System and method of generating a high efficiency biphasic defibrillation waveform for use in an implantable cardioverter/defibrillator (ICD)

[0002] This application claims the benefit of U.S. patent application No. 09/073,394, filed May 5, 1998, now U.S. Pat. No. 6,233,483, which claims the benefit of U.S. Provisional Patent Application No. 60/046,610, filed May 14, 1997.

[0003] The present invention relates to implantable medical devices, and more particularly to an implantable cardioverter defibrillator (ICD) configured to provide a high efficiency defibrillation waveform.

[0004] An ICD continues to be a relatively large device for implantation in the human body. The size of the ICD is primarily determined by the battery and capacitors used therein. The size of the battery (or batteries, in some instances) and capacitors, in turn, is determined by the shock energy requirements for a defibrillation pulse. Thus, a design approach which reduces the energy requirements for defibrillation results in a direct reduction in the overall ICD size.

[0005] In existing ICD devices, the defibrillation waveform or pulse used to deliver a defibrillation shock to the heart is generated by first charging the equivalent of a single capacitor (most ICDs use two capacitors connected in series to function as a single capacitor, thereby reducing the working voltage requirements for each capacitor of the series stack, as explained below) to a desired charge level (voltage) and then discharging the single capacitor through the cardiac tissue for a prescribed period of time during a first or positive phase of the defibrillation waveform, and then reversing the polarity of the discharge for a second prescribed period of time during a second or negative phase of the defibrillation waveform, thereby producing a biphasic stimulation pulse or waveform. It should be noted that in this context the term “single capacitor” is used to refer to a single capacitance, which may be, and usually is obtained by a hardwired connection of two capacitors in series such that the two series capacitors always function and act as though they were a single capacitor. (Two capacitors are connected in series in this manner in order to achieve a higher working voltage for the series-connected capacitor. That is, when two capacitors are connected in series, and each has a working voltage of, e.g., 375 volts (V), then the overall or total working voltage of the series combination becomes 750 V.)

[0006] The purpose of applying a defibrillation shock to the heart is to shock the heart out of a state of fibrillation, or other non-functional state, into a functional state where it may operate efficiently as a pump to pump blood through the body. To this end, the positive phase of the biphasic waveform is preferably a very high voltage that serves to synchronously capture as many heart membrane cells as possible. See, Kroll, “A minimum model of the signal capacitor biphasic waveform” Pace, November 1994. The negative phase of the biphasic waveform, in contrast, simply serves to remove the residual electrical charge from the membrane cells and bring the collective membrane voltage back to its original position or value. See, e.g., Kroll, supra; Walcott, et al., “Choosing The Optimal Monophasic and Biphasic Wave-Forms for Ventricular Defibrillation,”, Journal of Cardiovascular Electrophysiology (September 1995). A biphasic pulse generator of the type used in an ICD device is shown, e.g., in U.S. Pat. Nos. 4,850,357, issued to Bach, Jr.; and 5,083,562, issued to de Coriolis et al.

[0007] When a voltage shock is first applied to a membrane cell, the membrane does not respond to the shock immediately. Rather, the cell response lags behind the applied voltage. This time lag is more or less predictable in accordance with the Blair membrane model. See, e.g., Blair, “On the intensity-time relations for stimulation by electric currents. I” J. Gen Physiol., Vol. 15, pp. 709-729 (1932), and Blair, “On the intensity time relations for stimulation by electric currents. II”, J. Gen Physiol., Vol. 15, pp. 731-755 (1932); Pearce et al., “Myocardial stimulation with ultrashort duration current pulses,” PACE, Vol. 5, pp. 52-58 (1982). When the applied voltage comprises a biphasic pulse having a constant voltage level for the duration of the positive phase (a condition achievable only when the voltage originates from an ideal battery), the membrane cell response to the positive phase reaches a peak (i.e., is at an optimum level) at the trailing edge of the positive phase. Unfortunately, when the applied voltage originates from a charged capacitor, as is the case for an ICD device, the applied voltage waveform does not remain at a constant voltage level, but rather has a significant “tilt” or discharge slope associated therewith. Such tilt or slope causes the peak membrane cell response to occur at some point prior to the trailing edge of the positive phase, which is less than optimum. What is needed, therefore, is a way to optimize the applied voltage waveform so that a maximum membrane cell response occurs coincident with, or nearly coincident with, the trailing edge of the positive phase.

[0008] It is known in the art to switch the capacitors of an ICD from a parallel configuration during the positive phase of a biphasic defibrillation pulse to a series configuration during the negative phase of the biphasic defibrillation pulse. See, e.g., U.S. Pat. Nos. 5,199,429 (FIG. 7A) and 5,411,525. While such action produces a defibrillation waveform having a somewhat different shape, i.e., a waveform having a leading edge voltage of the second or negative phase which is approximately twice the trailing edge voltage of the first or positive phase, such action does little to achieve a maximum cell membrane response coincident with the trailing edge of the first or positive phase.

[0009] It is also known in the art to sequentially switch capacitors in an ICD device in order to allow waveform “tailoring”, e.g., prolong the positive phase duration by sequentially switching in a second charged capacitor as shown in FIG. 9 of U.S. Pat. No. 5,199,429, or by sequentially switching in second, third and fourth charged capacitors, as shown in FIG. 6C of U.S. Pat. No. 5,199,429. However, such “tailoring” still does not address the main concern of achieving a maximum cell membrane response coincident with the trailing edge of the positive phase.

[0010] It is thus evident that what is needed is a capacitor switching scheme and/or method for use within an ICD device which achieves a maximum cell membrane response near or coincident with the trailing edge of the positive phase.

[0011] It is also desirable to provide an ICD that is as small as possible. The limiting factor on ICD thickness is the diameter of the high-energy capacitors. As indicated above, current ICDs typically use two electrolytic capacitors. Current technology in electrolytic capacitors limits the stored voltage to about 370 V per capacitor. Therefore, the current approach is to use two large (≧180 μF) capacitors to achieve the stored energy of ≧25J required for defibrillation. Therefore, the thickness of the ICD is determined by the diameter of the large (≧180 μF) capacitors. There is thus a need for an ICD construction, which would permit the needed energy for defibrillation to be stored in the ICD, while allowing a thinner ICD thickness.

