Limitations and Late Complications of Third-Generation Automatic Cardioverter-Defibrillators
Background This study examines the limitations and complex management problems associated with the use of tiered-therapy, implantable cardioverter-defibrillators (ICDs).
Methods and Results The study group comprises the first 154 patients undergoing implantation of tiered-therapy ICDs at our institution. Pulse generators from three different manufacturers were used. In 39 patients, a complete nonthoracotomy lead system was used. The perioperative mortality was 1.3%. Of these 154 patients, 37% experienced late postoperative problems. Twenty-one patients required system revision within 36.5 months (mean, 8.57±11.3) of surgery. Reasons for revision were spurious shocks due to electrode fractures (3) or electrode adapter malfunction (2), inadequate signal from endocardial rate-sensing electrodes (3), superior vena cava or right ventricular coil migration (5), failure to correct tachyarrhythmias due to a postimplant rise in defibrillation threshold (5), or pulse generator failure (3). One of these patients required system removal for infection after revision of an endocardial lead. A further 32 patients received inappropriate shocks for atrial fibrillation with a rapid ventricular response or sinus tachycardia. Two of these patients also received shocks for ventricular tachycardia initiated by antitachycardia pacing triggered by atrial fibrillation. Ventricular pacing for bradycardia was associated with inappropriate shocks due to excessive autogain in 2 patients.
Conclusions Despite the major diagnostic and therapeutic advantages of tiered-therapy ICDs, a significant proportion of patients continue to experience hardware-related complications or receive inappropriate shocks.
Since the first human implant of an automatic internal cardioverter-defibrillator (ICD) in 1980 for the treatment of recurrent cardiac arrest, there have been rapid advances in device technology.1 These advances, coupled with the clinical efficacy of these devices in the prevention of sudden cardiac death, have led to a rapid expansion of the initially very stringent indications for ICD implantation.2 3 Earlier device implantation without protracted serial antiarrhythmic drug testing is increasingly practiced, not only for the treatment of cardiac arrest survivors but also for patients whose presenting arrhythmia is a hemodynamically tolerated sustained monomorphic ventricular tachycardia.4
The first available ICD devices were capable of high-energy defibrillation only irrespective of the nature of the ventricular arrhythmia. The current third-generation, multiprogrammable ICD pulse generators, in addition to high-energy defibrillation, incorporate facilities for antitachycardia and antibradycardia pacing, low-energy cardioversion, telemetry and intracardiac electrogram storage, and noninvasive programmed stimulation. A “second look” feature allows the device to abort therapy for arrhythmias that terminate spontaneously during charging.5 Concomitant with these improvements in ICD pulse generators, newer lead systems that can be implanted transvenously are likely to widen the indications for ICD therapy.6 7 8
The initial reports on these newer third-generation defibrillators and transvenous lead systems have focused on their advantages over earlier devices.9 10 11 However, as the technical aspects of device implantation become simplified, it is important that the limitations and complications associated with their use be evaluated. This study examines these problems and illustrates how they can be anticipated and in some cases avoided.
The study group comprises the first 154 patients undergoing implantation of third-generation ICDs at our institution for the treatment of sustained ventricular tachycardia (VT) or ventricular fibrillation (VF) or syncope associated with inducible VT not suppressed with antiarrhythmic drug therapy. Patient characteristics and baseline electrophysiological data are summarized in Table 1⇓. This patient group represents a highly selected population, of whom 68% presented with sustained monomorphic VT, with the majority of this subgroup (76%) having VT terminated by ventricular pacing either at baseline or on antiarrhythmic drugs. In contrast to earlier studies comprising mainly cardiac arrest survivors, there was a very low incidence of coronary artery bypass grafting (CABG) at the time of device implantation. However, a large proportion (34%) of these patients had undergone CABG in the past.
