Cardiac Pacer and Monitor System

The foregoing brief description of the present invention as well as the objects and advantages thereof, will become more apparent from the following more detailed description and additional advantages and objects inherent therein, and wherein:

FIG. 1 is a general block diagram of a system employing the present invention;

FIG. 2 is a timing diagram illustrating the relationship between the various signals employed within the system of FIG. 1;

FIG. 3 is a block diagram of the monitoring logic employed in an external version of the system of FIG. 1;

FIG. 3A is a schematic of the display reset circuit;

FIG. 3B is a block diagram of the monitoring logic employed in an internal version of the device of FIG. 1;

FIG. 4 is a further set of monitoring logic employed for detecting sensing failures;

FIG. 5 is a block diagram indicating the generation of timing signals employed in conjunction with the embodiments illustrated in FIGS. 1-3;

FIG. 6 shows an alternate method of detecting a sensing failure;

FIG. 7 is a logic circuit illustrating the implementation of the method of FIG. 6.

FIG. 8 is a timing diagram illustrating the relationship of the components described in FIG. 7;

FIG. 9 is a logic diagram illustrating the automatic detection of a failure;

FIG. 10 illustrates an alarm monitor for a sensing or capture failure detected by the logic of FIG. 9;

FIG. 11 is a timing diagram illustrating the relationship of the circuitry of FIG. 9.

Referring to FIG. 1, there is illustrated a cardiac pacer of the demand variety, employing a pulse generator and timer 10. In implanted units, the pulse generation rate is preset and fixed in accordance with a desired ("intrinsic") rate. A variable rate control may also be employed in external (temporary) pacers. The output of the generator is coupled to a constant current driver 12 which provides a current pulse of about 1 millisecond duration, between its output lines 14 and 16. The current pulse is used to initiate a contraction (i.e., stimulate) the ventricles of the heart. The possibly dangerous D.C. component of this pulse is isolated from the heart by means of the capacitor 18. The current path defined by the lines 14 and 16 is terminated in an electrode whose positive and negative terminals are labeled 20 and 22, respectively. The electrode is placed in contact with the heart tissue in a conventional manner. Current flow between the two terminals through adjacent cardiac tissue constitutes the stimulus which initiates a general ventricular contraction. The negative terminal is usually the site at which the contraction begins. Stimulating current then returns to the positive terminal. The electrode is "bipolar" if the terminals 20 and 22 are part of the same structure. If terminal 20 is the external shell of the pacer itself, the electrode is "unipolar" and the electrode structure has only terminal 22.

In the demand cardiac pacer indicated, the occurrence of spontaneous (nonstimulated) cardiac activity will suppress the release of any pacer output stimuli providing that the rate of occurrence of such activity exceeds the intrinsic rate. Spontaneous activity may be supraventricular in origin, in which case the heart's natural (atrial) pacemaker initiates a train of events resulting in ventricular contractions and consequent electrical activity known as the "QRS complex" (an R-wave). Irritable cardiac tissue in the ventricle may also "twitch" spontaneously and evoke electrical currents (also producing an R-wave). In either case, the spontaneous activity inhibits stimulus release for a time period T, (which is the reciprocal of the intrinsic rate), following the spontaneous activity. The inhibition occurs by resetting ("recycling") the pulse generator-timer 10. As shown in FIG. 1, the initial stimulus signal from the driver 12 is coupled to excitable cardiac tissue via isolation capacitor 18 (FIG. 2, line A). The response evoked by the stimulus (FIG. 2, line B) appears across electrode terminals 20 and 22 and is coupled by isolation capacitor 24 to be amplified by signal amplifier 28A. Ambient 50 or 60 cps electromagnetic noise may also be picked up by amplifier 28A. Such noise is filtered out by filter 28B. The filtered output is further amplified and shaped by R-wave amplifier 30. Elements 28A, 28B, 30 and 32 constitute a functional amplifying system 26. Element 32 is an intrinsic delay component. It will be understood that the delay component 32 is actually the time factor required for the amplifier system 26 to recover after being stimulated by the initial stimulus pulse. Elements 30, 32 and 36 together form the component R-wave amplification 27.