[0012] The present invention advantageously addresses the above and other needs.

[0013] The present invention generates a highly efficient first phase (which is usually a positive phase) of a biphasic defibrillation pulse by switching at least two charged capacitors, preferably three capacitors, from a parallel connection to a series connection during the first or positive phase of the defibrillation pulse. Such mid-stream parallel-to-series switch advantageously steps up the voltage applied to the cardiac tissue during the first phase. A stepped-up voltage during the first phase, in turn, gives an extra boost to, and thereby forces additional charge (current) into, the cardiac tissue cells, and thereby transfers more charge into the membrane of the excitable cardiac cell than would be transferred if the capacitors were continuously discharged in series.

[0014] Phase reversal, e.g., switching to a second or negative phase of the biphasic waveform) is timed to occur when the cell membrane voltage reaches its maximum value at the end of the first phase.

[0015] In accordance with one aspect of the invention, two capacitors are used within the ICD to produce a two-step waveform that outperforms the conventional one-step waveform. It will be shown that the two-step waveform requires a 15.6% lower leading edge, which may result in significantly less pain felt by the patient, and further translates into at 28.8% reduction in required stored energy. This reduction in leading edge amplitude and required stored energy is achieved by controlling the durations of the first and second steps in the two-step positive portion of the waveform.

[0016] In accordance with another aspect of the invention, three capacitors are used within the ICD in order to provide a thinner ICD. These three capacitors store the same energy as a two-capacitor ICD. These smaller capacitors have a smaller diameter and therefore the ICD can be made thinner.

[0017] Disadvantageously, using three capacitors instead of two creates its own set of problems that must be overcome by the present invention. Using three capacitors discharged in series results in: (a) high peak voltages (generally the peak voltage can be three times 370 V or 1110 V); and (b) a small discharge time constant, since the effective capacitance is that of a single capacitor divided by three (or 40 μF if 120 μF capacitors are used), resulting in a mismatch between the discharge (τ=R*C, with R≈50Ω) and tissue (τ_m≈3 ms) time constants. Advantageously, the present invention addresses both of these concerns.

[0018] In accordance with another aspect of the invention, the capacitors of the ICD are reconfigured from a parallel configuration to a series configuration during the positive portion of the defibrillation pulse. While this concept may be used effectively with a two-capacitor ICD, it is preferred for purposes of the present invention that at least three capacitors be used, thereby allowing the ICD to be somewhat thinner that it otherwise could be.

[0019] It is therefore a feature of the present invention to provide an ICD that generates a highly efficient stimulation waveform that transfers more charge to the membrane of an excitable cardiac cell than has heretofore been possible using conventional, series-discharge configurations.

[0020] It is a further feature of the invention to provide an ICD design that results in a thinner ICD than has heretofore been possible using a conventional two-capacitor ICDs.

[0021] The above and other aspects, features, and advantages of the present invention will be more apparent from the following more particular description thereof, presented in conjunction with the following drawings, wherein:

[0022] FIG. 1 illustrates a preferred defibrillation biphasic pulse or waveform generated in accordance with a two-capacitor ICD in accordance with the present invention;

[0023] FIG. 2 depicts the excitable cardiac membrane response to the waveform of FIG. 1;

[0024] FIG. 3 is a functional block diagram of a two-capacitor ICD device, which generates the waveform of FIG. 1;

[0025] FIG. 4 is a simplified schematic diagram of a three-capacitor ICD made in accordance with the invention;

[0026] FIG. 5 illustrates one type of defibrillation waveform that may be generated using the ICD of FIG. 4;

[0027] FIG. 6 depicts the excitable cardiac membrane response during phase 1 (positive phase) to the waveform of FIG. 5;

[0028] FIG. 7 illustrates another type of defibrillation waveform that may be generated using the ICD of FIG. 4;

[0029] FIG. 8 depicts the excitable cardiac membrane response during phase 1 (positive phase) to the waveform of FIG. 7;

[0030] FIG. 9 illustrates, for comparative purposes, the biphasic defibrillation waveform typically provided by a two-capacitor ICD of the prior art;

[0031] FIG. 10 illustrates, again for comparative purposes, the membrane response during phase 1 (positive phase) to the waveform of FIG. 9.

[0032] FIG. 11 shows the first phase of a parallel/series discharge waveform with the durations and time constants defined;

[0033] FIG. 12 shows a first contour plot of stored energy as a function of a scaling factor “K” (equivalent to C_A/C_Band the total capacitance (C_A/C_Bas scaled by τ_m/R_S);

[0034] FIGS. 13 and 14 show a second and third contour plot of the d₁and d₂, respectively, as a function of the scaling factor K and the total capacitance, wherein the optimal value occurs at the cross-hair;

[0035] FIGS. 15, 16 and 17 illustrate how the optimal values for d₁and d₂, tissue resistance (R_S) and tissue time constants (τ_m);

[0036] FIG. 18 is a graph of optimal durations for d₁and d₂as a function of tissue resistance (R_S) for desired (e.g., 60 μF) capacitor and a range of tissue time constants (τ_m);

[0037] FIG. 19 illustrates a single-step and a two-step (parallel/series) waveform of equal stored energy and their resulting cell membrane responses;

[0038] FIG. 20 illustrates the single-step and the two-step waveforms normalized to achieve the maximum cell member response; and

[0039] FIGS. 21 and 22 illustrate analogous results to those depicted in FIG. 20 albeit for extreme combinations of R_Sand C_A(=C_B).

[0040] The following description is of the best mode currently contemplated for practicing the invention.

[0041] The basic concept of the invention relating to forming an efficient defibrillation waveform can be practiced with two or more capacitors within the ICD. A preferred number of capacitors is three. However, the basic concept will first be explained in the context of a two-capacitor ICD.

[0042] In accordance with one aspect of the invention, then a biphasic pulse or waveform is generated by an ICD device having two capacitors that includes a positive phase of duration t₁ms and a negative phase of duration t₂ms, as shown in FIG. 1. First and second capacitors, C_Aand C_B, within the ICD device are initially charged to a voltage V₁and are connected in parallel. The biphasic defibrillation pulse begins by discharging the charged parallel capacitors through the cardiac tissue by way of defibrillation electrodes in contact with the cardiac tissue. Thus, a leading edge of the biphasic pulse starts at a first peak voltage of approximately V₁volts (the charge on the first and second capacitors when first connected to the electrodes).

