ABSTRACT
Background
The rapid evolution of cardiac implantable electronic devices (CIEDs) has increased remote transmission data, leading to excessive non‐actionable alerts (NAA) and alert fatigue.
Objective
Optimize alert parameters to minimize NAA and evaluate the impact on clinical outcomes.
Methods
We included 536 participants (mean age 75 (15) years, 60.4% male, 83.4% white) with CIEDs. In 413 patients, CIEDs were reprogrammed to censor alerts as follows: atrial fibrillation (AF) episodes < 5.5 h, persistent AF > 1 month with prior alerts, AF < 24 h on anticoagulation or with prior appendage occlusion, and non‐sustained ventricular tachycardia (NSVT) in defibrillator platforms. NAAs were tracked 90‐days pre‐ and post‐reprogramming. Incident ischemic stroke and sudden cardiac death (SCD) were assessed over a median 1.8‐year follow‐up. Logistic regression models examined associations between reprogramming and outcomes.
Results
Reprogramming was implemented for AF alerts (69.5%, n = 287) and NSVT alerts (30.5%, n = 126). After reprogramming, NAAs significantly decreased from 6.68 (SD = 10.02) to 2.27 (SD = 4.58), p < 0.001. During follow‐up, ischemic stroke rates in AF patients were similar between reprogrammed (5.2%, n = 15) and control groups (5.4%, n = 5). In those with NSVT alerts, SCD incidence was lower in reprogrammed (2.3%, n = 3) versus controls (9.3%, n = 3). In logistic regression models adjusted for demographics, CHA₂DS₂VASC score, anticoagulation status, and prior stroke history, there was no statistically significant difference in stroke risk between groups (OR 0.82 [0.27–2.51]).
Conclusions
Guideline‐based alert parameters in CIED patients significantly reduced NAA burden with no increasing in adverse outcomes in patients with device‐detected AF or NSVT alerts. This approach may reduce noise and safely improve efficiency.
Keywords: AF, device clinic, implantable cardiac devices, NSVT, remote monitoring
1. Introduction
Remote monitoring of cardiac implantable electronic devices (CIEDs) has become the standard of care, supported by a Class I recommendation based on robust evidence from multiple randomized controlled trials [1, 2, 3, 4, 5]. Remote transmissions include routine interrogations scheduled every 3 to 6 months, alongside automated alerts triggered by specific device‐detected clinical events meeting predefined parameters. While advancements in remote monitoring technology have significantly enhanced data transmission capabilities, estimated to generate nearly 10 million routine transmissions annually in the United States [6]. Although valuable, this surge in information often reduces the signal‐to‐noise ratio, with critical alerts buried among a flood of less relevant data. As a result, device clinics are frequently overwhelmed by numerous non‐actionable alerts (NAAs), which are often clinically irrelevant and hinder efficient patient care. Additionally, reimbursement policies that emphasize volume over efficiency further complicate care delivery, underscoring the need for both technical and policy solutions.
Despite sharing common monitoring principles, CIED systems vary significantly among manufacturers in terms of programmable alerts and proprietary algorithms. The absence of standardized alert configurations has left many devices relying on preset settings from the manufacturers, limiting the potential for uniformity. Clinical guidelines providing specific recommendations for alert customization remain scarce [7]. This lack of standardization highlights the need to optimize configurations to improve efficiency and reduce alert fatigue in clinical practice.
We reprogrammed alert parameters in a clinically rational manner (as agreed upon unanimously by electrophysiologists at our institution) with the goal of safely reducing the frequency of NAAs related to subclinical atrial fibrillation (AF) and non‐sustained ventricular tachycardia (NSVT) in defibrillator platforms. This study is a retrospective analysis of the effects of those programming changes. Additionally, we assessed the effects of these programming changes on the incidence of ischemic stroke in patients with AF and sudden cardiac death (SCD) in patients with NSVT.
2. Methods
2.1. Study Population
As part of our effort to integrate clinical guidelines, supported by prior literature and expert opinions from electrophysiologists, we implemented an alert reprogramming protocol to censor NAA in all CIEDs at a single academic medical center. For this study, we included 536 patients above 18 years old with CIEDs. Eligible devices included permanent pacemaker, implantable cardioverter‐defibrillator (ICD), or implantable loop recorders that is not implanted for cryptogenic stroke. Patients who did not have alerts for AF or NSVT on their CIEDs were excluded from the study. Additionally, those with less than 3 months of remote transmission data or follow up data were excluded as well.
