(Circulation. 1995;92:1651-1664.)
© 1995 American Heart Association, Inc.
Articles |
The Proarrhythmic Potential of Implantable Cardioverter-Defibrillators
From the Department of Cardiology, Cleveland Clinic Foundation, Cleveland, Ohio.
Correspondence to Sergio L. Pinski, MD, Cleveland Clinic Foundation, Department of Cardiology, Desk F15, 9500 Euclid Ave, Cleveland, OH 4419. Email pinskis@ccsmtp.ccf.org.
 |
Abstract |
|---|
 Top Abstract IntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
Abstract The implantable cardioverter-defibrillator (ICD) is remarkably effective in preventing sudden cardiac death in high-risk patients, but it also has the capacity to provoke or worsen cardiac arrhythmias. Tachyarrhythmias or bradyarrhythmias may result from the delivery of antitachycardia or antibradycardia therapies by tiered-therapy defibrillators. This proarrhythmia, although rarely fatal, increases the morbidity associated with defibrillator therapy. Proarrhythmia is related as much to suboptimal programming as to technical limitations of the device. The proarrhythmic potential of ICD therapy can be minimized by tailoring the "electrical prescription" according to characteristics of the clinical arrhythmia and individual ICD idiosyncrasies.
|
Introduction |
|---|
|
TopAbstractIntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
The concept of a therapeutic modality having the potential to cause effects directly opposite to those intended is neither new nor confined to the field of cardiovascular medicine. The aggravation or provocation of cardiac arrhythmias by an intervention intended to be antiarrhythmic is of particular concern because of its potentially life-threatening consequences. In recent years, better awareness of the proarrhythmic risks of pharmacological antiarrhythmic agents1 2 3 has fostered the development of nonpharmacological therapeutic modalities for ventricular tachyarrhythmias. The implantable cardioverter-defibrillator (ICD) is one such modality, which, since its clinical introduction in 1980,4 has evolved to a complex multiprogrammable device offering different tiers of therapy for ventricular tachyarrhythmias as well as pacing for bradycardias.5 The efficacy of the ICD in preventing sudden cardiac death in high-risk patients is well documented,6 7 but, like pharmacological agents, these devices can also induce or aggravate cardiac arrhythmias. Frequently, ICD-induced proarrhythmia is a consequence of inadequate device programming and not a manifestation of the intrinsic limitations of current ICD therapy. Therefore, careful programming is crucial in order to avoid this problem. In this review we will discuss the various forms of ICD-induced proarrhythmia with specific reference to their incidence, mechanisms, diagnosis, and management.
|
Classification |
|---|
|
TopAbstractIntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
The classification of ICD proarrhythmia can be based on a combination of the provoked arrhythmia and the clinical propriety and type of therapy delivered. Table 1
summarizes the various causes and mechanisms involved in ICD
proarrhythmia. This review is organized into two sections,
dealing first with ICD-induced tachyarrhythmias and then
with ICD-induced bradyarrhythmias. We define ICD-induced
proarrhythmia as an ICD therapeutic intervention that
results in the development of a new tachyarrhythmia or
bradyarrhythmia. At times, hemodynamically
significant new arrhythmias can result from a clinically
inappropriate absence of ICD intervention. Two examples of such a
"passive" proarrhythmia (deceleration of
ventricular tachyarrhythmias by
antitachycardia therapies and inappropriate inhibition of
bradycardia pacing caused by oversensing of T waves) are also
discussed, as they highlight the importance of optimum device
programming. A clinically inappropriate therapy is defined as one
delivered during a cardiac rhythm for which that therapy was not
intended (ie, rhythms other than sustained ventricular
tachycardia [VT] or ventricular fibrillation
[VF] or bradycardia below the programmed pacing rate). Although
exacerbation or provocation of ventricular and
supraventricular arrhythmias during the
postoperative period after ICD implant is relatively
common,8 9 10 these arrhythmias are not a
direct
result of device intervention and will not be discussed in this
review.
|
|
ICD-Induced Tachyarrhythmias |
|---|
|
TopAbstractIntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
Any of the therapies delivered by ICDs have the potential to induce tachyarrhythmias. When the device intervention resulting in proarrhythmia is clinically appropriate, proarrhythmia is a manifestation of the hazards inherently associated with antitachycardia therapies. Prevention of this form of proarrhythmia is based on tailoring antitachycardia therapies to the individual patient. If proarrhythmia occurs, minimization of its clinical sequelae is based on ensuring prompt and definitive backup defibrillation. On the other hand, prevention of proarrhythmia resulting from the delivery of clinically inappropriate therapies depends on the optimization of sensing parameters and detection algorithms of the device, the suppression of the triggering rhythms, or both.
|
Tachyarrhythmias Induced by Clinically Appropriate Therapies |
|---|
|
TopAbstractIntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
Acceleration of VT by Antitachycardia Pacing
Antitachycardia pacing is an effective treatment for sustained monomorphic VT and is an attractive alternative to the discomfort and battery draining effects of cardioversion shocks. However, acceleration of a previously tolerated VT and precipitation of VF are well-described complications of pacing therapy for VT11 and prohibited its widespread use before the advent of implantable devices with backup defibrillation capabilities.12
Several investigators have reported on the
incidence and predictors of
VT acceleration by antitachycardia pacing at the time of
electrophysiological testing both on a
per-episode and a per-patient basis (Table 2).
