Chronotropic Incompetence

The term chronotropic incompetence (CI) describes an insufficient HR relative to the metabolic needs of the organism, during exercise or under stress caused by emotional disturbance, pain, and so on. The definition of CI has not been standardized. It has been defined as the inability to reach, during an exercise test, a maximal HR more than 75% to 80% of the theoretical maximal HR calculated by Astrand’s formula (220 – Age). Others consider that the absolute peak HR during exercise must be more than 100 or 120 beats/min.³ Wilkoff determined the HR response to exercise in a mathematical model, which includes age, resting HR, intensity of exercise, and functional capacity (FC) in the equation:

In contrast to others, this method is not limited to a single value of HR at peak exercise and takes the activity of the patient into consideration. It does, however, require the performance of a cardiorespiratory exercise test with measurements of expired gases during maximal exercise up to the anaerobic threshold to avoid underestimating CI. Because of the limited relevance of definitions based purely on statistics, a functional definition is usually adopted in clinical practice; this definition includes determining whether CI is symptomatic and whether it limits exercise capacity.

The incidence of CI varies with the definition applied, though it may reach 60% among permanently paced patients. Its origin is atrial or ventricular. Sinus node dysfunction (SND) is the most frequent cause of atrial CI, observed in 25% to 40% of cases, and the incidence increases with age.⁴ A few patients presenting with permanent atrial fibrillation maintain a normal chronotropic function during exercise. However, they have inappropriate tachycardia, bradycardia, or both in alternans. The severity of CI in patients presenting with high-degree AV block increases with age and with the increasingly distal location of a block along the His-Purkinje system. Rarely dissociated from disorders of impulse formation or propagation, CI may be associated with autonomic dysfunction, as well as with myocardial ischemia.⁵ Further investigation is warranted if myocardial ischemia is suggested by clinical presentation, particularly since its incidence is threefold higher among patients with coronary artery disease.

CI can be effectively managed using one of two main pacing modes: (1) In the presence of AV block without associated SND, the P-wave–synchronous VDD or DDD modes allow the synchronization of ventricular contraction with atrial systole with a programmable AV delay. The maximum tracking rate must be adapted to the patient’s characteristics, including habitual physical activities, age, and underlying cardiac status. In order to prevent the development of pseudo-Wenckebach periodicity or 2 : 1 AV block during exercise, the upper tracking rate must not exceed the total atrial refractory period (TARP), which is postventricular atrial refractory period (PVARP) plus AV delay. Most pacemakers allow fixed programming of these intervals or, using automatic adaptive algorithms, their variation in response to HR. The AV delay can also be modulated beat-by-beat in response to the preceding cycle. Adaptation of the AV delay is often preferred to modulate TARP, as a shortening of the PVARP gives rise to the risk of triggering pacemaker-mediated tachycardia (PMT) during exercise. State-of-the-art algorithms effectively prevent the onset of PMT. (2) Sensor-driven pacing allows the adaptation of HR to exercise in the atrium (AAIR), the ventricle (VVIR), or both chambers (DDDR, DDIR). Several sensors have been tested; however, a detailed description of each is beyond the scope of this chapter. Currently, the most frequently used sensors are accelerometers and motion sensors, minute ventilation sensors (which track variations in thoracic impedance), and QT interval sensors. Minute ventilation is less sensitive than motion sensing, though its performance is higher during sustained physical exercise. Some devices have combined these sensors. Much work is in progress to develop more physiological and automatically adaptive (closed-loop) sensors. Verification of the programming of the sensor adapted to the patient’s activity—by analysis of the stored histograms or by a 24-hour ambulatory electrocardiogram (ECG) or exercise test, if necessary—is important.

Hyperchronotropism

Excessively rapid atrial or ventricular pacing may be a source of LV dysfunction. In animal models, long-term, rapid ventricular pacing has been shown to cause a more rapid and severe deterioration of LV function than does atrial pacing. Therefore, all causes of excessively rapid permanent pacing must be identified and corrected to eliminate the risk of induced ventricular dysfunction. These causes include PMT, adverse effects of preventive algorithms against atrial arrhythmias, ventricular response to atrial tachycardias, sensing anomalies, poorly programmed sensors and, rarely, dysfunction of electronic internal circuits (runaway pacemaker). These complications are prevented or detected during programming or interrogation of the pulse generator. A proper programming of the PVARP at rest and during exercise and its extension after ventricular extrasystoles prevent the onset of PMT. If these measures fail, algorithms of PMT termination by atrial pacing or extension of the PVARP are helpful. Finally, all pacemaker models allow programming of mode switch at a selected atrial rate to prevent the tracking of supraventricular tachycardias by the ventricles.