2.2. Alerts Reprogramming Changes
Reprogramming changes were implemented in 413 patients between April 2022 and December 2022. In these patients, CIEDs alerts programming for sensing AF in all CIEDs, and NSVT on defibrillator platforms were adjusted during an in person visit. AF alerts were triggered by the duration and persistence of AF. Alerts were programmed off for the following: AF episodes lasting less than 5.5 h, persistent AF episodes for more than 1 month which had generated an alert previously, AF episodes < 24‐h in patients with known paroxysmal AF on oral anticoagulants (OAC) or underwent appendage closures or surgical excision. Alerts for NSVT in patients with defibrillators were programmed off as well. The number of NAA for each patient were tallied 3 months before the reprogramming date and 3 months after. We identified a group of 123 patients that did not attend an in‐person visit during the study timeframe. Consequently, these patients did not have any alert reprogramming changes during the time window of interest and were chosen as a control group for the study. Clinical outcomes were assessed retrospectively through a medical chart review, with a follow‐up period of 1.8 years. Patients' devices were from the following manufacturers/monitoring platforms: Boston Scientific/Latitude (Marlborough, Massachusetts), Medtronic/‐Carelink (Minneapolis, Minnesota), Abbott Medical/Merlin (Abbott Park, Illinois), Biotronik/Biotronik home monitoring (Lake Oswego, Oregon). All CIEDs data were processed remotely using a third‐party vendor (CV Remote Solutions, Greensboro, NC, USA) with access to manufacturer specific online platforms and read‐only electronic health record access. This retrospective study was reviewed and approved by the Wake Forest Human Research Ethics Committee.
2.3. Clinical Outcomes and Covariates
Patient demographics including age (continuous in years), sex (male and female), race (White, Black and Other) were collected from the electronic medical record. Anticoagulant medication intake was reported if patients had been on therapy at the time of enrollment and for at least 6 months before outcome ascertainment. In patients with AF alerts, incident ischemic stroke was defined as clinically adjudicated diagnosis of ischemic stroke by a neurologist based on clinical symptoms and confirmatory imaging either by brain CTA or MRI. Sudden cardiac death diagnosis was based on in‐hospital documentation of VT/VF before death or device transmission data at the time of death or primary cause of death in death certificate in those with NSVT alerts. Other variables including CHA2DS2‐VASc score, and prior stroke was collected from electronic medical records.
2.4. Statistical Analysis
Demographics and clinical characteristics of the participants were compared between reprogramming and control group using Analysis of variance (ANOVA) for continuous variables
(reported as mean and standard deviation for normally distributed data or median and interquartile range for skewed data) and Chi‐square for categorical variables (presented as number and percentage). A paired samples t‐test was conducted to evaluate the effect of reprogramming on the number of alerts over a 90‐day period. To explore the potential impact of reprogramming on clinical outcomes, we conducted multivariable logistic regression analyses. Separate models were constructed to examine the association between reprogramming and ischemic stroke in AF patients, as well as SCD in those with NSVT alerts. Patients with AF and NSVT alerts were analyzed as distinct groups to assess clinical outcomes independently. Odds ratios (OR) and 95% confidence intervals (CI) for the risk of clinical outcomes in the reprogrammed group were reported compared to the control group as a reference. For the risk of ischemic stroke in patients with AF alerts, two multivariable‐adjusted models were constructed: Model 1 adjusted for age, sex, and race, and Model 2 adjusted for model 1 plus CHA₂DS₂VASC score, anticoagulation status, and prior stroke history. In a separate analysis, a logistic regression model adjusted for demographics was used to examine the association between censoring NSVT alerts and risk of SCD.
All statistical analyses were performed using Jamovi version 2.6.19 (Sydney, Australia), and a p values of less than 0.05 were considered significant.
3. Results
For the total 536 patients (413 reprogrammed and 123 control group), 70.5% (n = 378) of patients had AF alerts as an indication for inclusion compared to 29.5% (n = 158) with NSVT alerts. Demographic characteristics included mean (SD) age 73.4 (11.8) years, white (83.4%), males (60.4%) with an average CHA₂DS₂VASC score of 4 (2). (Table 1)
Table 1.
