOAC in all cases (unless patient has true valvular AF* or a mechanical heart valve) as NOACS are associated with a 50% lower risk of intracranial haemorrhage and in vent of iCH outcomes are better on NOACS than Vitamin K antagonists [17]. If the patient had been on a NOAC pre-ICH consider switching agent or lowering the dose of anticoagulant where appropriate (e.g. using the lower dose of dabigatran if on higher dose) . In general a period of 4-8 weeks has been suggested with repeated neuroimaging before restarting anticoagulation in the AF patient who has suffered stroke due to ICH.   

Ongoing Management of Atrial Fibrillation

ESC guidelines promote an integrated management to the AF patient and in their ‘circle of ideal integration’ list the importance of access to a many specialty model with the patient at the centre of decision making. A criticism personally would be the very obvious glaring omission of geriatric medicine, given that AF is primarily a disease of later life and that many patients with AF have significant co-morbidities, polypharmacy, frailty syndromes and  may have already had a stroke , be suffering with cognitive impairment and pose some of the more difficult decisions when deciding to initiate anticoagulation and its subsequent management  .

While the available research on models of integrated care for AF have had mixed results the components of such care have been heterogenous, none have included a specialist in gerontology and it is intuitive that a properly constructed integrated model of care that includes cardiology, gerontology / stroke-neurology , advanced nurse practitioners, clinical pharmacy and clinical nutrition would improve on patient education , medication adherence, better decision making re stroke prevention, AF symptom management, reduced drug errors or unnecessary prescription and reduced burden of AF and symptoms . It is important at each visit that the patients’ CHA₂DS₂-VASC , HAS-BLED, creatinne clearance as calculated by the Cockroft -Gault equation and the European Heart Rhythm Association (EHRA) Symptom Score (table 3) [18]

Table 3 Modified EHRA symptom score for AF

mEHRA score	symptoms	description
1	No symtoms
2a	Mild	Mild. Daily activity not affected and symptoms not troublesome to the patient
2b	Moderate	Moderate. Daily activity not affected but symptoms troublesome to the patient
3	Severe	Normal daily activity affected
4	Disabling	Normal daily activity discontinued

 Adapted from  Wynn GJ et al. EP Europace, 2014:16(7):965–972, https://doi.org/10.1093/europace/eut395

The frequency of return assessments at an AF service  is largely determined by patients symptoms and uncontrolled co-morbidities such as hypertension which can increase risk  of bleeding , or the presence of troublesome bleeding itself. In all other cases a good rule of thumb is to see the patient a month after initiation of oral anticoagulation and thereafter at monthly intervals suggested by the simple formula of [19]

Interval of Routine follow up in months  =   creatinine clearance (Cockroft Gault)

                                                                                                       10

Many older patients with AF have  significant co-morbidities such as cancer, hypertension , dementia, polypharmacy or are trauma patients or awaiting surgery . Management of AF within this context can be challenging as oral anticoagulation can pose a risk . The EHRA-PATHS project funded by Horizon 2020 aims to address the need for a more holistic approach to the older patient with AF through development of pathways of care that trigger both actions and KPIs for the domains of detected / suspected co-morbidity [20]. It is planned to do a RCT on the new proposed pathways of care and you can read more about this important project in the management of AF at  https://ehra-paths.eu/

References

1. Jansen S et al Atrial fibrillation is associated with impaired mobility in community-dwelling older adults.  J Am Med Dir Assoc. 2014 Dec;15(12):929-33

• Chugh SS et al. Worldwide epidemiology of atrial fibrillation: a Global Burden of Disease 2010 Study. Circulation. 2014 ;129(8):837-47

• Meyre P et al. Risk of hospital admissions in patients with atrial fibrillation: a systematic review and meta-analysis. Can J Cardiol 2019;35:13321343.

• Wilson JMG & Jungner G (1968). Principles and practice of screening for disease. Public Health Papers 34. Geneva, World Health Organization

• Perez, M.V. et al. Large-Scale Assessment of a Smartwatch to Identify Atrial Fibrillation. New England Journal of Medicine 381, 1909–1917.. doi:10.1056/nejmoa1901183

• Rizas KD et al. Smartphone-based screening for atrial fibrillation: a pragmatic randomized clinical trial. Nat Med. 2022 Sep;28(9):1823-1830. doi: 10.1038/s41591-022-01979-w. Epub 2022 Aug 28. PMID: 36031651.

