“Pill‐in‐Pocket” anticoagulation for stroke prevention in atrial fibrillation

Uninterrupted anticoagulation for atrial fibrillation (AF), regardless of AF burden, is deeply rooted in practice since the early anticoagulation trials. However, uninterrupted anticoagulation is not without risks, and may not be beneficial for allcomers with a history of AF. Indeed, contemporary data that support a critical duration threshold of AF that benefits from anticoagulation, and a temporal association between stroke and multihour AF episodes, compel the study of a more targeted approach to AF anticoagulation. In this review, we discuss data that support further investigation of “pill in the pocket” anticoagulation for AF, and introduce the pivotal Rhythm Evaluation for Anticoagulation Therapy for Atrial Fibrillation (REACT‐AF) trial that will robustly evaluate this strategy.


| INTRODUCTION
Medicine is not "one size fits all."Indeed, there are few examples when all patients with a history of a disease are treated with the same doses of the same medications without regard to disease burden, comorbidities, or efficacy of treatment.An exception is our current approach to stroke prevention in atrial fibrillation (AF), as chronic anticoagulation is recommended for nearly all patients with clinically detected AF and other vascular risk factors, regardless of the number of risk factors, burden of AF, or history of successful intervention. 1,2r many, this translates into a lifetime of exposure to the risks and costs of anticoagulation therapy, sometimes with questionable benefit.
As will be discussed in forthcoming sections, recent data suggest that AF stroke risk is not "all or none" based solely on a history of arrhythmia.Accordingly, this calls into question the "one size fits all" status-quo.Fortunately, the availability of wearable consumer-grade devices capable of frequent, intermittent AF monitoring, combined with rapid onset oral anticoagulants, provides an opportunity to test an altogether new strategy-one of targeted, time-delimited "pill in pocket" anticoagulation for stroke prevention (Figure 1).Such a personalized medicine strategy, if proven safe and effective, could maintain the benefits while reducing the risks of anticoagulation, thereby providing AF patients with an alternative means of stroke prevention that decreases the cost of treatment and improves quality of life.

| TEMPORAL ASSOCIATION BETWEEN AF AND STROKE; AF BURDEN AND STROKE RISK
Several fundamental issues need to be addressed for this form of personalized medicine to be realized.First are conflicting data surrounding the temporal association between AF and stroke.4][5] One explanation for the disparity in timing of stroke in relation to episodes of AF invokes the concept of the atrial myopathy, which can be associated with stroke even in the absence AF. 6 However, other explanations should be considered.Importantly, the aforementioned data cannot distinguish the role of potential confounders, such as hypertension and diabetes, that lead to stroke even when AF is absent.In addition, many of the device-detected AF episodes noted in these studies were short in duration and would therefore not be expected to result in ischemic brain injury.Finally, no contemporaneous effort was made to adjudicate stroke mechanism, so the true incidence of AF-related cardioembolic strokes in these cohorts remained undetermined.
In contrast to the studies described previously, two large casecrossover studies that controlled for potential confounders did show a clear temporal association between stroke and multihour AF episodes. 7,8In one of these studies, we evaluated 891 stroke patients from the Optum Health Record data set who also had a remotely monitored CIED with rhythm data available for analysis.
In this study, each patient served as their own control whereby heart rhythm data was compared between time points 1−30 days (case period), and 90−121 days (control period), before their stroke.In total, the odds ratio for an episode of AF > 5.5 h in the case versus control period was 3.71, and the highest risk of stroke was within 5 days of an AF episode (Figure 2). 7Furthermore, this temporal relationship was amplified in those not on oral anticoagulation and reduced to nonsignificance in those on anticoagulation.These data support the concept of a high-risk period for embolic stroke that peaks shortly after an AF episode occurs, wanes in the days and weeks following its conclusion, and is reduced with oral anticoagulation.The second critical aspect of a targeted anticoagulation approach is to define the actionable threshold of AF duration that benefits from anticoagulation.This concept of an actionable threshold is relatively new, as chronic anticoagulation regardless of AF duration is rooted in practice since the early anticoagulation trials which demonstrated no difference in stroke risk between those with "intermittent" versus "sustained" AF. 9 In contrast, data from the more modern era of AF management shows a clear step-up in risk as one moves from paroxysmal to persistent to permanent AF.1][12] For example, in an analysis of 21768 patients in the Optum Health Record data set with a CIED included in the Medtronic CareLink ® database who were not on anticoagulation, duration of AF was significantly associated with rates of stroke or systemic embolism.Indeed, the rate of stroke or systemic embolism per-year in patients with no AF, 6 min to 23.5 h of AF and >23.5 h of AF were 0.81 (95% CI: 0.74−0.89),1.0 (95% CI: 0.83−1.20),and 1.43 (95% CI: 1.13−1.82),respectfully.In addition, this study found a significant interaction between duration of AF and CHA 2 DS 2 -VASc in predicting annual rates of stroke and systemic embolism (Figure 3). 11gether, contemporary data demonstrating a temporal association between AF and stroke, and an actionable AF threshold, measured in hours, that is associated with higher risk of stroke, provide a viable framework to consider intermittent use of rapid-onset oral anticoagulants given only during a high-risk window in response to a multihour AF episode.

