Appendix

Appendix

Contents

Data tables

Table 2: Overview of the Al-Khatib et al. (2010) study

Table 3: Summary of results of the Al-Khatib et al. (2010) study

Table 4: Overview of the Boriani et al. (2013) study

Table 5: Summary of results of the Boriani et al. (2013) study

Table 6: Overview of the Crossley et al. (2009) study

Table 7: Summary of results of the Crossley et al. (2009) study

Table 8: Overview of the Crossley et al. (2011) study

Table 9: Summary of results of the Crossley et al. (2011) study

Table 10: Overview of the de Ruvo et al. (2015) study

Table 11: Summary of results of the de Ruvo et al. (2015) study

Table 12: Overview of the Landolina et al. (2012) study

Table 13: Summary of results of the Landolina et al. (2012) study

Table 14: Overview of the Luthje et al. (2015) study

Table 15: Summary of results of the Luthje et al. (2015) study

Table 2 Overview of the Al-Khatib et al. (2010) study

Study component

Description

Objectives/hypotheses

To determine if remote monitoring (using CareLink) of ICDs with or without CRT compared with quarterly in-clinic device interrogations improves patient outcomes and satisfaction with their ICD care.

Study design

Prospective, single-centre, randomised study.

Setting

Device clinics at a single US medical centre. Patients were enrolled between December 2006 and November 2007.

Intervention and comparator

Patients were randomly assigned in equal proportions to have either quarterly in-clinic ICD interrogations, classified as standard of care, or remote monitoring of ICDs using the CareLink transmission monitor.

Data on QoL (measured with the EQ‑5D), patient satisfaction with ICD care, cardiac problems, ICD-related issues, and medications was collected at baseline, 6 months (by telephone for the intervention group) and 12 months after enrolment.

Intervention

Patients were advised to keep a log of dates and reasons for hospital admissions, and emergency room and EP clinic visits. They were asked to use the remote monitoring system every 3 months, and they were seen in the device clinic at 12 months and at any time for device-related issues. Device programming was at the discretion of the treating physician. None of the heart failure capabilities of the devices (for example, impedance, heart rate variability) were used in managing patients' conditions.

Control

Patients randomised to the control arm were seen in the ICD clinic every 3 months and at any time that their ICD physician decided to see them for a device-related issue.

Inclusion/exclusion criteria

Inclusion

Patients were eligible if they were 18 years and over, had an ICD with or without CRT for an approved indication, were planning to have their device followed-up at the medical centre, had a landline telephone, and were able to provide informed consent.

Exclusion

None stated.

Primary outcomes

Primary end point was a composite of cardiovascular hospitalisation, emergency room visits for a cardiac cause, and unscheduled visits to the EP clinic for device-related issues at 1 year.

Secondary end points included use of evidence-based medications, health-related QoL, cost, cost-effectiveness, and patient satisfaction with ICD care.

Statistical methods

Wilcoxon rank sum test to compare continuous variables and Chi-square test to compare categorical variables. Cumulative event rates were calculated using Kaplan–Meier and outcomes in the 2 arms of the study were compared using the log-rank test and intention-to-treat principle.

Patients included

n=151 (76 CareLink; 75 control)

73% male (CareLink); 72% male (control)

Median age (years): 63 (CareLink); 63 (control)

One patient was lost to follow‑up and 4 withdrew (1 for lack of transport to clinic, 1 for a language barrier, and 2 moved to a nursing home), 7 patients died.

69 patients completed monitoring in the remote arm, and 70 in the standard care arm.

Results

There was no significant difference in the composite of cardiovascular hospitalisation, emergency room visits for a cardiac cause, and unscheduled visits to the EP clinic for device-related issues at 1 year (32% in the remote arm compared with 34% in the control arm; p=0.8), mortality, or cost between the 2 arms. QoL and patient satisfaction were significantly better in the control arm than in the remote arm at 6 months: 83 compared with 75 (p=0.002) and 88 compared with 75 (p=0.03) respectively, but not at 12 months.

Conclusions

There were no significant differences in cardiac-related resource utilisation at 1 year. But, given the small number of patients in this study, the real clinical and health economics impact of remote monitoring needs to be verified by a large, multicentre, randomised controlled trial.

Abbreviations: CRT, cardiac resynchronisation therapy; EP, electrophysiology; ICD, implantable cardioverter defibrillator; QoL, quality of life.

Table 3 Summary of results from the Al-Khatib et al. (2010) study

CareLink remote monitoring (n=76)

Face-to-face follow‑up (n=75)

Analysis

Primary outcomes

Rate of composite hospitalisations, cardiac-related emergency room visits and device-related unscheduled visits to the EP clinic

32%

34%

p=0.77

Number of hospitalisations

23%

24%

p=0.88

Most (66%) were for decompensated heart failure; 4 were for ICD shocks; 3 for right ventricular lead fracture; and 1 for generator replacement.

