Implantable Cardioverter-Defibrillators: Technical Aspects

Implantable Cardioverter-Defibrillator Generator

ICD generators have decreased significantly in size because of the significant advancements in battery, capacitor, and microprocessor technologies. Most of the volume of current ICDs is occupied by its battery and capacitors. The newest generations of devices have added device-initiated long-range telemetry device interrogation, automatic alert notifications, and bioimpedance measurement capability along with other advanced features.

Current devices provide all these options with a generator size of about 30 cm ³ in volume. The ICD generator casing is made of electrically active titanium, considered to be the preferred material because of its conductivity, strength, biocompatibility, corrosion resistance, and light weight. The casing serves to protect the circuitry from the corrosive effects of body fluids and at the same time act as an active high-voltage electrode.

The header is generally made of a clear material, polymethylmethacrylate, so that the connections with the leads can be visually confirmed during implantation and can be inspected if ICD system troubleshooting is required for component malfunction or suspicion of failure. The lead connections gain access to the ICD circuitry through feedthrough wires, which penetrate the casing through sealed openings.

The interior of the ICD consists of one or more batteries, capacitors, a DC-DC converter, hybrid with a microprocessor, telemetry communication coil, and their connections. The sensed ventricular signals, generally 5 to 25 mV in amplitude, enter the generator through the leads, are filtered, and then are analyzed by the algorithms programmed into the hybrid. The hybrid consists of electronic circuitry embedded in a silicon wafer, specifically designed for the analysis of these signals and identification of either tachycardia or fibrillation.

Battery

Unlike other batteries, ICD batteries have many performance requirements. Along with its compact size, it must be capable of charging the capacitors relatively quickly by delivering high-energy currents of approximately 75 amperes to permit for quick therapy. It should also be able to hold a large charge with a low current drain on the order of milliamperes for longevity. Factors affecting the current drain include pacing and defibrillation needs and the quiescent current needed for ongoing tasks such as powering the hybrid, monitoring intrinsic rhythm, and logging the data in the memory. Additionally, the battery performance over time must be predictable to allow for adequate warnings before it is depleted.

The battery used by many current ICD generators is a lithium silver vanadium oxide (Li/SVO) cell. There are two kinds of Li/SVO batteries, anode limited (Li) and cathode limited (SVO). Anode-limited (Li) batteries have two voltage plateaus: a very early short-lived one followed by a later long-lived one. Current ICD batteries are often cathode-limited (SVO) batteries and carry a charge between 1000 and 2000 mA•h at their beginning of life (BOL). This generates approximately 3.2 V at full charge. At middle of life (MOL), these batteries suffer from inherent internal impedance buildup, resulting from the accumulation of a film on the Li electrode. Although this may result in a prolonged charge time, pulsing the battery frequently by charging the capacitors is usually sufficient to prevent this phenomenon.

The battery status is generally estimated by its measured output voltage, which is a crude estimate of its remaining charge capacity. With a significant decline in open-circuit voltage and a rise in internal resistance, the ability to deliver adequate current to charge the capacitors becomes impaired, causing a significant prolongation in charge time. This is called the elective replacement indicator (ERI) or recommended replacement time (RRT). Once the ERI is reached, generator replacement can be electively scheduled, usually within 3 months, depending on the frequency of therapy (which determines the depletion rate). A later indicator, end of life (EOL), reflects a significantly lower voltage and indicates a more urgent need for generator replacement because of the associated long capacitor charge times required to achieve an appropriate shock energy. In addition to voltage criteria, the time required to charge the capacitors to full energy is also used as a measurement of the battery status and may activate the ERI. In certain devices, charging the capacitors stops after 20 seconds and delivers the stored energy as shock therapy.

Newer ICDs have incorporated newer battery technologies to improve battery performance and longevity. Balancing the cell to an appropriate electron reduction slows the progression of internal battery impedance over time. Additionally, the development of hybrid cathode batteries (Li/SVO blended with carbon monofluoride: Li/CFx-SVO) and lithium manganese dioxide (LiMnO ₂ ) have improved service life and resulted in a stable charge time throughout the lifetime of the battery. In addition, LiMnO ₂ batteries have no midlife impedance rise and stable voltage during the lifetime of the device, with a slow gradual decay toward the ERI independent of the rate of energy use. Significant work has been done to optimize device efficiency in an effort to decrease the current drain needed for daily operation of the device, memory management telemetry, and remote monitoring.

