Continuing evaluation of atrial fibrillation detection after cryptogenic stroke: 2-year findings from a multicentre study with Confirm Rx ICM

Abstract

Background

The detection of atrial fibrillation (AF) after a cryptogenic stroke (CS) carries important therapeutic implications. In this study, we aimed to accurately assess the incidence of AF among CS subjects by using an insertable cardiac monitor (ICM).

Methods

A prospective, single-arm, multicentre registry was conducted to identify AF in 155 CS subjects using the Confirm Rx ICM (Abbott, California, USA) across 20 global sites. Inclusion criteria comprised participants aged 40 years or older who had experienced CS within a 90-day window. At each follow-up visit, expert electrophysiologists reviewed and adjudicated ICM detected AF episodes. The primary endpoint was the cumulative incidence of true device-detected AF (lasting more than 30 s) at 6 months, evaluated with Kaplan-Meier methods.

Results

AF incidence was 21.3% (95% CI 15.3% to 29.1%) at 6 months, increasing to 48.8% (95% CI 34.7% to 64.9%) at 24 months. Subjects with AF detection experienced an average of 50.9 true AF episodes per subject per year. The median time from implantation to AF detection (>30 s) was 72 days (IQR 7–261). Among subjects with 30 s AF detection, anticoagulation therapy was initiated in 65.2% (30/46) of subjects. Oral anticoagulation medication was prescribed in 8.3% (9/109) of subjects without AF. Recurrent ischaemic stroke or transient ischaemic attack occurred in 5 subjects (3.2%, 5/155).

Conclusion

These results show that ICM-driven long-term continuous AF monitoring is associated with high diagnostic yield in CS subjects.

Trial registration number

NCT03505801

Correspondence to Dr Fabio Quartieri; [email protected]

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Previous randomised trials demonstrated that continuous monitoring with insertable cardiac monitor (ICM) significantly enhances atrial fibrillation (AF) detection in cryptogenic stroke subjects compared with standard care, though reported detection rates vary across studies.

WHAT THIS STUDY ADDS

• This prospective, multicentre study demonstrates high AF detection rates using Confirm Rx ICM in cryptogenic stroke subjects: 21.3% (95% CI 15.3% to 29.1%) at 6 months and 48.8% (95% CI 34.7% to 64.9%) at 24 months. Subjects with AF detection experienced an average of 50.9 true AF episodes per subject per year, with shorter but more frequent episodes compared with previous studies. Age and hyperlipidaemia emerged as significant independent predictors for 30 s AF detection.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings provide evidence supporting extended cardiac monitoring in cryptogenic stroke subjects and suggest the need for refined approaches to AF burden assessment. The high detection rate and distinctive AF patterns observed may influence clinical decision-making regarding monitoring strategies and anticoagulation initiation thresholds.

Introduction

Cryptogenic stroke (CS), a heterogeneous condition, is defined as an ischaemic stroke where no identifiable cause is found despite comprehensive diagnostic evaluation.¹ The classification of stroke subtypes according to the Trial of ORG 10172 in Acute Stroke Treatment criteria is widely used; CS is diagnosed only after ruling out major causative mechanisms.² According to the American Heart Association (AHA), ischaemic stroke is considered cryptogenic when the cause of ischaemic stroke remains undetermined after a standard evaluation involving basic laboratory tests, brain imaging, neurovascular imaging and cardiac evaluation.³

According to a meta-analysis of 50 studies, nearly 25% of subjects experiencing ischaemic stroke or transient ischaemic attack (TIA) have newly detected atrial fibrillation (AF) through sequential cardiac monitoring methods.⁴ Atrial high-rate episodes (AHRE), including subclinical AF detected by implanted devices in individuals with no prior AF history, occur in approximately 20% of subjects 1 year after cardiac device implantation. AHRE and subclinical AF are associated with clinical AF and stroke risk.⁵ The prospective ASSERT study linked subclinical atrial tachyarrhythmia to ischaemic stroke among older adults with hypertension, no history of AF and recent pacemaker or defibrillator implantation, compared with those without detected atrial tachyarrhythmia.⁶ Additionally, the risk of stroke associated with AF is similar for symptomatic and asymptomatic subjects.⁷

Further randomised trials, such as CRYSTAL-AF, have demonstrated that continuous monitoring with insertable cardiac monitor (ICM) significantly enhances the detection of AF in subjects with CS compared with standard care, which typically involves intermittent short-term monitoring.^8–10 These findings highlight the limitations of conventional monitoring approaches and suggest the potential benefit of long-term continuous monitoring strategies. The identification of AF following CS carries significant therapeutic implications; more than 80% of North American providers initiate anticoagulation on discovering occult AF. AHA guidelines recommend prolonged cardiac monitoring (COR: 2a and LOE: B-R) in CS subjects and anticoagulation if occult AF is found, regardless of duration.^{3 11} Early detection of AF and appropriate anticoagulant therapy in high-risk subjects has the potential to substantially reduce stroke rates and morbidity.¹² Therefore, AF detection is integral to a comprehensive approach to patient care.

