Abstract
Aims
Generator longevity is the key issue for patients, and is also important for payers, yet implanters of Cardiac Implantable Electronic Devices (CIEDs) face a challenge when selecting the appropriate device since battery longevity is only known for previous generation devices and whilst projected longevities are available for current devices, these are not in comparable formats. This study presents a new framework that facilitates an estimation of longevities for all CIEDs of both previous and existing generations that could simplify personalization of the device choice.
Methods
Longevity can be calculated based upon a simple concept entitled the “power consumption index” (PCI = t x I/C, where t is a constant of 1 hour, I is the current required by the device and C, its battery capacity). We retrieved published data from the user manuals of all commonly used pacemakers including single chamber, dual chamber, cardiac resynchronization and leadless devices. C and the components of current I including background current (Ibackground) and the pacing current (Ipacing) were calculated prior to calculation of the PCI for each device. Subsequently, a set of fictitious patient pool conditions via a Monte-Carlo simulation were used to model CIED survival curves which were then compared with real-life data from the Swedish device registry of previous generation CIEDs. Finally, we modeled survival curves for current generation devices using the PCI model.
Results
Using the PCI approach we were able to calculate longevities for all pacemaker devices under a variety of settings. The modeled Ibackground matched the data reported by manufacturers, and, under a variety of settings, regression analysis showed a low average error rate between industry-reported and modelled longevities (ratio: modelled longevity/industry reported longevity −1) = 0.1 ± 4.0% and 0.1 ± 0.7% for previous and existing SR/DR devices, 1.0 ± 5.0% and 0 ± 3.0% for previous and existing CRT-P, and 0 ± 4.0% for leadless pacemakers, respectively).
More than 50% of the PCI and thereby a significant contributor to longevity was accounted for by Ibackground. Ipacing was the second largest contributor (20% for standard single and dual chamber devices, 30% for CRT-P and 40% for leadless devices). Certain pacing algorithms and IEGM storage considerably impacted specific devices with longevity losses of up to 1 year. The Monte-Carlo analysis demonstrated consistency between projected longevities by the PCI model and real-life data for historical devices and the calculated longevities that stemmed from this were consistent with the real-world data from Sweden.
Conclusion
The PCI model combining power consumption and battery capacity allows a comparison of longevity across CIEDs and programming options. Such a tool could help implanters improve personalization of device prescription for their patients and payers to make more informed decisions about tailoring device purchases and programming most appropriate for their population.
Citation: Defaye P, Boveda S, Billuart J-R, Witte KK, Paton MF (2025) Cardiac implantable electronic devices’ longevity: A novel modelling tool for estimation and comparison. PLoS One 20(9): e0333195. https://doi.org/10.1371/journal.pone.0333195
Editor: Hamed Tavolinejad, University of Pennsylvania Perelman School of Medicine, UNITED STATES OF AMERICA
Received: April 18, 2025; Accepted: September 10, 2025; Published: September 29, 2025
Copyright: © 2025 Defaye et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Third-party data was publicly sourced for this study from the Swedish ICD & Pacemaker Registry (https://www.pacemakerregistret.se/icdpmr/start.do). The authors confirm others can replicate the study findings in their entirety by directly obtaining data from the Swedish ICD & Pacemaker Registry and following information outlined in the Methods section. The authors had no special access privileges that others would not have when attempting to access the minimal data from the Swedish ICD & Pacemaker Registry. All other relevant data for this study are within the paper and its Supporting information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: P.D. receives grants and honoraria from Medtronic, Boston Scientific, Abbott, and Microport, S.B. is a consultant for Medtronic, Boston Scientific, Microport, JRB is an employee of Microport CRM who supervised Maxime Corneloup as an intern, KKW has received research funding from the British Heart Foundation, the National Institute for Health Research, the Medical Research Council. He has also received grants and honoraria for teaching and consultancy work from Medtronic, Cardiac Dimensions, Novartis, Abbott, BMS, Pfizer, Bayer and has received an unconditional research grant from Medtronic to support a PhD program at the University of Leeds. These competing interests do not alter our adherence to PLOS ONE policies on sharing data and materials.
Abbreviations: CIED, Cardiac Implantable Electronic Devices; C, battery capacity; CRT-P, Cardiac resynchronization therapy with pacemaker; I, current drain; Ibackground, background current; Ipacing, pacing current; Iremote/IEGM/Algo, current for optional settings; Iremote, current for remote monitoring; IIEGM, current for IEGM storage; PCI, Power Consumption Index; RVPa, right ventricular pacing avoidance; Device manufacturers, Abt (Abbott), Btk (Biotronik), Bsc (Boston Scientific), Mdt (Medtronic), Mcp (Microport)
Introduction
Longevity of Cardiac Implantable Electronic Devices (CIEDs) is an important issue for patients, who wish to avoid further surgery, and purchasers, who wish to optimize cost-effectiveness, and is therefore a relevant consideration for clinicians. It is appreciated that there are common discrepancies between declared (future) longevities of generators and their subsequent survival curve once implanted [1,2]. Despite calls for more transparency and industry-wide standardized reporting of longevity [3–5], comparisons of longevity between devices and manufacturers in different settings remains challenging. Given that 30–40% of all CIED procedures are generator replacements, there exists the risk of conflicts of interest for both manufacturers, and, in fee-for-service healthcare environments providers, that limit enthusiasm for transparency [6].