Third-generation ICD pulse generators from three different commercial manufacturers were used: Medtronic PCD model 7217B (61 patients), Ventritex Cadence V 100 (76 patients), and CPI Ventak PRx (17 patients). All of these pulse generators are multiprogrammable, and their features are summarized in Table 2⇓.
Several medical as well as nonmedical factors influenced the choice of generator used. These factors included the compatibility of the device with previously implanted defibrillation patch electrodes, the advantage of stored intracardiac electrograms in the management of patients with complex arrhythmias,12 the potential advantage of biphasic shocks13 in patients with previously failed implantation due to high defibrillation thresholds (DFTs) (Cadence only), the desirability of implanting a nonthoracotomy lead system (Medtronic PCD), and device availability. The transvenous defibrillation lead system used was the Medtronic Transvene System, which for the purposes of this study was used with the Medtronic PCD only. This lead system comprises two separate high-voltage defibrillation electrodes. The larger of the two leads incorporates both the right ventricular bipolar sensing electrodes and a defibrillation coil. Active screw-in fixation is used for stability. The second defibrillation lead lies free in the superior vena cava at its junction with the right atrium.
Primary system implantation was performed in 141 patients and pulse generator replacement in the remaining 13. One hundred patients underwent a planned thoracotomy approach. In these patients, epicardial rate-sensing electrodes were implanted during thoracotomy in 29 patients and transvenously placed endocardial electrodes were used in the remainder. In 54 patients, implantation of a complete nonthoracotomy lead system was attempted. The attempted nonthoracotomy approach was performed using a graded protocol. Intraoperative defibrillation threshold testing was carried out in all patients receiving a primary implant. Satisfactory DFTs were defined as successful conversion of three of four episodes of VF using 18 J or less. In some patients in whom the morbidity of thoracotomy or system revision was judged to be prohibitive, DFTs of up to 24 J were accepted. The fact that all three devices have a maximum stored energy of at least 34 J allowed a safety margin of at least 10 J in all patients. If satisfactory DFTs could not be obtained using the transvenous defibrillation leads in combination with a subcutaneous patch, a single left ventricular extrapericardial patch electrode was implanted via thoracotomy. If this was unsuccessful, a second right ventricular patch was placed behind the sternum. In all except one patient, the right ventricular endocardial lead was used for rate sensing. Perioperative prophylactic antibiotic coverage consisting of vancomycin and cefazolin was administered routinely.
Before discharge from the hospital, patients underwent a predischarge electrophysiology study. In patients whose preoperative VT was pace-terminable, the efficacy of antitachycardia pacing (ATP) was tested. Burst rather than ramp ventricular pacing was used in the vast majority of patients. If ATP was unsuccessful, this function was not activated at discharge and further assessment was made at postdischarge noninvasive programmed stimulation (NIPS).
After discharge from the hospital, all patients were seen at 1- to 2-month intervals or more frequently if ICD discharges occurred inappropriately or with a high frequency. Within 1 to 2 months of discharge from the hospital, ICD efficacy was tested by noninvasive programmed stimulation unless an appropriate spontaneous discharge had confirmed device efficacy. At each visit, the device was interrogated and the number of tachycardia detections, their therapy, and efficacy were recorded. Patients who underwent device explantation for any reason, for example, cardiac transplantation, were censored from the study at the time of explantation.
Classification of Events
Therapies were classified as appropriate or inappropriate, based on all the available data including patient symptoms, telemetry data from the ICD including stored intracardiac electrograms, 12-lead surface ECG data as well as Holter or ECG event recordings, or rhythm strips recorded by emergency medical personnel. Therapy was classified as inappropriate only when the evidence for this was unequivocal. When the number of clinical events on the therapy log since the last interrogation exceeded the ability of the device to provide stored intracardiac electrograms for corroboration, the available stored events were taken as representative of all therapies delivered since the last visit for statistical purposes.