This study aims to evaluate AF incidence in subjects with CS using an ICM, based on interim results from the CS substudy of the SMART Registry.

Methods

A prospective, single-arm, multicentre registry was conducted to identify AF in subjects with CS using the Confirm Rx ICM from Abbott (California, USA). Eligible participants were aged 40 years or older and had experienced a stroke or TIA within the previous 90 days, with diagnosis supported by consistent symptoms and findings on either brain MRI or CT. Stroke was classified as cryptogenic after comprehensive diagnostic testing, including 12-lead ECG, ≥24 hours of ECG monitoring, transoesophageal echocardiography (TEE) and screening for thrombophilic states (in subjects <55 years), along with MR angiography (MRA), (CT angiography (CTA) or catheter angiography of the head and neck. Additionally, TEE or intracardiac echocardiography (ICE) was performed to rule out intracardiac thrombus, and CTA or MRA of the head and neck was conducted to exclude other stroke pathologies. Ultrasonography of cervical arteries and transcranial Doppler ultrasonography were permitted for subjects over 55 years as alternatives to MRA or CTA. Subjects with TIA were included if presenting symptoms included speech problems, limb weakness or hemianopsia. Exclusion criteria encompassed a history of AF or atrial flutter, eligibility or ineligibility for permanent oral anticoagulant therapy at enrolment, and indication for a pacemaker or implantable cardioverter–defibrillator.

Prior to enrolment, all subjects underwent thorough diagnostic assessments to systematically exclude potential embolic sources, in accordance with minimal standards outlined in American and European Guidelines. The inclusion and exclusion criteria (see online supplemental table 3) were carefully selected to closely match the study population from the Crystal AF trial (online supplemental file 2).⁹ Prior to analysis, all enrolled subjects underwent a systematic review to verify complete adherence to inclusion/exclusion criteria and protocol-specified procedures. Subjects who did not meet all protocol requirements were excluded from the analysis population. This approach aimed to offer a historical comparison with the two randomised groups in the Crystal AF study: those monitored continuously with ICM and those monitored with alternative methods, such as repeated ECGs or Holter tests.

The subject’s medical history, the date when index stroke or TIA occurred and the use of medications were recorded. Expert electrophysiologists reviewed and adjudicated AF episodes detected by the ICM during follow-up visits.

The primary endpoint is the cumulative incidence of true device-detected AF (lasting more than 30 s) at 6 months postinsertion of the Confirm Rx ICM, as determined by expert electrophysiologists at the participating site. The secondary endpoint is the cumulative incidence of true device-detected AF (lasting more than 30 s) at 24 months postinsertion of the Confirm Rx ICM, as determined by expert electrophysiologists at the participating site. The CS substudy followed subjects for 24 months.

AF monitoring strategies

Subjects underwent assessment at scheduled and unscheduled visits, with cardiac arrhythmias adjudicated by the site investigators. The ICM used (Abbott, Sylmar, California, USA) automatically detects and records AF, regardless of heart rate or symptoms. Merlin.net remote monitoring platform was used to transmit device data remotely. Follow-up visits were scheduled at 1, 6, 12, 18 and 24 months, with unscheduled visits in the event of symptom occurrence or after the transmission of ICM data, if advised by the investigator. If subjects reported an episode of AF since the previous visit, information was collected, and source documentation was acquired for adjudication.

Statistical analysis

Continuous variables were summarised by mean±SD if they had a normal distribution or by median and IQR if they had a skewed distribution. Categorical variables were summarised by frequency and percentage. The Kaplan-Meier method was used to analyse time to AF detection data and estimate the rate of AF detection, and the log-log method was used to estimate the 95% CI. For subjects who experienced an AF detection, the AF-free survival time was calculated as the number of months from ICM implant to the date of AF detection. Subjects were censored at the time of withdrawal for subjects who withdrew from the study (due to death, device explant, lost to follow-up or subject request) without AF detected. For subjects who did not have AF detected at the time of data cut-off, subjects were censored on the date of cut-off. Univariate and multivariate logistic regression models were used to evaluate significant predictors for 30 s AF detection at 24 months. Analyses were conducted using SAS V.9.4 software.

Patient consent

Patients were involved in this study through the informed consent process after device implantation. While they were not involved in the initial research design, their feedback influenced the implementation of remote monitoring protocols and follow-up schedules. Patients receive ongoing communication about their device data through scheduled remote monitoring sessions at 18 and 24 months postinsertion. No public involvement was included in this study design.

