Atrioventricular Synchrony Delivered by a Dual-Chamber Leadless Pacemaker System

Abstract

BACKGROUND:

A dual-chamber leadless pacemaker system has been designed for atrioventricular synchronous pacing using wireless, beat-to-beat, implant-to-implant (i2i) communication between distinct atrial and ventricular leadless pacemakers. The atrioventricular synchrony achieved across various ambulatory scenarios has yet to be systematically evaluated.

METHODS:

A prospective, single-arm, unblinded, multicenter, international clinical trial of the leadless pacemaker system was conducted in patients with a conventional dual-chamber pacing indication enrolled from February 2022 to March 2023. Leadless pacemaker systems were implanted, and 12-lead Holter electrocardiographic recordings were collected 3 months after implantation over various postures/activities: sitting, supine, left lateral recumbent, right lateral recumbent, standing, normal walk, and fast walk. An independent Holter core laboratory performed a manual adjudication of the percent of atrioventricular synchronous beats using the standard 300-millisecond PR interval limit. Atrium-to-ventricle and ventricle-to-atrium i2i communication success rates were also assessed. Post hoc summary statistics describing the relationships between atrioventricular synchrony and i2i success, posture/activity, implantation indication, atrioventricular event, and heart rate were calculated.

RESULTS:

In the evaluable population (n=384 of 464 enrolled [83%]; 61% male; age, 70 years; weight, 82 kg; 60% ejection fraction; 95% of beats evaluable), the mean atrioventricular synchrony of 98% of beats observed across all postures using the standard 300-millisecond limit was greater than both atrial-to-ventricular i2i (94%) and ventricular-to-atrial i2i (94%; P<0.001), exceeding both i2i values in 95% of patients. Atrioventricular synchrony was achieved in >95% of evaluable beats across all postures/activities, implantation indications, atrioventricular paced/sensed event combinations, and heart rate ranges (including >100 bpm).

CONCLUSIONS:

This dual-chamber leadless pacemaker system demonstrated atrioventricular synchrony in 98% of evaluable beats at 3 months after implantation. Atrioventricular synchrony was maintained across postures/activities and remained robust for heart rates >100 bpm.

Clinical Perspective.

What Is New?

• The first dual-chamber leadless pacemaker system that uses wireless, beat-to-beat, implant-to-implant communication between distinct atrial and ventricular devices provides reliable atrioventricular synchrony.

• This in-depth evaluation of subjects enrolled in a prospective, international, clinical trial demonstrated consistent atrioventricular synchrony at 3 months across multiple body postures, levels of activity, and sensing/pacing scenarios, including at elevated heart rates >100 bpm, according to continuous external electrocardiographic monitoring and using strict, programming-based criteria.

• Device-generated implant-to-implant diagnostics can be used as a lower bound surrogate to an otherwise cumbersome manual atrioventricular synchrony measurement.

What Are the Clinical Implications?

• Wireless implant-to-implant communication used in the new dual-chamber leadless pacemaker system provides patients with atrial pacing and atrioventricular synchrony similar to traditional transvenous pacemaker systems.

• The high proportion of atrioventricular synchrony (>95%) observed across various pacing indications, body positions/activities, and heart rates expands the potential application and benefits of leadless pacemaker technology to patients who need cardiac pacing.

Despite decades of clinical experience and advanced implantation tools, traditional transvenous pacemakers are still associated with complication rates of ≈10% to 15% in the first 6 months alone in contemporary studies, attributed predominantly to transvenous leads and the subcutaneous pulse generator pocket.^1,2 Leadless pacemakers (LPs) are contained entirely within the target chamber and were designed to simplify the implantation procedure and mitigate complications.^3–6

Single-chamber leadless pacing in the right ventricle (RV) is now an established form of therapy.⁷ However, ≈80% of patients requiring pacemakers, particularly those experiencing sinus node dysfunction or heart block, require reliable sensing with synchronous atrial and ventricular pacing. Recent advances in LP size, battery technology, and fixation mechanisms have opened the door for a right atrial (RA) LP combined with an RV LP to achieve dual-chamber leadless atrioventricular synchronous pacing.

Expanding leadless pacing to multiple chambers necessitates continual, bidirectional communication across multiple devices to maintain true atrioventricular synchrony. A novel dual-chamber LP system, Aveir DR (Abbott, Abbott Park, IL), has been developed with safe and secure implantation, stable electrical performance, and beat-to-beat wireless communication to maintain atrioventricular synchrony, all demonstrated in a preclinical setting.^8–10 A prospective, multicenter, international, pivotal clinical trial of this leadless pacing system has since been conducted, demonstrating clinical safety and robust performance in the first 300 enrolled patients.¹¹ Atrioventricular synchrony (PR interval ≤300 milliseconds) in ≥70% of beats during the sitting posture at 3 months was achieved in 97.3% of patients, meeting a prespecified primary performance end point of that trial (83% of patients). However, the study continued to enroll patients beyond that initial analysis.

Here, we provide an updated, comprehensive post hoc analysis of the complete study cohort, with an in-depth evaluation of the acute atrioventricular synchrony at 3 months after implantation across a broad spectrum of patient implantation indications, postures, activities, heart rates, and atrioventricular pacing scenarios among all enrolled subjects with de novo implants and evaluable atrioventricular synchrony data. In addition, the relationship between atrioventricular synchrony and the implant-to-implant (i2i) communication needed to maintain it was evaluated.

