Literature Analysis
This summary of the evidence has been formatted by the type of comparator, outcome category, (e.g., diagnostic yield, undesirable effects), and study design. The literature analysis emphasizes the research designs representative of the body of evidence (systematic reviews, meta-analyses) and the primary studies most applicable to specific outcomes e.g., randomized controlled trials (RCTs) for diagnostic yield. Multiple publications from the same dataset were grouped together in the analysis regardless of study design. The GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) approach domains of study limitations (risk of bias), indirectness (applicability), imprecision, inconsistency, and publication bias formed the basis of appraisal of the certainty of evidence [see Attachments, Tables 1 and 2].
Description of Studies
This description of studies is supplemented by tables 3–8 (see Attachments) that further summarize the characteristics, findings, and assessments of included studies.
This evidentiary review included 12 systematic reviews, most with meta-analysis [Table 3]. These reviews synthesized studies for all types of TNACMD. Only 2 reviews were limited to the analysis of RCTs. The number of primary studies and participants in the reviews ranged from 3 to 50 and 1149 to 135,300, respectively. The patient populations varied from asymptomatic older adults to individuals diagnosed with stroke or TIA. The detection of AF was a common primary outcome [Table 4]. Implantable cardiac monitors were the most frequently employed comparator. Five systematic reviews assessed clinical outcomes. Three syntheses reported on undesirable effects. Two reviews described the effect of TNACMD on clinical management. Five of the systematic reviews provided data about potential effect modifiers, patient selection, or the timing of TNACMD.
A total of 9 RCTs were included in the evidence review [Table 5]. Most trials were multi-centered and took place in North America. The number of participants varied widely from 21 to 11,931. The mean/median ages of participants ranged from 64.1 to 75 years. These studies included multi-morbid patients without known AF who had cryptogenic ischemic stroke, TIA, risk factors for stroke following cardiac surgery, or received opportunistic screening. Five trials assessed long-term cardiac monitors (LTCM) i.e., patch devices. Three RCTs evaluated Holter type monitoring devices. Four studies investigated event monitors e.g., ELR. There were no RCTs concerning MCT or MCOT devices that met eligibility criteria. Comparator interventions included usual care, implantable cardiac monitors, and between different types of TNACMD. Most studies reported diagnostic yield (detection of AF) as a primary endpoint [Table 6]. Primary and secondary outcomes were assessed at follow-up time points between 14 days and 15 months. A single large retrospective cohort study of Medicare beneficiaries (N = 287,789) was also included in the analysis.
There were 6 non-randomized studies of an intervention (NRSI) identified that investigated outcomes associated with the use of MCT devices on single cohorts that received TAVR [Table 7]. Four studies were conducted prospectively. Two trials evaluated the impact of TNACMD when used during the peri-procedural period (i.e., before and after TAVR). A single study assessed the value of a 1-week monitoring period prior to the implantation of a TAVR device. Three case series reported on the results of the extended application of real-time MCT ECG monitoring immediately following TAVR implantation. Five of the trials were performed at single centers in the USA or Canada. A fifth study involved two locations. Multiple types of MCT devices were used in the studies, with BodyGuardian™ and Pocket-ECG the most frequently reported. Monitoring periods were predominantly in the range of 28-30 days. Most studies reported on the incidence and type of arrhythmic events and consequential therapeutic changes [Table 8]. Other less frequently reported outcomes included feasibility (compliance, notifications) and mortality. All the studies were small, ranging from 23-192 participants obtained from convenience samples.
Effects of TNACMD
TNACMD compared to Standard of Care (SOC)
Diagnostic Yield
Two evidence syntheses analyzed data from primary studies that compared various TNACMD with SOC including short-term cardiac monitoring.4,5 Dahal et al.4 performed a meta-analysis of 4 RCTs (N = 1149) that evaluated the prolonged monitoring (LTCM, MCOT, ELR) ≥7 days in patients with cryptogenic stroke or TIA. The reviewers found prolonged cardiac monitoring of ≥7 days compared to shorter cardiac monitoring of ≤48 hours duration increased the detection of AF (≥30 seconds duration) in patients after cryptogenic stroke or TIA (13.8% vs. 2.5%; odds ratio [OR] 6.4; 95% confidence interval [CI], 3.50, 11.73; P < 0.00001; I2 = 0%]. It also increased the odds of AF detection of any duration (22.6% vs. 5.2%; 5.68 [3.3, 9.77]; P < 0.00001; I2 = 0%). This meta-analysis has several limitations including lack of patient-level data, different duration of follow-up among studies, and variations in type of device. It was not possible to ascertain which device and follow-up duration worked better compared to the others.
