Multi-path transthoracic defibrillation and cardioversion

Abstract

A defibrillation system for synchronized cardioversion of a patient includes a first housing that includes a measurement circuit configured to receive electrocardiogram (ECG) signals and measure ECG parameters based on the ECG signals, and a first processor configured to analyze the ECG parameters, and initiate communication of a synchronization signal for a second processor for delivery of one or more defibrillation pulses and further includes a second housing that is separate from and external to the first housing and that includes a shock delivery circuit, and the second processor which is configured to receive the communication of the synchronization signal from the first processor, and control the shock delivery circuit to deliver the one or more defibrillation pulses in response to the synchronization signal.

Claims

1. A defibrillation system for synchronized cardioversion of a patient comprising: a first housing comprising: a first shock delivery circuit configured to be controlled by a first processor for delivery of a first shock; a measurement circuit configured to receive electrocardiogram (ECG) signals and measure ECG parameters based on the ECG signals, and the first processor configured to: analyze the ECG parameters, and initiate communication of a synchronization signal for a second processor for delivery of a second shock; and a second housing that is separate from and external to the first housing, the second housing comprising: a second shock delivery circuit, and the second processor, wherein the second processor is configured to: receive the communication of the synchronization signal from the first processor, and control the second shock delivery circuit to deliver the second shock in response to the synchronization signal.

2. The defibrillation system of claim 1 wherein the delivery of the second shock is based on the ECG parameters.

3. The defibrillation system of claim 1 comprising: three or more electrodes configured to attach to the patient and provide two or more discharge pathways; and at least one measurement circuit configured to: couple to the three or more electrodes, and measure one or more electrical parameters of the patient, wherein at least one of the first processor and the second processor is configured to: based on the one or more measured electrical parameters, determine internal impedances, for the patient, corresponding to the two or more discharge pathways, and control at least one parameter of at least one of the first shock and the second shock based on the determined internal impedances.

4. The defibrillation system of claim 3 wherein the one or more electrical parameters comprise a current that passes through the patient.

5. The defibrillation system of claim 3 wherein the one or more electrical parameters comprise an induced voltage at a plurality of locations on the patient.

6. The defibrillation system of claim 3 wherein the at least one of the first processor and the second processor is configured to independently control the at least one parameter for each of the first shock and the second shock based on the internal impedances.

7. The defibrillation system of claim 3 wherein the at least one parameter comprises at least one of amplitude, tilt, duration first phase duration, second phase duration, current, voltage, and first phase average current.

8. The defibrillation system of claim 3 wherein at least the first processor is configured to use electrical impedance tomography (EIT) and determine the internal impedances as an internal impedance distribution.

9. The defibrillation system of claim 8 wherein one or more of the first processor and the second processor is configured to determine a predicted response of the myocardium of the patient to at least one of the first shock and the second shock based on a model and the internal impedance distribution.

10. The defibrillation system of claim 9 wherein the model comprises a biodomain model.

11. The defibrillation system of claim 9 wherein the predicted response comprises an indication of one or more excitable gap regions of the myocardium of the patient.

12. The defibrillation system of claim 11 wherein at least the second processor is configured to control at least one of the first shock delivery circuit and the second shock delivery circuit to deliver one or more electrical pulses in succession prior to at least one of the first shock and the second shock wherein a current density from the one or more electrical pulses is configured to be directed towards the one or more excitable gap regions of the myocardium of the patient via a predetermined physical arrangement of the three or more electrodes.

13. The defibrillation system of claim 3 wherein the measurement circuit comprises individual ECG monitoring channels and further wherein at least a portion of the three or more electrodes are coupled to the individual ECG monitoring channels.

14. The defibrillation system of claim 13 wherein one or more of the first processor and the second processor is configured to determine an epicardial activation wavefront distribution.

15. The defibrillation system of claim 13 wherein one or more of the first processor and the second processor is configured to determine an epicardial activation wavefront path.

16. The defibrillation system of claim 13 comprising at least sixteen ECG monitoring channels and at least sixteen electrodes.

17. The defibrillation system of claim 16 wherein the first processor is configured to sample ECG signals at a sample rate of approximately 250 Hz.

18. The defibrillation system of claim 13 wherein the three or more electrodes are distributed between at least one anterior electrode pad assembly and at least one posterior electrode pad assembly and further wherein the one or more of the first processor and the second processor is configured to sample the ECG signals with alternating samples from the at least one anterior electrode pad assembly and the at least one posterior electrode pad assembly.

Description

DESCRIPTION OF DRAWINGS

(1) FIG. 1 is a schematic of the circuitry for a biphasic defibrillator implementation.

(2) FIG. 2 is a block diagram of the implementation of FIG. 1.

(3) FIG. 3 is a plot of a biphasic waveform produced by the implementation of FIG. 1.

(4) FIG. 4 is a cross section of the thorax at the elevation of the heart used in finite element modeling showing the finite element mesh decomposition.

(5) FIG. 5a is a simplified version of FIG. 4 with electrode locations and coordinate axes and plans.

(6) FIG. 5b is a simplified version of FIG. 4 showing a preferred set of electrode locations.

(7) FIG. 5c is a simplified version of FIG. 4 showing an alternative set of electrode locations.

(8) FIG. 5d is a simplified version of FIG. 4 showing a further alternative set of electrode locations.

(9) FIG. 6 shows an isoconductance plot of the anterior thorax.

(10) FIG. 7 shows the placement of the anterior electrodes of FIG. 5b relative to the heart.

(11) FIGS. 8a and 8b show the placement of the electrodes of FIG. 5b,

(12) FIG. 9 shows an example of an annular electrode configuration.

(13) FIG. 10 shows examples of electrodes arranged in regular tessellations.

(14) FIG. 11 shows examples of electrodes arranged in semi-regular tessellations.