Although implanters and their patients appreciate the concept of battery capacity as a prime criteria for device longevity [7] they also recognize that energy drain plays a role [8,9]. However, the potential lifetime of the device is also determined by how energy is stored, and how efficiently it is delivered [10], along with the patient’s characteristics, all of which, add to the frustrating situation of complex and non-standardized user-manual declared longevity, with different programming as baseline across companies, making personalization of generator prescription impossible even in the presence of similar battery capacity. If one could reliably describe consumption and index this to battery capacity and pacing requirements, there remains the possibility of a reliable comparison of devices.
Based upon previous work defining the power consumption of CIEDs [11], which included a calculation for the inverse of longevity, we have developed the Power Consumption Index (PCI) (as defined by PCI = t x I/C (where t is a constant equal to 1 hour)) that aims to describe the intrinsic power consumption of the overall system (the pacing system coupled with the battery) during a normalized period (1 hour). The reciprocal of the PCI therefore allows a derivation of longevity (Fig 1 and S1 File).
Fig 1. PCI and nominal longevity.
Note: The Power Consumption Index for two illustrative devices. On the Y axis, the PCI index (τ x I/C with τ = 1 h) is split according to each current contribution (τ x I background/C, τ x I PACING/C, etc.). On the Y’ axis, the green lines represent the corresponding longevity in years (a nonlinear “1/x scale”; the inverse of the PCI index τ x I/C multiplied by 106 (I is in µA) and divided by 365x24 gives the corresponding longevity (in years).
The Power Consumption Index for two illustrative devices. On the Y axis, the PCI index (t x I/C with t = 1 h) is split according to each current contribution (t x I background/C, t x I PACING/C, etc..). On the Y’ axis, the green lines represent the corresponding longevity in years (a nonlinear “1/x scale”; the inverse of the PCI index t x I/C multiplied by 10^6 (I is in mA) and divided by 365x24 gives the corresponding longevity (in years).
Hence, the objectives of this study were 1) to create and test a “universal” model, based on the concept of PCI, that offers the opportunity to model longevity for any CIED; 2) to validate this by comparing the modeled survival curves of previous generation CIEDs with real world data and; 3) apply this to predict product survival curves for new generations of CIEDs.
Methods
Firstly, we collected battery capacity and estimated current drain for a variety of devices from manufacturers’ user manuals to calculate PCI values and thereby longevity. The user manuals are available from the five major CIED manufacturers: Abbott (Abt, formerly St Jude Medical, Sylmar, CA, USA), Boston Scientific (BSc, St Paul, MN, USA), Biotronik (Btk, Berlin, Germany), Microport CRM (Mcp, formerly LivaNova and Sorin, Clamart, France), and Medtronic (Mdt, Minneapolis, MN, USA). We also used the only web-based longevity calculator available from BsC [12]. Detailed results are provided in S1 File.
Battery capacity
Nominal voltage and battery capacities at ERI were systematically retrieved from user manuals for single and dual-chamber (SR and DR) pacemaker, cardiac resynchronization therapy (cardiac resynchronization therapy with pacemaker [CRT-P]) and leadless pacemaker (PM) models. Representative models of previous and new generations of CIEDs were selected based on experts’ opinion (Table 1: Device models).
Current drain modelling
Industry-reported CIEDs’ longevities in user manuals depend upon the programming (including the activation of specific algorithms such as rhythm storage, remote monitoring, and sensors). Current is not provided in manuals and therefore modeling was required to calculate the PCI for each CIED. Because nominal settings differ from one manufacturer to another, an additional step was necessary to specifically identify and estimate the current for background device activity and pacing (Ibackground, Ipacing) and optional settings (Iremote IIEGM Ialgo). For Ibackground, and Ipacing the evaluation was done by a regression analysis. For Iremote/IEGM/algo the current was estimated via the difference of longevity between activated and deactivated settings for each option.
Two categories of optional settings were considered. The first considered algorithms directly influencing the Ipacing drain such as automatic threshold management or reduction of right ventricular pacing (RVP) percentage (Ialgo) and the second explored optional settings such as remote monitoring and IEGM storage (Iremote and IIEGM). Unlike Ibackground or basic pacing, each factor needed to be analyzed individually (S1 File).
In addition, for the 20 previous generation single chamber and dual chamber pacemakers under investigation, 674 settings were considered, for 11 new generation devices 243 settings were explored and for 8 previous and 5 current CRT-P devices, respectively 294 and 177 settings. Finally for 3 leadless devices 156 settings were considered.
Estimation of previous devices’ nominal longevities, using the PCI model
A literature review was performed to retrieve standard pacemaker programming parameters (for example, heart rate, threshold, pulse amplitude, pulse duration) in clinical studies, registries and clinical practice. Data related to specific algorithms (reduction of ventricular pacing (hysteresis and MVPTM, aCRTTM), automatic management of pacing output) were also collected via clinical studies, when available. Where the effect on RV pacing of the algorithm was missing, for example with hysteresis and aCRTTM, these were imputed using data provided for clinical practice and assumed to be stable over the lifetime of the device. We also assumed that managed output algorithms would achieve and maintain the ideal of 1V. Then, under nominal conditions, the PCIs and the corresponding nominal longevities were estimated for the CIEDs examined (details in S1 File) via the battery capacity values and the current drain modeling using the formulas PCI = t x I/C and L = 1/PCI.
