Remote monitoring of cardiac implantable electronic devices

Abstract

Patients with cardiac implantable electronic devices (CIEDs) have been followed with periodic clinic visits. The number of patients with CIEDs has been increasing and CIEDs have become more complex. The workload of both medical staff and patients for CIED follow-up has also been increasing. Remote monitoring (RM) technology has undergone many developments, and RM has been used since 2008 in Japan. The benefits of RM are evident, but there are also problems with the technology. Different systems and various skills are required for RM management compared to conventional follow-up methods.

1 Introduction

Patients with cardiac implantable electronic devices (CIEDs) have been followed with periodic clinic visits and have received direct interrogation by a programmer which checks the battery, lead impedance, sensing amplitude, pacing threshold, and arrhythmic events. The number of patients with CIEDs has been increasing and CIEDs have become more complex. Follow-up frequency varies depending on the facility, physician preference, and available resources. Checks at clinics every 3–6 months have been recommended with increased frequency in response to product advisories and recalls [1]. The workload of both medical staff and patients for CIED follow-up has also been increasing. Technology-assisted medicine has provided many benefits. Remote monitoring (RM) technology has undergone many developments from the original transtelephonic monitoring of pacemakers for battery levels to currently available CIEDs with wireless telemetry capabilities. Various developments have occurred over the past decade, from fax reports to a social networking service system, from wired interrogation to wireless interrogation, and from one-direction transmission to bidirectional transmission. In Japan, RM has been used since 2008. Currently, 5 CIED companies in Japan use RM, and 27,700 patients in total have been followed as of December 2013. Because RM is a new technology, it has both benefits and problems.

2 Technological aspects of remote monitoring

RM data are transmitted from CIEDs to a transmitter station (Fig. 1) either by wired or wireless communication. Two types of transmitters exist: stationary and mobile transmitters. Only stationary transmitters are used in Japan. In addition to scheduled data transmission, alert-triggered data can be transmitted depending on the CIED [2]. Such a transmitter is linked by a telephone line to a central secure server website to store the transmitted data for further analysis. Companies use various telephone line types, including analog landlines, digital landlines, and a global system for mobile communications (GSM) network. It is sometimes difficult to set up a landline transmitter because a cable must connect the transmitter and the landline connector. Since RM data are likely to be transmitted at night, the transmitter should be set up in the bedroom rather than the living room. However, in traditional Japanese homes, there are no landline connectors in bedrooms. A GSM transmitter would be appropriate for such homes. After successful transmission of RM data, medical staff can check detailed RM data on a website from anywhere. The volume and nature of transmitted data are almost the same as those of the data obtained from direct interrogation. Medical staff can receive alert notifications by fax, SMS, voice message, or email. Occasionally, we can receive precise RM data immediately following an event such as appropriate implantable cardioverter defibrillator (ICD) therapy, inappropriate ICD therapy, and CIED abnormality. Medical staff can also manually or automatically activate message calls to patients to remind them of abnormalities. However, since elderly patients may be unaware of an abnormal signal, a telephone call may be necessary to notify them of abnormalities. The Table 1 shows RM characteristics by company.

Fig. 1. Pictures of the unique transmitters from each company.

1 Comparison of remote monitoring systems among 5 companies.

	Medtronic CareLink®	Biotronik Home Monitoring®	St. Jude Medical Merlin.net™	Boston Scientific Latitude®	Sorin SMARTVIEW™
Transmitter	CareLink Monitor	CardioMessenger	Merlin@Home	LATITUDE Communicator	SMARTVIEW monitor
Communication between device and transmitter	Wireless or Wired	Wireless	Wireless	Wired	Wireless
Data transmitter	Landline	Landline and global system for mobile communications	Landline and global system for mobile communications	Landline	Landline
Recommended setup location	Bedroom	Bedroom	Bedroom	Anywhere	Bedroom
Effective range to communicate between device and transmitter	3m	2m	3m	–	2m
Frequency of transmission	Scheduled FU; alert events; on patient demand	Scheduled FU; daily FU; alert events	Scheduled FU; alert events	Scheduled FU, weekly check, on patient demand	Scheduled FU; alert events; on patient demand
Time required for data transmission	About 5min	About 5min	About 5min	About 2min	About 5min
Matching between device and transmitter	Unnecessary for special workload	Unnecessary for special workload	Pairing between device and transmitter	Pairing between device and transmitter	Pairing between device and transmitter
Time zone for communication between device and transmitter
Scheduled FU	AM 0:00–1:30	Time set on device (usually midnight), once per day	AM 2:00–4:00	AM 7:00 (button blinking on and off)	AM 0:00–5:00
Alert events	Once per 3h	Once per 76min	Only AM 2:00–4:00	–	AM 0:00–5:00
Feedback to patient via transmitter	Confirmation for successful interrogation and transmission	LED indicating normal status or call to clinic	LED indicating call to clinic, automated phone calls	Automatic instructive text messages on LCD screen and action light	–
Synchronism between transmiter and server	Once per 15 days	None	Once per 7 days	Once per 8 days	Once Per 7 days
Others	Multiuser access capable special monitor (Express), OptiVol lung fluid status alert, any already implanted devices available for RM	Digital data export capability intrathroracic impedance measurement	CoRVUE fluid status alert ST level measurement	Body weight and blood pressure monitoring capability wireless communication is not available in Japan, instead of west countries	–