Number of cardiac-related emergency room visits

7%

5%

p=0.74

Number of device-related unscheduled visits to the EP clinic for issues

7%

7%

p=0.98

Selected secondary outcomes

Rate of atrial fibrillation and flutter detected by the ICD during follow‑up

45%

26%

p=0.01

No significant differences in these events at baseline.

Health-related quality of life (using EuroQoL thermometer)

75 (at 6 months)

80 (at 12 months)

83 (at 6 months)

80 (at 12 months)

p=0.002

p=0.47

None of the baseline QoL measures were significantly different between the 2 arms.

There were no significant differences in the EuroQoL score at 6 or 12 months.

Patient satisfaction with ICD care

75 (at 6 months)

88 (at 12 months)

88 (at 6 months)

88 (at 12 months)

p=0.03

p=0.09

Patient satisfaction with their ICD care at baseline was similar between the 2 arms.

Cost-minimisation analysis

Patients' devices were interrogated remotely 3 times, each costing $102.79 (total $308.37) and once in clinic costing $66.36 or $89.92 (if the need for ICD programming is included). Total costs $374.73–$398.29.

Patients had 4 clinic visits each costing $66.36 or $89.92 (if the need for ICD programming is included). Total costs $265.44–$359.68.

The remote monitoring strategy was more expensive than current standard care (difference $38.61–$109.29). Given the age of this population (mean 63 years), the analysis did not include costs for lost time from work for patients' visits to the clinic. Travel-related expenses could further decrease the difference between the 2 arms.

Deaths

4 (5%)

3 (4%)

p=0.99

Cause of death was non‑cardiac in 5 patients and unknown in 2. There were no device infections.

Abbreviations: EP, electrophysiology clinic; ICD, implantable cardioverter defibrillator; QoL, quality of life.

Table 4 Overview of the Boriani et al. (2013) study

Study component

Description

Objectives/hypotheses

To evaluate if remote monitoring can reduce time from device-detected events to clinical decisions.

Study design

Multicentre, randomised controlled trial.

Setting

32 centres from 6 countries (France, Hungary, Israel, Italy, Spain and Switzerland). Patients enrolled between May 2009 and April 2010 with a median follow‑up of 12 months.

Intervention and comparator

Intervention: RM strategy using CareLink network service. Patients had face-to-face follow‑ups at baseline and at 8 months, with remote follow‑ups at 4 and 12 months, and activation of automatic alerts.

Control: Standard management by scheduled face-to-face follow‑ups at baseline and at every 4 months.

Audible alerts for device integrity issues or for inactivated VF detection or therapy were activated in both groups.

Inclusion/exclusion criteria

Inclusion

Patients in sinus rhythm with first implantation of CRT‑D for systolic heart failure. Left ventricular systolic dysfunction (LVEF ≤35%), NYHA functional class III–IV.

Exclusion

Exclusion criteria were reported in Burri et al. (2010) as part of a trial design paper. Patients were excluded if they were not able to fully understand the instructions on remote monitoring using CareLink, had permanent AT/AF, had a CRT/CRT‑D device implanted before, had medical conditions that would limit study participation, were <18 years, were enrolled in or intended to participate in another clinical trial that may have an impact on the study end points, met any exclusion criteria required by local law, were unable or refused to sign a patient informed consent form, had a life expectancy of <1 year in the opinion of the physician, and if they were pregnant or breastfeeding.

Primary outcomes

The delay between event onset to clinical action relating to that event.

Secondary outcomes included: time from a clinical decision for any relevant event to the resolution of that event, QoL, in-hospital visits, automatic alert transmission and annual rate of all-cause hospitalisations.

Statistical methods

Continuous Gaussian variables were compared by the Student's t test for independent samples, whereas skewed distributions were compared using the Mann–Whitney nonparametric test. Differences in proportions were compared by applying Chi-square analysis. Rates of events were computed per 100 person years, as number of occurred events out of patient exposure time and reported separately for each arm. The exposure time was computed from the date of randomisation to the date of the last available information for each patient, either dropped out or died. Rates were compared using the Comparison Incidence Rates (Large Sample) Test. An alpha-level of 0.5 was used.

Patients included

n=148 (76 RM and 72 control).

Average age 68 years (RM), 67 years (control).

75% male (RM), 72.4% male (control).

Results

The median delay from device-detected events to clinical decisions was considerably shorter in the RM group compared with the control group: 2 (25th–75th percentile, 1–4) days compared with 29 (25th–75th percentile, 3–51) days respectively, p=0.004. In-hospital visits were reduced in the remote group (2.0 visits/patient/year compared with 3.2 visits/patient/year in the control group, 37.5% relative reduction, p<0.001). Automatic alerts were successfully transmitted in 93% of events happening outside the hospital in the RM group. The annual rate of all-cause hospitalisations per patient did not differ between the two groups (p=0.65).

Conclusions

RM in CRT‑D patients with advanced heart failure allows physicians to promptly react to clinically relevant automatic alerts and significantly reduces the burden of in-hospital visits.