[0043] During a first portion of the positive phase of the biphasic pulse, the amplitude of the biphasic pulse decays from the first peak voltage V₁to a voltage V₂in accordance with a first time constant τ₁. The first time constant τ₁varies as a function of (C_A+C_B)R, where C_Ais the value of the first capacitor, C_Bis the value of the second capacitor, and R is an effective resistance associated with the discharge through the first and second electrodes.

[0044] A second portion of the positive phase begins by connecting the first and second capacitors in series. This sudden series connection increases the defibrillation pulse to a second peak voltage of approximately 2 (V₂) volts (the sum of the voltages on each of the first and second capacitors at the time the series connection is made), as illustrated in FIG. 1. The amplitude of the biphasic pulse decays during the second portion of the positive phase from the second peak voltage 2 (V₂) to a voltage V₃in accordance with a second time constant τ₂. The second time constant τ₂varies as a function of (C_AC_B/C_AC_B) ) R. Advantageously, the voltage at the trailing edge of the positive phase, V₃, occurs at a time that is near the maximum cell membrane response.

[0045] The negative phase of the biphasic waveform begins by inverting the polarity of the series-connected first and second capacitors. Such negative phase thus commences at a third peak voltage of approximately −V3 volts, and decays thereafter towards zero in accordance with the second time constant τ₂. After a prescribed time period t₂, the negative phase ends.

[0046] The biphasic waveform produced in accordance with the two-capacitor ICD is illustrated in FIG. 1. The first portion of the positive phase may terminate when either: (1) the voltage decreases below a threshold voltage V₃; or (2) a prescribed time period t_ahas elapsed.

[0047] The tissue membrane voltage that results when the waveform of FIG. 1 is applied to excitable cardiac tissue membranes is as shown in FIG. 2. This membrane voltage is obtained by modeling the tissue membranes as taught in the Blair reference, previously cited. As shown in FIGS. 11-20, the optimum duration for t_awill be described in more detail.

[0048] A functional block diagram of the pulse generation circuitry used to generate the biphasic waveform of the two-capacitor ICD is shown in FIG. 3.

[0049] As seen in FIG. 3, a cardiac tissue-stimulating device 10 includes a power source 12, e.g., at least one battery, a timing and control circuit 14, a charging circuit 16, an isolation switch network SW1, a series parallel switch network SW2, at least two capacitors C_Aand C_B, an output switch network SW3, and at least two electrodes 20 and 22. The electrodes 20 and 22 are adapted to be positioned within or on the heart. The electrodes 20 and 22 are connected to the output switch SW3 through conventional leads 21 and 23, respectively.

[0050] A voltage sense amplifier 24 senses the voltage held on the capacitor C_B(which will be the same voltage as capacitor C_Awhen C_Aand C_Bare connected in parallel). In some embodiments of the invention, a current sense amplifier 26 may also be used to sense the current flowing to or returning from one of the electrodes 20 or 22. In FIG. 3, such current is sensed by differentially measuring the voltage across a small current-sense resistor R_Sconnected in series with electrode 22. The outputs of the voltage sense amplifier 24 and the current sense amplifier 26 are directed to the timing and control circuit 14.

[0051] A suitable cardiac activity sensor 28 is also employed within the device 10 in order to detect cardiac activity. The function of the sensor 28 is to sense cardiac activity so that an assessment can be made by the timing and control circuitry whether a defibrillation pulse needs to be generated and delivered to the cardiac tissue. Such sensor 28 may take many forms, e.g, a simple R-wave sense amplifier of the type commonly employed in implantable pacemakers. The details of the sensor 28 are not important for purposes of the present invention.

[0052] The power source 12 is connected to provide operating power to all components and circuitry within the device 10. The power source 12 also provides the energy needed to generate the biphasic defibrillation pulse. That is, energy stored within the power source 12 is used to charge capacitors C_Aand C_B, through the charging circuit 18, up to the desired initial defibrillation starting pulse voltage V₁. Such charging is carried out under control of the timing and control circuit 14. Typically, V₁may be a relatively high voltage, e.g., 350 volts, even though the power source 12 may only be able to provide a relatively low voltage, e.g., 3-6 volts. The charging circuit 16 takes the relatively low voltage from the power source 12 and steps it up to the desired high voltage V₁, using conventional voltage step-up techniques as are known in the art. This stepped-up voltage V₁is then applied through the isolation switch SW1 to both capacitors C_Aand C_Bat a time when C_Aand C_Bare connected in parallel, i.e., when SW2 is in its “P” position, and at a time when the output switch is in its open, or OFF, position. As the capacitors C_Aand C_Bare being charged, the voltage sense amplifier 24 monitors the voltage level on the capacitors. When the desired voltage V₁has been reached, the timing and control circuitry 14 turns off the charging circuit 16 and opens the isolation switch SW1, thereby holding the voltage V₁on capacitors C_Aand C_Buntil such time as a defibrillation pulse is needed.

[0053] When a defibrillation pulse is called for by the timing and control circuit 14, the output switch SW3 is placed in its positive phase position, POS, thereby connecting the parallel connected capacitors C_Aand C_B(on which the starting voltage V₁resides) to the cardiac tissue through the electrodes 20 and 22. Such connection starts the discharge of capacitors C_Aand C_Bthrough the cardiac tissue in accordance with the first time constant τ₁as described above in connection in FIG. 1.

[0054] After a period of time t_a, or as soon as the voltage across the parallel-connected capacitors C_Aand C_Bhas decreased to the threshold value V₂(as sensed by the voltage sense amplifier 24), the timing and control circuit switches SW2 to its series-connected or “S” position, thereby connecting the capacitors C_Aand C_Bin series across the electrodes 20 and 22. Such series connection doubles the voltage across the electrodes 20 and 22 to a value of 2(V₂) Thereafter, the discharge of the series-connected capacitors C_Aand C_Bcontinues through the cardiac tissue in accordance with the second time constant τ₂as described above. This discharge continues until the end of the positive phase.

[0055] The positive or first phase ends at a time t₁from the beginning of the positive phase (as measured by timing circuits within the timing and control circuit 14), or when the voltage has decayed to a value V₃(as sensed by voltage sense amplifier 24). Alternatively, the positive phase may end as a function of the sensed current (as sensed by the current sense amplifier 26), e.g., at a time when the sensed current has decreased from a peak value by a prescribed amount or percentage.

