Paroxysmal atrial fibrillation after ischemic stroke: how

Review

THEMED ARTICLE y Stroke For reprint orders, please contact [email protected]

Paroxysmal atrial fibrillation after ischemic stroke: how should we hunt for it? Expert Rev. Cardiovasc. Ther. 11(4), 485–494 (2013)

Azmil H Abdul-Rahim* and Kennedy R Lees Medical Sciences, Gardiner Institute, Western Infirmary, 44 Church Street, Glasgow, G11 6NT, Scotland, UK *Author for correspondence: Tel.: +44 141 211 1861 Fax: +44 141 211 2895 [email protected]

Paroxysmal atrial fibrillation (PAF) is a consideration in patients presenting with ischemic stroke despite sinus rhythm on admission ECG. Once PAF is identified, secondary prevention with anticoagulation prevents further stroke. Various methods are available to detect PAF, which is often asymptomatic. The evidence to guide our choice of detection strategies in stroke patients is weak. Current data suggest that prolonged monitoring increases the likelihood of PAF detection. Further research is required to establish the utility of various detection methods and identify their optimum timing, duration and cost–effectiveness. Newer technologies allow extremely brief paroxysmal episodes of AF to be detected, shorter than the traditional 30 s threshold required for diagnosis. However, risk associated with these very brief and hard to detect PAF episodes may be low. Here, the authors provide a broad review of the comparative utility and practicality of available detection methods, discuss ways to optimize outcome and highlight the current and future implications for detecting PAF in stroke patients. Keywords: atrial fibrillation • cardiac embolism • detection • diagnostic methods • paroxysmal • screening • stroke • stroke prevention

Atrial fibrillation (AF) is a common and treatable cause of ischemic stroke [1,2] . Strokes caused by AF have higher morbidity and mortality compared with most other etiologies [3–5] . Once AF is identified, treatment with anticoagulation reduces the annual risk of recurrent stroke in patients with AF by two-thirds [6] . Thus, in patients presenting with ischemic stroke, AF is an important risk factor to recognize. AF is often paroxysmal and asymptomatic, which can easily escape diagnosis [7] . Paroxysmal AF (PAF) is traditionally defined as recurrent AF (≥2 episodes, each lasting more than 30 s) that terminates spontaneously within 7 days [8–10] . PAF confers the same elevated risk as sustained AF. Up to 23% of the patients with cryptogenic ischemic stroke had PAF with long-term outpatient monitoring [11–13] . Newer anticoagulation drugs, such as dabigatran and rivaroxaban, promise to expand the arsenal of secondary stroke prevention caused by AF, but the benefits of these drugs can be accomplished only if the underlying PAF is identified [14–16] . Therefore, identifying the optimum method of screening stroke patients with PAF is of great importance.

www.expert-reviews.com

10.1586/ERC.13.21

Nevertheless, there is considerable uncertainty on the ‘gold standard’ approach to detect PAF post-stroke. The latest consensus guidelines from the American Heart Association/American Stroke Association for prevention of stroke in patients with acute ischemic stroke recommend at least 24-h cardiac monitoring but do not specify a method [17] . The European Stroke Organization recommends continuous ECG monitoring for the same period of time, with the addition of 24-h Holter ECG in selected patients, which is commonly performed during hospitalization [18] . The NICE UK and Scottish Intercollegiate Guidelines Network consensus statements on management of patients with stroke offer no specific g uidance on how to detect PAF [19,20] . In this review, the authors provide an overview of the various methods to detect PAF in post ischemic stroke patients and the clinical evidence of their comparative utility and practicality, discuss the uncertainties regarding the gold standard to screen stroke patients for PAF, highlight the current thoughts to optimize detection outcome and identify areas for promising research and developments.

© 2013 Expert Reviews Ltd

ISSN 1477-9072

485

Review

Abdul-Rahim & Lees

Methods to detect PAF in ischemic stroke patients History & clinical examination

AF is an arrhythmia that presents in approximately 1–2% of the general population [21,22] . It is associated with symptoms such as palpitation, chest pain, breathlessness and presyncope or syncope. The authors frequently ask stroke patients about these symptoms to elicit the possibility of underlying AF. However, many patients with AF are asymptomatic and such symptoms have poor correlation with episodes of cardiac arrhythmia [23–26] . A notable proportion of affected patients are unaware that they have AF, even though the diagnosis was previously documented in their medical notes [27] . Interestingly, self-reported diagnosis of AF by patients has been shown to carry a stroke risk similar to AF detected on routine ECG [28] . Collateral history from relatives and especially previous documentation are useful and may guide the diagnosis. AF is characterized by an irregularly irregular pulse. Pulse palpation to detect sustained AF in the general population is reasonably sensitive (87–97%), but not very specific (70–81%) [29] . However, the sensitivity of pulse palpation to detect PAF in stroke patients is unknown. Clinical signs will be absent in patients with PAF who are in regular sinus rhythm at presentation. Both positive and negative findings on pulse palpation routinely require further testing using ECGs. Clinical examination is therefore an imperfect tool to elicit the underlying diagnosis of PAF in post-stroke patients. Noninvasive detection methods ECG

On an ECG, AF is characterized by an absence of consistent P wave and presence of rapid oscillating waves that vary in size, shape and timing and are generally associated with an irregular ventricular response when atrioventricular conduction is intact [30] . Thus, a 12-lead ECG is the minimum evidence required to diagnose AF. It is readily available, noninvasive and relatively inexpensive. The main limitation is that it needs some degree of privacy to perform. New diagnosis of PAF based on a single ECG upon stroke presentation is achievable, but positive only in 2–3% of patients: Sposato et al. reported three new cases of PAF in 198 patients [31] , Rem et al. found six in 184 [32] and Francis et al. identified two in 64 [33] . Since an ECG gives an electrical snapshot of heart rhythm at a particular time, it is insufficient to detect paroxysmal arrhythmias. Interestingly, serial ECGs performed repeatedly (e.g., two- to four-times per day) should yield higher detection of PAF. A study by Douen et al. found that serial ECGs detected 15 new PAF cases in 143 patients (10.3%) during the first 3 days of stroke presentation [34] . Another study conducted in stroke patients with negative 24-h Holter demonstrated that transtelephonic ECG monitoring where an ECG was recorded and transmitted via telephone signals once daily for 1 month after stroke presentation, identified nine new cases of PAF in 98 patients (9.2%) [35] .