The logical effect of such a saturation phenomenon is to inhibit gate 36, for the duration of the saturation period. Thus, in a normal demand pacer, both the response evoked by the stimulus (FIG. 2, line A) as well as any spontaneous activity which may occur, will not be detected by the block 27 during the saturation or recovery period. Following the termination of this period, any spontaneous cardiac activity will be detected, amplified, appear at the input of gate 36 and at the output of amplifiers 38 and 30. If such activity occurs, the timing element 10 is "recycled" by output R (i.e., reset) through optional switch 40, and the subsequent stimulus (FIG. 2, line Q) will be released only at the end of an interval (T) following the instant at which recycling occurs. Thus, if the heart begins to beat naturally with frequency of at least 1/T, the generator 10 will be continuously reset and the pacer remains inactive. Amplifier 38 has an additional inverted output R which is used in the detection of sensing failures, as described below.

The pacer configuration also permits fixed pacing merely by resetting the switch 40 to the FIXED RATE position, thereby opening the loop and interrupting the feedback. In this mode, the pacer will operate at a fixed rate, continuously, until disabled.

Switch 40 is used only in external pacers. In demand pacers, the recycling output of amplifier 38 is directly connected to element 10 (and the switch is replaced by a magnetically actuated reed).

In a system such as described above, any failure of the pacer to properly stimulate the heart ("loss of capture") or failure to sense spontaneous activity and properly recycle the pacer timing circuits ("sensing failure") is undetected and thus will not prevent continued malfunction. If sensing failure occurs, the pacer is never recycled and it thus continues to stimulate at its fixed, intrinsic rate. This condition is potentially dangerous and may initiate ventricular fibrillation in a recently infarcted heart. If loss of capture occurs, the stimuli are ineffective. Then, if there is no underlying or independent cardiac rhythm because of various conditions generically known as "heart blocks", the heart beats at its "idioventricular" rate, which may be too low a rate to sustain adequate blood circulation. Pacers, heretofore, have not been designed to be "aware" of the occurrence of such potentially dangerous events, take corrective action if they do indeed occur, and simultaneously warn an attending doctor of the occurrence of these events.

The essential requirement for detection of "loss of capture" or "sensing failure" is the ability to detect the response of the heart (FIG. 2, line B) to the pacer output stimulus (FIG. 2, line A). This response is normally not seen because it occurs within 15 to 50 milliseconds after the stimulus, i.e., within the saturation or recovery period of the conventional amplifying system 26 and 27. Therefore, to make this response "visible", an additional compensated, high fidelity, A.C. coupled differential amplifier 42 is capacitively coupled to electrode terminals 20 and 22. To overcome the saturation of this amplifier by the relatively large stimulation pulse, fast recovery switches 44 are coupled to the amplifier 42 and to the constant current driver 12. The fast recovery switches act to eliminate, within 8 milliseconds after the stimulus, any electrical charges which remain on various capacitive components within the various amplifier units and current drivers.

Referring to FIG. 2, lines A to T, a timing diagram for use in conjunction with FIG. 1, the stimulus signals P.sub.ST and ST (FIG. 2, line A) occur at periodic intervals, assuming the absence of spontaneous cardiac activity. The response of the heart, FIG. 2, line B normally comprises of a depolarization potential D and a repolarization potential R, as distinct from the spontaneous "R-wave", occurs thereafter. The saturation of the amplifier unit 26, as indicated by the signal shown on FIG. 2, line C, at the output of delay 32, holds the pacer refractory for the blocking interval. If the pacer is in its DEMAND mode, the presence of a spontaneous R-wave, FIG. 2, line D, providing it occurs after the end of the pacer rafractory period but before the end of the stimulus release interval T (measured from the prior stimulus) will reset the generator 10. The next stimulus pulse ST, FIG. 2, line A, (A.sub.1), thus cannot occur until the end of the predetermined time period T. It is apparent that if the rate of occurrence of spontaneous activity continues to exceed the intrinsic pacer rate (1/T), no stimuli will be released, i.e., pacer activity is suppressed and the pacer remains in its "standby" condition.

On the assumption that pacer stimuli are not thus suppressed, the stimulus pulse ST will be applied to the timing circuit of FIG. 5 (explained in further detail below) to generate a series of time signals. The ST signal is reshaped by amplifier 208 to generate the P.sub.ST signal. The first of these timing signals, P.sub.R, FIG. 2, line F, will be applied after a delay determined only by the width of the P.sub.ST pulse, to the fast recovery switch 44, thereby unblocking the amplifier system 42 and the driver 12. The unblocking is effective within the duration of the P.sub.R pulse (8 milliseconds). The differential amplifier 42 is thus able to pass (or detect) the response, FIG. 2, line B or FIG. 2, line E, providing that it occurs 8 milliseconds or more after ST has terminated.