Classification of Deaths
Sudden cardiac death was defined as death within 1 hour of onset of symptoms or resuscitated cardiac arrest from which the patient did not regain consciousness before death. Other cardiac deaths were defined as nonsudden. Death from other causes was defined as noncardiac.
Of the 154 patients in whom a third-generation device was implanted, 152 were discharged alive from the hospital. There were no intraoperative deaths. There were two surgical deaths (1.3%) due to a worsening of the patient’s ventricular arrhythmia after surgery.14 These patients, with left ventricular ejection fractions of 12% and 15%, respectively, had undergone thoracotomy for epicardial patch placement and required multiple cardioversions with subsequent hemodynamic collapse. Neither patient, despite the presence of severe left ventricular dysfunction, had overt heart failure before surgery. There was one late postoperative infection (0.6%) after a rate-sensing lead revision.
Follow-up data are available on all patients. During a mean follow-up of 15.3±9.7 months (range, 1.5 to 42), 99 patients (65.1%) received therapy from their devices (Fig 1⇓). In 56 patients (36.8%), device therapies were always appropriate, that is, they were delivered for sustained ventricular arrhythmias only. Nearly one fifth (18.4%) of patients received both appropriate therapy for VT or VF and inappropriate therapies for other reasons. In 15 patients (9.8%), device therapy was delivered for inappropriate reasons only.
A total of 56 patients (36.8%) had late postoperative ICD-related problems. Twenty-one of these patients required a system revision a mean of 8.6±11.3 months after device implantation (see Table 3⇓). The majority of system revisions occurred in the early stages after implantation (13 of 21 occurring within the first 6 months). Seven of these 21 patients received inappropriate shocks; a further 34 patients (22.3%) received inappropriate shocks for the reasons listed in Table 3⇓. The primary cause of inappropriate therapy was the occurrence of supraventricular arrhythmias. Three patients had more than one device-related problem. Fig 2⇓ shows the Kaplan-Meier survival curves for these patients. The 2-year survival without the need for a system revision and without inappropriate shocks was 86% and 57%, respectively.
Not all of the problems listed in Table 3⇑ are of equal seriousness. The loss of sensing function and substantial rises in DFT are the most life-threatening, as both problems may result in failure to restore sinus rhythm in the event of a cardiac arrest. Repetitive spurious shocks in response to supraventricular arrhythmias are also potentially life-threatening in patients with precarious ventricular function. Device infection also represents a major problem, as it usually involves the removal of all device components. The rest of the problems outlined in Table 3⇑ are less serious, although they may still be associated with significant morbidity.
Major Therapeutic Advantages
Of 85 patients with pace-terminable VT at preimplantation electrophysiological evaluation, effective ATP was demonstrated in 83 patients (98%) either at predischarge or subsequent electrophysiological testing. Among these 83 patients, ATP therapy was used appropriately in 58 (69%) during follow-up. A total of 3325 episodes of VT were treated with ATP as initial therapy. ATP therapy was successful in terminating 3089 of these episodes, giving an overall success rate of 92.9%. There was, however, a wide variation in interpatient effectiveness, with some patients having several hundred events successfully pace-terminated while in others only a small proportion of events responded to ATP therapy (Fig 3⇓). However, on a per-patient basis, the majority of individuals (62%) had more than 90% of all episodes successfully terminated by ATP, whereas only 12% had fewer than 1 in 10 episodes terminated.
Aborted Shock Function
The aborted shock function was used on 288 occasions in 47 patients (30.9%) (Fig 4⇓). The mean number of aborted shocks per patient was 6.1 (range, 1 to 45). The total number of aborted shocks for inappropriate triggers such as lead noise or atrial fibrillation was almost twice that for nonsustained VT or VF (184 versus 104). The mean number of aborted shocks for nonsustained VT was 3.4 per patient (range, 1 to 24), while that for inappropriate triggers was 6.3 per patient (range, 1 to 45).