Results

Study population

A total of 171 subjects were enrolled in the study across 20 global sites from September 2021 to 31 January 2024. The study is currently in the follow-up phase, with the last subject follow-up visit expected in January 2025. As of the data cut-off date on 12 April 2024, 16 subjects did not meet inclusion/exclusion criteria, and therefore, 155 subjects were considered for analyses. 115 subjects had completed the 6-month follow-up visits, while a cumulative total of 4 had withdrawn, 1 missed the 6-month visit and 35 were pending their 6-month visit (online supplemental figure 1). For the 12-month follow-up, 80 subjects had completed their visits, with a cumulative total of 7 withdrawals, 6 missed visits, and 27 pending visits. At the 18-month follow-up, 59 subjects had completed their visits, with 10 cumulative withdrawals, 11 missed visits and 13 pending visits. By the 24-month follow-up, 29 subjects had completed their visits, with 15 withdrawals, 4 missed visits and 32 pending visits, resulting in a total follow-up duration of 151.6 subject years.

ICMs remained in 94.8% (147/155) of subjects. ICMs were removed due to battery end of service (3), subject death (2), at subject request (1), pacemaker implantation (2).

Baseline characteristics

Among 171 enrolled subjects, 155 subjects completed all required diagnostic tests (MRI or CT, 12 lead ECG, 24 hours ECG, TEE/ICE, head and neck CTA or MRA). Ultrasonography of cervical arteries and transcranial Doppler ultrasonography were permitted for subjects over 55 years as alternatives to MRA or CTA. All analyses presented are based on the 155 subjects who fully met the study criteria, ensuring data integrity and protocol compliance.

The baseline characteristics of the study population are detailed in table 1. The mean age of the participants was 64±10 years, with 34.2% being women. Hypertension was present in 49.0% of the subjects. The average CHA2DS2-VASc score among the participants was 3.5±1.5. All enrolled subjects had experienced a recent episode of cryptogenic ischaemic stroke within the preceding 90 days.

Table 1

Baseline demographics

Characteristic	N=155
Age mean±SD (years)	64±10
Female gender, % (n/N)	34.2% (53/155)
Europe	17.4% (27/155)
North America	3.2% (5/155)
Asia	79.4% (123/155)
Index event stroke	100% (155/155)
Hypertension	49.0% (76/155)
Hyperlipidaemia	27.1% (42/155)
Diabetes	23.9% (37/155)
Coronary artery disease	9.0% (14/155)
CHA2DS2-VASc score, mean±SD	3.5±1.5
0	1.4% (2/142)
1	6.3% (9/142)
2	16.9% (24/142)
3	26.8% (38/142)
4	26.8% (38/142)
5	9.2% (13/142)
6	11.3% (16/142)
7	1.4% (2/142)
Unknown	8.4% (13/155)
24 hours ECG monitoring	100% (155/155)

AF detection

Kaplan-Meier analysis showed a 21.3% incidence (95% CI 15.3% to 29.1%) of true AF episodes lasting more than 30 s at 6 months (figure 1). Notably, the first AF episode was detected in two subjects on the same day as the ICM insertion. By the 6-month mark, AF episodes were detected in 30 subjects cumulatively. The median time from the insertion of ICM to AF detection (episodes lasting more than 30 s) was 10 days (IQR 3–43 days) among a total of 30 subjects who had experienced their first AF episodes through 6 months.

Figure 1. Time to first detection of atrial fibrillation (30 s). AF, atrial fibrillation.

Kaplan-Meier analysis demonstrated a 49.2% incidence (95% CI 35.1% to 65.3%) of true AF episodes lasting more than 30 s at 24 months (figure 1). The median time from the insertion of ICM to AF detection (episodes lasting more than 30 s) was 72 days (IQR 7–261 days) among a total of 46 subjects who had experienced their first AF episodes through 24 months.

Incidence of AF lasting over 6 min at 6 months was 15.9% (95% CI 10.7% to 23.3%) and at 24 months was 36.2% (95% CI 25.4% to 49.6%). The incidence of AF lasting over 1 hour, 6 hours and 24 hours is also shown in figure 2.

Figure 2. Time to first detection of atrial fibrillation (30 s, 6 min, 1 hour, 12 hours, 24 hours). AF, atrial fibrillation.

Clinical actions

Various clinical actions were taken in response to 30 s AF detections throughout the study.

Throughout the study, AF was detected in a total of 46 subjects. Clinical actions were taken for 31 of these subjects. Among those who received clinical interventions, 30 subjects were prescribed new medications, 3 subjects underwent right atrial ablation and 5 subjects had left atrial ablation.