3 Remote monitoring advantages 3.1 Reduced workload and costs associated with CIED follow-up for patients and medical services

RM can reduce the workload associated with CIED follow-up. Before RM became available, patients with CIEDs had to visit hospitals for periodic CIED checks. It is burdensome for patients living in rural areas to visit hospitals for CIED interrogation. Visiting hospitals is also burdensome for patients' families, because about half of the patients must be escorted by family members, primarily the patients' children, who must take time off work to escort the patients. It is also becoming difficult to make appointments for CIED checks because of the increasing number of patients with CIEDs caused by the aging society and expanded indications for CIED implantation. Patients must spend several hours and sometimes even half a day in the hospital to have their CIEDs checked, but the rate of intervention is very low [3].

However, the burden of visiting a hospital is greatly reduced by using RM. RM use reduces the burden of overloaded clinics and saves valuable time and resources. Varma et al. [4] reported that RM reduced total in-hospital CIED evaluations by 45% without affecting morbidity. In their study, 1339 patients with high-energy CIEDs were randomized 2:1 to home monitoring (HM) or conventional follow-up. Thirty-one patients in the HM group (3.4%) and 21 patients in the conventional group (4.9%) died (P=0.226). The overall adverse event rate was 10.4% for HM versus 10.4% for conventional care over 12 months (non-inferiority P<0.005, 1-sided; P<0.010, 2-sided) (Fig. 2). Crossley et al. [5] also reported that wireless RM with automatic clinician alerts was associated with a significant reduction in mean length of cardiovascular hospital stay. The CONNECT (Clinical Evaluation of Remote Notification to Reduce Time to Clinical Decision) study was a multicenter, prospective, randomized study that included 1997 patients with high-energy CIEDs who were followed for 15 months. The study revealed a decrease in mean length of stay per cardiovascular hospitalization visit from 4.0 days in the in-office arm to 3.3 days in the remote arm (P=0.002). Hindricks et al. [6] reported that in prophylactic ICD recipients with automatic daily RM, extension of the 3-month in-office follow-up interval to 12 months appeared to safely reduce the ICD follow-up burden during a 27-month period after implantation. The 12-month interval resulted in a major reduction in total number of in-office ICD follow-ups (1.60 vs. 3.85 per patient-year; P=0.001). No significant difference was found between the 2 groups in mortality, hospitalization rate, or hospitalization length during the 2-year observation period. Landolina et al. [7] reported that RM reduces the number of emergency department/urgent office visits and total healthcare use in patients with high-energy CIEDs. Thus, introducing RM in patients with high-energy CIEDs can safely reduce the office visit burden.

Fig. 2. Number of office visits per patient and adverse event-free survival rates in the control and Home Monitoring groups. (A) Cumulative hospital-based encounter for implantable cardioverter defibrillator evaluations (sum total of scheduled plus unscheduled). Before the 3-month window (90±30 days), these consisted of unscheduled visit evaluations. After the scheduled 3-month in-person device check clinic visits for both groups, there is a progressive divergence of curves. In the Home Monitoring (HM) group, the slow rise in the curve reflects the incidence of unscheduled evaluations. (B) There was no difference in safety between conventional and HM-based follow-up.

In patients with pacemakers, RM also reduces the workload associated with CIED follow-up. Mabo et al. [8] reported that RM was a safe alternative to conventional care and significantly decreased the number of ambulatory visits during long-term follow-up of permanently paced patients. Over 18.3 months of follow-up, 17.3% of the patients in the RM group and 19.1% of the patients in the control group experienced at least one major adverse event (non-inferiority P<0.01). The number of interim ambulatory visits in the active group was 56% lower (P<0.001) than that in the control group. Halimi et al. [9] reported that early discharge with RM after pacemaker implantation or replacement was safe and facilitated patient monitoring in the month following the procedure. Folino et al. [10] reported that remote follow-up of pacemakers is a reliable, effective, and cost-saving procedure in elderly, debilitated patients.

Patients are satisfied with RM both in terms of its ease of use and continuous connection to the follow-up center [7,11]. Landolina et al. [7] reported a more favorable change in quality of life from the baseline to the 16th month in patients with RM compared to patients in the control group (P=0.026). RM has great benefits for the patient's peace of mind, psychological well-being, and safety, especially following an advisory, and is therefore considered an important alternative to conventional follow-up [12–15]. Raatikainen et al. [16] examined physicians' and nurses' time required for follow-up by office visit and RM. The physicians' time required to review RM data was significantly shorter than the time needed to complete CIED follow-up visits in the clinic (8.4±4.5 vs. 25.8±17.0 min, P<0.001). Nurses also spent more time on office visits than on RM follow-up (45.3±30.6 vs. 9.3±15.9 min, P<0.001). Thus, RM is cost-effective [10,16,17].

During an office visit, RM data are also used instead of direct interrogation. By checking RM data before the patient visits, the time required for direct interrogation and intervention can be reduced, especially in patients with problems. Determining the cause of a problem and how it should be managed can take time. If a problem is first detected during an office visit, the patient may have to wait for a long time until the problem is resolved. However, this would not be the case if the problem is detected and resolved before the patient visits. RM causes significantly fewer stresses than conventional follow-up.