(15) FIG. 12 shows examples of electrodes arranged in demi-regular tessellations.

(16) FIG. 13 shows a simulation of the effects on the heart due to a monophasic defibrillation pulse as modeled using the Virtual Electrode Theory.

(17) FIG. 14a, 14b depict, diagrammatically, what occurs when the area of the electrodes is varied.

(18) FIG. 15 is an electrical schematic of an H-Bridge Class D configuration circuit.

(19) FIG. 16 is an electrical schematic of circuitry driving an H-bridge Class D configuration circuit.

(20) FIG. 17 are waveforms produced by the H-bridge Class D configuration circuit.

(21) FIG. 18 is a plan view of an integrated defibrillation pad including an ultrasonic gel window for application of an ultrasonic probe.

(22) FIG. 19 is a block diagram of an implementation with dual defibrillators.

(23) FIG. 20 is a plan view of two integrated defibrillation electrode pads, electrically connected with a common connector in a dual defibrillator system.

(24) FIG. 21 is a perspective view of a patient to which an integrated defibrillation pad with anatomical markings has been applied.

(25) FIG. 22 is a block diagram of an integrated resuscitation system implementation.

(26) FIG. 23 is a decision flowchart for the integrated resuscitation system.

(27) FIGS. 24A and 25A show two different examples of the spatial distribution of epicardial voltages across a region of the heart immediately subsequent to defibrillation shock. Different shades of gray represent different voltage levels, with the darkest region at −90 mV (−50 mV in FIG. 25A) at the extreme left in the figure indicating the region of greatest depolarization, and the darkest region at 0 mV (+10 mV in FIG. 25A) at the extreme right indicating the region of greatest hyperpolarization. The white region and the bands immediately adjacent are the excitable gap. Isopotential contour lines are drawn 5 mV apart on the basis of the assumption that the normal resting potential is 285 mV and action potential amplitude is 100 mV.

(28) FIGS. 24B and 25B show two different examples of the spatial distribution of epicardial phase across a region of the heart immediately following a defibrillation shock. Different shades of gray represent different isochronous regions (i.e., regions at approximately the same cardiac phase). The regions are 10 msec apart, with the region in white representing the time immediately after the defibrillation shock (0 msec), and the region in the darkest gray representing 120 msec after the shock.

(29) FIGS. 24C and 25C are ECG signals representing possible behavior immediately after a defibrillation pulse, and show onset of tachycardia.

(30) FIG. 25D is the same as FIG. 25A but with the excitable gap region shown as a cross-hatched area.

(31) FIG. 25E is the same as FIG. 25B but with a cross hatched area representing the area of an applied quenching stimulus.

(32) FIG. 26 provides a simplified, perspective view of the ventricles, showing the generally spiral orientation of the muscle fibers.

(33) FIG. 27 provides a simplified, perspective view of the ventricles, showing two possible current vectors (for a two-vector approach) generally aligned with the muscle fibers.

(34) FIGS. 28A and 28B show an implanted defibrillator and electrodes that provide another implementation of the two vector approach in which the vectors are aligned with muscle fibers.

(35) FIG. 29 shows an electrical schematic of the implementation of FIG. 28B.

DETAILED DESCRIPTION

(36) There are a great many possible implementations of the invention, too many to describe herein. Some possible implementations that are presently preferred are described below. It cannot be emphasized too strongly, however, that these are descriptions of implementations of the invention, and not descriptions of the invention, which is not limited to the detailed implementations described in this section but is described in broader terms in the claims.

(37) One implementation of the invention is depicted in FIG. 1. The defibrillation waveform delivered to the patient is a biphasic or multiphasic waveform as described in U.S. Pat. No. 6,096,063. As described in that patent, the electromagnetic (EM) energy delivery means 1 is comprised of storage capacitors 2, 3 which are charged to a therapeutically effective voltage by a charging circuit 4 under control of the processing means 5 while relays 6, 7, 8 and 9 and the H-Bridges 10, 11 are open. As a means of reducing both size and cost, charging circuit 4 is used to charge both storage capacitors 2, 3 simultaneously. The first electrode pair 1 and the second electrode pair 2 are self adhesive pads, such as STAT-PADZ (ZOLL Medical Chelmsford Mass.), that are adhered to the patient's chest 3, shown in cross-section in FIG. 1.

(38) Upon determination by processing means 5, using any existing methods known to those skilled in the art, of the appropriate time to deliver the defibrillation energy to the patient, relay switches 12, 13, 14 and 15 are opened, and relay switches 6, 7, 8 and 9 are closed. Then, the electronic switches 16, 17, 18, and 19 of H-bridge 10 and 24, 25, 26, and 27 of H-bridge 11 are closed to allow electric current to pass through the patient's body in one direction, after which electronic switches 16, 17, 18, and 19 of H-bridge 10 and 24, 25, 26, and 27 of H-bridge 11 are opened and 20, 21, 22, and 23 of H-bridge 10 and 28, 29, 30 and 31 of H-bridge 11 are closed to allow the electric current to pass through the patient's body in the other direction. Relay switches 12, 13, 14 and 15 are combined in double-pole double-throw configuration (DPDT) to reduce size and cost. DPDT relay 12, 13 serves the purpose of isolating the current sources for the electrode pairs during discharge. Electronic switches 16-31 are controlled by signals from respective opto-isolators, which are, in turn, controlled by signals from the processing means 5. As shown in FIG. 2, processing means 5 is preferably a microprocessor, such as a Hitachi SH-3 40 combined with a read only memory device (ROM) 41, random access memory (RAM) 42, Clock 43, real time clock 44, analog-to-digital 45 and digital-to-analog 46 converters, power supply 47, reset circuit 48, general purpose input/output 49, and user interface in the form of a display 49 and input keys 50 and other circuitry known to those skilled in the art. A measurement means 52 is provided for measurement of electrical, electrocardiographic, physiological or anatomical parameters of the patient, the processing means 5 controlling the waveform parameters of at least one of the discharge pathways based on this measurement. Relay switches 6, 7, 8, and 9 which are also controlled by the processing means 5, isolate patient 3 from leakage currents of H-bridge switches 16-31 which may be about 500 microamperes.