[0056] As soon as the positive phase ends, the timing and control circuit 14 switches the output switch SW3 to the negative phase position, NEG, thereby reversing the polarity of the discharge of the series-connected capacitors C_Aand C_Bthrough the cardiac tissue. The negative phase lasts thereafter for a time period t₂determined by the timing and control circuitry.

[0057] The functions represented by the functional block diagram of FIG. 3 may be implemented by those of skill in the art using a wide variety of circuit elements and components. It is not intended that the present invention be directed to a specific circuit, device or method; but rather that any circuit, device or method which implements the functions described above in connection with FIG. 3 to produce a defibrillation waveform of the general type shown in FIG. 1 be covered by the invention.

[0058] Turning next to FIG. 4, there is shown a simplified schematic diagram of an ICD having three 120 μF capacitors C1, C2 and C3. The manner of charging the capacitors while they are connected in parallel is the same or similar to that shown in FIG. 3. When the capacitors C1, C2 and C3 have been charged to a high voltage, e.g., 370 V, a stored energy of approximately 25 Joules is realized. Once the capacitors have been charged by the ICD, the capacitors are configured for a parallel discharge. This is accomplished by closing switches S1, S2, S3 and S4, while maintaining switches S5 and S6 open. The parallel discharge takes place from time t=0 until a time d₁. Once d₁elapses, one of two options may be used to discharge the remaining charge.

[0059] In accordance with a first option, or Option 1, after d₁has elapsed (i.e., after the capacitors are discharged in parallel until time d₁), all of the capacitors are discharged in series for the remainder of the pulse. This is accomplished by opening S1, S2, S3 and S4 and closing S5 and S6. At a later time, d₂, the “H Bridge” circuit 40 (FIG. 4) is used to reverse the polarity of the output. At yet a later time, d, the output pulse is truncated.

[0060] The waveform generated in accordance with Option 1 is illustrated in FIG. 5. The tissue membrane voltage associated with the waveform of FIG. 5 is modeled and computed, using the Blair model, as shown in FIG. 6. For the example shown in FIGS. 5 and 6, the optimum value of d₁is nominally about 3.5 ms. The optimum choice of d₂is when the elapsed time at d₂is about 1.5 times the elapsed time at d₁, or when the elapsed time at d₂(from t=0) is about 5.25 ms.

[0061] In accordance with a second option, or Option 2, the capacitors C1 and C2 remain in parallel and are in series with C3 until time d₂. This is accomplished by opening S3 and S4 and closing S6. After d₂all the capacitors are in series (S1 and S2 also open, S5 closed) until C3 runs out of charge at a time d₄. After d₄, the diode D₁bypasses the depleted capacitor and the time constant of discharge is of C1 and C2 in series. At a time d₃, where d₂<d₃<d₄, the polarity of the output is reversed using the H Bridge 40. The pulse is truncated at time d. The resulting waveform is shown in FIG. 7. The resulting membrane voltage is modeled and computed and shown in FIG. 8.

[0062] For the example shown in FIGS. 7 and 8, the optimum values of d₁is 2.7 ms, d₂is 1.5 times d₁(or about 4 ms) , d₃is d₂+1.25 ms. The value of d₄is computed to be about 7.6 ms. The choice of d can be in the range of 1.5 to 2.0 times that of d₃.

[0063] With either Option 1 or Option 2, the choice of the values d₁, d₂and d₃are primarily functions of the ICD's capacitance value, the discharge pathway impedance, and the tissue time constant (τ_m).

[0064] The advantage of Option 2 is that the peak waveform voltage is lower than Option 1 yet a minute increase in membrane voltage over Option 1 is achieved. However, Option 1 is simpler to implement and diode D₁is not needed since all the capacitors are discharged equally.

[0065] The advantages of either Option 1 or Option 2 are better appreciated by comparing the results of such discharge, as presented in FIGS. 5, 6, 7 and 8, with the corresponding discharge achieved with a two-capacitor ICD series discharge, as is commonly used in a conventional ICD of the prior art. The discharge waveform achieved with a conventional two-capacitor ICD using series discharge, and the resulting membrane voltage, is shown in FIGS. 9 and 10, respectively. Note, that to store equal energy to the three capacitor ICD, each capacitor of the two-capacitor ICD must have 1.5 times the capacitance value, or two capacitors each with C=180 μF.

[0066] As can be seen from a comparison of FIGS. 9 and 10 with FIGS. 5 and 6 (Option 1), and 5A and 5B (Option 2), for equal stored energy, the value of the peak membrane voltage for Option 2 is 1.18 times higher than the membrane voltage realized using the conventional waveform. Similarly, Option 1 yields a membrane voltage that is 1.17 times higher than is realized using the conventional waveform. In other words, a 25 Joule ICD with three 120μF capacitors and a switching network as in Option 2 performs equally to a 34.4 Joule conventional ICD with two 180μF capacitors. This represents a remarkable improvement in performance.

[0067] As shown in FIG. 11, the two-step waveform has been reproduced. Although identical in nature to that shown in FIG. 1, the designators have been changed slightly for purposes of the in depth analysis that will follow.

[0068] As described above in conjunction with FIG. 3, two capacitors, C_A& C_B, have been charged to the same initial voltage, V₀₁. The system resistance (as seen by device) is given by R_S. For purposes of this discussion, the myocardium has been modeled as a parallel-RC circuit with myocardial tissue time constant, τ_m.

[0069] The amplitude of each step of the positive portion of the defibrillation waveform, shown in FIG. 11, can be characterized with the following basic equations:

[0070] V_s1(t₁)=V₀₁·exp[−t₁/τ_s1] 0≦t₁≦d₁

[0071] V_s2(t₂)=V₀₂·exp[−t₂/τ_s2] 0≦t₂≦d₂

[0072] wherein:

[0073] V_S1is the exponential decay during the first period, t₁, (i.e., Step1);

[0074] V_S2is the exponential decay during the second period, t₂, (i.e., Step2);

[0075] τ_S1is the time constant of C_Aand C_Bin parallel;

[0076] τ_S2is the time constant of C_Aand C_Bin series;

[0077] V₀₁is the initial voltage during Step1 on the capacitors C_Aand C_Bonce fully charged to the source voltage, V₀₁; and

[0078] V₀₂is the initial voltage during Step2 remaining on the capacitors C_Aand C_Bnow configured in series.