3.2 Early detection of clinical events by remote monitoring

RM can detect various clinical events earlier than conventional follow-up. Data obtained from direct interrogation can be detected remotely from patients at home. With automatic RM, various events can also be detected immediately. Varma et al. [4] reported more rapid detection of actionable events by RM than by conventional monitoring in patients with high-energy CIEDs. The median period from onset to physician evaluation of combined first atrial fibrillation (AF), ventricular tachycardia (VT), and ventricular fibrillation (VF) events with RM was 1 day, which was much shorter than the median period of 35.5 days with conventional care (AF: median, 5.5 vs. 40; interquartile range, 1–51.25 vs. 15.5–59; VT: median, 1 vs. 28; interquartile range, 1–6 vs. 6.5–69.25; VF: median, 1 vs. 36; interquartile range, 1–7 vs. 10–75; supraventricular tachycardia [SVT]: median, 2 vs. 39; and interquartile range, 1–19.5 vs. 8.5–69). RM also detected clinically asymptomatic (silent) problems early for combined first AF, VT, VF, or SVT events (median, 1 vs. 41.5; interquartile range, 1–6 vs. 10.5–70.25) (Fig. 4, bottom). System-related problems occurred infrequently (RM group, 14 vs. conventional group, 3). These problems included an elective replacement indicator (1 RM patient vs. 0 conventional patients) and atrial/ventricular lead out-of-range impedance (13 RM patients vs. 2 conventional patients). Statistical comparison was not possible because of the low incidence (Fig. 3). Crossley et al. [5] reported that wireless RM with automatic clinician alerts significantly reduced the time to a clinical decision in response to clinical events compared with that in standard office visits. The median time from a clinical event to clinical decision per patient was reduced from 22 days in the office arm to 4.6 days in the RM arm (P<0.001). Landolina et al. [7] reported that the time from an ICD alert condition to data review was reduced from 24.8 days in the standard arm to 1.4 days in the remote arm (P<0.001). Additionally, in patients with pacemakers, Mabo et al. [8] reported that the median delay in medical intervention with inter-quartile was 17 (4–48) days in the RM group vs. 139 (33–201) days in the control group, representing a mean 117-day gain in event detection (P=0.001). Thus, in patients with CIEDs, RM can detect various events earlier than conventional follow-up. Conversely, it is necessary to assess whether earlier detection of events results in clinical benefits for patients with RM or whether earlier detection of events excessively increases office visits, which may reduce clinical benefits. One study showed that early event detection may contribute to clinical benefits. ECOST [18] showed that patients with RM had 52% fewer inappropriate shocks, fewer hospital admissions for inappropriate shocks (3 vs. 11, P=0.02), and 76% fewer capacitor charges, leading to longer battery life.

Fig. 3. Periods of event onset to evaluation in the control and Home Monitoring groups. Home Monitoring enabled earlier physician evaluation of arrhythmias (left) and silent events (right). AF=atrial fibrillation, VT=ventricular tachycardia, VF=ventricular fibrillation, and SVT=supraventricular tachycardia.

Fig. 4. Survival comparison on and off the network by device type. For both implantable cardioverter defibrillator (ICD) and cardiac resynchronization therapy with defibrillator (CRT-D) recipients, annual and total survival rates were significantly better for patients who were transmitting device information to the network.

3.3 Possible reduction in mortality by remote monitoring

RM may reduce mortality in patients with CIEDs. Saxon et al. [19] reported that RM is associated with excellent survival. For 69556 ICD and cardiac resynchronization therapy with defibrillator (CRT-D) patients with RM registered in a network, 1- and 5-year survival rates were higher than those in 116,222 patients who received CIED follow-up in office visits (50% reduction; P<0.0001) (Fig. 4). Klersy et al. [20] reported that RM confers a significant protective clinical effect in patients with chronic heart failure (HF) compared with usual care. In total, 6258 patients and 2354 patients were included in randomized controlled trials (RCTs) and cohort studies, respectively. The median follow-up periods were 6 months for RCTs and 12 months for cohort studies. Both RCTs and cohort studies showed that RM was associated with significantly fewer deaths (RCTs: relative risk [RR]: 0.83, P=0.006; cohort studies: RR: 0.53, P<0.001) and hospitalizations (RCTs: RR: 0.93, P=0.030; cohort studies: RR: 0.52, P<0.001). Although there is no clear evidence that RM reduces mortality, RM may have such benefits.

3.4 Easy access to database systems

RM provides easy access to database systems. Classically, CIED data have been stored on paper in health records. Much time has been needed to gather, analyze, and research CIED data. However, RM data are stored in a firmly protected server and can be easily downloaded, resulting in reduced workload associated with research. RM is also helpful for checking a setting in a situation without a programmer. Precise setting data can be easily downloaded from the server.

4 What is obtained by remote monitoring? How should we act?

RM has many benefits. However, the total volume of RM data is extensive, and we have to understand what data are important and where the important data are located. Here, we describe what is obtained by RM and how we should act based on RM data. However, some suggestions may be limited to recommendations without any evidence because the obvious evidence based on CIED information is scarce.

4.1 Arrhythmic events

Arrhythmic events, particularly supraventricular arrhythmic events, are frequently reported in RM data [3,5,21].

