SecureSense Défibrillateurs Abbott

Programming an ICD is complex and must merge the characteristics of the device recipient with the specific ways the various devices detect, count and discriminate. It may be, at times, challenging to extrapolate the results of a study that used the ICD of a manufacturer to the devices of the 4 competitors, given the fundamental differences in their conception and function. In absence of an optimal, universal programming, the results of large studies are concordant with regard to the need to a) limit the incidence of inappropriate or unnecessary therapies, without jeopardizing the patient’s safety, and b) give priority to ATP instead of electric shocks, which might improve the quality of life as well as increase survival. The formulation of guidelines to outline the programming principles of ICD became necessary to standardize the practices of the various implanting medical centers. It is based on a detailed analysis of the literature in order to define the broad rules of programming, as well as describe the different standard programming of the devices made by the 5 main manufacturers. It is noteworthy that the proposed programming is often markedly different from the nominal programming recommended by the manufacturers.

Anti-bradycardia programming

In Abbott defibrillators and in absence of indication for ventricular pacing, the programming recommendation is VVI at 40 bpm for single chamber devices, and DDD with the VIP algorithm ON to promote spontaneous atrioventricular conduction for dual chamber devices. The aim, in this instance (class I indication) is to eliminate unnecessary and deleterious right ventricular (RV) pacing, known to increase mortality and rates of hospitalization.

Patients presenting with sinus node dysfunction must receive dual chamber (class I indication) instead of single chamber ICD with a view to improve the quality of life and lower the risk of pacemaker syndrome, atrial fibrillation and cerebral vascular accidents. 

In recipients of triple chamber defibrillator, the device programming must enable a percentage of biventricular stimulation approaching 100%, with the QuickOpt™ algorithm programmed ON, which automatically adjusts the atrioventricular delay and interventricular delays despite the absence of a formal confirmation of its efficacy.

Detection

For a primary prevention indication, a 30-cycle ventricular tachycardia (VT) zone between 187 and 250 bpm and a 30-cycle ventricular fibrillation (VF) zone >250 bpm must be programmed. For a secondary prevention indication, these 2 zones may be supplemented by an additional VT zone 10 to 20 bpm slower than the clinical VT. These programming guidelines correspond to those formulated in the main document:

1. In primary or secondary prevention, whether in the VT or VF zone, the device programming must allow a minimum duration of 6 seconds (or approximately 30 cycles depending on the rate) before the detection counter is full, to limit the overall number of appropriate or inappropriate therapies delivered (class I indication)
2. In primary prevention, the slowest zone of therapy must be programmed between 185 and 200 bpm to limit the overall number of therapies delivered (class I indication)
3. In secondary prevention, when the cycle of the clinical VT is known, a VT zone 10 à 20 bpm slower than that tachycardia should be programmed (class IIA indication).

In the early days, the programming priority of ICD was a rapid detection and treatment of ventricular arrhythmias because of several considerations: a) most patients underwent implantation of the device for a secondary prevention indication after reanimation from cardiac arrest, b) the charge time of monophasic shocks could exceed 10 seconds, c) the energy required to defibrillate increased with the duration of the tachyarrhythmia, and d) a risk of undersensing of the ventricular EGM also increased with the length of the tachyarrhythmia. Furthermore, depending on the practices at the various implanting medical centers, the tachycardia zones could be programmed relatively slow with the intent to treat all clinical ventricular tachyarrhythmias.

Along with the evolution of the technology and the changes in indications (higher prevalence of implants for primary prevention indications), various large studies showed that the programming of an excessively short delay between the onset of tachyarrhythmia and the first therapy delivery or of excessively slow zones of tachycardia increased the risk of inappropriate as well as unnecessary therapies delivered for episodes of ventricular tachycardia (VT), which would have ended spontaneously, had a greater number of cycles been programmed. 

The MADIT RIT study, in recipients of Boston Scientific defibrillators implanted for primary prevention indications, revealed that the programming of a very short VT detection zone starting at 170 bpm was likely to increase significantly the risk of appropriate though avoidable therapies, the number of inappropriate therapies, and the rate of hospitalizations and mortality. Other studies have also shown that the programming of a prolonged duration or a greater number (30/40) of cycles needed to fill the VF counter, lowered the overall number of therapies delivered without increasing significantly the risk of syncope. 

