Should we screen for atrial fibrillation?

Thank you for an interesting pro et contra account on atrial fibrillation (AF) screening [1]. Although comprehensive and balanced, I would like to add some points that may complement the debate, and that may also be of interest in relation to the SAFER study, the world’s largest randomised controlled trial (RCT) of systematic screening for AF, recently launched in the UK [2, 3].

First, it is stated in the pro-argument that “current evidence provides a strong case for introducing AF screening now”. Still, the author recognizes that “there is likely to be an optimal screening duration and frequency for detecting paroxysmal AF” [1]. The latter is a crucial point, however largely unknown. We simply do not know how we should screen, for how long we should screen or who we should screen – in order to reach the desired outcomes (reduced stroke incidence or mortality). Only RCTs, such as the STROKESTOP and SAFER study [2, 4], can adequately respond to these crucial question.

Meanwhile, data from observational studies may provide us with some hints. In our Akershus Cardiac Examination (ACE) 1950 study in Norway[5], we performed a screening study using the exact same method as in STROKESTOP, and as planned in the SAFER study (Zenicor© handheld one-lead ECG; 30 seconds, twice daily and when symptoms, for two weeks). In the ACE 1950 study, a general population aged 65 years, was recruited. Screening was offered all those with one or more additional risk factor(s) for stroke (other than sex and age), according to the CHA2DS2-VASc score (i.e. ≥2 in men and ≥3 in women) (n=1,510). Still, in this ‘increased risk’ population, we found a relatively low prevalence of previously unknown AF; 0.9% at the age of 65 years [5].

On the other hand, as for the STROKESTOP study, our study demonstrated that handheld ECG seems suitable for the purpose of population screening. It was well accepted; as many as 99.5% of the participants recorded ≥20 ECGs during two weeks, and the rate of ECG ‘incidentalomas’ was negligible. Additional ECG monitoring (Holter or R test), due to low quality of the handheld ECGs, was only needed in 8 out of 1,510 participants [5].

One relevant question of interest with regard to screening by handheld ECG, is: What proportion of all cases of unknown AF is found by a two-week (or more), intermittent ECG strategy? While long-term continuous ECG (e.g. implanted loop recorders) may provide us with the ‘true prevalence’ of AF, we do not know whether all the AF potentially revealed by continuous surveillance, carry the same stroke risk as ‘traditional’, clinically detected AF. Bear in mind, the latter is the kind of AF largely represented in the trials we build our current practice on. In fact, evidence already exists pointing towards a positive association between increased AF burden (i.e. ‘time in AF’) and stroke risk [6, 7].

So, do we want to know of any little irregularity of the atria? No, probably not. And if so, can intermittent handheld ECG be viewed as a reasonable compromise between a single standard ECG and long-term continuous ECG, for the purpose of screening? Yes, probably. We may hypothesize that ‘multiple snapshots’ of 30 seconds during 2 weeks may detect more of the ‘relevant’ (i.e. at increased risk of stroke) unknown AF, while avoiding to pay notice to every single atrial run. The problem is; we do not how low burden of AF it is relevant to find (and treat). Fortunately, there are ongoing RCTs investigating this, but only in selected populations with existing implanted cardiac devices (and only from the comparable case of AHREs; atrial high-rate episodes, as identified from intracardiac electrograms) [8, 9]. Unfortunately, no studies have, to my knowledge, systematically investigated the proportion of ‘total unknown AF’ identified by intermittent handheld ECG, compared to long-term implantable ECG devices. If not already planned for, would it not be of interest to assess this in a subgroup of the SAFER study?

Lastly, I would like to add a point of concern regarding the use, or misuse, of the Wilson and Jungner WHO criteria for disease screening [10]. The apparent fulfilment of all or most of these criteria are widely being used as an argument that AF screening should or must be implemented [11]. However, as pointed out by UK screening experts [12], Wilson and Jungner regarded these principles only as a preliminary checklist, or a prerequisite to justify the search (and funding) of further evidence from RCTs. This is where we stand now, waiting for the evidence.

To conclude, we should not rush into any form of systematic screening for AF, before randomised controlled trials provide us with much needed answers. Meanwhile, we should continue raising awareness of AF and related risks, and at the same time continue the search for AF among our patients – be it by conventional or new methods of recording ECGs.