The dual-chamber LP consists of an atrial LP (ALP) and a ventricular LP (VLP; Figure 1A) that communicate bidirectionally and on a beat-to-beat basis to continually deliver atrioventricular synchronous pacing with a novel i2i communication modality.⁸ Transmissions are sent from ALP to VLP (A-to-V) and from VLP to ALP (V-to-A) immediately before each paced event and after each sensed event to avoid interference with noncommunication (ie, pacing) operations. Beat-to-beat transmissions allow the receiving LP to initiate the appropriate blanking periods, timers, and atrioventricular delays in response to a paced/sensed event. Examples of i2i communication are shown in Figure 1B. It is important to note that the i2i signal characteristics can be tuned to enhance transmission success using the i2i setting level, which incorporates the transmitting signal pulse amplitude, transmitting signal pulse duration, and receiving sensing threshold (range, 1–7; nominal, 4).

Figure 1.

Schematics of dual-chamber devices and implant-to-implant communication. A, The leadless pacemaker system includes both atrial (ALP) and ventricular leadless pacemakers (VLP). B, Example of implant-to-implant, or i2i, communication messages for As-Vp and Ap-Vs event sequences. Ap indicates atrial pacing; As, atrial sensing; RA, right atrial; RV, right ventricular; Vp, ventricular pacing; and Vs, ventricular sensing.

During successful bidirectional i2i communication, the system can function in DDD(R) mode. Specifically, each RV sensing/pacing event triggers a V-to-A message to initiate the VA interval for RA pacing, as necessary. The subsequent RA sensing/pacing event triggers an A-to-V message to initiate the programmed atrioventricular delay for RV pacing, as necessary. This back-and-forth communication adjusts the relative timing of RA and RV pacing to account for any sensed intrinsic activity or programming change. Persistent i2i communication interruption, however, could disrupt this harmony and result in delayed ventricular pacing or withheld atrial pacing.

The system mitigates such asynchrony by monitoring the receipt of each transmission and automatically transitioning to 1 of 3 “safeguard modes.” If A-to-V i2i communication fails, the system effectively functions in DDI mode. In contrast, if V-to-A i2i communication fails, the system effectively functions in VDD mode. If i2i communication fails in both directions simultaneously, the system effectively functions in VDI mode.⁸ Collectively, these 3 safeguard modes guarantee ventricular pacing at the programmed rate during i2i interruption while providing atrial tracking and pacing whenever possible, despite changes in intrinsic rhythm or atrioventricular conduction.

The atrioventricular synchrony that ultimately results from i2i, however, can be definitively assessed only by manually verifying atrial and ventricular activation timing through external electrocardiographic adjudication. Although such a manual assessment may not readily be feasible in a clinical implantation setting, the A-to-V and V-to-A i2i success rates (ie, i2i throughput, percent of transmissions sent that were successfully received) are device-based diagnostics that can be interrogated by the programmer. These i2i success rates can serve as feedback metrics, upstream from atrioventricular synchrony, to guide implantation site selection and provide LP communication monitoring.⁹

METHODS

Clinical Study Design

An evaluation of atrioventricular synchrony at 3 months after implantation was performed as part of a prospective, single-arm, multicenter, international clinical trial of the dual-chamber leadless pacing system (NCT05252702).¹¹ The study was performed according to the principles outlined in the Declaration of Helsinki and the Good Clinical Practice guidelines of the European Commission. This report follows the Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines relevant to post hoc analyses. All patients provided written informed consent, and the study protocol was approved by each institutional ethics committee.

Study Population

As previously detailed, inclusion criteria were the standard indications for dual-chamber pacing; exclusion criteria were mechanical tricuspid valve prosthesis, inferior vena cava filter, preexisting pacing or defibrillation leads, or electrically active implantable medical devices.¹¹ Atrioventricular synchrony was evaluated for de novo patients with dual implantation (ie, no existing implanted pacemaker hardware, leadless or otherwise), as detailed later.

Device Implementation

After enrollment, the dual-chamber LP was implanted per standard instructions, with implantation sites and device programming left to the discretion of the implanting physician. At the 3-month postimplantation visit (90±30 days), device programming was modified for atrioventricular synchrony evaluation to promote atrial or ventricular pacing at each cycle as follows: Pacing mode was set to DDD or DDDR; paced/sensed atrioventricular delays were set to ≤250 milliseconds; the Ventricular Intrinsic Preference (Abbott) feature was deactivated to avoid promoting intrinsic atrioventricular conduction and extension of atrioventricular delays beyond 250 milliseconds; the hysteresis rate was deactivated to avoid promoting intrinsic atrial activity; and the base rate was set to ≥10 bpm above the intrinsic atrial rate (or per physician discretion) if atrial pacing was not present.

For each subject, a US Food and Drug Administration/European Union–certified 12-lead Holter monitor (Mortara H12+, Mortara Instrument, Milwaukee, WI) was used to record continuous ECGs at the 3-month visit for each of the following prescribed postures and limited activities in order: sitting, supine, left lateral recumbent, right lateral recumbent, standing, normal walk (ie, comfortable pace), and fast walk (ie, faster, brisk pace). Holter ECGs for each posture were recorded for 2 minutes, with the exception of 5 minutes for the sitting position. LP diagnostics were cleared immediately before each posture period with the programmer (Merlin Patient Care System, Abbott), and LPs were interrogated after each posture period to determine interim i2i success rates.