Slow VT. In all except 1 patient, detection of ventricular arrhythmias was appropriate for the programmed device parameters. However, one of the most troublesome of problems not yet adequately solved by current ICD technology is the inability to reliably differentiate between VT and supraventricular rhythms when there is rate overlap between the two.
Because of the constraints involved in choosing optimal rate cutoffs in patients with overlapping sinus and VT rates, 12 patients (7.9%) had spontaneous slow VT below the programmed detection rate of the device. In 8 patients (5.2%) in whom such recurrent slow VTs could not be satisfactorily managed despite multiple antiarrhythmic drug trials and complex device reprogramming, endocardial ventricular mapping with a view to transcatheter radiofrequency ablation was undertaken. In 5 patients (3.2%), ablation of the target slow VT was possible. In these patients, despite the fact that faster nonclinical tachycardias were often still inducible by programmed ventricular stimulation after ablation, the technique provided useful adjunctive therapy to reduce the frequency of multiple ICD discharges for slow VT.
Supraventricular tachycardias. At the other end of the spectrum, 32 patients (21%) received inappropriate shocks for either sinus tachycardia or atrial fibrillation with a rapid ventricular response. ATP was used as initial therapy in 26 patients (17%). In some cases, ATP resulted in apparent successful termination of the supraventricular tachycardia (SVT) due either to spontaneous slowing of the heart rate or possibly by concealed retrograde conduction in the AV node during ventricular pacing. This “pseudotermination” of supraventricular arrhythmias by ATP could be seen clearly in patients with stored intracardiac electrograms (Fig 5⇓). In those with event counters only, inappropriate therapies with pseudotermination were classified as successful therapy by the device log. In this patient population, therefore, the ICD log will tend to overestimate device efficacy.
In a majority of patients, antitachycardia pacing for supraventricular rhythms was simply ineffective. However, in 2 patients, burst ventricular pacing during SVT induced VT, as has previously been recognized.15 These episodes of VT but not the background atrial arrhythmia were terminated by shock therapy (example in Fig 6⇓). The cycle of ATP, VT, and shock was then repeated until the device could be either inactivated or reprogrammed. In patients in whom ATP was not activated or failed to slow SVT, shocks of increasing energy were delivered. However, in only 1 patient did these shocks terminate the supraventricular arrhythmia. In 7 patients, a series of 15 or more inappropriate shocks was delivered during a single clinical episode. In 1 patient, this complication was sufficiently disturbing to initiate requests for device inactivation and require psychiatric counseling. In an additional patient, fortuitously wearing a Holter monitor recorder at the time, there was evidence of progressive ST-segment elevation after repeated shocks for sinus tachycardia with subsequent ECG and enzymatic evidence of nontransmural myocardial infarction.
Drug therapy to prevent spurious shocks. At the time of initial discharge from the hospital, all patients with a history of atrial fibrillation or at high risk for spurious discharges for SVTs (for example, young patients with high levels of activity) were discharged on sinus slowing or AV nodal–blocking drugs (digoxin, 50%; β-adrenergic blocker, 36%; calcium channel blocker, 7%; class Ia antiarrhythmic agents, 20%; and amiodarone, 6%).
Those patients receiving inappropriate therapy for SVTs were taking the following medications at the time of first inappropriate therapy: digoxin, 64%; β-adrenergic blocker, 37%; and calcium channel blocker, 18%. Initially, 67% of patients were taking one, 18% two, and 3.0% three of these agents. After adjustments in therapy, all were receiving rate-lowering medications, with 52% receiving one, 42% two, and 6% three of the above agents, with 10% of patients also receiving amiodarone therapy.
Rise in defibrillation thresholds. In 4 patients, ICD system revision was required because of high DFTs. In 2 of the patients, both with nonthoracotomy lead systems and initial DFTs of ≤18 J, multiple episodes of VT required initiation of therapy with amiodarone. When the DFTs were retested on amiodarone, 34-J shocks failed to restore sinus rhythm, and the systems were therefore revised with implantation of supplementary extrapericardial defibrillation patch electrodes. In 1 patient, a rise in DFT led to a failure to convert a clinical tachyarrhythmia, and this patient was successfully resuscitated by emergency medical technicians.
