Expanding Indications for Permanent Pacemakers

Pacing for Hemodynamic Benefit in Cardiomyopathies

The use of pacing to alter the hemodynamic state in patients with cardiomyopathies is an emerging technique in cardiovascular medicine. Dual-chamber (atrial and ventricular) pacing changes the temporal relation between atrial and ventricular contraction and the sequence of ventricular activation, and both factors probably contribute substantially to the hemodynamic benefit achieved with pacing.

Pacing in Hypertrophic Obstructive Cardiomyopathy

The pathophysiology of hypertrophic obstructive cardiomyopathy involves a dynamic pressure gradient over the left ventricular outflow tract caused by a sphincter action of the contracting septal musculature and the anterior mitral leaflet [1]. It has long been known [2] that right ventricular apical pacing in hypertrophic obstructive cardiomyopathy may reduce this gradient; it probably does so by altering the activation sequence of the septum, which contributes to the generation of the gradient (Figure 1). However, reports on large series of patients treated with dual-chamber permanent pacing for this indication have appeared only in recent years.

Fananapazir and colleagues [3, 4] initially published their clinical experience with 44 consecutive patients who had hypertrophic obstructive cardiomyopathy refractory to medical treatment. These patients received dual-chamber pacemakers programmed to sense and pace in both the atrium and the ventricle (DDD); the atrioventricular intervals were short enough to fully activate the ventricle from the pacing site at the right ventricular apex (according to electrocardiographic criteria). Follow-up ranged from 1.5 to 3 months, and remission of symptoms occurred in 42 of the 44 patients (angina, dyspnea, presyncope, and palpitations all abated significantly, and the incidence of syncope tended to decrease). Significant improvement was noted in New York Heart Association functional class and in average exercise duration, and the average left ventricular outflow tract gradient decreased significantly with an increase in cardiac output. These effects persisted even after cessation of pacing during normal sinus rhythm; some changes on the surface electrocardiogram (T-wave morphology) and signal-averaged electrocardiogram also persisted. Therefore, the effect of long-term pacing could not be attributed solely to the alteration of the septal activation sequence by ventricular pacing, and other myocardial changes must have occurred during the phase of long-term pacing that contributed to the persistently improved hemodynamic state that followed cessation of pacing therapy.

Fananapazir and colleagues have published their expanded experience more recently [5]. Notably, they did not require an initial hemodynamic response to temporary pacing as a criterion for selecting patients for permanent pacing.

Jeanrenaud and coworkers [6] supported these results with longer follow-up (up to 62 months) of eight patients with hypertrophic obstructive cardiomyopathy in whom the gradient was persistently reduced even after cessation of pacing. In addressing the issue of optimizing the atrioventricular interval during their initial hemodynamic study, these investigators found that the optimal interval varied among individual patients but was generally short.

A recent report by Matsumoto and colleagues [7] provides evidence of the importance of apical septal activation in the benefit of pacing for patients with hypertrophic obstructive cardiomyopathy. These authors showed that any favorable hemodynamic effect was lost when ventricular pacing was done from the high septum rather than from the right ventricular apex.

Optimal atrioventricular interval programming is crucial in patients with hypertrophic obstructive cardiomyopathy, but this issue remains challenging and incompletely understood. The optimal interval is determined by two opposing factors. Because the benefit of pacing in these patients appears to involve altered septal motion resulting from apex-to-base septal activation, the atrioventricular interval must be short enough to completely capture the left ventricle before any native ventricular activation occurs over the atrioventricular node and the His-Purkinje system. This may necessitate extremely short programmed atrioventricular intervals, especially during exercise, when native conduction over the atrioventricular node tends to accelerate. On the other hand, it is well known that the impaired relaxation caused by hypertrophic obstructive cardiomyopathy often requires a relatively longer atrioventricular interval to achieve maximal left atrial contribution to ventricular filling [8, 9]. Thus, the very short atrioventricular intervals ideal for apex-to-base septal capture may compromise left ventricular filling in patients with hypertrophic obstructive cardiomyopathy.

Thus, the optimal atrioventricular interval is not simply the shortest programmable interval, and it varies widely. No simple formula is yet available for determining the optimal interval, and such determinations probably have to be made for each individual patient. Attempts to optimize intervals are usually made on the basis of Doppler echocardiographic measurements, including left ventricular outflow tract gradient, systemic blood pressure, and transmitral flow characteristics [10, 11], but the selection criteria for optimal intervals are still controversial and poorly defined.