Lower Ventricular Pacing Rate

In the early days of cardiac pacing, Sowton et al studied the optimal lowest pacing rate by measuring CO and left and right filling pressures in patients presenting with complete AV block.⁶ The highest CO and lowest filling pressures were associated with pacing rates between 55 and 90 beats/min and a mean of 71 beats/min. CO/ventricular pacing rate curves showed roughly two distinct profiles. Flat curves (Figure 35-3, left) were most often associated with normal myocardial contractility and a stable CO as HR increased. Peaking curves (see Figure 35-3, right) were often observed in the presence of LV dysfunction, showing greater variations in CO with changes in HR. At the fastest HR, CO levels fell, reflecting the shortened diastolic filling time and alteration in LV compliance.

Figure 35-3 Left, Flat curve. Right, Peaked curve.

(From Sowton E: Haemodynamic studies in patients with artificial pacemakers, Br Heart J 26;737–746, 1964.)

In clinical practice, the lowest pacing rate is often set between 50 and 70 beats/min, as observed on average in healthy individuals at rest. At less than 50 beats/min, CO decreases proportionally with HR, which must be avoided in the presence of HF. In patients with coronary artery disease or ventricular hypertrophy and diastolic dysfunction, excessively rapid pacing decreases coronary perfusion and ventricular filling. Furthermore, excessively rapid pacing shortens the life of the pulse generator.

Maximal Ventricular Pacing Rate

The upper ventricular pacing rate, whether atrial tracking driven or sensor driven, must also be tailored on the basis of the patient’s habitual activities and underlying cardiovascular status. An excessively slow upper pacing rate causes exercise-induced chronotropic insufficiency, and an excessively rapid pacing rate may be deleterious. Kindermann et al ascertained the theoretical, optimal, maximal pacing rate corresponding to the increase in maximum VO₂ (oxygen rate) measured during cardiorespiratory exercise testing in patients with a normal or less than 45% LV ejection fraction (EF), the majority of whom were paced in DDD mode.⁷ The mean optimal pacing rate was 86% of the theoretical maximal HR in patients with a normal LVEF, and 75% in patients with a depressed LVEF. It is noteworthy that programming of the upper pacing rate was excessively high in the majority of patients. Furthermore, in most patients, maximal VO₂ rapidly reached a plateau during exercise testing, regardless of systolic LV function. According to Fick’s principle, a VO₂ that remains stable despite an increase in HR (in presence of a stable arteriovenous O₂ difference) indicates a decrease in SV.

A decrease in CO along with an excessively rapid HR is associated with three main mechanisms: (1) development of myocardial ischemia, (2) decrease in LV diastolic filling, and (3) unfavorable force/HR ratio. The complex interaction of these mechanisms explains why one cannot predict the optimal ventricular pacing rate from measurements of baseline LVEF only. Likewise, the optimal pacing rate during exercise calculated as a percentage of the theoretical maximal HR must be interpreted cautiously. Under some conditions, including ongoing myocardial ischemia, arrhythmias, and so on, the optimal maximal HR estimated on the basis of the maximal CO may not be optimal for individual patients.

Heart Rate and Cardiac Resynchronization Therapy

The long-term effects of atrial pacing at a rate faster than the sinus rate in recipients of cardiac resynchronization therapy (CRT) devices programmed in DDD/DDDR mode are only indirectly understood. First, continuous atrial pacing might increase the inter-atrial conduction delay and accentuate the left heart AV dyssynchrony (see Atrial Conduction Delay). In addition, several studies have found a link between increased mortality and long-term elevation of HR in unpaced patients with or without HF. Finally, a slower HR may promote angiogenesis. For the time being, and pending additional information, the recommendation is to keep the backup atrial pacing rate below the sinus rate, except in the presence of chronotropic insufficiency.

Prevention and Management of Atrial Conduction Delay

In patients with long inter-atrial conduction times, different technical solutions may help prevent pacing-induced AV dyssynchrony in the left heart. This applies mainly to patients presenting with permanent, complete AV block or to those patients who need ventricular stimulation for hemodynamic indications such as hypertrophic obstructive cardiomyopathy. These patients may benefit from one of the following two choices:

Figure 35-5 Adverse hemodynamic consequences of major intra-atrial conduction delay in a patient permanently paced in the DDD mode for bradycardia. A standard AV delay of 150 ms results in a short filling time and single pulse mitral flow (left panel). Increasing the AV delay to 250 ms restores normal AV synchrony (right panel).