A health technology assessment (HTA) employed a network meta-analysis, with SOC (short-term Holter monitoring) as the common comparator, in assessing the relative effectiveness of LTCM compared with ELRs.5 This HTA included 12 non-randomized studies of interventions (NRSI) that reported on patients with palpitations, syncope, dizziness, stable chronic heart failure, ischemic and non-ischemic stroke, cryptogenic stroke, TIA, or known/suspected dysrhythmias. Both types of devices were more effective than SOC. There was no substantial difference between them in their ability to detect symptoms (risk difference 0.01; 95% CI: –0.18, 0.20). Using GRADE for network meta-analysis, the quality of the evidence was evaluated as low. Additional constraints were that all primary studies were observational designs, and only indirect comparisons could be obtained.
Six RCTs reported on the detection of AF by different TNACMD as compared to SOC that included 24-hour Holter monitoring.6-11 These studies included patients without known AF who had cryptogenic ischemic stroke, TIA, risk factors for stroke following cardiac surgery, or received asymptomatic screenings. Monitoring periods ranged from 14 to 90 days, with detection endpoints extending to as much as 15.3 months. The TNACMD strategies were superior in detecting AF versus short-term monitoring at all follow-up periods (TNACMD = median 16.3% [14-19.6], SOC = 2.1% [1.7-5]) except for opportunistic screening (TNACMD = 0.7%, SOC = 0.6%).
Cumulative AF or atrial flutter (AFL) was calculated in 3 RCTs. Within the first 30 days of monitoring, AF/AFL of >6 and >24 hours durations were detected in 8.6% and 3.1% of the TNACMD group versus 0% in the SOC group, respectively.7 Paroxysmal atrial fibrillation (PAF) of any duration was identified in 1 RCT at 28 days (14% in the TNACMD group and 2.1% in the SOC group)9 and through 90 days (44% in the TNACMD group and 4% in the SOC group).8
Clinical Management
Dahal et al.4 meta-analyzed data from 4 RCTs (N = 1149). Patients diagnosed with cryptogenic stroke or TIA who underwent prolonged monitoring (LTCM, ELR, MCOT) were more likely to be on anticoagulation at follow-up than those receiving SOC (2.21[1.52, 3.21]; P < 0.0001; I2, 0%).
There were 4 RCTs that explored the impact of TNACMD on the initiation of oral anti-coagulant (OAC) therapy.6,8-10 Higgins et al.8) found absolute differences favoring the TNACMD group of 16% at 14- and 90-day time points. Another RCT reported an absolute difference of 7.5% in favor of TNACMD compared to SOC at 90 days.6 Two RCTs provided insufficient details (uncertainties about the absolute effects and time frame) to permit conclusions about the effects of TNACMD on clinical management.9,10
Mortality, Stroke, Major Adverse Cardiac Events
A single systematic review and meta-analysis of 4 RCTs involving 1149 patients with cryptogenic stroke or TIA who underwent prolonged monitoring (LTCM, ELR, MCOT) or SOC (short-term Holter monitoring) found no differences in mortality (1.33 [0.29, 6.00]; P = 0.71; I2 = 0%] or recurrent stroke or TIA (0.78 [0.40, 1.55]; P = 0.48; I2, 0%).4
Two trials investigated mortality, finding no significant difference with SOC at 90 days between patients assigned to a 7-day external loop recorder (ELR)8 or 14-day patch monitor.9
There was no evidence that screening for AF using a 14-day continuous TNACMD reduces recurrent stroke or TIA.8,9 The data were from 2 small RCTs (N = 100, 116), and event rates were low. This may have resulted in underestimating the effect of TNACMD.
Healthcare Utilization
There was no indication, from a single RCT, that screening for AF using a 14-day continuous TNACMD in people ≥70 years of age seen in primary care practice reduces stroke-related hospitalizations.10 However, the event rates were low, which may have resulted in underestimating the effect of TNACMD.
Undesired Effects
There were no serious adverse events associated with the use of the ELR and patch devices.7,8,10 A single RCT reported 10.4% of participants experienced a device-related adverse event due to skin irritation from the adhesive material of the wearable patch within 30 days of randomization.7
Adherence
Data from a single trial found that among the patients in the intervention group who started to undergo monitoring, 233 of 284 (82.0%) completed 3 or more weeks of monitoring.6
Effect Modifiers
No studies were identified.
