Author + information
- Received November 10, 2003
- Revision received February 3, 2004
- Accepted March 16, 2004
- Published online July 7, 2004.
- Ruediger Becker, MD*,* (, )
- Jutta Ruf-Richter, RN*,
- Julia C Senges-Becker, MD*,
- Alexander Bauer, MD*,
- Slawomir Weretka, MD*,
- Frederik Voss, MD*,
- Hugo A Katus, MD* and
- Wolfgang Schoels, MD*
- ↵*Reprint requests and correspondence:
Dr. Ruediger Becker, University of Heidelberg/Cardiology, Bergheimer Strasse 58, 69115 Heidelberg, Germany.
Objectives The purpose of this study was to analyze the utility of patient-alert features in implantable cardioverter defibrillators (ICDs).
Background Various alert features producing acoustic warning signals have been implemented in newer generation ICDs, but their role in early detection of system-related complications has not been systematically evaluated.
Methods In 240 patients implanted with Medtronic ICD devices, the following alert features were routinely activated: pacing lead impedance <200 or >2,000 Ω, high-voltage lead impedance <10 or >200 Ω, low battery voltage (elective replacement indicator), long charge time (>18 s), >3 shocks delivered per episode, and all therapies in a zone delivered. Alert events occurring during follow-up were assessed in relation to actual findings (hospital charts, chest X-rays, ICD printouts including sensing/pacing/defibrillation threshold tests, episode data) to determine incidence, sensitivity, and specificity of the alert function.
Results During 12.2 ± 8.9 months, 24 alert events occurred in the 240 patients (pacing lead impedance, n = 4; high-voltage lead impedance, n = 7; low battery voltage, n = 1; >3 shocks, n = 6; all therapies, n = 6). A total of 22 serious complications (necessitating reprogramming or device/lead replacement) were observed, 14 of which were primarily identified through a patient alert (lead fracture, n = 11; connector defect, n = 1; T-wave oversensing, n = 1; battery depletion, n = 1). This reflects a sensitivity of 64% and a specificity of 96% of the alert function for serious complications. With 14 of 24 patient alerts being caused by serious complications, the positive predictive value reached 58%.
Conclusions Patient-alert features are a useful additional tool facilitating early detection of serious ICD complications, but they do not substitute for regular ICD follow-up, because of their low sensitivity.
System-related complications still occur in a significant proportion of patients with implantable cardioverter defibrillators (ICD). The majority of complications are lead-related and occur within three months after device implantation (1–11); however, even during long-term follow-up, system-related complications are encountered (1,3,12–16). Keeping in mind that many of these complications are potentially life-threatening, immediate detection would be desirable to ensure maximal safety. Therefore, Patient Alert (Medtronic Inc., Minneapolis, Minnesota) has been developed, which is a programmable feature that enables a patient to be audibly alerted once one of various parameters exceeds the normal range. The present study was designed to assess the utility of this feature based on an analysis of incidence, cause, and clinical consequences of alert events.
All patients implanted with respective ICD systems (Medtronic 7227, 7229, 7230, 7231, 7271, 7272, 7273, 7275) between February 1999 and July 2002 and followed up for at least three months postoperatively were included in this retrospective study.
Standard patient-alert settings
The following alert features were routinely activated before hospital discharge:
1) Pacing lead impedance <200 or >2,000 Ω
2) High-voltage lead (HVB) impedance <10 or >200 Ω
3) Low battery voltage (elective replacement indicator)
4) Long charge time (>18 s)
5) >3 Shocks delivered in one episode
6) All therapies in a zone delivered
Any abnormality in one of these parameters would audibly alert the patient once daily at a programmable time. The alert time was individually adapted to the patients' waking hours (usually 8 to 10 a.m.). All patients were instructed to immediately present to our ICD clinic if an alert signal should occur.
Routine follow-up schedule
After ICD implantation, all patients underwent a prehospital discharge test, including measurement of standard lead parameters and induction of ventricular fibrillation (VF). Routine outpatient follow-up was scheduled for one and three months postoperatively and every three months thereafter. Chest X-rays were performed every six months.
Data acquired during routine three monthly follow-up and unplanned visits
The following data were acquired:
1) Patient history
2) Physical examination
3) 12-lead electrocardiogram
4) Device interrogation
a) Episode data
b) Battery voltage/charge time
c) (P/)R-wave measurement
d) Pacing threshold test
e) Lead impedance test
f) Oversensing tests, if suspected
Data from routine follow-ups and unplanned visits (UPVs) prompted by patient-alert events were analyzed with respect to incidence, cause, and clinical consequences. A decrease in pacing impedance <200 Ω was considered as an indicator of insulation failure, and a sudden increase >2,000 Ω was categorized as a sign of lead fracture. A serious complication was defined as a device or lead dysfunction necessitating surgical revision or immediate reprogramming (e.g., lead fracture, exit block, undersensing, oversensing with inappropriate therapy, battery depletion). Sensitivity and specificity of the patient-alert function were calculated for the detection of seriouscomplications. Non-serious complications disclosed through a patient alert were considered as false positive alerts.
Statistics were performed using SAS for Windows Version 6.12 (SAS Institute, Cary, North Carolina). For comparison between groups, a chi-square test was applied. A p value <0.05 was considered statistically significant.