It is easily organized since stroke patients would normally be admitted to specialized stroke units instead of general wards. It is well established that patients with stroke have a better outcome with treatment in a specialized stroke unit compared with an unmonitored ward [36] . This survival benefit may arise from early recognition of malignant arrhythmias and poststroke complications. Hence, many centers are presently equipped with a telemetric monitoring system [37] . If the patient is correctly connected to the system, the continuous in-patient telemetry will be able to detect asymptomatic PAF relatively accurately. The main drawbacks are that the system is expensive and requires close attention from staff (i.e., demanding the reviewer’s competence in interpreting rhythm abnormalities in a busy ward environment). It also restricts patients’ movements and limits the maximum monitoring period. Some units require physical wire connection to the monitor; patients have to be disconnected from the monitor to undergo investigations and physical therapy, such as further imaging and physiotherapy sessions. Rem et al. reported six new cases of PAF in 184 stroke patients (3%) when using 48–72 h of continuous in-patient cardiac telemetry [32] . Lazzaro et al. found no case of PAF despite the same method and period of monitoring in 133 patients [11] . Conversely, a study conducted by Sposato et al. using prolonged continuous in-patient telemetry performed in the telemetry wards including the coronary care unit and the intensive care unit over the period of several weeks, yielded diagnosis of new PAF in 20 out of 110 stroke patients (18%) [31] . Approximately 70% of these new PAF diagnoses were made within the first 72 h of monitoring. Taking a step back, the findings from these studies seem conflicting. One could suggest a number of reasons that could explain this divergence, ranging from selection bias (which patients to be continuously monitored via telemetry), variable levels of staffing with diverse competency to recognize paroxysmal rhythm abnormalities and the use of various software algorithms to assist staff to identify arrhythmias. The introduction of a structured algorithm or automated analysis for telemetric data further improves the detection rate of new PAF in stroke patients. Kallmünzer et al. reported that continuous in-patient telemetry with a structured algorithm yielded a significantly higher detection rate for new PAF compared with serial ECG and standard telemetry without a structured reading, 6.6% (18 out of 271) versus 2.9% (8 out of 271) versus 2.6% (7 out of 271), respectively [38] . Rizos et al. also found that the addition of automated analysis of ECG data (i.e., unsupervised AF detection algorithm) increases the detection of new PAF in stroke patients, 38 new PAF cases in 496 stroke patients (7.7%), as compared with normal continuous ECG monitoring that only identified 27 new PAF in the same cohort (5.4%) [39] . However, competent ECG readers must verify positive automated analysis results because false positives can occur. Holter monitor

Continuous in-patient telemetry

Continuous in-patient cardiac telemetry appears to be the logical extension of serial ECGs to detect PAF in stroke patients. 486

The Holter monitor (ambulatory ECG) continuously records an ECG signal in a defined period of time, usually 24–48 h, although recordings up to 7 days are available in some centers [40] . The Expert Rev. Cardiovasc. Ther. 11(4), (2013)

Paroxysmal atrial fibrillation after ischemic stroke

modern Holter monitor is battery operated and digitally records two or three different leads to allow observation of different wave morphology from multiple vectors. The continuous ECG tracing is then downloaded and analyzed with a software algorithm to assist the reviewer in identifying any rhythm abnormalities. This eliminates: dependence on ward staff to review the monitor periodically, eliminates alarm sensitivity as in continuous in-patient cardiac telemetry and patient participation from the transmission. Many devices nowadays also include patient-activated event markers to allow increased correlation between symptoms and rhythm abnormalities [41] . Patients are also asked to keep a symptom diary to aid the investigation. Current devices are relatively small (size: 70 × 95 × 20 mm; weight: 190 g) [41] , but some patients may still find these cumbersome to wear for prolonged periods. Stroke patients may be less able to cooperate with the monitoring procedure due to any neurological deficit sustained. For example, stroke patients may find it challenging to change the Holter monitor’s leads daily by themselves or to ensure constant adequate leadsto-body contact as compared with healthy adults. The greatest limitation is the short recording period, typically 1–2 days. The absence of real-time data analysis can also be an important clinical drawback. In a comparative study by Lazzaro et al., Holter monitoring was found to be more reliable in detecting new cases of PAF in ischemic stroke patients compared with a continuous in-patient cardiac monitor, 6% (6 out of 133) versus 0% (0 out of 133, respectively) [11] . However, when comparing Holter monitor with continuous monitoring with intermittent analysis by trained staff, no additional benefit was demonstrated [42] . It seems inevitable that prolonged monitoring using a Holter device is likely to maximize detection of paroxysmal episodes of AF. Evidence shows that each additional 24-h period of monitoring leads to new diagnosis of PAF in a further 2–4% of stroke patients [11,43] . A 7-day Holter monitoring period yields diagnosis of new PAF in 28 out of 224 stroke patients (12.5%) [44] . Prolongation of Holter monitoring increases the detection of PAF in patients with ischemic stroke and the standard 24-h period of monitoring would therefore underestimate the prevalence of PAF in this group of patients. Encouragingly, a 7-day period of outpatient Holter monitoring to detect new PAF after ischemic stroke is also found to be cost effective [45,46] . Intermittent external patient or event-activated recorders

Intermitted patient or event-activated recorders are also referred to as event monitors or external loop recorders (ELRs). They may be used for prolonged periods (e.g., 4–6 weeks), with episodic recordings triggered by patients or automatically by the device. The device may be worn continuously or applied to the chest wall at the time of recording. Some ELRs record data only when activated by the patient. The latest devices have automatic triggers that recognize brady or tachyarrhythmias [41] . The main advantage of ELRs over Holter monitors is that they are smaller and allow longer periods of ECG monitoring to detect PAF. Only brief ECG tracings around the defined events are recorded when activated: it may be difficult to establish the duration and overall www.expert-reviews.com