| INTERMITTENT, PASSIVE AF MONITORING USING DIGITAL HEALTH TECHNOLOGIES
Given the often-asymptomatic nature of AF, long-term passive cardiac monitoring is a key component of a targeted anticoagulation strategy.Photoplethysmography (PPG) based methods for AF detection are accurate in both smartphones and smartwatches. 13,14 the positive predictive value of detecting a true AF event is dependent on the pretest probability of AF in the cohort assessed, the use of these devices to guide intermittent anticoagulation in a sample with known-AF is particularly attractive.Indeed, the current generation of AF-sensing smartwatches allows for inexpensive, noninvasive, long-term assessment of pulse irregularities using intermittent PPG, confirmed with on-demand single-lead ECG.
F I G U R E 2 Temporal association between atrial fibrillation and stroke.Data from Singer et al. 7 reveal a clear temporal association between episodes of atrial fibrillation and cardioembolic stroke.Results of this case-crossover study, which accounted for potential confounding factors, demonstrate that the odds ratio for ischemic stroke was highest in Days 1−5 following an episode of atrial fibrillation (defined as at least one day with >5.5 h of atrial fibrillation).
F I G U R E 3 Stroke risk in non-anticoagulated patients by AF burden and CHA 2 DS 2 -VASc score.Data from Kaplan et al. 11 demonstrate that stroke risk in patients with atrial fibrillation not constant, but is instead a function of vascular risk factors (CHA 2 DS 2 -VASc score) and burden of arrhythmia.AF patients with the highest number of risk factors and atrial fibrillation burden incur the greatest risk of stroke. 11AF, atrial fibrillation.
Several limitations of PPG based AF detection, however, should be recognized.For example, Apple Watch passively checks for an irregular tachogram for 1 min every 2 h if the user is at rest.After an irregular tachogram is detected, the algorithm will increase the frequency of sampling up to every 15 min.If 5/6 consecutive pulse checks are irregular, the user will be notified of an irregular pulse. 15 updated Apple Watch algorithm, cleared for those with a known AF history, provides more frequent pulse assessments which are used to calculate a weekly AF burden, though no notifications are delivered to the patient.The Fitbit Rhythm Detection algorithm includes 5 min overlapping detection intervals that require at least 11 consecutive, analyzable irregular tachograms to occur within a 24 h period to produce an irregular heart rhythm notification.
It is important for patients and healthcare providers to understand that the irregular rhythm notification algorithms discussed above do not offer true continuous monitoring for AF, are only FDA-approved for those without a history of AF, do not reliably pick up short episodes of AF, and suspend detection during periods of excessive movement. 15,16Lastly, the use of wearables for a "pill in pocket" approach to anticoagulation requires a high degree of patient compliance, particularly during sleeping hours when the majority of AF episodes may be detected due to less arm movement. 14embolism, and major bleeding.However, IMPACT included patients with a reduced left ventricular ejection fraction, and utilized vitamin K antagonists, rather than rapid-acting anticoagulants, thereby limiting generalizability of results. 5nversely, two single-arm pilot studies have assessed the feasibility of the "pill in pocket" approach using CIEDs with encouraging results.In the REACT.COM pilot study, 59 AF patients with an implantable cardiac monitor reinitiated their previously prescribed NOAC for 30 days in response to an AF episode >1 h.