(39) Resistive circuits 55, 56 that include series-connected resistors 57, 58, 59 and 60, 61, 62, respectively, are provided in the current path, each of the resistors being connected in parallel with shorting switch 63-68 controlled by processing means 5. The resistors are preferably of unequal value and stepped in a binary sequence such that with the various combinations of series resistance values, there are 2.sup.n different combinations, where n is the number of resistors. Immediately prior to delivering the therapeutic defibrillation energy a smaller amplitude “sensing” pulse is delivered by closing H-bridge switches 16-19 and 24-27 and the resistor shorting switches 63-68 are all open so that current passes through the resistors in series. The current sensing transformers 69 and 70 sense the current that passes through the patient through their respective electrode pairs 1a, 1b, 2a and 2b, from which the processing means 5 determines the resistance of the patient 3.

(40) The initial sensing pulse is integral with, i.e., immediately followed by, a biphasic defibrillation waveform, and no re-charging of storage capacitor occurs between the initial sensing pulse and the biphasic defibrillation waveform. If the patient resistance sensed during the initial sensing pulse is low, all of the resistor-shorting switches 63-68 are left open at the end of the sensing pulse so that all of the resistors 57-62 remain in the current path (the resistors are then successively shorted out during the positive phase of the biphasic defibrillation waveform in the manner described below in order to approximate a rectilinear positive phase). Thus, the current at the beginning of the positive first phase of the biphasic defibrillation waveform is the same as the current during sensing pulse. If the patient resistance sensed during the sensing pulse is high, some or all of the resistor-shorting switches 63-68 are closed at the end of the sensing Pulse, thereby shorting out some or all of the resistors.

(41) Thus, immediately after the sensing pulse, the biphasic defibrillation waveform has an initial discharge current that is controlled by microprocessor 46, based on the patient impedance sensed by current-sensing transformer 69, 70. The current level of the sensing pulse is always at least 50 percent of the current level at the beginning of positive first phase, and the sensing pulse, like the defibrillation pulse, is of course a direct-current pulse.

(42) By appropriately selecting the number of resistors that remain in the current path, the processing means reduces (but does not eliminate) the dependence of peak discharge current on patient impedance, for a given amount of charge stored by the charge storage device. For a patient impedance of 15 ohms, the peak current is about 25 amperes, whereas for a patient impedance of 125 ohms, the peak current is about 12.5 amperes (a typical patient is about 75 ohms.)

(43) During the positive phase of the biphasic waveform, some or all of the resistors 57-62 that remain in series with the patient 3 are successively shorted out. Every time one of the resistors is shorted out, an upward jump in current occurs in the waveform, thereby resulting in the sawtooth ripple shown in the waveform of FIG. 3. The ripple tends to be greatest at the end of the rectilinear phase because the time constant of decay (RC) is shorter at the end of the phase than at the beginning of the phase. Of course, if all of the resistors have already been shorted out immediately after the end of the sensing pulse, the positive phase of the biphasic waveform simply decays exponentially until the waveform switches to the negative phase.

(44) As is shown in FIG. 3, at the end of the positive phase, the current waveform decreases through a series of rapid steps from the end of the positive phase to the beginning of negative phase, one of the steps being at the zero crossing. Processing means 5 accomplishes this by (1) successively increasing the resistance of resistive circuit 55, 56 in fixed increments through manipulation of resistor-shorting switches 57-62, then (2) opening all of the switches in H-bridges 10-11 to bring the current waveform down to the zero crossing, then (3) reversing the polarity of the current waveform by closing the H-bridge switches that had previously been open in the positive phase of the current waveform, and then (4) successively decreasing the resistance of resistance circuit 55, 56 in fixed increments through manipulation of resistor-shorting switches 57-62 until the resistance of resistance circuit 55, 56 is the same as it at the end of the positive phase.

(45) In one implementation a variable resistor 71, 72 is provided in series with the other resistors 57-62 to reduce the sawtooth ripple. Every time one of the fixed-value resistors 57-62 is shorted out, the resistance of variable resistors 71, 72 automatically jumps to a high value and then decreases until the next fixed-value resistor is shorted out. This tends, to some extent, to smooth out the height of the sawtooth ripple from about 3 amps to about 0.1 to 0.2 amps, and reduces the need for smaller increments of the fixed-value (i.e., it reduces the need for additional fixed-value resistor stages).

(46) A cross-sectional view of the human thorax is shown in FIG. 4. Each of the constituent tissues are subdivided into cells for use in finite element simulations of the fields and currents generated by defibrillation pulses. Electrode pairs 1a, 1b, 2a, and 2b are also depicted in the figure. FIG. 5a-d depicts a simplified version of the cross section of FIG. 4. A line is defined in the figure, the Cardiac Center of Mass (CCOM) line 75, which runs through the CCOM point 76 and is parallel to the patient's back. In preferred implementations, at least one, and preferably two, electrodes are located posterior to the CCOM line. Additionally, the midpoint/COM (MCOM) line 77 is the line defined by midpoint of the lateral extent (MLE) of the posterior electrode or electrodes 78, 79 and the CCOM point 76. The electrode plane 81, is defined by the plane resulting in the least mean squared error distance to the centroids 82 of the electrodes distal 84 to the MLE 78. There is further defined a Projected Cardiac Area (PCA), that is the area in the electrode plane 81 of the shape formed by the intersection of the electrode plane 81 with the locus of lines 80 parallel to the MCOM line 77 and tangent to the surface of the heart 83. The area, shape and position of the electrodes are such that the area of each individual electrode is less than 70% of the PCA and the sum of the areas of the electrodes distal 84 to the MLE 78 is greater than 80%, and preferably 100%, of the PCA.