Validation of PCI model, using Monte-Carlo simulations
In order to validate the survival curves the model was applied not only according to nominal parameters but a variety of settings in order to reflect real-world patient characteristics and programming. A pool of fictitious patient sets (100,000 patients) was created via a Monte-Carlo simulation (programmed in Python™) with age, indication (sinus node dysfunction [SND], intermittent atrio-ventricular block [AVB], complete AVB) and programmed parameters based on available literature [13–20]. The parameters used for the Monte-Carlo model are described in S2 File. Right ventricular pacing avoidance [RVPa] algorithms were assumed to be applied for SND patients eligible for DDD pacing. Remote monitoring was not standard for previous generation devices. We hypothesized a 50% adoption rate of remote for new generation current devices. The impact of settings (such as capture management, additional IEGM storage, RVPa for intermittent AVB for pacemaker, aCRT™ and MPP™ for CRT) on energy consumptions were studied.
For each device, longevity was calculated per patient using via the PCI energy consumption formula. Missing information which could not be derived from manuals was hypothesized, while assuming similarities among same generation devices. When longevity exceeded patient life expectancy, data were censored (since end of service uncommonly occurs simultaneously with death, and residual battery life of the device is rarely collected at death). The distribution of PCI and corresponding longevity across the pool of fictitious patients allowed the drafting of product survival curves for each cardiac implant. The PCI model was then validated for previous generation’s devices by comparing these modeled product survival curves and real-life data. For real-world survival curves, we used the Swedish registry [21] which was started in 1989 on the initiative of the Swedish Society of Cardiology. All the implanting clinics in Sweden report to the registry that compiles quarterly and annual reports of pacemaker use in Sweden. Every year there are about 5000 pacemaker procedures in Sweden. The real life product survival curves were extracted from these reports
Results
Battery capacity retrieved from manufacturers’ manuals
Battery capacities for single and dual chamber pacemakers have remained, on average, unchanged between previous generation and contemporary devices (close to 1 Ah), with disparities among manufacturers (Table 2).
In CRT-P devices, most manufacturers use the same battery capacity as for their conventional pacemaker, close to 1 Ah. Only BSc offers a battery above 1 Ah (using the same technology as Accolade™ DR EL) with 1.5 Ah at elective replacement indicator (ERI).
Battery capacity for single and dual chamber leadless pacemakers is very different from standard pacemakers. For VVI pacing, the Micra™ SR capsule is equipped with a 0.12 Ah battery, while the Aveir™ capsule battery is 0.24 Ah. Aveir™ is also available in a dual chamber configuration with the same ventricular capsule combined with a smaller atrial capsule (equipped with a 0.17 Ah battery capacity).
Current drain modeling
Background current (Ibackground) and pacing current (Ipacing).
For conventional pacemakers, the difference between industry-reported longevities and those derived by regression was 0.1 years ± 4% for previous generation devices and −0.1 years ± 0.7% for new generation devices. For devices with a variety of configurations (Mdt and BSc via its longevity calculator website), the regression coefficient (R2) exceeded 90%, across all settings applied.
The Ibackground derived by regression analysis (Table 3) matched those reported by manufacturers for most devices with few exceptions. There was, however, a significant change between previous and new generation devices. Previous generation devices relied on a Ibackground exceeding 9–10 µA (except for Identity™ for which the Ibackground ranged between 5.72 µA and 6.19 µA, the Evia™-T for which the Ibackground ranged between 6 µA and 6.66 µA and Symphony™ with an Ibackground at 6µA). For new generation CIEDs, the Ibackground did not exceed 7 to 7.5 µA, except for BSc devices, which reached 9.7 µA to 10.4 µA. Similar results were observed for CRT-P. For leadless pacemakers, the Ibackground was much lower than for conventional pacemakers (ranging from 0.8 µA to 0.94 µA for SR and 1.75 µA for DR devices, including 0.81 µA to 1µA for communication between the atrial and the ventricular capsules).
The analysis conducted in S1 File (Pacing current) shows that for the following settings (60 bpm, pulse width 0.4 ms, lead impedance 500 ohms, with 100% pacing), the Ipacing was relatively consistent across all categories of contemporary pacemakers with average current drains of 1.98 ± 0.12µA at 2.5V pacing output and 4.37 ± 0.24 µA, at 3.5V.
Current from optional settings (Ialgo/remote/sensorIEGM).
Reduction of ventricular pacing algorithms (RVPa) or Adaptive CRT (aCRTTM) directly modulates the percentage of RV pacing. Depending on the amount of pacing avoided, Ipacing decreases from 1.98–4.37 µA, respectively at 2,5V and 3,5V (100% pacing), down to 0,89−1,97µA (45% of pacing) and 0,59−1,31µA (30% of pacing). User manuals report that these savings are achieved without energy cost. On the other hand, multipoint point pacing for CRT (MPPTM) increases Ipacing on the left ventricular channel up to 3,96 and 8,74µA respectively at both 2,5V and 3,5V.
Threshold algorithms whose sole objective is to guarantee capture, increase pacing outputs and pacing current (user manuals do not specify impact on longevity). Other automatic threshold algorithms aim at optimizing outputs and do this either daily (Capture ManagementTM from Mdt and Capture ControlTM from Btk) or on beat-to-beat basis (Auto captureTM from Abt, Automatic captureTM from Bsc). User manuals do not report an energy cost for the daily algorithms but an energy cost of 1 µA can be derived for beat-to-beat algorithms. The analysis conducted for BsC device (see S1 File: Algorithms influencing pacing current) shows that the beat-to-beat algorithm saves energy (current drain) only if the percentage of pacing is high (>60%) or outputs exceed 2,5V when the algorithm is deactivated.