The PROVIDE study, in 1,670 recipients of Abbott ICD implanted for primary prevention indications, compared a group of patients whose device was programmed in the usual fashion (cycles limited to 12 in a single VT zone and 12 in the VF zone) versus a group with the programming of a) 25 cycles between 180 and 214 bpm (VT zone 1), b) 18 cycles between 214 and 250 bpm (VT zone 2), and c) 12 cycles >250 bpm (VF zone). In the second group, the overall number of shocks and mortality were significantly lowered, emphasizing the foremost importance of the programming of these settings, and imposing an unequivocal change in the programming routine. 

The programming recommended for the Abbott ICD is not simply based on the results of that last study and includes the observation of a significant improvement in the prognosis and survival of ICD recipients by the programming of a large number of cycles, including in the VF zone. The 30 cycles recommended in the VT or VF zones is greater than the 12 cycles used in the PROVIDE study. Likewise, this number of cycles and the limit of the VF zone are far superior to the nominal values recommended by Abbott. 

Therapies

In the VF zone, the programming of a) an ATP sequence (burst, 85%) during/before the charge and of b) the highest shock amplitude is recommended. In the VT2 zone, ≥1 burst of ATP (85%) must be programmed. And in the VT1 zone (secondary prevention indication), ATP must also be programmed. This programming scheme is concordant with the guidelines formulated in the main document:

1. In all patients presenting with structural heart disease who are at risk of developing an organized VT, 1 sequence of ATP should be programmed (unless it is known to be unsuccessful or arrhythmogenic) in the zones of VT and VF with a view to obviate the need to deliver an electric shock (class I indication).
2. A burst of ≥8 stimuli instead of a ramp (class I indication) and a cycle length between 84 and 88%of the tachycardia cycle length (class I indication) should be programmed.

Recent studies have observed that, while the delivery of an electric shock is the only effective means of terminating a very rapid polymorphous ventricular tachyarrhythmia, the need to deliver ≥1 electric shocks to terminate an organized ventricular arrhythmia, regardless of its rate and zone (VT or VF), is associated with a significant degradation of the prognosis. The choice of ATP instead of a shock for this type of arrhythmia is, therefore, a programming priority (class I indication) that improves the quality of life, and lowers the power consumption as well as mortality. The efficacy of bursts seems greater than that of ramps (PITAGORA study) with a more favorable arrhythmia termination / acceleration to polymorphous arrhythmia ratio. While a single burst can usually be programmed in the VF zone, the number of bursts that must be programmed in the VT zone is debatable, though the delivery of >3 bursts is rarely successful.

Discrimination

In the case of Abbott single chamber ICD, it is recommended to limit the programming to the settings based on the morphology analysis (sudden onset and stability set on “Passive”). In the case of dual or triple chamber ICD, all the discrimination settings, including morphology, stability and onset must be programmed ON, since the diagnosis of VT requires that they all be satisfied. Discrimination must be effective to a ≥230 bpm limit. This programming recommendation corresponds to those formulated in the main document:

1. Except in patients who present with an abnormal atrioventricular conduction, the discrimination algorithms must be programmed to a rate of 230 bpm to limit the risk of inappropriate therapy (class I indication).
2. It is recommended to turn off the clocks (time-out) that force the delivery of therapies beyond a set duration, including when the device diagnoses an SVT (class IIB indication).

The combination of a) relatively high programming of tachycardia zones, b) counters requiring a large number (30) of cycles, and c) reliable VT/SVT discrimination algorithms up to 200 to 230 bpm has considerably decreased the incidence of inappropriate therapies. The need to discriminate the origin of tachycardias is highest in patients at risk of overlap between the VT and SVT rates, i.e. patients presenting with slow VT and episodes of atrial fibrillation with a rapid ventricular response or with sinus tachycardia. Conversely, the discrimination algorithms must be turned off in patients presenting with complete atrioventricular block. With the devices made by St. Jude Medical, the discrimination is only possible within the VT zones and impossible in the VF zone. Therefore, the choice of lower limit of the VF zone has a major effect on the discrimination of arrhythmias, explaining in part the recommendation of programming a VF zone starting at 250 bpm. A 230-bpm limit has been recommended to implement this discrimination, after the observation, in some patients, of rates 200 bpm during episodes of AF, and to not discriminate tachyarrhythmias whose rates are >230 bpm, since faster SVT are rare and since the risk of misclassification of ventricular arrhythmias is high.