Two sampling tests are shown in FIG. 1. These tests are made to indicate the occurrence of an adequate response by the heart to the pacer stimulus ST. The two components of this response (D and R, FIG. 2, line B) are tested individually. The SP.sub.A pulse serves to test the depolarizing response D, the second SP.sub.B pulse is used to test the repolarizing response R. The signals SP.sub.A and SP.sub.B are supplied from the circuit of FIG. 5 and occur at times set by delay elements 200 and 202, respectively. These times are determined by observation of the total response (FIG. 2, line B) at the output of amplifier 42 and are set so that SP.sub.A occurs at the peak of the D response and SP.sub.B occurs at the peak of the R response. The depolarizing test is effected through a first gate 46, while the repolarizing test is effected through a second gate 48. The signal levels are appropriately signed such that the presence of the response signals during the sampling periods will produce an appropriate gated output signal to the amplifier 52. In practice, the gates are actually individual dual input FET switches whose outputs are directly tied together, (constituting a logical buffer) and connected to the input of amplifier 52.

It is noted that although testing of both depolarization and repolarization signals is shown, it is often simpler and more convenient to perform the sampling test solely on the depolarizing signal, since the depolarization response is sometimes larger and more stable. It is possible, and in some cases, necessary, however, to test the repolarization signal alone, or to use both tests for greater reliability.

The outputs of the gates 46 and 48 are fed to an amplifier 52. The gain of the amplifier 52 is set to provide a safety factor with respect to the minimum amplitudes to be permitted for the R and D responses. These minimum amplitudes are defined by the threshold controls (plus and minus) on comparator elements 56 and 58 which are discussed further below. The normal gain setting (safety factor) is such that it provides a signal amplitude to the amplitude comparators which is, preferably, between 2 and 3 times the selected threshold levels. The output of amplifier 52 is supplied to a comparator stage 54. The comparator stage includes threshold adjustments which are matched to the patients R and D cardiac responses to pacer stimuli ST. These responses may vary for each individual, and it is thus preferable, although not essential, to appropriately set these threshold controls at the time of pacer installation. Thus, the gain of both amplifier 52 and the threshold levels of the comparator stage 54 act in conjunction to provide appropriate comparator outputs, FIG. 2, line L. When the signals at the output of amplifier 52 (FIG. 2, line K) do not reach threshold (points K.sub.1 and K.sub.2, FIG. 2, line K), no corresponding output will emerge from the comparator stage 54. As shown, the comparator stage 54 includes a positive (+) D comparator 56 and a negative (-) R comparator 58 which are used as previously discussed. Comparator circuit 54 is designed such that any voltage more negative or more positive than the corresponding threshold voltage selected, gives rise to constant, unipolar (positive) pulse outputs whose duration is that of either SP.sub.A or SP.sub.B. It thus functions to measure the absolute value (ABV) of the R and D responses.

The outputs of the comparator stage 54 are amplified in amplifier 60 to provide the absolute value signals ABV for subsequent processing. The absence of an ABV signal (cf: FIG. 2, line K or L) will indicate a failure of the comparator stage 54 input signal to meet either of its preset R and D threshold levels, an event defined as "loss of capture".

In accordance with the present invention, an additional condition, defining "loss of capture" with even greater certainty, is to determine that both the R and D responses occur within a maximum interval with respect to each other, i.e., that the total evoked response is biphasic. In practice, the occurrence and detection of the R response is physiologically determined by, and hence must follow, the D response, since active cardiac tissue also recovers following excitation and contraction and such recovery is accompanied by generation of an oppositely phased repolarization (R) potential. In rare cases, however, under conditions where a heart is irritable, cardiac potentials may spontaneously arise at some remote focus and propagate to the stimulating-recording electrode. To reduce the possibility of such potentials being interpreted as true responses, a test for biphasicity is readily accomplished by elements 60A and 60B. Element 60A is a monostable which is turned on for 0.15 seconds by the output of comparator stage 54. The output of the D comparator 56 necessarily appears first in time and sets 60A. The output of the R comparator 58 will normally appear about 100 milliseconds after the output of 56. It will have no effect on 60A which is already set, but will pass through gate 60B. Element 60A then resets about 50 milliseconds later. An output from 60B thus coincides with the occurrence of the R response at the selected sampling time, within the total period of 60A set time, and implies the earlier occurrence of the D response. Switch 60C selects this output when placed in position Y. In position X, 60C selects either the R or the D response. Since the R response is invariably a consequence of an active D response, the switch 60C in X position will pass both R and D responses. In practice, X position of 60C sampling of the R response is sufficient to detect loss of capture. However, confusion may arise in the event of the occurrence of spontaneous potentials with the same polarity as the R response and which also happen to occur at the selected sampling time. Thus, in accordance with the invention, the switch 60C is placed into the position Y and the circuitry of 60A, 60B and 60C minimizes this possibility by imposing the additional requirement that the D response must also first be recorded and sampled.