Migration of Transvenous Defibrillation Leads
Of 54 patients with transvenous defibrillation electrodes, lead displacement or retraction occurred in 5 patients (superior vena cava lead, 2 patients; right ventricular lead, 3 patients). In 1 patient, perforation of the right ventricle occurred without pericardial tamponade, which required a thoracotomy for system revision. This patient developed severe pericarditis, which persisted for several months after hospital discharge.
Sensing Lead Problems
Sensing lead problems resulted from either insulation breaks, Y-adaptor malfunction, or a marked diminution of sensed R-wave amplitude (Table 3⇑). The former two problems led either to inappropriate or aborted shocks that were documented by intracardiac electrogram recordings. In 1 patient, mechanical noise induced by intermittent contact of a redundant endocardial rate-sensing lead against a newly implanted sensing electrode (ringing) resulted in multiple inappropriate shocks. Only one of these sensing problems occurred with the nonthoracotomy lead system.
Pulse Generator Problems
In 4 patients, the implanted pulse generator was prematurely explanted and replaced. In 2 patients this was done because of a faulty circuit capacitor and in another patient because of premature battery depletion at 6 months (no patient in this series has yet required an elective pulse generator change). In all cases, pulse generator malfunction was detected at routine follow-up interrogation or testing. No adverse clinical outcome resulted. In the third patient, the device, for unknown reasons, failed to detect an ECG-documented episode of VT, which was well within the programmed detection parameters. The device was therefore replaced, but no abnormality was found on factory testing.
Autogain-Related Sensing Problems
Two of the three devices in this study incorporate an automatic gain control feature. The autogain feature automatically maximizes gain settings during bradycardia pacing to ensure that the underlying rhythm interpreted as asystole is not in fact VF with low-amplitude intracardiac signals and ignores the immediate postpacing signal (postpacing refractoriness). When sinus rhythm resumes, however, the transiently augmented amplitude of the intracardiac electrogram may result in double or triple counting, with subsequent inappropriate shock delivery during sinus rhythm. This was observed in 2 patients who received inappropriate shocks after ventricular pacing. In 1 patient, it was resolved by programming an increase in the number of tachycardia intervals for VT or VF detection, thereby allowing sufficient time for the autogain function to gradually reduce gain settings before detection and therapy. In the second patient, this approach proved ineffective and required placement of a separate permanent pacemaker for bradycardia pacing.
Other Management Issues
In 5 patients (3.2%), all of whom had coronary artery disease and easily inducible VT at preimplant evaluation, 25 episodes of VT after ventricular paced beats terminating prolonged pauses were documented by stored intracardiac electrograms (Fig 7⇓). These paced beats were all appropriately timed and did not result from failure to sense sinus beats, as has previously been described.16 This mechanism of VT induction, recognized by careful analysis of stored intracardiac electrograms, can easily be overlooked because of the very small size of the ventricular pacing artifact in most cases (Fig 7⇓). These episodes of VT are probably promoted by the long cycle length of the pauses, since they may be effectively prevented by increasing the backup bradycardia pacing rate. By contrast, the use of antiarrhythmic drug therapy, which may further slow the sinus rate, may be counterproductive.
Deaths and System Removal
Twenty-three patients (15.1%) died during follow-up, 19 (12.5%) from cardiac causes. Five other patients had their devices removed, 3 at the time of cardiac transplantation and 1 at the time of corrective arrhythmia surgery. The final patient had the device removed because of infection related to a sensing lead revision procedure. This patient subsequently underwent successful ventricular arrhythmia surgery.
