Oversensing

T-wave oversensing

T wave oversensing currently remains a significant problem in the management of ICD-implanted patients since it can be accompanied by the occurrence of inappropriate therapies particularly during exertion (when RT and TR intervals correspond to the VF zone due to sinus tachycardia). T wave oversensing is associated with a typical alternating pattern between 2 morphologically different signals, namely a high frequency signal (R wave) and a low frequency signal (T wave). For each cardiac cycle, the device counts the R wave and the T wave as a second additional signal resulting in a doubling of the heart rate. The alternating interval duration (RT intervals and TR intervals) is usually pronounced for slow rates (short RT intervals and long RT intervals) although often less during exertion (RT and RT intervals roughly equivalent) or for patients with long QT syndrome. 

Any oversensing of the T wave should be considered as an emergency, with mandatory modification of the programming to avoid the occurrence of inappropriate therapies. An electrical shock or an antitachycardia pacing sequence delivered as a result of T wave oversensing may be accompanied by a proarrhythmogenic effect with the risk of inducing polymorphic ventricular arrhythmia. Indeed, the electric shock synchronizes either on the R wave or on the T wave, with a 50% probability that the shock is delivered on the T wave, hence during the vulnerable ventricular period. Patients with T wave oversensing and high defibrillation threshold are particularly at risk. If the defibrillation threshold is high and approaches the maximum capabilities of the device, the upper vulnerability value is also high. A shock delivered on the T wave therefore has a very high probability of inducing ventricular fibrillation (concept of upper limit of vulnerability) which will subsequently be very difficult to terminate even with a maximum defibrillation shock (high defibrillation threshold).  

Starting with the Lumax 740 platform, various modifications have been made to the BiotronikTM devices, considerably reducing the incidence of T wave oversensing. Compared to the Lumax 540 platform and prior models, different elements of the ventricular sensing process have been modified: 1) the analog-to-digital converter has been modified with integration of a 10-bit parallel converter allowing better adjustment to rapid signals and better signal amplitude resolution; this new converter also reduces energy consumption and potentially prolongs the life of the device; 2) the filters have also been modified, the low-pass filter is no longer programmable and the high-pass 2 filter no longer exists; 3) for the old platforms, the incoming signal was systematically rectified (a negative signal resulted in a signal of the same amplitude but with a positive deflection), which is no longer the case starting with the Lumax 740 platform. When a signal reaches the sensing threshold, the device opens a 110 ms sensing window; during this interval, the device searches for the largest absolute value (positive or negative) of the signal amplitude which then corresponds to the value of the measured R wave; this value has consequently been prolonged from 80 to 110 ms to allow a better assessment of long-duration signals. After this 110 ms window, the sensing threshold is set at 50% (modifiable) of this measurement as in previous generations; the starting value can be much higher than in the previous platforms, the R wave can be measured up to a value of 25 mV (which considerably reduces the risk of T wave oversensing). The threshold remains fixed at 50% for 240 ms; the threshold subsequently decreases after 350 ms (110 + 240 ms) to 25% (modifiable) of the amplitude; on previous platforms, this duration was 360 ms; the ensuing decrement is 12.5% every 156 ms; this decrement is relatively large so as to reach maximum sensing values to accurately sense a VF episode despite a potentially high initial value; the values of 12.5% and 156 ms are not modifiable; 4) the sensing circuit after ventricular pacing has also been modified; after ventricular pacing, the device analyzes the amplitude of the evoked response for 110 ms without possibility of double sensing during a period of 120 ms (modification relative to the previous platforms or this blanking was programmed to Auto); there is an absolute blanking of 19.5 ms following the stimulus to avoid sensing of the spike; the behavior is thereafter identical to that occurring after ventricular sensing.