Atrioventricular Synchrony Evaluation

An independent core laboratory (IQVIA Biotech, Morrisville, NC) performed an evaluation of the atrioventricular synchrony associated with each posture at 3 months after implantation. Although only the atrioventricular synchrony evaluation associated with the sitting posture was prespecified and included in the primary study end points, post hoc evaluation of all other postures/activities followed the same procedure. In brief, atrial and ventricular paced/sensed activations were automatically adjudicated with custom software, with adjudications manually confirmed by multiple cardiologists. Events were classified as atrial sensing, atrial pacing, ventricular sensing, ventricular pacing, and atrioventricular combinations thereof (eg, atrial sensing-ventricular pacing) based on waveform morphology and observed pacing spikes. The PR interval was calculated as the duration between isoelectric departure for each atrial and subsequent ventricular event. For the primary analysis, beats with a PR interval ≤300 milliseconds were classified as atrioventricular synchronous. The 300-millisecond threshold was chosen from previous clinical studies evaluating atrioventricular synchrony in alternative LPs^12,13 to allow for potential intrinsic atrioventricular conduction within the range of programmed atrioventricular delays and beyond which patients may present with clinical symptoms.¹⁴

A secondary analysis was performed with a prespecified stricter definition of atrioventricular synchrony based on the patient-specific atrioventricular delays programmed for each device. In this strict definition, beats were classified as atrioventricular synchronous if the PR interval was >100 milliseconds and less than both the programmed paced atrioventricular delay+20 milliseconds and the sensed atrioventricular delay+50 milliseconds. The 100-milliseconds-lower PR limit reflects the lower physiological boundary in this dual-chamber pacemaker population. A 20-millisecond margin was added to both the paced and sensed atrioventricular delays to account for potential discrepancies in device electrogram-based compared with manual ECG-based identification of atrioventricular events and any latency from pacing. An additional 30-millisecond margin was allotted for the sensed atrioventricular delay (ie, 50 milliseconds total) to allow intrinsic activation of the entire atrium, including the specific ALP site. This strict atrioventricular synchrony definition presents a more accurate, programming-specific set of boundaries for assessing expected device functionality and is more aligned with corresponding transvenous DDD pacemaker evaluations than a broad, programming-agnostic upper limit of 300 milliseconds and avoids attributing atrioventricular synchrony to coincidental ventricular events occurring after atrial events.

For each posture, the atrioventricular synchrony percentage was calculated as the number of atrioventricular synchronous beats as a percentage of all evaluable beats (Figure S2). Evaluable beats were defined as having an identifiable ventricular paced/sensed event preceded by either an identifiable atrial paced/sensed event or the confirmed absence of one (ie, flatline). In addition, evaluable beats excluded atrial/supraventricular/ventricular tachycardias, premature atrial/ventricular contractions, rates exceeding the maximum tracking rate, and atrioventricular events not classifiable because of low electrocardiographic signal amplitude, noise, or fusion. The mean ventricular heart rate was calculated as the mean across all evaluable beats for each posture.

Patients with a substantial deviation from the aforementioned prescribed device programming at the 3-month evaluation visit were excluded from the analysis. To maintain consistent populations for comparison of the standard and strict atrioventricular synchrony definitions, patients with the following atrioventricular delay programming were also excluded: (1) paced or sensed atrioventricular delay <120 milliseconds, to account for the lower 100-millisecond bound of the strict atrioventricular synchrony definition+20-millisecond margin described above; (2) paced atrioventricular delay+20 milliseconds >300 milliseconds (ie, paced atrioventricular >280); or (3) sensed atrioventricular delay+50 milliseconds>300 milliseconds (ie, sensed atrioventricular >250 milliseconds). However, the patients with longer atrioventricular delays were already excluded according to the study protocol. Independently of patient exclusions, individual patient posture recordings were also excluded if the majority (>50%) of beats in the recording sample were not evaluable. Note that some patients could not perform every posture and activity.

In addition to the acute posture/activity recordings collected in clinic at 3 months, out-of-clinic durations of all distinct instances of i2i communication interruption from 1 to 3 months were collected from device interrogations at 3 months. Although not directly associated with any specific posture/activity, the 2-month time window allowed potential communication interruption instances longer than the 2-minute in-clinic recordings to be captured.

Statistical Analysis

Statistical analyses were performed with MATLAB (The MathWorks, Natick, MA). Percentages are reported as mean (95% CI), with confidence CIs calculated from a binomial distribution and capped at 100%. Continuous variables are reported as mean±SD. Correlations between atrioventricular synchrony and continuous variables were quantified by the Spearman rank correlation coefficient.

Comparisons of proportions (eg, contingency tables of patients with atrioventricular synchrony ≥90%) were evaluated with χ² tests, followed by post hoc χ² tests of group pairs, when warranted, with Holm-Bonferroni correction for multiple comparisons. Comparisons of atrioventricular synchrony across multiple categorical groups (eg, at each posture) were evaluated with Kruskal-Wallis tests, followed by post hoc Dunn tests with Sidak correction for multiple comparisons. Paired comparisons of atrioventricular synchrony between 2 groups (eg, standard versus strict atrioventricular synchrony criteria at sitting posture) were evaluated with Wilcoxon signed-rank tests.