Abbreviations: AF, atrial fibrillation; CRT‑D, cardiac resynchronisation therapy defibrillators; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; QoL, quality of life; RM, remote monitoring; VF, ventricular fibrillation.

Table 5 Summary of results from the Boriani et al. (2013) study

CareLink remote monitoring (76)

Face to face (n=72)

Analysis

Primary outcome

Delay between event onset to clinical action

Alerts (n): 166

Median delay (25th–75th percentile) between alert triggering to event review (days): 3 (1–10)

Events matching alert criteria (n): 114

Median delay (25th–75th percentile) between alert triggering to event review (days): 37 (14–71)

p<0.001

Selected secondary outcomes

Time from actionable device-detected event to clinical decisions

Device detected events: 37

Median time from event onset to related clinical decisions (days): 2

Device-detected events: 19

Median time from event onset to related clinical decisions (days): 29

p=0.004

In-hospital visits (scheduled, unscheduled and emergency visits)

144 (2 visits/year)

225 (3.2 visits/year)

p<0.001

Hospital admissions

19

22

p=0.65

QoL

Baseline score (25th–75th percentile): 41 (16 to 62)

Change in score from baseline to 8‑month follow‑up (25th to 75th percentile): −17 (−32 to −2)

Baseline score (25th–75th percentile): 40 (18 to 53)

Change in score from baseline to 8‑month follow‑up (25th–75th percentile): −10 (−23 to 0)

p=0.38

p=0.45

Change in clinical status from enrolment to 12‑month follow‑up

54% improved, 35%unchanged and 11% worsened

48% improved, 38% unchanged and 14% worsened

p=0.69

Deaths

n=5

n=2

Successful alert transmission

144/155 (93%) of events (excluding alert transmissions that failed due to hospital admission)

NA

Device-detected events

Patients who had at least 1 event satisfying the criteria for triggering a device alert: 57 (75%)

Observed rate of OptiVol (events/year): 1.6

Rate of AT/AF burden and fast ventricular rate during AF episodes (events/year): 0.7

Patients who had at least 1 event satisfying the criteria for triggering a device alert: 48 (67%)

Observed rate of OptiVol events/year: 1.5

Rate of AT/AF burden and fast ventricular rate during AF episodes (events/year): 0.2

p=0.28

p=0.59

p<0.001

Abbreviations: AF, atrial fibrillation; AT, atrial tachyarrhythmia; NA, not applicable; QoL, quality of life.

Table 6 Overview of the Crossley et al. (2009) study

Study component

Description

Objectives/hypotheses

To evaluate remote pacemaker interrogation for the earlier diagnosis of clinically actionable events compared with traditional transtelephonic monitoring and routine in-person evaluation.

Study design

Prospective, randomised, parallel, unblinded, multicentre, open-label clinical trial.

Setting

Study enrolment was from May 2004 to March 2007, in 50 US centres.

Intervention and comparator

Patients were randomized in a 2:1 manner to the RM arm or the control arm using permuted block randomisation.

Intervention

Patient data in the RM arm were sent using the CareLink network service at 3, 6, and 9 months.

Comparator

The control arm patients did a TTM transmission at 2, 4, 8, and 10 months. At 6 months, patients with dual-chamber pacemakers were seen in person, and a TTM transmission was done by patients with single-chamber pacemakers.

All patients: Pacemaker programming was at the discretion of the responsible physician except for 3 parameters. The study ended with a face-to-face follow‑up visit at 12 months. Unscheduled transmissions and in-person evaluations were included in the analysis.

Inclusion/exclusion criteria

Inclusion criteria

Patients had to have a pacemaker compatible with the Medtronic CareLink remote monitoring service and were enrolled after the implantable pulse generator system was deemed stable. Patients with both single- and dual-chamber pacemakers were enrolled, at least 30 days after system modification, including new device implant, device upgrade, or lead changes. Patient had to have access to an analogue phone line, and be able to operate the TTM monitor and the CareLink Monitor.

Exclusion criteria

Exclusion criteria were reported in Chen et al. (2008) as part of a trial design paper. Patients were excluded if they were enrolled in another pacemaker clinical study that might confound the results of this trial, and if they were being considered for an ICD.

Primary outcomes

Time-to-first diagnosis of a CAE (patients without a CAE were censored at the exit date due to death, lost to follow‑up, or study closure).

CAEs were defined as events that needed a clinical decision for possible change of medication or further medical assessment.

Statistical methods

The Peto and Peto modification of the Gehan–Wilcoxon test was done. An intent-to-treat analysis was performed. Only events diagnosed by the clinician counted toward the primary objective. A p value <0.05 indicated that the freedom from first diagnosis of CAE was significantly different when patients were followed with remote interrogation (RM) compared with those being followed with TTM and having scheduled face-to-face follow‑up (control).