R-wave double counting

In rare patients with severe intra-ventricular conduction disorder and wide QRS, the ventricular EGM can exceed the duration of the post-ventricular refractory period resulting in the same signal being sensed twice. The double counting of the R wave can occur during a sinus rhythm, during a premature ventricular contraction or solely during a slow VT  Certain drugs (particularly sodium channel blockers for elevated heart rates) or certain metabolic disorders (hyperkalemia) can favor this oversensing by prolonging the duration of the QRS. The double counting of the R wave can also occur in a patient with a dual-chamber ICD, a prolonged PR interval and loss of right ventricular capture. The defibrillator counts both the paced ventricular event and the spontaneous ventricular activity conducted from the atrium. Similarly, in a patient with a triple-chamber ICD and loss of right ventricular capture, the device can count both the ventricular paced event and the right ventricular depolarization arising from left ventricular capture.
The EGM pattern during a double counting of the R wave is relatively characteristic with alternation between 2 ventricular interval durations. The second signal is usually sensed at the end of the ventricular refractory period (the R1R2 interval is exactly equal to the programmed ventricular refractory period or within a limit of  + 20 ms) and always corresponds to the VF zone. The classification of the second interval (R2R1) is contingent on the programming of the tachycardia zones and the heart rate (higher probability of being in the tachycardia zone if the rate is high and the tachycardia zones are programmed low).
The post-ventricular ventricular refractory period is the parameter incorporated in the various ICDs to solve this oversensing problem without jeopardizing the sensing quality of a VF episode. In older platforms, double counting was relatively common especially when this type of device was connected to an integrated bipolar lead. Two reasons were invoked: a very short post-ventricular ventricular refractory period (in the order of 80 ms) and the large spacing of the 2 sensing electrodes on an integrated bipolar lead favoring the prolongation of the duration of the ventricular EGM. 

The double counting of the R wave is an exceptional occurrence on the latest generation of ICDs. The risk of oversensing has been considerably reduced by a prolongation of the post-ventricular ventricular refractory period from 80 ms to 110 ms on the new platforms. This value can also be changed using a code known to the employees of BiotronikTM. The prolongation of the ventricular refractory period generally allows eliminating the problem of double counting and must therefore be proposed in first intention, while bearing in mind that excessive prolongation can lead to an increased risk of undersensing of a true VF. Lowering ventricular sensitivity may sometimes resolve the problem, although this option may also generate a risk of VF undersensing. Moreover, this option is often ineffective since the second ventricular signal can be of at least equal amplitude to the first. Setting a very high VF zone to avoid inappropriate therapies in this setting also does not appear suitable. In the rare instances where the refractory period cannot be sufficiently prolonged, the implantation of a new pacing/defibrillation lead can be proposed. 
In patients with very broad QRS, it is essential during implantation to carefully analyze the pattern and width of the ventricular intra-cardiac electrogram and to verify the absence of any double ventricular counting. It is also probably more appropriate to implant, in this setting, a dedicated bipolar lead rather than an integrated bipolar lead which favors double counting. 

Myopotentials oversensing

Two types of myopotentials can be oversensed by an ICD: 

1) diaphragmatic myopotentials: the use of a high self-adjusting sensitivity allows optimizing the quality of the sensing of the low-voltage VF signals, but also increases the risk of oversensing of diaphragmatic myopotentials at the end of diastole when the sensitivity reaches its maximum. Diaphragmatic myopotential oversensing is rare but has been increasingly observed in patients implanted with an integrated bipolar lead positioned at the apex of the right ventricle. Permanent ventricular pacing is associated with an increased risk of oversensing of these myopotentials since, after pacing, the time spent at maximum sensitivity is prolonged especially at slow heart rates. An integrated bipolar lead favors the phenomenon due to a wider sensing antenna. Diaphragmatic myopotentials correspond to low-amplitude, high-frequency signals, most often detected exclusively on the sensing channel (absent on the far-field channel). The two main characteristics of this type of signal are that their amplitude varies with the respiratory cycle and that can be replicated by specific maneuvers (deep inspiration, Vasalva, forced cough). Oversensing occurs initially at the end of diastole when sensitivity is maximal. Sensing of the true R wave (of high amplitude) modifies the sensitivity level and interrupts, at least temporarily, the oversensing of these small signals, which explains why prolonged oversensing only occurs in pacemaker-dependent patients (absence of spontaneous R wave, sensitivity level permanently at maximum). Oversensing may be avoided by reducing the sensitivity level with the need to verify the accurate sensing of VF signals. In paced patients, an increase in the minimal pacing rate may also have a favorable effect. In some instances, it may be necessary to implant a new defibrillation (DF4 system) or pacing (DF1 system) lead at a remote distance from the diaphragm (septum or infundibulum).
 