Cardiac deaths occurred a mean of 11.3±9.7 months after ICD implantation. The mean left ventricular ejection fraction of these patients was 27.1±10.7%. A total of 13 deaths were due to progressive heart failure. Six patients died suddenly 6.1±5.1 months after ICD implantation, resulting in a 96.4% 1-year actuarial sudden death–free survival. All 6 patients had a history of clinically significant heart failure, with mean left ventricular ejection fraction of 31.8±8.9%. Two of these 6 patients suffered acute hemodynamic collapse, were resuscitated by emergency medical technicians, and subsequently died without regaining consciousness 48 and 72 hours later, respectively. In 1 patient, the first documented rhythm was ventricular pacing, with no evidence of shocks from his device. An additional patient awaiting cardiac transplantation was admitted to the hospital in VT with shock after his device had fired several times; he died almost immediately afterward in electromechanical dissociation after an external shock. One patient with ischemic heart disease died within 20 minutes of the onset of acute chest pain and could not be resuscitated. Two patients died suddenly at home with unwitnessed collapse, with device interrogation in 1 patient revealing several successful shock therapies followed by bradycardia pacing. In none of these patients was there any evidence that the device had failed to detect arrhythmias or deliver appropriate therapy.
For the earliest generation of ICDs, there was appropriate skepticism about the ability of the devices to reliably detect and convert life-threatening ventricular arrhythmias. In this series, however, there was only a single unexplained episode of failure to detect VT. Thus, currently available ICD devices are extremely reliable and efficient at detecting and terminating VTs. However, third-generation ICDs, despite their technical sophistication, are associated with a significant incidence of management problems. ICD therapy is now regarded as the treatment of choice for cardiac arrest survivors, coupled with coronary revascularization and adjunctive drug therapy where appropriate. The availability of newer ICDs with smaller pulse generators, biphasic shock waveforms, and improved lead systems will lower the morbidity of device implantation and contribute to more widespread use in the future.17 However, the present study underscores the fact that the use of these devices may be associated with significant problems.
The majority of problems observed in this study occurred in patients who also had appropriate detection and termination of VT or VF. This is not surprising, since particularly in patients with relatively slow VT, tachyarrhythmia detection rates must be programmed in a zone where the risks of inappropriate therapy for sinus tachycardia or AF with a rapid ventricular response pose the greatest risk. Furthermore, in a large majority of patients, inappropriate therapy as observed in this study reflects the limitations of current ICD technology rather than device malfunction or failure. The problems associated with effective ICD therapy in these patients may well be an acceptable price to pay for the protection against sudden death, which the device offers. However, a smaller group of patients experience ICD-related problems without appropriate therapy. This factor will require consideration in the evaluation of clinical trials on the prophylactic use of ICD therapy in high-risk patients.
As with previous generations of devices, there are two broad categories of problems associated with current ICD therapy: those related to hardware complications and those related primarily to the limitations of tachyarrhythmia detection algorithms.18 19 Some hardware-related problems with rate-sensing and defibrillation leads have already been solved (for example, problems with noise from faulty Y-adaptors) and are no longer relevant. Similarly, the high incidence of lead dislodgment with the early use of nonthoracotomy lead systems seen in this and other series has been largely overcome by the use of better anchoring sleeves and the heightened awareness of the risks of displacement.6 20 These problems are likely, therefore, to arise less frequently in the future. However, at present, there are no data on the long-term durability of current transvenous lead systems. The early complications observed in this and other studies may be compounded by long-term integrity problems, as seen with previous transvenous lead systems.21 This concern is of particular importance in younger patients with well-preserved left ventricular function who may require defibrillator therapy for decades.
Another problem shared with earlier generations of devices is elevation of DFTs by antiarrhythmic therapy, particularly amiodarone. With the routine use of biphasic waveforms, this problem may be partially resolved.