(From Daubert JC, Pavin D, Jauvert G, Mabo P: Intra- and interatrial conduction delay: Implications for cardiac pacing, Pacing Clin Electrophysiol 27:507–525, 2004.)

Figure 35-6 Instantaneous hemodynamic changes induced by bi-atrial synchronous DDD pacing versus standard DDD pacing in a patient with long inter-atrial conduction time. Left, Switch from standard to bi-atrial pacing instantaneously corrects the baseline atrioventricular dyssynchrony in the left heart. Right, Switch from biatrial to standard DDD causes a prominent decrease in aortic velocity-time integral.

(From Daubert JC, Pavin D, Jauvert G, Mabo P: Intra- and interatrial conduction delay: Implications for cardiac pacing, Pacing Clin Electrophysiol 27:507–525, 2004.)

It is generally possible to increase the programmed AV delay at the cost of a long TARP and a decrease in the maximal atrial tracking rate during exercise. Depending on the position of the lead and the distribution of intra-atrial morphologic abnormalities, atrial pacing might lengthen or shorten atrial conduction time, with a high interindividual variability. A septal implantation of the atrial lead often allows the shortening of conduction delays. If an atrial lead has already been implanted in the appendage, an alternative dual right atrial pacing, or bi-atrial stimulation, using a second lead placed at the coronary sinus ostium or within the coronary sinus, can be considered. One must also pay attention to the prolongation of the AV delay, which, during CRT, might compromise resynchronization by promoting spontaneous AV conduction.

Late Atrial Sensing and Cardiac Resynchronization Therapy

In the presence of major intra-atrial conduction delay between the sinus node and the atrial pacing site, atrial sensing is delayed. The interval between atrial sensing and ventricular sensing is foreshortened, with a risk of inhibited biventricular pacing or fusion beats. Repositioning the atrial lead or modulating or ablating the AV node may be necessary.

The clinical effects of fused spontaneous and paced QRS in recipients of CRT systems remain uncertain. Indirect arguments suggest that fusion complexes are not necessarily adverse.¹¹ In limited studies, at least a similar hemodynamic benefit was observed in patients at rest in whom resynchronization was achieved by LV stimulation alone, with fusion via the right bundle branch, or by biventricular stimulation.¹² Furthermore, the absence of first-degree AV block predicts a response to CRT. Either the presence of AV block is a prognostic factor associated with more severe heart disease or its absence is associated with a higher likelihood of fusion or “concealed resynchronization.”¹³ Nevertheless, pending further studies, the current recommendation is to avoid or limit biventricular simulation with fusion to the extent possible.

Atrioventricular Synchrony During Exercise

During exercise, sympathetic activity facilitates AV nodal conduction, and a shortening of AV delay that is inversely proportional to the HR is observed. Implantable pacemakers are able to reproduce this phenomenon in patients with AV block.¹⁴ The rate-adaptive AV delay shortens AV delay, usually linearly with acceleration of the HR, up to the shortest programmable value.¹⁵ This offers several benefits: (1) improvements in hemodynamic function and cardiopulmonary performance while preserving proper AV synchrony during exercise¹⁶; (2) ability to increase the maximal synchronous HR during exercise by shortening TARP; and (3) subjective improvement in the quality of life by the adaptation of the AV delay during exercise.¹⁷

Conflicting data have been reported regarding the programming of hemodynamically optimal AV delays in recipients of CRT systems during effort, probably reflecting heterogeneous patient characteristics, including degree and type of dyssynchrony, lead positions, and underlying heart disease.¹⁸ The current recommendation is to deactivate the rate-adaptive AV delay in the absence of AV conduction disorder.

Ventricular Conduction Delay and Atrioventricular Synchrony

An additional delay is inserted between left atrial and LV activation in cases of marked slowing of intraventricular or VV conduction with left bundle branch block morphology, most often observed in the context of advanced heart disease. Despite the absence of PR interval abnormality of surface ECG (Figure 35-7), unequivocal left AV mechanical dyssynchrony is present on ultrasound Doppler examination (Figure 35-8). Although conventional DDD pacing with a short AV delay can correct LV filling (transmitral Doppler), it should not be programmed in patients with HF because of the potential adverse effects of RV pacing. Moreover, this pacing increases only left heart filling, whereas biventricular stimulation restores AV synchrony on both sides of the heart (see Figure 35-17) as well as intraventricular synchrony and improves clinical outcomes.