Acceleration occurs in 0% to 26% of episodes and in up to 43% of
patients.11 13 14 15 16 17 18 19
The discordance reflects the imperfectly
reproducible nature of this phenomenon. The efficacy and safety of
antitachycardia pacing techniques are inversely related. A
higher incidence of tachycardia termination can be achieved
with more aggressive protocols (ie, shorter initial and subsequent
coupling intervals, more extrastimuli per attempt, more attempts) but
often at the price of a high incidence of acceleration. Waldecker et
al,16 in a study of 215 episodes of induced VT in 100
patients, reported an acceleration incidence of 6%, 17%, and 36% and
an efficacy of 16%, 43%, and 54%, for single, double, and bursts of
extrastimuli, respectively. Randomized studies have not demonstrated
differences in the incidence of acceleration for ramp or burst pacing
algorithms of similar degree of
aggression.18 19 20 21 22
Acceleration is more likely to occur in tachycardias of shorter
and more variable cycle length.17 23 Antiarrhythmic
drugs may decrease the incidence of acceleration by slowing the
tachycardia rate.14 23 The risk of acceleration
appears to be independent of left ventricular ejection
fraction.17
|
Several studies suggest that spontaneous VTs
are less likely to be
accelerated by antitachycardia pacing than induced ones
(Table
2
).19 24 25 26
Leitch et al24 reported a
4.3% incidence of acceleration in 614 spontaneous episodes of VT in 5
of 15 patients. However, the seemingly more favorable acceleration
rates for spontaneous tachycardias may reflect selection bias
as antitachycardia pacing was activated only in
those patients in whom that therapy was effective at
electrophysiological testing. Gillis et
al19 compared the incidence of acceleration for induced
and spontaneous VT in patients with ICDs and reported incidences of 6%
and 1% of episodes, respectively. Similar findings were reported by
Siebels and Kuck.27 These differences in the incidence of
VT acceleration between acute testing and clinical follow-up may be
explained by differences in tachycardia characteristics
(induced VTs are generally faster), and changes in modulating factors
(eg, ischemia, autonomic tone) possibly triggered by repetitive
VT induction.
Several pathophysiological mechanisms can account for pacing-induced VT acceleration. In an in vitro model of reentrant VT in rabbit hearts, Brugada et al28 described the following mechanisms: induction of another wavefront in the same tachycardia circuit (double wave reentry), change to a functionally determined circuit (reentry around a functional line of block without involvement of a fixed obstacle), and change of the reentrant circuit to reentry within a different, faster anatomic pathway. The relative importance of these mechanisms in the acceleration of clinical VT is not clear. In a more clinically relevant model of "figure-of-eight" postinfarction reentrant ventricular tachycardia in the dog, El-Sherif and associates29 emphasized the role of continuing the antitachycardia pacing beyond the number of extrastimuli that terminate ventricular tachycardia in producing acceleration of VT. They showed that after tachycardia termination by the first few beats in a train, subsequent stimulated beats could induce new arcs of functional block and thus create different reentrant pathways. If the new circuit has a shorter revolution time, tachycardia acceleration and occasionally degeneration into VF could occur.
Complete avoidance of antitachycardia pacing proarrhythmia may demand a tradeoff between patient safety and comfort. Ideally, the pacing algorithm should be programmed to the least level of aggression that reliably terminates VT. The relative merits of programming the "electrical prescription" based on the results of predischarge electrophysiological testing versus initial empiric programming of a generic antitachycardia pacing algorithm in all patients are uncertain.27 In many implanting institutions, demonstration of VT termination without any instances of acceleration is mandatory for activation of outpatient antitachycardia pacing. With this approach, the incidence of antitachycardia pacing–induced acceleration of spontaneous VT can be minimized. However, because of the lack of reproducibility of acceleration and the more favorable ratio of efficacy to acceleration for spontaneous than for induced VT, this practice may be too restrictive, depriving some patients of the potential benefit of antitachycardia pacing. Furthermore, electrophysiological testing, even when performed noninvasively, is expensive and may need to be repeated during follow-up.30 Prospective trials are required to assess the relative efficacy, safety, and costs of these alternative strategies.
To minimize the potentially deleterious clinical effects of
VT
acceleration, it is important to program a safe backup defibrillation
output and to understand the response of each device to an accelerated
arrhythmia (see below). Prompt delivery of a high-output
shock after acceleration of VT will usually restore sinus rhythm before
the occurrence of syncope. Patients with antitachycardia
pacing should be assessed periodically for the development of VT
acceleration. This can be suspected by the occurrence of ICD shocks
preceded by palpitation and is generally confirmed by careful scrutiny
of data stored in the ICD memory, including therapy sequencing, RR
intervals recorded during the episode, and stored electrograms (Fig
1). The ease of diagnosis will vary among devices; some
provide data that is more comprehensive and easier to interpret than
others.26 31 Once acceleration is identified,
treatment
options include inactivation of antitachycardia pacing or
refinement of the pacing protocol either empirically or under
electrophysiological guidance.
|
Acceleration of VT by Cardioversion Shocks
Acceleration of VT
or degeneration to VF is relatively common for
low-energy cardioversion, which is variously defined as shocks of
as much as 2 to 5 J (Fig 2). In patients undergoing
ICD implantation, Winkle et al32 demonstrated
per-episode acceleration rates of 23%, 14%, 10%, and 15% for
shocks of 1 J, 5 J, 10 J, and 25 J, respectively. The incidence and
predictors of VT acceleration by low-energy cardioversion
have been best studied for arrhythmias induced during
electrophysiological testing; the incidence
of acceleration on a per-patient basis was 35% in one
study33 and on a per-episode basis varied from 6% to
31% (Table
3).13 15 32 34 35
Most
investigators have reported an increased risk of acceleration for
faster tachycardias. In one study, the risk of VT acceleration
was 2.2% and 9.8% for tachycardias with rates below and above
180 beats per minute, respectively.35 Acceleration also
appears to be more frequent in patients with more depressed left
ventricular function or higher cardioversion
thresholds.33 In a randomized crossover study, Bardy et
al36 found similar acceleration rates of inducible VT for
low-energy cardioversion and antitachycardia pacing
(17% versus 21%, respectively). Interestingly, an acceleration
response to one therapy did not predict the same response to the other.