Anticoagulants prescription

Among subjects with 30 s AF detection, anticoagulation therapy was initiated in 65.2% (30/46) of subjects. For those with 6 min AF detection, anticoagulation therapy was initiated in 64.7% (22/34) of subjects. For subjects with longer AF episodes, the initiation rates were: 76.2% (16/21) for 1-hour detection, 76.9% (10/13) for 6-hour detection and 62.5% it was initiated in 76.2% (16/21) of subjects; for those with 6 hours AF detection, it was initiated in 76.9% (10/13) of subjects; for those with 24 hours AF detection, it was initiated in 62.5% (5/8) of subjects. Anticoagulation therapy was prescribed in 8.3% (9/109) of subjects without detected AF episodes.

Recurrence of stroke

During a mean follow-up of 11.7±7.4 months, ischaemic stroke or TIA recurred in 5 subjects (3.2%, 5/155), with a mean time to recurrence of 9.6±5.7 months from enrolment. None of these five subjects had detected AF at the time of their recurrent events. Three subjects experienced strokes within the first 6 months after enrolment (at 4.3, 6.2 and 6.3 months), one subject at 13.5 months and another at 17.7 months.

Duration of AF

46 subjects with detected AF experienced an average of 50.9 true AF episodes per subject per year. The maximum duration of AF of a single episode in a single day (within the first 6 months) had a skewed distribution among patients, with a median equal to 5.08 min (IQR 2.03–12.61 min) (figure 3).

Figure 3. Maximum AF duration of a single episode in an individual subject (within the first 6 months). **One subject with maximum AF duration indicated as >24 hours is included here. AF, atrial fibrillation.

Independent predictors for incidence of AF

The results of the logistic regression model show that elderly age (OR 1.09, 95% CI 1.04 to 1.14, p=0.0001) and hyperlipidaemia (OR 2.37, 95% CI 1.06 to 5.29, p=0.035) are significant independent predictors for incidence of 30 s and AF in CS subjects (table 2). For 6 min AF detection, only elderly age remained significant (OR 1.08, 95% CI 1.03 to 1.13, p=0.0012). For 1-hour AF detection, ischaemic condition (OR 14.20, 95% CI 1.30 to 154.83, p=0.0295) and peripheral vascular disease (OR 25.74, 95% CI 2.00 to 331.81, p=0.0128) emerged as significant predictors. For 24-hour AF, hyperlipidaemia (OR 14.77, 95% CI 1.57 to 139.14, p=0.0186) and prior PTCA (OR 13.72, 95% CI 1.53 to 122.98, p=0.0192) were significant predictors (online supplemental table 1).

Table 2

Logistic regression results (incidence of 30 s AF vs no AF)

Parameters	Univariate analysis			Multivariable analysis
	Parameter estimate (95% CI)	P value	Sample size	Parameter estimate (95% CI)	P value	Sample size
Age	1.083 (1.039 to 1.129)	0.0002	155	1.09 (1.04 to 1.14)	0.0001	155
Unstable angina, Yes vs No	0.785 (0.080 to 7.753)	0.8360	155
Prior CABG, Yes vs No	0.000 (0.000 to 1)	0.9888	155
CAD, Yes vs No	1.894 (0.618 to 5.805)	0.2637	155
CHA2DS2-VASc score	1.189 (0.930 to 1.521)	0.1673	142
Ischaemic vs non-ischaemic	2.431 (0.332 to 17.806)	0.3818	155
COPD asthma, Yes vs No	1.189 (0.105 to 13.445)	0.8888	155
Diabetes, Yes vs No	1.186 (0.535 to 2.627)	0.6744	155
LVEF	0.987 (0.946 to 1.029)	0.5286	138
Hyperlipidaemia, Yes vs No	2.277 (1.080 to 4.799)	0.0305	155	2.37 (1.06 to 5.29)	0.0350	155
Hypertension, Yes vs No	1.740 (0.866 to 3.496)	0.1198	155
Liver disease, Yes vs No	0.785 (0.080 to 7.753)	0.8360	155
Other CV, Yes vs No	0.691 (0.181 to 2.635)	0.5882	155
MI, Yes vs No	1.606 (0.259 to 9.947)	0.6105	155
Prior PTCA, Yes vs No	0.946 (0.177 to 5.061)	0.9483	155
PVD, Yes vs No	4.909 (0.434 to 55.533)	0.1986	155
Renal disease, Yes vs No	0.946 (0.177 to 5.061)	0.9483	155
Female vs male	0.501 (0.230 to 1.093)	0.0826	155	0.46 (0.20 to 1.07)	0.0714	155
Stroke	0.000 (0.000 to 1)	0.9779	155
Thyroid, Yes vs No	3.733 (0.602 to 23.126)	0.1570	155

AF, Atrial Fibrillation; CABG, Coronary Artery Bypass Grafting; COPD, Chronic Obstructive Pulmonary Disease; LVEF, Left Ventricular Ejection Fraction; PTCA, Percutaneous Transluminal Coronary Angioplasty; PVD, Peripheral Vascular Disease.