Rate adaptive pacing usually relies on a G-sensor (accelerometer) to adapt pacing rate according to effort. Adaption of pacing rate can be optionally enhanced with the combination of a minute ventilation (MV) sensor. User manuals describe an estimation of the energy consumption by the MV sensor of (0.69 µA- 0.77 µA). The impact of rate adaptive technology on pacing is unknown and somewhat unpredictable.
For most suppliers, IEGM storage is embedded as a standard function and the energy cost related to EGM is already included in the current background. Only Mdt reports a specific impact on longevity (see S1 File: IEGM). While 6-month storage has minimal effects (0,11–0,34 µA for EnpulseTM, 0,04–0,11 µA for AzureTM), the optional additional use of pre-arrhythmia EGM storage, increases current drain and reduces projected service life by approximately by 34% or 4 months per year for Enpulse (equivalent to 5,67−6,16 µA for EnpulseTM 1,3–1,7µA for AzureTM).
For remote monitoring, assuming 2−4 transmissions per year, the current consumption is around 1,14−1,75 µA for RF solutions (BsC, Btk) while it is 0,09−0,59 µA for Bluetooth connectivity (Mdt, McP). Btk provides a unique solution as its devices transmit data daily (alerts are managed via its website) with a fixed energy cost close to 1,75 µA.
Estimation of nominal longevities, using the PCI model
After deriving current drain, the PCIs were computed and corresponding longevities were modelled, at nominal settings, across all devices. Standard settings, PCI value per device and corresponding longevity, evolution between previous and new generation devices, as well as a sensitivity analysis are reported in S2 File. Fig 2 compares each device per category (SR, DR, CRT-P) according to the PCI chart presented previously. Across all devices, the PCI values range from 26.9 (corresponding to a longevity of 4.2 years) (Enpulse) and 7.6 (with an estimated longevity of 15.1 years) for the single chamber Aveir ventricular device if the programmed output is set to 1.5V).
Fig 2. PCI and longevity with nominal settings for pacemaker devices.
The Power Consumption Index and longevity of previous and current devices describing the contribution of each setting. For conventional pacemakers, the settings considered were: basic rate: 60bpm, pacing threshold: 2.5V, pulse duration: 0.4ms and impedance: 500 ohms for both A&V. For VVI pacemakers, ventricular pacing: 90%. For dual chamber pacemakers, atrial pacing: 70% for SND, 30% for AVB (51% on average) and ventricular pacing assumptions accounted for the difference between AAI/DDD mode and other RVP algorithms (29% vs 47%). Options such as sensor, IEGM storage and remote monitoring (2 transmissions/year) are reported. For leadless VVI and DDD pacemakers, the settings used were: basic rate: 60bpm, pulse duration: 0.25ms for Micra™ and 0.4 ms for Aveir™, impedance: ~ 600 ohms for ventricle and ~300 ohms for the atrium. Pacing outputs were not reported in studies and two options were considered: 1.5 V or 2.5V reflecting the level of confidence of practitioners in adapting output (thresholds observed were typically low: 1.25 V at implant and 0.75 V weeks after). Pacing percentages were the same as for conventional pacemakers. Hysteresis mode was applied for DDD. For CRT-P, BIV was the standard pacing mode (60bpm, 50% A, 100% BIV, 500 ohms) with alternative options such as aCRTTM or MPPTM pacing. Ext. IEGM: extended IEGM, FUP: interrogation of cardiac implant via inductive during a face to face follow-up (one per year).
The Power Consumption Index and longevity of previous and current devices describing the contribution of each setting.
For conventional pacemakers, the settings considered were: basic rate: 60bpm, pacing threshold: 2.5V, pulse duration: 0.4ms and impedance: 500 ohms for both A&V. For VVI pacemakers, ventricular pacing: 90%. For dual chamber pacemakers, atrial pacing: 70% for SND, 30% for AVB (51% on average) and ventricular pacing assumptions accounted for the difference between AAI/DDD mode and other RVP algorithms (29% vs 47%). Options such as sensor, IEGM storage and remote monitoring (2 transmissions/year) are reported. For leadless VVI and DDD pacemakers, the settings used were: basic rate: 60bpm, pulse duration: 0.25ms for Micra™ and 0.4 ms for Aveir™, impedance: ~ 600 ohms for ventricle and ~300 ohms for the atrium. Pacing outputs were not reported in studies and two options were considered: 1.5 V or 2.5V reflecting the level of confidence of practitioners in adapting output (thresholds observed were typically low: 1.25 V at implant and 0.75 V weeks after). Pacing percentages were the same as for conventional pacemakers. Hysteresis mode was applied for DDD. For CRT-P, BIV was the standard pacing mode (60bpm, 50% A, 100% BIV, 500 ohms) with alternative options such as aCRTTM or MPPTM pacing. Ext. IEGM: extended IEGM, FUP: interrogation of cardiac implant via inductive during a face to face follow-up (one per year).