Fewer data exist regarding the frequency of acceleration of spontaneous
VT by low-energy cardioversion in patients with implanted
antitachycardia devices (Table 3). In small series
including selected patients, acceleration occurs in 4.7% of
episodes37 and in 15% to 18% of
patients.12 37 38
|
|
The mechanisms by which low-energy cardioversion results in VT acceleration or degeneration are not clearly defined. Shocks that create potential gradient fields of a critical strength in tissues with a critical degree of refractoriness may result in circulating wave fronts of ventricular activation that manifest as ventricular tachyarrhythmias.39 Shock delivery synchronized with the QRS complex is designed to reduce that risk. Synchronization of cardioversion by ICDs is based on the local electrogram. As the timing of the local electrogram relative to global ventricular depolarization varies depending on the site of VT origin and the sensing lead(s) position,40 a "synchronized" shock delivered to hearts with a wide dispersion of refractoriness and nonuniform conduction properties may encounter areas of myocardium in their vulnerable period. This is more likely during very rapid VTs as the repolarization phase of one beat may extend into the depolarization phase of the subsequent beat.41 The optimal timing of a cardioversion shock in relation to the local electrogram and surface QRS complex is not known. Waspe et al13 did not find a relationship between the timing of shock delivery and occurrence of VT acceleration. However, using a crossover design, Li et al42 found that VT was accelerated by internal cardioversion shocks in 87% of patients when shocks were delivered simultaneously with QRS onset but only in 20% when shocks were delivered 100 ms into the QRS complex. In future devices, the ability to control the timing of shock delivery relative to the sensed ventricular electrogram may decrease the incidence of VT acceleration by cardioversion shocks. Another mechanism by which very low energies (<0.5 J) may result in tachycardia acceleration is the creation of conduction delay or block in areas of the reentrant circuit with subsequent development of an alternative shorter pathway.43
The diagnosis and management of cardioversion-induced acceleration is similar to that of pacing-induced acceleration. In both circumstances, a sound knowledge of the response of the ICD in question to acceleration is helpful in treating this problem.
Device Response to Acceleration of Ventricular
Tachyarrhythmias
Acceleration generally results in a tachycardia rate
that
falls in a faster "therapeutic zone," with subsequent delivery of
an appropriately more aggressive therapy. However, if the programmed VT
zone is wide, acceleration can result in a faster tachycardia
in the same zone. This could result in further futile attempts at
antitachycardia pacing and potential
hemodynamic compromise. Some ICDs incorporate features
designed to prevent this scenario. The PRx allows the programming of an
"acceleration percentage"; if the rate change exceeds the
programmed percentage, maximum energy shocks are delivered even when
the tachycardia remains in the same zone. In the PCD and Jewel
devices, VT is clas- sified as accelerated when its cycle length at
redetection has decreased by 
60 ms irrespective of the "zone."
If acceleration occurs after the delivery of
antitachycardia pacing, subsequent programmed pacing
sequences are interrupted and the next level of programmed therapy is
delivered instead.
Deceleration of Ventricular
Tachyarrhythmias by Antitachycardia
Therapies
Infrequently, antitachycardia pacing fails to
terminate VT but instead results in deceleration below the device
cutoff rate. The exact incidence and mechanisms for this phenomenon are
unclear. The slowed VT may have the same morphology as the original
one, suggesting a change in the
electrophysiological properties of the same
circuit. Alternatively, the new VT may have a different morphology,
suggesting termination of the original one by the extrastimuli with
subsequent manifestation of a different circuit. Deceleration is
usually a transient phenomenon but occasionally can be pronounced and
persistent enough to satisfy the algorithm for sinus rhythm
redetection, thus inhibiting further device intervention. We have
studied patients in whom, at
electrophysiological testing, the delivery
of antitachycardia pacing consistently resulted in
a persistent slow VT that was hemodynamically
deleterious (Fig 3). In other patients, the slower VT
reaccelerates to its baseline rate after redetection of sinus rhythm
has been satisfied. The ICD then interprets the VT as a new episode,
preventing appropriate progression in the therapeutic algorithm.
|
A
similar phenomenon may occur in patients in whom a shock results in
the deceleration of a ventricular
tachyarrhythmia to one that falls in a slower
"therapeutic zone." Devices differ in their response to this
phenomenon. For example, the Cadence and the Guardian 4215 never
"step down" (ie, will never deliver antitachycardia
pacing during an arrhythmia episode in which shocks were
previously delivered), whereas the type of response can be programmed
in the Res-Q (ratchet function). A "step-down" in the level
of response could compromise the safety of ICD therapy by delaying the
delivery of definitive shock therapy. The step-down algorithm in
the PCD is potentially deleterious, as VT detection is suspended for 64
sensed events after delivery of defibrillation therapy. This long time
to redetect VT after a defibrillation shock may result in
hemodynamic compromise (Fig 4).44 This
idiosyncrasy was improved in the
Jewel by shortening the suspension of therapy to 17 sensed events after
a defibrillation shock.
|
Induction of Supraventricular
Tachyarrhythmias by Therapies Delivered for
Ventricular Tachyarrhythmia
Shocks for VT or VF are delivered
regardless of the phase of the
atrial electrical cycle, so that atrial fibrillation or flutter may be
precipitated when delivery of energy occurs during the atrial
"vulnerable period" (Fig 5). The incidence is
relatively high (15% to 20%) when using low-energy cardioversion
shocks delivered via temporary defibrillation
catheters.13 15 Cardioversion shocks delivered via
permanently implanted systems result in the induction of atrial
fibrillation in up to 6% of episodes and 29% of
patients.45 46 47 Most episodes are
transient, but they may
be sustained in patients with the substrate to support atrial
fibrillation or flutter.
|
The factors associated with the induction of
atrial arrhythmias
by cardioversion shocks have not been well studied. In animal studies,
it appeared to be more frequent for shocks greater than 0.5 J, but the
entire range of clinically relevant energies was not
investigated.41 In one clinical study, high-energy
shocks (>15 J) were more likely to induce atrial fibrillation than
low- or intermediate-energy shocks.46 In another
study, however, only shocks 
3 J resulted in the induction of atrial
fibrillation.47 There are scarce data comparing the
incidence of induction of atrial fibrillation or flutter by different
lead systems and current pathways. Minimization of atrial involvement
in the current pathway would intuitively tend to decrease the chance of
atrial arrhythmia induction.48 Two nonrandomized
studies reported conflicting findings. In one, induction of atrial
fibrillation was more frequent for shocks delivered via epicardial than
nonthoracotomy defibrillator systems,46 whereas in the
other, atrial fibrillation occurred only when low-energy shocks
were delivered via an electrode system that included a right atrial
coil.47
Induction of atrial fibrillation with a rapid
ventricular
response by cardioversion shocks is a particularly undesirable outcome.
If the ICD is unable to discriminate between the
supraventricular arrhythmia and VT, the therapy
algorithm will continue, resulting in inappropriate shocks (Fig
6). The case of monitored sudden death reported by
Birgersdotter-Green et al49 best exemplifies the
potentially deleterious consequences of such an event. A 5-J
cardioversion shock induced sustained atrial fibrillation. The
ventricular response to atrial fibrillation increased above
the detection rate, triggering the delivery of VT and VF therapies.