Baseline population characteristics and clinical outcomes.
| Variable: | Total (n = 536) | Programmed (n = 413) | No programming (n = 123) | p value |
|---|---|---|---|---|
| Age, years | 75 ± 15 | 75 ± 16 | 77 ± 13.5 | 0.141 |
| Sex, males | 323 (60.4) | 255 (61.9) | 68 (55.3) | 0.187 |
| Race, whites | 447 (83.4) | 345 (83.5) | 102 (82.9) | 0.310 |
| CHA₂DS₂VASC score | 4 ± 2 | 4 ± 3 | 5 ± 2 | < 0.001 |
| Prior stroke | 80 (14.9) | 54 (13.1) | 26 (21.1) | 0.023 |
| Deceased | 56 (10.5) | 37 (8.9) | 19 (15.5) | 0.041 |
| Anticoagulation: | ||||
| Oral anticoagulants | 237 (44.2) | 193 (46.7) | 44 (35.7) | 0.179 |
| Aspirin | 135 (25.2) | 87 (21.1) | 48 (39) | < 0.001 |
| Both OAC & aspirin | 126 (23.5) | 108 (26.1) | 18 (14.6) | 0.031 |
| LAAO | 12 (2.2) | 7 (1.7) | 5 (4.1) | 0.124 |
| None | 26 (4.9) | 18 (4.4) | 8 (6.5) | 0.980 |
| Device alerts: | ||||
| AF | 378 (70.5) | 287 (69.5) | 91 (74) | 0.686 |
| NSVT | 158 (29.5) | 126 (30.5) | 32 (26) | 0.468 |
| Ischemic strokea | 20 (3.7) | 15 (3.6) | 5 (4) | 0.831 |
| SCDb | 6 (1.1) | 3 (0.7) | 3 (2.4) | 0.112 |
Note: Categorical data presented as numbers (percentage %), continuous data presented as Mean ± SD for normally distributed variables or Median ± IQR for skewed data. All patients with NSVT alerts) had implantable cardioverter‐defibrillators (ICDs) or cardiac resynchronization therapy defibrillators.
Abbreviations: AF, atrial fibrillation; LAAO, left atrial appendage occlusion device; NSVT, non‐sustained ventricular tachycardia; SCD, sudden cardiac death.
Ischemic stroke was assessed in AF alerts patients.
SCD was assessed in NSVT alerts patients.
A breakdown of device types and vendors across the study sample is presented in (Table 2). Pacemakers were the most common devices in both groups, comprising 41.2% of the reprogrammed cohort and 52.8% of the control group, followed by ICDs at 37.8% and 34.1%, respectively. The NAA burden was assessed for 90 days before and after reprogramming. Alert reprogramming was performed between April 2022 and December 2022, therefore data collection extended from January 2022 through March 2023. Adjusted alert settings led to significant reduction in NAA. The mean number of alerts decreased from 6.68 (10.02) to 2.27 (4.58) post‐reprogramming, with a median reduction of 3 alerts (p < 0.001). The mean difference was 4.41 (SE = 0.480). (Figure 1) This reduction was consistent across all device types. Implantable loop recorders (ILR)/Insertable cardiac monitors (ICM) demonstrated the greatest relative reduction in NAAs by 80.9%, followed by ICDs with 73.3% relative reduction rate. Overall, reprogramming led to a 74.2% total reduction in NAA volume. (Table 3) During follow‐up, the incidence of ischemic stroke in AF patients was 5.2% (n = 15) of the reprogrammed group compared to 5.4% (n = 5) in the control group. (Table 4) Of the 15 patients in the reprogrammed group who suffered a stroke during follow‐up, 13 were taking OAC and the two patients that were not had had their OAC stopped due to bleeding concerns. All five stroke cases in the control group were on OAC. In those with NSVT alerts, SCD occurred in 2.3% (n = 3) of the reprogrammed group versus 9.3% (n = 3) in the control group. (Table 5). Due to the low number of cases, the logistic regression models did not yield statistically significant results. However, there was a trend suggesting that reprogramming was not associated with an increased risk of ischemic stroke in AF patients. Adjusting for age, race, sex, anticoagulation status, prior stroke, and CHA₂DS₂‐VASc score, the results remained in the same direction but did not reach statistical significance (OR [95% CI]: 0.82 [0.27–2.51], p = 0.73). Similarly, in patients with NSVT alerts, an adjusted logistic regression model accounting for demographic factors showed no clear increase in SCD risk with alert censoring, however, results were not statistically significant (OR [95% CI]: 0.28 [0.05–1.53], p = 0.14). (Table 6).
