I'll be speaking about the issue of pacing and, in particular, right ventricular (RV) vs biventricular (BiV) pacing in heart failure patients.

I will first discuss hemodynamic consequences of conduction abnormalities and RV pacing. I'll speak about the mechanism of resynchronization therapy and the hemodynamic effects of BiV and left ventricular (LV) pacing, and its impact on our selection of candidates for resynchronization therapy.

As we look at the hemodynamic effects of conduction abnormalities -- and I will extend and include RV pacing in this -- a number of phenomena occur. I'll discuss ejection fraction (EF), then the issues of stretch and energetic inefficiency, as well as mitral papillary muscle dysfunction with mitral regurgitation. This is from Kass, and I refer you to an outstanding review that he did in Reviews in Cardiovascular Medicine.

This indicates the various hemodynamic effects occurring as a result of discoordination. In the upper panel is the early segment, in the setting of left bundle branch block that is the septum. There is early contraction and shortening in this region. This causes some stretching of the opposite area. Then, the lateral wall -- in this case the late area -- contracts. This is at the time when it has followed the prestretching that occurred as a result of the early contraction area (the septum), and so it performs at a higher load. This is described by Kass as a "sloshing" of blood from early, to late, to early activated regions. The bottom panel is pressure-volume relationships and curves. Going from sinus rhythm to RV pacing, there is a decrease in the stroke volume -- at a price that one pays hemodynamically. This is our first look at what may be important in terms of the results of this discoordination as well as the consequence of RV pacing.

There is an energetic price one pays. Overall the work is less, but the efficiency is also less; there is energy wasted. The premature activation of the early area, the septum, is inefficient and consumes energy and is less efficient in terms of work; whereas, the late area -- operating at this higher load because of the prestretching that occurs -- pays a higher metabolic price.

In addition to these phenomena, molecular changes have been described -- both at the gap junction protein level, such as Connexin 43, which exists in ventricular myocardium, as well as other excitation contraction coupling proteins, such as sarcoplasmic reticulum Ca2+ ATPase (SERCA 2a) and phospholamban. There is also an alteration of kinase expression seen during RV pacing.

Next, some very early hemodynamic studies. This was one of the early studies described by Askenazi in 1984, looking at a population of patients with normal QRS and PR intervals and comparing atrial pacing with atrioventricular (AV) sequential pacing. One saw a decrease in the change in pressure/change in time (dP/dt) as a result of RV pacing, although in the presence of AV coordination.

Some indirect evidence -- I will be very up front that this is very indirect. These were studies not designed to examine the role of RV pacing in particular, but we will see their results. We will go from studies that had a relatively small percentage of LV dysfunction and heart failure to studies that involve a population with a higher percentage of LV dysfunction.

The Mode Selection Trial (MOST) was a landmark study, a 6-year, prospective, randomized comparison of single-chamber, ventricular rate-modulated (VVIR) pacing, vs dual-chamber, rate-modulated (DDDR) pacing. This involved over 2000 patients with sinus node dysfunction, EF of 56%, Class I and II heart failure 85%, age 73 years. This involved a low-rate pacing at 60 beats per minute (bpm), an AV interval between 120 and 200 milliseconds (ms), and a median follow-up of 33 months.

One analysis, which was reported separately from the original publication by Sweeney, examined the relationship between the percentage pacing and the percentage of patients with heart failure hospitalizations. This bar graph shows the percentage of patients with heart failure hospitalization on the y-axis, and the subgroups by the percentage of time they spent pacing; you go from a population that has less than 10% of the time pacing vs those -- both VVIR as well as DDDR -- with pacing of greater than 90%. One sees a stepwise increase in the percentage of patients hospitalized for heart failure. Again, this is not strictly a heart failure population, but is simply examining the incidence of this event.

Looked at another way, this was a post hoc analysis taking patients with greater than 40% overall cumulative percentage pacing vs those with less than 40% pacing. This is the event-free survival from heart failure, and the progressive incidence of heart failure accumulates over time in the patient population with greater than 40% ventricular pacing. So, although there may be an association and some difference in the populations, this is a progressive disorder that occurs during the follow-up duration.

I would like to turn to the Dual Chamber and VVI (single-chamber ventricular paced/ventricular sensed) Implantable Defibrillator (DAVID) trial, which was designed to look at the efficacy of dual chamber (atrial triggered, ventricular inhibited) defibrillator (DDD) vs VVI pacing. In this population, the end point was the composite end point of time to death or first hospitalization. This was reported as part of the main manuscript for the DAVID trial investigators.

This patient population was aged 65 years, mean EF 27% -- quite a different population from that studied in the MOST trial. This was an implantable cardioverter defibrillator (ICD) population. New York Heart Association Class I and II was 88%. The QRS duration was, on average, 120 ms, with roughly 31% having at least 130 ms duration, the PR being in the middle range of 184 ms.