Atrial fibrillation persisted after all the available therapies were
exhausted. The patient developed ventricular flutter
shortly thereafter, but the ICD did not respond because it was in the
"dormant" state after delivery of all programmed therapies. This
resulted in a fatal outcome.
|
|
Tachyarrhythmias Induced by Clinically Inappropriate Therapies |
|---|
|
TopAbstractIntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
Mechanisms of Proarrhythmic Effects of Clinically Inappropriate Antitachycardia Therapies
The proarrhythmic risks of clinically inappropriate antitachycardia therapies vary according to the type of therapy delivered, the underlying rhythm, and patient idiosyncrasies (Table 4
).6 24 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
In general, the risks
are higher for antitachycardia pacing and low-energy
shocks than for high-energy shocks. Antitachycardia
pacing is most often used in patients with sustained monomorphic VT
occurring in the setting of chronic coronary artery disease.
These VTs are generally reproducibly initiated by programmed
ventricular stimulation. Therefore, it is not surprising
that sequences of antitachycardia pacing are remarkably
effective in inducing VT when delivered during
supraventricular rhythms, acting as a "very
aggressive" programmed ventricular stimulation
protocol.
|
It is well known that an electrical stimulus delivered during the vulnerable period of ventricular repolarization may induce VF.66 The upper limit of vulnerability hypothesis helps explain the effects of shocks delivered during normal rhythms. According to this hypothesis, there is a range of energies that will induce VF, with the lower and upper limits being called the VF threshold67 and the "upper limit of vulnerability,"68 respectively. Limited clinical data suggest that for any given patient and defibrillation system, the upper limit of vulnerability is similar (slightly lower) to the defibrillation threshold.69 Consequently, high-energy shocks delivered inappropriately during sinus rhythm in the absence of factors associated with sympathetic discharge or cardiac ischemia are rarely arrhythmogenic. To date, all reported cases have involved misdirected shocks delivered during magnet testing of the original CPI device.6 59 On the other hand, proarrhythmic effects of ICD shocks delivered during sinus tachycardia and other supraventricular arrhythmias are more common. This increased susceptibility to proarrhythmia during faster rhythms may be related to one or more of the accompanying factors: a concomitant increase in sympathetic discharge,70 rate-related decreases in ventricular refractoriness, or provocation of myocardial ischemia,71 all of which are associated with an increased vulnerability to electrical arrhythmogenesis. Similarly, shocks in sinus rhythm occurring as a result of a committed device's response to nonsustained VT can be proarrhythmic. One can postulate that the myocardial and autonomic milieu responsible for the generation of the nonsustained VT and the increased dispersion of refractoriness that accompanies the premature beats67 result in an increased vulnerability to shock-induced proarrhythmia.
Delivery of Inappropriate Antitachycardia Therapies
Caused by Failure to Differentiate Supraventricular
From Ventricular Arrhythmias
ICDs are designed to detect and treat
life-threatening
ventricular tachyarrhythmias with maximum
sensitivity. In all devices, tachycardia recognition is based
on the measurement of heart rate. Assuming correct sensing of the
cardiac depolarization signal, this ensures 100% sensitivity for
tachycardias above the programmed cutoff rate. However, the
major drawback of rate-only detection is its poor specificity.
Failure to differentiate between supraventricular (most
frequently atrial fibrillation with a rapid ventricular
response or sinus tachycardia) and ventricular
arrhythmias is the most frequent cause of spurious ICD
therapies. Many of these spurious interventions can be avoided by
adequate device programming. The potential for the occurrence of
clinically significant supraventricular
tachyarrhythmias should be assessed in each patient and the
device programmed accordingly. When using exclusively or predominantly
"shock-only" ICDs, the incidence of these events during
long-term follow-up has ranged between 16% and
21%.55 72 With earlier devices, the incidence may
have
been underestimated because of the lack of data logging and electrogram
storage capabilities.55 The problem is of growing
importance, as the greater versatility of tiered-therapy ICDs has
expanded their use to patients with relatively slow and
well-tolerated VTs, whose rates are more likely to overlap with
those of supraventricular origin. In those patients,
supraventricular arrhythmias will frequently
trigger the delivery of antitachycardia pacing or
low-energy cardioversion, which, as explained above, have more
potential for arrhythmia induction than high energy shocks.
Schmitt et al64 analyzed the significance of
supraventricular tachyarrhythmias in 86
patients implanted with a tiered-therapy ICD. During follow-up,
inappropriate therapy delivery for supraventricular
tachyarrhythmias was documented in 16% of patients and was
suspected in an additional 21%. In only 35% of the patients who
received documented or suspected spurious therapies for atrial
fibrillation, had this arrhythmia been present before
implant.
To avoid false detection of sinus tachycardia, the cutoff rate for tachycardia detection should ideally be programmed above the patient's maximal exercise-induced sinus rate. However, in many patients (more frequently in those on antiarrhythmic drugs), the sinus and VT rates overlap. In a study of 108 patients, Paul et al73 showed that in the absence of antiarrhythmic drugs, only 11% had sinus rates exceeding the rate of VT, but this proportion increased to 35% in patients on single-drug antiarrhythmic regimens and to 63% in patients on combination antiarrhythmic drug therapy. On the other hand, ß-adrenergic blockade had a favorable effect by blunting the sinus rate without affecting the rate of VT. Control of the ventricular response should be the mainstay in the prevention of inappropriate therapy for chronic atrial fibrillation. For patients with paroxysmal atrial fibrillation, combined therapy to prevent recurrences of paroxysmal atrial fibrillation and to slow AV conduction if paroxysms do occur may be required.74 Catheter ablation of the AV junction with pacemaker implantation may be considered in recalcitrant cases.75
Detection "enhancements" have been incorporated into tiered-therapy ICDs to improve the accuracy of automatic arrhythmia diagnosis. The sudden onset criterion helps to discriminate between sinus tachycardia and VT,76 while the rate stability criterion is useful in differentiating atrial fibrillation with a rapid ventricular response from monomorphic VT.77 The incorporation of these enhancements into the detection algorithms represents a tradeoff between sensitivity and specificity because an algorithm with perfect specificity may fail to detect some episodes of hemodynamically significant VT. Few data exist concerning the diagnostic accuracy and optimal settings for enhancement criteria. Furthermore, the performance of the different implementation of similar enhancing criteria among various devices has not been analyzed. Physicians must be aware of the values and limitations of each added detection criterion. Their implementation can only be recommended in selected patients who have demonstrated inappropriate device interventions for supraventricular tachyarrhythmias or who are at high risk for this complication. Furthermore, the fact that induced arrhythmias may interact with these detection enhancements in a markedly different way than spontaneous ones must be taken into account. Therefore, successful performance during electrophysiological testing cannot guarantee successful performance in real-life situations.