Introduction

The last consensus of experts on the optimization of ICD programming recommends, in recipients of Abbott devices, the systematic programming of the SecureSense™ algorithm. This algorithm was initially conceived to enable an early diagnosis of ICD lead dysfunction, the weak link of defibrillation systems. Partial conductor or insulator fractures and connection issues modify infrequently the pacing and defibrillation impedance or the measurements of pacing and sensing threshold. Consequently, a follow-up based exclusively on these measurements is often associated with a belated diagnosis of lead dysfunction that can be the source of inappropriate therapies. On the other hand, the occurrence of brief, characteristic episodes of ventricular oversensing (intermittent sensing of short, disorganized, non-physiologic cardiac signals) is often an early sign of lead dysfunction. These signals are highly variable in amplitude and rate, and may saturate the amplifiers (conductor break). In the early phase of a lead or insulator rupture, oversensing may be limited to the bipolar sensing channel and be absent on the high-voltage channel. The functional principle of the SecureSense™ algorithm is based on this discordance and on a comparison between the bipolar sensing channel and the “discrimination channel”. A noise counter increases when short cycles are limited to the bipolar channel and are absent on the discrimination channel. When oversensing is suspected, the therapies are inhibited and an alert (vibration of the pulse generator and/or remote monitoring alert) is triggered to prompt an expeditious intervention. The algorithm combines a protection against undersensing on the discrimination channel, to guarantee the highest safety, and an optimal ability to detect ventricular arrhythmias. The SecureSense™ algorithm was originally conceived to diagnose noises originating from the lead, caused by a) a fracture of the pacing or sensing conductors, b) an insulator rupture, or c) an insufficient tightening or a faulty insertion of the lead connector in its receptacle. This algorithm seems highly promising by allowing also the diagnosis of other types of oversensing (P and T waves, double counting of the R wave) and by eliminating inappropriate therapies in this context. 

This chapter explains the broad principles followed by the functional details of this algorithm and by illustrative tracings. 

Functional principles of the SecureSense™ algorithm

 The SecureSense™ algorithm interferes directly with the decision, by the implantable device, to treat or not treat a rapid episode, including a range of very rapid rates that threaten the patient’s prognosis, as it operates in the VT as well as in the VF zones. The main objective of this algorithm is to increase the device’s specificity by inhibiting inappropriate therapies that might be delivered because of oversensing. It is, consequently, of foremost importance that the highest sensitivity is preserved, and that all episodes of sustained ventricular arrhythmia be effectively diagnosed and appropriately treated. 

The SecureSense™ algorithm is based on 6 broad functional principles:

1. it is based on an analysis of the congruence of the length of the ventricular cycles present on the bipolar and the discrimination channels. A noise counter is integrated in the decision to deliver or inhibit the therapies. Once active, it increases with each short cycle sensed by the bipolar channel, and is reset to 0 after 2 short cycles sensed by the discrimination channel;
2. the algorithm suspects oversensing and inhibits the therapies, should a discordance between the 2 channels be detected, i.e. the presence of short cycles on the bipolar channel, which increase the arrhythmia counter, in absence of short cycles on the discrimination channel;
3. the algorithm suspects a true ventricular arrhythmia and does not inhibit the programmed therapies when both channels are concordant: i.e. short cycles are detected on the bipolar channel as well as on the discrimination channel;
4. the algorithm diagnoses non-sustained episodes of noise when the noise counter increases without saturating the arrhythmia counter; this allows the identification of oversensing issues, even when sporadic;
5. a protection has been put in place against the risk of inhibition of the therapies in presence of undersensing by the discrimination channel, which is indispensable to preserve the highest sensitivity of the device to diagnose ventricular arrhythmias;
6. this algorithm operates together with the programming of patient warnings (vibrations of the pulse generator) and remote monitoring alerts (Merlin.net™), which shorten the delay between the diagnosis of oversensing by the device and its medical management.

Details of the SecureSense™ algorithm function  

The discrimination channel

This channel is used for the SecureSense™ algorithm as well as to optimize the discrimination between arrhythmias of ventricular versus supraventricular origin. Sensing in the discrimination channel may take place between the RV coil and the pulse generator (coil to can) or between the RV distal electrode (tip to can) and the can (programmable configuration). It is, therefore, in both instances, a «unipolar» or «far-field» sensing (with a single pole inside the heart). In nominal programming, the discrimination channel senses between the coil and the can. It is noteworthy that, while the distal electrode (tip) - can configuration may be a tempting means of discriminating the origin of arrhythmias, it should probably be omitted in order to detect lead breakdowns reliably. Since the distal electrode is common to both channels that are being compared, a dysfunction of that electrode may become apparent as short cycles oversensed on both the high voltage and the discrimination channels (resetting of the noise counter to 0 leading to an absence of inhibition of the therapies). 