The pacer of the present invention is designed to sense a plurality of loss of capture events before instituting corrective and warning action. To this end, a loss of capture detection monitor, illustrated generally in FIGS. 3 and 3B, includes therein an N stage counter 62 coupled to a 2.sup.N line decoder 64, which acts to decode the output of the N stages counter 62 to a one out of Y signal, where Y=2.sup.N. For example, if a three stage binary counter (N=3) is employed, then the decoder 64 will have Y=2.sup.N or 8 output lines, as shown. The counter and decoder are conventional components.

A control flip-flop 66 is coupled to a specific output of the decoder 64, in accordance with the desired and preset number of loss of capture events which will ultimately provide corrective and warning signals. The flip-flop 66 is of conventional design, and in the illustration given (FIGS. 3 and 3B), has its set input coupled to the fourth line of the decoder through a buffer (OR) gate 68, by a selector switch 70. In the implantable model of this FIG. 3B, switch 70 may be replaced by a permanent connection to a selected value chosen by the surgeon. An external pacer of this design, FIG. 3A, will retain the selector switch. It should be noted that the first line of the decoder is numbered "0" but it represents a valid counter condition (001 in a three stage binary counter) and a 1 event condition. Thus, the fourth line is numbered "3" representing three events, with the fourth event giving rise to an output to flip-flop 66. Greater or lesser event conditions can be set by adjustment of the switch 70, in the external unit, or by presetting a permanent connection in the internal unit.

The operation of the detection monitor of FIGS. 3 and 3B will now be set forth. The stimulus pulse P.sub.ST, derived from the signal ST, will activate amplifier 72 (FIG. 2, line A) and thus enter the counter 62 through the coincidence (AND) gate 74, and accumulate one count, (FIG. 2, line M.sub.1). The gate 74 will have had its other input preconditioned by virtue of the normal state of flip-flop 66 providing a "high" output is present on output line FF.sub.CR. Immediately thereafter, should the comparator stage 54 pass a signal exceeding the selected threshold values an ABV signal will be generated. The ABV signal will enter the buffer (OR) gate 76 and reset the counter. Should the ABV signal not occur, the counter 62 will not be reset. Thus, if the counter 62 remains set after the threshold tests have been made at the times determined by delay elements 200 and 202, such a set condition indicates the absence of an adequate R or D response (or both) to ST, i.e., "loss of capture". An identical test is made to test the heart's response to the next ST. This second ST pulse will cause counter 62 to accumulate a second count, and initiate the events which enable the next test. If this test indicates that either the R or D signals are above the selected threshold, an ABV signal is generated and counter 62 is reset to zero. Thus, the line "3" output of counter 62 is activated by a fourth indication of "loss of capture" providing that there have been three previous, consecutive failures. For ease of illustration, it is presumed that the switch position is set at position "1" and switch 60C is in position A. Thus, referring to FIGS. 2, 3 and 3B, assuming that both the R and D responses to the first stimulus exceed the selected threshold, the counter 62 will be reset to zero at the time of occurrence of the SP.sub.A pulse. If a second stimulus evokes a subthreshold R response but an adequate D response, the counter will be reset to zero at the time of occurrence of the SP.sub.B pulse. In neither of these situations has there been "loss of capture". As illustrated, the third stimulus fails to evoke either an adequate R or D response. The same sequence of events would have occurred with 60C in position Y. Thus an ABV signal is absent, and the first counter stage is set (FIG. 2, line M). Assuming now that the next stimulus again fails to capture, the ABV signal is again absent. Thus the counter accumulates the count of 2, with the first stage being reset by the second P.sub.ST, causing the second stage to set (FIG. 2, lines M.sub.1 and M.sub.2). An output pulse thus appears on line "1" coincident with the next P.sub.ST, (FIG. 2, line N). This pulse will be applied through switch 70, via buffer 68, to the flip-flop 66, causing the output FF.sub.CS to go high (FIG. 2, line O) and FF.sub.CR to go low. Gate 74 is thus blocked from further count pulses and the counter 62 stage 2 holds this condition until reset, as described further below.