In addition to these univariable analyses, the combined impact of all relevant variables on atrioventricular synchrony was explored by an ANOVA of a linear mixed-effects model. The model was fit with 3-month atrioventricular synchrony as the outcome and the following linear predictors with no associations: i2i success (A-to-V and V-to-A), posture/activity, pacing indication, mean heart rate, RA implantation site, RV implantation site, and atrioventricular event (percent Ap and percent Vp). In all tests, differences were considered statistically significant at P<0.05.

RESULTS

Study Population

From February 2022 to March 2023, 464 de novo patients were enrolled, and 452 of 464 (97.4%) underwent an implantation attempt across 77 centers in the United States, Canada, Europe, and Asia-Pacific region, with both LPs successfully implanted in 446 of 452 patients (98.7%). Of successfully implanted patients, 424 of 446 (95.1%) completed the 3-month atrioventricular synchrony assessment visit.

Of patients with atrioventricular synchrony evaluated, 415 of 424 (97.9%) completed at least 1 posture/activity, and 398 of 415 (95.9%) had proper device programming. Of the 2786 possible posture/activity recordings (ie, 398 remaining patients×7 posture/activities), 101 recordings (3.6%) were not obtained because some patients did not perform every posture/activity; however, all 398 patients completed at least 1 posture recording. Of the remaining 2685 posture/activity recordings, 277 recordings (10.3%) were excluded because the majority of beats in the recording were not evaluable, resulting in an exclusion of 14 more patients (3.5%).

In total, atrioventricular synchrony was evaluated in 2408 posture or activity recordings from 384 patients (6.1±1.6 posture or activity recordings per patient of 7 possible). In these recordings, 94.9% (95% CI, 92.7–97.1) of beats were classified as evaluable (mean [95% CI] across patients). The percent of evaluable beats across patients did not differ according to sex (P=0.92), history of atrial fibrillation (P=0.20), or indication for pacing (P=0.71).

Patient disposition is detailed in Figure S1. Characteristics for the patients included in the evaluation (age, 69.5±13.5 years; 61.2% male) are provided in the Table.

Table.

Patient Characteristics, Device Programming, and Acute Measurements

Atrioventricular Synchrony Compared With i2i Success

To evaluate the relationship between i2i communication success and the resulting atrioventricular synchrony, each metric was calculated as the mean across all completed postures for each patient. Over the analyzed patient population (N=384), the resulting mean success rates for atrioventricular synchrony, A-to-V i2i, and V-to-A i2i were 98.1% (95% CI, 96.7–99.4) of beats, 93.6% (95% CI, 91.2–96.1) of transmissions, and 94.1% (95% CI, 91.7–96.4) of transmissions, respectively, as shown in Figure 2A. These i2i rates were achieved with mean programmed A-to-V and V-to-A i2i setting levels of 4.8±1.4 and 5.2±1.3 of 7, respectively. V-to-A i2i was statistically greater than A-to-V i2i (P<0.005), and it is important to note that atrioventricular synchrony was greater than both V-to-A i2i and A-to-V i2i (P<0.001 for both).

Figure 2.

Relationship between atrioventricular (AV) synchrony and i2i transmission success. A, atrioventricular synchrony, Atrial-to-ventricular (A-to-V) implant-to-implant (i2i) success, and ventricular-to-atrial (V-to-A) i2i success, calculated from the mean percent of beats across all postures for each patient. B, Color-coded atrioventricular synchrony vs A-to-V i2i and V-to-A i2i. Bars show mean (95% CI). Brackets indicate statistically significant differences. Sample sizes indicate number of patients.

Atrioventricular synchrony was also evaluated with stricter, programming-specific criteria rather than a more liberal 300-millisecond cutoff (Supplemental Material). Across all posture recordings for all patients, atrioventricular synchrony was achieved in 96.3% (95% CI, 94.4–98.2) of beats using the programming-specific criteria (1.7% lower than 300-millisecond criterion, patient-wise mean difference; P<0.001).

The impact of i2i communication interruption on atrioventricular synchrony can be visualized in Figure 2B, which shows that most outlier patients with atrioventricular synchrony of 70% to 90% (orange) or <70% (red) were associated with A-to-V or V-to-A i2i success rates <80%. Note that only 2 of 384 patients (0.5%) exhibited mean atrioventricular synchrony <70% across all postures. The ability of the device-based i2i values to ultimately serve as the lower bound of the resulting atrioventricular synchrony was revealed using the difference between atrioventricular synchrony and the lower of the V-to-A and A-to-V i2i values. Atrioventricular synchrony was 7.1 (95% CI, 4.6–9.7) mean absolute percentage points greater than the lower of the 2 i2i success rates for each patient. Overall, atrioventricular synchrony exceeded both i2i values in 94.5% of patients (363/384) and was no lower than 5% below the lower i2i value in 99.0% of patients (380/384).