Review

burden of PAF episodes. ELRs can also provide nearly real-time ECG data when the patient transmits recent symptomatic events via telephonic or wireless internet transmission. The main disadvantage of these devices arises from the need for patient compliance to activate the device during symptomatic episodes, unless an automatic trigger is programmed or built into the device. A considerable percentage of patients do not complain with continuous application of the device due to local skin irritation (with the leads) and poor skin contact during physical activity [40,41] . Stroke patients may find it especially awkward to wear and operate ELR devices. This arises from participation in physical therapy such as physiotherapy and occupational therapy sessions or because they already have major problems with dressing or with instruments due to physical or intellectual deficits. With any device that is not worn continuously (e.g., a handheld device that needs to be applied to chest wall at the time of symptoms), short paroxysmal episodes that terminate before the device is applied or asymptomatic AF will be missed. Nevertheless, ELRs offer ample data storage since only the activated events are recorded rather than continuous tracings of normal rhythm and so deliver the benefit of extended monitoring. In two separate comparative studies, 24-h Holter monitor failed to detect new PAF episodes, while 4–7 days duration of ELR monitoring yielded detection of new PAF in 6–8% of stroke patients: Barthélémy et al. reported four new cases of PAF in 52 patients while Jabaudon et al. found five new AF in 88 patients [47,48] . The detection of new PAF cases is further increased to 14% (18 out of 27) of stroke patients, when 7-day monitoring using ELRs was repeated at 3 months and 6 months after the onset of ischemic stroke [49] . Mobile cardiac outpatient telemetry

Mobile cardiac outpatient telemetry (MCOT) offers home-based, prolonged real-time cardiac monitoring. This was developed to combine the benefits and overcome the drawbacks of Holter monitors and standard ELRs. Patients wear a small transmitting device similar in size to standard ELRs. The device transmits ECG data either continuously in real-time or at regular intervals to a central laboratory, where reviewers can then analyze the rhythms. MCOTs solve the issue of data storage and allow lengthier monitoring periods (up to 30 days) to detect PAF in poststroke patients. The system is usually equipped with software that detects rhythm abnormalities automatically to assist data reviewers, who over-read the tracings in the laboratory. Any patient-activated data are also transmitted and annotated on the recordings to aid symptom–rhythm correlations. MCOT, however, is limited by cost and real-world limitations similar to ELRs including loss of signal due to inappropriate system use, signal artifact and patient noncompliance for a variety of reasons [41] . A consecutive series of 56 patients with cryptogenic stroke, collected by Tayal et al. using 21-day MCOT monitoring yielded detection of new PAF in 23% (13 out of 56) of the patients [12] . Similarly, a retrospective study by Elijovich et al. using 30-day MCOT monitoring resulted in 20% (4 out of 20) PAF detection rate in cryptogenic stroke patients [13] . The high rate of PAF 487

Review

Abdul-Rahim & Lees

detection by MCOT in the cryptogenic stroke population may reflect the extended monitoring period compared with other detection methods. Invasive detection methods Implantable loop recorders

It is well accepted that the longer a patient is monitored, the greater the likelihood of detecting paroxysmal AF [29,50,51] . Implantable loop recorders (ILRs) are small leadless devices (typically size 60 × 20 × 8 mm) that record ECG signals and are implanted subcutaneously to overcome patient compliance and to allow a lengthier monitoring period [40] . The ILR device can be activated automatically when it detects arrhythmias, especially PAF, or may be triggered by the patient when he places an activator over the device. The ECG data can then be downloaded and reviewed appropriately. The main constraints of ILRs are the need for implantation: they require an invasive procedure with associated risk of complications such as infection, bleeding, discomfort etc.. ILRs cost more than other noninvasive methods. ILRs also carry the risk of false-positive detections because of premature atrial or ventricular complexes, artifacts and sinus arrhythmias [40] . In a prospective study on 24 patients with cryptogenic stroke, Dion et al. found no case of new PAF detected after 14.5 months of monitoring period using ILRs [52] . However, the criticism is that the ILR software used for this study was programmed to detect only PAF with a rapid ventricular response rate (more than 165 beats/min). Therefore, many older patients with PAF may have been missed. Intra-cardiac recording systems

Pacemakers and implantable cardioverter–defibrillators can also be used as continuous monitoring devices. These devices have battery life from 5 to 12 years, depending on the device type and therapy delivered. When programmed appropriately, these devices can identify PAF with more than 95% sensitivity and specificity [41,53] . They also offer therapy depending on the device type. However, they carry similar limitations as ILRs, but with greater risk of complications. One study reported that detection of PAF in stroke patients who had pacemakers or cardiodefibrillators found a new diagnosis of PAF in 45 out of 163 patients (28%) over 1 year of follow-up [54] . However, the study may have included patients that were already at high risk for AF, due to the confounding cardiac diseases that necessitated implantation of such devices. It is difficult to extrapolate this result to a wider ischemic stroke population. Comparison between detection methods

Concerns have been raised about the low agreement between numerous studies to assess optimal methods, timing and duration of monitoring to detect PAF in poststroke patients. The interval between stroke onset and the timing of monitoring to detect PAF ranged from 72 h to 3 months after stroke onset [12,31,35,47,52] . Meanwhile, the duration of monitoring, regardless of the detection method used, ranged from 22 h to 14 months [11–13,43,47,48] . 488

Hence, it is difficult to compare like-to-like between these studies. Nevertheless, two principal observations can be made, regardless of the detection method used. First, a higher incidence of PAF was reported in studies that evaluated earlier monitoring than those performed at a later stage after stroke onset. Second, a higher incidence of PAF was found in studies that monitored patients for a longer duration. Clinicians also face challenges on deciding which method to adopt to detect PAF in stroke patients, since there is no clear agreement in the literature. One study found that Holter monitor is better in detecting new PAF in stroke patients as compared with continuous in-patient cardiac monitoring [11] , whereas another study claimed the opposite [39] . A separate study confirmed the superiority of continuous in-patient cardiac monitoring over Holter monitoring and this superiority was increased when a structured algorithm was introduced to the system [38] . Interestingly, another study reported that the use of Holter monitor does not provide any additional benefit in comparison with continuous telemetry monitoring with intermittent analysis by trained staff [42] . These differences highlight the practical difficulties and lack of an established gold standard to detect PAF after stroke. Thus, the evaluation of the definitive efficacy of different PAF detection methods in this population remains difficult. Optimizing the detection outcome