[0079] The analysis that follows directly will explain how to determine the absolute and approximate solutions for the optimal durations, d₁and d₂, to maximize induced myocardial potential, V_m(t), when the two capacitors are arranged in a parallel-series, two-step arrangement.

[0080] Consider the myocardial responses to V_s1(t₁) [Step1] and V_s2(t₂) [Step2] separately. Note that the following derivations (Equations 1-4) make absolutely no assumptions regarding any specific relationships between the characteristics of Step1 and Step2.

[0081] The “Step1” myocardial response, V_m1, to the Step1 waveform, V_s1, is described by: $\begin{array}{cc}\frac{d{V}_{\mathrm{m1}}\left(t_1\right)}{d{t}_1}+\frac{V_{\mathrm{m1}}\left(t_1\right)}{{\tau }_m}\propto \frac{V_{\mathrm{s1}}\left(t_1\right)}{{\tau }_m}& \text{(Eq. 1)}\end{array}$

[0082] with the initial condition: V_m1(0)=0.

[0083] The solution to this differential equation is: $\begin{array}{cc}{V}_{\mathrm{m1}}\left(t_1\right)=\{\begin{array}{cc}\frac{V_{01}}{{\alpha }_1}·\left(\exp \left[\frac{-t_1}{{\tau }_{\mathrm{s1}}}\right]-\exp \left[\frac{-t_1}{{\tau }_m}\right]\right)& {\tau }_{\mathrm{s1}}\ne {\tau }_m\\ \frac{V_{01}}{{\tau }_{\mathrm{s1}}}·\left(t_1·\exp \left[\frac{-t_1}{{\tau }_{\mathrm{s1}}}\right]\right)& {\tau }_{\mathrm{s1}}={\tau }_m\end{array}\mathrm{where}{\alpha }_1=1-\left({\tau }_m/{\tau }_{\mathrm{s1}}\right).& \text{(Eq. 2)}\end{array}$

[0084] The “Step2” myocardial response, V_m2, to the Step2 waveform, V_s2, is governed by: $\begin{array}{cc}\frac{d{V}_{\mathrm{m2}}\left(d_1,t_2\right)}{d{t}_2}+\frac{V_{\mathrm{m2}}\left(d_1,t_2\right)}{{\tau }_m}\propto \frac{V_{\mathrm{s2}}\left(t_2\right)}{{\tau }_m}& \text{(Eq. 3)}\end{array}$

[0085] with the initial condition: V_m2(d₁,0)=V_m1(d₁), where d₁represents the final duration of Step1.

[0086] This initial condition ensures that there is a continuity of myocardial voltage when transitioning from the end of Step1 into the start of Step2. The solution to this differential equation is: $\begin{array}{cc}{V}_{\mathrm{m2}}\left(d_1,t_2\right)=V_{\mathrm{m1}}\left(d_1\right)·\exp \left[\frac{-t_2}{{\tau }_m}\right]+\{\begin{array}{cc}\frac{V_{02}\left(d_1\right)}{{\alpha }_2}·\left(\exp \left[\frac{-t_2}{{\tau }_{\mathrm{s2}}}\right]-\exp \left[\frac{-t_2}{{\tau }_m}\right]\right)& {\tau }_{\mathrm{s2}}\ne {\tau }_m\\ \frac{V_{02}\left(d_1\right)}{{\tau }_{\mathrm{s2}}}·\left(t_1·\exp \left[\frac{-t_2}{{\tau }_{\mathrm{s2}}}\right]\right)& {\tau }_{\mathrm{s2}}={\tau }_m\end{array}& \text{(Eq. 4)}\end{array}$

[0087] where α₂=1−(τ_m/τ_s2), and V₀₂is proportional to V_S2(0)

[0088] Equation (4) describes a curve with a single maximum value. The step durations, d₁=d₁^optand d₂=d₂^opt, that maximize this shock-induced myocardial voltage, V_m2(t₁, t₂) can be determined by solving the simultaneous equations given by: $\begin{array}{cc}\frac{\partial V_{\mathrm{m2}}\left(d_1^{\mathrm{opt}},d_2^{\mathrm{opt}}\right)}{\partial d_1^{\mathrm{opt}}}=0\frac{\partial V_{\mathrm{m2}}\left(d_1^{\mathrm{opt}},d_2^{\mathrm{opt}}\right)}{\partial d_2^{\mathrm{opt}}}=0& \text{(Eq. 5)}\end{array}$

[0089] From Equation (5), two equations that describe d₂^optas a function of d₁^optcan be found (the following derivations assume τ_s1≢τ_mand τ_s2≢τ_m): $\begin{array}{cc}{d}_2^{\mathrm{opt}}=\frac{{\tau }_m}{{\alpha }_2}·\ln \left\{1+\left(\frac{{\alpha }_2}{{\alpha }_1}·\frac{V_{01}}{\partial V_{02}/\partial d_1^{\mathrm{opt}}}\right)·\left(\frac{1}{{\tau }_{\mathrm{s1}}}\exp \left[\frac{-d_1^{\mathrm{opt}}}{{\tau }_{\mathrm{s1}}}\right]-\frac{1}{{\tau }_m}\exp \left[\frac{-d_1^{\mathrm{opt}}}{{\tau }_m}\right]\right)\right\}& \text{(Eq. 6)}\end{array}$ $\begin{array}{cc}{d}_2^{\mathrm{opt}}=\frac{{\tau }_m}{{\alpha }_2}·\ln \left\{\frac{{\tau }_{\mathrm{s2}}}{{\tau }_m}\left[1-\left(\frac{{\alpha }_2}{{\alpha }_1}·\frac{V_{01}}{V_{02}\left(d_1^{\mathrm{opt}}\right)}\right)·\left(\exp \left[\frac{-d_1^{\mathrm{opt}}}{{\tau }_{\mathrm{s1}}}\right]-\exp \left[\frac{-d_1^{\mathrm{opt}}}{{\tau }_m}\right]\right)\right]\right\}& \text{(Eq. 7)}\end{array}$