Given the complexity of optimizing atrioventricular intervals, pacemakers with a wide range of programmable atrioventricular intervals and the capacity for rate adaptation of the intervals (intervals that can be shortened with exercise) should be used in patients with hypertrophic obstructive cardiomyopathy to maintain the hemodynamic advantage of ventricular activation through the apical septum during exercise [12, 13].

The inability of some patients with hypertrophic obstructive cardiomyopathy to tolerate short atrioventricular intervals because of inadequate filling of the ventricles, and the inability of others to achieve complete (apex-to-base) septal capture despite maximally short atrioventricular intervals may limit the efficacy of permanent pacing in patients with this cardiomyopathy. Various methods have been used in attempts to slow native atrioventricular conduction in patients not initially benefiting from permanent dual-chamber pacing; these methods include the use of drugs that produce atrioventricular nodal slowing [6] and ablation of the atrioventricular node in resistant cases [14, 15]. An ablation enables exclusive pacemaker-mediated activation of the left ventricle from the right ventricular apex, and longer atrioventricular intervals are programmed in an attempt to enhance left ventricular filling.

Several questions about pacing in hypertrophic obstructive cardiomyopathy still await definitive answers. What is the role of the initial hemodynamic testing with temporary pacing before permanent pacemaker implantation? How long will the beneficial hemodynamic effects of pacing last? What is the mechanism of benefit during and after pacing? How long can the postpacing effect persist? And does pacing result in cellular changes in the myocardium?

The association between the distribution of ventricular hypertrophy in the various forms of hypertrophic obstructive cardiomyopathy and the effect of pacing also awaits further clarification. A few anecdotal reports [16, 17] indicate that pacing may even benefit patients with nonobstructive hypertrophic cardiomyopathy.

On the basis of these preliminary results, we believe that pacing has a role in refractory hypertrophic obstructive cardiomyopathy. In pacing for this indication, we use an atrioventricular interval short enough for the complete capture of the ventricles, and we use Doppler echocardiography to measure the ventricular gradient and various indicators of left ventricular filling to further optimize the atrioventricular interval. Additional controlled studies are needed to better define the long-term benefits and the effects on long-term prognosis of this novel mode of therapy.

Pacing in Dilated Cardiomyopathy

Pacing is even less clearly understood in dilated cardiomyopathy than in hypertrophic obstructive cardiomyopathy. Hochleitner and colleagues [18, 19] are the primary advocates of pacing therapy in dilated cardiomyopathy, and they have published the results of pacing in 17 patients with idiopathic dilated cardiomyopathy who had medically refractory heart failure and severe symptoms. Pacing was in the dual-chamber mode with an atrioventricular interval of 100 ms, and patients were followed for up to 5 years. All patients had significant improvement in functional class that was maintained throughout the follow-up period and modest improvement in ejection fraction that tended to decrease with time. No patient needed further hospitalization because of exacerbation of heart failure. During the first few months after implantation, these beneficial effects dissipated within 2 to 4 hours of interruption of pacing, but in later months, the favorable effects persisted despite pacing withdrawal. The median survival of patients in this group was 22 months, and no patient died as a result of progressive heart failure. The mechanisms of the beneficial effect of pacing in patients with dilated cardiomyopathy are still poorly understood. Notably, dual-chamber pacing (as opposed to ventricular pacing only) and the use of relatively short atrioventricular intervals are probably essential for any beneficial effect [18-22]. A decrease in the duration of diastolic mitral regurgitation due to short atrioventricular interval pacing, which results in an increase in the diastolic filling time of the ventricles, has been suggested as a mechanism for increasing cardiac output in these patients [20]. Atrioventricular intervals as short as 6 ms were shown to be beneficial in 12 patients with dilated cardiomyopathy who were selected specifically because they had short diastolic filling times caused by diastolic mitral or tricuspid regurgitation [20]. To date, however, no systematic hemodynamic studies of this mechanism have been done, and explanations that go beyond the effects on diastolic mitral regurgitation and diastolic filling time may apply [21, 22] (Figure 2).