Patients included

n=897 (RM 602; control 295)

52% male (RM); 48% control

Mean age (years): RM 68; control 69

382 patients with at least 1 CAE (RM 271;111 control) included in primary analysis.

Results

The mean time to first diagnosis of CAEs was earlier in the RM arm (5.7 months) than in the control arm (7.7 months). Three (2%) of the 190 events in the control arm and 446 (66%) of 676 events in the RM arm were identified remotely.

Conclusions

Remote pacemaker interrogation follow‑up using the Medtronic CareLink network service detects CAE events that are potentially important more quickly and more frequently than transtelephonic rhythm strip recordings.

Abbreviations: CAE, clinically actionable events; ICD, implantable cardioverter defibrillator; RM, remote monitoring; TTM, transtelephonic monitoring.

Table 7 Summary of results from the Crossley et al. (2009) study

CareLink) remote monitoring (n=602)

Transtelephonic monitoring (n=295)

Analysis

Efficacy

271 (45%) patients had ≥1 CAE

111 (38%) patients had ≥1 CAE

Primary outcome

Time to first CAE

Mean 5.7 months

Median 4.9 months

446 (66%) of 676 CAEs were detected during remote 'follow‑up'.

Mean 7.7 months

Median 6.3 months

3 (2%) of 190 CAEs were detected during a TTM transmission, all others were found during face-to-face follow‑up evaluations.

Average follow‑up of 375±140 days.

Significant difference in median time to first CAE between groups; p<0.0001.

Selected secondary outcomes

Number of CAEs reported per patient

1.123

0.644

The most frequent CAE reported was non-sustained VT, followed by AT/AF episodes lasting 48 hours or more.

Safety

Not reported

Serious adverse events

Not reported

Abbreviations: AT/AF, atrial tachycardia/atrial fibrillation; CAE, clinically actionable events; TTM, transtelephonic monitoring; VT, ventricular tachycardia.

Table 8 Overview of the Crossley et al. (2011) study

Study component

Description

Objectives/hypotheses

To determine if wireless remote monitoring with automatic clinician alerts (CareLink) reduces the time from a clinical event to a clinical decision in response to arrhythmias, cardiovascular disease progression, and device issues compared with patients having standard face-to-face follow‑up care. A secondary objective was to compare the rates of cardiovascular health care use in patients in the remote arm with those in the face-to-face follow‑up arm.

Study design

Multicentre, prospective, randomised evaluation.

Setting

Patients were enrolled from November 2006 to May 2008. The last follow‑up visit was in August 2009. The study was done in 136 US centres.

Intervention and comparator

CareLink programming

To limit the number of device transmissions sent in the remote arm, a conservative approach was used to select alert thresholds. Only values needing clinician attention and possible intervention were specified. Exactly 1 automatic clinician alert could be sent for any 1 clinical event between face-to-face follow‑up device interrogations. Clinicians had access to the entire set of device-collected diagnostics for all study patients.

Intervention

Patients in the remote arm had a home monitor, and their face-to-face follow‑ups at 3, 6, 9, and 12 months were replaced with remote visits, including a remote device transmission. All automatic clinician alerts were enabled for patients in the remote arm. Audible patient alerts were disabled except for those related to lead and device integrity. These patients also had face-to-face follow‑ups at 1 and 15 months.

Control

Patients in the control had face-to-face follow‑ups at 3, 6, 9 and 12 months. Only audible patient alerts associated with lead and device integrity were enabled.

Inclusion/exclusion criteria

Adult patients with an implanted Medtronic wireless ICD or CRT‑D system using the CareLink Network. After successful insertion of an ICD or CRT‑D, patients were randomly assigned in a 1:1 manner, stratified by device type, to wireless remote monitoring or face-to-face follow‑up care.

Inclusion criteria

Being able and willing to replace regularly scheduled face-to-face follow‑ups with remote follow‑ups; and being able to attend all required follow‑up visits.

Exclusion criteria

Patients were excluded for: permanent AF (constant AF for which there were no plans to try to restore sinus rhythm); chronic warfarin therapy; having had a previous ICD, CRT device, or pacemaker; under 18 years; and having a life expectancy <15 months.

Primary outcomes

The primary outcome, time to clinical decision, was defined as the time from device detection of a clinical event to a decision being made in response to the event, as reported by the clinician or as shown by device data obtained at interrogation. The key secondary objective was to compare cardiovascular HCU rates.

Statistical methods

A Wilcoxon rank-sum test was used to compare the median time from event onset to clinical decision between treatment arms. To allow for multiple HCU events per patient, an Andersen–Gill proportional hazards regression model was used to compare the hazard rates for each type of HCU event (hospitalisation, ED, unscheduled office or urgent visit) between arms.

Patients included

n=1997 (1014 remote; 983 face-to-face)

70.5% male (remote); 71.7% male (face-to-face)

Mean age (years): 65.2 (remote); 64.9 (face-to-face).