On average, the PCI for conventional pacemakers is lower for current generations compared with previous generations leading to an increase in longevity for both SR and DR devices (10.8 years vs. 15.4 years for SR and 11.2 years vs. 14.3 years for DR). Unlike standard pacemakers, the average PCI for CRT-P increased (from 12.5 to 14.1) with the introduction of remote monitoring leading to a reduction of longevity (from 8.3 years to 7.8 years). For leadless devices, the PCI reached 14.2 (corresponding to a longevity of 8.8 years) for dual chamber and 11.7 (10.6 years) for single chamber devices demonstrating the consequence of energy cost of transmission between capsules in the two-chamber system.
The split of PCI per current highlights a strong impact of the Ibackground for all categories of pacemaker: more than 50% of PCI is due to Ibackground. The reduction of total PCI for SR/DR between previous and new generations of conventional CIEDs resulted primarily from the reduction of the Ibackground. For conventional pacemakers, Ipacing accounted for only 20% of the total PCI for SR/DR pacemakers and for 30% of the PCI for CRT-P.
Among the contemporary devices, Accolade™ SR/DR had the highest PCI, and thus the lowest estimated longevity (PCI of 13.5 and longevity of 8.5 years for SR, PCI of 14.2 and longevity of 8 years for DR). This is because this device had the highest Ibackground (10µA) as compared with other devices in the same category. On the other hand, Accolade™ DR EL, even with activated remote monitoring, had the lowest PCI (8.9), and thus the longest longevity (12.8 years), thanks to the high battery capacity at 1.6Ah.
For the other devices, differences were primarily due to optional settings such as extended IEGM, sensor or remote monitoring. In the past, IEGM storage negatively impacted Mdt device longevity. This has been significantly improved upon for standard pacemakers. Moreover, remote monitoring power consumption is three times higher for RF solutions than Bluetooth solutions (PCI for remote monitoring 1.2 vs 0.4). For example, the estimated longevity for Edora™ 8 reached 9.7 years for SR and 9.1 years for DR. Azure™ benefits from a low remote monitoring power consumption from Buetooth and reached 12.7 years for SR and 11.7 years for DR (extended IEGM turned “Off”). Alizea™ SR and DR benefit from additional battery capacity especially if remote monitoring is switched off, such that nominal longevity reached 12.9 years, similar to that of the Accolade™ DR EL device. The impact of activating the G-sensor was not different between devices.
For leadless SR and DR devices, assumptions included a basic rate of 60bpm, a pulse duration of 0,25 ms for Micra™ (Mdt) and 0,4 ms for Aveir™ (Abt), an impedance of ~600 ohms for ventricular pacing and ~300 ohms for atrial pacing [22–25]. Pacing outputs were not reported in studies and two options were considered: 1.5V or 2.5V, reflecting the level of confidence of practitioners in adapting output (thresholds observed were typically low: 1.25 V at implant and 0.75 V weeks after). Pacing percentages were the same as the one for conventional pacemakers. Hysteresis mode was applied for DDD.
Unlike conventional pacemakers, PCI related to pacing in leadless pacemakers accounted for more of the total PCI (40% on average, higher with 2.5V outputs). This is the consequence of a lower battery capacity and thus, a greater proportion of total available energy is required for pacing. Consequently, PCI and longevity significantly changed depending on pacing output assumptions (Aveir™ SR: 15.1 years if at 1.5V vs. 10.6 years if at 2.5V; Micra™ SR:10.2 years vs. 6.6 years, respectively; Aveir™ ventricular device (DDD mode): 12.5 years vs. 10.6 years; Aveir™ atrial capsule (DDD mode) 7.3 vs. 5.4 years, respectively).
For CRT-P, biventricular pacing is the standard pacing mode albeit there are additional options such as aCRTTM or MPPTM pacing. Biventricular pacing with the Visionist CRT-P device is associated with a PCI of 10.7 and a longevity of 10.6 years, whereas the Percepta™ device with aCRT activated achieves a PCI of 12.3 and longevity of 9.3 years. Not surprisingly Quadra Allure™ with MPP™ activated suffered a considerable increase in PCI and a corresponding reduction in longevity (PCI: 16.9; longevity of 6.8 years).
Sensitivity analysis of longevity related to fluctuations of currents (see S2 File) reveal a standard deviation close to 3–4% (ratio: sigma divided by nominal longevity) across all devices. For a small, clinically achievable sample size this led to a 95% CI for nominal longevity of 0,3–0,7 years, whereas a simulation of 100 revealed a 95%CI of 0,04–0,10 years and for a simulation of 40,000 the 95%CI was 0,003–0007 years.
Validation of PCI model, using Monte-Carlo simulations
Previous generation devices.
Modeled survival curves with standard assumptions fitted Swedish registry data for multiple models with few exceptions (S2 File). For conventional pacemakers, modeled survival curves departed from real-life data for following models: Enrythm™ DR, Evia™ DR-T and Identity™ DR 5370. For Enrythm™, the programming of IEGM storage was the only parameter explaining the difference. For Evia™ DR-T and for Identity™ DR Adx, the difference between modeled survival curves could be explained by the automatic threshold management mode. Overall, the model showed a good fit for CRT-P available in the Swedish registry (InSync III™, Invive™, Frontier II™, Anthem™). This comparison could not be pursued further since real-world programming of implants is not available on the Swedish registry website. Nevertheless, the model showed consistency between real-world and modelled survival curves for most CIEDs from previous generations (Fig 2).
Estimation of current devices’ longevity, using the PCI model
Conventional pacemakers.