Capacitors

The rate of energy delivery required for cardiac defibrillation is much greater than can be delivered directly from ICD batteries. Hence, capacitors are used to store the energy over a longer period of time (seconds) and deliver it over a shorter period of time (milliseconds). This generates a strong electric field over a short period of time capable of interrupting ventricular arrhythmias. Capacitors are defined by their capacitance (C), which is a measure of the amount of electrical charge that the capacitor can store for a given voltage. Multiple capacitors are charged in a parallel configuration by the battery through a DC-DC converter, and then they are connected in series to be able to deliver the stored energy (30–80 J) but at high voltages (up to 890 V) within 10 to 20 ms. The rate with which the capacitor is being charged depends on its capacitance and the internal impedances of the battery and the circuitry. However, the rate with which the capacitor delivers its charge to the patient depends on its capacitance and the lead/can/body impedance.

To achieve a high stored charge, current ICDs incorporate new designs such as the use of multiple capacitors in series and the use of capacitors with convoluted surfaces to increase surface area. Most capacitors used in the ICD industry are aluminum and tantalum electrolytic capacitors because of their ability to charge within the prescribed time limitations. With time, the dielectric layers in these capacitors become deformed, causing current to flow through them (capacitor leak). This will result in a suboptimal charge and prolonged charge time. Capacitor reformation, where the capacitors are discharged through high-impedance circuitry to allow for a longer discharge time, will usually regenerate the dielectric layer and fix the leak.

Electrodes

There are three essential functions of defibrillator systems: (1) detection of the tachycardia, (2) pacing stimulation of the heart, and (3) shock stimulation of the heart. The electrodes are the noninsulated segments of the leads that deliver these functions. The technology used for detection (sensing) and pacing is similar to that used by pacemakers. However, the bipole for sensing and pacing is sometimes from the tip of the ventricular lead to the distal shocking electrode instead of from the tip to the ring electrode. This is called an integrated bipolar system instead of a true bipolar system.

High-voltage coil electrodes are ideal for high-energy shocks. The lower impedance of the coils and the conductor cables to the coils permit a high-current charge discharge from the capacitors. In addition, their wide surface area creates a broader electrical field that may enable the defibrillation of more myocardial mass and reduce the current density near the coil. To better distribute the current drain along the length of the coil, the proximal and distal ends of each coil are sometimes connected via a shunt wire.

Nonthoracotomy or Transvenous Leads

Nonthoracotomy ICD leads (NTLs) were designed to carry high defibrillation energy to the inside of the heart. They can have a dedicated proximal sensing/pacing ring electrode (true bipolar) or use the distal high-voltage shocking coil as the proximal sensing/pacing electrode (integrated bipolar). True bipolar leads often offer better discrimination for sensing and are less susceptible to far-field oversensing and postshock undersensing. The latter is caused by the remoteness of the distal coil from the distal electrode. They often have better pacing performance as well because of the lower current drain and being less susceptible to postshock loss of capture. On the other hand, integrated bipolar leads may offer better defibrillation performance because of the shorter tip-to-distal coil distance. They also have a simpler lead design and a lower incidence of T wave oversensing.

NTLs can have one RV (single-coil) or two (dual-coil) high-voltage shocking electrode(s). The distal coil is usually placed in the RV apex or along the RV septum and the proximal coil in the SVC. The effect of a dual-coil ICD lead on the DFT is multifactorial. It changes the shock vector, and it lowers high-voltage shock impedance, resulting in shortening the shock waveform duration. Although dual-coil ICD lead systems have predominated in the United States and Europe, their clinical superiority over the single-coil lead system is not well established. Defibrillation efficiency is slightly improved, usually by 1 to 3 J. This may be justified in patients with anticipated high DFTs. However, this potential benefit needs to be weighed against the added complexity of the lead construction and its potential effect on lead reliability and the potential for more complex extraction when needed. Additionally, a low right atrial position of the proximal coil, in the case of a severely enlarged right ventricle, may in fact increase the DFT by the negative current vector effect. In other patients, there may be a need to implant other coils (a coronary sinus, middle cardiac vein, subcutaneous, or azygos coil) when maximal shocks are ineffective and are almost always more effective in reducing the required defibrillation energy by 5 to 15 J.