These patients were randomized to VVI at 40 bpm (VVI-40) vs DDDR at 70 bpm (DDDR-70). The most common AV delay was 180 ms, resulting in 60% of ventricular-based beats, vs 1% VVI in this population.

When we look at the composite end point -- mortality and heart failure -- one can see a statistically significant difference between patients with VVI-40 vs patients with DDDR-70.

The authors in this study looked at this in terms of the primary end point. Here is a Kaplan-Meier survival curve of the probability of mortality as well as heart failure hospitalization. The higher incidence is seen in the DDDR pacing group compared with that of the VVI group. The authors speculate in this article that this may be due to the increased atrial pacing and heart rate, the reduction of the PR interval, or the role of RV pacing. Although we observe this difference, we are not certain which of these factors is predominant.

The last of these 3 studies is the Multicenter Automatic Defibrillator Implantation Trial (MADIT) II substudy. I am grateful to Jonathan Steinberg for loaning me this information in his subanalysis. MADIT-II was a study involving over 1200 patients. It looked at patients with prior myocardial infarction and EF less than or equal to 30% who were randomized to ICD vs conventional medical therapy. It demonstrated a reduction in mortality of 31% in favor of the ICD-treated group.

In the analysis, as it was reported in the original publication, there was a trend toward an increased rate of hospitalization for new or worsened heart failure in the ICD-treated group -- 19.9% compared with 14.5% in the conventional treated group -- despite the overall benefit of the ICD for all-cause mortality. One of the hypotheses was that RV pacing induced conduction abnormalities that may result in mechanical dyssynchrony similar to that seen in left bundle branch block; therefore, this analysis was performed.

Of the patients randomized to the ICD arm, 76% had records that could be evaluated for the percentage pacing. This population consisted of Class II to IV heart failure in 67%, and QRS duration greater than 120 ms of 12%, with a mean EF of 23%. The device information was dual chamber mode in 46%, a lower rate of 53 bpm, and a mean AV interval of 190 ms.

Roughly a third of patients had ventricular pacing of 50% or greater in this analysis.

A variety of clinical characteristics that were examined were statistically significant in terms of relationship in the population having less than 50% vs those with 50% or greater pacing. This is not a randomized study in terms of the percentage of ventricular pacing. There are distinct differences in these 2 populations, so the conclusions we draw have to take this strongly into consideration.

Looking at the end points of new and worsened heart failure and the composite end point of new and worsened heart failure or death, we see that these occurred in 21% and 23%, respectively, with appropriate therapy for ventricular tachycardia and ventricular fibrillation occurring in 25%.

These Kaplan-Meier curves show the relationship of the percentage ventricular pacing. This analysis was done at the 50% cut-off. Those being less than 50% having an improved survival in terms of new or worsened heart failure, compared with those with a greater than or equal to 50% ventricular pacing.

A similar relationship is seen with a composite end point of heart failure or death.

Looking at a multivariate analysis, cumulative ventricular pacing remains a statistically important variable of this composite end point.

We can see this when we calculate the overall hazard ratios in terms of these various end points.

In terms of conclusions, they demonstrated an association between the utilization of RV pacing in this high-risk postinfarction patient group and that of the occurrence of congestive heart failure, as well as that of ventricular tachycardia and ventricular fibrillation.

Looking at these 3 studies in composite, we would say that an association has been identified, but really not proof, that RV pacing increases the risk of developing heart failure.

Nonetheless, one has to consider important options in patients who have intact conduction. One may want to consider using RV pacing with a long AV delay to prevent AV pacing.

There are a variety of strategies to perform this, including using DDIR pacing or VVI pacing, as has been suggested, to provide back-up pacing, or to invoke various parameters or features or try to prolong the AV interval.

One question that comes up in the patient with absence of conduction -- which is preferred, RV or BiV pacing? Dr. Steinberg will speak about one of the trials that addresses some of these important patient groups.

I would like to turn to the subject of the differences of hemodynamics seen in RV pacing compared with BiV pacing. First, I want to comment about the mechanics that occur during BiV pacing. It is felt that because the LV and septal walls are incoordinate in the presence of RV pacing or underlying left bundle branch block, that these can be coordinated. The left coronary vein lead or LV site of pacing recruits the lateral wall; the RV apical lead then recruits the septal wall. This coordination results in an important hemodynamic advantage.

Immediately upon turning on pacing, one can identify it in an acute hemodynamic study and in an increase in dP/dt. Similarly, looking at the acute pressure-volume relationship, one can see a shift to the left, shown by the dotted line, that occurs when one goes from RV only or a left bundle branch block pattern to BiV pacing. A similar phenomenon is seen with LV pacing alone, rather than BiV pacing in combination.