Whatever the mechanism, it is clear that the optimal atrioventricular interval for the achievement of best cardiac output is different in patients with dilated cardiomyopathy than in patients (or laboratory animals) with better ventricular function, in whom longer atrioventricular intervals have been reported to be as good as or better than short ones [23, 24].

In patients with structurally normal hearts or hearts with mildly to moderately reduced ventricular function, the abnormal sequence of ventricular activation associated with right ventricular apical pacing has been shown to be associated with deleterious effects on ventricular function in comparison with the normal activation sequence of normal conduction [25]. Recent preliminary reports that left ventricular performance is better with high septal than with apical pacing [26] make it tempting to speculate that high septal pacing may have the potential to further improve the results of pacing therapy for patients with dilated cardiomyopathy.

Information is scarce about the percentage of patients with dilated cardiomyopathy likely to benefit from dual-chamber pacing and on the approach to selecting these patients. Some investigators [27], in fact, have found dual-chamber pacing to have no beneficial effect in these patients.

We believe that there may be a population of patients with dilated cardiomyopathy who may benefit from dual-chamber pacing with short atrioventricular delay. However, further studies are needed on the effectiveness of pacing and its effect on survival in these patients (in whom the mortality rate is high) and on the criteria for selecting patients who may benefit from this approach.

Pacing for the Prevention of Atrial Fibrillation

Paroxysmal atrial fibrillation is a major public health problem because of its relatively high incidence in elderly persons. Information on the effect of certain pacing modes in the prevention of paroxysms of atrial fibrillation has evoked interest in the possible role of pacing in the prevention of atrial fibrillation.

In patients who receive permanent pacemakers for sinus node dysfunction, the incidence of atrial fibrillation is lower with atrial-based or dual-chamber pacing systems that maintain atrioventricular synchrony than it is with devices programmed to the ventricular-only pacing mode [28, 29]. Some preliminary evidence [12, 30] also shows that atrial pacing can reduce the frequency of preexisting paroxysmal atrial fibrillation in patients with the sick sinus syndrome. Therefore, atrial-based pacing, through mechanisms not yet completely understood, appears to have a suppressive effect on the occurrence of atrial fibrillation in some patients [31, 32].

Some preliminary data indicate that preventing sinus bradycardia is important in reducing the incidence of atrial fibrillation. Accordingly, preliminary observations have shown that rate-adaptive atrial and dual-chamber pacing (pacing modes that prevent relative bradycardia and inadequate heart rate increase with exercise) maintain sinus rhythm better than their non-rate-adaptive counterparts [33, 34]. This suggests a potential role for excessive catecholamine increase in response to bradycardia (or relative bradycardia during exercise) in the induction of atrial fibrillation. Rate-adaptive pacing of the atrium presumably prevents inappropriate relative bradycardia and, through overdrive suppression, may decrease atrial ectopy and dispersion of atrial refractoriness, resulting in reduced vulnerability to atrial arrhythmias. Another mechanism contributing to atrial fibrillation that may be modulated by atrial pacing is “vagally mediated paroxysmal atrial fibrillation,” described by Coumel and colleagues [35]. Their descriptions further support the presumed association between bradycardia and atrial fibrillation.

Sutton [36] advocates rate-adaptive atrial-based (atrial rate-adaptive pacing [AAIR] or dual-chamber rate-adaptive pacing [DDDR]) pacing to prevent atrial arrhythmias in any patient who needs pacing and is at risk for atrial arrhythmias or who has any evidence of chronotropic incompetence. It should be emphasized, however, that the relative incidence of the “bradycardia-dependent” initiation of atrial fibrillation is not yet known and, to our knowledge, the approach recommended by Sutton has not yet been proved to be effective in large numbers of patients or in any peer-reviewed studies.

Although the above information has been derived from patients who have had pacemakers implanted for other indications (usually symptomatic sinus node dysfunction), there are theoretical reasons to believe that the same considerations apply to patients with relative bradycardia or attenuated heart rate response to exercise, which is not otherwise an indication for pacing. Atrial pacing may reduce the incidence of episodes of paroxysmal atrial fibrillation in some of these patients (who essentially have mild or subclinical forms of the sick sinus syndrome). Therefore, paroxysmal atrial fibrillation may become an indication for pacing in these patients even in the absence of “traditional” indications, although this hypothesis is based primarily on theoretical reasoning and is supported by little actual data.