ICD leads are designed with either coaxial or multilumen constructions. A coaxial design, where layered conductors are separated by layers of polyurethane insulation material, was formerly used in some of the original designs. This was associated with insulation failure caused by metal ion oxidation of the middle polyurethane insulation layer, identical to that seen in coaxial polyurethane pacemaker leads in the Medtronic Transvene family of leads. Clinically, this manifested as low stimulation impedance and under- and oversensing. Current ICD leads employ multilumen construction designs, where conductors run in parallel through a single insulating body. This design has allowed these leads to be more resistant to this mode of failure and thinner. However, some variations of this design have been the subject of other failure mechanisms, such as in the cases of Sprint Fidelis (Medtronic) and Riata leads (St. Jude Medical/Abbott). In this multiluminal configuration, the conductor of the distal pace/sense electrode is a coil conductor, whereas the rest are cable conductors. Although the latter are more resistant to mechanical stress, coil conductors are still needed to allow the shaped stylet to go through the lead and position the lead in the heart.

To insulate these conductors, manufacturers use a combination of insulators: silicone, polyurethane, fluoropolymers, and silicone-polyurethane copolymers. Silicone has the best biostability characteristics, but it lacks tensile strength and tear resistance, making it flimsy and prone to abrasions, and it has a high coefficient of friction, complicating implantation. Polyurethane, on the other hand, has excellent tensile strength and a lower coefficient of friction; however, it has poor biostability and is prone to degradation, including environmental stress cracking and metal ion oxidation. Instead, a thin layer of Teflon-like fluoropolymer (ethylene tetrafluoroethylene [ETFE] or polytetrafluoroethylene [PTFE]) is often used to coat the conductors to prevent corrosion. However, this increases the stiffness of the lead and does not have strong physical characteristics. Recently a silicone-polyurethane copolymer (Optim, St. Jude Medical) was used as a component of the insulation. Bench studies of this copolymer showed that it exhibits the biostability and flexibility of silicone and the durability and abrasion resistance and low friction coefficient of polyurethane.

Currently available leads either have silicone insulation backfilling in between and behind the metal defibrillation coils or have a flat wire coil technology to decrease tissue ingrowth. Until recently, some defibrillation coils have also been covered with expanded PTFE (Gore ePTFE). This ePTFE coating is slippery, permits the defibrillation current to be transmitted, and prevents tissue ingrowth and an inflammatory fibrous reaction. Each of these techniques appear to reduce tissue ingrowth and facilitate transvenous lead extraction.

Invariably, NTLs have a high-energy coil located near the distal end and lying within the RV cavity. Manufacturers have released NTLs with similar constructions, although some details differ ( Table 121.1 ). The physics of DC current flow, however, requires at least one other electrode to complete the shocking circuit. The development of smaller ICD generators has allowed for pectoral implantation, which has enabled the use of the generator casing as the second electrode (i.e., a hot can). An animal study compared the defibrillation efficacy of a hot can ICD system placed in a left pectoral or subaxillary location with right pectoral and left or right abdominal locations. The left pectoral and axillary subcutaneous positions were superior to all other locations. The right pectoral location was superior to both abdominal locations. These results imply that alternative ICD implantation sites are feasible in the event of inability to implant a left prepectoral device because of left subclavian venous occlusion, history of left mastectomy/radiation, left-sided arteriovenous fistula, or other reasons to avoid the left prepectoral area.

Sensing

All current ICDs use ventricular heart rate as the cornerstone variable in tachycardia recognition. However, to determine heart rate, the interval between each depolarization of the ventricle must be measured. It is not interval recognition but the detection of individual electrogram events that becomes the basic building block in this process. The process begins with the placement of the sensing lead. The sensed electrogram depends on the health of the myocardium in close proximity to the lead, the far-field structures of the diaphragm, anterior chest wall, and right atrium, and other electrical devices such as pacemakers, cellular phones, and other sources of electromagnetic interference. Detection of the electrogram events is completely dependent on the quality of the signal, and the quality of the signal is determined primarily at the time of the lead placement. Additional aspects potentially dependent on the position of the lead are measures of electrogram morphology such as electrogram event width.

In dual-chamber devices, the accurate recognition of arrhythmias adds another layer of complexity and hopefully specificity with the inclusion of atrial lead data input into the generator. The intracardiac location of the atrial lead (far away from the annulus to minimize the far-field RV signal) and a short interelectrode distance between the distal and proximal electrodes improve the signal-to-noise ratio of the sensed atrial signal and may thus improve the accuracy of the data used for tachycardia discrimination.