Survival curves for current devices are shown in Fig 3 and reported in S3 File. The 95% confidence intervals are extremely low, due to the size (100,000) of the simulated population, allowing a comparison across devices (S3 File). The aggregated survival curves for conventional pacemakers showed wide differences between devices and manufacturers. Accolade™ SR and DR have the shortest estimated lifespan while the extended longevity DR version of Accolade™ offered the best longevity. The impact of programming options is reported in S2 File.
Fig 3. Modelled survival curves for current generation devices without (3a) and with (50% remote transmission adoption and 2 yearly transmission) (3b) activated.
(*) No daily check, 2 Radiofrequency transmissions/year. The figures show the modelled curves for current generation devices.
The figures show the modelled curves for current generation devices.
Among possible settings, the beat-to-beat automatic threshold management algorithm (Auto captureTM, Automatic captureTM) tended to straighten product survival curve with energy savings on side of the inflexion point and energy cost on the other side, suggesting that optimal programming could extend median product longevity. Reduction of ventricular pacing for intermittent AVB via AAI/DDD mode (available for Btk, Mcp and Mdt devices) extended median longevity for the corresponding devices and reduced the difference between Accolade™ DR EL and Azure™ XT DR or Alizea™ DR. The number of remote transmissions had a marginal effect on longevity. Two devices (Edora™, Alizea™) showed a reduced longevity simply by activation of remote monitoring (by a one-off increase of current for Edora™ or by a reduction of battery capacity for Alizea™).
Leadless pacemakers.
In the single chamber segment, Aveir™ V significantly outperformed Micra™ pacemaker thanks to its larger battery capacity. Nevertheless, the Micra™ equipped with Capture™ management was associated with a survival curve at 1,5V that matched the Aveir™ survival curve at 2,5V (AutocaptureTM is not available on Aveirtm).
For two chamber leadless devices, two survival curves were considered, one for each capsule. The Aveir™ atrial capsule had a shorter lifespan compared with the ventricular capsule. The simulation emphasizes the need to optimize pacing outputs for the atrial capsule and the importance of using the RV pacing avoidance algorithm.
Discussion
The current study firstly describes a novel way to estimate generator longevity by combining current and battery capacity, and then validates this model across a variety of programming options and previous generation of devices by comparing the modeled data with observed longevity from a country-wide registry, and finally provides estimations of the longevities of currently implanted devices for which there are no reliable observed data.
Longevities from user manuals are difficult to use for implant decision making because manufacturers provide these with a variety of settings, pacing options and configurations. In addition, the lack of a common framework does not facilitate an understanding of the determinants of longevity and a comparison between devices. Calculations of longevity not only should use settings reflecting clinical practice but also split power consumption according to unavoidable current usage (background current) and optional algorithms to help practitioners in their implant decision and subsequent programming. We summarize here the key findings and a few recommendations.
Battery capacity and background current
Battery capacity as a standalone criterion is irrelevant. This study illustrates this using the examples of leadless pacemakers which, despite much lower battery capacity, achieve a reasonable longevity compared with a standard pacemaker. On the other hand, background current plays a key role (at least 50% of PCI is due to Ibackground across all devices). Leadless pacemakers for example, benefit from a substantially lower background current compared with standard pacemakers. Therefore, both battery capacity and background current need to be combined in order to provide the foundation for longevity assessment. This is the purpose of the index proposed here. The PCI combines background current (i.e., C/(365*24*10^-6 Ib) and battery capacity and thereby the energy capacity (in years) available for pacing and programming options. Applying this criterion could contribute to personalized device selection, with a focus on device longevity for a wide range of patients (see Table 4).
Pacing current
The present analysis did not reveal major differences between manufacturers within each category of device, but, by exploring the differences between conventional pacemakers and leadless pacemakers, demonstrated the relative importance of energy consumption and its impact on device longevity as long as lower pacing outputs (<1.5V) can be achieved without loss of capture. As pacing output reaches 2.5V, power consumption increases and longevity is significantly impacted particularly in those devices with lower battery capacity, emphasizing the need to target pacing output close to 1−1.5V. Particularly in leadless dual chamber devices, the atrial capsule which has a modest energy capacity has the potential to limit longevity if atrial pacing is above 80% due to the additional costs of inter-capsule communication. Presently therefore, a dual chamber approach to SND requires careful consideration, especially given the younger age of the affected population. On the other hand, AV Block, with reliable intrinsic atrial activity is likely to be a more useful application for these until further technological developments improve the energetic demand of this connection.
Current drain from optional programming.
Conventional pacemakers are often chosen specifically for their programming options but the impact of these on longevity is highly variable and depends upon the manufacturer and the device generation. Amongst RVP avoidance algorithms, AAI/DDD mode may be more effective in reducing RV pacing than other algorithms provided it can deal with intermittent AVB.
Automatic threshold management has the potential for an adverse effect on battery longevity if the threshold is low and should be deactivated in this situation. The benefit of automatic threshold management is also limited in patients where RVP avoidance is effective [17]. ‘Daily’ algorithms are also preferable since beat-to-beat algorithms have a larger energy cost. For leadless pacemakers, as longevity is very sensitive to pacing outputs, daily automatic threshold management may be more critical for longevity than for standard pacemakers.
Whether extended IEGM storage and remote monitoring are required should be carefully reviewed on a patient-by-patient basis. For some devices, activation of remote monitoring generates a “one off” energy cost; fortunately, the impact of the number of transmissions on longevity remains marginal.