[0090] Setting Equations (6) and (7) equal to each other and simplifying produces the following implicit equation for d₁^opt: $\begin{array}{cc}\left(\frac{{\tau }_m}{{\tau }_{\mathrm{s2}}}·\frac{{\alpha }_1}{V_{01}}\right)=\left(\frac{1/{\tau }_{\mathrm{s1}}}{\partial V_{02}/\partial d_1^{\mathrm{opt}}}+\frac{{\tau }_{\mathrm{s2}}/{\tau }_m}{V_{02}\left(d_1^{\mathrm{opt}}\right)}\right)\exp \left[\frac{-d_1^{\mathrm{opt}}}{{\tau }_{\mathrm{s1}}}\right]-\left(\frac{1/{\tau }_m}{\partial V_{02}/\partial d_1^{\mathrm{opt}}}+\frac{{\tau }_{\mathrm{s2}}/{\tau }_m}{V_{02}\left(d_1^{\mathrm{opt}}\right)}\right)\exp \left[\frac{-d_1^{\mathrm{opt}}}{{\tau }_m}\right]& \text{(Eq. 8)}\end{array}$

[0091] Further simplifications of Equation (8) require that V₀₂(d₁) be explicitly defined.

[0092] When the two system capacitors (C_A& C_B) are configured into a parallel arrangement during Step1 and then reconfigured into a series arrangement during Step2, the system time constants can be explicitly defined as:

[0093] τ_s1=R_S·(C_A+C_B) τ_s2=R_S·(C_AC_B)/(C_A+C_B)   (Eq. 9)

[0094] Furthermore, V₀₂(d₁) is explicitly defined as:

[0095] V₀₂(d₁)=2·V_s1(d₁) =2·V₀₁·exp[−d₁/τ_s1]   (Eq. 10)

[0096] where Equation (10) codifies the notion that, in a parallel-series arrangement, the leading edge voltage of Step2 equals twice the trailing edge voltage of Step1.

[0097] Substituting Equation (10) into Equation (8) and solving explicitly for d₁^optand subsequently d₂^opt[via Equation (6) or (7)] yields: $\begin{array}{cc}{d}_1^{\mathrm{opt}}=-\frac{{\tau }_m}{{\alpha }_1}·\ln \left\{\left(\frac{{\tau }_m}{{\tau }_{\mathrm{s1}}}\right)\left(\frac{2{\alpha }_1-{\alpha }_2}{{\alpha }_1-{\alpha }_2}\right)\right\}& \text{(Eq. 11)}\\ d_2^{\mathrm{opt}}=+\frac{{\tau }_m}{{\alpha }_1}·\ln \left\{\left(\frac{1}{2}\right)\left(\frac{2{\alpha }_1-{\alpha }_2}{{\alpha }_1-{\alpha }_2}\right)\right\}& \text{(Eq. 12)}\end{array}$

[0098] The maximum myocardial voltage attained using these optimal parallel-series step durations can then be determined by substituting Equations (10)-(12) into Equation (4) and simplifying: $\begin{array}{cc}{V}_{\mathrm{m2}}\left(d_1^{\mathrm{opt}},d_2^{\mathrm{opt}}\right)={V_{01}\left(\frac{1}{2}\right)}^{-\frac{1}{{\alpha }_2}}{\left(\frac{{\tau }_m}{{\tau }_{\mathrm{s1}}}\right)}^{\frac{1}{{\alpha }_1}-1}{\left(\frac{2{\alpha }_1-{\alpha }_2}{{\alpha }_1-{\alpha }_2}\right)}^{\frac{1}{{\alpha }_1}-\frac{1}{{\alpha }_2}}& \text{(Eq. 13)}\end{array}$

[0099] Note that Equations (11)-(13) are valid for any independent values of C_Aand C_B.

[0100] According to this simple RC model of defibrillation, successful defibrillation is achieved when the myocardial voltage (as embodied herein by V_m1and V_m2) is “depolarized” to its threshold value, V_th. An equation that describes the minimum relative magnitude for V₀(i.e., the voltage to which each of the capacitors is charged in preparation for the defibrillation shock) that successfully drives V_m2to V_thcan be obtained from Equation (13) by setting V_m2=V_thand solving for V₀₁(which, for these parallel-series shocks, is equivalent to V₀).

[0101] Since the total stored energy in capacitors C_Aand C_Bis given by: $\begin{array}{cc}{E}_{\mathrm{stored}}=\frac{1}{2}\left(C_A+C_B\right)·V_0^2& \text{(Eq. 14)}\end{array}$

[0102] then the optimal relationship between C_Aand C_Bthat maximizes myocardial voltage for a given total stored energy can be found by substituting C_A=k·C_Binto Equation (14) and then solving for k in ∂E_stored/∂k=0. The result is:

[0103] k^opt=C_A/C_B=1   (Eq. 15)

[0104] The above result implies that C_Ashould equal C_Bin order to achieve maximum myocardial impact for any given total energy. The relationship C_A=C_Bis equivalent to τ_s1=4·τ_s2[see Equation (9)], from which simplified versions of Equations (1l)-(13) can be derived: $\begin{array}{cc}{d}_1^{\mathrm{opt}}=-\frac{{\tau }_m}{{\alpha }_1}·\ln \left\{\left(\frac{1}{3}\right)\left(1+\frac{{\tau }_m}{2{\tau }_{\mathrm{s2}}}\right)\right\}& \text{(Eq. 16)}\\ d_2^{\mathrm{opt}}=+\frac{{\tau }_m}{{\alpha }_2}·\ln \left\{\left(\frac{1}{3}\right)\left(1+\frac{2{\tau }_{\mathrm{s2}}}{{\tau }_m}\right)\right\}& \text{(Eq. 17)}\\ V_{\mathrm{m2}}\left(d_1^{\mathrm{opt}},d_2^{\mathrm{opt}}\right)=2{{\left(V_{01}\left(\frac{{\tau }_m}{2{\tau }_{\mathrm{s2}}}\right)\right)}^{\frac{1}{{\alpha }_2}-1}\left[\left(\frac{1}{3}\right)\left(1+\frac{{\tau }_m}{2{\tau }_{\mathrm{s2}}}\right)\right]}^{\frac{1}{{\alpha }_1}-\frac{1}{{\alpha }_2}}& \text{(Eq. 18)}\end{array}$