4.1.1 Real-time electrocardiogram

We can recognize real-time intracardiac electrocardiogram (ECG) at the time of data transmission. Real-time ECG is useful because we can compare the intracardiac ECG in arrhythmic events to that in stable conditions. The morphology of sinus rhythm is useful for deciding whether an arrhythmic event is a supraventricular or ventricular arrhythmia.

4.1.2 Premature ventricular contractions

In almost all CIEDs, the premature ventricular contraction (PVC) frequency is recorded. Emergency action for PVC events is not necessary. If antiarrhythmic drugs are administered for PVC, the PVC counter is useful to judge the effect of therapy. However, the PVC counter is not always precise. CIEDs might misrecognize PVC for premature atrial contraction or T waves because of an atrial blanking or atrial refractory period. Thus, the accuracy of CIEDs is thought to be inferior to that of Holter ECG.

4.1.3 AF

If an atrial lead is implanted, RM can detect AF. Intervention is not needed for AF events in patients in whom AF has been detected and anticoagulant drugs have been administered. However, in patients without a history of AF, we can start anticoagulant drugs early and prevent stroke. There are both symptomatic and asymptomatic patients with paroxysmal AF (PAF). It is sometimes difficult to detect AF in asymptomatic patients with paroxysmal AF. However, patients with CIEDs are continuously monitored and AF is likely to be detected [22].

Some studies have examined the interaction of AF events detected by CIEDs and stroke. Capucci et al. [23] reported that patients with longstanding AF >24 h had a 3-times higher risk of stroke than did control patients. Boriani et al. [24] reported that implementation of CIED data on AF presence/duration/burden can contribute to improved clinical risk stratification. Glotzer et al. [25] reported that thromboembolic event (TE) risk is a quantitative function of atrial tachycardia (AT)/AF burden and that AT/AF burden >5.5 h on any of 30 prior days appeared to double TE risk. The TREND study was a prospective, observational study for patients with >1 stroke risk factor (HF, hypertension, age >65 years, diabetes, and prior TE) who received implantation of pacemakers or defibrillators that monitor AT/AF burden (defined as the longest total AT/AF duration on any given day during the prior 30-day period). AT/AF burden subsets were defined as zero, low (<5.5 h [median duration of subsets with nonzero burden]), and high (≥5.5 h). Adjusted hazard ratios (95% confidence intervals [CIs]) in the low and high burden subsets were 0.98 (0.34–2.82, P=0.97) and 2.20 (0.96–5.05, P=0.06), respectively. Healey et al. [26] reported that subclinical atrial tachyarrhythmias, without clinical AF, occurred frequently in patients with pacemakers and were associated with a significantly increased risk of ischemic stroke or systemic embolism. Subclinical atrial tachyarrhythmias were defined as episodes of atrial rate >190 beats per minute for >6 min. In total, 2580 patients were enrolled in the ASSERT study. The patients were aged ≥65 years with hypertension and no history of AF and had recently undergone pacemaker or defibrillator implantation. Subclinical atrial tachyarrhythmias detected with implanted CIEDs occurred within 3 months in 261 patients (10.1%). Subclinical atrial tachyarrhythmias were associated with an increased risk of clinical AF (hazard ratio, 5.56; P<0.001) and ischemic stroke or systemic embolism (hazard ratio, 2.49; P=0.007). However, it is unknown whether early intervention for patients with CIED-based AF has a clinical benefit.

Attention must also be given to AF in patients with CRT. In some patients with CRT, if AF occurs, intrinsic conduction becomes manifest and the pacing rate of CRT decreases (Fig. 5). AF may induce decompensated HF. Santini et al. [27] reported that CIED-detected AT/AF is associated with worse prognoses and that continuous CIED diagnostic monitoring and Web-based alerts may inform the physician of AT/AF occurrence and identify patients at risk for cardiac deterioration or patients with suboptimal rate or rhythm control.

Fig. 5. Cardiac Compass in a patient with cardiac resynchronization therapy with a defibrillator. Atrial fibrillation resulted in decreases in biventricular pacing rate and intrathoracic impedance.

4.1.4 Non-sustained VT

Non-sustained VT (NSVT) events are frequently detected, and detailed data such as intracardiac electrogram, AA interval, and VV interval can be obtained (Fig. 6). However, although CIEDs can detect NSVT, it may in fact be supraventricular tachyarrhythmia related to an atrial blanking or atrial refractory period. It is easy to detect and diagnose NSVT in patients with an ICD or CRT-D, but it is sometimes difficult to accurately diagnose NSVT in patients with pacemakers because an atrial marker may not appear because of blanking and refractoriness.

Fig. 6. Remote monitoring data of non-sustained ventricular tachycardia in a patient with an implantable cardioverter defibrillator. (A) A-A and V-V intervals during non-sustained ventricular tachycardia (NSVT). (B) Intracardiac electrogram during NSVT.

Intervention for NSVT detected by CIED depends on the patient‘s characteristics. In patients with ICDs or CRT-Ds, special management does not seem necessary. NSVT events also occur in patients with pacemakers. Faber et al. [28] reported that many pacemaker patients present with VT and that intracardiac ECGs and alert functions from pacemakers may enhance physicians' awareness of patients' intrinsic arrhythmic profiles. Although NSVT is detected, special care may not be needed for patients without structural heart disease or symptoms because of their good prognoses [29]. However, in patients with structural heart disease, NSVT is a risk factor for sudden cardiac death [30,31], and special care is needed. The findings in those studies were based on Holter ECG results, and it is unknown whether arrhythmic events detected with CIEDs are a risk. Further studies are needed.