Despite the uncertainties in the type and duration of m onitoring, detection of PAF is important to facilitate effective secondary prevention with anticoagulation after stroke. The effectiveness of various detection methods described above would be improved if more intensive screening were focused on patients who are at high risk of developing PAF. Older age and the presence of multiple vascular risk factors are strongly associated with the presence of AF in the general population [30] . However, we lack clinical predictors to risk stratify stroke patients with risk of undiagnosed PAF with enough certainty to allow directed screening with different detection methods. A large retrospective study of stroke patients using data from four different trials in the Virtual International Stroke Trials Archive [55,56] suggested that advancing age, women, patients with congestive heart failure and those without hypertension are at increased risk of PAF [57] . Thus, more targeted screening for PAF may be focused on these groups of patients. Nevertheless, further research is needed to validate these clinical predictors in stroke patients and identify other predictors that can eventually be incorporated into a useful risk score to detect PAF, similar to the CHADS2 and CHADSVASc scoring systems to risk stratify AF patients for future stroke risk. We must consider that once a stroke or transient ischemic attack has already occurred, it may still be cost effective and clinically important to detect PAF even among the lower-risk strata. Other novel predictors are being identified to optimize detection of PAF in ischemic stroke patients. For example, one study demonstrated that frequent atrial ectopics identify stroke patients who are likely to have or develop PAF [49] . Patients with 70 atrial ectopics or more in the first 24 h of telemetry had a 26% (13 out of 60) rate of Expert Rev. Cardiovasc. Ther. 11(4), (2013)


new PAF in the subsequent ELR monitoring period. Meanwhile, patients with fewer than 70 atrial ectopics in the first 24 h had only a 6.5% (5 out of 77) rate of new PAF in the following ELR monitoring period. From a different perspective, the findings also support the general thought that complementary strategies to detect PAF such as long-term rhythm monitoring follow-up must follow any negative initial cardiac monitoring in patients with ischemic stroke of unknown etiology. Other findings such as a dilated left atrium on echocardiograms or embolic appearance on neuroimaging may also support the presence of PAF. A more cutting-edge example is the application of brain natriuretic peptide (BNP) as a predictive biological marker of delayed AF in stroke patients. High plasma BNP level is found in stroke patients who are subsequently shown to have AF [58,59] . However, more research is needed to agree on the optimal cutoff BNP level to ensure its robustness as a predictor of PAF in this population. Ultimately, the use of biomarkers will improve the diagnosis of cardioembolic sources of ischemic stroke, thus rapidly guiding optimal diagnostic tests to detect PAF and accelerating the start of effective secondary stroke preventions. Promising areas for future developments

Despite interesting findings from various observational studies, we still lack evidence from large, randomized, multicenter studies to guide our choice of monitoring methods and the ideal timing and duration of monitoring required to detect PAF cost-effectively in ischemic stroke patients. There is also the important question of whether these types of cardiac monitoring would offer extra benefits over routine clinical follow-up. Effective monitoring methods should be able to detect new PAF that would otherwise remain undiscovered in stroke patients, or at the very least, reduce the time between stroke onset and detection of PAF. Once PAF is detected, secondary prevention therapy with a nticoagulant medication can be commenced swiftly. Our understanding of the benefits of anticoagulation therapy for secondary stroke prevention is based on PAF diagnosed using conservative definitions and traditional detection methods [6] . Several recent studies using newer and prolonged detection methods as discussed above, reported very brief episodes of PAF, much shorter than the traditional 30-s threshold of diagnosis [11,13,31,34,35] . In these studies, detection of PAF after stroke with such new and sensitive methods led to initiation of oral anticoagulation therapy. However, the issue of how much PAF burden is clinically important in conferring future stroke risk remains unresolved. The risk associated with extremely brief and hard to detect PAF may be low. False-positive recordings of very brief PAF due to atrial oversensing detected using new automated methods may lead to overtreatment and may harm our patients in the long run. No data from randomized clinical trials (RCTs) are available to guide appropriate treatment for these very short episodes of PAF or to inform their prognosis. On the other hand, these very brief episodes of PAF may be clinically relevant in cryptogenic stroke patients. Thus, trials are needed to guide diagnosis and treatment approach in stroke patients with very brief episodes of PAF and also to consider the threshold incidence below which detection is not worthwhile. www.expert-reviews.com

Review

An example of an exciting development in this area is the ongoing CRYSTAL-AF study, a multicenter RCT that evaluates the diagnostic yields of long-term cardiac monitoring using an ILR compared with routine clinical follow-up [60,101] . Four hundred and fifty cryptogenic stroke patients confirmed with negative initial 24-h ECG monitoring will be enrolled in the study. The study promises answers on the effectiveness of outpatient monitoring to detect PAF beyond routine clinical follow-up. The result of this study will also begin to address key issues surrounding newer detection methods for PAF: is it better to monitor continuously for a longer duration, even if it means possibly overlooking very brief episodes of PAF, or is it better to monitor much shorter duration with novel methods that are capable of detecting even the shortest of PAF episodes? The study is expected to complete in May 2013 and is anticipated with great interest. Nevertheless, more randomized controlled trials are needed to assess other types of monitoring methods against routine clinical follow-up, to elicit the ideal methods and duration of cardiac monitoring in poststroke patients. In addition, CRYSTAL-AF may tell us how to detect short episodes of PAF but may be underpowered to tell us how to treat patients with low burden, brief runs of PAF. Another factor that needs to be considered is the cost–effectiveness of various strategies. Each of the detection methods have their own pros and cons, but cost will always affect the final decision. Kamel et al. have demonstrated that prolonged outpatient Holter monitoring (for at least 7 days) for secondary stroke prevention is cost effective [45] . Hence, more cost-effective analyses should also be encouraged, along with RCTs, to assess the effectiveness of other detection methods. Since the optimal duration and methods of monitoring to detect PAF is unknown, a more targeted screening of stroke patients who are at high risk of PAF may certainly improve the yield of detection, regardless of monitoring strategies adopted. It may also improve the cost–effectiveness of the diagnostic method chosen. A more directed screening will require a huge effort to validate clinical predictors of PAF including demographics, clinical history and presentations, ECG and imaging results and perhaps the application of the cutting-edge biomarkers. Ideally, these clinical predictors should eventually be incorporated into a useful and robust clinical predictive score for PAF, allowing presumptive early secondary prevention for high-risk patients while waiting for confirmatory detection of PAF. Apart from methods to detect PAF as discussed above, there are emerging devices that can be used for screening for PAF in the general population in home settings. Examples of these are the blood-pressure monitors that have a facility to detect PAF and a finger probe similar to that used for pulse oximetry, which uses the principle of photoplethysmography [61–63] . These devices could be used by patients to self-screen or by healthcare professionals to screen patients for PAF opportunistically. The devices benefit from the ability to modify thresholds of detection to achieve maximum sensitivity to detect PAF. However, their application and practicality in ischemic stroke patients is debatable. The current practice of secondary stroke prevention is to start oral anticoagulation therapy once PAF is confirmed. 489