The durations of distinct instances of i2i interruption were quantified for the out-of-clinic period before the 3-month atrioventricular synchrony assessment. From 1 to 3 months after implantation across all patients, 86.6% (95% CI, 82.5–90.7) of all instances of A-to-V i2i interruption and 96.8% (95% CI, 95.0–98.7) of all instances of V-to-A i2i interruption were <6 seconds in duration; moreover, 98.9% (95% CI, 97.7–100.0) and 98.9% (95% CI, 97.8–100.0) of all instances of A-to-V and V-to-A interruption, respectively, were <30 seconds in duration.

Atrioventricular Synchrony by Posture/Activity

Atrioventricular synchrony was compared across each distinct posture/activity: sitting, supine, left lateral recumbent, right lateral recumbent, standing, normal walk, and fast walk. As shown in Figure 3A, atrioventricular synchrony using the standard 300-millisecond criteria (blue) was similar across all postures/activities in this population (overall P=0.08), with posture-wise means ranging from 96.9% (supine) to 98.7% (sitting, fast walk). As shown in Figure 3B, the proportion of patients with atrioventricular synchrony at least 90% using the standard 300-millisecond criteria (blue) was different across all postures/activities (overall P<0.001), ranging from 90.5% (supine) to 97.7% (normal walk). Specifically, the proportion of patients with atrioventricular synchrony at least 90% was higher for sitting, standing, and normal walk than supine (P<0.05 for each pair).

Figure 3.

Atrioventricular (AV) synchrony by patient posture/activity. A, Proportion of beats with atrioventricular synchrony, binned by posture/activity, and (B) proportion of patients with atrioventricular synchrony in ≥90% of beats. Atrioventricular synchrony metrics are shown for the standard atrioventricular synchrony criteria (300 milliseconds, blue) and strict atrioventricular synchrony criteria (by programmed atrioventricular delays [AVDs], orange). Bars show mean (95% CI). Brackets indicate statistically significant differences. Sample sizes indicate number of patients performing each posture/activity.

Atrioventricular synchrony was also evaluated with the use of stricter programming-specific criteria rather than the more liberal 300-millisecond cutoff. Also shown in Figure 3A, atrioventricular synchrony using these strict programming-specific criteria (orange) exceeded 94% for all postures/activities yet exhibited group-wise statistical differences (overall P<0.005). Also shown in Figure 3B, the proportion of patients with atrioventricular synchrony at least 90% using programming-specific criteria was different across all postures/activities (overall P<0.005), ranging from 85.7% (supine) to 94.1% (standing). Sitting and standing postures were again associated with higher proportions than supine (P<0.05 for each pair).

For standard compared with strict atrioventricular synchrony criteria, significant differences were observed in atrioventricular synchrony for each posture (P<0.05 for each pair; Figure 3A) and proportion of patients with atrioventricular synchrony exceeding 90% for supine, normal walk, and fast walk (P<0.05 for each pair; Figure 3B).

Atrioventricular Synchrony by Patient Indication

Atrioventricular synchrony values for each posture/activity were also compared for patient subpopulations with sinus node dysfunction (n=251) compared with atrioventricular block (n=118), shown in Figure 4. No aggregate differences across postures were observed within either the sinus node dysfunction population (P=0.29) or the atrioventricular block population (P=0.56). Across indications, however, atrioventricular synchrony was statistically greater in patients with sinus node dysfunction compared with patients with atrioventricular block at all postures except right lateral and standing (P<0.05 for each pair).

Figure 4.

Atrioventricular (AV) synchrony by patient posture/activity for patients with sinus node dysfunction vs atrioventricular block. Bars show mean (95% CI). Brackets indicate statistically significant differences across indications for each posture/activity. Sample sizes indicate number of patients performing each posture/activity.

Atrioventricular Synchrony by Heart Rate

A significant correlation between atrioventricular synchrony and the mean heart rate for each patient-posture recording was not observed in this population (Spearman ρ=0.00, P=0.89). The mean heart rate across all patients and postures was 77±13 bpm, with 5.8% of all recordings exceeding a 100-bpm mean heart rate. Specific heart rate ranges that could have a unique impact on atrioventricular synchrony were illustrated by binning each patient-posture recording according to the mean recorded heart rate, as shown in Figure 5. Mean atrioventricular synchrony values for all heart rate bins, including >100 bpm, exceeded 97%. The impact of heart rate on atrioventricular synchrony using the programming-specific criteria is detailed in the Supplemental Material.

Figure 5.

Atrioventricular (AV) synchrony by heart rate. Proportion of beats with atrioventricular synchrony for each recording across all postures, binned by mean heart rate. Bars show mean (95% CI). Sample sizes indicate number of posture recordings with mean heart rate of interest.

Atrioventricular Synchrony by Implantation Site

Atrioventricular synchrony values were also compared across RA and RV LP implantation sites, shown in Figure 6. Atrioventricular synchrony exceeded 96% for each site, with no differences observed across RA implantation sites (P=0.30; Figure 6A) or RV implantation sites (P=0.99; Figure 6B). An in-depth analysis of all RA/RV implantation site combinations (eg, each RA+RV site combination), although not shown for brevity, also revealed no atrioventricular synchrony difference (P=0.45). The impact of implantation site on atrioventricular synchrony using the programming-specific criteria is detailed in the Supplemental Material.

Figure 6.