4.1.5 VT/VF

Physicians are notified of the occurrence of VT/VF by e-mail or an alert during RM, and they can obtain detailed information about the VT/VF, including intracardiac electrogram, AA interval, VV interval, and AV interval data (Fig. 7). A decision can then be made as to whether the event is appropriate or inappropriate for therapy (Fig. 8) and how to manage the arrhythmic event in patients at home. Early management of an ICD storm may be possible. However, patients with ICD storms are known to have poor prognoses [32].

Fig. 7. Emergency event in a patient followed by remote monitoring. (A) An emergency e-mail showing that the patient received an ineffective maximum energy shock. (B) Summary and intracardiac electrogram in ventricular tachycardia on the website.

Fig. 8. T wave oversensing detected by remote monitoring. Intracardiac electrogram on the website showed T wave oversensing.

4.2 CIED malfunction

Although CIEDs are effective in patients with bradyarrhythmia or tachyarrhythmia, they sometimes fail to operate properly. CIED abnormality results in life-threatening events [33]. Such events can be avoided by early detection of a CIED abnormality.

4.2.1 Battery abnormality

Batteries usually run out gradually but occasionally run out suddenly. These events are usually asymptomatic. Some CIEDs have a beeping sound to alert for a serious abnormality, but others do not have such a function. Even if the CIED has this function, elderly patients may be unaware of the beeping sound. Acute battery depletion can result in serious events in patients totally depending on CIEDs. In all companies, battery abnormality is set to a red alert, and it is easy to notice on the website (Fig. 9). We have experienced CIEDs in which battery abnormality is likely to occur. If a battery abnormality is detected, immediate action is necessary.

Fig. 9. Elective replacement indicator event detected by remote monitoring. Battery depletion was easily noticed on the website.

4.2.2 Lead fracture

Lead abnormality is also usually asymptomatic, and patients are unaware of it until adverse events have occurred following lead abnormality. Pacing failure results in life-threatening events in patients without a native beat, and ICD lead abnormality results in a short circuit [34,35] or inappropriate ICD shock, both of which are followed by life-threatening adverse events. In one report, ICD shock resulted in poor prognoses [36]. Various studies have shown that RM can detect lead abnormality earlier than conventional follow-up [4,5,7,37]. Various methods can detect lead abnormality. Recently available CIEDs have a function for early detection of lead abnormality. A red or yellow alert indicates abnormalities of lead impedance, pacing threshold, and R wave sensing. Some noise sensing can be detected as high ventricular or high atrial episodes. In the initial stage of lead abnormality, lead impedance is sometimes not changed. It is therefore important to analyze an intracardiac electrogram of NSVT or AF to detect lead abnormality (Fig. 10).

Fig. 10. Lead abnormality detected by remote monitoring. (A) Intracardiac electrogram on the website showed that a ventricular fibrillation event was induced by noise. (B) Gradual lead impedance elevation. Emergency e-mail was received. (C) Intracardiac electrogram on the website showed that a non-sustained ventricular tachycardia event was induced by noise. (D) Sudden elevation of pacing threshold in the right ventricular lead.

4.2.3 Electromagnetic interference

Electromagnetic interference (EMI) influences CIEDs. In conventional follow-up, EMI is detected by abnormal CIED therapy such as inappropriate ICD therapy or oversensing and by a beeping sound. However, in RM, EMI is detected by an emergency e-mail of VF occurrence, and RM detects EMI much earlier than conventional follow-up (Fig. 11). EMI causes resetting of the CIED setting, which is also detected early by RM.

Fig. 11. Electromagnetic interference detected by remote monitoring. Emergency e-mail of lead integrity alert was received. Intracardiac electrogram on the website showed that the lead integrity alert was induced by electromagnetic interference.

4.2.4 Sensing or pacing threshold abnormality

Previously, A wave and R wave amplitude and pacing threshold measurements were performed directly in office visits. However, recently available CIEDs have a function for automatic A wave and R wave amplitude and pacing threshold measurements. RM can obtain these data from patients at home, thus allowing long intervals between office visits. A trend graph of lead information can also be checked on the website, which is useful for early detection of lead abnormality. However, if various situations do not meet the condition, the automatic measurement is canceled. In such cases, it is necessary to check whether the function is operating correctly.

4.3 CIED tuning

Recently available CIEDs have multiple functions to deliver physiological pacing, minimize battery depletion, and ensure safety. Optimal settings such as AV interval, pacing output, and pacing mode settings may be different at implantation and during the follow-up period.

In conventional follow-up, these settings can be detected and adjusted only in office visits. However, RM can obtain accurate information from patients at home. Optimal tuning is particularly important for patients with CRT who have poor prognoses.

4.3.1 Minimizing ventricular pacing

Newer CIEDs except for CRT devices have a function for minimizing the ventricular pacing rate in order to reduce AF occurrence or prevent ventricular dyssynchrony [38,39]. The ventricular pacing rate can be seen on the website. Instead of employing the minimizing ventricular pacing rate function, such as MVP or VIP, the ventricular pacing rate may remain high in some patients. In these patients, it may be reasonable to turn off these functions to save battery power or avoid ventricular pacing in vulnerable periods (Fig. 12).