(47) In a preferred implementation, the electrodes are positioned as shown in FIGS. 5b, 7, 8a and 8b. FIG. 7 shows the relative location of the electrodes 1a, 2a and the thoracic cage and the heart 4. FIGS. 8a and 8b show the electrode placement on a typical patient. FIG. 5c depicts a lateral placement of the electrode pairs. In another implementation, the electrodes may be so configured as concentric, as shown in FIG. 9. The electrodes may also be placed so that the current pathways are essentially parallel, as shown in FIG. 5d (in which the locations of electrodes 1b and 2b have been reversed from FIG. 5b).

(48) The conductances of the various tissues as shown in FIG. 4 are approximately as follows:

(49) TABLE-US-00001 Tissue type Conductivity (ohms-cm) Skin 3.4 Blood 6.5 Lung 0.7 Skeletal Muscle 1.5 (transverse) 4.2 (longitudinal) Fat 0.5 Cardiac Muscle 7.6 Bone 0.06

(50) Conductivities of the various tissues can vary by as much as a factor of 100. To accommodate this, waveform parameters of the energy delivered to each of the discharge pathways is independently controllable. For example, this may be accomplished in the just-described implementation by providing two high voltage capacitors 2, 3 and by appropriately switching the resistors 57-62 that remain in series with the patient 3. By appropriately selecting the number of resistors that remain in the current path, the dependence of peak discharge current on patient impedance can be reduced (but not eliminated), for a given amount of charge stored by the charge storage device. For example, for a patient impedance of 15 ohms, the peak current is about 25 amperes, whereas for a patient impedance of 125 ohms, the peak current is about 12.5 amperes (a typical patient is about 75 ohms.)

(51) Alternatively, independent control may also be achieved by providing only one high voltage capacitor for more than one of the electrode pairs while still providing separate resistor networks 57-59 and 60-62 for each current pathway. Another waveform parameter that may be adjusted is waveform duration, which is controllable by switch networks 10, 11. The average first phase current can also be independently adjusted, e.g., by providing a second charging circuit 4 to charge a second group of one or more capacitors to a voltage independent from the first group of one or more capacitors. Waveform parameters for independent adjustment include, but are not limited to, tilt, duration, first phase duration, second phase duration, current, voltage, and first phase average current.

(52) As can be seen in the isoadmittance curves shown in FIG. 6, the conductances of the internal organs, muscle and bone vary significantly, much more so than do conductances at the body surface. In a preferred implementation, electrical impedance tomography (EIT) is used to determine these internal conductances or impedances. Electrical impedance tomography (EIT) is used to determine the resistivity distribution of the thorax in at least two dimensions, and the calculated resistivity distribution is then used to determine the waveform parameters for each current vector. For example, the amplitude of the defibrillation pulse for each electrode pair can be independently adjusted to achieve the optimal current distribution in and around the myocardium. Using such a method, the current actually delivered to the organs themselves can be controlled at the surface of the body on as fine a level of detail as determined by the number, location and size of the electrodes located on the body surface.

(53) In the most basic implementation, only three electrodes with three possible electrode pairs is sufficient to use EIT methods to determine waveform parameters. In the preferred implementation shown in FIG. 5b, four electrodes are used, for a total of six [(n−1)!] possible electrode pairs. This number is chosen for ease of implementation and cost; implementations with more electrode pairs are possible.

(54) The EIT system is governed by Poisson's equation:
∇.Math.ρ.sup.−1∇V=I,

(55) Where V is the voltage, ρ is the resistivity distribution and I is the impressed current source distributions within the region being studied and the boundary conditions are V.sub.0 and J.sub.0. In the case of EIT, high frequency, low amplitude signals, e.g., 60 KHz and ˜1 microampere respectively, are used. Since there are no current sources of this frequency in the body, then ρ=0, and Poisson's equation becomes Laplace's equation:
∇.Math.ρ.sup.−1∇V=0

(56) In the field of EIT, several types of problems are studied: 1. The “forward problem”, where ρ, V.sub.0 and J.sub.0 are given and the goal is to determine the voltage and current distributions V and J. 2. The “inverse problem”, where V and J are given and the goal is to determine ρ. 3. The “boundary value” problem where V.sub.0 and J.sub.0 are given and the goal is to determine ρ, V and J.

(57) In a preferred implementation, ρ, V and J are determined using boundary value problem methods, then once ρ is determined, the optimal V.sub.0 and J.sub.0 are determined using a modified inverse problem where the desired V and J in and near the myocardium are given and the defibrillation waveforms for each of the electrode pairs is generated.

(58) In general principle, the process of EIT involves injecting a current by an electrode, and the induced voltage is measured at multiple points on the body surface. In the preferred implementation, what is termed the “multireference method” is used for configuring the current voltage pairs. (Hua P, Webster J G, Tompkins W J 1987 Effect of the Measurement Method on Noise Handling and Image Quality of EIT Imaging, Proc. Annu. Int. Conf. IEEE Engineering in Medicine and Biology Society 9 1429-1430.) In the multireference method, one electrode is used as the reference electrode while the remaining electrodes are current sources with the induced voltages being measured on each electrode simultaneously while the current is being delivered. The amplitude of the current sources are individually varied and each electrode is treated as a reference lead in succession. Finite element methods are then used to convert the calculus problem (∇.Math.ρ.sup.−1∇V=0) into a linear algebra problem of the form YV=C, where Y, V, and C are the conductance, voltage, and current matrices respectively. Y, V, and C are also sometimes known as the master matrix, node voltage vector, and node current vector respectively. Mesh generation is performed on the two or three-dimensional physical model with triangular or quadrilateral elements for two dimensional problems and hexahedral shapes for three-dimensional problems. Boundary conditions are then set such as at the reference node or driving electrodes for Dirichlet (known surface voltages) or Neuman (known surface currents) boundary conditions. A number of methods have been used to compute the master matrix such as Gaussian elimination or Cholesky factorization.