References:
1. Lown M, Moran P, Should we screen for atrial fibrillation? BMJ, 2019. 364:l43.
2. SAFER study webpage. [Accessed 2019 February 18th,]; Available from: https://www.safer.phpc.cam.ac.uk.
3. NIHR awards £3m for new research to investigate screening to prevent one in ten strokes (press release, 16 May 2018). [Accessed 2019 February 18th,]; Available from: http://www.phpc.cam.ac.uk/pcu/nihr-awards-3m-for-new-research-to-investi....
4. Svennberg E, Engdahl J, Al-Khalili F, Friberg L, Frykman V, Rosenqvist M, Mass Screening for Untreated Atrial Fibrillation: The STROKESTOP Study. Circulation, 2015. 131(25):2176-84.
5. Berge T, Brynildsen J, Larssen HKN, Onarheim S, Jenssen GR, Ihle-Hansen H, . . . Tveit A, Systematic screening for atrial fibrillation in a 65-year-old population with risk factors for stroke: data from the Akershus Cardiac Examination 1950 study. Europace, 2017. 10.1093/europace/eux293 [Epub ahead of print].
6. Van Gelder IC, Healey JS, Crijns H, Wang J, Hohnloser SH, Gold MR, . . . Connolly SJ, Duration of device-detected subclinical atrial fibrillation and occurrence of stroke in ASSERT. Eur Heart J, 2017. 38(17):1339-44.
7. Vanassche T, Lauw MN, Eikelboom JW, Healey JS, Hart RG, Alings M, . . . Connolly SJ, Risk of ischaemic stroke according to pattern of atrial fibrillation: analysis of 6563 aspirin-treated patients in ACTIVE-A and AVERROES. Eur Heart J, 2015. 36(5):281-7a.
8. Lopes RD, Alings M, Connolly SJ, Beresh H, Granger CB, Mazuecos JB, . . . Healey JS, Rationale and design of the Apixaban for the Reduction of Thrombo-Embolism in Patients With Device-Detected Sub-Clinical Atrial Fibrillation (ARTESiA) trial. Am Heart J, 2017. 189:137-45.
9. Kirchhof P, Blank BF, Calvert M, Camm AJ, Chlouverakis G, Diener HC, . . . Vardas P, Probing oral anticoagulation in patients with atrial high rate episodes: Rationale and design of the Non-vitamin K antagonist Oral anticoagulants in patients with Atrial High rate episodes (NOAH-AFNET 6) trial. Am Heart J, 2017. 190:12-8.
10. Wilson JM, Jungner YG, Principles and practice of mass screening for disease. Public Health Papers No. 34. 1968, Geneva: World Health Organization.
11. Neubeck L, Orchard J, Lowres N, Freedman SB, To Screen or Not to Screen? Examining the Arguments Against Screening for Atrial Fibrillation. Heart Lung Circ, 2017. 26(9):880-6.
12. Raffle AE, Gray M, Screening: evidence and practice. 1st ed. 2007, New York, the United States: Oxford University Press Inc (textbook). 317.

Disease labeling of populations with common phenomena of ageing (Non valvular AF) is indeed a risky business. As is the albeit common view presuming causality in a highly confounded statistical association. The non biased term is to consider stroke as commonly associated with atrial fibrillation and the issue linked to this is the problem of defining where anticoagulation can credibly reduce or defer recurrent injury and death in those at risk.

Any prevalence of atrial fibrillation needs a causal process established and risk factor, for the aetiology of AF (not just the consequences such as stroke or haemorrhage) defined. Stroke associated with atrial fibrillation is clearly not always thromboembolic and anticoagulation reduces but does not eliminate stroke in chronic atrial fibrillation patients. Anticoagulation may defer stroke (again not prevent stroke or death) but also causes deaths in a minority due to unprovoked bleeding. These patients are not easily identified by crude population bleeding risk scores.

The older the sample frame the more prevalent unrecognised and asymptomatic paroxysmal AF will be. What to do about any AF is far from clear. It is not helped by confusing any AF with persistent AF, in a young heart from an old heart, from a structurally normal heart from one with multiple structural abnormalities. Reacting to any AF in any circumstance is fraught with error, bias and confounded by risk to the patient. That 90% of a small sample accepted anticoagulation for any AF in Strokestop is in itself hardly surprising. It does not confirm an informed or even valid choice and actually probably suggests that the risk was grossly over egged and the benefit over simplified. It is critical to remember that CHADS VASC is a risk equation developed in chronic persistent non valvular atrial fibrillation . As with many population scores while helpful in those where it was intended to be used, even in those it will over estimate the stroke rate. Hence new risk scores continue to be tested in different sample frames.

Th ARREST AF trial and many others shows anticoagulated patients continue to have strokes due to poor control of atherosclerotic risk factors (HBP, DM, renal impairment), unchecked obesity/hypo-ventilation/OSA smoking and lack of fitness. This seems sensible as none of these respond to anticoagulation. Equally we know paroxysmal AF patterns can be present for decades without consequences in younger individuals many of whom reach late middle age tolerating brief and frequently uncontrolled periods in AF without any consequences beyond transient irritation for minutes, hours or days. Equally atrial flutter is frequently severely symptomatic yet has reduced thromboembolic risk and the potential for ECG patterns of flutter/fibrillation has been long documented and are seen daily and again likely to have reduced thromboembolic risk. While improving tools for marginally safer and as effective anticoagulation (NOAC agents) have emerged over the last decade these are not effective treatments for ischaemic or atherosclerotic stroke.The mystery of the linkage of AF to thromboemblism is just not as simple as often portrayed.