In a well-designed study, Swerdlow et al78 prospectively evaluated the rate stability and sudden onset criteria in 100 patients with PCD defibrillators at electrophysiological study, exercise testing, and clinical follow-up. The stability criterion of 40 ms afforded optimal discrimination between VT and atrial fibrillation; it allowed correct detection with minimal delay of all episodes of spontaneous or induced VT while rejecting 96% of spontaneous episodes of paroxysmal atrial fibrillation and 99% of the spontaneous episodes of chronic atrial fibrillation with rates above 120 beats per minute. The sudden onset criterion had a less impressive performance; the optimum sudden onset ratio of 87% rejected sinus tachycardia 98% of the time, but failed to detect 0.5% of spontaneous VTs. These VTs gradually accelerated above the cutoff rate or arose during periods of sinus tachycardia. These settings should be considered as useful starting points if the enhancement criteria are to be enabled but should not be interpreted as universal recommendations. Other devices (eg, PRx, Res-Q) allow the programming of a "sustained high rate" criterion, which overrides the other enhancement criteria and will allow the delivery of therapies for a tachycardia episode that has not fulfilled one of the enhancement criteria but has persisted for the programmed number of intervals of "sustained high rate."
Several non–rate-related algorithms, incorporating analysis of cardiac activation sequence, electrogram characteristics, and hemodynamic parameters, have been proposed to increase the specificity of arrhythmia detection by ICDs.79 These approaches may ultimately prove to have discriminative performance superior to that of conventional algorithms. Currently, only a morphology criterion, the probability density function (PDF), is available in CPI devices. Although theoretically useful, studies have shown that its overall accuracy is low. Therefore, if it is intended to be used, PDF should be tested before implementation.80 81
Delivery of Inappropriate Therapies for Nonsustained
Ventricular Arrhythmias by Defibrillators With a
Committed Response
ICDs differ in their mode of response to a detected
tachyarrhythmia. Committed devices always deliver therapy
once detection criteria are met without taking a "second look"
after capacitor charge, whereas noncommitted devices discharge only if
the presence of tachycardia is reconfirmed just before
delivery. Committed behavior often results in the delivery of shocks
during sinus rhythm in response to nonsustained VT of sufficient
duration to satisfy the detection algorithm.82 Several
investigators documented proarrhythmic events resulting from this
committed behavior.54 59 60 Treatment
options include
pharmacological suppression of nonsustained VT, reprogramming to a
noncommitted mode if possible, or extension of the detection time in
strictly committed devices. This last maneuver, however, will also
delay the delivery of therapy for sustained
tachyarrhythmias, with potential
hemodynamic compromise.
Delivery of Inappropriate Antitachycardia Therapies
Caused by Oversensing of Signals
The detection of VF is a technically
demanding process. The ICD
system must identify signals of low and continuously varying amplitude
that are generated by this rhythm and at the same time reject
extraneous noise. Furthermore, avoidance of T wave sensing must be
accomplished despite a sensing refractory period that must be short
enough to detect tachyarrhythmias.83 The most
common signals that can be oversensed by ICDs include electronic
"noise" generated by structural defects or loose connections in
the sensing system, T waves, and pacing artifacts from a separate
pacemaker. The consequences, if any, of such signals will depend on the
duration of the episodes of oversensing (ranging from very transient to
permanent), the frequency content of the oversensed signals, the
response mode of the device (committed or noncommitted), and the
programmed detection and treatment parameters. The
documentation of delivery of antitachycardia therapies for
rhythms below the cutoff rate of the ICD is diagnostic of
oversensing. Oversensing can also be demonstrated by analysis
of real-time electrograms, sensing marker channels, or beeping tones,
depending on the device. When intermittent malfunction is suspected,
analysis should be repeated during the use of various muscle
groups and generator or lead manipulation.
Despite advances in ICD system design and manufacture, conductor or insulation breakdown in the sensing lead(s), failure of sensing lead adapters, and loose-set screws remain common causes of oversensing.84 85 86 In a series of 241 ICD patients, electrical noise accounted for inappropriate therapies in 7.55 Electrical noise is frequently intermittent so that in patients with noncommitted ICDs, shocks for electrical noise are usually aborted.82 87 Even when delivered, inappropriate antitachycardia therapies triggered by electronic noise are rarely proarrhythmic for several reasons. First, electrical noise has a high frequency response. Thus, it is almost always detected in the VF zone, triggering high-energy shocks with energies in excess of the upper limit of vulnerability. Second, when they occur during periods of normal sinus rates, the chances of a random shock falling in the vulnerable period are reduced. Third, other arrhythmogenic influences such as ischemia or increased sympathetic tone are unlikely to be present concomitantly.
Oversensing of T waves is an unusual cause
for inappropriate delivery
of antitachycardia therapies. T wave sensing occurs when
the amplitude of the T wave exceeds the sensing threshold and is more
common with devices that utilize a "lock on gain" amplifier
system.50 88 The resulting double counting will
trigger
the delivery of inappropriate therapy if the spontaneous rate is at
least half the cutoff rate. Oversensing of the T wave of spontaneous
beats by ICDs cannot be prevented by lengthening the sensing refractory
period, which in order to permit detection of VF is short (
150 ms),
and generally nonprogrammable. With some devices (PCD, Jewel, Res-Q),
the maximum sensitivity can be decreased to overcome this problem, but
this mandates retesting to ensure adequate detection of VF. The use of
negative chronotropic agents or an increase in the detect rate to
prevent the heart rate from satisfying the detection algorithm during
persistent T wave oversensing are less desirable alternative
options.