In patients with long interatrial conduction delay (which may predispose to atrial arrhythmias), atrial fibrillation and other atrial arrhythmias can be prevented with dual-site atrial pacing. Daubert and colleagues [10] have studied this population. The prevalence of interatrial conduction delay (usually diagnosed by measuring the interval from the onset of right atrial electrical activation to left atrial activation, as determined by left atrial electrocardiography or echocardiographically measured mitral valve opening time) is still unclear, and its relative importance in producing atrial fibrillation remains unknown. According to Daubert and colleagues, the incidence of interatrial delay is relatively high and may approach 35% in patients with atrial disease. This delay may lead to interatrial asynchrony and thereby promote various atrial arrhythmias, including atypical left atrial flutter, which eventually evolves into chronic atrial fibrillation [37, 38]. By pacing both atria simultaneously, with one electrode in the right atrial appendage and another in the coronary sinus, Daubert and colleagues [38-40] achieved atrial resynchronization and substantially reduced the incidence of atrial arrhythmias. In patients with interatrial delay and atrioventricular nodal disease, “triple-chamber” pacing was used, with leads in the right atrial appendage, the coronary sinus, and the right ventricle.

Atrial resynchronization by dual-site atrial pacing is therefore a novel approach that may help to prevent atrial fibrillation in some patients. Further studies are being done to better define efficacy and to determine patient selection criteria for this approach.

On the basis of preliminary results and observations, it appears that pacing may decrease the frequency of paroxysms of atrial fibrillation in some patients, including those with chronotropic incompetence, vagally mediated fibrillation, and long interatrial conduction times. Controlled clinical studies are needed to determine the clinical usefulness of pacing in these patients.

Pacing in Heart Transplant Recipients

Although the occasional need for pacing in heart transplant recipients has long been recognized [41, 42], the indications and timing for permanent pacemaker implantation in these persons was poorly defined until recently.

Bradyarrhythmias, primarily isolated sinus node dysfunction, are the most common clinically significant arrhythmias complicating cardiac transplantation [43]. Their overall incidence during the first few weeks after transplantation varies from 18% to 50%; as many as 36% of patients require temporary pacing, and 4% to 25% (but fewer than 10% in most series) ultimately need permanent pacing [41-51].

Several predisposing factors have been noted to contribute to sinus node dysfunction in heart transplant recipients; these include long ischemic time of the donor heart, ischemia of the donor sinus node, long aortic cross-clamping time, and bypass time [41, 44-49]. The association of rejection with bradyarrhythmias is more controversial [41, 47, 49, 50, 52-54]. Although bradycardia should raise the suspicion of rejection, most instances of bradycardia cannot be explained by rejection.

Initially, heart transplant recipients with sinus node dysfunction frequently received permanent pacemakers shortly after surgery. This was because reports suggested that early sinus bradycardia was associated with the unreliability of lower escape mechanisms and with subsequent episodes of asystole and sudden death [41, 44, 55]. However, according to most of the recent large series, early sinus node dysfunction appears to be a transient phenomenon in most cases. It typically occurs within the first 2 weeks after transplantation and has a peak incidence in the second week [43]. Miyamoto and colleagues [49] documented the course of prolonged bradyarrhythmias (mostly sinus node dysfunction) that developed within 5 days after transplantation. Of these, 69% resolved within 7 days and 76% resolved within 20 days. Other series supported the conclusion that more than 50% recovery occurs within 3 weeks [43, 44, 48]. Therefore, there is little doubt that many cases of early sinus node dysfunction are transient.

Long-term follow-up of patients receiving permanent pacemakers in the first few weeks after cardiac transplantation usually shows more than 50% recovery of sinus node function with pacemaker nondependence by a few months after implantation [45, 49, 50, 56-58]. In one series [58], none of the patients who had received the pacemaker before the 16th day after surgery continued to require chronic pacing. Only two series showed more than 50% long-term pacemaker dependency in patients who received pacemakers after transplantation [46, 47]. It is unclear whether the different results in these two series were caused by different patient populations, different criteria for pacemaker implantation, or late timing of pacemaker implantation. For patients whose impaired sinus node function was initially recovered, few reports note the later development of recurrent sinus node dysfunction and the need for late pacemaker implantation [49, 59]. Similarly, late increased mortality and higher incidence of sudden death in this group have not been reported.