Review

Abdul-Rahim & Lees

Table 1. Different modalities to detect paroxysmal atrial fibrillation. Modality

Advantages

Disadvantages

Medical history

Basic skills for physicians

Poor symptom and rhythm correlation Needs investigation to confirm suspicion

Clinical examination

Basic skills for physicians

May miss paroxysmal rhythm abnormalities Needs investigation to confirm suspicion

Serial ECG

Readily available Relatively inexpensive Detects asymptomatic event during particular ECG recording

Time consuming Needs some degree of privacy to perform May miss paroxysmal rhythm abnormalities

Continuous in-patient telemetry

Accurate diagnosis Continuous recording Detects asymptomatic events Additional structured algorithms or automated analysis to assist detection

System is costly Demands in-patient monitoring (staff input) Restricts patient movements

Holter monitor

Simple to use Continuous recording Detects asymptomatic events Newer devices have patient-activated event markers to allow symptom–rhythm correlation (but limited)

Short monitoring period Patient encouraged to keep symptom diary

External loop recorder

Longer monitoring period Allows patient input for symptom–rhythm correlation

Does not detect asymptomatic events Patient compliance required

Mobile cardiac outpatient telemetry

Longer monitoring period Continuous monitoring Detects asymptomatic events

System set-up is costly Patient compliance required

Implantable loop recorder

Lengthier monitoring period Detects asymptomatic events Telephonic or internet-based transmission Software to assist detection and interpretation

Costly Associated risk of invasive procedure False-positive detections

Intracardiac recording (pacemakers and cardioverter– defibrillators)

Prolonged monitoring period Detects asymptomatic events High sensitivity and specificity Offers therapy

Costly Greater risk of associated complications (compared with implantable loop recorders)

Noninvasive detection methods

Invasive detection methods

Anticoagulation therapy is routinely commenced as soon as the patient is both medically and neurologically stable. If the drug of choice is warfarin, this is often started on days 2 or 3 poststroke, achieving therapeutic anticoagulation by days 7–10. This is to minimize the risk of worsening secondary hemorrhagic transformation, if any, that frequently occurs between days 1 and 4 poststroke, particularly following treatment with intravenous thrombolysis [64,65] . From this perspective, detection of PAF too early may not be essential, as the decision to start anticoagulation therapy will be deferred until the patient is more stable. However, patients with AF are also proven to be at risk of early recurrent ischemic stroke even after thrombolysis [66] . Early PAF detection would therefore ensure the swift start of anticoagulation therapy at the time of greatest recurrence risk. The optimal timing of diagnosis and treatment initiation remains uncertain. 490

Expert commentary

AF is a common and treatable cause of ischemic stroke. Identification of PAF as a cause of cardioembolic stroke significantly reduces the risk of recurrent stroke through effective secondary prevention (Table 1) . However, the evidence to guide strategies to detect PAF in ischemic stroke patients is unclear. The authors can conclude that the longer monitoring is carried out, the more cases of PAF are detected and that early detection is desirable (Figure 1) . However, there are at least three key questions that need to be addressed regarding detection methods of PAF in stroke patients. First, what is the ideal timing and duration of the chosen detection method? Second, who are the best candidates that would benefit from a prolonged detection process? Third, what is the prognostic significance of very brief episodes of PAF detected by these novel methods? Several lines of ongoing research promise to shed light on these questions, thus ensuring the best care for the patients. Expert Rev. Cardiovasc. Ther. 11(4), (2013)

Until further data are available, the authors believe that complementary strategies, such as outpatient long-term monitoring or prolonged ambulatory cardiac rhythm follow-up, should follow any negative initial in-patient monitoring in patients with ischemic stroke of unknown etiology. Outpatient cardiac monitoring with Holter monitors to detect PAF after stroke is cost effective but the optimal duration of monitoring is uncertain. Five-year view

Promising areas for future developments to optimize the detection of PAF in ischemic stroke patients have been extensively discussed above. The authors anticipate valuable results from the CRYSTAL-AF study, the first multicenter RCT to evaluate the effectiveness of outpatient monitoring using ILRs to detect PAF compared with routine clinical follow-up. It may also provide vital information on the timing episodes of PAF in stroke patients. The ongoing CMACS study will address similar issues but using the MCOT systems [102] . Nevertheless, more evidence is needed to assess other novel detection methods available and their cost–effectiveness because the recurrence risk may vary with type and duration of PAF detected. Regardless of the detection method used, a robust and validated risk scoring of clinical predictors of PAF in stroke patients could help us to direct screening towards higher risk patients and perhaps even allow presumptive early secondary prevention to be commenced while waiting for confirmatory detection of PAF. Ultimately, this could be part of a holistic approach to ischemic stroke management. Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial

Cumulative percentage of PAF detected


Review

Stahrenberg et al. [44] (Holter monitor)

Jabaudon et al. [48] (Serial ECG)

Schuchert et al. [67] (Holter monitor)

Douen et al. [34] (Serial ECG)

14 12 10 8 6 4 2 0 1

2

3 4 5 Time (days after stroke)

6

7

Figure 1. Detection rate of paroxysmal atrial fibrillation versus time after stroke. PAF: Paroxysmal atrial fibrillation.

conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript.