| PRIOR TRIALS
Over the course of 14 months, only 18 patients (31%) experienced recurrent AF with an average of 2 episodes per patient.The overall reduction in "time on" anticoagulation was 94% with no strokes and two major bleeds occurring off anticoagulation. 17The second trial, TACTIC-AF, enrolled 48 PPM and ICD patients and reinitiated anticoagulation in response to a continuous AF episode >6 min or a total AF burden of >6 h over 24 h. 18Two-thirds of patients had recurrent AF defined by these thresholds, again with no strokes and one major bleed off anticoagulation.Together, these trials enrolled 96 patients with 112 patient-years of follow-up with no strokes, thereby providing a basis for further study. 17,18| PIVOTAL TRIAL: REACT-AF The highest levels of clinical evidence are necessary to change clinical practice, especially when it surrounds a subject as critical as stroke prevention in AF.The Rhythm Evaluationfor Anticoagulation Therapy for Atrial Fibrillation (REACT-AF) trial will compare the current standard of care of chronic anticoagulation with a smartwatch-guided, targeted, timedelimited approach to anticoagulation (Figure 4).The study will enroll 5350 patients from 80 to 100 US sites.Major inclusion criteria include CHA 2 DS 2 -VASC score of 1−4 for males and 2−4 for females without prior stroke or TIA, low burden of AF as defined by symptoms and external monitoring ≥7 days, and current NOAC use.Major exclusion criteria include contraindications to NOAC, history of heart failure, and existing or planned CIED.Patients in the treatment group will be provided with a smartwatch with arrhythmia detection capability, and will be instructed to wear the watch for a minimum of 14 h a day and during sleep, if possible.In response to a single episode of continuous AF >1 h, participants will receive a notification to take their previously prescribed NOAC, which will be continued for 30 days following the last AF episode.
The primary endpoint is non-inferiority for a combination of stroke, F I G U R E 5 Emerging landscape of atrial fibrillation stroke prevention.A proposed treatment algorithm if smartwatch-guided "pill in the pocket" anticoagulation is proven to be safe and effective.Continuous NOAC: patients with high atrial fibrillation burden and/or high vascular risk with low bleeding risk.Left atrial appendage occlusion: patients with high AF burden and/or high vascular risk with high bleeding risk;smartwatch-guided "pill-in-the-pocket" anticoagulation: low-intermediate vascular risk, low-intermediate atrial fibrillation burden, across the spectrum of bleeding risk.AF, atrial fibrillation; LAAO, left atrial appendage occlusion; NOAC, novel oral anticoagulant.
arterial embolism, and all-cause mortality and the secondary endpoint is superiority for major bleeds.The pivotal REACT-AF trial will therefore provide the public with valuable information regarding the safety and efficacy of smartwatch guided "pill-in-the-pocket" anticoagulation for AF.
If shown to be effective, results will expand stroke prevention options for select patients with AF (Figure 5).Enrollment in REACT-AF is scheduled to commence Spring 2023.

| CONCLUSION
The "one size fits all" model of chronic anticoagulation for stroke prevention in AF is plagued with the challenges of low adoption rates, high discontinuation rates, and legitimate concerns over bleed risk.
The growing recognition of a critical AF duration threshold, coupled with emerging data on the temporal association between AF and stroke, establishes a firm foundation to test the concept of targeted, time-delimited "pill-in-pocket" anticoagulation.
We believe that "pill in pocket" anticoagulation may have particular relevance, given the increasing evidence base for early rhythm control strategies that significantly lower AF burden. 19,20The development of rapid onset oral agents and widespread availability of consumer-grade digital health devices capable of frequent, passive rhythm assessments make this novel approach feasible and scalable.
Whether this novel intervention affords the same protection against stroke while reducing major bleeds will be answered by REACT-AF.

F
I G U R E 1 "Pill-in-Pocket" anticoagulation."Pill-in-pocket" anticoagulation is a novel paradigm whereby a patient initiates oral anticoagulation only in response to a documented episode of atrial fibrillation.The safety and efficacy of this strategy will be tested in the pivotal REACT-AF trial.Republished with permission from: Passman, R. (2021)."Pill-in-Pocket" anticoagulation for atrial fibrillation: fiction, fact, or foolish?"Circulation 143(23): 2211-2213.PEIGH and PASSMAN | 2153 One prior randomized trial evaluated the "pill in pocket" approach to stroke prevention.The IMPACT trial randomized 2718 ICD and CRT-D patients with no AF history to receive either time-delimited oral anticoagulation based upon remotely-detected AF or standard continuous oral anticoagulation.The trial did not demonstrate a benefit of targeted anticoagulation for the composite endpoint of stroke, systemic F I G U R E 4 Study schema for REACT-AF.AF, atrial fibrillation; NOAC, novel oral anticoagulant; REACT-AF, rhythm evaluation for anticoagulation therapy for atrial fibrillation.