Table 2.
Type of cardiac implantable electronic devices.
| Device type | Total (n = 536) | Programmed (n = 413) | No Programming (n = 123) |
|---|---|---|---|
| Pacemakers (PPM) | 235 (43.8%) | 170 (41.2%) | 65 (52.8%) |
| Implantable cardioverter‐defibrillator | 198 (36.9%) | 156 (37.8%) | 42 (34.1%) |
| Insertable cardiac monitor/implantable loop recorder | 103 (19.2%) | 87 (21.1%) | 16 (13.0%) |
| Vendor: | |||
| Medtronic | 122 (22.8%) | 94 (22.8%) | 28 (22.8%) |
| Biotronik | 167 (31.2%) | 56 (13.6%) | 111 (90.2%) |
| Boston scientific | 87 (16.2%) | 35 (8.5%) | 52 (42.3%) |
| Abbott | 158 (29.5%) | 42 (10.2%) | 116 (94.3%) |
| Other | 9 (1.7%) | 6 (1.5%) | 3 (2.4%) |
Figure 1.
Effects of alerts reprogramming on number of non‐actionable alerts. Figure illustrates a significant reduction in NAAs within 90 days following alert reprogramming. The mean number of alerts dropped substantially (6.68 ± 10.02 to 2.27 ± 4.58, p < 0.001), t‐tests and non‐parametric analyses confirming the findings (p < 0.001).
Table 3.
Non‐actionable alerts reductions by device type.
| Device type | Pre‐reprogramming alerts | Post‐reprogramming alerts | % Reduction |
|---|---|---|---|
| ILRs/ICMs | 978 | 187 | 80.9% |
| ICDs | 1,842 | 492 | 73.3% |
| Pacemakers | 623 | 210 | 66.3% |
| Total | 3,443 | 889 | 74.2% |
Abbreviations: ICD, implantable cardioverter‐defibrillator; ICM, insertable cardiac monitor; ILR, implantable loop recorder.
Table 4.
Baseline characteristics of patients with AF alerts by reprogramming status.
| Variable: | Total (n = 378) | Programmed (n = 287) | No programming (n = 91) | p value |
|---|---|---|---|---|
| Age, years | 78 ± 12 | 78 ± 13 | 77.3 ± 9.17 | 0.410 |
| Sex, males | 224 (59.4) | 177 (61.6) | 47 (51.6) | 0.084 |
| Race, Whites | 332 (87.8) | 253 (88.1) | 79 (86.8) | 0.443 |
| CHA2DS2‐VASc score | 4 ± 2 | 4 ± 3 | 5 ± 2 | < 0.001 |
| Prior stroke | 59 (15.6) | 39 (13.5) | 20 (21.9) | 0.061 |
| Deceased | 45 (11.9) | 28 (9.7) | 17 (18.6) | 0.024 |
| Anticoagulation: | ||||
| Oral anticoagulants | 200 (52.9) | 159 (55.4) | 41 (45.1) | 0.081 |
| Aspirin | 51 (13.5) | 27 (9.4) | 24 (26.3) | < 0.001 |
| Both OAC & aspirin | 103 (27.2) | 85 (29.6) | 18 (19.7) | 0.077 |
| LAAO | 12 (3.2) | 7 (2.4) | 5 (5.4) | 0.154 |
| None | 12 (3.2) | 9(3.1) | 3 (3.2) | 0.932 |
| aIschemic stroke | 20 (5.3) | 15 (5.2) | 5 (5.4) | 0.920 |
Note: Categorical data presented as numbers (percentage %), continuous data presented as Mean ± SD for normally distributed variables or Median ± IQR for skewed data.
Abbreviations: AF, atrial fibrillation; OAC, oral anticoagulants.
In the reprogrammed group, 13 of 15 stroke cases were on OAC; 2 had discontinued due to bleeding. All 5 stroke cases in the control group were on OAC.
Table 5.