(59) The Newton-Raphson algorithm may also be used for reconstruction of the resistivity distribution. The algorithm is an iterative algorithm particularly well suited to non-linear problems. The Newton-Raphson method minimizes an error termed the “objective function”. Here, it is defined as the equally weighted mean square difference between the measured and estimated voltage responses:
Φ(ρ)=(½)(V.sub.e(ρ)−V.sub.0).sup.T(V.sub.e(ρ)−V.sub.0).

(60) Using methods known to those skilled in the art, an algorithm is utilized whereby a distribution is first estimated, then the theoretical voltage response to a given current input is calculated using the finite element method. The estimated voltages are subtracted from the measured voltages to obtain the objective function. If the objective function is less than an error threshold, the estimated distribution is deemed to be an acceptable estimation. If not, the following equation is used to update the resistivity distribution:
Δρ.sup.k=−[V.sub.e(ρ.sup.k).sup.TV.sub.e′(ρ.sup.k)].sup.−1{V.sub.e′(ρ.sup.k).sup.T[V.sub.e′(ρ.sup.k)−V.sub.0]}

(61) This sequence is repeated until an acceptable estimation is achieved.

(62) In a preferred implementation, a table lookup method is provided to determine the estimated voltage matrix V.sub.e(ρ). The table values are based on average patient resistivity distributions and assuming correct placement of the electrode. Better accuracy can be achieved by providing anatomical markings 126 on the electrode pad as shown in FIG. 21.

(63) Accuracy may also be improved by providing a secondary imaging method such as ultrasound to take advantage of its higher imaging resolution to calculate the positions of the internal organs relative to the electrodes. If a secondary imaging method such as ultrasound is used to determine the positions of internal tissues, EIT can be used to determine the resistivities of each tissue type.

(64) In other implementations, an average resistivity value is determined for the tissue regions as defined by the secondary imaging method. This is accomplished by first defining a tissue region such as the lungs or myocardium by standard image processing methods. Next, the calculated resistivity distribution is overlayed onto the secondary image. All nodes of the resistivity distribution that are contained within a particular tissue region are combined together into a single resistivity measure for that tissue region. The method of combination may be an averaging, median, or other statistical or image processing method.

(65) The optimal V.sub.0 and J.sub.0 are determined using a modified inverse problem where the desired V and J in and near the myocardium are given and the defibrillation waveforms for each of the electrode pairs is generated.

(66) Improved current delivery (and impedance measurements) can be achieved by close-packing a large number of electrodes. Many arrangements of electrodes are possible. In a preferred implementation, the configuration of electrodes is determined with the assistance of the theory of tessellation. A regular tiling of polygons (in two dimensions), polyhedra (three dimensions), or polytopes (n dimensions) is called a tessellation. Tessellations can be specified using a Schläfli symbol. The breaking up of self-intersecting polygons into simple polygons is also called tessellation, or more properly, polygon tessellation. There are exactly three regular tessellations composed of regular polyhedra symmetrically tiling the plane, as shown in FIG. 10. Tessellations of the plane by two or more convex regular polygons such that the same polygons in the same order surround each polygon vertex are called semi-regular tessellations, or sometimes Archimedean tessellations. In the plane, there are eight such tessellations, shown in FIG. 11. There are fourteen demi-regular (or polymorph) tessellations, which are orderly compositions of the three regular and eight semi-regular tessellations. These polyhedra are shown in FIG. 12. Other demi-regular tessellations are Penrose Tilings. In three dimensions, a polyhedron that is capable of tessellating space is called a space-filling polyhedron. Examples include the cube, rhombic dodecahedron, and truncated octahedron. There is also a 16-sided space-filler and a convex polyhedron known as the Schmitt-Conway polyhedron, which fills space only aperiodically. Space-filling polyhedron can be utilized to better fit the electrodes to the three-dimensionality of the human thorax. In the preferred implementation, the electrode tessellation pattern is a cubic or hexagonal regular tessellation.

(67) In another implementation, using the previously-described EIT methods, it is possible to deliver arbitrarily complex spatial and temporal distributions of current to the heart limited only by the number and size of electrodes on the thorax. During fibrillation, either pre or post-shock, the direction of the activation wavefront can vary and is not predictable, as is the case with a normal sinus rhythm. In FIG. 24B, the post-shock wavefront travels in a counter-clockwise, while in FIG. 25B, the rotation is clockwise. The arrows in the figure represent the direction of movement of the wavefronts; the shading of a region represents the phase of the cardiac cycle in the region (e.g., white representing 0 msec of the cycle, and darkest gray representing 110 msec). The wave may approach a region of the heart where only a small percentage of the cells are able to depolarize, preventing effective muscle contraction, but allowing the wave to continue. Indeed, the wave does continue, and by the time it travels completely around the heart, more cells have repolarized, and the wave can continue. The wave will, in fact, continue, and since no blood is being pumped, it is usually a fatal condition unless there is an intervention with a defibrillator. The heart can support multiple such incoherent waves traveling in different directions, so we will apply the term wavelets to these various waves. As the wavelets travel around the heart, the rate at which they complete a ‘circuit’ is called the VF cycle rate. As the wavelet passes a particular point and depolarizes the cells, the cells will recover, or repolarize before the wavelet returns.