The availability of backup VVI pacing in newer ICDs has not abolished the need for the concomitant use of permanent pacemakers in some patients.89 The multiple potential deleterious interactions between pacemakers and ICDs have been extensively reviewed.90 Specifically, double or triple counting (pacing artifact[s] and the evoked ventricular electrogram) during a paced rhythm can result in fulfillment of the detection algorithms with delivery of inappropriate antitachycardia therapy. Proarrhythmia aside, sensing of pacemaker stimuli by ICDs may present a more ominous problem if, during ventricular fibrillation, these signals inhibit arrhythmia detection and result in failure of appropriate device intervention. Counting of pacemaker artifacts may be avoided by appropriate planning and testing at implantation. Pacemaker leads should be placed as far apart as possible from the rate-sensing bipole of the ICD. Problematic double counting occurring during follow-up generally can be overcome by reprogramming the pacemaker to its minimum safe output or to a maximum rate less than half of the ICD rate cutoff. Switching to AAI(R) mode should be considered in patients with reliable AV conduction. If troublesome oversensing persists, pacemaker lead repositioning will usually solve the problem.
Induction of Ventricular Tachyarrhythmias
by Inappropriate Antibradycardia Pacing
Competition between an
asynchronous pacemaker and the spontaneous
cardiac rhythm may induce ventricular
tachyarrhythmias if the pacemaker stimulus captures the
ventricle during its vulnerable period.91 This occurrence
is extremely uncommon in patients with antibradycardia pacemakers as
attested by the routine use of magnets to check pacemaker function.
However, patients with ICDs may be more vulnerable to the effects of
asynchronous pacing. Induction of sustained monomorphic VT by the
resulting long-short stimulation sequences may then occur in
patients with the appropriate substrate to support such an
arrhythmia (Fig 7).
|
Transient sensing failure during sinus rhythm usually results from spontaneous variation in signal amplitude. Frequently, the amplitude of the ventricular electrogram of a sinus beat after a premature beat is often significantly lower than that of the preceding sinus beats.92 In tiered-therapy ICDs with "co-dependent" sensing, transient sensing failure during sinus rhythm can also result from an abrupt change in the amplifier gain setting.93 The sense amplifier adapts continuously to the characteristics of the incoming signal by adjusting either the gain setting or the sensing threshold. These adjustments are performed either on a cycle-by-cycle basis (eg, PCD, Jewel) or after averaging a fixed number of preceding signals (Cadence). The latter, "slower reacting" strategy is more likely to lead to a transient inability to sense the next sinus signal after a series of large amplitude signals (eg, nonsustained VT), which triggers a decrease in the gain to prevent amplifier saturation. With either mechanism of undersensing, a pacing stimulus is delivered at the programmed escape interval after the last large amplitude beat.
The prevention of inappropriate antibradycardia pacing depends on the underlying mechanism and the sensing characteristics of the device in question. When it occurs in patients with ICDs with separate amplifiers for bradycardia and tachycardia sensing (eg, PRx), it can be solved by increasing the sensitivity in the bradycardia channel.63 It is more difficult to circumvent when it occurs in patients with ICDs with "co-dependent" sensing (eg, Cadence). Increases in the bradycardia pacing rate aimed at avoiding duplication of the coupling interval sequences that resulted in arrhythmia induction may be effective.50 In some patients the antibradycardia pacing function may need to be inactivated (possible with Cadence, Guardian 4215, Jewel, PRx), but this would obviously present another problem in patients with postshock bradyarrhythmias. However, some devices (PRx, Guardian 4215) allow the selective programming of antibradycardia pacing immediately after a shock, while the function is disabled at other times.
|
ICD-Induced Bradyarrhythmias |
|---|
|
TopAbstractIntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
Postshock Bradyarrhythmias
It is widely appreciated that external defibrillation may be complicated by bradyarrhythmias.94 The potential occurrence of marked bradycardia or prolonged asystole after ICD shocks has been a matter of concern, as hemodynamic collapse or recurrence of VT or VF may result. There is controversy regarding the incidence of significant bradycardia after internal defibrillation. Niazi et al95 analyzed the incidence of bradycardia after 157 shocks in 50 patients at time of epicardial defibrillator implant and found only two instances of prolonged asystole. The average duration of postdefibrillation pauses was 1.2 seconds. The first postdefibrillation spontaneous QRS occurred from a sinus mechanism in 49% of cases. No preoperative clinical variables correlated with the duration of postdefibrillation asystole. On the other hand, Ciccone et al34 reported a 23% incidence of bradyarrhythmias after delivery of low-energy transvenous cardioversion in patients with VT. Other investigators have implicated postdefibrillation bradyarrhythmias as the cause of cardiac arrest in patients with ICDs without backup pacing.96 In at least one case, there was evidence that a single appropriate ICD shock resulted in complete heart block requiring resuscitation.97
The mechanisms responsible for shock-induced bradycardias are not well understood. Some studies suggest a direct depressant effect of high current density fields, but their extrapolation to the clinical setting is dubious because the high potential gradients involved are achieved only within a short distance of the defibrillating electrodes in patients (and unlikely to be present in close proximity to the conduction system). In cultured chick embryo myocardial cells there is a direct correlation between the strength of the electrical discharge and the duration of postshock asystole.98 In animal models, transient conduction block can be demonstrated immediately after defibrillation shocks in areas close to the defibrillation electrodes and with high current density electric fields.99 The depression of automaticity in cultured cells100 and the localized conduction block99 are more significant for monophasic than biphasic waveform shocks of similar magnitude. It is not known if postdefibrillation bradyarrhythmias are less common in humans after biphasic shocks. Interestingly, bradyarrhythmias are almost never seen when high-energy defibrillation shocks are applied during normal rhythm, suggesting that the tachyarrhythmia being converted (probably via the induction of ischemia) is responsible for the postshock bradyarrhythmia. Internal shocks have been reported to occasionally convert chronic atrial fibrillation.34 This could result in prolonged asystole in patients with underlying sick sinus syndrome.101
It is clear that backup pacing capability is desirable (cost considerations aside) for most patients with ICDs. To avoid uncommon but potentially life-threatening postshock bradyarrhythmias, backup ventricular pacing should in general be activated at a relatively slow rate (40 to 50 beats per minute) and high output (see next section) in all patients, regardless of their prior occurrence during implant or testing. It should be noted that with some devices (eg, Cadence, Res-Q), a slow pacing rate can occasionally prolong the detection of ventricular fibrillation of very low amplitude by extending the time required to adjust the sensing automatic gain control level to the electrogram amplitude. One may speculate that backup antibradycardia pacing will be a standard feature in all available ICDs in the near future.