2. pectoral myopotentials: in an ICD, the pulse generator being positioned in the pocket near the pectoral muscles (and thus not part of the sensing circuit), the pectoral myopotentials should therefore not generate oversensing. The amplitude of these myopotentials is greater when recorded at the level of the far-field channel which includes the pulse generator (sensing between the right ventricular coil and the generator). On the other hand, if there is an insulation break (typically an erosion leading to a current leak) at the level of the pocket portion of the lead (friction between the pulse generator and the lead), then the sensing (near-field) channel may oversense the pectoral myopotentials, which can lead to pacing inhibition and/or the occurrence of inappropriate therapies. Analysis of the EGMs in this setting reveals the presence of very rapid non-physiological (high-frequency) signals. Oversensing can be replicated by isometric movements of the arm ipsilateral to the generator or by manipulation of the lead in the pocket. When there is suspicion of pectoral myopotential oversensing, a chest X-ray must be performed along with complete control verification of the device (impedance values, pacing and sensing thresholds). The presence of an abnormally low impedance value (pacing and/or defibrillation) or a sudden decrease in this value is suggestive of an insulation break. In very rare cases, myopotential oversensing can be observed when a DF1 system has been implanted with inversion of the pins in the connector.

50Hz oversensing

The potential risk of electromagnetic interference with an implantable defibrillator has been frequently described including in the hospital environment, in the patient’s home or during his/her professional activities. Interference may occur by conduction if the patient is in direct contact with the emitting source or by radiation if the patient is within an electromagnetic field. The most recent ICDs are protected against the vast majority of sources of interference that the patient may encounter in his or her daily life. The parasitic signals are typically filtered, the analysis being restricted to a narrow bandpass corresponding to the physiological signals (high-pass and low-pass filters). However, the high adaptive sensitivity required in an ICD for accurate signal detection during ventricular fibrillation may favor the sensing of non-physiological signals corresponding to the same bandpass. Electromagnetic interference signals may not be appropriately filtered and lead to more or less severe consequences ranging from the occurrence of inappropriate therapies to pacing inhibition in a pacemaker-dependent patient, inappropriate mode switching due to false diagnosis of supraventricular arrhythmia, rapid ventricular pacing synchronized to atrial oversensing, suspension of therapy detection, or fallback to asynchronous mode. Exceptionally, interference with a high intensity electromagnetic field can cause permanent damage to the circuits. 
The diagnosis of electromagnetic interference is based on the concordance between exposure to a source at the time of the episode and oversensing of characteristic signals (rapid, regular and saturating the baseline). Electromagnetic interference at the mains frequencies (60 Hz in the USA and 50 Hz in Europe) occurs when the patient is in physical contact with poorly insulated electrical equipment. If the oversensing is prolonged, a single electric shock is most often curative since the patient usually interrupts his activity immediately. Electromagnetic interference is more frequent for an integrated bipolar lead than for “true” bipolar sensing, the sensing antenna being wider. The characteristic high-frequency, non-physiological signals are sensed on the various available channels (possible diagnosis of dual tachycardia, AF + VF) and are generally of greater amplitude on the far-field channel than on the sensing (near-field) channel.  

The main preventive measure consists of identifying the emitting source and avoiding the use of certain poorly insulated instruments.

Shock lead dysfunction

In the presence of a suspected lead dysfunction, various exams and measurements must be performed: 1) chest X-ray: radiographic abnormalities are not systematic and a typical pattern of lead fracture is not observed in over 50% of cases; 2) repetitive pacing and defibrillation impedance measurements: the latest generations of ICDs perform periodic (daily) impedance measurements with good correlation between sub-threshold measurements and actual measurements on an effectively-delivered high amplitude shock. The presence of an abnormal value or significant variations in the daily measurements (break in the impedance curve) may reveal a lead dysfunction albeit with moderate sensitivity. Indeed, a significant number of patients present a lead dysfunction revealed by the presence of oversensing episodes without abnormal impedance measurements or abrupt variations in values. A low impedance value is suggestive of an insulation break (current leakage) while a high value is suggestive of a conduction wire break (loss of continuity of the defibrillation circuit); 3) evaluation of the sensing and pacing thresholds: the alteration in standard pacing parameters is often delayed; the sensitivity relative to a decrease in ventricular sensing or an increase in pacing thresholds in predicting lead fractures is therefore very low; 4) analysis of the different endocardial electrograms: the pattern of endocardial EGMs associated with a lead fracture is suggestive but non-specific: intermittent sensing of disorganized, rapid, non-physiological cardiac cycles with possible saturation of the amplifiers (conductor wire break) or low amplitude (sensing of myopotentials due to insulation break). These signals exhibit significant variability in both amplitude and frequency, are intermittent in the cardiac cycle and are most often recorded in the VF zone with values at the limit of the post-ventricular refractory period. They can affect the sensing channel and/or the far-field channel depending on the fracture site and only may become apparent after an electrical shock has been delivered on an actual VF episode.