1980 patients were included in analysis (1,005 remote; 975 face-to-face).

Results

The median time from clinical event to clinical decision per patient was reduced from 22 days in the face-to-face arm to 4.6 days in the remote arm (p<0.001). The HCU data showed a decrease in mean length of stay per CV hospitalisation visit from 4.0 days in the face-to-face arm to 3.3 days in the remote arm (p=0.002).

Conclusions

Wireless remote monitoring with automatic clinician alerts compared with standard face-to-face follow‑up significantly reduced the time to a clinical decision in response to clinical events and was associated with a significant reduction in mean length of CV-related hospital stay.

Abbreviations: AF, atrial fibrillation; CRT, cardiac resynchronisation therapy; CRT‑D, cardiac resynchronisation therapy with defibrillation; CV, cardiovascular; ED, emergency department; HCU, health care utilisation; ICD, implantable cardioverter-defibrillators; LOS, length of stay; NA, not applicable.

Table 9 Summary of results from the Crossley et al. (2011) study

CareLink remote monitoring (n=1014)

Face-to-face standard care (n=983)

Analysis

Efficacy

172 (17%) patients had ≥1 CE

145 (15%) patients had ≥1 CE

Primary outcome

Median time from an event to clinical decision

4.6 days

22 days

Reduction of 17.4 days (79%)

A sensitivity analysis including multiple events of the same type between an event onset and a device interrogation/visit also showed a significant reduction from the time an event occurs to a clinical decision (p<0.001).

Selected secondary outcomes

Automatic clinician alert transmissions

329 of 575 clinical events were triggered by an automatic clinician alert.

NA

246 clinical events did not trigger an automatic clinician alert because the alert was programmed off (7%) or the alert was not reset after being previously triggered (93%). Automatic clinician alerts were triggered but not successfully transmitted for 149 (45%) clinical events, mainly because the home monitor was not set up to send out transmissions.

Clinicians classified automatic clinician alerts (140 events) as: 'meaningful' (62%); 'timed appropriately' (84%); 'it could have waited longer' (12%); 'didn't want to know at all' (2%).

Mean LOS during CV hospitalisation

3.3 days

4 days

The mean LOS during a CV hospitalisation was significantly reduced (18%, p=0.002) in the remote arm.

Annualised rate of CV HCU visits per patient

Hospitalisation: 0.50

ED: 0.24

Unscheduled clinic visit: 2.24

Hospitalisation: 0.47

ED: 0.21

Unscheduled clinic visit: 1.95

Hospitalisation: p=0.524

ED: p=0.325

Unscheduled clinic visit: p=0.099

Mean LOS per hospitalisation for patients with a clinical event during follow‑up

3.2 days

4.3 days

p=0.007

Mean LOS per hospitalisation for patients without a clinical event during follow‑up

3.3 days

3.9 days

Not significantly different

Mortality

Not significantly different for patients with an ICD (p=0.31) or patients with a CRT‑D (p= 0.46).

Abbreviations: AF, atrial fibrillation; CE, clinical event; CI, confidence interval; CRT‑D, cardiac resynchronisation therapy with defibrillation; CV, cardiovascular; ED, emergency department; HCU, health care utilisation; ICD, implantable cardioverter-defibrillators; LOS, length of stay.

Table 10 Overview of the de Ruvo et al. (2015) study

Study component

Description

Objectives/hypotheses

To estimate and compare the event-free rate at 1 year in 4 remote monitoring (RM) systems, and investigate the effect of periodicity of RM transmissions on early detection of clinical- and device-related events.

Study design

Prospective, single centre, non-randomised study.

Setting

A single Italian medical institution. Patients were enrolled between January 2009 and January 2011.

Intervention and comparator

Intervention/comparator

21 patients with an ICD were monitored with 4 different HM devices: CareLink, BHM, LAT and SJM.

Remote follow‑ups were configured quarterly, except for the BHM (daily transmissions).

All 4 RM technologies available on the market were used, assigned to patients before implant, and activated at discharge. In-hospital follow‑ups were done for all technologies 1 and 12 months after implantation.

Inclusion/exclusion criteria

Inclusion

Patients with a standard indication for ICDs with or without CRT. Written informed consent was obtained from participating patients.

Exclusion

None stated.

Primary outcomes

The primary end point was time to investigator's first evaluation of a true-positive clinical- or device-related event during the first year after implant, whichever was first seen during a remote follow‑up (whether or not it was triggered by an automatic alert) or during an in-person visit. An episode was classified as false positive if it did not trigger medical intervention other than device reprogramming. The number of RM transmissions, alerts, and the mean intervals between consecutive RM transmissions were also registered and compared.