Atrioventricular (AV) synchrony by implantation site. Atrioventricular synchrony is shown across all postures, grouped by (A) right atrial (RA) implantation site and (B) right ventricular (RV) implantation site. Bars show mean (95% CI). No statistical differences were observed across RA sites (P=0.30) or RV sites (P=0.99). Sample sizes indicate number of patients. RAA indicates right atrial appendage.

Atrioventricular Synchrony by Atrioventricular Event

The distribution of atrioventricular events in this population, as the mean percent of beats across all postures, is provided in the Table. In terms of predominant atrioventricular event for each of the 2408 posture recordings across all patients, 20.9% of postures were predominantly Ap-Vp, 43.2% were Ap-Vs, 23.6% were As-Vp, and 11.1% were As-Vs. As shown in Figure 7A, greater atrioventricular synchrony was associated with Ap events (Ap-Vp or Ap-Vs) compared with As events (As-Vp or As-Vs; P<0.05). In contrast, as shown in Figure 7B, lower atrioventricular synchrony was associated with Vp events (Ap-Vp or As-Vp) compared with Vs events (Ap-Vs or As-Vs) (P<0.001). These observations were supported by comparisons across pairs of atrioventricular event combinations shown in Figure 7C (overall P<0.001). Although statistical differences were observed, atrioventricular synchrony was delivered in at least 97% of beats for each atrioventricular event combination. The impact of atrioventricular event on atrioventricular synchrony using the programming-specific criteria is detailed in the Supplemental Material.

Figure 7.

Atrioventricular (AV) synchrony by atrioventricular event. Atrioventricular synchrony is shown across all postures, separated by (A) dominant atrial event, (B) dominant ventricular event, and (C) dominant atrioventricular event combination. Bars show mean (95% CI). Brackets indicate statistically significant differences. Sample sizes indicate number of posture recordings dominated by each event. Ap indicates atrial pacing; As, atrial sensing; Vp, ventricular pacing; and Vs, ventricular sensing.

Multivariable Analysis

The linear mixed-effects model of atrioventricular synchrony demonstrated consistent findings with the univariable analyses. Specifically, posture/activity, A-to-V i2i success, V-to-A i2i success, percent Ap, and percent Vp demonstrated a significant impact on atrioventricular synchrony (P<0.001), but pacing indication, heart rate, and implantation site did not. Detailed results of the linear mixed-effects model are provided in Table S1.

DISCUSSION

Many patients with a pacemaker require synchronous atrial and ventricular pacing/sensing, particularly those with sinus node dysfunction or heart block. This first dual-chamber LP system maintains atrioventricular synchrony through a novel beat-to-beat, i2i communication channel. The safety and performance of this system were evaluated in a prospective, multicenter, international clinical trial. The primary safety and performance end points for the first 300 patients were met and quantified as (1) 90.3% freedom from device- or procedure-related complications in the first 90 days, (2) 90.2% atrial capture threshold ≤3.0 V (at 0.4-millisecond pulse width) and atrial sensing amplitude ≥1.0 mV, and (3) 97.3% with atrioventricular synchrony in ≥70% of beats in the sitting posture.¹¹ This evaluation extends these findings to include an in-depth report of the acute atrioventricular synchrony measured at 3 months after implantation for all 384 de novo clinical trial subjects with proper programming and complete data across 2408 posture or activity recordings.

The atrial and ventricular LPs each contribute to maintaining atrioventricular synchrony by sending a transmission to one another (eg, A-to-V, V-to-A) at each local paced and sensed event. If 100% of such i2i transmissions are successful, atrioventricular synchrony would be observed in 100% of beats. During bouts of i2i communication interruption, novel safeguard algorithms provide dynamic shifts in pacing mode to minimize any resulting asynchrony. An important finding was that ≈99% of i2i interruption instances resolved within 30 seconds, when the system returns to DDD behavior before changes in atrial rate or PR interval are likely given an opportunity to disrupt atrioventricular synchrony.

As a consequence of the safeguards, the fleeting nature of most i2i interruptions, and potential for intrinsic ventricular activation, atrioventricular synchrony should always meet or exceed i2i success in either direction, absent any confounding sensing or capture issues. Across a broad spectrum of patient implantation indications, postures, activities, heart rates, implantation sites, and atrioventricular event combinations in this patient population, A-to-V and V-to-A i2i transmissions were successful in 93.6% and 94.1% of beats, respectively, yet atrioventricular synchrony was achieved in 98.1% of beats. Moreover, atrioventricular synchrony exceeded both i2i success rates in 94.5% of patients and was no lower than 5% below both i2i values in 99.0% of patients. Note that the apparent patient anomalies in which the mean atrioventricular synchrony did not exceed mean i2i success rates (eg, 3 orange points in the upper-right corner of Figure 2B) were associated with atrioventricular synchrony values just under the 90% cutoff (ie, 88.8%, 87.0%, and 83.3%).

Although atrioventricular synchrony can be definitively verified only by manual electrocardiographic adjudication, i2i success rates are device-based diagnostics that can be readily interrogated by the programmer and serve as a lower bound for atrioventricular synchrony estimation. Therefore, implantation site selection and subsequent device monitoring can be facilitated by using i2i success as a conservative surrogate for atrioventricular synchrony.