Key issues • Atrial fibrillation is a common and treatable cause of ischemic stroke. • Once paroxysmal atrial fibrillation (PAF) is identified, effective secondary prevention using anticoagulation medication prevents recurrent stroke. • PAF is often asymptomatic, making diagnosis difficult. Various detection methods to detect the abnormal rhythm are available. • The evidence to guide strategies to detect PAF in ischemic stroke patients is weak. • Further research is required to establish the utility of various PAF detection methods, the ideal timing and the duration of monitoring and their cost–effectiveness. • New technologies are emerging to optimize the detection of PAF in stroke patients. However, caution should be exercised when extrapolating the benefits of anticoagulation therapy for very brief episodes of PAF detected using new detection methods.

References Papers of special note have been highlighted as: • of interest •• of considerable interest 1

2

Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 22(8), 983–988 (1991). Marini C, De Santis F, Sacco S et al. Contribution of atrial fibrillation to


incidence and outcome of ischemic stroke: results from a population-based study. Stroke 36(6), 1115–1119 (2005). 3

Kimura K, Minematsu K, Yamaguchi T; Japan Multicenter Stroke Investigators’ Collaboration (J-MUSIC). Atrial fibrillation as a predictive factor for severe stroke and early death in 15,831 patients with acute ischaemic stroke. J. Neurol. Neurosurg. Psychiatr. 76(5), 679–683 (2005).

4

Karatas M, Dilek A, Erkan H, Yavuz N, Sözay S, Akman N. Functional outcome in stroke patients with atrial fibrillation. Arch. Phys. Med. Rehabil. 81(8), 1025–1029 (2000).

5

Sandercock P, Bamford J, Dennis M et al. Atrial fibrillation and stroke: prevalence in different types of stroke and influence on early and long term prognosis (Oxfordshire community stroke project). BMJ 305(6867), 1460–1465 (1992).

491

Review 6

7

8

9

10

11

Abdul-Rahim & Lees

Saxena R, Koudstaal P. Anticoagulants versus antiplatelet therapy for preventing stroke in patients with nonrheumatic atrial fibrillation and a history of stroke or transient ischemic attack. Cochrane Database Syst. Rev. 4, CD000187 (2004). Savelieva I, Camm AJ. Clinical relevance of silent atrial fibrillation: prevalence, prognosis, quality of life, and management. J. Interv. Card. Electrophysiol. 4(2), 369–382 (2000). Calkins H, Brugada J, Packer DL et al. HRS/EHRA/ECAS expert Consensus Statement on catheter and surgical ablation of atrial fibrillation: recommendations for personnel, policy, procedures and follow-up. A report of the Heart Rhythm Society (HRS) Task Force on catheter and surgical ablation of atrial fibrillation. Heart Rhythm 4(6), 816–861 (2007). McNamara RL, Brass LM, Drozda JP Jr et al.; American College of Cardiology; American Heart Association. ACC/AHA key data elements and definitions for measuring the clinical management and outcomes of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Data Standards (Writing Commitee to Develop Data Standards on Atrial Fibrillation). J. Am. Coll. Cardiol. 44(2), 475–495 (2004). Fuster V, Rydén LE, Asinger RW et al.; American College of Cardiology/American Heart Association Task Force on Practice Guidelines; European Society of Cardiology Committee for Practice Guidelines and Policy Conferences (Committee to Develop Guidelines for the Management of Patients With Atrial Fibrillation); North American Society of Pacing and Electrophysiology. ACC/AHA/ESC Guidelines for the Management of Patients With Atrial Fibrillation: Executive Summary A Report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines and the European Society of Cardiology Committee for Practice Guidelines and Policy Conferences (Committee to Develop Guidelines for the Management of Patients With Atrial Fibrillation) Developed in Collaboration With the North American Society of Pacing and Electrophysiology. Circulation 104(17), 2118–2150 (2001). Lazzaro MA, Krishnan K, Prabhakaran S. Detection of atrial fibrillation with concurrent holter monitoring and continuous cardiac telemetry following ischemic stroke and transient ischemic

492

attack. J. Stroke Cerebrovasc. Dis. 21(2), 89–93 (2012). 12

Tayal AH, Tian M, Kelly KM et al. Atrial fibrillation detected by mobile cardiac outpatient telemetry in cryptogenic TIA or stroke. Neurology 71(21), 1696–1701 (2008).

13

Elijovich L, Josephson SA, Fung GL, Smith WS. Intermittent atrial fibrillation may account for a large proportion of otherwise cryptogenic stroke: a study of 30-day cardiac event monitors. J. Stroke Cerebrovasc. Dis. 18(3), 185–189 (2009).

14

Patel MR, Mahaffey KW, Garg J et al.; ROCKET AF Investigators. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N. Engl. J. Med. 365(10), 883–891 (2011).

15

Prasad V, Kaplan RM, Passman RS. New frontiers for stroke prevention in atrial fibrillation. Cerebrovasc. Dis. 33(3), 199–208 (2012).

16

Connolly SJ, Ezekowitz MD, Yusuf S et al.; RE-LY Steering Committee and Investigators. Dabigatran versus warfarin in patients with atrial fibrillation. N. Engl. J. Med. 361(12), 1139–1151 (2009).

17

Adams HP Jr, del Zoppo G, Alberts MJ et al.; American Heart Association; American Stroke Association Stroke Council; Clinical Cardiology Council; Cardiovascular Radiology and Intervention Council; Atherosclerotic Peripheral Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups. Guidelines for the early management of adults with ischemic stroke: a guideline from the American Heart Association/American Stroke Association Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups: the American Academy of Neurology affirms the value of this guideline as an educational tool for neurologists. Stroke 38(5), 1655–1711 (2007).

•

18

The recommendations for the management of acute ischemic stroke from various organizations: American Heart Association/American Stroke Association, Scottish Intercollegiate Guidelines Network (SIGN) and NICE. Guidelines for management of ischaemic stroke and transient ischaemic attack 2008. Cerebrovasc. Dis. 25(5), 457–507 (2008).

•

Also describes the recommendations for the management of acute ischemic stroke from various organizations: American Heart Association/American Stroke Association, SIGN and NICE.

19

Scottish Intercollegiate Guidelines Network (SIGN). Management of patients with stroke or TIA: assessment, investigation, immediate management and secondary prevention. A national clinical guideline. SIGN Guideline No. 108 (2008).