Statistical methods

Sample distributions of continuous variables were tested for normality with the Shapiro–Wilk test. Categorical variables were reported as percentages. Comparisons among RM groups were done with the Kruskal–Wallis rank test for continuous variables using Bonferroni's correction for pair-wise multiple comparisons. Chi-square test was used for comparison of baseline categorical variables. Event-free rates were estimated with the product-limit method and Kaplan–Meier plots generated. Comparisons among groups were done with the log-rank test. Areas under Kaplan–Meier curves were calculated with the restricted mean method. Uni- and multivariate Cox proportional hazard models were used to investigate the association between event-free rate and frequency of RM transmissions.

Patients included

n=211 patients with an ICD (65 CareLink; 61 BHM; 49 LAT; 36 SJM)

72% male (CareLink); 70% male (BHM); 70% male (LAT); 83% male (SJM).

Mean age (years): 70 (CareLink); 70 (BHM); 66 (LAT); 67 (SJM).

Results

Event-free rates were 49% with BHM, 57% with LAT, 57% with CareLink, and 58% with SJM (log-rank, p=0.23). BHM generated 304 (IQR, 184–342) transmissions/patient /year, LAT 9 (8–11), CareLink 7 (5–10), and SJM 8 (7–14; p<0.000001). Eighty actionable events occurred at 1‑year follow‑up, 69 (86%) with RM systems; BHM was associated with a higher cumulative rate of actionable events. Daily transmissions were independently associated with an increased probability of event detection compared with periodic transmission systems. The chance of event detection was reduced by 20% (p=0.036) for a 1 month increase of the between-transmission interval (27% for actionable events, p=0.004).

Conclusions

Although all RM systems effectively detected major events, daily transmission was associated with a higher probability of early event detection.

Abbreviations: BHM, BIOTRONIK Home Monitoring; CareLink, Medtronic CareLink; CRT, cardiac resynchronisation therapy; ICD, implantable cardioverter defibrillator; IQR, interquartile range; LAT, Boston Latitude; RM, remote monitoring; SJM, St. Jude Merlin.

Table 11 Summary of results from the de Ruvo et al. (2015) study

CareLink (n=65)

BIOTRONIK Home Monitoring (n=61)

Boston Latitude

(n=49)

St. Jude Merlin

(n=36)

Analysis

Primary outcome

Event notification through RM

46/46

62/69

33/34

38/40

First event notification to

physicians was provided by RM in 179 events (94%) either with automatic alerts or scheduled remote reports with abnormal data.

Selected secondary outcomes

False-positive RM detected events

11

1

4

0

16 remotely detected false positive events (8%).

False-negative RM detected events

0

7 (4 deaths, 2 worsening

HF, 1 undetected AF episode

associated with atrial under-sensing)

1 (left ventricular lead dislodged)

2 (1 death, 1 atrial sensing

issue)

10 events were not detected remotely.

p>0.06

p≤0.008 after

Bonferroni's correction

Actionable events detected by RM

12/14

31/34

20/24

6/8

80 events (42%) were actionable, 69 (86%) of which were detected remotely.

Cumulative rates of

actionable events

22%

37%

45%

16%

p=0.005 log-rank test

A statistically significant difference between BHM and CareLink was detected (p=0.007).

Median maximum interval between transmissions (IQR) in days

93 (82–126)

9 (3–25)

70 (63–

96)

86 (58–93)

The maximum expected duration of unmonitored periods

was significantly shorter in the BHM system (p<0.0001).

Abbreviations: AF, atrial fibrillation; BHM, BIOTRONIK Home Monitoring; HF, heart failure; IQR, interquartile range; LAT, Boston Latitude; RM, remote monitoring; SJM, St. Jude Merlin.

Table 12 Overview of the Landolina et al. (2012) study

Study component

Description

Objectives/hypotheses

The EVOLVO study aimed to test the hypothesis that remote management, using CareLink, can reduce emergency healthcare use (emergency department or urgent face-to-face assessments) in patients with HF who have implanted wireless-transmission-enabled ICD/CRT‑D with specific diagnostic features for HF, thereby increasing efficiency compared with standard management consisting of scheduled face-to-face follow‑up and patient response to audible ICD alerts.

Study design

Prospective, randomised, open, multicentre study

Setting

Six centres in Italy. Patients enrolled from May 2008 to July 2009 and followed up for a 16‑month period.

Intervention and comparator

Intervention

RM strategy using CareLink network service with audible alerts disabled, at 4 and 12 months with face-to-face follow‑ups at 8 and 16 months.

Comparator

Standard management consisting of scheduled visits at 4, 8, 12 and 16 months and patient response to audible alerts.

Inclusion/exclusion criteria

Inclusion criteria

Left ventricular systolic dysfunction or LVEF≤35% documented at the moment of implantation; implantation with a wireless-transmission–enabled Medtronic ICD or CRT‑D endowed with thoracic impedance measurement capabilities (OptiVol algorithm); ability and willingness to have remote follow‑up instead of scheduled routine face-to-face follow‑up visits; and ability to attend all required follow‑up examinations at the study centre.

