Atrial fibrillation (AF) is a major cause of hospitalisation, morbidity and mortality. It is associated with an almost two-fold risk of death and an almost five-fold increase in the risk of stroke. Moreover, AF aggravates heart failure and, in turn, heart failure promotes AF. The burden of AF is projected to increase in the future as a result of demographic changes, improved survival following adverse cardiovascular events and associated comorbidities. However, the true epidemiology and impact of the rising incidence of AF on healthcare systems have not been determined so far, since the use of monitoring systems currently available for the detection of AF is limited in most patient groups and especially in the general population.
Symptoms are insensitive and non-specific for the identification of AF. Studies1 on antiarrhythmic drugs showed that 70% of AF episodes are not noticed by patients. Studies using the Holter monitor (Mortara Instrument, Inc., Wisconsin, US) demonstrated that asymptomatic episodes are 10–12 times more frequent than symptomatic events, yet patients who are asymptomatic may still be at increased risk of stroke and mortality and may also suffer from comorbidities, such as heart failure, hypertension and diabetes. Acute stroke is a common first presentation of patients with AF, given that arrhythmia often develops asymptomatically. Significantly, the available data indicate that paroxysmal AF carries the same risk of stroke as persistent or permanent AF. Therefore, improved technologies for the early detection of this growing epidemic are imperative.
Non-invasive Monitoring for the Detection of Atrial Fibrillation
AF is commonly detected by routine electrocardiogram (ECG) or 24–48 hour ambulatory Holter monitor ECG recordings. These set-ups are predominantly limited by relatively short or intermittent periods of sampling and patient compliance issues (such as skin irritation or interference with daily activities).2 The diagnostic yield of recently introduced, extended monitoring methods, including trans-telephonic ECG transmission, seven-day Holter monitoring and 30-day cardiac event recording, is still limited. The sensitivity of these systems ranges from 31–70%, while the negative predictive value range is even lower (21–65%). Newer methods, using automated detection algorithms, could enhance monitoring capacity and outsource long-term Holter monitoring from the hospital – but, in reality, at a relatively high cost.
Of note, there is a clear relationship between the duration of monitoring and the diagnostic yield.3 Ziegler et al.4 demonstrated that both the identification of patients with AF and the assessment of AF burden improve as the frequency or duration of monitoring increases and as the patient’s AF burden increases. This observation has important implications for pharmacological and interventional therapeutic approaches intended to protect patients from AF and related cardiovascular events, since AF is associated with an almost two-fold risk of cardiovascular morbidity and mortality. Thus, continuous monitoring, using relatively simple devices for the reliable detection of AF irrespective of symptoms, could improve the management of AF.
Rhythm Monitoring Using Implantable Therapeutic Devices
Device-based therapies – including pacemakers, implantable cardioverter defibrillators (ICDs) and cardiac resynchronisation therapy defibrillators (CRT-D) – are capable of rapid identification, recording and transmission of electrographic data. Continuous monitoring modalities from implantable devices have provided substantial insights into the characteristics of AF.
In a MOST subanalysis, atrial high-rate episodes of AF detected by pacemakers of ≥5 minutes’ duration were associated with premature death. Israel et al.5 found that 38% of patients with pacemakers who experienced episodes of AF lasting longer than 48 hours were completely asymptomatic. Using pacemaker memory with stored atrial electrograms as a continuous monitor, recurrence of AF was detected over 18 months in 88% of patients seemingly under successful ‘rhythm control’ by optimised antiarrhythmic drug therapy. Therefore, studies relying on infrequent ECG recordings may severely underestimate the AF recurrence rate and overestimate the success of antiarrhythmic drugs. It has been demonstrated that only continuous measurement of atrial tachyarrhythmia burden can provide full disclosure of rhythm control in patients over a long follow-up period.
Ricci et al.6 detected AF in 42 of 166 patients using implanted pacemakers or ICDs over a period of 18 months. Only clinically relevant AF detections (first-detected or persistent AF; AF for >10% of the time on at least five consecutive days) were regarded. Interestingly, AF was not known to exist before device implantation in more than half of these patients. Confirming previous observations, 73% of AF episodes were completely asymptomatic and not perceived by patients. Through remote monitoring technology,7
AF detection was immediately transmitted to the institution following the patient. Thus, AF detection occurred much earlier than with routine follow-up, enabling the treating physician to start therapeutic interventions such as anticoagulation, antiarrhythmic drug modification or cardioversion an average of 148 days earlier.
In addition, long-term, continuous Holter monitor function, using an implantable device with accurate atrial tachyarrhythmia detection, currently provides the most accurate assessment of the efficacy of catheter ablation for AF. This is significant, since catheter ablation has emerged as a promising modality that may in the future allow discontinuation of oral anticoagulation after a successful procedure in selected patients. This has not been demonstrated to date; however, evaluation of the efficacy of long-term therapy, based on freedom from tachyarrhythmia recurrence as documented by periodic ECGs and ambulatory Holter monitoring, may be inadequate because of prolonged and unpredictable arrhythmia-free intervals between episode clusters. Remote monitoring might prove useful as a triage tool for an individualised follow-up schedule, which could improve the management of patients with arrhythmia (while also containing costs and workload for cardiologists). This is important, as the extended monitoring capabilities of therapeutic implantable devices (pacemakers or ICDs) provide a more complete disclosure on the effect of therapeutic interventions such as radiofrequency ablation and pharmacological drug treatment. This has been demonstrated by our group8–10 and others.2,5
The IMPACT study is the first prospectively planned, multicentre, randomised trial (including patients with implanted dual-chamber ICD and CRT-D devices) designed to test the hypothesis that the initiation and withdrawal of oral anticoagulant therapy, guided by continuous ambulatory monitoring of cardiac electrical activity, improves clinical outcomes by reducing the combined rate of stroke, systemic embolism and major bleeding relative to conventional management. The study is designed to optimise therapy by protecting patients against thromboembolism proximate to episodes of documented atrial high-rate episodes of AF, while reducing the risk of bleeding potentiated by anticoagulation when AF is not present.
