1. Basic concepts

• Basic concepts : AV delay and VV delay optimization
• AV delay optimization
• VV delay optimization
• LV pacing alone or biventricular pacing ?
• AdaptivCRT algorithm

AV DELAY AND VV DELAY OPTIMIZATION

Biventricular resynchronization provides significant clinical benefit, a reverse remodeling with reduction of the cardiac volume, and a decrease in morbidity and mortality in heart failure patients with wide QRS. The main limitation of this therapy is that all studies found a significant percentage of patients that do not respond favorably to the resynchronization therapy. Different approaches have been proposed to reduce the percentage of non-responders. Once the patient is implanted, a sub-optimal adjustment of the CRT device can contribute to alter the quality of the response. The principle of the CRT is to change the sequence of activation in a patient with electrical conduction disorder by adjusting the activation delays between a right atrial lead, a right ventricular lead and a left ventricular lead. Two programmable parameters are accessible in this context: 1) the AV delay that determines the activation timing between the right atrium and the right ventricle, with independent programming of a detected AV delay (after the detection of a spontaneous atrial (AS cycle -BV)) and a paced AV delay following an atrial pace (AP-cycle BV). It is possible to program a variable AV delay with a linear reduction of the AV delay paralleling the increase in heart rate; 2) The VV delay regulates the activation delay between the right ventricle and left ventricle; simultaneous activation (VV delay to 0), a right pre-activation (RV à LV, X ms) or a left pre-activation (LV à RV, X ms) are programmable; it is not possible to program a variable VV delay with different values at rest and during exercise. Acute hemodynamic studies have clearly demonstrated a significant benefit provided by the optimization of the AV and / or VV delay. The clinical demonstration of this benefit is much less convincing.

AV DELAY OPTIMIZATION

The atrial contraction contributes to 20-30% of the cardiac output at rest in heart failure patients with systolic dysfunction, this contribution increasing during exercise. Heart failure patients with electrical conduction disorder often displays atrioventricular asynchrony with a shortening of the filling time, a merging of the E and A waves and diastolic mitral regurgitation.
In resynchronized patients, programming a short AV delay allows for anticipating the E wave, a dissociation of E and A waves and prolongation of the filling time. The AV delay should not be set too short because this would result in the amputation of the A wave by mitral closure. Adjusting the AV delay is recommended after implantation of a CRT pacemaker or defibrillator even if the level of clinical evidence is modest.
There are large inter-individual variations in the intra-atrial conduction and intra-ventricular disorders generating marked differences in terms of optimal AV delay justifying theoretically a tailored approach for each patient. The sensed and paced AV delays are independently programmable and must also be optimized independently. A limitation of the optimization of the AV delay is that it is usually performed at rest in supine position and for a given heart rate. These conditions differ significantly from those observed in everyday life. During exercise, unlike patients with healthy heart where the optimal AV delay shortens with increasing heart rate, it seems that resynchronized patients response to stress is not stereotypical. In some patients the optimal AV delay during exercise is longer than at rest, in others it is shorter. The systematic use of the automatic AV delay algorithm probably ensures continuous capture during exercise but is not necessarily associated with an additional hemodynamic benefit. Therefore its programming should be discussed for each patient. Biventricular resynchronization allows for reverse remodeling with a progressive reduction over time of the tele-systolic and end-diastolic volume sand pressures. Therefore, the optimization of the AV delay should be ideally repeated periodically.
The optimal AV delay allows for a maximum contribution of the left atrial contraction to left ventricular filling, prolongs the filling time, improves the cardiac output in the absence of diastolic mitral regurgitation.
If the AV delay is set too long, the atrial contraction occurs too early in diastole, limiting the atrial contribution to the ventricular filling. The atrial contraction is superimposed with the initial diastolic phase. The cardiac echocardiography finds a fusion between E wave and A wave and a short filling time with a persistent diastolic mitral regurgitation.
If the AV delay is set too short, the ventricular contraction occurs too early resulting in premature mitral closure interrupting the current filling and limits atrial contribution to ventricular filling. The echocardiography finds a premature E wave, a long filling time and split E and A waves with a  truncated A wave by the mitral closure. The decrease in end-diastolic pressure and the decreased preload lead to a reduction in the dP / dt max and the cardiac output.
Before starting the AV delay optimization, some elements must be known. In patients with complete atrioventricular block and high grade AV block or with a very long PR interval, changes in AV delay will have no direct effect on the degree of ventricular capture and fusion. In contrast, in patients with preserved atrioventricular conduction, prolonging the AV delay will cause a progressive fusion with spontaneous activation. Adjusting the AV delay must be performed under electrocardiographic control by integrating the idea that in the group of patients with no complete AV block, which represents the majority of the patients, tuning the AV delay will vary the delay between the atrial systole and the ventricular systole but also will directly interfere with the ventricular activation sequence and the degree of ventricular fusion. To overcome this difficulty, the AV delay is often systematically programmed short (between 90 and 120 ms after a sensed atrial activity and between 130 and 150 ms after atrial pacing).