[0105] Finally, the optimal capacitance for a given R_sand τ_mis determined by finding the value of C_Athat minimizes E_stored, that is, solving for C_Ain ∂E_stored/∂C_A=0 (with k=1). The result is: $\begin{array}{cc}{C}_A=C_B=\frac{{\tau }_m}{R_s}& \text{(Eq. 19)}\end{array}$

[0106] or equivalently, the optimal capacitance (for a given R_sand τ_m) is that which satisfies: $\begin{array}{cc}\frac{1}{2}{\tau }_{\mathrm{s1}}=2{\tau }_{\mathrm{s2}}={\tau }_m& \left(\mathrm{Eq}.20\right)\end{array}$

[0107] Under these ideal conditions, the optimal step durations are:

[0108] d₁^opt=+2τ_m·1n[3/2]≈0.811·τ_m  (Eq. 21)

[0109] d₂^opt=+τ_m·1n[3/2]≈0.405·τ_m  (Eq. 22)

[0110] Further insights into the preceding theoretical calculations can be gleaned from corresponding graphical analyses. The relative stored energy required for defibrillation (E_stored) for all possible parallel-series two-step waveforms is graphically illustrated in the contour plot of FIG. 12. In this plot, the x-axis is indexed by the total capacitance (C_A+C_B, scaled by τ_m/R_s) while the y-axis is indexed by the ratio of the two capacitances (k=C_A/C_B). Although perhaps seemingly non-intuitive axis definitions, they efficiently provide complete coverage of the entire parameter space of all possible capacitor combinations for two-step waveforms. As indicated by the horizontal line 100 and the vertical line 102 overlaid on this plot (and as consistent with the conclusions of Equations (15) and (19)), the most efficient two-step positive portion for the biphasic shock is delivered when:

[0111] k=1.0; and

[0112] C_A+C_B=2·τ_m/R_s;

[0113] which occurs at point 104 in FIG. 12.

[0114] The contours then step out from this optimal point in 1% increments, thus providing an indication as to the relative sensitivity of the energy efficiency to deviations in either total capacitance or capacitance ratio. In fact, energy efficiency remains quite robust: for example, energy efficiency remains within 1% of optimal for:

[0115] ˜1.5·τ_m/R_s<(C_A+C_B)<˜2.7·τ_m/R_S; and

[0116] ¹⁸0.7<k<¹⁸1.4.

[0117] Two-dimensional contour plots of optimal Step1 and Step2 durations (normalized by τ_m, i.e., d₁^opt/τ_mand d₂^opt/τ_m) as given by Equations (11) and (12) are presented in FIGS. 13 and 14, respectively.

[0118] Similar to FIG. 12, FIGS. 13 and 14 have respective horizontal lines 110, 120 and vertical lines 112, 122 from have been overlaid on these contour maps as well. Their respective intersections 114, 124 appropriately correspond to the “0.811” and “0.405” coefficients found in Equations (21) and (22), respectively.

[0119] Since R_Sand τ_mrepresent patient-specific variables that directly impact the choice of durations used for these stepped waveforms, it is perhaps useful to present example values for d₁^optand d₂^optfor a representative range of values for R_s(30-90 Ω), τ_m(2-4 ms), and C_A(30-90 μF). The tables shown in FIGS. 15-17 provide such a set of example values, wherein values for d₁^optand d₂^optare computed from Equations (16) and (17), respectively.

[0120] Given the limits of the ranges used for R_s, τ_m, and C_Ain the tables shown in FIGS. 15-17, d₁^optand d₂^optrange from lows of 1.286 and 0.422 ms (when τ_m=2 ms, C_A=30 μF, and R_s=30 Ω) to highs of 3.704 and 2.689 ms (when τ_m=4 ms, C_A=90 μF, and R_s=90 Ω), respectively.

[0121] To summarize the above, for the ranges of:

[0122] τ_m=2-4 ms;

[0123] R_s=30-90 Ω;

[0124] C_A=C_B=30-90 μF

[0125] Then, the optimum durations fall in the ranges:

[0126] d₁^opt=1.286-3.704

[0127] d₂^opt=0.422-2.689

[0128] Of course, d₁^optand/or d₂^optcould move outside of these ranges if any one or more of R_s, τ_m, and C_Aexceed the limits used for these tables. In those cases, Equations (16) and (17) could be used to compute exactly the optimal step durations for any combination of R_s, τ_m, and C_A.

[0129] In another embodiment, the device could also determine d₁^optand d₂^optbased on measured values for R_s, and/or a programmed value for τ_m, based on a particular value for C_Aand C_B.

[0130] By way of example, if the capacitance value for C_Aand C_Bis set to 60 μF, so that Equation 19 is satisfied for a tissue resistance, R_sequal to nominally 50 ohms and a tissue time constant, τ_m, then for a range for τ_m, of 2 ms to 4 ms, and a range for R_sof 30-90 ohms, then:

[0131] If τ_m=2.0 ms and Rs=90 ohms, then:

[0132] C_A+C_B)*R_s/τ_m=5.4

[0133] d₁^opt=0.962 * τ_m(=1.923 ms)

[0134] d₂^opt=0.809 * τ_m(=1.618 ms)

[0135] If τ_m=4.0 ms and Rs=30 ohms, then:

[0136] (C_A+C_B)*R_s/τ_m=0.9

[0137] d₁^opt=0.643 * τ_m(=2.573 ms)

[0138] d₂^opt=0.211 * τ_m(=0.844 ms)

[0139] To further assist with interpreting the results embodied in FIGS. 13 and 14 and the table shown in FIGS. 15-17, FIG. 18 graphs a subset of those data as simple functions of R_sand τ_m. In particular, FIG. 18 presents a pair of graphs: the left and right halves plot d₁^optand d₂^opt, respectively, as functions of R_sfor three representative values of τ_m(2, 3, and 4 ms). For these graphs, C_A=C_B=60 μF (thus k=1.0). Consistent with the data in the tables shown in FIGS. 15-17 both d₁^optand d₂^optincrease in value with increasing R_sor τ_m. Moreover, this figure helps illustrate how d₁^optappears significantly more sensitive to relative changes in τ_mthan in R_s, while d₂^optappears to have the opposite sensitivity.