Postshock Increase in Pacing Thresholds Leading to
Ventricular Noncapture
The pacing threshold is frequently increased
transiently after the
delivery of defibrillation shocks. Occasionally, this can lead to
clinically relevant loss of capture and severe bradycardia or
asystole.102 Animal and human studies suggest that higher
shock energies,102 concomitant administration of class I
antiarrhythmic drugs,103 and longer duration of the
arrhythmia before shock therapy104 tend to
exacerbate the increase in pacing threshold. Most studies have
addressed the effects of internal defibrillation shocks on the pacing
threshold of a separate pacemaker. The effects on the pacing threshold
of shocks from tiered-therapy ICDs in which the same lead is used
for shocking and pacing are less well defined. Cohen et
al105 found transient lack of capture (lasting 4.9±5.1
seconds) after 8 of 22 episodes of internal defibrillation in patients.
Calkins et al106 reported that 4 of 30 patients (13%)
with ICDs and separate pacemakers developed transient loss of capture
after internal shocks. In one patient, failure to capture was observed
up to 56 seconds after ICD discharge. In a study in dogs assessing
endocardial and epicardial pacing after high-energy shocks, the
duration of capture loss was current dependent and significantly
increased by the prior administration of flecainide.103 In
contrast, Khastgir et al107 reported that in patients
undergoing implant of an epicardial ICD system there were no
significant increments in the ventricular pacing threshold
10 and 60 seconds after delivery of a 20-J defibrillation shock. Time
to capture at an output 1.1 times threshold was short (<2 seconds in
all 10 patients tested). Chronic administration of amiodarone
did not influence the results. Using a tripolar catheter with similar
design to the current Endotak system, Winkle et al108
could not demonstrate a change in pacing threshold after transvenous
defibrillation shocks.
Transient changes in the tissues interfacing with the pacing electrodes (probably induced by partial shunting of the high current delivered by the ICD) have been implicated as responsible for the increase in pacing threshold105 On rare occasions, an ICD shock can result in the dislodgment of a freshly implanted ventricular pacing lead. It has been suggested that the use of active fixation leads in patients with ICDs and concomitant pacemakers may minimize this occurrence.109
From the previous studies, it can be
concluded that high-output
stimuli may have to be delivered for the first few beats after
defibrillation to ensure capture. Clinically, the possibility of loss
of capture after a shock can be minimized by programming a wide safety
margin in the pacing output. This is especially important in
pacemaker-dependent patients or in those with marked postshock
bradycardia. In pacemaker-dependent patients, programming of high
outputs will compromise device longevity. Such patients will be better
served by the use of two separate devices with pacing capabilities. The
more versatile permanent pacemaker should be programmed with standard
settings, whereas backup pacing from the ICD can be programmed at a
lower rate and very high output (Fig 8). Alternatively,
some ICDs (PRx, Guardian 4215, Res-Q) allow programming of a higher
pacing output for variable periods of time after shock
delivery.
|
Shock-Induced Reset of a Separate Pacemaker Resulting in
Bradycardia
The strong electromagnetic field induced by a
defibrillation shock
may reset DDD(R) and VVIR pacemakers to the VVI or VOO mode. The
pacemaker will operate in the reset mode until reprogrammed. Pacemakers
store programmed parameters in volatile memory susceptible
to the influence of electromagnetic interference. To avoid erratic
behavior when communication between the internal microprocessor and the
memory location is interrupted by electromagnetic interference, the
pulse generator converts automatically to a mode
("power-on-reset" mode) whose instructions are stored in
nonvolatile (read-only) memory.110 The possible
deleterious effects of the power-on-reset mode in the pacing
polarity have received most attention. In most pulse generators with
programmable electrode configuration, the reset mode is unipolar, in
order to ensure continuous pacing regardless of the polarity of the
attached lead. In patients with ICDs, this could result in inhibition
of detection of VF by the large unipolar pacemaker
artifacts.111 The potential for this complication can be
minimized by the use of "bipolar committed" pacemakers (ie, those
that maintain the bipolar pacing configuration even in the reset mode)
in patients with ICDs. The effects of reset on the pacing rate are less
well appreciated. With most pacemakers, reset will result in pacing in
the VOO or VVI mode at relatively slow rates (Fig 9). In
selected patients, this may result in hemodynamic
compromise or pacemaker syndrome.112 Although the
occurrence of reset of separate pacemakers by internal defibrillator
shocks has been low with epicardial ICD
systems,106 113 it
may be more common with transvenous ones. This uncommon interaction may
be difficult to predict and prevent. The sensitivity of pacemakers to
electromagnetic interference varies among models and manufacturers, but
this has not been systematically studied. Implanting both lead systems
as far apart as possible and programming a lower energy shock could be
useful in preventing this interaction. The possibility of pacemaker
reset by ICD shocks should be explored at the time of implant and
predischarge testing in all patients with separate devices. In patients
prone to this interaction, the function of the pacemaker should be
assessed by transtelephonic monitoring as soon as possible after an ICD
shock. Prompt documentation of pacemaker reset will minimize the time
the patient is inadvertently paced in a less
physiological mode.
|
Inappropriate Inhibition of Bradycardia Pacing From a
Tiered-Therapy ICD Due to Oversensing of T Waves
Oversensing of T
waves after paced beats is relatively frequent
with tiered-therapy ICDs.50 The addition of
bradycardia pacing in these devices presents seemingly unavoidable
contradictions. For example, in the absence of sensed complexes, two
potentially life-threatening diagnoses must be considered: asystole
requiring bradycardia pacing and fine VF requiring amplifier gain
adjustments for proper detection. In tiered-therapy ICDs, the
sensing function is transiently suspended (blanking) after pacing.