Thus caution is needed with simple views that screening everyone for anything and treating the majority is as has been seen so often to be the case far from informed. Cost utility of screening is complex but more directly patients choices are not informed to be told a simplistic tale of one size fits all. Unfortunately presuming trials are there to "strengthen the case" reveals confirmation bias exists before any actual data is available. Show us the data first then we might consider its value and truly inform patients. Till then I would advise you use your watch for telling the time?

Dear Sirs,

The debate the authors have present in the article regarding the need for an AF screening programme is thought-provoking both from the point of view of disease burden, and the continuing appearance of screening in UK healthcare.

It seems an attractive prospect to implement a screening programme for this disease which is highly identifiable, treatable and as the authors have highlighted – is expanding rapidly with our ageing population. I am encouraged to read, through cohort study, that anticoagulation in so called screen-detected AF is associated with favourable outcomes in terms of survival and stroke risk[1]. However, surely these initial findings must be substantiated or challenged further in a randomised control trial before we seriously consider the role of a new screening programme.

I have my own concerns with AF screening. Firstly, would be the downstream cost (on finances and resources) of implementing the programme. If one is implemented, there will be a significant rise in patients requiring treatment by anticoagulation, be it warfarinisation or with newer, DOACs. In our current healthcare economy do we have the capacity to manage this batch of new patients i.e. anticoagulation clinic provision, ample trained clinicians for accurate diagnosis at screening, GPs monitoring more INRs, acute management of therapy-associated bleeding? However, if the benefit in terms of mortality reduction in heart failure, MI, ischaemic stroke and death from AF outweighs the resource burden, then perhaps the logistical challenge of screening is one we should accept. Nonetheless, this will be no small challenge given the estimated prevalence of AF in 2060 at more than 1 million people in the UK[2]. Another caveat to this is the interval between screening, additionally the authors’ note on paroxysmal AF being identified adds a source of possible unnecessary treatment and uncertain diagnosis.

The average UK citizen is subjected to an increasing variety of screening programmes throughout their lifetime, some more effective and cost-efficient than others. The NHS England abdominal aortic aneurysm (AAA) programme for example, was reviewed as continuing to be cost-effective[3] in 2016 and was previously observed to cost £5758 per year of life gained[4]. Although, with patients attending multiple different screening appointments, is uptake dwindling as they have to juggle more and more interaction with healthcare services on top of their own personal commitments? An experienced clinician can identify a possible AF simply through taking a patients pulse. The authors alluded to the rising use of self-monitoring devices and smart watches etc. in the role of identifying AF. Yet, is it time we handed some control to patients and encourage them to be more self-interested with their own health and fitness. How many patients/relatives/carers are able to monitor the pulse of themselves or another, a very simple skill? Public campaigns such as ‘Think FAST!’[5] have promoted awareness of the cardinal signs of a stroke, but to what degree are the public aware of the existence and simple detection of AF – one of the major causes of ischaemic stroke?

I welcome the idea of a screening programme for AF, provided it had the backing of more substantial, RCT based evidence. Ultimately, screening would see a rise in the number of people with an new diagnosis – do we have the resources to tackle that and are there more simple methods than a typical screening programme to identify atrial fibrillation and for patients to become more involved?

Yours sincerely,

Dr Jordan Wardrope,
MBChB, BMSc (Hons),
Junior Doctor in General Surgery

References:
1. Freedman B, Camm J, Calkins H, Healey JS et al. Screening for Atrial Fibrillation: a report of the AF-SCREEN International Collaboration. Circulation 2017; 135:1851-67
2. Lane DA, Skjøth F, Lip GYH, Larsen TB et al. Temporal trends in incidence, prevalence, and mortality of atrial fibrillation in primary care. J Am Heart Assoc 2017;6:e005155
3. Jacomelli J, Summers L, Stevenson A, Lees T et al. Impact of the first 5 years of a national abdominal aortic aneurysm screening programme. Br J Surg 2016;103(9):1125-31
4. Glover MJ, Kim LG, Sweeting MJ, Thompson SG et al. Cost-effectiveness of the National Health Service Abdominal Aortic Aneurysm Screening Programme in England. Br J Surg 2014;101(8):976-82
5. Dombrowski SU, Mackintosh JE, Sniehotta FF, Araujo-Soares V et al. The impact of the UK ‘Act FAST’ stroke awareness campaign: content analysis of patients, witness and primary care clinicians’ perceptions. BMC Public Health 2013;13(1):915