[0140] While FIGS. 12-17 provide a comprehensive overview of all possible parallel-series two-step waveforms, it is also useful to consider some specific examples that can aid in illustrating the relative improvements gained by using such a parallel-series two-step capacitor arrangement over the traditional one-step arrangement.

[0141] FIG. 19 graphically compares the positive portion of the biphasic shock waveform shapes (V_s, top two waveforms, 150 and 160) and associated tissue responses (V_m, bottom two waveforms, 152 and 162) for one-step, 150, and parallel-series two-step, 160, shocks having equal stored energies and leading-edge voltages.

[0142] For this example, shown in FIG. 19:

[0143] τ_m=3 ms, R_s=50 Ω, C_A=C_B=60 μF

[0144] (thus, Equations 15 & 19 are satisfied).

[0145] The one-step shock is generated by essentially keeping C_Aand C_Bin a parallel arrangement for its entire shock duration, for a constant effective capacitance of 120 μF. As is evident from the tissue responses (i.e., comparing the one-step response 152 to the two-step response 162), two-step the myocardial voltage (162) reaches a higher higher final cell membrane potential (+18.6%) in a shorter total duration (3.65 vs. 4.16 ms 12.3%) as compared to the final cell membrane potential (152) using the one-step shock. A consequence of this improved tissue response is that this two-step waveform requires a lower effective leading-edge voltage (and hence a lower stored energy) to achieve the same defibrillation efficacy as its equivalent one-step waveform.

[0146] FIG. 20 illustrates this scenario by resealing the results presented in FIG. 19 such that the strength of each shock is sufficient to produce tissue responses of equal amplitudes. Consistent with the results presented in FIG. 19, this two-step positive portion of the biphasic shock waveform 164 theoretically requires a 15.6% lower leading-edge voltage than its one-step counterpart 154, which translates into a 28.8% reduction in required stored energy, and a potentially lower pain waveform for the patient since the leading edge of the shocking pulse is reduced.

[0147] FIGS. 21 and 22 illustrate analogous results to those depicted in FIG. 20, but for relatively extreme combinations of R_sand C_A. In FIG. 21, R_s=30 Ω and C_A=C_B=30 μF, while in FIG. 22, R_s=90 Ω and C_A=C_B=90 μF. As is evident in FIGS. 21 and 22, the shape of the optimal parallel-series two-step waveform depends strongly on the magnitudes of R_sand C_A. Furthermore, the relative improvement in energy efficiency also strongly depends on these values.

[0148] For example, in FIG. 21, the two-step waveform 166 induced an equivalent final tissue response as its one-step waveform 156, but with an 8.8% shorter duration (2.1 vs. 2.3 ms), a 6.5% lower leading-edge voltage, and a 12.6% reduction in required stored energy.

[0149] In FIG. 22, the relative improvements were a 14.3% shorter duration (5.3 vs. 6.3 ms), a 25.9% lower leading-edge voltage, and a 45.0% reduction in required stored energy. Thus, these comparisons suggest that there would be especially great incentive for utilizing two-step waveforms instead of traditional one-step waveforms when the magnitudes of R_sand C_Aare large, while the incentive is relatively minimal when the magnitudes of R_sand C_Aare small. Unfortunately, because of the inherent limitations of this theoretical model, it is not possible to directly compare amplitude-based results (e.g., leading-edge voltage, required stored energy) derived for differing R_sor τ_m. For this reason, the results of FIGS. 20-22 are all self-normalized (that is, there is no relationship between the amplitudes in these graphs).

[0150] Finally, while Equations (16) and (17) provide exact formulas for determining d₁^optand d₂^optwhen k=1 (i.e. , C_A=C_B) , it is sometimes helpful and/or practical to also identify various approximations to such solutions. Consider the following infinite series expansion of the natural logarithm: $\begin{array}{cc}\ln \left[x\right]=2·[\left(\frac{x-1}{x+1}\right)+\frac{1}{3}·{\left(\frac{x-1}{x+1}\right)}^3+\frac{1}{5}·{\left(\frac{x-1}{x+1}\right)}^5+\dots ]& \left(23\right)\end{array}$

[0151] Utilizing just the first term of this expansion, Equations (16) and (17) can be simplified to: $\begin{array}{cc}{d}_1^{\mathrm{opt}}\approx \frac{2{\tau }_m}{3-{\alpha }_1}=\frac{2{\tau }_{\mathrm{s1}}·{\tau }_m}{2{\tau }_{\mathrm{s1}}+{\tau }_m}\Rightarrow \frac{1}{d_1^{\mathrm{opt}}}\approx \frac{1}{2{\tau }_{\mathrm{s1}}}+\frac{1}{{\tau }_m}=\frac{1}{4R_s{C}_A}+\frac{1}{{\tau }_m}& \left(24\right)\\ d_2^{\mathrm{opt}}\approx \frac{2{\tau }_m}{3-2{\alpha }_2}=\frac{{\tau }_{\mathrm{s2}}·2{\tau }_m}{{\tau }_{\mathrm{s2}}+2{\tau }_m}\Rightarrow \frac{1}{d_2^{\mathrm{opt}}}\approx \frac{1}{{\tau }_{\mathrm{s2}}}+\frac{1}{2{\tau }_m}=\frac{1}{2}·\left(\frac{4}{R_s{C}_A}+\frac{1}{{\tau }_m}\right)& \left(25\right)\end{array}$

[0152] In words, these relationships suggest that the optimal step durations can be well approximated by computing variously weighted parallel combinations of system and myocardial time constants. And despite using only one term of Equation (23), these approximations are relatively quite accurate over a broad range of τ_s1/τ_mand τ_s2/τ_mratios (only their ratios, not their absolute values, impact their accuracy). For example, the relative error for d₁^optis less than 5% for 0.4<τ_s1/τ_m<5, while the relative error for d₂^optis less than 5% for 0.2<τ_s2/τ_m<3. When Equation (20) is also satisfied (that is, when system and myocardial time constants are ideally matched), these relative errors are each only 1.35%. In all cases, these approximation calculations underestimate the true values by these respective relative errors.

[0153] While the invention herein disclosed has been described by means of specific embodiments and applications thereof, numerous modifications and variations could be made thereto by those skilled in the art without departing from the scope of the invention set forth in the claims.