Figure 35-7 Intracardiac recordings from four permanent leads at time of device implant. The P-R interval is normal on surface electrocardiogram (ECG). However, the intracardiac ECGs show marked VV conduction delay (140 ms) resulting in major left atrioventricular dyssynchrony (LA – LV = 240 ms). RA, Right atrium; LA, left atrium (distal coronary sinus); RV, right ventricle; LV, left ventricle (posterolateral vein); IEGM, intracardiac electrogram.

Figure 35-8 Effect of left bundle branch block (LBBB) on AV synchrony. Left ventricular outflow tract (orange triangles) and mitral Doppler (blue triangle, E wave; white triangle, A wave) are shown during the cardiac cycle. LBBB is responsible for an increased systolic and decreased diastolic interval (see also Figure 35-15) with fusion of A and E waves. The P-R interval is unchanged.

(Modified from Cazeau S, Gras D, Lazarus A, et al: Multisite stimulation for correction of cardiac asynchrony, Heart 84:579–581, 2000.)

Hemodynamic Consequences

Ventricular Pacing and Clinical Outcomes

Pacing Mode and Clinical Outcomes

Several pacing characteristics and their consequences, including percentage and duration of ventricular pacing, pacing mode with respect to conduction abnormalities, site of ventricular stimulation, and duration of the paced QRS, also determine the development of HF.³³ An increase in the paced QRS duration in patients with SND or AV block is associated with an increased likelihood of development of HF.³⁴ This association may be explained by a more prominent mechanical ventricular dyssynchrony at some pacing sites, though it might also be a manifestation of more severe myocardial and conduction system disease.

Choosing the right pacing mode is crucial. In the DAVID trial, which enrolled patients presenting with severe LV dysfunction and a dual-chamber implantable cardioverter defibrillator (ICD), the incidence of hospitalizations for management of HF was higher among those paced in the DDDR mode, at a basic pacing rate of 70 beats/min with a short AV delay, than in patients paced in the VVI mode, at a backup rate of 40 beats/min, which was associated with a higher percentage of ventricular pacing in the former group.³⁵

Furthermore, a post hoc analysis of the Mode Selection Trial (MOST) (Figures 35-11 and 35-12) revealed that in patients paced in DDDR or VVIR for SND, both the rates of atrial fibrillation and hospitalizations for HF increased proportionally to the percentage of ventricular pacing.^36,37 This underscored the importance of limiting RV apical pacing and the risk of loss of benefits conferred by pacing modes that preserve AV synchrony.

Figure 35-11 Cumulative survival free from first hospitalization for management of heart failure according to percent of ventricular pacing during the first 30 days in patients with sick sinus syndrome. A, DDDR mode. B, VVIR mode. (timeline in months).

(From Sweeney MO, Hellkamp AS, Ellenbogen KA, et al: Mode Selection Trial Investigators: Adverse effect of ventricular pacing on heart failure and atrial fibrillation among patients with normal baseline QRS duration in a clinical trial of pacemaker therapy for sinus node dysfunction, Circulation 107:2932–2937, 2003.)

Figure 35-12 Relation between risk of atrial fibrillation (AF) and cumulative percent ventricular pacing (Cum % VP) estimated by Cox model in a patient with sick sinus syndrome. Dashed lines represent 95% confidence intervals for point-by-point estimates of the hazard ratio for a 1% change in Cum % VP. Left, DDDR. Right, VVIR.

(From Sweeney MO, Hellkamp AS, Ellenbogen KA, et al: Mode Selection Trial Investigators: Adverse effect of ventricular pacing on heart failure and atrial fibrillation among patients with normal baseline QRS duration in a clinical trial of pacemaker therapy for sinus node dysfunction, Circulation 107:2932–2937, 2003.)

Cardiac Pacing and Obstructive Hypertrophic Cardiomyopathy

The benefit conferred by RV apical stimulation in obstructive hypertrophic cardiomyopathy (OHCM) is mediated by ensuing changes in the contraction sequence of the various LV segments, the septum in particular. This mitigates the intraventricular gradient, the Venturi effect inside the LV outflow tract, and the systolic anterior motion of the mitral valve. The hemodynamic benefit depends on complete RV capture and proper AV synchronization. While a short AV delay facilitates the former, it may hamper the latter, particularly in the presence of a long inter-atrial delay, which may worsen the clinical status of these patients, whose hemodynamic function depends prominently on ventricular filling. By lengthening AV nodal conduction time, pharmaceuticals such as β-adrenergic blockers and calcium antagonists may enable the programming of a longer AV delay while maintaining complete RV capture. Should drug therapy be unsuccessful, radiofrequency ablation or modulation of the AV node is another option.