Fig. 12. Change of ventricular pacing rate. (A) Pacing rate one week after implantation (DDD). Ventricular pacing rate was very low. (B) Pacing rate at follow-up (MVP). Pacing mode was switched from DDD to MVP before discharge. Ventricular pacing rate was increased because atrio-ventricular conduction had deteriorated. We noticed this event by remote monitoring.

4.3.2 Maximizing CRT pacing rate

Decreases in the CRT pacing rate can reduce the effect of CRT, especially in patients with AF [40,41]. Newer CIEDs have a function for maximizing the CRT pacing rate, such as ventricular sense response or sliding AV interval, and RM can detect a low CRT pacing rate or long ventricular sensing events. Although the AV interval or medication can be adjusted to maximize the CRT pacing rate in office visits, it is not clear when the setting should be changed for optimization. However, sufficient adjustment is particularly important for non-responder patients.

4.3.3 Pacing mode optimization

Optimal pacing mode can vary during long follow-up. For example, paroxysmal AF can develop into chronic AF. During paroxysmal AF, the optimal pacing mode is DDD. However, after development of chronic AF, VVI or VVIR may be appropriate. The F wave is likely to be undersensed, which will lead to A pacing with ventricular blanking and ventricular blanking causing ventricular pacing in vulnerable periods, which sometimes causes life-threatening arrhythmia (Fig. 13).

Fig. 13. Ventricular tachycardia induced by ventricular pacing on T wave. (A) Ventricular pacing in a short coupling interval (1) induced ventricular tachycardia (VT) (2). The implantable cardioverter defibrillator terminated the VT (3). (B) Ventricular pacing in short coupling interval was introduced because of an atrial refractory period after premature ventricular beat.

Fine-tuning may be useful in patients with CIEDs, but it is unknown how early various settings should be changed. In the future, tuning may become easy if remote programming technology is developed.

4.4 Physiological data

Newer CIEDs with RM provide various types of information not only about arrhythmia but also about physiological situations, including heart rate variability, activity, atrial pacing rate, ventricular pacing rate, supraventricular arrhythmic events, ventricular arrhythmic events, ventricular rate during AF, apnea events, body weight, blood pressure, ST change, and intrathoracic impedance (ITI). Although it is unclear how to manage patients based on such information, the information can be analyzed in an integrated fashion and the patient's status can be understood (Fig. 14).

Fig. 14. The Cardiac Compass showed a summary of various physiological information.

4.4.1 Heart rate variability

Heart rate variability (HRV) is calculated from the RR interval, which has been associated with autonomic tone. In patients with HF, HRV is likely to be decreased and can predict mortality [42–44]. Results of those studies were based on the RR interval in Holter ECG data. Autonomic tone is related to sinus node, and AA interval may be more accurate than RR interval for assessing autonomic tone. Thus, HRV based on CIED data may be more useful than HRV based on Holter ECG data. Landolina et al. [45] reported that CIED-monitored HRV is useful for identifying, early after implantation, patients with a low likelihood of long-term benefits from CRT and at high risk for cardiovascular events. Adamson et al. [46] reported that HRV continuously measured from an implanted CRT-D was lower in patients at high risk for hospitalization and mortality and may be useful in the clinical management of patients with chronic HF. Their study included 397 patients with New York Heart Association class III or IV HF. Continuous HRV was measured as the standard deviation of 5-min median atrial–atrial intervals (SDAAMs) sensed by CIEDs. SDAAM <50 ms when averaged over 4 weeks was associated with increased mortality risk (hazard ratio: 3.20, P=0.02), and SDAAMs were persistently lower over the entire follow-up period in patients who required hospitalization or in patients who died. HRV has also predicted sudden death in patients with ischemic heart disease but not in patients with non-ischemic heart disease [42,47–49].

4.4.2 Activity

Almost all CIEDs have an acceleration sensor. RM can show activity based on an acceleration sensor. In patients with HF, activity increases with increased HRV [50,51].

4.5 Early detection of decompensated HF

HF is a common cause of hospitalization and represents a considerable economic burden to society [52]. Despite therapeutic advances, most of these events are readmissions for acute deterioration of chronic HF [53]. Approximately 25% of discharged patients are readmitted within 30 days, and HF is associated with a higher rate of hospital readmission than any other medical or surgical cause of hospitalization [54]. The ability to identify patients at risk for hospitalization would be useful. Most HF-related hospitalizations are related to fluid accumulation, and careful fluid status surveillance and symptom monitoring are important.

4.5.1 Telemonitoring

Clinical signs and symptoms of HF usually appear late in the course of decompensation and are largely unreliable in routine follow-up of patients with HF. In some reports, telemonitoring based on symptoms and body weight failed to reduce HF hospitalization [55,56] (Fig. 15). There are various possible reasons why telemonitoring failed to reduce HF hospitalization. One possibility is that intervention based on symptoms or body weight occurs too late, because intracardiac pressure is already elevated at the time of symptom appearance or increased body weight [57]. Intracardiac pressure is increased 10–20 days before HF hospitalization. Another possibility is low adherence rate. Chaudhry et al. [55] found in their TRLR-HF trial that 14% of the patients who were randomly assigned to undergo telemonitoring never used the system and that only 55% of the patients were still using the system at least 3 times per week in the final week of the study period. These findings are important given that considerable resources, which would be difficult to leverage outside a clinical trial, were directed toward optimizing patients' engagement with the system. Thus, a data gathering system by self-action has limitations compared to an automatic data gathering system for obtaining various data.