TNACMD compared to Implantable Cardiac Monitors
Diagnostic Yield
Six systematic reviews evaluated the effectiveness of implantable loop recorder (ILR) versus TNACMD for identifying AF in patients with any type of stroke (e.g., cryptogenic, ischemic) or TIA.12-17 Broadly, these reviews concluded that for patients with sufficient cognitive and physical ability to comply with the use of TNACMD, a 1-month duration can capture a significant proportion of AF and should be considered in place of ILRs. The choice of specific type of TNACMD should be made based on availability, patient’s preferences, and clinical judgment. ILRs can be considered in patients where a prolonged duration of monitoring (>30 days) is anticipated, if a TNACMD fails to detect any AF after 4 weeks of monitoring, or if there are anticipated issues with compliance.
A single RCT sought to determine, in patients with a recent ischemic stroke, whether 12 months of ILR monitoring detected more occurrences of AF compared with conventional external loop recorder (ELR) monitoring for 30 days.18 There was a statistically significant between-group difference in the rate of detection of AF or flutter lasting >2 minutes by 12 months (15.3% in the ILR group vs 4.7% in the prolonged ELR group).
Clinical Management
Al Qurashi et al.13 conducted a systematic review and meta-analysis to assess the efficacy of implantable cardiac monitors (ICMs) compared to a range of TNACMD in detecting post-stroke AF and the subsequent use of oral anti-coagulant therapy. The review included 3 RCTs of 1233 patients with any type of stroke and no prior AF or flutter. The ICM arm had significantly higher usage of oral anticoagulants as compared to the TNACMD arm. (RR = 2.76, 95% CI: 1.89, 4.02, P < 0.05). Limitations of this review included the small number of eligible studies and the need to assume the baseline characteristics of patients in all the included trials were similar.
Mortality, Stroke, Major Adverse Cardiac Events
A single systematic review and meta-analysis reported showed there were no significant differences between ICMs and TNACMD in the reduction of mortality (RR = 0.67, 95% CI: 0.31, 1.45; P = 0.42; I2 = 0%) and in the recurrence of stroke (RR = 0.73, 95% CI: 0.48, 1.10; P = 0.14; I2 = 0%).13
Healthcare Utilization
No studies found.
Undesired Effects
A systematic review with meta-analysis found ICM usage was associated with a higher incidence of mild to moderate adverse events (RR = 10.52, 95% CI:1.35, 82.14; P = 0.02) and a higher number of severe adverse events as compared to the use of TNACMD (RR = 7.61, 95% CI: 1.36, 42.51; P = 0.02).13
Effect Modifiers
Noubiap et al.17 performed a systematic review and meta-analysis that explored factors influencing AF detection rates. The reviewers found 1) a steady increase of AF rates with duration of monitoring; 2) higher rates of AF in patients with ESUS compared to those with cryptogenic stroke; and 3) the association of older age, CHA2DS2-VASc score, PA-TDI Interval, left atrial enlargement, P wave maximal duration, prolonged PR interval and atrial runs with higher rates of AF detection.
Comparison of Different Types of TNACMD
A network meta-analysis found that LTCM and ELR devices were equally effective in their ability to detect symptomatic cardiac arrhythmias.5
Reynolds, et al.19 retrospectively analyzed the USA National Medicare claims data of 287,789 beneficiaries without a preceding established arrhythmia diagnosis. The aim of the study was to make comparisons among different device types (ELRs, LTCM, Holter, and MCT) across a range of outcomes. The device-specific analysis showed that compared to Holter, ELR, MCT, or other LTCM manufacturers, a specific LTCM (Zio® XT 14-day patch, iRhythm Technologies, San Francisco, CA) had the highest adjusted odds of AF detection and lowest adjusted odds of TNACMD retesting. Findings were consistent for specific arrhythmia diagnoses of ventricular tachycardia, atrioventricular block, and PAF. The annualized all-cause inpatient hospitalizations during the follow-up period were lowest in the LTCM and Holter cohorts (mean = 0.45 stays for both cohorts), followed by the ELR and MCT cohorts (mean 0.60 stays for both cohorts). Annualized follow-ups for all-cause ED visits were lowest in the LTCM cohort (mean = 0.70 visits). Meanwhile, outpatient visits were lowest in the Holter and ELR cohorts (mean = 24.5 visits for both cohorts); however, patients in the LTCM cohort experienced the smallest increase in outpatient visits over baseline (mean change = 3.11 for the LTCM cohort). This study included limitations inherent in retrospective observational designs, the inability to adjust for the duration of monitoring, potential for industry funded bias (6 of 8 authors were affiliated with the sponsor), and the likelihood of confounding due to the influence of patient, provider, and system factors.