Despite its attractive premise, RV apical stimulation for OHCM has yielded mixed results. The PIC trial observed a 50% decrease in left intraventricular gradient and a 20% increase in exercise capacity, though the alleviation of dyspnea and chest pain was similar in DDD and AAI modes, suggesting a placebo effect.³⁸ Except in a subgroup of patients older than 65 years, the Multicenter Pacing Therapy (M-PATHY) trial found neither subjective nor objective differences in exercise capacity between RV-paced patients versus non–RV-paced patients.³⁹ In the absence of more convincing evidence—and given the availability of other, more effective treatments, including surgery or even alcohol ablation of the septum—the role played by cardiac stimulation in OHCM is currently limited (class IIb recommendation in professional guidelines).⁴⁰

Pacing Mode Selection

A pacing mode that would, similar to AAI, completely eliminate ventricular stimulation would be ideal for patients presenting with SND and intrinsic conduction. However, the risk of new AV conduction disorders mandates standby ventricular pacing. Different technical solutions have been proposed to favor natural AV conduction. In the DDD(R) mode, the percentage of ventricular pacing is not effectively decreased by the programming of a fixed, long AV delay because of the physiological variability of AV nodal conduction. The AV hysteresis algorithm enables a search for intrinsic conduction by the extension of the AV interval and further promotes normal ventricular depolarization, though its performance is limited. Algorithms such as AAIsafeR, managed ventricular pacing, and reverse mode switch, which automatically switch the pacing mode from AAI(R) to DDD(R) in the event of a blocked P wave or critical lengthening of the P-R interval, have been designed. An advantage of AAIsafeR is the programmability of the pacing mode switch criteria. The disadvantages of these algorithms involve mostly patients presenting with prolonged spontaneous AV delay. In the presence of values greater than 300 ms, the upper responsive rate must be limited, and a risk exists of AV dyssynchrony as the HR increases.

The influence of RV pacing minimization algorithms on clinical outcome is still controversial. The randomized INTRINSIC RV study, which included recipients of ICDs, programmed either in the VVIR mode with a backup rate set at 40 beats/min or in the DDDR mode with the AV hysteresis algorithm, found no difference in clinical outcomes between the two groups.⁴¹ The SAVE PACe study, which enrolled patients with SND paced in the DDDR mode with or without an algorithm of RV pacing minimization, found a 9.1% median percentage of ventricular pacing with the algorithm compared with 99% without the algorithm. While a 40% decrease in relative risk in the incidence of persistent AF was observed, the rates of death and hospitalizations related to HF were similar in both groups.⁴²

Alternative Right Ventricular Pacing Sites

When permanent ventricular capture is needed, the use of alternative RV pacing sites should be considered to prevent the adverse effects of RV apical pacing and improve ventricular function. The technical feasibility and safety of implanting the RV lead onto the septum have been confirmed in several studies. The short-term to intermediate-term capture threshold, impedance, and sensing characteristics of leads implanted in the apical, septal, or RV outflow tract (RVOT) positions are similar.^43,44 The mean procedural times, complication rates, and risks of lead dislodgment are also similar. The risk of perforation seems low, and the extraction procedures are of similar complexity. The results of studies that compared the short-term and long-term hemodynamic benefits conferred by alternative RV pacing sites versus apical pacing are conflicting for various reasons, including the imprecise definition of the site of lead implant (RVOT versus septum), heterogeneous patient populations and evaluation methods, and, sometimes, insufficient duration of follow-up.⁴⁵

Right Ventricular Upgrade

CRT has been used as an upgrade for patients presenting with severe LV dysfunction and worsening HF due to RV pacing. Several small controlled studies have documented improvements in invasively measured hemodynamic function conferred by CRT upgrades, as well as the alleviation of ventricular dyssynchrony and mitigation of LV remodeling (decrease in ED and ES diameters), increase in LVEF, and decrease in NYHA functional class and increase in distance covered during 6-min walk.^53,54 The effects on survival remain unknown, and long-term clinical results have been mixed. Several multicenter studies are ongoing to further address these uncertainties.⁵⁵