Baseline characteristics of patients with nsvt alerts by reprogramming status.
| Variable: | Total (n = 158) | Programmed (n = 126) | No programming (n = 32) | p value |
|---|---|---|---|---|
| Age, years | 66 ± 12 | 65.5 ± 11.7 | 67.7 ± 13.7 | 0.374 |
| Sex, males | 99 (62.7) | 78 (61.9) | 21 (65.6) | 0.746 |
| Race, Whites | 115 (72.8) | 92 (73) | 23 (71.8) | 0.721 |
| Prior stroke | 21 (13.3) | 15 (11.9) | 6 (18.7) | 0.930 |
| Deceased | 11(7) | 9 (7.1) | 2 (6.2) | 0.674 |
| Anticoagulation: | ||||
| Oral anticoagulants | 37 (23.4) | 34 (26.9) | 3 (9.3) | 0.571 |
| Aspirin | 84 (46.8) | 60 (47.6) | 24 (75) | 0.412 |
| Both OAC & aspirin | 23 (14.6) | 23 (18.2) | 0 | — |
| None | 14 (16) | 9 (7.1) | 5 (15.6) | 0.340 |
| SCD | 6 (3.8) | 3 (2.3) | 3 (9.3) | 0.061 |
Note: Categorical data presented as numbers (percentage %), continuous data presented as Mean ± SD for normally distributed variables or Median ± IQR for skewed data.
Abbreviations: NSVT, non‐sustained ventricular tachycardia; OAC, oral anticoagulants; SCD, sudden cardiac death.
Table 6.
Impact of alerts reprogramming on clinical outcomes in AF and NSVT patients.
| Reference level | Unadjusted model OR (95% CI) | p value | Model 1 OR (95% CI) | p value | Model 2 OR (95% CI) | p value | |
|---|---|---|---|---|---|---|---|
| Ischemic strokea | No programming | 0.81 (0.28–2.34) | 0.695 | 0.78 (0.27–2.26) | 0.641 | 0.82 (0.27–2.51) | 0.730 |
| SCDb | No programming | 0.24 (0.05–1.23) | 0.088 | 0.28 (0.05–1.53) | 0.144 | — | — |
Note: Odds ratios (OR) and 95% confidence intervals (CI) are presented for reprogramming compared to no reprogramming as the reference group. Model 1: Adjusted for age, sex, and race. Model 2: Adjusted for Model 1, plus CHA2DS2VASC score, anticoagulation status, and prior stroke history.
Ischemic stroke defined as clinically adjugated diagnosis of ischemic stroke by a neurologist based on clinical symptoms and confirmatory imaging either by CTA or MRI.
SCD: Sudden cardiac death, diagnosis based on cause of death in death certificate if available or in‐hospital documentation of VT/VF before death or device transmission Logistic regression models were used as follows.
4. Discussion
The lack of standardized alert driven parameters for CIED remote monitoring creates a challenge in managing high volume data and the resultant clinician alert fatigue. In this study we demonstrated that clinically guided reprogramming of CIED alert parameters significantly reduces NAA without compromising patient safety or increasing adverse clinical outcomes. Following reprogramming, the mean number of NAAs decreased by 74.2%; thus is a substantial improvement in alert burden related to subclinical AF as well as NSVT on ICD platforms. Importantly, the reduction in NAAs was achieved without compromising patient safety. These findings have important implications for improving the efficiency of device clinics, mitigating clinician alert fatigue, and optimizing patient care in remote CIED monitoring.
The main endpoint of effective remote monitoring follow‐up is to identify actionable clinical events and patterns while ensuring timely responses. However, managing the large volume of alert transmissions poses a significant challenge for efficient triage [8]. The concept of personalizing remote transmission data to reduce the noise of NAA through alert‐driven protocols has been explored through multiple approaches, including machine learning and artificial intelligence [9, 10, 11]. For instance, Rosier et al. utilized an AI algorithm that incorporated medical records and remote data to stratify AF patients, resulting in an 86% reduction in notification workload [12].
Despite these advancements, the alert‐driven system continues to face challenges, particularly the lack of standardization across vendors [13]. Different platforms generate varying types and quantities of alerts, even with uniform clinician‐applied settings. In our sample, 66.5% of all CIEDs and platforms achieved a complete (100%) reduction in the number of NAA after implementing the intended alerts protocol. For alerts related to NSVT, 93.9% of patients experienced full (100%) censoring of these alerts. In contrast, complete censoring of the specified AF alerts was achieved in 50.1% of patients. For the remaining devices, although less than a 100% reduction in NAA was observed, our protocol was still effective in significantly reducing the total number of these alerts. These findings highlight the effectiveness of the alert protocol in safely reducing alert burden, while identifying areas for further refinement in specific configurations.