The issue of inappropriate therapy for supraventricular arrhythmias remains one of the major problems associated with ICD therapy. The 26.6% rate of inappropriate shocks is similar to that reported for earlier generations of devices.2 3 Two of the three devices in this series incorporate an optional rate stability criterion as part of the programmable arrhythmia detection algorithm. The rate stability algorithm detects beat-to-beat variations in cycle length and was initially developed to distinguish regular supraventricular arrhythmias from AF. However, great caution must be exercised in using this algorithm to distinguish VT from AF with a rapid ventricular response, since VT may be associated with variability in cycle length, making safe and reliable differentiation between the two arrhythmias by rate stability alone difficult in some patients. For this reason and to ensure maximum sensitivity for VT detection, we did not use the rate stability function, and this accounts in part for the continuing high incidence of inappropriate therapies observed in this study. Bardy et al6 reported that only 3 of 84 patients followed for 11±7 months received inappropriate therapy for AF and that no patient had an undetected ventricular tachyarrhythmia when the rate stability criterion was used. The availability of ICDs with dual-chamber sensing capability will also undoubtedly improve the specificity of algorithms for VT and SVT discrimination.
Mode of Death
Of the six sudden deaths in this series, at least three were related to primary pump failure or ischemia. These sudden deaths in patients with poorly compensated left ventricular function suggest that sudden deterioration in hemodynamic state, mediated by bradyarrhythmia or tachyarrhythmia, although appropriately treated by the ICD, may still be fatal. It has been suggested that as many as 50% of sudden deaths in patients with advanced congestive heart failure may be due to bradyarrhythmias.22 23 The mechanisms of such bradyarrhythmias may be inherently different from the primary bradyarrhythmias associated with sick sinus syndrome or intermittent complete heart block, which can be effectively countered by cardiac pacing. Our observations suggest that even the most versatile ICD devices with bradycardia pacing do not uniformly prevent sudden cardiac death in patients with severely compromised left ventricular function. Since current ICDs store only events that trigger tachyarrhythmia but not bradyarrhythmia therapy, a more precise understanding of the modes of death in patients with ICDs will be gained only when these devices incorporate additional memory for storage of bradyarrhythmic events as well.
Reduction in Shocks With ATP
The success of ATP in reducing the need for ICD shocks in selected patients with frequently recurring VT in this study is striking. Without this function, implantation of an ICD in many of these patients would have been impractical. The high success rate (92.9%) of conversion of VT by ATP in this highly selected group of patients is virtually identical to the 88% to 92.4% reported from other centers.6 10 24 25 It is likely that these early reports of successful ATP also represent highly selected patient subgroups. Furthermore, this high percentage of success is due in part to a relatively small number of patients in whom numerous episodes of VT were reliably terminated by ATP. The observation that 62% of all patients had more than 90% of their VT episodes terminated by ATP may better reflect the overall effectiveness of antitachycardia pacing.
The patient population in this study represents a highly selected subgroup of patients in whom the results may not be applicable to the broad spectrum of ICD recipients. When third-generation devices first became available in limited numbers, their use was largely reserved for patients with severe left ventricular dysfunction and frequently recurring drug-resistant VT, which could be terminated by ATP or low-energy cardioversion. In addition, these devices were implanted as pulse generator replacements in patients who had already demonstrated problems with earlier-generation devices. Thus, the percentage of patients receiving inappropriate therapy for these conditions could be greater in this study than might be expected in an unselected group of patients receiving ICD therapy. Furthermore, the inappropriate shock rate might have been lower if the rate stability function, available in two of the three ICDs used, had been used although at the potential cost of nondetection of VT in this high-risk population.
Despite the increasing availability of telemetry and stored electrograms, it is not always possible to determine whether therapy was appropriate or inappropriate, especially when large numbers of events have occurred. In this study, therapy was classified as inappropriate only when the evidence was unequivocal. Therefore, in this patient group, as with previous studies, we may have overestimated the number of ventricular arrhythmia events effectively terminated by the device.