The temporal burden of and freedom from tachyarrhythmic episodes, requiring initiation or cessation of anticoagulant therapy, are governed by individual patient risk stratification using the Congestive heart failure, hypertension, aged ≥75 years, diabetes, stroke/transient ischaemic attack [TIA]/hromboembolic events [TE] (CHADS2) score (predicting the risk of stroke in patients with AF based on age and the presence of: congestive heart failure; hypertension; diabetes mellitus; and prior stroke or ischaemic attack), which has been incorporated into most clinical practice guidelines and performance measures. It is hoped that the results of the IMPACT study will: firstly, help define the clinical utility of wireless remote cardiac rhythm surveillance; and, secondly, prove useful in guiding anticoagulant therapy in some patients at risk of stroke in the future.
Leadless Implantable Cardiac Monitors for the Detection of Atrial Fibrillation
Only a limited number of patients within the population, with a history or at risk of AF, need to be treated using device-based therapies. Therefore, continuous monitoring in order to detect and/or quantify AF in these patients is not yet possible, limiting the diagnostic and therapeutic possibilities in this high-risk population.
Recently, the first leadless implantable cardiac monitor (ICM) with AF detection capabilities was validated.11 The Reveal XT Performance Trial (XPECT)11 tested an algorithm dedicated to AF detection and incorporated into a subcutaneous ICM already recommended by current guidelines,12 jointly produced by the Task Force for the Diagnosis and Management of Syncope, European Society of Cardiology, the European Heart Rhythm Association, the Heart Failure Association and the Heart Rhythm Society, for the evaluation of unexplained syncope. This ICM also features detection algorithms for bradyarrhythmias and ventricular tachyarrhythmias. In a validation study, the dedicated AF detection algorithm reliably detected the presence or absence of AF and the AF burden was accurately quantified.
The sensitivity, specificity, positive predictive value and negative predictive value in the XPECT trial were approximately 96%, 85%, 79% and 97%, respectively.2 Considering the potential complications of AF, high-sensitivity AF detection should be the primary request for any AF detection algorithm.
The AF burden measured in this study using the ICM fitted very well with the reference value derived from Holter monitor recordings. The overall accuracy of the ICM for detecting AF was 98.5%. Importantly, no data exist on the implementation of such devices in the clinical routine of AF monitoring. However, the published performance figures are high relative to conventional, non-invasive monitoring methods. This ICM is, therefore, a potentially promising tool for the detection of AF independently of symptoms. In the future, this monitoring approach could guide individualised therapeutic concepts beyond the currently available risk scores (e.g. CHADS2, or congestive heart failure/LV dysfunction, hypertension, aged ≥75 years, diabetes, stroke/TIA/TE, vascular, disease (prior myocardial infarction [MI], peripheral arterial disease [PAD] or aortic plaque), aged 65–74 years, sex category (i.e. female gender) (CHA2DS2-VASc), predicting the risk of stroke and thromboembolism in patients with AF based on gender, age and the presence of: congestive heart failure; hypertension; prior stroke, ischaemic attack or thromboembolism; vascular disease; and diabetes). In addition, it might foster greater integration between clinical care and research.
Clinical studies using the ICM are already under way. The Study of Continuous Cardiac Monitoring to Assess Atrial Fibrillation After Cryptogenic Stroke (CRYSTAL AF) study will assess the benefit of continuous monitoring using an ICM compared with an optimal standard of care in patients with cryptogenic stroke. In addition, patients with AF fitted with an ICM are already included in a multicentre, international registry – INSIGHT (R)XT – to further address the clinical importance and usefulness of this device for AF management in different subgroups. However, further prospective studies focusing on rhythm control are necessary to demonstrate the ultimate clinical value of ICMs in special subgroups of AF patients: firstly, before and after AF ablation procedures; and, secondly, during pharmacological treatment. In summary, the leadless ICM is a promising tool, which might in future improve the management of AF through the advanced detection and characterisation of cardiac rhythm disorders, in turn guiding specific therapy more effectively.
Conclusion and Future Clinical Implications
Extended Holter monitor recordings are essential to provide an accurate diagnosis and to inform subsequent therapeutic decisions in the management of AF. New, non-invasive systems could allow outsourcing of long-term Holter monitoring from the hospital, potentially improving the extent and recording results of Holter monitoring. Subcutaneous implantable systems, with automated detection algorithms for atrial arrhythmias, represent another opportunity for long-term rhythm monitoring, facilitating not only an objective evaluation of the efficacy of different therapies but also the advancement of therapy itself in the future.