To address these limitations, further research should explore the broader applicability of alert‐censoring strategies in larger populations. Additionally, integrating multiple approaches such as guideline‐based alert parameters, machine learning‐based filtering, and medical record integration may further optimize device clinic efficiency while enhancing patient care. Notably, our study is the first to assess the clinical ramifications (i.e. adverse outcomes) when censoring these alerts, marking an important step toward safely optimizing alert management systems.
In our sample, the incidence of stroke in patients with AF was comparable between the reprogrammed and control groups, at 5.2% and 5.4%, respectively. Notably, the observed stroke incidence was lower than the expected risk based on the CHA₂DS₂VASC score scores in our sample [14]. The approach to censoring alerts for subclinical AF was guided by prior literature and clinical trials examining the association between AF episode duration and stroke risk.
For example, the TRENDS trial demonstrated that patients with AF episodes lasting less than 5.5 h had a stroke risk similar to those without any detected AF or atrial tachycardia [15]. Similarly, while the ASSERT study found that AF episodes lasting at least 6 min were associated with a 2.5‐fold increased risk of ischemic stroke over 2.5 years, a sub‐analysis of ASSERT revealed a statistically significant increase in stroke risk only for episodes exceeding 24 h [16, 17]. Furthermore, evidence from the ARTESIA and NOAH‐AFNET trials further supports the low stroke risk associated with brief or infrequent subclinical AF episodes [18, 19]. The ARTESIA trial found only modest stroke risk reduction with anticoagulation for subclinical AF, which was offset by an increased bleeding risk [18]. A subsequent ARTESIA sub‐study demonstrated that for patients with subclinical AF lasting between 6 min and 24 h, the duration of the longest episode did not significantly impact stroke risk or the treatment benefit of anticoagulants [20]. Similarly, the NOAH‐AFNET trial highlighted the limited benefit of anticoagulating patients with atrial high‐rate episodes, given their low baseline stroke risk [19]. These findings create the rationale for censoring alerts for AF episodes lasting less than 24 h, especially in patients already receiving anticoagulation or those with AF durations under 5.5 h.
Approximately 95% of our sample was adequately anticoagulated, and all but two incident stroke cases occurred in anticoagulated individuals; both of those patients had discontinued oral anticoagulation due to bleeding concerns. These results suggest that censoring subclinical AF alerts may be a safe approach when patients' stroke risk is otherwise clinically optimized either medically or surgically. Moreover, our protocol include alerts were triggered solely by the duration and persistence of AF and did not account for ventricular rate control during AF episodes. Future studies may explore the utility of incorporating rate‐based alert parameters to further refine clinical stratification.
A substantial body of evidence has established an association between NSVT and adverse cardiovascular outcomes, such as stroke, coronary heart disease, SCD, and overall mortality [21, 22]. In our sample we chose to censor alerts in patients with ICD systems for NSVT [defined as ventricular high‐rate episodes not requiring therapy (ATP or shock)]; detection criteria for VHR were as per the patient's chronic programming. In this context, the occurrence of NSVT episodes does not typically require acute management. This approach thus serves to reduce the burden of unnecessary alerts, while not compromising patient safety.
The incidence of SCD was not significantly different and was, in fact, lower in the reprogrammed group (2.3%, with NSVT alerts censored) compared to the control group (9.3%). These results, combined with the absence of specific interventions for these events, supports the notion that censoring these alerts are unlikely to influence SCD risk in patients with ICDs [23, 24, 25]. Further investigation into refining NSVT alert settings on pacemaker platforms, perhaps by integrating additional clinical parameters such as LVEF and cardiac biomarkers, is needed.
4.1. Study Limitations
While this study's strengths include an extended follow‐up period and clinically adjudicated outcomes, our findings should be interpreted in light of several limitations. First, this was a single‐center study, which may limit the generalizability of the results to other populations. Second, the control group was selected from patients who were not reprogrammed during the time window of interest, it is possible that the reasons for not being reprogrammed (lack of in‐person device clinic appointment attendance) may bias and confound the results. Additionally, the small number of cases weakens the strength of the observed associations and limits drawing strong conclusion from the regression models results. By the end of follow‐up, nearly 10% of patients in the sample were deceased and some had their devices deactivated for monitoring and shock delivery as part of end‐of‐life protocols. Furthermore, the lack of death certificates for all patients, along with reliance on scheduled remote transmissions every 3 months, could also lead to an underestimation of SCD cases. Additionally, our study focused exclusively on safety‐related clinical outcomes (ischemic stroke in patients with AF alerts and SCD in those with NSVT alerts) without evaluating other potentially relevant endpoints such as the frequency of ICD shocks or anti‐tachycardia pacing. These device‐based therapies may also be influenced by alert reprogramming which represents a relevant clinical question and a potential area for future investigation.