It therefore seems reasonable to postpone the decision to implant a permanent pacemaker for sinus node dysfunction after transplantation until near the time of hospital dismissal or about 3 weeks after transplantation, whichever comes first, as long as temporary epicardial pacing leads can be used. Clinical and electrophysiologic variables are poor predictors of future pacemaker dependency, and only observation during hospitalization clarifies the need for pacemaker implantation. Even with relatively late implantation, many patients become nondependent within a few months.

The incidence of atrioventricular nodal and infranodal conduction system disease is much lower in heart transplant recipients, and its natural history is therefore much less clear. Atrioventricular nodal dysfunction is reported as the indication for pacing in fewer than 10% of patients with pacemakers in recent series [45, 46, 49], although other studies have found incidences as high as 25% to 28% in this population [48, 50, 58]. Even in patients with sinus node dysfunction after transplantation, the incidence of concomitant atrioventricular nodal dysfunction is low [60].

Although there seems to be a high rate of long-term resolution of atrioventricular block in transplant recipients [50, 58], the time course is not known, and no definite recommendation can be made about the appropriate timing for pacemaker implantation in these patients.

The issue of the best pacing mode for cardiac transplant recipients remains unsettled. Because atrioventricular nodal dysfunction is relatively rare and because most recipients have isolated sinus node dysfunction, the atrial-only rate-adaptive pacing mode has been suggested [47, 61]. On the other hand, because most patients need only backup pacing and will eventually become nondependent, one may argue that the simple ventricular-only pacing mode would suffice for most of them [57].

A few groups [62, 63] have reported using the transplant recipient's native atrium (when chronotropically competent) as a “physiologic sensor,” with two electrodes in the atria. One electrode is for sensing the recipient's sinus rhythm and the other is for “tracking” with a minimal delay in the donor's atrium. A recent study [64] did not provide support for this approach, and the long-term reliability and chronotropic competence of the recipient's sinus mechanism have long been questioned [42]. Given these uncertainties, we tend to select the most flexible dual-chamber, rate-adaptive pacing systems, namely, DDDR. These pacemakers can be programmed to the DDDR pacing mode with a long atrioventricular interval to achieve adequate support of the atria without pacing the ventricles while still providing the added safety of “backup” ventricular pacing.

Other technical issues that should be addressed in this population are beyond the scope of this review and are summarized elsewhere [12, 65].

Permanent pacing is indicated in heart transplant recipients with sinus node dysfunction that persists for about 3 weeks after transplantation. Although one can argue in favor of atrial-based pacing only, we tend to use dual-chamber pacing in these patients.

Pacing in the Long QT Syndrome

The congenital long QT syndrome is an inherited anomaly of cardiac repolarization that may result in recurrent syncope and sudden death. Initial treatment with various medications and left cervical sympathectomy does not always yield satisfactory results [66]. In the long QT syndrome, whether acquired or congenital, rapid pacing is known to prevent torsade de pointes [67]. The mechanism of this effect is probably related to a reduction in bradycardia-dependent phenomena, including dispersion of refractoriness and development of afterdepolarizations, to which patients with the syndrome are prone [66]. β-Blocker therapy can be effective in the long QT syndrome because it insulates the heart from the sudden sympathetic discharge that often precipitates an arrhythmic event. Although β-blockers are generally effective in reducing the incidence of symptomatic ventricular arrhythmias, they have an unacceptable failure rate in both adults [68] and children [69] with the long QT syndrome. This may be related in part to their bradycardiac effects, which may predispose to torsade de pointes through the mechanisms discussed above. This problem is important, because patients with the long QT syndrome tend to have high incidences of sinus bradycardia and atrioventricular conduction disturbances [66, 69]. Therefore, combining β-blockers with pacing to prevent bradycardia may conceivably improve the results of β-blocker therapy.