A total of 240 patients were included and followed up for 12.2 ± 8.9 months. The patients' clinical characteristics are summarized in Table 1. A total of 32 (13.3%) complications were encountered, of which 22 (9.2%) were judged to be serious (Table 2). During the follow-up period, 24 patients (10%) experienced an alert event: pacing lead impedance, n = 4; HVB impedance, n = 7; low battery voltage, n = 1; >3 shocks, n = 6; and all therapies, n = 6. A total of 14 of 22 serious complications were primarily identified through Patient Alert (sensitivity 64%, specificity 96% [208 of 216]) (Table 2). On the other hand, 14 of 24 alert events were caused by serious complications (positive predictive value 58%) (Tables 3 and 4). ⇓⇓Complications identified through alert events or not associated with alert events are specified in Table 2. Alert events were significantly more common after device replacement (using existing leads) than after first implantation or complete system (device + lead) reimplantation (9 of 37 = 24.3% vs. 15 of 203 = 7.4%, p = 0.004). Furthermore, patients with abdominal devices were more commonly affected by alert events (5 of 9 = 55.6% vs. 19 of 231 = 8.3%, p < 0.001). Other clinical parameters (ICD indication, underlying heart disease, age, and so forth) did not differ between patients with and those without alert events.
The lead fractures not associated with alert events (n = 5) were detected as follows: massive oversensing with delivery of all VF therapies prompting a UPV before the alert signal would occur (n = 1); oversensing with ≤3 inappropriate shocks either prompting UPVs (n = 2) or diagnosed at routine follow-up (n = 2). In all cases (n = 5), the sensing integrity counter had collected >300 short RR intervals (120 to 130 ms) within three months.
Of note, five patients claimed to have perceived an alert signal, but upon device interrogation, no alert event was documented. These “phantom” alerts were excluded from statistical analysis.
Presentation of characteristic cases
Two months after routine follow-up in May 2003, a patient implanted with a single lead system (Medtronic 7229Cx device/6943 lead) due to ventricular tachycardia experienced an alert event triggered by a pacing lead impedance >2,000 Ω. During ICD testing, normal sensing and pacing parameters were observed (pacing threshold 2 V/0.2 ms; R wave 7 mV); no oversensing was documented, despite rigorous myopotential testing. However, the pacing lead impedance alternated between 410 to 440 Ω and >2,000 Ω. During surgical revision, an incomplete lead fracture necessitating lead replacement was found. Without a patient-alert feature, this complication might have been missed even during consecutive follow-up visits, because single impedance measurements as performed during routine check-up could have yielded normal results.
Two weeks after routine follow-up in April 2002, a patient implanted with an abdominal system in 1995 due to resuscitated VF perceived an alert signal. The initial device (Medtronic 7218D) had been replaced with a Medtronic 7227D in August 2000 due to battery depletion, leaving both leads in place (right ventricular [RV] apex, Medtronic 6936; superior vena cava [SVC], Medtronic 6937). Device interrogation disclosed that the alert had been triggered by a high voltage lead impedance >200 Ω. During myopotential testing, marked oversensing was documented in the far field electrogram derived between SVC and RV high voltage lead (HVA/HVB). This suggested a conductor wire breach of either the RV or the SVC high voltage lead. Although the chest X-ray was unsuspicious, surgical lead revision was performed. Selective intraoperative impedance measurements (unipolar) on both high voltage leads yielded an RV lead impedance >200 Ω, confirming lead fracture.
To the best of our knowledge, this is the first study to evaluate Patient Alert as an ICD feature designed to facilitate early detection of system-related complications. During a mean follow-up of one year, 10% of the patients experienced an alert event. The majority of alert events were caused by serious complications necessitating reprogramming or device/lead replacement. As expected, most complications detected through alert events were lead-related (86%). However, eight of 22 (36%) serious complications occurring during the follow-up period were not associated with alert events. According to our data, Patient Alert contributes to the early detection of system-related complications, but it does not substitute for regular follow-up visits, because of its relatively low sensitivity (64%).
The incidence of ICD-related complications encountered in clinical studies depends on factors such as cohort size, follow-up duration, definition of complications, device/lead selection, mode of implantation, and so forth. In fact, published overall complication rates range from 3.9% to 53%, with lead-related complications varying between 2.1% and 22%. With 13.3% (32 of 240) overall and 7.9% (20 of 240) lead-related complications, the present study was comparable with other mid-term follow-up trials in newer generation ICDs (2,4,5).
Based on the parameters monitored, the alert feature holds the potential for detection of lead- and device-related complications such as lead fracture, dislodgment, insulation defect, and sudden battery depletion. However, in the absence of automatic sensing/pacing threshold tests, sensing and pacing defects with preserved lead continuity are missed. Therefore, in its present form, the feature's sensitivity in diagnosing system-related complications is necessarily limited. Nevertheless, the prompt detection of serious complications in a considerable proportion of patients, as demonstrated in this study, clearly justifies its routine use. According to our data, the integration of sensing integrity counters (= short interval counters) might enhance the sensitivity of the alert feature.
- implantable cardioverter defibrillator
- right ventricle/ventricular
- superior vena cava
- unplanned visit
- ventricular fibrillation
- Received November 10, 2003.
- Revision received February 3, 2004.
- Accepted March 16, 2004.
- American College of Cardiology Foundation
- Ellenbogen K.A,
- Wood M.A,
- Shepard R.K,
- et al.
- Rosenquist M,
- Beyer T,
- Block M,
- et al.
- Hoffmann E,
- Steinbeck G
- Kron J,
- Herre J,
- Renfroe E.G,
- et al.
- Brady P.A,
- Friedman P.A,
- Trusty J.M,
- et al.
- Nunain S.O,
- Roelke M,
- Trouton T,
- et al.