Exclusion criteria

Reported in Marzegalli et al. (2009) as part of a trial design paper. Patients were excluded if they were under 18 years, were unwilling or unable to give informed consent, had a life expectancy of <12 months, or were participating in another clinical study that may have an

impact on the end points of the present study.

Primary outcomes

Primary Outcomes

All visits (ED and urgent face-to-face follow‑ups) with an interval of <24 hours between the decision to see the patient and the visit. The events anticipated to prompt these visits were ICD alerts for system integrity, atrial and ventricular arrhythmias, decrease in intrathoracic impedance signifying possible fluid accumulation, and patient symptoms. To determine whether remote monitoring was associated with a different rate of ED and urgent face-to-face follow‑up for HF, arrhythmias, or ICD-related events from patients in the standard arm.

Secondary Outcomes

Visits representing the primary end point were further subdivided: visits related to episodes of worsening of HF, and visits for arrhythmias or ICD-related episodes. The rate of total healthcare use (any face-to-face follow‑up visit, emergency department visit, and hospitalisation needing at least 1 overnight stay) for HF, arrhythmias, or ICD events was also compared between groups.

Visits were scrutinized and classified as necessary or unnecessary for the clinical management of the condition.

The study also tested whether remote monitoring reduced the time from any alert condition to the ICD data review, modified the patient's clinical status as measured by the Clinical Composite Score, or modified the patient's QoL as measured by the Minnesota Living With Heart Failure Questionnaire.

Statistical methods

An intention-to-treat analysis was done for all objectives. Primary and secondary hypotheses were tested using the combined Mantel–Haenszel estimate stratified by centre and other potential confounders.

Normality of distribution was tested with the nonparametric Kolmogorov-Smirnov test. Differences between mean data were compared using a t test for Gaussian variables and an F test to check the hypothesis of equality of variance. The Mann–Whitney nonparametric test was used to compare non-Gaussian variables. Differences in proportions were compared by application of Chi-square analysis or the Fisher exact test as appropriate.

Patients included

n=200 patients (99 RM; 101 standard care)

81.8% male (RM); 75.2% male (standard care)

Mean age (years): 66 (RM), 69 (standard care)

Results

Over 16 months, the primary end point was 35% less frequent in the remote arm (75 compared with 117; incidence density, 0.59 compared with 0.93 events per year; p=0.005). A 21% difference was seen in the rates of total healthcare visits for HF, arrhythmias, or ICD-related events (4.40 compared with 5.74 events per year; p<0.001). The time from an ICD alert condition to review of the data was reduced from 24.8 days in the standard arm to 1.4 days in the remote arm (p<0.001). The patients' clinical status was similar in the 2 groups, whereas a more favourable change in QOL was seen from baseline to 16 months in the remote arm (p=0.026).

Conclusions

The results showed that RM can reduce ED or urgent face-to-face follow‑up and, in general, total healthcare use in patients with HF with modern ICD/CRT‑D. Compared with standard follow‑up through face-to-face follow‑up and audible ICD alerts, RM resulted in increased efficiency for healthcare providers and improved quality of care for patients.

Abbreviations: CRT‑D, cardiac resynchronisation therapy with defibrillator; CI, confidence interval; ED, emergency department; HF, heart failure; ICD, implantable cardioverter defibrillators; IRR, incident rate ratio; LVEF, left ventricular ejection fraction; QoL, quality of life; RM, remote monitoring.

Table 13 Summary of results from the Landolina et al. (2012) study

CareLink remote monitoring (n=99)

Face-to–face monitoring (n=101)

Analysis

Primary outcome

All emergency department and urgent face-to-face follow‑ups

75 visits

0.59 events per year

117 visits

0.93 events per year

IRR 0.65; 95% CI, 0.49–0.88 (p=0.005)

IRR adjusted for centre, use of CRT, and ischaemic origin.

Selected secondary outcomes

ED and urgent face-to-face follow‑ups for worsening of HF

48 visits

0.38 events per year

92 visits

0.73 events per year

IRR=0.52; 95% CI, 0.37– 0.75 (p<0.001)

ED and urgent face-to-face follow‑ups for arrhythmias or ICD-related episodes

27 visits

0.21 events per year

25 visits

0.20 events per year

IRR=1.14; 95% CI, 0.65–1.99 (p=0.649)

Healthcare use for HF, arrhythmias or device-related events

4.4 events per year

5.74 events per year

IRR=0.79; 95% CI, 0.71– 0.89 (p<0.001)

Hospitalisations needing at least 1 overnight stay

0.45 events per year

0.39 events per year

p=0.464

Wireless remote notifications (CareLink) and audible alerts (control)

2.5 events per year

2.4 events per year

IRR=1.04; 95% CI, 0.89 –1.23 (p=0.602)

Rate of appropriate additional visits due to alerts

86% (72/84)

53% (42/79)

p<0.001

Median time from alert to ICD data review

1.4 days (25th–75th percentile 0.8–7.3)

24.8 days (25th–75th percentile 9.5–48.8).

p<0.001

Change in QoL at 16 months

−2 (25th–75th percentile −17 to 8)

+2 (25th–75th percentile −7 to 10)

p=0.026

Change in clinical status from the time of enrolment to the 16‑month follow‑up visit

17% improved, 49% were unchanged, and 34% worsened.