Discussion

In this international prospective observational study, we analysed 155 subjects diagnosed with CS after comprehensive diagnostic assessments to thoroughly rule out all potential sources of embolism, in accordance with current guidelines and consensus papers.^1–3

Main study results

We found that (1) ICMs provided a high rate of AF detection and (2) ICM-driven AF detection triggered OAC prescription and other clinical interventions in a relevant proportion of patients.

AF detection rate

In our study, the incidence of detected AF (>30 s) was 21.3% at 6 months and 48.8% at 2 years. These rates differ from previous studies in CS patients. When interpreting these differences, several factors warrant consideration. First, the duration of non-invasive cardiac monitoring prior to ICM implantation is a critical determinant of subsequent AF detection rates, as longer preimplant monitoring periods might select populations with different propensities for AF detection. While technical differences between monitoring devices might contribute to varying detection rates, this comparison is inherently limited by potential differences in preimplant screening durations across studies. The pattern of AF detection in our cohort showed some distinct characteristics. We observed more frequent but shorter episodes (median maximum daily AF duration 39.5 min) compared with CRYSTAL-AF (median 11.2 hours). This observation aligns with contemporary understanding that AF patterns can be highly variable, ranging from brief, frequent episodes to longer, less frequent occurrences. Interestingly, three other studies confirmed that ICM used in our study has been shown in a randomised study to provide a high arrhythmia detection yield.^13–15 Ziegler et al¹⁴ evaluated a cohort of 1247 US patients with CS and monitored with the same ICM used in the Crystal AF study, and they found an AF detection rate of 12.2% at 6 months and 21.5% at 2 years, which are similar to Crystal AF study findings and much lower than our study findings. Boriani et al¹⁵ described the detection of AF using the same ICM used in our study in a cohort of US adults who were hospitalised with ischaemic stroke, which was cryptogenic in about 90% of patients. They reported that the AF detection rate was 33.9% at 2 years, which was slightly lower than the incidence found in our study but much higher than that found in Crystal AF⁹ and that reported by Ziegler et al.¹⁴ These differences across studies likely reflect the complex interplay of multiple factors, including patient characteristics, monitoring duration and AF patterns.

Recent clinical trials and anticoagulation strategy

The optimal management strategy for device-detected AF after CS requires careful consideration in light of recent evidence. The NOAH trial investigated edoxaban in patients with subclinical AF without prior stroke; however, it did not demonstrate a benefit of edoxaban in terms of reduction in the primary endpoint and actually showed harm with increased major bleeding events.¹⁶ Meanwhile, ARTESIA showed that apixaban reduced stroke risk in patients with device-detected subclinical AF ≥6 min. The recent subgroup analysis in patients with prior stroke further supports this approach.¹⁷ Although these trials reported lower event rates than our study (our recurrent stroke/TIA rate was 3.2%), this difference likely reflects our higher-risk population of CS patients. Notably, in our cohort, all five recurrent events occurred in patients without detected AF at the time of their events, and only one was on OAC therapy. This observation, combined with evidence from ARTESIA supporting anticoagulation in subclinical AF, suggests that a more aggressive anticoagulation strategy might be warranted in poststroke patients with device-detected AF. However, the optimal AF duration threshold for initiating anticoagulation remains uncertain, particularly given our finding of frequent but shorter AF episodes (median maximum daily duration 39.5 min) compared with previous studies. Further research is needed to determine whether lower AF duration thresholds for anticoagulation might benefit this high–risk population.

Anticoagulants prescription

Anticoagulation therapy has been prescribed in our study in 65.2% of subjects with 30 s of AF detection. This rate is lower than that (92%) observed in the Crystal AF study, but it is important to outline that in the Crystal AF study, 100% of patients with AF episode duration longer than 1 hour received OAC therapy, but only 70% of patients with shorter AF durations (<1 hour) were prescribed with OAC. Several aspects can explain our OAC prescription rate (65.2%); first of all, while we observed more AF episodes than Crystal AF, the duration of AF in our cohort was shorter than that observed in Crystal AF, and that could have guided some investigators towards a more conservative anticoagulation approach. In our study, the median maximum daily AF duration was 39.5 min, while it was 11.2 hours in the Crystal AF study. It is also important to outline the fact that after CRYSTAL AF, some studies^{16 18 19} have questioned the clinical benefit of initiating OAC therapy after detection of short subclinical AF episodes.