Two previous preclinical ovine studies of this system’s i2i communication reported long-term A-to-V and V-to-A i2i transmission success rates exceeding 99%⁸ and atrioventricular synchrony exceeding 99% across a variety of postures/activities, similar to this clinical study.⁹ The slightly higher i2i success rates observed in those 2 preclinical studies may be attributed to differences in ovine compared with human intra-LP distances and orientations, which are anatomically constrained and may affect i2i signal reception.

Successful i2i communication, which is verified intraprocedurally during implantation, could theoretically be hindered by the ambulatory postures, activities, and cardiovascular conduction scenarios associated with daily living. Consequently, atrioventricular synchrony at 3 months after implantation was assessed across a variety of prescribed postures/activities: sitting, supine, left lateral recumbent, right lateral recumbent, standing, normal walk, and fast walk. atrioventricular synchrony was maintained regardless of posture or activity, with ≥90% of all patients exhibiting atrioventricular synchrony in at least 90% of beats.

The independence of atrioventricular synchrony from posture/activity was documented for both sinus node dysfunction and atrioventricular block patient subpopulations alike. This is in contrast to reports of a leadless VDD pacemaker (Micra atrioventricular, Medtronic, Minneapolis, MN) that uses accelerometer-based tracking of atrial contractions to time single-chamber ventricular pacing. A comparable evaluation of 40 patients with atrioventricular block in the MARVEL 2 study (Micra Atrial Tracking Using a Ventricular Accelerometer 2) of the Micra leadless VDD pacemaker defined atrioventricular synchrony as the percent of paced or sensed ventricular events within 300 milliseconds after a surface ECG-confirmed P wave.¹² That study reported the atrioventricular synchrony delivered at rest (89.2%) to be significantly reduced at sitting and standing postures (75.4% and 69.8%) and to trend lower for normal and fast walk (72.7% and 74.7%), situations in which the hemodynamics associated with synchrony are more critical. This reduction was potentially attributed to orthostatic tachycardia or changes in venous hemodynamics that can dampen the accelerometer-based atrial contraction signal. In contrast, the dual-chamber system studied here provides atrial pacing and maintained atrioventricular synchrony in patients with atrioventricular block at the corresponding sitting and standing postures (98.0% and 98.1%), as well as during normal and fast walk activities (97.9% and 97.1%).

Inherent in the comprehensive array of postures/activities, ranging from stationary to a fast walk, was a broad spectrum of resulting mean heart rates, including ≥100 bpm. A correlation between atrioventricular synchrony and mean heart rate was not observed, and mean atrioventricular synchrony exceeded 97% for all rate ranges. It is notable that each recording was conservatively classified by the mean heart rate but likely included much higher rates. The consistent atrioventricular synchrony with elevated heart rates with this dual-chamber leadless system contrasts that observed with leadless VDD pacemakers, which show a significant decline in atrioventricular synchrony at heart rates >100 bpm.¹⁵ Another observational study of leadless VDD pacemakers reported significantly lower atrioventricular synchrony for sinus rates >80 bpm (33% versus 91%); at higher heart rates, reports demonstrated that accelerometer signals associated with ventricular filling and atrial contraction can fuse, thus confounding proper mechanical identification of atrial events.¹³

This patient population included a broad spectrum of RA and RV implantation sites. Differences in atrioventricular synchrony were not observed across specific RA implantation sites, RV implantation sites, or RA+RV implantation site combinations. However, these anatomical LP locations may not fully capture the factors that can affect i2i communication and atrioventricular synchrony (eg, relative distance and orientation of the LPs) due to patient-to-patient differences in cardiac anatomy. A more comprehensive, systematic evaluation of fluoroscopy collected at implantation is warranted. It is important to note that the i2i setting level can be manually adjusted beyond the automatic recommendations to achieve a patient-specific balance of i2i success and battery longevity.

The analysis also included a range of atrioventricular event combinations (ie, Ap, As, Vp, Vs) as a natural result of exercise-induced heart rate variability and the broad patient population, which included individuals with sinus node dysfunction and atrioventricular block indications alike. Although atrioventricular synchrony was delivered in at least 97% of beats across all distinct atrioventricular event combinations, greater atrioventricular synchrony was associated with either atrial pacing or ventricular sensing. During atrial pacing, greater atrioventricular synchrony could result from the VLP maintaining both the V-A and V-V delays, despite transient A-to-V i2i interruption. Greater atrioventricular synchrony during atrial pacing may also, in part, be an artifact of adjudication because pacing spikes are more identifiable than sensed P waves, which may go unnoticed. In contrast, greater atrioventricular synchrony during ventricular sensing may be explained by its lack of dependence on A-to-V i2i communication to initiate the atrioventricular delay for timely ventricular activity when intact atrioventricular conduction was present. This scenario may help explain cases in which the proportion of beats with atrioventricular synchrony exceeds that with A-to-V communication success.

The criteria used to classify each beat as atrioventricular synchronous are critical to defining true atrioventricular synchrony. Aside from using a standard PR interval upper limit of 300 milliseconds, this analysis introduced a new, programming-specific definition of atrioventricular synchrony that accounts for the intended functionality of each implanted pacemaker. With these strict criteria, true atrioventricular synchrony was still achieved in 96.3% of beats, only 1.7% lower than when the more liberal, programming-agnostic 300-millisecond criterion was used. This additional and more accurate quantification of true atrioventricular synchrony further illustrated the functionality of beat-to-beat i2i communication for true dual-chamber leadless pacing.