As a consequence of the setting of flip-flop 66, several actions transpire. As shown in FIG. 2, line I, an alarm signal ST.sub.2C generated by elements 204 and 205 of the timing circuit of FIG. 5 is provided. ST.sub.2C occurs at a time set by 204 and has a fixed duration (2 milliseconds). Referring to FIG. 1 and FIGS. 3 and 3B, the setting of flip-flop 66 output FF.sub.CS to a high state places a high signal on line 78, thereby energizing the ST pulse doubling circuit 80. The pulse doubling circuit 80 is coupled to the constant current driver 12 and increases the level of the stimulation signal ST, FIG. 2, line A, point A.sub.2. Although the term "doubler" is used, it is understood that an increase in magnitude of the stimulating signal ST is intended, and that a range of increased magnitudes may be chosen, of which doubling is the preferred. Increasing the magnitude of the stimulating pulse in this manner provides a high probability for regaining capture. At the same time, the high signal on line 78 provides a conditioning signal to the coincident (AND) gate 82, thereby permitting the pulsing circuit 84 to inject an additional alarm signal ST.sub.2C into the driver 12. As a result, the pacer now provides a supplemental signal displaced from the principal stimulation signal ST. This signal (FIG. 2, line S, FIG. 2, line A, point A.sub.3) may be seen by a clinician on a standard electrocardiographic recording, and will continue to be present until the counter 62 and flip-flop 66 are reset as discussed below. Thus, a permanent record is maintained of a loss of capture event, whereas such an event would have previously gone unreported in prior art devices. Thus, the unit is not only self-correcting, but diagnostic as well. The use of the increased magnitude signal would not normally be sufficient by itself as a diagnostic indicator because of the variations in recorded amplitude as a function of electrode position with respect to the heart. The supplemental alarm pulse, however, is a time distinguishing event, and clearly indicatable on EKG recording equipment. The pulse ST.sub.2C is set to occur in the "absolute refractory period" of the heart following the doubled stimulus. The usual absolute refractory period of a human heart is normally about 50-70 milliseconds, and any stimuli occurring in that interval following an adequate ST will be ineffective. ST.sub.2C, set to occur during this period, will therefore serve only as a warning or alarm pulse. Its width is set at 2 milliseconds so as to make it more readily visible on the EKG recording.

After EKG checking by the clinician, the counter 62 and flip-flop 66 are resettable, in the external unit version of FIG. 3, by external actuation by momentarily closing the manual reset switch 86. In an implanted pacer FIG. 3B, switch 86 is embodied as a magnetic reed switch which can be actuated by an RF generator which is brought near to the switch activating the RF detector 395 and driver element 396, causing the closure of switch 86. As a result in either case, FIG. 3 or 3B, a momentary signal, FIG. 2, line Q, is applied to coincident (AND) gate 88, the reset input to flip-flop 66, amplifier 90 and buffer (OR) gate 76, thus resetting the counter 62. In external pacers, the FF.sub.CR output is also applied to gate 301, shown in FIG. 3A, with a gated ABV pulse, and then to buffer gate 302 where it is gated with the differentiated FF.sub.CR signal to reset a visible display 305 of the state of counter 62. The flip-flop 66 also changes state, FIG. 2, line O, rendering FF.sub.CS low and FF.sub.CR high. Thus, the counter 62 is enabled for beginning its count sequence again, and the doubling and warning circuits 80 and 84, respectively, are disabled.