Fig. 15. Probability of freedom from readmission for any reason or death from any cause. No significant difference was seen between the 2 groups in the rate of readmission for any reason or death from any cause.

4.5.2 Direct cardiac pressure monitoring

Technology for intracardiac pressure monitoring has been developed. One system is left atrial pressure monitoring. Ritzema et al. [58] reported that physician-directed patient self-management of left atrial pressure has the potential to improve hemodynamics, symptoms, and outcomes in patients with advanced HF. Forty patients with HF and acute decompensation were implanted with investigational left atrial pressure monitors. For the first 3 months, patients and clinicians were blinded to the readings, and treatment continued per usual clinical assessment. Thereafter, left atrial pressure and individualized therapy instructions guided by these pressures were disclosed to the patients. Survival without decompensation was 61% at 3 years, and events tended to be less frequent after the first 3 months (hazard ratio: 0.16, P=0.012). Technology for direct pulmonary artery (PA) pressure monitoring has also been developed. Abraham et al. [59] reported that PA pressure monitoring resulted in a significant reduction in hospitalization for HF patients. During the follow-up periods in their study (mean periods of 15 months), the treatment group had a 37% reduction in HF-related hospitalization compared with the control group (158 vs. 254, HR: 0.63; P<0.0001) (Fig. 16). Thus, intracardiac monitoring may be an alternative to telemonitoring for reducing HF hospitalization, but such technology has not been available so far. The most beneficial point is detecting preclinical HF before symptoms appear (Fig. 17). However, an implanted device in the body is overwrapped by connective tissue, which may lead to monitoring failure because the pressure transducer may be overwrapped by connective tissue. The longevity of implantable pressure-monitoring devices is an unresolved issue.

Fig. 16. Cumulative heart failure-related hospitalizations during the entire period of randomized single-blind follow-up. The treatment group had a 37% reduction in heart failure-related hospitalization compared with the control group (158 vs. 254, HR 0.63, 95% CI 0.52–0.77; P[less than]0.0001).

Fig. 17. Reactive and proactive intervention for heart failure. Proactive intervention for heart failure may be useful for preventing heart failure hospitalization.

4.5.3 Intrathoracic impedance

CIEDs can now gather and store various types of data related to HF, such as data on ventricular pacing rate, CRT delivery rate, HRV, and ITI. We have already mentioned ventricular pacing rate, CRT delivery rate, and HRV. Here, we will focus on ITI.

Recent studies have suggested that ITI may be a useful parameter to track day-to-day changes in pulmonary fluid status [60,61]. ITI was measured over time by an implanted CIED and was correlated inversely with changes in left ventricular (LV) end diastolic pressure in a canine HF model [60] as well as with changes in pulmonary capillary wedge pressure (PCWP) [61] and right ventricular pressure [62] (Fig. 18). Additionally, ITI was negatively correlated with NT-proBNP [63]. OptiVol alert, which is a fluid status algorithm calculated from ITI, can detect impending fluid accumulation at an early stage [61]. HF hospitalization was reduced for patients with OptiVol alert. Abraham et al. [64] reported that the sensitivity and unexplained detection rate of ITI monitoring were superior to those seen for acute weight changes. ITI monitoring represents a useful adjunctive clinical tool for managing HF in patients with implanted CIEDs. Catanzariti et al. [65] reported that the alert capability seemed to reduce the number of HF hospitalizations by allowing timely detection and therapeutic intervention. In their study, 532 HF patients with implanted ICDs with an OptiVol feature were followed for 11±7 months. The audible alert was on in 430 patients (ON group) and disabled in 102 patients (OFF group). HF hospitalizations were required for 29 patients (7%) in the ON group and for 20 patients (20%) in the OFF group (P<0.001). The rate of combined cardiac death and HF hospitalization was lower in patients with Alert ON (log-rank test, P=0.007). However, another study suggested that the OptiVol alert could not reduce HF hospitalization [66]. van Veldhuisen et al. reported that using an implantable diagnostic tool to measure ITI with an audible patient alert did not improve outcomes and increased HF hospitalizations and outpatient visits in HF patients. They studied 335 patients with chronic HF who had undergone implantation of ICDs alone (18%) or CRT-Ds (82%). Patients were randomized to have information available to physicians and patients as an audible alert in case of preset threshold crossings (access arm) or not (control arm). The primary end point was a composite of all-cause mortality and HF hospitalizations. During a mean period of 14.9±5.4 months, the primary end point occurred in 48 patients (29%) in the access arm and in 33 patients (20%) in the control arm (HR, 1.52; P=0.063). This was related mainly to more HF hospitalizations (HR, 1.79; P=0.022), whereas the numbers of deaths were comparable (19 vs. 15; P=0.54). Sensitivity and positive predictive value of OptiVol alert were not sufficient to be reliable predictive parameters of HF [67], [68]. In another report, the combination of CIED diagnostics and OptiVol alert was necessary to increase the positive predictive value [69]. Whellan et al. [70] reported that combined HF CIED diagnostics identified patients at higher risk of subsequent HF hospitalization [70]. PARTNERS HF (Program to Access and Review Trending Information and Evaluate Correlation to Symptoms in Patients With HF) was a prospective, multicenter, observational study in patients receiving CRT-Ds. Data from 694 CRT-D patients who were followed for 11.7±2 months were analyzed. Ninety patients had 141 adjudicated HF hospitalizations with pulmonary congestion at least 60 days after implantation. Patients with positive combined HF CIED diagnostics had a 5.5-fold increased risk of HF hospitalization with pulmonary signs or symptoms within the next month (hazard ratio: 5.5, 95% CI: 3.4–8.8, P<0.0001). OptiVol alert combined with other parameters is also better for predicting HF events than only OptiVol alert. A meta-analysis showed that RM in patients with HF can reduce mortality and HF hospitalization [20], as previously mentioned.