•

Describes the recommendations on the management of acute ischemic stroke from various organizations: American Heart Association/American Stroke Association, SIGN and NICE.

20

National Institute for Clinical Excellence UK (NICE). Stroke: Diagnosis and initial management of acute stroke and transient ischaemic attack (TIA). NICE Clinical Guideline No. 68 (2008).

•

Describes the recommendations on the management of acute ischemic stroke from various organizations: American Heart Association/American Stroke Association, SIGN and NICE.

21

Mant J, Wade D, Winner S. Health care needs assessment: stroke. In: Health Care Needs Assessement: The Epidemiologically Based Assessment Reviews. Stevens A, Raftery J, Mant J, Simpson S (Eds). Radcliffe Medical Press, Oxford, UK, 141–244 (2004).

22

Camm AJ, Lip GY, De Caterina R et al.; ESC Committee for Practice Guidelines (CPG); Document Reviewers. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation: an update of the 2010 ESC Guidelines for the management of atrial fibrillation. Developed with the special contribution of the European Heart Rhythm Association. Eur. Heart J. 33(21), 2719–2747 (2012).

•• Describes the current European Cardiology Society recommendations and standard of care for atrial fibrillation. 23

Barsky AJ. Palpitations, arrhythmias, and awareness of cardiac activity. Ann. Intern. Med. 134(9 Pt 2), 832–837 (2001).

24

Mehall JR, Kohut RM Jr, Schneeberger EW, Merrill WH, Wolf RK. Absence of correlation between symptoms and rhythm in ‘symptomatic’ atrial fibrillation. Ann. Thorac. Surg. 83(6), 2118–2121 (2007).

25

Warfarin versus aspirin for prevention of thromboembolism in atrial fibrillation: stroke. Lancet 343(8899), 687 (1994).

Expert Rev. Cardiovasc. Ther. 11(4), (2013)


•

Describes the superiority of warfarin over aspirin in preventing stroke in atrial fibrillation.

26

Furberg CD, Psaty BM, Manolio TA, Gardin JM, Smith VE, Rautaharju PM. Prevalence of atrial fibrillation in elderly subjects (the Cardiovascular Health Study). Am. J. Cardiol. 74(3), 236–241 (1994).

27

28

29

30

31

32

33

34

35

Meschia JF, Merrill P, Soliman EZ et al. Racial disparities in awareness and treatment of atrial fibrillation: the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study. Stroke 41(4), 581–587 (2010). Soliman EZ, Howard G, Meschia JF et al. Self-reported atrial fibrillation and risk of stroke in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study. Stroke 42(10), 2950–2953 (2011). Harris K, Edwards D, Mant J. How can we best detect atrial fibrillation? J. R. Coll. Physicians Edinb. 42(Suppl. 18), 5–22 (2012). Atrial fibrillation: national clinical guideline for management in primary and secondary care. National Collaborating Centre for Chronic Conditions. NICE, UK (NICE). (2006). Sposato LA, Klein FR, Jáuregui A et al. Newly diagnosed atrial fibrillation after acute ischemic stroke and transient ischemic attack: importance of immediate and prolonged continuous cardiac monitoring. J. Stroke Cerebrovasc. Dis. 21(3), 210–216 (2012).

ischemic stroke patients to a stroke care monitoring unit versus a conventional stroke unit: a randomized pilot study. Stroke 34(1), 101–104 (2003). 38

Kallmünzer B, Breuer L, Hering C et al. A structured reading algorithm improves telemetric detection of atrial fibrillation after acute ischemic Stroke. Stroke 43(4), 994–999 (2012).

39

Rizos T, Güntner J, Jenetzky E et al. Continuous stroke unit electrocardiographic monitoring versus 24-hour Holter electrocardiography for detection of paroxysmal atrial fibrillation after stroke. Stroke 43(10), 2689–2694 (2012).

40

41

42

43

44

Rem JA, Hachinski VC, Boughner DR, Barnett HJ. Value of cardiac monitoring and echocardiography in TIA and stroke patients. Stroke 16(6), 950–956 (1985). Francis DA, Heron JR, Clarke M. Ambulatory electrocardiographic monitoring in patients with transient focal cerebral ischaemia. J. Neurol. Neurosurg. Psychiatr. 47(3), 256–259 (1984). Douen AG, Pageau N, Medic S. Serial electrocardiographic assessments significantly improve detection of atrial fibrillation 2.6-fold in patients with acute Stroke Stroke 39(2), 480–482 (2008). Gaillard N, Deltour S, Vilotijevic B et al. Detection of paroxysmal atrial fibrillation with transtelephonic EKG in TIA or stroke patients. Neurology 74(21), 1666–1670 (2010).

36

Organised in-patient (stroke unit) care for Stroke. Cochrane Database Syst Rev. 4, CD000197 (2007).

37

Sulter G, Elting JW, Langedijk M, Maurits NM, De Keyser J. Admitting acute


45

46

Seet RC, Friedman PA, Rabinstein AA. Prolonged rhythm monitoring for the detection of occult paroxysmal atrial fibrillation in ischemic stroke of unknown cause. Circulation 124(4), 477–486 (2011). Zimetbaum P, Goldman A. Ambulatory arrhythmia monitoring: choosing the right device. Circulation 122(16), 1629–1636 (2010). Gumbinger C, Krumsdorf U, Veltkamp R, Hacke W, Ringleb P. Continuous monitoring versus HOLTER ECG for detection of atrial fibrillation in patients with stroke. Eur. J. Neurol. 19(2), 253–257 (2012). Shafqat S, Kelly PJ, Furie KL. Holter monitoring in the diagnosis of stroke mechanism. Intern. Med. J. 34(6), 305–309 (2004). Stahrenberg R, Weber-Krüger M, Seegers J et al. Enhanced detection of paroxysmal atrial fibrillation by early and prolonged continuous holter monitoring in patients with cerebral ischemia presenting in sinus rhythm. Stroke 41(12), 2884–2888 (2010). Kamel H, Hegde M, Johnson DR, Gage BF, Johnston SC. Cost-effectiveness of outpatient cardiac monitoring to detect atrial fibrillation after ischemic stroke. Stroke 41(7), 1514–1520 (2010). Wachter R, Stahrenberg R, Gröschel K. Letter by Wachter et al regarding article ‘Cost–effectiveness of outpatient cardiac monitoring to detect atrial fibrillation after ischemic stroke’. Stroke 42(3), e36; author reply e37 (2011).