Sensing in the discrimination channel and automatic control of sensitivity  

Like the bipolar channel, the discrimination channel operates with an automatic adaptive sensitivity: instead of a fixed value, the sensing threshold adapts automatically, according to the amplitude of the preceding R wave, with the level of sensitivity increasing thereafter throughout the cycle, until it reaches the lowest programmed value (maximum sensitivity) in search of a possible low-amplitude signal. The main difference with respect to the bipolar channel is the total absence of programmable settings: the post-sensing ventricular refractory period measures 150 ms, the threshold start is at 62.5% without decay delay, the maximal sensitivity is 0.3 mV and the low-frequency attenuation filter is deactivated. 

Markers and intervals on the discrimination channel

The number 2 of the marker VS2 corresponds to a sensed event in the second sensing channel. Whether the cycle is short (tachycardia) or long (slow rhythm), all signals sensed in the discrimination channel are marked VS2 (no VS, -, VF or VT marker). The intervals between 2 VS2 cycles are not marked on the tracing, which may on occasion complicate the explanation of the algorithm function. 

Activation of the noise counter on the discrimination channel

Sensing on the SecureSense™ discrimination channel is not constantly activated (which would slightly increase the power consumption of the device). The noise counter is automatically activated after the detection, on the bipolar channel, of 2 out of 3 instantaneous cycles corresponding to the VT or the VF zones. Because of the presence of a 350-ms warm-up interval during which sensing is precluded, sensing on the discrimination channel begins 350 ms after the second short cycle on the bipolar channel. 

Deactivation of the noise counter on the discrimination channel

Once the noise counter activated, it increases with each short cycle on the bipolar channel and is deactivated only after 255 intervals classified VS or VP. Two out of three short cycles are needed to reactivate the counter.

The noise counter

Once activated, the counter is incremented after the occurrence of short cycles on the bipolar channel: the noise counter is incremented by 1 with each detection of an instantaneous short interval on the bipolar channel; a short cycle on the bipolar channel may correspond to a F, T or - marker if the instantaneous cycle is short and the average of the 4 preceding cycles is long. 

The counter is reset to zero after the occurrence of short cycles on the discrimination channel: the noise counter is reset to zero after the detection of 2 short instantaneous cycles on the discrimination channel; these cycles do not need to be consecutive. The definition of a “short” cycle on the discrimination channel varies as a function of the number of programmed zones. In presence of a single programmed VF zone, a cycle is short if <400 ms. In presence of one or two programmed VT zones, a cycle is short if shorter than the longest programmed VT cycle +30 ms (if for example, the VT zone 2 is programmed between 350 and 270 ms and VT zone 1 between 420 and 350 ms, a cycle is short if ≤450 ms). 

Noise counter and delivery of therapies 

When the VT or VF counter is full, the analysis of the noise counter has a direct influence on the decision to deliver or withhold the delivery of therapy:

• if the noise counter is <10: a) the VT/SVT discrimination scrutinizes the VT zone before delivering a first therapy; b) the first therapy is delivered in the VF zone if it consists of ATP delivered before or during the charge. If the ICD’s capacitors are being charged, the noise counter is verified one more time at the end of the charge and, if <10 the shock is delivered whereas if it is ≥10 it is dumped. Once a first therapy delivered, the SecureSense™ algorithm is withheld until the end of the episode, i.e. until the return to sinus rhythm.
• if the noise counter is ≥10 when the VT or VF counter is full, oversensing is diagnosed and the therapies are inhibited. The count of SecureSense™ is verified with each redetection of VT or VF (every 6 cycles). If the counter remains ≥10, the therapies are inhibited.

Noise counter and non-sustained oversensing

Once activated, the noise counter increases continuously in presence of intermittent short cycles detected on the bipolar channel. If the short cycles are intermittent, the VT or VF counter never fills up. However, if the noise counter reaches a count of 10 (5 on the early models) or a multiple of 10, a non-sustained lead noise is diagnosed and a marker (SNS on the new models) is inscribed on the tracing. An alert and a patient notification may also be delivered.