One other form of failure detection which is designed into the pacer of the present invention is the detection of sensing failure. This condition occurs when the pacer releases a stimulus at a time when the occurrence of spontaneous activity should have suppressed it. This type of failure is detected in the manner illustrated generally in FIG. 4. It is assumed that spontaneous activity occurs at some time between (a) the end of the refractory period of the heart following an adequate stimulus, and (b) the release of the next ST at the end of interval T (the reciprocal of the pacer's intrinsic rate). This spontaneous activity generates a signal R on line 34, R on line 34A, or R' on line 39A from amplifier 39, as well as R' on line 39B. R, as noted previously, is generated in spontaneous activity occurs during the non-refractory period of amplifier 26. R' is a conditioned output derived from the essentially non-refractory amplifier 42. Thus, the spontaneous activity which also appears at the input to 42 gives rise to a signal at the output of 42. This signal is applied to gate 37 together with an RINH signal derived from the 0.25 second monostable multivibrator 208A (FIG. 5). The effect of R INH is to block outputs from 42 for 0.25 seconds after the stimulus ST. Thus, R' is necessarily coincident with R, and is in effect a redundant R signal. However, since R' (and R') are derived from amplifier 42 which has better frequency response and higher fidelity than the amplifier 26 which produces R (and R), the R' output may be present in the absence of R, i.e., 42 will produce a faithful reproduction of a spontaneous potential seen at the stimulating-recording electrode while 26 will not, or will miss the signal altogether. Normally, since R is used to recycle the pulse generator and timer 10, this will result in a "failure to sense". Any lack of "agreement" between R and R' will be an indication of this failure. Any component failure in either 42 or 26 will produce the same effect. However, if one of these signals is absent due to component failure, the other is likely to be present, since the probability of coincident failure is low. R and R' are buffered together through buffer gate 41. The output of gate 41 is amplified by amplifier 43, whose output is used to set flip-flop 96, enabling gate 327. Gate 327 remains enabled for a period of 0.6 seconds, since the same spontaneous activity acts to reset flip-flop 96 via the 0.6 second delay element 94. If an ST pulse is generated within this 0.6 second interval, a sensing failure has occurred, since the normal pacemaker rate is normally not set to exceed 90 beats per minute, corresponding to an escape interval T=0.67 seconds. (An ST pulse would normally not have occurred prior to the end of this minimum escape interval following a spontaneous R wave.) Thus, if a sensing failure does occur and the ST pulse is incorrectly released, P.sub.ST passes through G327 and flip-flop 328 is set by signal A, and remains set until reset by the action of switch 86 (as previously described). The flip-flop 328 set output (FF.sub.SS, FIG. 2, line R) conditions the coincident (AND) gate 98 to allow passage of the timed SF marker pulse, FIG. 2, line J provided from the timing circuit of FIG. 5. The SF marker pulse is timed so as to occur sufficiently displaced from the stimulus signal ST and warning signal ST.sub.2C so as to be distinguishable therefrom, and shown at point A.sub.4 on FIG. 2, line A. The coincidence of the two signals at gate 98 passes a pulse to the sensing failure pulsing circuit 100 which energizes the constant current driver 12 to place the signal, FIG. 2, line T on the output terminals 20 and 22, in the manner described heretofore in connection with the warning signal ST.sub.2C. Thus, a sensing failure pulse will also be established on the EKG trace. The sensing failure may be permanently established notwithstanding recovery of sensing, until manually or clinically reset, as was done in connection with the loss of capture warning signal ST.sub.2C. Further, a counting/decoding system, equivalent to that employed in FIGS. 3 and 3B, may also be employed in FIG. 4, to establish a sensing failure condition only after a predetermined number of failures have occurred.

The ST signal can also be employed along with the complementary output of flip-flop 96 to count predetermined numbers of sensing failures in appropriate circuitry 104, shown generally, but corresponding to that employed in FIGS. 3 and 3B. The sensing failure signal SF as supplied by the driver 12 is sufficiently narrow in width (0.1 milliseconds or less) so as to not exhibit any stimulating characteristics.

Referring to FIG. 5, a timing circuit is shown for generating each of the above discussed wave forms in timed relationship. Each spaced signal is generated from the original ST signal from the generator 12, coupled via amplifier 208 and appropriately delayed by delays 200, 202, 204, 206, 209. The pulse width time durations are determined by the conventional monostable multivibrators 201, 203, 205, 207, 208A and 210, employed to generate the pulses used throughout the system.

It may also be noted that if a sensing failure occurs, simultaneous indications of loss of capture may also be displayed if the uninhibited stimulus ST occurs within the total (absolute plus relative) refractory period of the heart (about 100 to 250 milliseconds) which follows the occurrence of spontaneous activity. In such an event, the ST pulse initiates the usual loss of capture test, and since the R and D responses may not be evoked during this refractory period, loss of capture may also be indicated.