•  Hindricks G. et al . Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS): The Task Force for the diagnosis and 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, European Heart Journal, Volume 2021;42(5):373–498, https://doi.org/10.1093/eurheartj/ehaa612

• Stroke Risk Stratification Schemes in Atrial Fibrillation in the Era of Non- Vitamin K Anticoagulants: Misleading and Obsolete, At Least for the “Low-Risk” Patients? Manolis AS, Manolis TA, Manolis AA, Melita H. Curr Drug Targets. 2017;18(16):1852-1865
• Refinement of ischemic stroke risk in patients with atrial fibrillation and CHA2 DS2 -VASc score of 1. Huang D1, Anguo L, Yue WS, Yin L, Tse HF, Siu CW Pacing Clin Electrophysiol. 2014 Nov;37(11):1442-7.

1. Müller P et al Echocardiography-derived total atrial conduction time (PA-TDI duration): risk stratification and guidance in atrial fibrillation management. Clin Res Cardiol. 2021;110(11):1734-1742. doi: 10.1007/s00392-021-01917-9.

1. Chao TF et al. Incident co-morbidities in patients with atrial fibrillation initially with a CHA2DS2-VASc score of 0 (males) or 1 (females): implications for reassessment of stroke risk in initially ‘low-risk’ patients. Thromb Haemost 2019;119:11621170

1. Oldgren J et al. Early Versus Delayed Non–Vitamin K Antagonist Oral Anticoagulant Therapy After Acute Ischemic Stroke in Atrial Fibrillation (TIMING): A Registry-Based Randomized Controlled Noninferiority Study Circulation 2022-10-04 146(14): 1056-1066

1. Hong KS et al. Rivaroxaban vs warfarin sodium in the ultra-early period after atrial fibrillation-related mild ischemic stroke: a randomized clinical trial. JAMA Neurol. 2017;74:1206–1215. doi:

10.1001/jamaneurol.2017.2161

1. Labovitz AJ et al.  Early apixaban use following stroke in patients with atrial fibrillation: results of the AREST trial. Stroke.2021;52:1164–1171. doi: 10.1161/STROKEAHA.120.030042

1. Korompoki E et al. Long-term antithrombotic treatment in intracranial hemorrhage survivors with atrial fibrillation. Neurology. 2017 Aug 15;89(7):687-696. doi: 10.1212/WNL.0000000000004235. Epub 2017 Jul 19. PMID: 28724590; PMCID: PMC5562962.

1. Biffi A et al. Oral Anticoagulation and Functional Outcome after Intracerebral Hemorrhage. Ann Neurol. 2017 Nov;82(5):755-765. doi: 10.1002/ana.25079. Epub 2017 Oct 31. PMID: 29028130; PMCID: PMC5730065.

1. Boulouis G et al. Outcome of intracerebral haemorrhage related to non-vitamin K antagonists oral anticoagulants versus vitamin K antagonists: a comprehensive systematic review and meta-analysis. J Neurol Neurosurg Psychiatry. 2018 Mar;89(3):263-270. doi: 10.1136/jnnp-2017-316631. Epub 2017 Oct 13. PMID: 29030422.

1. Wynn GJ et al. The European Heart Rhythm Association symptom classification for atrial fibrillation: validation and improvement through a simple modification, EP Europace 2014:16(7): 965–972, https://doi.org/10.1093/europace/eut395

1. Steffel et al . 2021 European Heart Rhythm Association Practical Guide on the Use of Non-Vitamin K Antagonist Oral Anticoagulants in Patients with Atrial Fibrillation, EP Europace, 2021;23(10):1612–1676, https://doi.org/10.1093/europace/euab065

• Heidbuchel H et l. ESC and EHRA lead a path towards integrated care for multimorbid atrial fibrillation patients: the Horizon 2020 EHRA-PATHS project. Eur Heart J. 2022 Apr 14;43(15):1450-1452. doi: 10.1093/eurheartj/ehab672. PMID: 34694355.