(68) It is desirable to be able to deliver electrical current to specific regions of the myocardium so as to either reduce the extent of the excitable gap region 92 (FIG. 25D) or to terminate pre or post-shock wavelets. The extent of the excitable gap region 92 can be reduced or abolished by delivering current sufficient to depolarize the region of the myocardium that would have been in the excitable gap prior to the defibrillation shock. Using the methods of EIT just described, the resistivity distribution is calculated. Then, using a biodomain model, e.g., as described in IEEE Trans Biomed Eng 46: 260-270, 1999 (Entcheva), a predicted response of the myocardium to an electrical stimulation by the device can be calculated. The bidomain model may be a system of two reaction-diffusion equations, one for the intracellular and the other for the extracellular space linked by transmembrane current. The model simplifies the complex, non-linear behavior of the sarcolemmal ion channels, assuming a passive membrane behavior modeled as a parallel combination of a resistor and a capacitor in order to minimize computational complexity. The model is a system of two differential equations with a constant resistance membrane, Rm:
Ñ×(giÑΦi)=β(Vm/Rm),
Ñ×(geÑΦe)=β(Vm/Rm),
where Vm and Rm are the transmembrane voltage and membrane resistance, respectively, and gi and ge are the intracellular and extracellular conductivity tensors modeling the fiber architecture of the myocardium and β is the cell surface to volume ratio. The excitable gap region 92 is located based on the bidomain calculations, and is shown in FIG. 25D as a hatched region.

(69) Using the previously described EIT methods, the heart is stimulated by one or more pulses in succession before the defibrillation shock by current that is focused in the region of the myocardium that the bidomain model predicted as occupying the excitable gap (hatched region in FIG. 25D). Delivering the pre-shock current to those regions not adequately affected by the defibrillation shock will result in less total energy being required for the combined energies of the pre-defibrillation stimulation and defibrillation shock. A tessellated electrode pattern may be used, e.g., incorporating 16 electrodes for the anterior and posterior electrode pad assemblies.

(70) In another implementation, an electrode configuration is provided whereby at least some portion of the stimulating electrodes are also connected to filtering and amplification of individual electrocardiographic (ECG) monitoring channels. In a preferred implementation, at least 16 ECG channels are available in the device with the input multiplexed between one electrode on the anterior pad and one electrode on the posterior pad. A sample rate for ECG analysis is preferably 250 Hz, therefore the A/D is sampled at 500 Hz with alternating samples from the anterior and posterior electrode. Employing EIT and solving the forward problem as discussed previously, the distribution of activation wavefronts on the epicardium can be calculated. The path of the activation wavefront can also can calculated, for instance in FIGS. 24B and 25B where the arrows show the path of the activation wavefronts. Using the previously described EIT methods, the heart is stimulated by one or more pulses in succession following the defibrillation shock by current that is focused in the region of the myocardium that lies in the path of the activation wavefront (e.g., the hatched region in FIG. 25E). Depolarizing tissue in front of the activation wavefront will quench the arrhythmogenic activation wavefront when it reaches the refractory, recently stimulated region.

(71) Referring to FIG. 1, variable resistors 71, 72 are replaced by high voltage transistor switches and provided for each of the 16 electrodes. Eight of the 16 anterior electrodes are connected to one H-bridge 10, and the remaining eight are connected to the other H-bridge 11. Alternatively, all 16 may be connected to one H-bridge. By switching the transistor switches 71, 72 at a rate significantly higher than the H-bridge, preferably >50 times faster, the switching of each high voltage transistor 71, 72 can be pulse-width modulated to produce an average current for each electrode. As long as the switching duration is less than approximately 200 microseconds, the myocardium will respond to the average of the pulse-width modulated (PWM) currents. As a result, it is possible also to switch the H-bridge with a period of 120 microseconds, with positive and negative phases of the bridge with a duration of 55 microseconds and a 10 microsecond interphase delay. Each individual high voltage transistor is pulse-width modulated with pulse durations of 5-10 microseconds during both the positive and negative phases of the bridge so as to be able to produce a potentially bipolar voltage at any of the electrodes necessary for a fully functional EIT system.

(72) One possible theory to explain the improvement that some implementations of the invention may achieve in defibrillation efficacy (understanding, of course, that the invention is not limited to this theory) is as follows: As stated previously, the theory of Virtual Electrode Polarization (VEP) describes the phenomena by which, because of current flow within a partially conductive medium (the myocardium) contained within another partially conductive medium (blood of the cardiac chambers, lungs, interstitial fluids and other organs within the thoracic cavity), myocardial polarization during defibrillation is characterized by the simultaneous presence of positive and negative areas of polarization adjacent to each other. “Phase Singularity” as defined within the context of VEP is a critical point that is surrounded by positively polarized (equivalent to “depolarized” in the conventional electrophysiology nomenclature), non-polarized and negatively polarized (equivalent to “hyperpolarized”) areas. These phase singularities are the source of re-initiation of fibrillation. Post shock excitations initiate in the non-polarized regions between the positively and negatively polarized areas through a process termed “break excitation.” The break excitations propagate through the shock-induced non-polarized regions termed “excitable gaps”, and if the positively polarized regions have recovered excitability, then a re-entrant circuit at which fibrillation may initiate is formed. With biphasic defibrillation, the second phase of the shock nullifies the VEP effect by depolarizing the negatively polarized tissue. Since less energy is needed to depolarize repolarized tissue than further depolarize already depolarized tissue, effective biphasic defibrillation achieves nearly complete depolarization of the myocardium by reversing the negative polarization while maintaining the positive polarization. There remain, however, excitable gaps even with biphasic and multiphasic waveforms, albeit reduced in scope relative to monophasic waveforms, and there still remains the potential for significant improvement of the efficacy of biphasic defibrillation waveforms.