Example of progressive AV delay adjustment in a resynchronized patient with preserved AV conduction; progressive fusion appears with the prolongation of the AV delay.
Various techniques have been proposed to optimize the AV delay:

• Echocardiography
  Different echocardiographic methods have been proposed to optimize the AV delay: the Ritter's method (that has not been validated in a population of patients with heart failure), the search for a maximum aortic or mitral VTI, a maximum dP / dt max, and the iterative method. The latter is widely used in clinical practice, the goal being to obtain the longest filling time with no amputation of the A wave based on trans-mitral flow analysis.
• Other methods
  Various estimates of the cardiac contractility or cardiac output can be used: wave pulse, blood pressure, dP / dt max, electrocardiographic appearance ... The clinical applicability in daily practice is often limited.
• Automatic optimization algorithm embedded in the device
  If repeated optimizations of AV delay are necessary and must be done in various conditions of pre load, the ideal solution would be that the pacemaker realizes it itself. The AdpativCRT function is available in the latest generation of Medtronic defibrillators; the operating principles of this new algorithm will be discussed at the end of the present chapter.

VV DELAY OPTIMIZATION

Some patients do not respond to CRT and continue to display a significant mechanical ventricular dyssynchrony after implantation. Adjusting the VV delay results in sequential biventricular pacing and has a direct impact on the sequence of ventricular activation. Modifying the VV delay may be proposed to reduce the persistent asynchrony in non-responding patients. This parameter appears interesting in theory in patients with a suboptimal position of the LV lead, or a latency and a prolonged conduction time at the stimulation site. If the optimization of VV delay allows for significant acute hemodynamic benefit, the question of the clinical relevance of this parameter remains debated and not confirmed by clinical studies. As for setting the AV delay, it is likely that the remodeling process directly affects the optimization of VV delay and that the optimization of this parameter must be repeated over time and in various pre-load conditions.
The same tools can be used to optimize AV and VV delay. Cardiac echocardiography is often used in clinical practice. The aortic VTI reflecting the cardiac output, the dP/dt max reflecting the cardiac contractility or the measurement of the degree of ventricular asynchrony are mostly used. Once again, the AdaptivCRT function also proposes to automatically optimize the VV delay. In the light of the practical limits of the VV optimization, the repeated automatic adjustment of this parameter by the device itself looks promising. However, it is still necessary to demonstrate its clinical relevance.

This example shows the effect of VV delay on the ventricular electrical activation; if it is easy to show that the electrocardiographic appearance is actually different from one configuration to another, it is much more difficult to determine what configuration will provide the best clinical response.

LV PACING ALONE OR BIVENTRICULAR PACING ?

One of operating principle of the AdaptivCRT algorithm consists in choosing between left ventricular with fusion and biventricular pacing.
No study has ever demonstrated superiority of biventricular pacing on a pure left ventricular stimulation. In contrast, acute hemodynamic studies have consistently found significant benefit with isolated left ventricular pacing. Similarly, clinical studies found a benefit more or less identical in terms of NYHA class, exercise capacity and ventricular remodeling to those observed with biventricular pacing. Nevertheless, large studies demonstrating the benefits provided by resynchronization were all performed with biventricular stimulation and not with left ventricular stimulation.
In LV pacing configuration, the resynchronization of the two ventricles can be obtained by the fusion between left ventricular paced activation of the right ventricle intrinsic activation. If it seems that the optimal acute hemodynamic benefit may be obtained with a certain degree of fusion (limited data on a very limited number of patients), this optimal degree of fusion is difficult to define and to maintain during exercise (changes in the heart rate and PR interval).
Isolated left ventricular pacing is an attractive option, particularly if the implanted device is a CRT pacemaker. Indeed, it may be performed by using a conventional dual chamber pacemaker without implantation of right ventricular lead which increases the cost / effectiveness ratio and reduces the risk of complication. However, in AV block pacemaker dependent patients, implanting only a left ventricular lead seems risky given the higher percentage of lead dislodgement and high pacing threshold. In patients implanted with a CRT defibrillator, the implantation of a right ventricular lead is essential. However, programming the device in a « LV stimulation only » configuration avoids the consumption associated with right ventricular pacing.

ADAPTIV-CRT ALGORITHM

As seen previously, the ideal for repeated optimizations of the pacing configuration would be that the device itself performs this automatically. This optimization procedure has no additional costs, and is "effortless" for the physician and the different clinical departments (echocardiography, electrophysiology...). Moreover, the majority of measurements made by the device are reproducible. The optimization algorithm AdaptivCRT was developed with this goal. The demonstration of its favorable clinical impact on resynchronized patients remains, however, unproved.