Band-Pass Filtering

The sense amplifier processes signals presented to the pulse generator by the sensing electrodes and allows signals of certain frequencies to be presented to the detection logic, whereas others are filtered out. This band-pass filter consists of a high-frequency cutoff to filter out myopotential signals and a low-frequency cutoff to filter out repolarization T wave signals. The midrange is intended to represent a band of frequency containing true signal events. The intent is to prevent extraneous signals from fooling the device into falsely detecting tachyarrhythmias. Unfortunately, there is some frequency overlap between repolarization and depolarization waves, atrial and ventricular events, postpacing polarization and depolarization of the ventricles, myopotentials, cardiac depolarizations, and environmental signals and cardiac events.

Frequency and Amplitude

Both a large signal amplitude and high-frequency content improve the chances of the signal being detected. In addition, there is a strong relation between the amplitude of the signal presented to the band-pass filter and the frequencies contained within that signal. Larger signals usually improve the specificity of electrogram event detection by placing the frequency content into the center frequencies of the band-pass filter. However, a signal that is relatively small, such as 4 to 6 mV, but has good frequency content as represented by the slew rate measured on the pacing system analyzer of more than 1 V/s, will probably do better than a larger signal such as 8 to 10 mV with a much poorer frequency content and slew rate of less than 0.1 V/s.

Autogain and Autothreshold

One of the larger challenges present in ICD detection algorithms is the marked variability in the amplitudes of the signals presented to the ventricular lead: naturally conducted beats through the His-Purkinje system are 5 to 25 mV, paced amplitudes are 5000 mV with a polarization voltage after each pacer spike, premature ventricular contractions and ventricular tachycardia (VT) amplitudes range from 2 to 25 mV, amplitudes of signals during asystole are 0 to 0.15 mV, and ventricular fibrillation (VF) amplitudes range from 0.2 to 20 mV. The transitions between each of the rhythms must be accurately identified.

The most difficult problems are distinguishing frequent excursions from low to high and back to low amplitude electrogram events. The two approaches to this problem are the use of autogain and autothreshold algorithms. The autogain technique uses fixed-amplitude voltage thresholds and amplifies the signal at continually varying gains according to various algorithms to single-count both small and large signals. The autothreshold technique uses one amplifier gain and a continuously varying threshold to accomplish the same end. Usually, the threshold will vary within a single cycle, decaying after each sensed event after an adjustment proportional to the amplitude of the signal. Sometimes the floor or minimum voltage limit can be programmed to distinguish between noise in the signal and low-amplitude VF signals. After a paced beat, the autogain is set to the maximum, and the autothreshold is set to the minimum. In some devices, a time delay for the threshold to initiate its decay can be programmed to reduce the detection of a large T wave from being misidentified as a second R wave.

Estimation of Adequate Amplitude

The ventricular sensing electrode is placed during a nontachycardic rhythm. Therefore the adequacy of that placement must be determined before the VF is induced for defibrillation adequacy testing. Although there is marked variability in the average electrogram event amplitude between VF episodes and within a single VF episode, there are relations between the amplitude of the sinus rhythm electrogram amplitude and the average, minimum, and maximum VF electrogram events of subsequently induced arrhythmias. As an estimate of adequacy, sinus rhythm electrograms of at least 5 mV reliably predict that fewer than 10% of the electrogram events will be undersensed and that failure of tachyarrhythmia detection will be essentially eliminated.

Detection Algorithms

Once an electrogram event is marked or detected, the detection algorithm interprets the pattern of the intervals between the events. The algorithms attempt to develop a hierarchy of detected rhythms that require different therapies such as bradycardia, VT requiring antitachycardia pacing (ATP), VT requiring low-energy shocks, and VF requiring higher-energy shocks. The exclusion of sinus tachycardia and atrial fibrillation with overlapping rates from ventricular arrhythmias is particularly difficult. Added to the ventricular rate criterion is the duration of the arrhythmia, the acuteness of the increase in rate, and the beat-to-beat variation in the cycle length to distinguish these rhythms. Derivatives of the electrogram morphology including templates, duration, and vector correlations, and with the addition of an atrial electrode, the pattern of the relation between the atrial and ventricular events have more recently improved the specificity of arrhythmia detection.