Third-generation ICDs offer significant advantages over previously available devices. In particular, ATP can be extremely effective in reducing the number of shocks delivered for frequently occurring VT. However, third-generation ICDs, including those with nonthoracotomy lead systems, are still associated with mechanical problems and inappropriate therapies. Given the extremely rapid evolution of ICD technology, it is likely that most of the technical difficulties associated with current ICD devices will be solved, leaving random component failure and operator error as the primary sources of future problems. Nevertheless, the results of this study underscore the complexity of these advanced devices and emphasize the need for additional clinical experience as well as prospective evaluation of different detection algorithms and treatment options to minimize the occurrence of inappropriate therapies while maintaining maximum sensitivity for the detection of VT and VF.
Dr Ruskin is a member of the Scientific Advisory Committee for Medtronic, Inc.
- Received September 22, 1994.
- Accepted November 20, 1994.
- Copyright © 1995 by American Heart Association
Kelly PA, Cannom DS, Garan H, Mirabal GS, Harthorne JW, Hurvitz RJ, Vlahakes GJ, Jacobs ML, Ilvento JP, Buckley MJ, Ruskin JN. The automatic implantable cardioverter-defibrillator: efficacy, complications and survival in patients with malignant ventricular arrhythmias. J Am Coll Cardiol. 1988;11:1278-1286.
Bardy GH, Hofer B, Johnson G, Kudenchuk PJ, Poole JE, Dolack L, Gleva M, Mitchell R, Kelso D. Implantable transvenous cardioverter-defibrillators. Circulation. 1993;87:1152-1168.
Yee R, Klein GJ, Leitch JW, Guiraudon GM, Guiraudon CM, Jones DL, Norris C. A permanent transvenous lead system for an implantable pacemaker cardioverter-defibrillator: nonthoracotomy approach to implantation. Circulation. 1992;85:196-204.
Jordaens L, Trouerbach J-W, Vertongen P, Herregods L, Polaert J, Van Nooten G. Experience of cardioverter-defibrillators inserted without thoracotomy: evaluation of transvenously inserted intracardiac leads alone or with a subcutaneous axillary patch. Br Heart J. 1993;69:14-19.
Fromer M, Brachmann J, Block M, Siebels J, Hoffmann E, Almendral J, Ohm OJ, den Dulk K, Coumel P, Camm AJ, Toubol P. Efficacy of automatic multimodal device therapy for ventricular tachyarrhythmias as delivered by a new implantable pacing cardioverter-defibrillator: results of a European multicenter study of 102 implants. Circulation. 1992;86:363-374.
Bardy GH, Troutman C, Poole JE, Kudenchuk PJ, Dolack GL, Johnson G, Hofer B. Clinical experience with a tiered therapy, multiprogrammable antiarrhythmia device. Circulation. 1992;85:1689-1698.
Marchlinski FE, Gottlieb CD, Sarter B, Finkle J, Hook B, Callans D, Schwartzman D. ICD data storage: value in arrhythmia management. PACE Pacing Clin Electrophysiol. 1993;16(pt II):527-534.
Bardy GH, Johnson G, Poole JE, Dolack L, Kudenchuk PJ, Kelso D, Mitchell R, Mehra R, Hofer B. A simplified, single-lead unipolar transvenous cardioversion-defibrillation system. Circulation. 1993;88:543-547.
Paul VE, Anderson M, Jones S, Skehan DJ, Camm AJ, Ward DE. Cardiologist implanted cardioverter defibrillators: early experience of three systems. Eur J Cardiac Pacing Electrophysiol. 1991;1:21-26.
Luu M, Stevenson WG, Stevenson LW, Baron K, Walden J. Diverse mechanisms of unexpected cardiac arrest in advanced heart failure. Circulation. 1989;80:1675-1680.
Packer M. Lack of relation between ventricular arrhythmias and sudden death in patients with chronic heart failure. Circulation. 1992;85(suppl I):I-50-I-56.