Operating principles

The AdaptivCRT algorithm is only available in DDD or DDDR mode and can be programmed on by selecting either: 1) the “Adaptive Bi-V” setting - the device automatically optimizes the pacing parameters (AV and VV delays) – or 2) the “Adaptive Bi-V and LV” setting – the device will choose between a pure LV pacing configuration with fusion and a regular biventricular pacing with optimization of the AV and VV delays. This algorithm can also be turned off by programming 3) “Nonadaptive CRT”.
This algorithm never leads to the use of extreme values of AV or VV delays. For AdaptivCRT function, possible sensed AV delays range between 80 ms and 140 ms. Possible paced AV delays range between 100 ms and 180 ms. The range of timing for the intraventricular VV delays varies from 0 ms to 40 ms (left or right pre-excitation).

AdaptivCRT operating function relies on the regular evaluation of 1) the atrio-ventricular conduction time, which corresponds to the delay between the EGM recorded by the right atrial lead and the EGM recorded by the right ventricular lead; 2) the width of the P-wave, which corresponds to the delay between the atrial EGM recorded on the bipolar channel of the right atrial lead and the end of the atrial EGM recorded by the shock channel; 3) the width of the QRS complex that corresponds to the delay between the EGM detected by the right ventricular bipole and the end of the EGM recorded on the shock channel.

The algorithm assesses the patient’s intrinsic atrio-ventricular conduction every minute and determines if the patient's AV interval is normal or prolonged. The AV interval measurement is performed by extending the sensed and paced AV delay to 300 ms to allow for intrinsic conduction. In the absence of spontaneous conducted ventricular event for more than 3 consecutive cycles, a prolonged AV conduction is diagnosed and the time interval between AV interval measurements doubles (for example, 2 min, 4 min, 8 min… and so on until a max of 16 hours is reached).

The P-wave and QRS width measurements are scheduled every 16 hours. This interval warrants a sampling at various time of the day. During the measurement, the device will switch the recording channel EGM 1 to RV coil (HVA)/ SVC coil (HVB)  (or HVA/atrial anode in the absence of SVC coil). After 5 beats, the delay between the atrial and the ventricular EGMs, the width of the P wave and the width of the QRS are measured. 

The first measurement of the P-wave and the QRS width is scheduled 30 minutes after the implant. After implantation, the P-wave and the QRS width can be measured anytime by programming the parameter AdaptivCRT.

If the AdaptivCRT parameter is set to « Adaptive Bi-V and LV”, it can switch automatically between the auto BIV and LV mode. The patient will be stimulated in pure LV mode, if the following conditions are respected: 1) the patient's heart rate must be less than or equal to 100 bpm; 2) the conduction delay between the spontaneous atrial EGM and the spontaneous ventricular EGM must be less than or equal to 200 ms; 3) the conduction delay between the paced atrial EGM and the spontaneous ventricular EGM must be less than or equal to 250 ms.
If one of these criteria is not found, the patient is stimulated in a biventricular mode.
 

Details of the algorithm operating function

The exact functioning of this algorithm is relatively confidential.

In a first step, the device assesses intrinsic conduction to determine if a patient's AV interval is normal or prolonged. Normal AV intervals are defined as less than 200 ms for atrial-sensed intervals and less than 250 ms for atrial-paced intervals.

In the presence of a normal AV conduction time and if the patient's heart rate is below 100bpm, the device will use the Adaptative LV pacing mode (LV pace only). The timing of the LV pace is automatically adjusted based on the intrinsic AV interval measurement that occurs every minute.

If the patient AV conduction time exceeds 133.3ms, the LV pacing occurs at about 70% of the intrinsic AV interval.

If the AV conduction time is inferior to 133.3ms, the LV pacing will be delivered 40 ms prior to the intrinsic QRS (calculated AV delay - 40ms).

When intrinsic AV intervals are prolonged, or when the patient's heart rate is above 100 bpm, or if a loss of LV capture is confirmed by LV capture management (LVCM), the Adaptive BiV mode will operate. 

The AV delay will then be calculated as follows:

• After a sensed atrial event, the AV delay is adjusted to pace 40 ms after the end of the P wave (measured on the shock channel) but at least 50 ms before the onset of the intrinsic QRS.
• After a paced atrial event, the AV delay is adjusted to pace 30 ms after the end of the P wave (measured on the shock channel) but at least 50 ms before the onset of the intrinsic QRS during atrial pacing (timing between the atrial stimulus and the bipolar right ventricular EGM).

During Adaptive BiV pacing, the optimal VV delay will be deducted from the QRS width.
If the QRS duration (timing between the bipolar RV EGM and the end of the QRS EGM on the shock channel) is included between 50 ms and 150 ms, the LV will be pre-excited. If the QRS width is included between 150 and 180 ms, a right ventricular pre-excitation is set. If the QRS width is not included between 50 and 180 ms, a LV or RV pre-excitation of 10 ms will be used.
The AV conduction times and the P wave width will also be used to optimize the VV delay. If the AV conduction time during spontaneous atrial rhythm is longer than the P wave width, then the VV delay will be set to 0 ms. 

2. Specificities by company