Fig. 18. Relationship between intracardiac pressure and intrathoracic impedance. (A) Example from one patient. Relationships between intrathoracic impedance, pulmonary capillary wedge pressure, and net fluid loss (I/O) during 4 days of intensive diuresis in the CCU. (B) Correlation between daily medians of ePAD and intrathoracic impedance. Examples are given for 2 patients with major HF events. Correlations are shown during the whole follow-up period (r=−0.31 and −0.51) (left) and within the 1-month period before a major HF event (r=−0.77 and −0.81) (right).

5 Important factors at the time of informed consent

Patients benefit greatly from RM, and they feel secure and safe because they feel connected to the hospital at all times. However, patients may have excessive expectations for RM. It is impossible to respond to all alerts immediately for the following reasons. One reason is medical staff. For medical staff, the workload of office visits can be reduced, but other workloads are increased. It is unclear whether the total workload decreases by introducing RM. It is thus difficult to respond to all alerts. To respond immediately to all alerts, dedicated staff is necessary for 24 h a day because alert e-mails are often received at night. We do not currently receive sufficient reimbursement for emergency events, and it is difficult to employ dedicated staff for alert e-mails. However, patients may have the misunderstanding that medical staff can monitor their RM data 24 h a day. Another reason is technological problems. Even if the hospital has dedicated medical staff for emergency events, some alerts are not sent immediately. A third reason is occasional connection failures between the CIED and the transmitter or between the transmitter and server. Thus, we have to inform patients that RM is not for emergency events and that they must visit the hospital if they have any symptoms.

It is also necessary to inform patients about security. RM data are sent though the Internet and stored in a server. These digital data are protected firmly, but the risk of hacking remains.

6 An RM management method example

Various workloads increase for managing RM. We have to obtain informed consent, instruct on RM transmitter use, enroll patients on a website, check the website for data delivery, call the patients if data have not been delivered, download and analyze the RM data, and inform the patients about analysis results. Thus, the workload cannot be managed completely by doctors; other staff members, including medical engineers, nurses and secretaries, must also help. A team was established in 2008 in Okayama University Hospital for managing RM. The RM team includes doctors, medical engineers, nurses, and secretaries. There are 2 major goals for the RM team. One is to manage RM with various staff to avoid focusing the workload on one person, and the other is to introduce this new technology to our institute as well as other associated hospitals.

RM team members are assigned different tasks. Mainly, doctors obtain informed consent, nurses provide instructions on RM transmitter use, medical engineers enroll the patients on the website and perform the primary RM data analysis, and doctors perform the secondary analysis. We have made one report for one analysis. For patients in associated hospitals, we have mainly helped to analyze RM data. Analysis reports are sent by e-mail or fax to associated hospitals[71]. By December 2013, we had followed more than 1000 patients with approximately 70 associated hospitals and made more than 8000 RM reports. Only ~5% of the data required intervention by medical staff. Thus, if we focus on only ~5% of the events, we can dramatically reduce the workload associated with CIED follow-up for both patients and medical staff and may also reduce the cost for CIED follow-up.

7 Reimbursement considerations

RM benefits not only patients with CIEDs but also medical staff. However, reimbursement for RM is insufficient for the workload of the medical staff. We have been able to claim reimbursement for RM since 2010 in Japan. We can now claim 5500 yen/4 months for office visits and 3600 yen for emergency cases. However, the reimbursement has been limited to office visits. This seems to be one reason why RM has not expanded despite the great benefits for patients with CIEDs. Matsumoto reported that the reimbursement fee should be higher in consideration of the workload of the medical staff [72]. The problem may be resolved in the future. Insufficient reimbursement has also been reported in other countries [73]. One main reason is thought to be a relative paucity of research data associated with prognosis. Although various evidence of RM has been reported, it is unknown whether RM can improve prognoses in patients with CIEDs.

8 Unresolved problems with remote monitoring

Some unresolved problems with RM remain: CIED data might not be secure from hackers accessing a server on the Internet. In addition, a useful database system is not available in Japan. It is time-consuming to enter each dedicated ID and password to access the server in each company. Moreover, to analyze RM data, we have to create our own database, which usually requires manual typing. It is a main source of workload for medical staff. Further, it is unclear whether RM can improve prognosis and whether cost-effectiveness is acceptable for clinical use. Finally, reimbursement is insufficient to manage RM in various institutes.

If these problems are resolved in the future, RM may be more accessible for both patients and medical staff.

Conflict of interest

None declared.