5. Conclusions
This study demonstrates that clinically guided reprogramming of CIED alert parameters significantly reduces the burden of NAA while maintaining patient safety. By optimizing alert settings for subclinical AF and NSVT, we achieved a substantial reduction in alert volume without increasing adverse patient outcomes. These findings underscore the potential of tailored alert management strategies to improve the efficiency of remote monitoring, reduce clinician fatigue, and enhance patient care. Future research should focus on validating these findings in larger, multicenter studies and explore the integration of different approaches to optimize cardiac remote monitoring.
Conflicts of Interest
Several co‐authors are affiliated with CV Remote Solutions, which supported the remote processing of CIED data in this study. No financial compensation was received for participation in this study, and no commercial influence was exerted on study design, data interpretation, or manuscript preparation.
Acknowledgments
The authors have nothing to report.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Slotwiner D., Varma N., Akar J. G., et al., “HRS Expert Consensus Statement on Remote Interrogation and Monitoring for Cardiovascular Implantable Electronic Devices,” Heart Rhythm : The Official Journal of the Heart Rhythm Society 12 (2015): e69–e100. [DOI] [PubMed] [Google Scholar]
- 2. Varma N., Epstein A. E., Irimpen A., Schweikert R., and Love C., “Efficacy and Safety of Automatic Remote Monitoring for Implantable Cardioverter‐Defibrillator Follow‐up: The Lumos‐T Safely Reduces Routine Office Device Follow‐up (TRUST) Trial,” Circulation 122 (2010): 325–332. [DOI] [PubMed] [Google Scholar]
- 3. Crossley G. H., Boyle A., Vitense H., Chang Y., and Mead R. H., “The CONNECT (Clinical Evaluation of Remote Notification to Reduce Time to Clinical Decision) Trial,” Journal of the American College of Cardiology 57 (2011): 1181–1189. [DOI] [PubMed] [Google Scholar]
- 4. Hindricks G., Taborsky M., Glikson M., et al., “Implant‐Based Multiparameter Telemonitoring of Patients With Heart Failure (IN‐TIME): A Randomised Controlled Trial,” Lancet 384 (2014): 583–590. [DOI] [PubMed] [Google Scholar]
- 5. Ferrick A. M., Raj S. R., Deneke T., et al., “2023 HRS/EHRA/APHRS/LAHRS Expert Consensus Statement on Practical Management of the Remote Device Clinic,” Europace : European Pacing, Arrhythmias, and Cardiac Electrophysiology : Journal of the Working Groups on Cardiac Pacing, Arrhythmias, and Cardiac Cellular Electrophysiology of the European Society of Cardiology 25 (2023): euad123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Diamond J., Varma N., and Kramer D. B., “Making the Most of Cardiac Device Remote Management: Towards an Actionable Care Model,” Circulation: Arrhythmia and Electrophysiology 14 (2021): e009497. [DOI] [PubMed] [Google Scholar]
- 7. Bawa D., Kabra R., Ahmed A., et al., “Data Deluge From Remote Monitoring of Cardiac Implantable Electronic Devices and Importance of Clinical Stratification,” Heart Rhythm O2 4 (2023): 374–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Varma N., Piccini J. P., Snell J., Fischer A., Dalal N., and Mittal S., “The Relationship Between Level of Adherence to Automatic Wireless Remote Monitoring and Survival in Pacemaker and Defibrillator Patients,” Journal of the American College of Cardiology 65 (2015): 2601–2610. [DOI] [PubMed] [Google Scholar]
- 9. Lazarus A., Gentils M., Klaes S., et al., “Filtering of Remote Monitoring Alerts Transmitted by Cardiac Implantable Electronic Devices and Reclassification of Atrial Fibrillation Events by a New Algorithm,” Cardiovascular Digital Health Journal 4 (2023): 149–154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Varma N., Love C. J., Michalski J., and Epstein A. E., “Alert‐Based ICD Follow‐up,” JACC: Clinical Electrophysiology 7 (2021): 976–987. [DOI] [PubMed] [Google Scholar]