20% improved, 36% were unchanged, 44% worsened.

No statistical difference

Abbreviations: CI, confidence interval; CRT cardiac resynchronisation therapy; CI, confidence interval; ED, emergency department; HF, heart failure; IRR, incident rate ratio; QoL, quality of life.

Table 14 Overview of the Luthje et al. (2015) study

Study component

Description

Objectives/hypotheses

To estimate the influence of remote monitoring with fluid monitoring using OptiVol alerts on the time-to-first heart failure related hospitalisations, ventricular tachyarrhythmia occurrence, and mortality when compared with standard clinical care.

Study design

A prospective, single-centre, randomised study

Setting

A single centre in Germany. Patients enrolled between December 2007 and April 2011, and followed up for 15 months.

Intervention and comparator

Intervention

Patients in the RM arm were connected to the Medtronic CareLink network.

Comparator

In the control group, standard face-to-face follow‑ups were done every 3 months.

Patients having CRT‑D or ICD implants or replacements were randomised to RM including OptiVol ON (remote arm) compared with RM OFF (standard arm) and followed for 15 months. In both groups, the audible

OptiVol alert was disabled.

Inclusion/exclusion criteria

Inclusion criteria

Patients >18 years needing an ICD or CRT‑D according to current guidelines, or patients with a previously implanted device without FM feature and a replacement indication for battery depletion were included in the study after written informed consent. HF or a history of hospitalisation for decompensated HF was not a prerequisite for inclusion.

Exclusion criteria

Permanent atrial fibrillation, a life expectancy ≤15 months, pregnancy, and participation in another study.

Primary outcomes

The primary outcome was the time taken to first hospitalisation due to worsened heart failure and was stated in Zabel et al. (2013) as part of a study design paper.

Statistical methods

Differences in baseline characteristics were evaluated using student's t-test, a Chi-square test, or a Fisher's exact test, as appropriate. For time-to-first HF-related hospitalisation, time-to-first ICD shock, and time to death a Cox proportional hazard analysis, and Kaplan–Meier survival analysis with log-rank test were done.

Patients included

n=176 (87 CareLink; 89 control)

80.5 % male (CareLink); 74.2% (control)

Mean age (years): 66 (CareLink); 65.9 (control).

Results

Cox proportional hazard analysis on the time-to-first HF-related hospitalisation showed a hazard ratio of 1.23 (0.62–2.44; p=0.551) favouring the control group. In the remote group, 13 patients (15%) had ICD shocks compared with 10 patients (11%) in the control group (p=0.512). Average time-to-first ICD shock was 212±173  in the remote arm and 212+143 days in the control arm (p=0.994). The Kaplan–Meier estimate of mortality after 1 year was 8.6% (8 deaths) in the remote group compared with 4.6% in the control group (6 deaths; p=0.502).

Conclusions

RM in combination with FM had no significant effect on HF-related hospitalisations, ICD shocks or mortality.

Abbreviations: ICD, implantable cardioverter defibrillator; CRT‑D, cardiac resynchronisation therapy and defibrillator; FM, fluid monitoring; HF, heart failure; RM, remote monitoring.

Table 15 Summary of results from the Luthje et al. (2015) study

CareLink remote monitoring (n=87)

Standard care group (n=89)

Analysis

OptiVol alerts

174 (78%) alerts (94 in patients with a CRT and 80 in patients with an ICD) in 68 patients (35 with a CRT and 33 with an ICD).

93 alerts were classified as true positive based on clinical assessment.

Patients hospitalised for worsened HF during follow‑up

20

22

One patient in the RM group was hospitalised before an OptiVol Alert was sent.

Mean number of emergency department visits

0.10±0.25

0.10±0.23

No significant difference

Mean number of urgent

care visits

0.30±0.50

0.10±0.30

p=0.0332

Total number of patients having ICD shocks

13 (15%)

10 (11%)

Number of patients having inappropriate ICD shocks

2 (2%)

2 (2%)

Average time-to-first ICD shock

212±173 days

212±143 days

p=0.994

No significant difference (Kaplan–Meier analysis)

Number of deaths

8 (1 sudden cardiac; 3 non-sudden cardiac; 1 non‑cardiac death.

3 deaths could not reliably be classified)

6 (1 sudden cardiac; 3 non-sudden cardiac; 2 non‑cardiac deaths)

The Kaplan–Meier estimate of all-cause mortality after 1 year was 8.6% in the RM group compared with 4.6% in the control group. No significant difference was seen for time to death between the 2 groups.

Abbreviations: ICD, implantable cardioverter defibrillator; CRT‑D, cardiac resynchronisation therapy and defibrillator; HF, heart failure.