Independent predictors for AF in CS patients

In our study, age was significantly associated with AF detection. This result adds further evidence to the observation of Crystal AF investigators,²⁰ who found that age was independently associated with ICM-driven AF detection. These findings were expected since several studies have shown that AF, including long-duration AF and asymptomatic AF, is more frequent with older age.^{21 22}

Clinical implications

Our findings should be interpreted within the current evidence framework regarding device-detected AF and stroke risk. Two distinct but related aspects need to be considered in interpreting these results. The association between device-detected AF and stroke risk represents the first key consideration. Several observational studies have demonstrated an association between AHRE (>5 min) and increased stroke risk.^{23 24} In the study recently published by Yaghi et al,²⁵ ICM use was associated with a significantly reduced risk of death with HR=0.70, 95% CI (0.55 to 0.89). Also, in the meta-analysis performed by Tsivgoulis et al²⁶ ICM use, compared with conventional monitoring in CS patients, was associated with increased AF detection yield, higher OAC initiation and decreased risk of recurrent stroke with ICM. However, it is important to note that this association does not necessarily establish a causal relationship between device-detected AF episodes and stroke events. The temporal relationship between AF episodes and stroke events remains complex and incompletely understood. In our study, notably, all five recurrent strokes occurred in patients without detected AF at the time of their events, further highlighting this complexity.

The therapeutic implications of device-detected AF constitute the second critical aspect requiring careful consideration. The optimal management strategy for device-detected AF, particularly regarding anticoagulation, remains uncertain. Current evidence for anticoagulation in device-detected AF comes primarily from two recent trials: the NOAH trial, studying patients with subclinical AF but no prior stroke, and the ARTESIA trial, examining patients with device-detected subclinical AF ≥6 min. However, it is crucial to note that these trials included patients with pacemakers or ICDs, not stroke patients with ICMs. Therefore, their findings cannot be directly extrapolated to our population of CS patients. Currently, there is limited evidence demonstrating that anticoagulation in device-detected AF after stroke prevents recurrent ischaemic events.

Our observation of a 65.2% anticoagulation rate in AF-detected patients reflects this ongoing uncertainty in clinical practice. The decision to anticoagulate remains individualised, considering factors such as AF burden and pattern, patient-specific stroke risk factors and the balance between stroke prevention and bleeding risk. Continuous AF monitoring through ICMs provides valuable information about arrhythmia burden and patterns, which can inform clinical decision-making,²⁷ and it may have clinical relevance to avoid AF-related silent brain infarcts that may deteriorate cognitive function²⁸ and in specific patients, such as patients with paroxysmal, infrequent and asymptomatic AF.^29–34

Limitations

Several limitations of our study should be acknowledged. First, our study had a single-arm observational design; therefore, we cannot directly compare our findings with alternative monitoring strategies. Second, patient enrolment averaged approximately four patients per centre per year; this number may seem low, but we believe it reflects the rigorous inclusion criteria and comprehensive screening process required to establish CS diagnosis. While this thorough screening process helped ensure a well-characterised study population, it may have affected the generalisability of our findings. Third, AF episodes were adjudicated by expert electrophysiologists at each participating site rather than by a central independent committee. Although all participating electrophysiologists followed standardised criteria for AF diagnosis and the adjudication process was carefully documented at each site, we acknowledge that central adjudication might have provided additional standardisation. Fourth, while we observed patients for up to 24 months, longer follow-up might provide additional insights into AF patterns and stroke recurrence. Finally, the relatively small sample size resulted in wide CIs for some of our analyses, particularly at later time points. It is difficult to compare studies that could have added different durations of non-invasive monitoring prior to ILR implant because these possible differences could have selected populations with a different propensity to have subsequent episodes, and that difference could have resulted in different incidences of ICM detected AF. Despite these limitations, we believe that our findings contribute meaningful data regarding AF detection in CS patients and align with emerging evidence from contemporary studies using similar monitoring approaches.

Conclusions

Long-term continuous monitoring with Confirm Rx ICM demonstrated a substantial AF detection rate of 21.3% at 6 months and 48.8% at 24 months in CS patients. These findings support the value of extended cardiac monitoring in this population, through the optimal management strategy for device-detected AF continues to evolve.

The authors thank Wenjiao Lin and Allison Havener for conducting statistical analyses.

Data availability statement

No data are available. The study protocol and statistical analysis plan will be made available on reasonable request to the corresponding author.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

Approval for the study protocol was obtained from all relevant institutional review boards or ethics committees across participating centres in South Korea, Japan, Germany, Italy, Slovakia, Singapore, Portugal and the USA. Written informed consent was obtained from all subjects prior to enrolment, in accordance with the Declaration of Helsinki and local regulations.

Footnote

Contributors FQ, Y-SB, J-SP, T-HK, KH, MM, K-WK and LK were involved in subject recruitment, data collection and manuscript preparation. FQ, LF, KL and AG contributed to study design and data interpretation. All authors participated in manuscript review and revision. As the guarantor of this work, FQ had full access to all study data and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors approved the final version of the manuscript. FQ is the guarantor.

Competing interests FQ received consultancy fees from Abbott. T-HK received consultancy fees from Abbott. K-WK received consultancy fees from Abbott. LF, KL and AG are Abbott employees.