(73) FIG. 13 shows the results of a simulation in a study by Efimov (Am J Physiol Heart Circ Physiol 2000; 279:H1055-70). The lighter grey region 90 is a region of positive polarization and the black region 91 is one of negative polarization. The white region 92 is the excitable gap region. FIGS. 14a and 14b depict, in schematic view, what occurs when the area of the electrodes is varied. As can be seen, by increasing the size of the electrodes, the contact angle, φ 93, of the electric field lines is increased in the region of the excitable gap, thereby reducing the areal extent of the excitable gap. Reduction of the areal extent of the excitable gap, improves the chances for a successful defibrillation and reduces defibrillation thresholds.

(74) In other implementations, the waveforms may each be composed of a sequence of pulses. The relative timing of the current vectors may be designed so that the pulse sequences are interposed with non-overlapping individual pulses.

(75) Referring to FIGS. 26 and 27, in another implementation the electrode pairs may be configured so that, at the heart, the electrical current vector 205 for one pair is substantially aligned with (e.g., parallel to the long axis of) the myocardial fiber orientation of the epicardium 200 on the anterior portion 202 of the heart, while the current vector 206 for the other pair of electrodes is substantially aligned with the fiber orientation of the epicardium 201 on the posterior portion 203 of the heart. In a three-dimensional conductive medium, the current distribution is also three-dimensional, with each point within the volume best described as a 3-space vectorial value specifying the current density. Thus, the term “current vector” is a simplification of this more general concept, and describes a volume-averaging of the current density that has been normalized to provide a vector that describes the average direction of current flow within a volume. Myocardium has been shown to have anisotropic conductivity, with the long axis along the fiber orientation having the higher conductivity. This method thus enhances current flow within the myocardium (as it lowers the impedance along paths through the myocardium, causing more current to flow through the myocardium). This method is also applicable to implementations containing more than two pairs of electrodes. For instance, four pairs of electrodes may be used with the electrical current vector for each aligned with the fiber orientation of a portion of the heart. In one implementation, the four portions of the heart are the left anterior, right anterior, right posterior, and left posterior.

(76) Referring to FIGS. 28A and 28B, the method described in the previous paragraph may also be applied to provide a “leadless” implanted defibrillator system wherein all the stimulating electrodes and the device housing of the implanted cardiac stimulation device is external to the thoracic cage but implanted subcutaneously. Commonly termed implanted cardioverter-defibrillators (ICDs), the devices have required for at least one of the stimulating electrodes to be placed on a catheter placed within the heart, typically in the right ventricle and superior vena cava in order to reduce the energy required to defibrillate, thus minimizing the size of the implanted device.

(77) Referring to FIG. 28A, two separate defibrillator housings 207, 208 are provided along with stimulating electrodes 209, 210, 212. The corresponding stimulating electrodes 209, 210 for the right atrial housing 207, and the corresponding stimulating electrode 212 for left atrial housing 208, result in current vectors 205, 206 substantially aligned with the fiber orientation of the myocardinum in the posterior and anterior portions of the heart, respectively. The right atrial stimulating electrodes 209, 210 (which electrically serve as one electrode) are positioned over the intercostal spaces to reduce defibrillation impedance and are mounted onto the adjacent ribs with mounting hardware of a suitable material such as stainless steel alloys (e.g., 316L), cobalt-alloy F90, or titanium, or sutured into place. Using two pairs of stimulating electrodes provides a more uniform current distribution across a larger region of the heart. Because of the human heart's angular orientation to the rib cage, the intercostal spaces are oriented perpendicular to the perimeter of the heart in projection. Thus, aligning a single stimulating electrode with the intercostal space may reduce the impedance and increase therapeutic current levels, but as a result myocardial current distribution becomes more non-uniform. This current density intensification is eliminated by placing at least two commonly connected stimulating electrodes 209, 210 in different intercostal spaces. The particular intercostal spaces may be adjacent or may have intervening, non-stimulating intercostal spaces.

(78) Referring to FIG. 29, the housings 207, 208, which are configured as “active” cans with all or some portion of the housing providing stimulation to the stimulating electrodes, are connected to conductors contained in a common sheath 211 composed of a biocompatible material such as a silicone-polyurethane copolymer. The stimulating electrode 212 corresponding to the left atrial housing 208 is located just below the xyphoid process. Preferably, this electrode is oriented parallel to the perimeter of the projection of the heart onto the rib cage to provide better current distribution. Microprocessors 213, 214 communicate via conductors 215, 216 to provide status information and synchronization of defibrillation pulses during defibrillation. Preferably, one of the microprocessors functions as a primary processor. The primary processor is responsible for functions such as self tests and communication with devices outside the body that are better performed in a centralized rather than distributed manner, ECG signal processing may also be localized to the primary processor, and the secondary processor may perform the single function of providing a defibrillation pulse at a time deemed appropriate by the primary processor, which communicates that time via conductors 215, 216. Preferably, one wire 215 is a single-wire serial interface providing a variety of status and data between the processors, while the second wire 216 is a hardware-level interrupt to provide microsecond-level control of the timing of the secondary defibrillation pulse. In a related implementation shown in FIG. 28B, conductors for electrodes 209, 210 that are part of the defibrillation circuit contained in housing 207 pass through left atrial housing 208, and the conductor 212 passes through housing 207, in order to provide an alternative configuration with a minimal number of wires within the patient.

(79) In another implementation, an imaging method such as ultrasound or magneto-resonant imaging (MRI) may be used to determined the exact location and angular orientation of the heart within the thoracic cavity prior to implantation of the device so as to obtain improved positioning to produce current vectors 205, 206 in better alignment with the fiber orientation of the epicardium. The defibrillator housings 207, 208 and stimulating electrodes 209, 210, 212 are positioned so as to provide closer alignment of the expected current vectors 205, 206 to the fiber orientations 200, 201. In the implementation where MRI is used, external transthoracic pacing electrodes may be applied to the patient in the positions such as is shown in FIG. 8a or 8b prior to the process of obtaining the MRI. A test pacing pulse or pulses may then be delivered to the patient that is synchronized with the MRI excitation pulses so as to produce a three dimensional representation of the current vector space that will show the myocardial fiber orientation for use in determining the positioning of the housings 207, 208 and stimulating electrodes 209,210, 212.