The effect of resynchronization on the timing, which I alluded to earlier, one can see graphically. In red in left bundle branch block pattern, the septum contracts early -- the area of early shortening -- followed by contraction of the lateral wall due to the delay in electrical and then mechanical function. The septal region then will see a stretching phenomenon, shown by the upper deflection in the direction of stretching. In contrast, in green, when one has cardiac resynchronization therapy (CRT), both the septal and lateral motions are inward, without the additional stretching that occurs in the left bundle branch block pattern.

There may be a benefit in terms of both altering the activation sequence and improving the acute hemodynamics.

I'll next talk about what has been observed and reported by Kass in terms of energetics. There is an inefficiency of energetics with left bundle branch block or RV pacing. In comparison, looking at BiV, or in this case LV pacing, one goes from a lower to a higher dP/dt with LV pacing. This is in comparison with the results of dobutamine, an inotropic agent, in terms of improving the dP/dt, but at a cost of the myocardial oxygen consumption, shown on the y-axis. The conclusion has been that CRT enhances systolic function while lowering, not increasing, the global myocardial energy demand.

In terms of the time course of these effects, we can see an improvement in dP/dt immediately and a further increase over time. However, when one stops the pacing -- noted by the off-immediate time -- there is an abrupt drop in dP/dt that continues to decline over time. This phenomenon results in a decrease in systolic function in 2 phases: the early phase, which is quite abrupt, and a further decline over time.

A phenomenon, remodeling, occurs as a result of cardiac resynchronization. It is shown here going from baseline to a decreased LV volume -- a result of the 1- and 3-month data. In the off-immediately or even the 1-week-off period, the LV modeling is relatively preserved. Over time, this will start to go back toward the original LV volume.

I would like to talk about how some of these hemodynamic data can be used in predicting responders. The clinical trials to date have used QRS duration. Other direct assessments of myocardial mechanical discoordination have been proposed and are being studied. These include tissue Doppler, echo Doppler, and other forms of imaging, such as magnetic resonance imaging.

I'll show a number of predictors that have not been shown to be valuable. Looking at QRS width, for example, one sees a large spread in terms of the benefit of dP/dt as well as other measurements.

Similarly, looking at a change in QRS -- going from wide to narrow -- is also not extremely predictive in terms of the benefit hemodynamically.

People have turned to other indices, in this case a wall motion index, looking at the septal-to-posterior wall motion. This -- initially reported by Pitzalis in predominantly a dilated cardiomyopathy group -- shows a significant relationship, in this relatively small population of patients, of this index vs changes in LV dimension.

Similarly, looking at tissue Doppler, there has been a significant correlation between changes in function, such as EF, and indices of tissue Doppler. Using a variety of these indices has been proposed as a potential mechanism of both following, and also of predicting, patients who might improve hemodynamically -- those with long QRS durations as well as, potentially, a population of patients who have relatively narrow QRS durations.

The other major issue is that of location. This has been studied in this initial study by Auricchio looking at the various sites that had the greatest change in pulse pressure and hemodynamic benefit, demonstrating a preference to the lateral region -- and other studies showing similar findings in subsequent chronic evaluations.

I would like to touch upon the mechanical vs the electrical resynchronization. In LV pacing vs BiV pacing, you can see the timing of electrical events; early, in blue, is compared with late, in red. The lateral wall is being paced in the left ventricle; as you would anticipate, there is an early site followed by a late site, spreading to the rest of the heart. This is compared with BiV pacing, where one has much more uniform appearance. Despite this difference in electrical activation demonstrated in this experimental study, one saw an improvement in hemodynamics. The authors have speculated on the possible mechanisms of this, but underscoring that this is a mechanical function, not necessarily electrical synchrony, that we are trying to achieve.

I would like to touch upon selection criteria. These are the currently accepted ones based on American College of Cardiology/American Heart Association (ACC/AHA) Guidelines: New York Heart Association Class III or IV; QRS duration of 130 ms or more; EF of 35% or less; and LV end-diastolic dimension of 55 mm or more.

This case is a patient with a distant myocardial infarction, significant heart failure and limitations of exertion, EF of 30%, excellent medical therapy, a widened QRS sinus rhythm, and LV dilatation.

Looking at the various therapies, we would agree that resynchronization therapy would be quite appropriate in this individual, in this case with ICD therapy. A number of studies underscore the role of ICD therapy in this population as well.

There are still a number of unanswered questions in terms of further understanding the mechanism, the type of benefit that could be obtained, how to adjust the variables defining who benefits, defining the role vs other therapy, and combination with other therapies.