Optimal Left Ventricular Stimulation Site

Theoretically, the optimal ventricular stimulation site yielding the greatest mitigation of dyssynchrony by CRT is the LV region activated last. Several short-term studies have shown prominent hemodynamic improvement by stimulation of that region, as determined by (1) the electrical delay between the QRS onset and the LV-sensed ECG, (2) three-dimensional endocardial mapping, (3) echocardiography (tissue Doppler, strain or speckle tracking), or (4) by magnetic resonance imaging. This site of stimulation is also associated with improvements in functional performance, in reverse remodeling, and in a longer adverse event–free survival. Poor placement of the LV lead could explain nonresponse to therapy in a significant proportion of patients. While it is rational to implant the LV lead in a lateral or posterolateral position, Ansalone et al observed that it corresponds to the zone of latest activation in fewer than 60% of cases. This is further complicated by (1) anatomic variations, (2) technical limitations such as diaphragmatic stimulation or unacceptably high capture threshold, or (3) the presence of scarred myocardium precluding the placement of the LV lead at the preferred site. The targeted hemodynamic and clinical improvements are, nevertheless, of prime importance; preoperative investigations, including echocardiography, magnetic resonance imaging (MRI), and computed tomography (CT) are likely to allow the identification of the best negotiated LV stimulation site, despite the clinical complexity that it might represent.

In contrast to the implantation of the previously described LV lead, optimal placement of the RV lead in biventricular stimulation is poorly understood. Lecoq et al have analyzed the various electrical characteristics associated with CRT.⁵⁶ The placement of the RV lead was modified to obtain the thinnest QRS after the implantation of the LV lead. A midseptal position was chosen in the majority of cases. Only the shortening of the QRS (not the baseline QRS duration) predicted the response to CRT. However, the predictive value of QRS shortening remains controversial, and more detailed information regarding the placement of the RV lead is awaited.⁵⁷

Left Ventricular Epicardial Stimulation Versus Endocardial Stimulation

The original intent to stimulate the LV endocardial surface was based on the normal endocardium-to-epicardium electrical activation sequence, which is associated with a speedy and synchronous ventricular depolarization (Figure 35-13). Van Deursen et al found, in an animal model, a shorter ventricular depolarization and less dispersion of repolarization with endocardial stimulation than with epicardial biventricular stimulation.⁵⁸ They also observed an acute hemodynamic benefit (Figure 35-14), which seemed less dependent on site or timing than with epicardial stimulation. Derval et al compared invasive and noninvasive acute hemodynamic measurements made in 35 patients presenting with nonischemic, dilated cardiomyopathies during stimulation from 10 endocardial and 1 epicardial LV sites and found wide interindividual variability in the hemodynamically optimal stimulation site.⁵⁹ Stimulating at the optimal site increased LV dP/dt significantly, compared with (1) standard LV epicardial stimulation via the coronary venous system, (2) lateral LV endocardial wall stimulation, or (3) stimulation at the latest site of LV activation ascertained echocardiographically. Stimulation of the endocardial site immediately opposite to the transvenously identified epicardial site had no significant effects on either dP/dt_max or LVES pressure. However, compared with epicardial stimulation, endocardial stimulation improved LV diastolic function, as ascertained by measurements of LV – dP/dt.

Figure 35-13 Three-dimensional reconstruction of electrical activation times in ventricles measured with epicardial and endocardial electrodes in canine heart with experimental left bundle branch block. First activation/pacing sites are in red (color bar indicates time scale in milliseconds). Endocardial biventricular pacing induces greater shortening of the total activation time more than epicardial biventricular pacing does. Endocardial biventricular pacing is associated with more synchronous LV activation than epicardial biventricular pacing. LBBB, Left bundle branch block; BiV, biventricular; epi, epicardial; LV, left ventricular; endo, endocardial.

(From Van Deursen C, van Geldorp IE, Rademakers LM, et al: Left ventricular endocardial pacing improves resynchronization therapy in canine left bundle-branch hearts, Circ Arrhythm Electrophysiol 2:580–587, 2009.)

Figure 35-14 A, Percent shortening of ventricular activation time. B, LV dP/dt_max during endocardial and epicardial biventricular (BiV) pacing compared with baseline atrial pacing in the presence of left bundle branch block. A larger and significant increase in LV dP/ dt_max is seen during endocardial pacing than during epicardial pacing and during both biventricular and left ventricular pacing. Lvshort, Short AV delay. *P < .05 versus baseline; #P < .05 endocardial (ENDO) versus epicardial (EPI) (mean values and standard deviations of 48 paired measurements).

(From Van Deursen C, van Geldorp IE, Rademakers LM, et al: Left ventricular endocardial pacing improves resynchronization therapy in canine left bundle-branch hearts, Circ Arrhythm Electrophysiol 2:580–587, 2009.)