Overall PCI and longevity
Overall there has been an increase in longevity of more than one year for conventional SR/DR pacemakers between previous and current generation devices, although there has been a reduction of longevity for CRT-P devices. For leadless devices longevity is similar to that of a standard single chamber pacemaker, but as described, the longevity of two chamber leadless devices is currently predicted to be substantially lower than a standard dual chamber pacemaker. Disparities between suppliers still prevail and should be taken into consideration by clinicians when deciding on device prescription.
Product survival curves
Our concept was successfully validated using real-life data from previous generation devices, supporting our proposal that the calculated forecasted product survival curves for current devices will be robust.
Next steps
In order to provide greater applicability of the PCI concept one could apply it to a prospective cohort where device programming and longevity outcomes are tracked, validating its predictive capacity beyond historical data and, in parallel conduct benchtop measurements for key programming features (e.g., automatic capture management, aCRT) to empirically validate modeled current assumptions. One could analyze how real-time remote monitoring logs correlate with PCI-predicted consumption, offering a feedback loop for future model refinement.
Limitations
Current drain modeling.
For some devices, pacing current data could not be obtained for all output values and were hypothesized by assuming similarities among same generation devices. These assumptions were mitigated via an analysis of the relationship between current and outputs highlighting consistency across devices. Options which did not specifically lead to an impact on longevity were assumed to require minimal current (e.g., RVP avoidance algorithms and standard IEGM storage) and are not considered as generating additional energy cost.
Harmonization of nominal conditions and calculation of nominal PCI: Validation of the model.
The literature review performed to retrieve standard pacemaker parameters provided information on the average settings of devices. As RVP avoidance algorithms were not consistently investigated through clinical trials, percentages of pacing achieved were assumed to be identical according to type (AAI/DDD mode, hysteresis mode such as SAV + ®, VIP®) for SND patients. On the other hand, the impact of AAI/DDD modes (MVP®, SafeR®, Vp suppression®) for intermittent AVB patients was investigated in clinical trials and therefore could be considered in the model. For these patients we considered a 1.5% increase per year of ventricular pacing to take into consideration progressive A-V node disease over time. Apart from this group, we had to assume that settings are stable beyond 2–3 years as we lacked clinical data over a longer time frame. Alternatively, another approach would have been to build a Markov model to incorporate these changes across time, but the probability of transition would have been ‘heuristic’ since such data are not available in literature. This would have added significant complexity to our model without changing our overall point and making the present manuscript less accessible.
Modeling survival curves.
The main limitation is the fact that real-life programming of implants was not accessible through the Swedish registry website and hence the model could not be tested with all the combination of options available. The consistency observed between real-life and modelled survival curves suggest that real-life programming does not significantly depart from the settings populated in the model, but this could not be verified.
Conclusions
Projected longevities of CIEDs are needed for device selection and optimal programming at the time of the implant. Information from user manuals remains difficult to apply, due to a lack of harmonization in the estimated longevities provided by manufacturers regarding settings and programming conditions. We present for the first time, a model based on the Power Consumption Index (PCI) which offers the potential to compare longevity between devices under multiple settings and programming conditions. Longevities estimated by the PCI model appear to be consistent with real-life data for multiple CIED models from different manufacturers. This information could provide implanters, and their patients, the opportunity for personalized pacemaker hardware prescription whilst also paving the way towards standardized reporting of CIED longevity.
Supporting information
S2 File. Power consumption index rational; Power consumption and nominal longevities; PCI and longevity model sensitivity analysis; Survival curves generated by the Monte-Carlo modelling; Settings and distribution used for Monte-Carlo simulations; Survival curves generated for previous generation devices, Impact of settings for previous generation standard pacemaker devices.
https://doi.org/10.1371/journal.pone.0333195.s002
(DOCX)
Acknowledgments
The authors thank Ernest W. Lau, MD for insightful discussion related to cardiac device implant longevity and Maxime Corneloup for programming Monte-Carlo simulations with Python™ software. The authors also want to thank the reviewers for their advice during the review process.
References
- 1. Schaer BA, Koller MT, Sticherling C, Altmann D, Joerg L, Osswald S. Longevity of implantable cardioverter-defibrillators, influencing factors, and comparison to industry-projected longevity. Heart Rhythm. 2009;6(12):1737–43. pmid:19959121
- 2. Alam MB, Munir MB, Rattan R, Adelstein E, Jain S, Saba S. Battery longevity from cardiac resynchronization therapy defibrillators: differences between manufacturers and discrepancies with published product performance reports. Europace. 2017;19(3):421–4. pmid:27009183
- 3. Munawar DA, Mahajan R, Linz D, Wong GR, Khokhar KB, Thiyagarajah A, et al. Predicted longevity of contemporary cardiac implantable electronic devices: A call for industry-wide “standardized” reporting. Heart Rhythm. 2018;15(12):1756–63. pmid:30063990
- 4. Boriani G, Merino J, Wright DJ, Gadler F, Schaer B, Landolina M. Battery longevity of implantable cardioverter-defibrillators and cardiac resynchronization therapy defibrillators: technical, clinical and economic aspects. An expert review paper from EHRA. Europace. 2018;20(12):1882–97. pmid:29757390
- 5. Censi F, Calcagnini G, Mattei E, Ricci RP, Zoni Berisso M, Landolina M, et al. Estimate and reporting of longevity for cardiac implantable electronic devices: a proposal for standardized criteria. Expert Rev Med Devices. 2021;18(12):1203–8. pmid:34854774
- 6. Dean J, Sulke N. Pacemaker battery scandal. BMJ. 2016;352:i228. pmid:26848134
- 7. Ellis CR, Dickerman DI, Orton JM, Hassan S, Good ED, Okabe T, et al. Ampere Hour as a Predictor of Cardiac Resynchronization Defibrillator Pulse Generator Battery Longevity: A Multicenter Study. Pacing Clin Electrophysiol. 2016;39(7):658–68. pmid:26875541
- 8. Lau EW. Technologies for Prolonging Cardiac Implantable Electronic Device Longevity. Pacing Clin Electrophysiol. 2017;40(1):75–96. pmid:27943326
- 9. Paton MF, Landolina M, Billuart JR, Field D, Sibley J, Witte K. Projected longevities of cardiac implantable defibrillators: a retrospective analysis over the period 2007-17 and the impact of technological factors in determining longevity. Europace 2020;22:149-155. doi: 10.1093/europace/euz222. Erratum in: Europace. 2020 Jan 1;22(1):155.