Protection against undersensing

To preserve the highest sensitivity and lower the risk of inhibiting the therapies due to undersensing on the discrimination channel, the algorithm incorporates a protection, which cancels the inhibition of therapies, should sensing on this channel be flawed. The algorithm is automatically reprogrammed to “Passive” during the episode, enabling the delivery of therapy. The algorithm is interrupted if one of these three events occurs during an episode: 1) ≥2 VS2 cycles with a <0.6 mV amplitude, 2) 

a pause > 2200 ms between two VS2 cycles, or 3) occurrence of a single cycle on the discrimination channel during the tachycardia. This protection prevents the inhibition of therapies by prolonged undersensing of an actual VT or VF on the discrimination channel. If the algorithm is interrupted, a specific warning of undersensing on the discrimination channel is notified on the programmer at the time of interrogation.

Protection against undersensing

To preserve the highest sensitivity and lower the risk of inhibiting the therapies due to undersensing on the discrimination channel, the algorithm incorporates a protection, which cancels the inhibition of therapies, should sensing on this channel be flawed. The algorithm is automatically reprogrammed to “Passive” during the episode, enabling the delivery of therapy. The algorithm is interrupted if one of these three events occurs during an episode: 1) ≥2 VS2 cycles with a <0.6 mV amplitude, 2) a pause between two >2,200-ms VS2 cycles, or 3) occurrence of a single cycle on the discrimination channel during the tachycardia. This protection prevents the inhibition of therapies by prolonged undersensing of an actual VT or VF on the discrimination channel. If the algorithm is interrupted, a specific warning of undersensing on the discrimination channel is notified on the programmer at the time of interrogation. 

To optimize the detection in the discrimination channel, the RV coil – can configuration can be reprogrammed to V tip-can. This sensing vector, however, may present limitations from the standpoint of prevention of inappropriate therapies due to lead fracture.

The bipolar and the discrimination channels show concordant short cycles and EGM characteristic of polymorphous tachycardia. The noise counter is activated after 2 instantaneous short cycles (classified -) on the bipolar channel + 350 ms (corresponding to the warm-up phase), followed by the appearance of the VS2 markers on the discrimination channel. The VF counter increases with each F-classified cycle and is filled when its programmed value reaches 12. The noise counter increases in parallel, with each short cycle on the bipolar channel, though is reset to 0 each time 2 short, not necessarily consecutive cycles, are detected on the discrimination channel. When the VF counter is full (12), the noise counter is 3, thus well below 10. This confirms that the episode is true VF, the therapies are not inhibited and the capacitors are charging. At the end of charge, the noise counter is, once again, verified. 

There is a characteristic discordance between the bipolar channel, where the cycles are short and disorganized, and the discrimination channel, which shows an unremarkable tracing. The noise counter is activated after 2 instantaneous short cycles (classified -) on the bipolar channel + 350 ms (corresponding to the warm-up phase). The first VS2 marker appears thereafter on the discrimination channel at the time of the next QRS complex. The VF counter increases with each F-classified cycle (multiple oversensing); it is full when it reaches the programmed value of 12. The noise counter increases in parallel with each short cycle on the bipolar channel. This counter is never reset to 0 since no short cycle is present on the discrimination channel. When the VF counter is filled (12), the noise counter is above or equal to 10 (11). The device diagnoses oversensing (RV lead noise) and inhibits the therapies. After each redetection (6 F-classified cycles) on the bipolar channel, the noise counter undergoes a new analysis. If, as in this example, the counter remains ≥10, the therapies are inhibited since the device has diagnosed persistent oversensing.

Episode of non-sustained oversensing

The same characteristic aspect of oversensing is apparent, with short cycles present on the bipolar channel and absent on the discrimination channel (the QRS complexes are accurately sensed). The episodes of oversensing are brief and the VF counter is never filled, while the noise counter has reached 10. The noise counter is never reset to 0 since there is no short cycle on the discrimination channel. An episode of non-sustained oversensing (NSLN, NSN on the new devices) is diagnosed. The NSLN or NSN marker appears thereafter on the tracing each time the noise counter reaches a multiple of 10. A remote monitoring alert is triggered after a programmable number of similar non-sustained episodes of noise. This algorithm, therefore, discloses the occurrence of oversensing issues, including when they are highly intermittent, representing an important contribution to an early diagnosis of lead dysfunction, which often manifests itself initially as very brief episodes of short ventricular cycles.