Trans-aortic, trans-septal, and trans-apical approaches have been described for permanent LV endocardial stimulation in humans. The first feasibility study was carried out by atrial trans-septal lead implantation, by Jais et al.⁶⁰ In several studies of small numbers of typical CRT system recipients, biventricular endocardial stimulation lowered NYHA functional class and increased LVEF. These data, based on small numbers of observations, need additional studies to clarify the risk/benefit ratio from the perspectives of long-term anticoagulation and complications associated with the various lead implantation techniques.

Single Left Ventricular Stimulation

CRT using a single LV lead is based on the expectation of fusion between stimulated LV and spontaneous RV depolarizations, resulting in a synchronous biventricular activation sequence (see Figure 35-20). Immediate hemodynamic and clinical improvements were observed in typical candidates for CRT. Furthermore, a fused QRS seemed to be associated with a higher LV dP/dt_max. Device Evaluation of CONTAK RENEWAL 2 and EASYTRACK 2: Assessment of Safety and Effectiveness in Heart Failure (DECREASE-HF), a randomized study, compared the 6-month echocardiographic outcomes in patients who underwent (1) synchronous biventricular, (2) sequential biventricular, or (3) single-site LV stimulation.⁶¹ A trend toward a greater decrease in LV dimensions with biventricular LV stimulation than with single-site LV stimulation was observed. In addition, mitral regurgitation worsened in 18% of patients who underwent single-site LV stimulation, perhaps because of greater LV dilation and less complete resynchronization, since ventricular fusion complexes were not required in that study. Furthermore, this means of stimulation is not applicable to ICD recipients or pacemaker-dependent patients, who need an RV lead, because of the notorious instability of the LV epicardial leads and the high capture threshold often associated with them. Finally, despite the availability of a dynamic AV delay function, the programming of synchronous LV-RV activation during spontaneous variations in AV conduction may be challenging, for example, during exercise.

Multiple-Site Left Ventricular Stimulation

CRT using a single LV lead may be insufficient to produce optimal ventricular resynchronization, particularly in the presence of prominent LV dilation. It has been hypothesized that stimulating the LV at two or more distant sites would further improve LV performance and increase the rate of response to CRT. While the feasibility and safety of this strategy has been confirmed, it requires longer procedures and fluoroscopic exposure. In experienced hands, the procedural success rate approaches 85%. In short-term hemodynamic studies in patients with HF, Pappone et al found that dual-site LV stimulation increases LV dP/dt and shortens QRS duration compared with single-site stimulation.⁶² In contrast, Padeletti et al observed no further hemodynamic improvement conferred by CRT delivered via dual-site LV stimulation compared with single-site LV stimulation.⁶³ The multicenter, single-blind, crossover TRIP HF trial included 40 patients who had indications for both CRT and for ventricular pacing because of atrial fibrillation and slow ventricular response.⁶⁴ After 3 months of biventricular stimulation, the patients were randomly assigned to biventricular stimulation versus triple-site ventricular stimulation; then they crossed over to the alternative assignment for three additional months. While no difference was observed in clinical outcomes between the two groups, a significantly greater decrease in LVED diameter and increase in LVEF were observed during triple-site ventricular stimulation. Furthermore, 40% of nonresponders to biventricular stimulation (ascertained on the basis of LVES diameter) became responders during dual-site LV stimulation.

Short-Term Hemodynamic Effects

In patients presenting with HF and left bundle branch block, Leclercq et al observed an immediate hemodynamic improvement during biventricular stimulation compared with spontaneous conduction or DDD RV apical pacing (Figure 35-16).⁶⁶ Instead of an increase in ventricular preload, the increase in CO was caused by an increase in myocardial contractility, since the filling pressures had decreased during CRT. Other studies confirmed improvements in hemodynamic function, which manifested as increases in LV dP/dt_max, LVEF, and pulse pressure, and in LV diastolic function, which manifested as a longer ventricular filling time, higher –dP/dt, and lower filling pressures.

Figure 35-16 Instantaneous decrease in pulmonary capillary wedge pressure (PCWP) and 32% increase in cardiac output (CO) by switch from A-A interval to biventricular DDD stimulation at fixed rate in a patient with ischemic cardiomyopathy. QRS duration is decreased by 35%.

(From Leclercq C, Cazeau S, Le Breton H, et al: Acute hemodynamic effects of biventricular DDD pacing in patients with end-stage heart failure, J Am Coll Cardiol 32:1825–1831, 1998.)