Eldar and colleagues [70] were the first to show the efficacy of combining permanent pacing with β-blocker therapy in eight patients with the symptomatic long QT syndrome. In a later publication [71], they extended their experience to 21 patients. These patients (previous therapy with β-blockers, antiarrhythmic medications, or sympathectomy had failed in most of them) received pacing at rates of 75 to 125 beats per minute (sufficient to reduce the QT interval to less than 440 ms) in combination with β-blockers. Of the 18 patients who received both therapies throughout the study, only 2 had recurrent cardiac events; events in the 3 patients who did not receive both therapies throughout the study were clearly related to the interruption of drug treatment or to pacemaker malfunction. Eldar and colleagues emphasized the need for dual-chamber pacing systems; these systems are necessary because of the relatively high incidence of atrioventricular block in this population, an incidence that may be aggravated by β-blocker treatment. It is noteworthy that these results, although impressive, did not derive from a randomized, controlled study.

Similar but somewhat less favorable results were reported by Moss and colleagues [72]. In their series of 30 patients, 21 (70%) were free of recurrent arrhythmic or suspected arrhythmic events after pacemaker implantation. Possibly contributing to the less favorable results of this study were a slower pacing rate (60 to 75 beats per minute for patients in whom pacing therapy failed) and a lack of particular attention to normalization of the QT interval when the pacing rate was determined. Moreover, 4 of the 9 patients in whom treatment failed had received atrial pacing only, which would not have been effective during episodes of atrioventricular nodal block. A long-term mortality rate of 16% was reported in children receiving permanent pacing for the long QT syndrome [69], but information on pacing technique and rate, concomitant therapy, and the selection of patients to receive pacing was lacking in this study.

We believe that a combination of β-blocker therapy and pacing has a role in patients with the symptomatic long QT syndrome who do not respond to β-blocker therapy alone or who develop serious bradycardia during β-blocker therapy. One may even argue in favor of this approach as primary therapy for patients with the long QT syndrome. Further controlled studies are needed to verify the beneficial effect of this therapy on survival.

Pacing in Neurally Mediated Syncope (Vasovagal Syncope)

Neurally mediated syncope is listed as a potential indication for pacing in the American Heart Association and American College of Cardiology guidelines for permanent pacing [73]. The effect of pacing in patients with this condition is still controversial, however, as is the role of temporary pacing during head-up tilt testing in selecting these patients for pacing. It is clear that, in many patients, neurally mediated syncope (defined as syncope induced by a head-up tilt test) has both cardioinhibitory and vasodepressor components, and it is usually difficult to quantify the relative contribution of each component.

Sutton and colleagues have long been the primary advocates of pacing in neurally mediated syncope of the cardioinhibitory subtype [74-77]. Ventricular-only pacing usually fails to ameliorate symptoms even if a bradycardiac response prevails [78], because the absence of atrioventricular synchrony aggravates the peripheral vasodilatation that generally accompanies this condition. However, accumulating evidence suggests that dual-chamber atrioventricular synchronous pacing is beneficial. Sutton and colleagues did tilt-table testing on 3 successive days in 10 patients with predominantly cardioinhibitory neurally mediated syncope. The studies done on the first and second days served as controls and as proof of the reproducibility of syncope. On the third day, tilt testing was repeated with a temporary pacemaker programmed to dual-chamber pacing mode. Syncope was prevented in 5 of 6 patients. The mean time that tilt-up was tolerated was prolonged from less than 1 minute to more than 3 minutes, and cardiac output and arterial blood pressure were significantly increased [74]. These investigators then implanted permanent pacemakers in patients with cardioinhibitory neurally mediated syncope and programmed them to operate as demand dual-chamber pacemakers. Symptoms were abolished in 21 of 40 patients during a mean follow-up period of 24 months [75]. At a mean follow-up of 39 months, the investigators documented symptomatic improvement in 84% of 37 patients, complete resolution of symptoms in 35% of the 37 patients, and a marked reduction in the total number of episodes [77]. They concluded that most cases of vasovagal syncope can be aborted by pacing and that, even if syncope occurs, pacing can prolong consciousness to avoid serious injury.

On the other hand, Sra and colleagues [79] published less favorable results found in 22 patients with bradycardia (rate, less than 60 beats per minute) during neurocardiogenic syncope induced by head-up tilt testing. The patients then received primarily atrioventricular sequential pacing (two patients with atrial fibrillation received ventricular pacing) at a rate 20 beats per minute higher than baseline heart rate. Although pacing failed to prevent a significant decrease in blood pressure during repeated tilt testing, the decrease was somewhat blunted and symptoms were milder. Many of the patients who had had syncope on their first test had only presyncope on repeated tests. Nevertheless, the authors concluded that pacing was not effective in preventing neurally mediated syncope. The difference in the conclusions of the two groups may reflect the different processes used to select patients for pacing and the different expected end points for pacemaker therapy.