During intermittent ventricular pacing, the amplifier can
still adjust its sensitivity to the amplitude of the spontaneous R
waves, but during continuous pacing only the local afterdepolarization
and T wave can be sensed. Usually, the sense amplifier adjusts to these
low-amplitude events by operating at maximal gain or minimal
threshold. If this amplification results in inappropriate sensing of
the T wave as a separate event, the delivery of the next pacing
stimulus is inhibited and the effective pacing escape interval is thus
lengthened. In pacing-dependent patients, this could result in
symptomatic bradycardias or pauses. Wide pacing pulse
widths, high pacing amplitudes, and high sensitivities all promote
inappropriate sensing after pacing. In general, oversensing of the
paced T wave can be circumvented by programming a longer postpace
refractory period, which is independently programmable in ICDs (Fig
10). However, a long blanking period after a paced
event (>350 ms) coupled with a fast pacing rate may impair the
detection of ventricular tachyarrhythmias. This
can be avoided by ensuring that (1) the difference between the
bradycardia escape interval and the blanking period is greater than the
tachyarrhythmia detection interval or (2) the bradycardia
escape interval is greater than twice the blanking period.
Alternatively, programming a narrower pulse width, a lower amplitude
pulse, or a higher sense threshold (eg, PCD) may eliminate
inappropriate T wave sensing. It is not recommended to program the
minimal sensitivity >0.6 mV, as this could hinder the speed and
accuracy of VF detection. If sensitivity is decreased to eliminate
inappropriate sensing, the capability of the ICD to detect induced VF
should be reassessed.
|
|
Conclusions |
|---|
|
TopAbstractIntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
ICDs are associated with a small but significant risk of proarrhythmia. ICD-induced proarrhythmic events are rarely fatal but increase the morbidity associated with device therapy. Most causes of the phenomenon are avoidable. Improved diagnostic capabilities in newer devices facilitate its recognition. Effective treatment depends on a firm understanding of ICD technology and individual device idiosyncrasies as well as knowledge of the arrhythmic profile of the individual patient. As devices continue to evolve to become more flexible and complex, and with dual-chamber defibrillators on the horizon, it is likely that a greater potential for proarrhythmia will exist. We hope that this will be countered by heightened physician awareness and understanding so that the full clinical benefits of device therapy in patients with arrhythmias can be realized.
Received February 9, 1995; revision received April 18, 1995; accepted April 18, 1995.
|
References |
|---|
|
TopAbstractIntroductionClassification ICD-Induced Tachyarrhythmias Tachyarrhythmias Induced by...Tachyarrhythmias Induced by...ICD-Induced Bradyarrhythmias Conclusions References |
|---|
-
Velebit V, Podrid P, Lown B, Cohen BH, Graboys
BH. Aggravation and provocation of ventricular
arrhythmias by antiarrhythmic drugs.
Circulation. 1982;65:886-894.
[Abstract/Free Full Text]
-
The Cardiac Arrhythmia Suppression Trial
(CAST) Investigators. Effect of encainide and flecainide on
mortality in a randomized trial of arrhythmia suppression after
myocardial infarction. N Engl J Med. 1989;321:406-412. [Abstract]
-
Roden DM. Risks and benefits of
antiarrhythmic therapy. N Engl J Med. 1994;331:785-791.
[Free Full Text]
-
Mirowski M, Reid PR, Mower MM, Watkins L, Gott VL,
Schauble JF, Langer A, Heilman MS, Kolenik SA, Fischell RE, Weisfeldt
ML. Termination of malignant ventricular
arrhythmias with an implanted automatic defibrillator in human
beings. N Engl J Med. 1980;303:322-324. [Medline]
[Order article via Infotrieve]
-
Klein LS, Miles WM, Zipes DP.
Antitachycardia devices: realities and
promises. J Am Coll Cardiol. 1991;18:1349-1362. [Abstract]
-
Winkle RA, Mead RH, Ruder MA, Gaudiani VA, Smith NA,
Buch WS, Schmidt P, Shipman T. Long-term outcome with the
automatic implantable cardioverter-defibrillator.
J Am Coll Cardiol. 1989;13:1353-1361. [Abstract]
-
Powell AC, Fuchs P, Finkelstein DM, Garan H, Cannom
DS, McGovern VA, Kelly E, Valhakes GJ, Torchiana DF, Ruskin JN.
Influence of the automatic implantable
cardioverter-defibrillator on the long-term prognosis of
survivors of out-of-hospital cardiac arrest.
Circulation. 1993;88:1083-1092.
[Abstract/Free Full Text]
-
Gartman DM, Bardy GH, Allen M, Misbach GA, Ivey TD.
Short-term morbidity and mortality of implantation of
automatic implantable cardioverter-defibrillator.
J Thorac Cardiovasc Surg. 1990;100:353-359. [Abstract]
-
Kim SG, Fisher JD, Furman S, Gross J, Zilo P, Roth
JA, Ferrick KJ, Brodman R. Exacerbation of
ventricular arrhythmias during the postoperative
period after implantation of an automatic defibrillator.
J Am Coll Cardiol. 1991;18:1200-1206. [Abstract]
-
Böcker D, Block M, Isbruch F, Wietholt D,
Hammel D, Scheld HH, Borggrefe M, Breithardt G. Comparison of
frequency of aggravation of ventricular
tachyarrhythmias after implantation of automatic
defibrillators using epicardial versus nonthoracotomy lead
systems. Am J Cardiol. 1993;71:1064-1068. [Medline]
[Order article via Infotrieve]
-
Fisher JD, Mehra R, Furman S. Termination of
ventricular tachycardia with bursts of rapid
ventricular pacing. Am J Cardiol. 1978;41:94-99. [Medline]
[Order article via Infotrieve]
-
Zipes DP, Heger JJ, Miles WM, Mahomed Y, Brown JW,
Spielman SR, Prystowsky EN. Early experience with an implantable
cardioverter. N Engl J Med. 1984;311:485-490. [Abstract]
-
Waspe LE, Kim SG, Matos JA, Fisher JD. Role of
a catheter lead system for transvenous countershock and pacing during
electrophysiologic tests: an assessment of the usefulness of catheter
shocks for terminating ventricular
tachyarrhythmias. Am J Cardiol.
1983:52:477-484.
- Keren G, Miura DS, Somberg JC. Pacing termination of ventricular tachycardia: influence of antia