Review

ECG monitoring for the detection of atrial fibrillation and flutter after acute stroke and transient ischemic attack. Stroke 35(7), 1647–1651 (2004). 49

Wallmann D, Tüller D, Wustmann K et al. Frequent atrial premature beats predict paroxysmal atrial fibrillation in stroke patients: an opportunity for a new diagnostic strategy. Stroke 38(8), 2292–2294 (2007).

50

Vasamreddy CR, Dalal D, Dong J et al. Symptomatic and asymptomatic atrial fibrillation in patients undergoing radiofrequency catheter ablation. J. Cardiovasc. Electrophysiol. 17(2), 134–139 (2006).

51

Klemm HU, Ventura R, Rostock T et al. Correlation of symptoms to ECG diagnosis following atrial fibrillation ablation. J. Cardiovasc. Electrophysiol. 17(2), 146–150 (2006).

52

Dion F, Saudeau D, Bonnaud I et al. Unexpected low prevalence of atrial fibrillation in cryptogenic ischemic stroke: a prospective study. J. Interv. Card. Electrophysiol. 28(2), 101–107 (2010).

53

Glotzer TV, Hellkamp AS, Zimmerman J et al.; MOST Investigators. Atrial high rate episodes detected by pacemaker diagnostics predict death and stroke: report of the Atrial Diagnostics Ancillary Study of the MOde Selection Trial (MOST). Circulation 107(12), 1614–1619 (2003).

54

Ziegler PD, Glotzer TV, Daoud EG et al. Incidence of newly detected atrial arrhythmias via implantable devices in patients with a history of thromboembolic events. Stroke 41(2), 256–260 (2010).

55

Ali M, Bath PM, Curram J et al. The Virtual International Stroke Trials Archive. Stroke 38(6), 1905–1910 (2007).

56

Ali M, Bath P, Brady M et al.; VISTA Steering Committees. Development, expansion, and use of a stroke clinical trials resource for novel exploratory analyses. Int. J. Stroke 7(2), 133–138 (2012).

57

Kamel H, Lees KR, Lyden PD et al.; Virtual International Stroke Trials Archive Investigators. Delayed detection of atrial fibrillation after ischemic stroke. J. Stroke Cerebrovasc. Dis. 18(6), 453–457 (2009).

47

Barthélémy JC, Féasson-Gérard S, Garnier P et al. Automatic cardiac event recorders reveal paroxysmal atrial fibrillation after unexplained strokes or transient ischemic attacks. Ann. Noninvasive Electrocardiol. 8(3), 194–199 (2003).

58

Okada Y, Shibazaki K, Kimura K, Iguchi Y, Miki T. Brain natriuretic peptide as a predictor of delayed atrial fibrillation after ischaemic stroke and transient ischaemic attack. Eur. J. Neurol. 17(2), 326–331 (2010).

48

Jabaudon D, Sztajzel J, Sievert K, Landis T, Sztajzel R. Usefulness of ambulatory 7-day

59

Montaner J, Perea-Gainza M, Delgado P et al. Etiologic diagnosis of ischemic stroke

493

Review

Abdul-Rahim & Lees

subtypes with plasma biomarkers. Stroke 39(8), 2280–2287 (2008).

modified microlife blood pressure monitor. Am. J. Hypertens. 22(8), 848–852 (2009).

60

Sinha AM, Diener HC, Morillo CA et al. Cryptogenic Stroke and underlying Atrial Fibrillation (CRYSTAL AF): design and rationale. Am. Heart J. 160(1), 36–41.e1 (2010).

63

Stergiou GS, Karpettas N, Protogerou A, Nasothimiou EG, Kyriakidis M. Diagnostic accuracy of a home blood pressure monitor to detect atrial fibrillation. J. Hum. Hypertens. 23(10), 654–658 (2009).

•

The CRYSTAL-AF study is the first multicenter randomized controlled trial to evaluate the effectiveness of outpatient monitoring using implantable loop recorders to detect paroxysmal atrial fibrillation compared with routine clinical follow-up. Study is expected to complete in May 2013.

64

Lees KR, Bluhmki E, von Kummer R et al.; ECASS, ATLANTIS, NINDS and EPITHET rt-PA Study Group. Time to treatment with intravenous alteplase and outcome in stroke: an updated pooled analysis of ECASS, ATLANTIS, NINDS, and EPITHET trials. Lancet 375(9727), 1695–1703 (2010).

61

Lewis M, Parker D, Weston C, Bowes M. Screening for atrial fibrillation: sensitivity and specificity of a new methodology. Br. J. Gen. Pract. 61(582), 38–39 (2011).

62

Wiesel J, Fitzig L, Herschman Y, Messineo FC. Detection of atrial fibrillation using a

494

65

Cucchiara B, Kasner SE, Tanne D et al.; SAINT Investigators. Factors associated with intracerebral hemorrhage after thrombolytic therapy for ischemic stroke: pooled analysis of placebo data from the Stroke-Acute Ischemic NXY Treatment (SAINT) I and SAINT II trials. Stroke 40(9), 3067–3072 (2009).

66

Awadh M, MacDougall N, Santosh C, Teasdale E, Baird T, Muir KW. Early recurrent ischemic stroke complicating intravenous thrombolysis for stroke: incidence and association with atrial fibrillation. Stroke 41(9), 1990–1995 (2010).

67

Schuchert A, Behrens G, Meinertz T. Impact of long-term ECG recording on the detection of paroxysmal atrial fibrillation in patients after an acute ischemic stroke. Pacing Clin. Electrophysiol. 22(7), 1082–1084 (1999).

Websites 101

Study of Continuous Cardiac Monitoring to Assess Atrial Fibrillation After Cryptogenic stroke. (CRYSTAL-AF). www.clinicaltrials.gov/ct2/ showNCT00924638

102

Long-term Cardiac Monitoring after Cryptogenic stroke. (CMACS). www.clinicaltrials.gov/ct2/ showNCT00932425

Expert Rev. Cardiovasc. Ther. 11(4), (2013)