- 11. Rosier A., Mabo P., Temal L., et al., “Remote Monitoring of Cardiac Implantable Devices: Ontology Driven Classification of the Alerts,” Studies in Health Technology and Informatics 221 (2016): 59–63. [PubMed] [Google Scholar]
- 12. Rosier A., Mabo P., Temal L., et al., “Personalized and Automated Remote Monitoring of Atrial Fibrillation,” Europace: European Pacing, Arrhythmias, and Cardiac Electrophysiology : Journal of the Working Groups on Cardiac Pacing, Arrhythmias, and Cardiac Cellular Electrophysiology of the European Society of Cardiology 18 (2016): 347–352. [DOI] [PubMed] [Google Scholar]
- 13. Vandenberk B. and Raj S. R., “Remote Patient Monitoring: What Have We Learned and Where Are We Going?,” Current Cardiovascular Risk Reports 17 (2023): 103–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Lip G. Y. H., Nieuwlaat R., Pisters R., Lane D. A., and Crijns H. J. G. M., “Refining Clinical Risk Stratification for Predicting Stroke and Thromboembolism in Atrial Fibrillation Using a Novel Risk Factor‐Based Approach,” Chest 137 (2010): 263–272. [DOI] [PubMed] [Google Scholar]
- 15. Glotzer T. V., Daoud E. G., Wyse D. G., et al., “The Relationship Between Daily Atrial Tachyarrhythmia Burden From Implantable Device Diagnostics and Stroke Risk: The TRENDS Study,” Circulation: Arrhythmia and Electrophysiology 2 (2009): 474–480. [DOI] [PubMed] [Google Scholar]
- 16. Healey J. S., Connolly S. J., Gold M. R., et al., “Subclinical Atrial Fibrillation and the Risk of Stroke,” New England Journal of Medicine 366 (2012): 120–129, 10.1056/NEJMx160004. [DOI] [PubMed] [Google Scholar]
- 17. Van Gelder I. C., Healey J. S., Crijns H. J. G. M., et al., “Duration of Device‐Detected Subclinical Atrial Fibrillation and Occurrence of Stroke in ASSERT,” European Heart Journal 38 (2017): 1339–1344. [DOI] [PubMed] [Google Scholar]
- 18. Healey J. S., Lopes R. D., Granger C. B., et al., “Apixaban for Stroke Prevention in Subclinical Atrial Fibrillation,” New England Journal of Medicine 390 (2024): 107–117. [DOI] [PubMed] [Google Scholar]
- 19. Kirchhof P., Toennis T., Goette A., et al., “Anticoagulation With Edoxaban in Patients With Atrial High‐Rate Episodes,” New England Journal of Medicine 389 (2023): 1167–1179. [DOI] [PubMed] [Google Scholar]
- 20. McIntyre W. F., Benz A. P., Healey J. S., et al. Risk of Stroke or Systemic Embolism According to Baseline Frequency and Duration of Subclinical Atrial Fibrillation: Insights From the ARTESiA Trial [published correction appears in Circulation. 2025;151:e29. [DOI] [PubMed]
- 21. Al‐Khatib S. M., Stevenson W. G., Ackerman M. J., et al. 2017 AHA/ACC/HRS Guideline for Management of Patients With Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society [published correction appears in Circulation. 2018;138:e419‐e420. [DOI] [PubMed]
- 22. Lin C. Y., Chang S. L., Chung F. P., et al., “Long‐Term Outcome of non‐Sustained Ventricular Tachycardia in Structurally Normal Hearts,” PLoS One 11 (2016): e0160181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Katritsis D. G., Siontis G. C. M., and Camm A. J., “Prognostic Significance of Ambulatory ECG Monitoring for Ventricular Arrhythmias,” Progress in Cardiovascular Diseases 56 (2013): 133–142. [DOI] [PubMed] [Google Scholar]
- 24. Katritsis D. G., Zareba W., and Camm A. J., “Nonsustained Ventricular Tachycardia,” Journal of the American College of Cardiology 60 (2012): 1993–2004. [DOI] [PubMed] [Google Scholar]
- 25. Patton K. K., “The Riddle of Nonsustained Ventricular Tachycardia and Sudden Cardiac Death: Are We Approaching a Solution?,” Circulation 122 (2010): 449–451. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.