Sutton and colleagues [80] recently proposed classifying neurally mediated syncope on the basis of continuous beat-to-beat recordings of arterial pressure and heart rate during tilt testing. Using this classification system, these investigators hope to define in a more detailed way the different hemodynamic patterns preceding syncope and thus improve their ability to predict the effects of different therapies, including pacing. This proposed classification system is summarized in Table 1. Although the relative frequency of each subdivision of patients with neurally mediated syncope is not yet clear, the authors anticipate that most patients will have types 1 and 2a and that types 2b and 3 will be rare. Patients with type 2b may be the most likely to benefit from pacing therapy, and patients with type 2a may benefit to a lesser extent, but this has not been proved. Notably, 17 of the 22 patients studied by Sra and coworkers had type 1 or 2a, but not 2b.

Thus, the question of whether pacing is beneficial in neurocardiogenic syncope is far from resolved. The Vasovagal Syncope International Study (VASIS) now under way in Europe is limited to patients classified by tilt testing as having type 2a or 2b and is the first to address this issue in a controlled, randomized manner [81]. The results of this study will probably have an important effect on the selection of patients with neurally mediated syncope for pacing therapy. In an in-depth review of this subject, Benditt and colleagues [82] recommended randomized, controlled trials to better understand the usefulness of permanent pacing in this population.

We believe that a subset of patients who have neurally mediated syncope with a predominant bradycardiac component may benefit from dual-chamber pacing. We currently select these patients by using temporary dual-chamber pacing during head-up tilt testing to show improvement with pacing. In the few patients for whom pacing seems justified, we tend to use dual-chamber pacing with rate hysteresis to prevent unnecessary pacing when no episodes occur and to maintain the ability to pace at relatively fast rates once an episode of neurally mediated syncope has begun.

Conclusion

In recent years, new frontiers have opened in cardiac pacing. Indications for pacing now include conditions not previously known to benefit from pacing therapy, conditions in which indications for pacing have not been clear, and conditions in which selection criteria for pacing have been poorly defined.

Encouraging results have been reported on the use of pacing for hemodynamic and symptomatic benefit in patients with drug-resistant hypertrophic obstructive cardiomyopathy. We believe that the available information probably justifies the use of pacing as an alternative to cardiac surgery in drug-refractory cases. Preliminary observations in patients with dilated cardiomyopathy await further confirmation. Studies are needed to clarify the mechanisms of hemodynamic benefit in these two groups of patients and to establish better selection criteria for pacing therapy and appropriate programming of atrioventricular delay.

Preliminary results and theoretical reasoning suggest that various atrial pacing techniques may help to minimize the recurrence of paroxysmal atrial fibrillation in some patients who do not have traditional indications or who have borderline indications for pacing. Further studies are needed to improve selection criteria, to define the best pacing mode, and to clarify the importance of atrial synchronization in the prevention of atrial fibrillation.

Cardiac transplant recipients are a growing group of patients, and the question of their need for pacing is often raised. We believe that with our current, improved understanding of the natural history of bradycardia in patients after transplantation, better selection criteria can be established to minimize the unnecessary early implantation of pacemakers. The decision to pace for sinus node dysfunction should probably be postponed for about 3 weeks after transplantation, if possible.

Recent studies of the treatment of patients with the long QT syndrome have suggested that pacing may be a mainstay of therapy that helps to prevent syncope and sudden death. According to some authors, it has become a first-line therapy. It should be emphasized that this conclusion has been reached on the basis of small, nonrandomized studies.

Better criteria for selecting patients for pacing are also being developed for patients with neurally mediated syncope. An improved understanding of the mechanisms of hemodynamic response and appropriate classification of patients according to the patterns of hemodynamic response will certainly improve the selection of patients with neurally mediated syncope of predominantly cardioinhibitory mechanism for permanent pacing.

Dr. Glikson: Heart Institute, Sheba Medical Center, Tel Hashomer 52621, Israel.