For further protection against faults, as previously described, additional redundancy may be provided. Thus, referring to FIG. 5, an additional monostable multivibrator 208A with a delay of 0.25 seconds, corresponding to recovery time of the R-wave amplifier 30, receives a stimulation signal P.sub.ST. The output of the multivibrator 208A is coupled to a further inhibitory gate 37 and amplifier 39. The output R of amplifier 38 and R' of amplifier 39 are buffered together in an OR gate buffer 41 and connected through the amplifier 43 to the point R, R' in FIG. 4. As stated previously, the use of redundant amplifiers protects against a failure in one R-wave amplifier, by providing a signal to FIG. 4, even where one of the R-wave amplifiers is inoperative.

An additional method of detecting a sensing failure is illustrated in FIGS. 6-8. This method makes use of the essential property of the pacer pulse generator and timer in its relation to the R-wave -- i.e., that since the R-wave must recycle the timer, a voltage change across a "timing capacity" within the timer must be observed in a specific time relation to the R-wave. If this voltage change does not occur, a sensing failure must follow. Thus, as illustrated; ZC = "Zero crossing" output in FIG. 6, wherein the voltage ESTO across the timing capacitance is shown as the saw-tooth wave form emerging from timing generator 10. For each zero crossing thereof, a ZC pulse is generated. The R'-R-wave timing is shown below the ZC signals.

The occurrence of an R-wave and the simultaneous absence of ZC means that the timing circuit has not been properly recycled. The output B is used in a manner identical to that described above to initiate the counting and marking of sensing failures. As before, the R-wave makes use of the redundant amplifiers in the pacer which produce R and R', and their complements. Logically, only (R + ZC) = sensing "miss". All of the other possibilities (R + ZC), (R + ZC) and R + ZC) are permitted.

The circuit for logical implementation of the method illustrated by FIG. 6 is shown in FIG. 7, and the timing relationship of such components in FIG. 8.

Referring to FIG. 7, the timer and pulse shaping circuit 10 (see FIG. 1) produces a saw-tooth output signal ESTO (FIG. 6) which is applied to a zero crossing detector 350, producing the ZC pulse (see FIG. 8). The ZC pulse is applied to the pulse widener 352 and to the inverter 354 to produce ZCw.

The signals R or R' (see FIG. 1) are applied through the OR gate buffer 356, pulse widener 358 and isolation amplifier 360 to produce the Rw signal. The output of 356 goes to pulse shaper 364 to provide the signal PS to delay 366 and isolation amplifier 368, thereby providing a sampling signal SS.

The signals ZCw and Rw are sampled by signal SS in gate 370A to provide the logic signal ZCw + R) to buffer 370D, and passes through an isolation amplifier 370E for indicating a sensing failure, B.

As indicated previously, in accordance with the invention, sensing failures may be encountered because one of the signal amplifiers (amplifier 26 more likely than amplifier 42) does not properly amplify spontaneous activity at its input, or because a component fails in either one of these amplifiers. In this case, the outputs R and R' will not agree, and a sensing failure is probable since pulse generator-timer 10 may not be properly recycled. This type of sensing failure is detected by gates 370B or 370C. If R and R' occur, they will be wide enough to surround the sampling signal SS. Thus, the absence of R with R' present, or the absence of R' with R present, will be indicated by the passing of SS through gates 370B and 370C respectively. The outputs of these gates are also taken to buffer 370D to produce an indication of sensing failure, B.

The sensing failure signals A and B, (FIG. 5) can be injected through OR gate buffer 330 directly into the optional counter as well as being introduced through buffer 99 to set the sensing failure indicator FF 328 (FIGS. 3 and 3B) through an OR gate buffer 99. Thus, a "miss" is detected as the absence of recycling (zero crossing) following the occurrence of an R-wave (B), (lack of coincidence between R and R' (B)) or by the premature release of a stimulus following an R-wave (A).