Patient and public involvement statement Subjects were involved in this study following device implantation through the informed consent process. While subjects were not involved in the initial research design phase, their feedback influenced the implementation of remote monitoring protocols and follow-up schedules. The Merlin.net remote monitoring platform was used to transmit device data remotely, with scheduled follow-up visits at 1, 6, 12, 18 and 24 months. Subjects receive ongoing communication regarding their device data through these scheduled remote monitoring sessions. No public involvement was included in the study design or conduct.

Provenance and peer review Not commissioned; externally peer reviewed.

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

References

¹ Saver JL. CLINICAL PRACTICE. Cryptogenic Stroke. N Engl J Med 2016; 374: 2065–74. doi:10.1056/NEJMcp1503946

² Adams HP, Bendixen BH, Kappelle LJ, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 1993; 24: 35–41. doi:10.1161/01.str.24.1.35

³ Kleindorfer DO, Towfighi A, Chaturvedi S, et al. 2021 Guideline for the Prevention of Stroke in Patients With Stroke and Transient Ischemic Attack: A Guideline From the American Heart Association/American Stroke Association. Stroke 2021; 52: e364–467. doi:10.1161/STR.0000000000000375

⁴ Sposato LA, Cipriano LE, Saposnik G, et al. Diagnosis of atrial fibrillation after stroke and transient ischaemic attack: a systematic review and meta-analysis. Lancet Neurol 2015; 14: 377–87. doi:10.1016/S1474-4422(15)70027-X

⁵ Boriani G, Vitolo M, Imberti JF, et al. What do we do about atrial high rate episodes? Eur Heart J Suppl 2020; 22: O42–52. doi:10.1093/eurheartj/suaa179

⁶ Healey JS, Connolly SJ, Gold MR, et al. Subclinical atrial fibrillation and the risk of stroke. N Engl J Med 2012; 366: 120–9. doi:10.1056/NEJMoa1105575

⁷ Boriani G, Bonini N, Vitolo M, et al. Asymptomatic vs. symptomatic atrial fibrillation: Clinical outcomes in heart failure patients. Eur J Intern Med 2024; 119: 53–63. doi:10.1016/j.ejim.2023.09.009

⁸ Gladstone DJ, Spring M, Dorian P, et al. Atrial fibrillation in patients with cryptogenic stroke. N Engl J Med 2014; 370: 2467–77. doi:10.1056/NEJMoa1311376

⁹ Sanna T, Diener H-C, Passman RS, et al. Cryptogenic stroke and atrial fibrillation. N Engl J Med 2014; 371: 1261. doi:10.1056/NEJMc1409495

¹⁰ Triantafyllou S, Katsanos AH, Dilaveris P, et al. Implantable Cardiac Monitoring in the Secondary Prevention of Cryptogenic Stroke. Ann Neurol 2020; 88: 946–55. doi:10.1002/ana.25886

¹¹ Joglar JA, Chung MK, Armbruster AL, et al. 2023 ACC/AHA/ACCP/HRS Guideline for the Diagnosis and Management of Atrial Fibrillation: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2024; 149: e1–156. doi:10.1161/CIR.0000000000001193

¹² Kalarus Z, Mairesse GH, Sokal A, et al. Searching for atrial fibrillation: looking harder, looking longer, and in increasingly sophisticated ways. An EHRA position paper. Europace 2023; 25: 185–98. doi:10.1093/europace/euac144

¹³ Ip J, Jaffe B, Castellani M, et al. Accuracy of arrhythmia detection in implantable cardiac monitors: A prospective randomized clinical trial comparing Reveal LINQ and Confirm Rx. Pacing Clin Electrophysiol 2020; 43: 1344–50. doi:10.1111/pace.14076

¹⁴ Ziegler PD, Rogers JD, Ferreira SW, et al. Long-term detection of atrial fibrillation with insertable cardiac monitors in a real-world cryptogenic stroke population. Int J Cardiol 2017; 244: 175–9. doi:10.1016/j.ijcard.2017.06.039

¹⁵ Boriani G, Auricchio A, Botto GL, et al. Insertable cardiac monitoring results in higher rates of atrial fibrillation diagnosis and oral anticoagulation prescription after ischaemic stroke. Europace 2023; 25: euad212. doi:10.1093/europace/euad212

¹⁶ Kirchhof P, Toennis T, Goette A, et al. Anticoagulation with Edoxaban in Patients with Atrial High-Rate Episodes. N Engl J Med 2023; 389: 1167–79. doi:10.1056/NEJMoa2303062

¹⁷ Shoamanesh A, Field TS, Coutts SB, et al. Apixaban versus aspirin for stroke prevention in people with subclinical atrial fibrillation and a history of stroke or transient ischaemic attack: subgroup analysis of the ARTESiA randomised controlled trial. Lancet Neurol 2025; 24: 140–51. doi:10.1016/S1474-4422(24)00475-7