Multiple mechanisms are behind these hemodynamic benefits. On the one hand, CRT resynchronizes at the AV, VV, and intraventricular levels, with the main hemodynamic benefit brought about by the correction of LV activation and resynchronization of segmental contraction. On the other hand, as Nelson et al observed, a greater efficiency of the cardiac pump is achieved by LV stimulation, which decreases to a modest extent the myocardial consumption of energy.⁶⁷ Improvement in VV coupling is another hemodynamically favorable and immediate mechanism of CRT (Figure 35-17).⁶⁸ LV activation and onset of diastole are expedited and do not limit LV filling by the space occupied by the right ventricle inside the pericardium when the filling pressures are high.⁶⁹

Figure 35-17 Correction of interventricular delay with biventricular pacing compared with left bundle branch block or conventional right ventricular apical pacing. Early left ventricular activation during biventricular pacing shortens the left pre-ejection time, shortening the contraction of both ventricles and that of systole. Ventricular filling is improved with biventricular pacing without the need to set a very short atrioventricular delay, as in conventional pacing. Pulm, Pulmonary.

(From Cazeau S, Gras D, Lazarus A, et al: Multisite stimulation for correction of cardiac asynchrony, Heart 84:579–581, 2000.)

Electrical Optimization

Attempts in some studies to obtain the thinnest QRS during CRT by changing the lead(s) position(s) has been associated with higher rates of therapeutic response, though this is still debated. This strategy is clearly limited by the presence of zones of slow conduction situated near the stimulation site, which might conceal components of the QRS, rendering them isoelectric and spuriously shortening the duration of the complex. The morphology of fusion complexes may also help in assessing the effects of CRT. A QRS identical to that produced with left-sided or right-sided stimulation raises the suspicion of conduction delays and latency of capture at the right-sided and left-sided leads, respectively. The programming of sequential LV-RV stimulation may allow a more synchronous activation of the ventricles (Figure 35-20). A change in the bipolar, extended bipolar, or inverted bipolar LV stimulation vector is effective in cases of high capture threshold or phrenic stimulation. The clinical merits of such programming remains to be shown.

Figure 35-20 Top, Simultaneous biventricular (BiV) pacing showing the effect of ventricular latency or slow conduction due to myocardial scar; the left ventricular (LV) depolarization wavefront is delayed. Major portions of the left ventricle are depolarized by the right ventricular (RV) wavefront with prolongation of the biventricular activation time. Bottom, Interventricular delay programmed with LV pre-excitation to compensate for the LV latency, shortening the BiV activation time. In the case of single-site LV pacing, the degree of fusion with spontaneous conduction on the right side of the heart depends on the programmed atrioventricular delay. This might improve hemodynamic function in patients with a markedly prolonged LV latency, which cannot be overcome by programming maximum interventricular intervals to pre-excite the left ventricle. LV, Left ventricle; RV, right ventricle; CRT, cardiac revascularization therapy; BiV, biventricular.

(From Barold B, Ilercil A, Herweg B: Echocardiographic optimization of the atrioventricular and interventricular intervals during cardiac resynchronization, Europace 10:iii88–iii95, 2008.)

Echocardiographic Optimization

In advanced heart disease, intraventricular and interventricular conduction delays are sometimes markedly lengthened, and a sequential activation of the ventricles may be necessary to achieve the best resynchronization. Echocardiography is the most commonly applied method of VV delay optimization, mostly based on aortic VTI as a surrogate measurement of LVEF. It is an iterative method, which consists of changing AV delay while searching for the highest VTI. Despite immediate improvements in LV dP/dt, SVEF, and LVEF conferred by optimization of VV delay, the multicenter Insync III Marquis (Medtronic, Minneapolis, MN), DECREASE-HF, and Resynchronization for Hemodynamic Treatment for Heart Failure Management II (RHYTHM II) ICD trials did not find clear clinical benefits compared with simultaneous biventricular stimulation.^61,74,75 In clinical practice, in the absence of a proven clinical advantage, the systematic optimization of the VV interval is not recommended after implantation of CRT systems.

Several questions remain, therefore, unanswered and are being further investigated, including (1) the putative benefit of VV optimization by algorithms based on electrical intervals, (2) the role played by ultrasound evaluations of intraventricular delays, using tissue Doppler, strain, speckle tracking, or three-dimensional echocardiography as a means of ventricular delays optimization, and (3) long-term clinical benefits conferred by new VV delay optimization methods.