Referring now to FIG. 9, the technique is illustrated whereby alarm conditions indicating sensing or stimulating failures may be permanently registered for later diagnostic clinician observation. As shown, a narrow pulse or "spike" detector 400 detects the spike signals (FIG. 2, line E) and transmits these spikes over the line 402 as signal SD. The spikes to be detected if present, are the ST, SF and ST.sub.2C components of the EKG (cf. FIG. 2, lines A and E). A first pulse (ST) detector 404 responds to the SD signal, separating out of the first pulse and applying it to sensing delay 406 and a capture delay 408 equal to the normal SF and ST.sub.2C delays respectively. The MV circuits 410 and 412 are monostable multivibrators set for a predetermined pulse duration of, for example 10 milliseconds. The monostable multivibrator 410 is driven by the output of sensing delay 406. Similarly, monostable multivibrator 412 having a predetermined pulse duration delay period of, for example, 10 milliseconds is driven by the output of the capture delay circuit 408. The output of each of these circuits, monostable multivibrator 410 and monostable multivibrator 412 produce respectively MVS and MVC signals coupled to AND gates 414 and 416, giving rise to the output signals CFR and SFR. The AND gates 414 and 416 also receive input signals from the spike detector 400 along the SD line. As will be evidenced from an examination of the timing wave form shown in FIG. 11, the MVC and MVS are independently timed signals appearing at proper moments to surround the warning signals. Thus, any signals other than warning signals are excluded from being coupled to the sensing failure flip-flop 418 or the capture failure flip-flop 420 shown in FIG. 10. Activation of these flip-flops from one stable state, indicating no failure condition, to the other stable state, results in a specific indication of sensing or capture failure. An indicator mechanism illustrated by way of example as light sources 422 and 424, respectively, respond to the failure state to give a visual alarm. Of course, it will be understood that other types of sensing failure indicators may be used and that the illustration of light sources is intended as only exemplary. The sensing failure flip-flop or capture failure flip-flop are reset by the manual activation of the manual reset switch 426.

Thus, unless the pacer is suppressed, a situation which inhibits the operation of FIG. 9 since no ST pulse is produced, the circuit of FIG. 9 provides a means for detecting and isolating sensing or capture failures for monitor indication. If the pacer is suppressed, normal testing procedures require that it be put into a "magnetic mode" by external magnet actuation of reed switch 40. This will put the pacer into a fixed rate mode of operation and artificially generate an ST signal. As a result, the circuit of FIG. 9 can be actuated in either case, and the automation of detection of a prior intermittent sensing or capture failure in the pacer system is achieved.

Of specific advantage, the circuit of FIG. 9, can be used as a receiver at the remote end of a remote data link such as a conventional voice band telephone, microwave link, etc. Thus, the failure signals may be detected by coupling the pacer to an acoustic coupler, transmitting the signal over the telephone or other suitable data link as shown in FIG. 9, and detecting the received signal at the spike detector 400. Thus the detection of the prior intermittent sensing or capture failure may be achieved automatically. If the data link is replaced by a direct connection to the patient, such as by using a differential amplifier connected to detector 400, the detection of sensing or capture failures can be automatically detected by the locally actuated detection circuits.

Although the pacer can be placed into its fixed rate ("magnetic") mode by an external magnet which actuates reed switch 40, the failure indicators remain set since their reset is desirable only under clinical conditions. As noted previously, this may be done only by a clinical device which can actuate RF detector 395 and driving element 396. The clinician may thus effect reset, and reset all indicators, by activating an RF transmitter. Obviously, other means for insuring that the patient can perform a local monitoring without resetting the indicators can be realized as well.

Although not shown, an RF shield is mounted about all components to prevent interference. The RF shield is designed so that it does not prevent actuation of detector 395 by a locally applied RF field. All of the components are solid state electronic devices, miniaturized and compacted along with a power source into a single surgically implantable unit, for use as the internal unit.

Thus, an autodiagnostic pacer, operating in either fixed mode or demand mode, has been described with the advantages of sensing loss of capture and failure, and providing increased magnitude stimulating signals (to correct loss of capture) and appropriate warning signals which may be permanently maintained, even after the event has corrected itself.

In the external unit, the pacer may be provided with various features enabling the clinician to make adjustments. Typical of these are the gain controls symbolized by circles with arrows therein which are coupled to various units shown in the figures. These controls, while not normally present in the implantable units, may exist therein in the form of miniature components which may be permanently adjusted to each patient by the attending physician or surgeon prior to implant.

It will be understood that although various preferred time durations for pulses and signals have been set forth, they are intended to be exemplary only and not limiting. Variations in timing will be apparent to those skilled in the art within the ambit of the inventive scope.

While the invention has been disclosed with reference to a limited number of embodiments, it will be apparent that variations and modifications may be therein, and it is intended in the following claims to cover each such variation and modification as falls within the true spirit and scope of the invention.