(80) In another implementation, resistance circuits 55, 56 are eliminated and the waveform shape, and thus also the first phase average current, is adjusted by pulse width modulating the switches in the H-bridges 10, 11. This configuration is the Class D amplifier configuration, known to those skilled in the art of amplifier design. In its simplest form, a switch-mode amplifier consists of an H-bridge and a load as shown in FIG. 15. Amplifiers are typically classified by their output stages. Of the common output-stage topologies (Classes A, B, AB, and D), Class D amplifiers exhibit the highest efficiency. A linear output stage (Class A, B, or AB) draws considerable bias current while sourcing and sinking current into a speaker making them not particularly well suited to high voltage designs. A nonlinear (Class D) output stage eliminates this bias current. In the preferred implementation, as shown in FIG. 16, the Class D amplifier consists of an input preamplifier 95 for isolating, filtering and level shifting the control voltage from the processing means 5, a sawtooth oscillator 96, a comparator 97, two MOSFET drivers 98, 99, and the H-bridge switches 100-103. The comparator samples the input signal, with the oscillator frequency determining the duration of the sampling period. Thus, the oscillator frequency is an important factor in the overall performance of a Class D amplifier. As shown in FIG. 17, the comparator output 104 is a pulse-width modulated (PWM) square wave that drives the H-bridge. The PWM square wave 104 is created by a comparator whose inputs are the sawtooth (V.sub.RAMP) 105 and the control signal (V.sub.IN) 106. The H-bridge then outputs the square wave differentially. For a given input level, the comparator output is a duty-cycle modulated square wave with period determined by the sawtooth frequency. The PWM square wave controls the H-bridge drivers 100-103, turning opposite pairs of MOSFETs off and on, thereby reversing current to the load within a single period. The output may be filtered by capacitor filters or inductor/capacitor filter combinations which remove high-frequency content from the H-bridge square wave output.

(81) Alternatively, the measurement of the thoracic cavity may be carried out using an ultrasound transducer capable of imaging the heart and surrounding tissue. An ultrasound transducer may be incorporated into an integrated defibrillation pad, as shown in FIG. 18. In a preferred implementation, an opening in the center of the electrode is provided that is covered over with an ultrasonic-conducting gel 107. The gel is a bilayer structure with a more aggressive adhesive provided on the face opposite to the patient for attaching the ultrasonic probe prior to use.

(82) In other implementations, there may be two or more separate defibrillators, as shown in FIG. 19. The first defibrillator 110 acts as the master defibrillator, while additional defibrillators 111 function as slave defibrillators whose energy is delivered synchronously with that of the master defibrillator 110. Synchronization is provided by communication means 114. Preferably, the communication means 114 is implemented as a simple switch. In a conventional defibrillator, the delivery of energy is initiated via the closure of a discharge switch 112 located on the front panel or on a set of defibrillation paddles. The closure of the switch initiates the defibrillation sequence under the control of processing means 5. Charging of the high voltage capacitors on both defibrillators 110, 111 is initiated via the charge-control user inputs 115. At the appropriate time, the clinician will press the discharge button 112. This causes the processing means 5 on the first defibrillator 110 to close a slave discharge switch that initiates the discharge sequence on the second defibrillator 111, at which time the first defibrillator 110 also initiates its discharge sequence. The wiring for the communication means 114 is preferably configured such that the wires are located within the same cable as the energy delivery wires, thus reducing any additional cabling. The communication means 114 may also incorporate digital communication methods which provide additional information about defibrillator status.

(83) The defibrillator pad 123 may integrate all connections into a single connector 120 as shown in FIG. 20. The defibrillator pads may be constructed such that a seam line 121 is located between the active areas 122 of the pad 123 where the seam line 121 is of higher compliance than the active areas such that the pad 123 can be folded during storage without creasing the active areas.

(84) In another implementation, a physiological parameter, e.g., the electrocardiograph (ECG), is measured in conjunction with the EIT image, and an estimate is made by the device of the chances for a successful defibrillation shock based analysis of ECG data. Depending on the estimate of shock success, decisions as to the proper treatment to provide the patient are made in a coordinated resuscitation effort that includes both defibrillation and chest compressions, which can be provided manually in response to prompts, or in a semi-automated or fully automated fashion. The block diagram and flow chart for such a system is shown in FIGS. 22 and 23.

(85) One or more additional electrodes 125 may be provided for diaphragmatic stimulation (DS) and may be incorporated into the anterior electrode such that the DS Electrode (DSE) is located over the patient's diaphragm as shown in FIG. 21. Diaphragmatic stimulation induces air exchange in the lungs during cardiopulmonary resuscitation (CPR) for improved oxygenation. The return path for the stimulation current from the DSE is through one of the pre-existing electrodes. Utilizing EIT or other imaging methods, the current distribution may be adjusted to achieve optimal stimulation, as described previously in this patent. The DSE may be integrated with defibrillation and cardiac pacing to provide a coordinated resuscitation effort in an automated or semi-automated fashion. The integrated resuscitation may also incorporate a means of providing chest compressions, such as a piston-based system manufactured by Michigan Instruments (Michigan) or a constricting band system manufactured by Revivant Corp. (California). FIG. 23 shows a decision flow chart of one possible integrated resuscitation protocol.

(86) Many other implementations of the invention other than those described above are within the invention, which is defined by the following claims. The invention applies to both defibrillation and cardioversion; in the claims, references to defibrillation should be interpreted as also encompassing cardioversion. Some implementations of the invention are broader than defibrillation and cardioversion.