- 10. Evans JM, Cleves A, Morgan H, Millar L, Carolan-Rees G. ENDURALIFE-Powered Cardiac Resynchronisation Therapy Defibrillator Devices for Treating Heart Failure: A NICE Medical Technology Guidance. Appl Health Econ Health Policy. 2018;16(2):177–86. pmid:29086228
- 11. Lau EW. Longevity decoded: Insights from power consumption analyses into device construction and their clinical implications. Pacing Clin Electrophysiol. 2019;42(4):407–22. pmid:30802980
- 12. Boston Scientific Longevity calculator, https://www.bostonscientific.com/en-EU/medical-specialties/electrophysiology/device-longevity/longevity-calculator.disclaimer.html
- 13. Udo EO, van Hemel NM, Zuithoff NPA, Barrett MJ, Ruiter JH, Doevendans PA, et al. Incidence and predictors of pacemaker reprogramming: potential consequences for remote follow-up. Europace. 2013;15(7):978–83. pmid:23419656
- 14. Auricchio A, Ellenbogen KA. Reducing Ventricular Pacing Frequency in Patients With Atrioventricular Block: Is It Time to Change the Current Pacing Paradigm? Circ Arrhythm Electrophysiol. 2016;9(9):e004404. pmid:27637555
- 15. Stockburger M, Defaye P, Boveda S, Stancak B, Lazarus A, Sipötz J, et al. Safety and efficiency of ventricular pacing prevention with an AAI-DDD changeover mode in patients with sinus node disease or atrioventricular block: impact on battery longevity-a sub-study of the ANSWER trial. Europace. 2016;18(5):739–46. pmid:26612882
- 16. Biffi M, Melissano D, Rossi P, Kaliska G, Havliĉek A, Pelargonio G, et al. The OPTI-MIND study: a prospective, observational study of pacemaker patients according to pacing modality and primary indications. Europace. 2014;16(5):689–97. pmid:24500787
- 17. Benezet-Mazuecos J, Iglesias JA, Rubio JM, Cortés M, de la Cruz E, de la Vieja JJ, et al. Limitations of the AutoCapture™ Pacing System in patients with cardiac stimulation devices. Europace. 2014;16(10):1469–75. pmid:24789975
- 18. Pierantozzi A, Landolina M, Agricola T, Lunati M, Pisanò E, Lonardi G, et al. Automatic adjustment of stimulation output in resynchronization therapy: impact and effectiveness in clinical practice. Europace. 2011;13(9):1311–8. pmid:21515591
- 19. Zoppo F, Gagno G. Left ventricle automatic pacing threshold management in CRT systems: A comprehensive review. J Cardiovasc Electrophysiol. 2020;31(9):2489–98. pmid:32573048
- 20. Martin DO, Lemke B, Birnie D, Krum H, Lee KL-F, Aonuma K, et al. Investigation of a novel algorithm for synchronized left-ventricular pacing and ambulatory optimization of cardiac resynchronization therapy: results of the adaptive CRT trial. Heart Rhythm. 2012;9(11):1807–14. pmid:22796472
- 21. Swedish ICD, Pacemaker Registry, https://www.pacemakerregistret.se/icdpmr/start.do
- 22. Reynolds D, Duray GZ, Omar R, Soejima K, Neuzil P, Zhang S, et al. A Leadless Intracardiac Transcatheter Pacing System. N Engl J Med. 2016;374(6):533–41. pmid:26551877
- 23. Piccini JP, Stromberg K, Jackson KP, Laager V, Duray GZ, El-Chami M, et al. Long-term outcomes in leadless Micra transcatheter pacemakers with elevated thresholds at implantation: Results from the Micra Transcatheter Pacing System Global Clinical Trial. Heart Rhythm. 2017;14(5):685–91. pmid:28111349
- 24. Knops RE, Reddy VY, Ip JE, Doshi R, Exner DV, Defaye P, et al. A Dual-Chamber Leadless Pacemaker. N Engl J Med. 2023;388(25):2360–70. pmid:37212442
- 25. Shantha G, Brock J, Singleton MJ, Schmitt AJ, Kozak P, Bodziock G, et al. A comparative study of the two leadless pacemakers in clinical practice. J Cardiovasc Electrophysiol. 2023;34(9):1896–903. pmid:37522245