¹⁸ Svendsen JH, Diederichsen SZ, Højberg S, et al. Implantable loop recorder detection of atrial fibrillation to prevent stroke (The LOOP Study): a randomised controlled trial. Lancet 2021; 398: 1507–16. doi:10.1016/S0140-6736(21)01698-6

¹⁹ Healey JS, Lopes RD, Granger CB, et al. Apixaban for Stroke Prevention in Subclinical Atrial Fibrillation. N Engl J Med 2024; 390: 107–17. doi:10.1056/NEJMoa2310234

²⁰ Brachmann J, Morillo CA, Sanna T, et al. Uncovering Atrial Fibrillation Beyond Short-Term Monitoring in Cryptogenic Stroke Patients: Three-Year Results From the Cryptogenic Stroke and Underlying Atrial Fibrillation Trial. Circ Arrhythm Electrophysiol 2016; 9: e003333. doi:10.1161/CIRCEP.115.003333

²¹ Boriani G, Vitolo M, Diemberger I, et al. Optimizing indices of atrial fibrillation susceptibility and burden to evaluate atrial fibrillation severity, risk and outcomes. Cardiovasc Res 2021; 117: 1–21. doi:10.1093/cvr/cvab147

²² Benjamin EJ, Go AS, Desvigne-Nickens P, et al. Research Priorities in Atrial Fibrillation Screening: A Report From a National Heart, Lung, and Blood Institute Virtual Workshop. Circulation 2021; 143: 372–88. doi:10.1161/CIRCULATIONAHA.120.047633

²³ Glotzer TV, Hellkamp AS, Zimmerman J, et al. 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 2003; 107: 1614–9. doi:10.1161/01.CIR.0000057981.70380.45

²⁴ Boriani G, Glotzer TV, Santini M, et al. Device-detected atrial fibrillation and risk for stroke: an analysis of >10 000 patients from the SOS AF project (Stroke preventiOn Strategies based on Atrial Fibrillation information from implanted devices). Eur Heart J 2014; 35: 508–16. doi:10.1093/eurheartj/eht491

²⁵ Yaghi S, Ryan MP, Gunnarsson CL, et al. Longitudinal outcomes in cryptogenic stroke patients with and without long-term cardiac monitoring for atrial fibrillation. Heart Rhythm O2 2022; 3: 223–30. doi:10.1016/j.hroo.2022.02.006

²⁶ Tsivgoulis G, Katsanos AH, Grory BM, et al. Prolonged Cardiac Rhythm Monitoring and Secondary Stroke Prevention in Patients With Cryptogenic Cerebral Ischemia. Stroke 2019; 50: 2175–80. doi:10.1161/STROKEAHA.119.025169

²⁷ Correction to: Searching for Atrial Fibrillation Poststroke: A White Paper of the AF-SCREEN International Collaboration. Circulation 2020; 141: e99. doi:10.1161/CIR.0000000000000762

²⁸ Kühne M, Krisai P, Coslovsky M, et al. Silent brain infarcts impact on cognitive function in atrial fibrillation. Eur Heart J 2022; 43: 2127–35. doi:10.1093/eurheartj/ehac020

²⁹ Koh KT, Law WC, Zaw WM, et al. Smartphone electrocardiogram for detecting atrial fibrillation after a cerebral ischaemic event: a multicentre randomized controlled trial. Europace 2021; 23: 1016–23. doi:10.1093/europace/euab036

³⁰ Vitolo M, Imberti JF, Maisano A, et al. Device-detected atrial high rate episodes and the risk of stroke/thrombo-embolism and atrial fibrillation incidence: a systematic review and meta-analysis. Eur J Intern Med 2021; 92: 100–6. doi:10.1016/j.ejim.2021.05.038

³¹ Lopez Perales CR, Van Spall HGC, Maeda S, et al. Mobile health applications for the detection of atrial fibrillation: a systematic review. Europace 2021; 23: 11–28. doi:10.1093/europace/euaa139

³² Skrebelyte-Strøm L, Rønning OM, Dahl FA, et al. Prediction of occult atrial fibrillation in patients after cryptogenic stroke and transient ischaemic attack: PROACTIA. Europace 2022; 24: 1881–8. doi:10.1093/europace/euac092

³³ McIntyre WF, Wang J, Benz AP, et al. Estimated incidence of previously undetected atrial fibrillation on a 14-day continuous electrocardiographic monitor and associated risk of stroke. Europace 2022; 24: 1058–64. doi:10.1093/europace/euab324

³⁴ Miyazawa K, Pastori D, Martin DT, et al. Characteristics of patients with atrial high rate episodes detected by implanted defibrillator and resynchronization devices. Europace 2022; 24: 375–83. doi:10.1093/europace/euab186