At the AF Symposium (11–13 January 2018, Orlando, USA) John Camm, St George’s University of London and Imperial College, London, UK, looked at the current status of screening for atrial fibrillation (AF). This is as an area of controversy with practitioners on both sides of the argument.
“Many people who suffer from embolic cerebral infarctions and ischaemic events have asymptomatic atrial fibrillation,” stated Camm. “Around 2% of the general western population have AF and in around 10–20% it is undiagnosed, meaning that screening could play a significant role in stroke prevention. This is particularly true as strokes associated with AF have a poor prognosis compared with strokes of other aetiologies.”
There are two main ways of screening: opportunistic and systematic. Opportunistic screening occurs when a patient attends a doctor for an unrelated condition, for example a health check or a flu vaccine, and the doctor or his/her staff take the opportunity to assess the cardiac rhythm using some simple technique, such as a rhythm strip, a blood pressure monitor with rhythm analysis, or a specific device for screening for AF, such as MyDiagnostick. Systematic screening involves ECG assessment of a group of patients, such as everyone aged 75, specifically to identify their heart rhythm. Large numbers of people are often invited to take part in organised screening programmes.
“The SEARCH-AF stroke prevention study (screening education and recognition in community pharmacies of atrial fibrillation) set out to see if opportunistic screening was a worthwhile endeavour in preventing stroke in those over 65 years in age. 1 It offered opportunistic screening using 1000 patients with an average age of 79±6 years and all had CHA 2DS2-VASc score ≥2,” said Camm. “Pharmacists used an AliveCor Kardia monitor (AliveCor) attached to a WiFi-enabled mobile phone to obtain ECGs (iECGs) to identify AF in patients that came in to pick up prescriptions.” Patients were given one ECG and of the patients that were screened 15 (1.5%) were found to have new AF. The modelled cost of the screening worked out at around A$30,000 per stroke prevented.A similar study in Sweden used systematic screening.
“A study, called STROKESTOP, looked at people between the ages of 75 and 76 years from two areas in the Stockholm region.2 They were screened with repetitive ECGs over a two-week period. Of the around 13,000 people invited only around 7,000 agreed to be screened,” explained Camm “Of those with unknown histories of AF the yield was 3%, double the yield from the SEARCH-AF opportunistic screening study. The STROKESTOP study also potentially improved outcomes for those with known AF, offering the opportunity to improve their medical care and treatment.”
Another study, REHEARSE-AF, was a systematic screening study but increased the number of ECGs offered.3 It again used the AliveCor device to screen 1,000 patients, ECGs were taken two-times a week over a period of 12 months. The study found that 4% of patients had undiagnosed AF— seemingly there is a link between the number of ECGs given and the number of people diagnosed with AF. The REHEARSE-AF cohort were above 65 years of age, had CHA2DS2-VASc score of more than 2. As 4% were identified as having AF and routine care picked up about 1% of patients who had AF, the yield was about 3%. The study was relatively expensive costing around US$10,000 per AF diagnosis.
In 2016 the European Society of Cardiology issued guidelines for screening for atrial fibrillation. They recommended opportunistic screening in those over 65 years old by pulse taking or ECG. They also considered systematic ECG screening in those over 75 years old or at a high-risk of stroke.
In order to satisfy screening councils at a national level, the Wilson and Junger criteria have to be fulfilled. 4 These criteria concern the knowledge of the disease, the knowledge of the test, the treatment of the disease and the cost of the screening process. When screening is judged against these criteria, not enough is known about the development of complications or adverse events in AF detected by screening rather than by patients who are symptomatic and present to the medical system. There is also currently not enough agreement with regards to the treatment of the disease and even when we do know that patients have AF many remain untreated with anticoagulation.
“From real-world data it is known that people with asymptomatic AF are at risk,” explained Camm, “For example, 5,500 patients were followed for three years in the UK and the patients with AF had a worse prognosis than those without. Those that were treated well with anti-coagulants did far better than those who are not treated or those who are given anti-platelet therapy, predominantly aspirin.”
Recently, the US Preventive Service issued a document for discussion concerning the issue of screening for AF—the draft document does not recommend screening because the current evidence is insufficient to assess the benefits and harms of screening; the evidence is of poor quality, conflicting and the balance of benefit and harm cannot be determined. Their document says that there is no trial directly assessing the benefit of screening for AF with ECG on clinical outcomes and that by screening, such as carrying out ECGs, small to moderate harm may be caused to the patient. They recommend that more trials take place and that patients are divided into two groups, the screened population and the unscreened population, and that they are then followed in order to identify irreversible harm such as mortality, myocardial infarction, and stroke and compared to those patients that were not screened.
The risk of stroke
The ASSERT study looked at subclinical atrial tachyarrhythmia in patients that had recently received an implantable pacemaker device. They were monitored for three months and then followed for a mean of 2.5 years for the primary outcome of ischaemic stroke or systemic embolism. The risk of developing AF was 2.5 times greater in the 10% of the population that had atrial high rate events in the first three months and the absolute risk of ischaemic stroke and systemic embolisms was about 1.7% per year for this population.
“That is about the lowest risk that you could consider using warfarin or a vitamin K antagonist because of possible haemorrhagic complications associated with VKA anticoagulation. But with novel oral anticoagulant therapy we can go down to a risk of 1% or less. It might be well worth considering an anti-coagulation strategy in patients of this sort but you will see various complications,” said Camm.
It is important to understand the temporal relationship between AF and any stroke or embolic event. A study by Turakhia et al looked at this.5 They took 9,850 patients from the Veterans Administration who had Medtronic devices with the CareLink system. Out of the patients, 187 had eligible strokes and Turakhia et al looked at the likelihood of stroke if AF had occurred and found that if AF occurred it was associated with an 18-fold risk of stroke. They also looked at the likelihood of stroke within a given time if AF had occurred—the chance of stroke was highest in the five to 10 days after an AF episode.
European guidelines suggest that patients with short episodes of asymptomatic AF discovered on an implantable device which monitors continuously should be further assessed by seeking ECG confimraitn and then treating according to guidelines. The decision to anticoagulate should be made on the basis of the burden of AF and the thromboembolic risk.
There are clinical trials currently underway dividing patients that have atrial high rate events in two groups, one will be anticoagulated with a NOAC and the other will have usual care. These large trials will eventually tell us how to effectively treat and care for these patients.
- Lowres N, et al. Thrombosis and Haemostas 2014;111:1167–1176
- Svennberg E, et al. Circulation 2015;131:2176-84
- Julian P.J. Halcox et al. Circulation. 2017;136:1784-1794
- G Wilson and G Jungner; Principles and Practice of Screening for Disease, WHO, 1968
- Turakhia M, et al. Circ Arrythm Electrophysiol 2015;8:1040