Limitations

Of the 464 de novo patients initially enrolled, 80 (17%) were excluded from the final analysis. However, the reasons for exclusion (eg, withdrawn consent, LP fixation, improper device programming, ECG signals with persistent noise or low amplitudes) were independent of i2i communication and thus not expected to bias the reported communication success or downstream atrioventricular synchrony. Separately, individual beats within each Holter recording window that were not evaluable by the core laboratory were excluded, and the results may not represent certain scenarios (eg, ectopic beats, intermittent electrocardiographic noise due to vigorous body movement). However, the majority (89.7%) of the 2685 collected recordings were analyzed, and only 3.5% of recorded patients were not included because of uninterpretable Holter monitors.

This analysis was limited by the relatively short Holter recording windows (2–5 minutes) and restricted set of prescribed postures/activities, which could capture neither the range nor distribution of activities in the full spectrum of daily life. In addition, such out-of-clinic settings may also include environmental electrical noise that could potentially interfere with i2i communication. An evaluation of atrioventricular synchrony during ambulatory 24-hour Holter periods without programming restrictions or prescribed postures/activities should be the subject of future analyses, as should longer-term i2i communication performance. Such an evaluation may also reveal the impact of heart rate changes or time spent at elevated heart rates on atrioventricular synchrony because the current approach relied on the mean heart rate for each recording.

Although atrioventricular synchrony was quantified electrically, the long-term downstream impact on patient symptoms or quality of life was not evaluated and requires further study. Potential errors inherent with human electrocardiographic adjudication pose another limitation; activity-related noise artifacts and underlying patient rhythm variation may have resulted in inaccurate event identification (eg, incorrect classification of events as evaluable). Direct comparisons with transvenous pacemaker in the form of a 2-arm study could potentially clarify this issue.

Conclusions

The first-in-human evaluation of the new dual-chamber LP demonstrated acute atrioventricular synchrony in 98.1% of beats, providing true dual-chamber DDD(R) pacing. A stricter, programming-specific definition of atrioventricular synchrony for each patient still yielded excellent results (96.3%). Atrioventricular synchrony was maintained across a wide array of postures, activities, heart rates, and pacing scenarios using beat-to-beat, wireless communication with a suite of backup safeguards. Device-based i2i communication success diagnostics were within 5% of the assessed atrioventricular synchrony in 99% of patients and thus may serve as a lower-bound surrogate to an otherwise cumbersome manual atrioventricular synchrony measurement.

ARTICLE INFORMATION

Sources of Funding

This study was funded by Abbott.

Disclosures

Dr Ip reports consulting fees from Abbott and Medtronic and serves as a Steering Committee member for Abbott and a Data Safety Monitoring Committee member for Boston Scientific. Dr Rashtian reports research grants and consulting fees from Abbott. Dr Exner reports consulting fees and research grants from Abbott and Medtronic and, unrelated to this article, reports consulting fees and research grants from Boston Scientific and GE Healthcare, as well equity in CorVista, Clarius, eMurmur, HelpWear, and ProtonIntel. Dr Reddy is a consultant to Abbott. Unrelated to this article, he serves as a consultant for and has equity in Ablacon, Acutus Medical, Affera-Medtronic, Apama Medical-Boston Scientific, Anumana, APN Health, Aquaheart, Atacor, Autonomix, Axon Therapies, Backbeat, BioSig, CardiaCare, CardioNXT/ AFTx, Circa Scientific, CoRISMA, Corvia Medical, Dinova-Hangzhou DiNovA EP Technology, East End Medical, EPD-Philips, EP Frontiers, Epix Therapeutics-Medtronic, EpiEP, Eximo, Farapulse-Boston Scientific, Field Medical, Focused Therapeutics, HRT, Intershunt, Javelin, Kardium, Keystone Heart, Laminar, LuxMed, Medlumics, Middlepeak, Neutrace, Nuvera-Biosense Webster, Oracle Health, Restore Medical, Sirona Medical, SoundCath, and Valcare; unrelated to this work, has served as a consultant for AtriAN, Biosense Webster, BioTel Heart, Biotronik, Boston Scientific, Cairdac, Cardiofocus, Cardionomic, CoreMap, Fire1, Gore & Associates, Impulse Dynamics, Medtronic, Novartis, Philips, Pulse Biosciences; and he has equity in DRS Vascular, Manual Surgical Sciences, Newpace, Nyra Medical, Surecor, and Vizaramed. Dr Doshi reports consulting fees from Abbott and serves as a steering committee member for Abbott. Dr Badie, J.R. Nevo, and A. Goil are employees of Abbott. Dr Defaye reports research grants from Abbott, Boston Scientific, and Medtronic. Dr Canby reports consulting fees from Medtronic. Dr Grazia Bongiorni serves as a steering committee member for Abbott. Dr Shoda serves as a Steering Committee member for Abbott. Dr Hindricks serves as a Steering Committee member for Abbott. Dr Knops reports research grants, consulting fees, and serving as a Steering Committee member for Abbott and Boston Scientific and speaking fees from Medtronic.

Supplemental Material

Figures S1–S6

Table S1