Non-Comparative Studies
Diagnostic Yield
Six NRSI reported on the incidence, duration, and type (e.g., AF, bradycardia, high-degree atrioventricular block) of arrhythmic events identified by TNACMD either before and/or after implantation of a TAVR device.20-25 One methodologically weak study found arrhythmic events were diagnosed in almost half of patients.20 The other 5 studies reported on a more narrow range (7.3% – 10%) of new arrhythmic events.
Clinical Management
Three NRSI described changes in therapeutic management – mostly permanent pacemaker implantation – as a result of extended cardiac monitoring in patients undergoing TAVR.20,21,23 The implementation of TNACMD prior to TAVR resulted in a change in therapeutic management (anti-coagulation medication, permanent pacemaker) for more than a quarter of patients.20 The use of TNACMD post-TAVR resulted in smaller proportions (3.1% – 6%) of patients undergoing permanent pacemaker implantation.21,23
Mortality, Stroke, Major Adverse Cardiac Events
A small non-comparative study (N=54) did not record any mortalities for patients using an MCT monitor for 30 days following TAVR.23
Undesired Effects
No studies were identified.
Adherence
Two NRSI assessed patient compliance with using MCT technologies prior to or immediately following the TAVR procedure.21,23 Overall adherence for at least 14 days was high (87.5% - 96%).
Effect Modifiers
No studies were identified.
Summary of Main Results
Prolonged cardiac monitoring with TNACMD is more effective than SOC in the detection of AF in patients after stroke or TIA, or with symptoms associated with an increased risk of AF. For patients with the ability to comply with the use of TNACMD, a 30-day duration can capture a significant proportion of AF and should be considered in lieu of ILRs. Patients using TNACMD are more likely to receive guideline-concordant anti-coagulant treatment compared to those undergoing SOC. In contrast, individuals with ICMs are more likely to be prescribed timely anti-coagulant therapy than those wearing TNACMD. There were no significant differences between SOC, ICMs and TNACMD in the reduction of mortality, or the recurrence of stroke or TIA. The likelihood of AF detection steadily increases with the duration of monitoring and patient age. It is uncertain whether any type of TNACMD is clinically superior to other externally wearable devices. TNACMD may improve the detection of conduction disturbances pre/post-TAVR, which may then result in meaningful changes in therapeutic management.
Applicability of Evidence
The included studies were highly applicable to the Medicare beneficiary population. The largest study explicitly focused on the Medicare population. Most trials were multi-centered and took place in North America. The mean/median ages of participants ranged from 64.1 to 84 years. The included studies reported on multi-morbid patients.
Quality of Evidence
The overall certainty of evidence was rated as moderate for clinical, healthcare management/utilization, and undesirable effects outcomes. There were some concerns about the risk of bias (selection and performance bias). There was no serious indirectness, inconsistency, imprecision, or risk of publication bias.
The certainty of evidence for the comparative effectiveness of different TNACMD was judged to be very low (observational data, high risk of confounding, and the potential for industry funding bias).
The certainty of evidence for the use of TNACMD for patients undergoing TAVR was low to very-low (non-comparative observational studies, small sample sizes, high risk of selection bias, and limited generalizability).
Clinical Guidelines and Positions of National and Specialty Organizations
A total of 22 publications developed by 15 professional societies or agencies were identified to supplement the summary of evidence [see Attachments, Table 9].34-55 These guidelines encompassed a broad range of disorders (stroke, TIA, cardiac arrhythmias, cardiac conduction delay, syncope, silent cerebrovascular disease, hypertrophic cardiomyopathy, severe aortic stenosis) and monitoring during AF ablation.
Collectively, there was a strong consensus in favor of extended cardiac monitoring in situations where AF cannot readily be diagnosed with standard ECG. There was limited explicit guidance regarding the optimal monitoring devices. Most guidance recommended that the frequency of symptoms should dictate the type of monitoring device. A single guideline proposed a clinical workflow commencing with continuous (short-term 24 hour and up to 7 days) ambulatory ECG monitoring, which if unsuccessful, is followed by intermittent external loop recording (long term from weeks to months). For those individuals remaining undiagnosed after prolonged noninvasive monitoring, ILR may be necessary.48 An American College of Cardiology expert panel favored prolonged ECG monitoring (>14 days) during the post-procedural period following TAVR due to the increased risk of arrhythmic events.55