This invention relates to a method of determining whether events in a heart's electrical activity result in corresponding blood flow events, and to an external defibrillator implementing such method.
Out of hospital survival from sudden cardiac death (SCD) is now a regular occurrences[1]. The most important factor influencing survival is the time from the onset of the condition to treatment being administered. Since automatic external defibrillators (AEDs) allow treatment by emergency services and trained personnel on site rather than requiring that the patient is transported to a medical facility, their use is becoming more widespread on a daily basis. More recently their use is being made a requirement on flights, in some public places and offices by government bodies. There is also clinical and public pressure to make them available to family members and smaller institutions.
A modern AED is a compact portable unit which automatically analyses electrocardiograph (ECG) rhythms obtained from the human thorax via electrodes in contact with the skin. A user attempting to respond to a suspected SCD emergency would simply get the AED and press the “ON” button and then follow the voice and text prompts and instructions generated by the device. A typical sequence of events would be for the device to prompt the user to apply the pads to the patients chest and make sure that the electrodes are properly connected to the device. Once the device is satisfied that this has been performed, the device will then automatically analyse the patients heart rhythm and make a diagnosis. If it determines that a high voltage shock is required, it will automatically charge to a predetermined energy level and advise the user to press a clearly marked “SHOCK” button. Upon pressing this button, the energy is delivered to the patient through the same electrodes used to monitor the ECG. Whether or not the device advises shock, the user is informed continuously about the state of the patient and/or the actions and states of the device.
In order to analyse the patients ECG these AED's use an algorithm which employs various parameters measured from the digitised ECG[2-3]. Typically these parameters include frequency and amplitude of the ECG rhythm and also some integration techniques such as slope, morphology and heart rate. Many algorithms also utilise zero content or baseline content and energy ratio calculations. Calculating the mathematical variance of some parameters can also improve the performance of these devices.
All of the AED's using this approach however suffer from the fact that they are analysing the ECG or parameters measured thereof. Although there is a great deal of literature to show how the ECG is related to the mechanical and haemodynamic response of the heart, there are many instances where ECG activity is present when there is very little or no cardiac output. Furthermore many heart rhythms can exhibit a random and fluctuating nature while maintaining a sufficient cardiac output. The algorithms therefore can and do advise and fail to advise incorrectly[4-6]. What is really required is a device which can correctly identify and differentiate a cardiac rhythm associated with no pulse or haemodynamic collapse from those with satisfactory cardiac output.
More recent work in the field of impedance cardiography[7] has enabled an impedance cardiograph (ICG) to be measured from the same two electrodes used to acquire the ECG and deliver the energy to the patient. Historically measuring the ICG had required four band type electrodes in order to achieve satisfactory accuracy. Utilising the electrodes already employed for ECG allows the ICG to be easily measured as another signal and source of information. The ICG contains information relating to the presence of cardiac output rather than electrical cardiac activity and can therefore be used to help classify heart rhythms more accurately and consistently. Referring now to the top trace in
dZ/dt=d[Z−Z
o
]/dt (1)
where Z is the raw impedance measured from the patients thorax and Zo is the baseline impedance of Z.
Note how the dz/dt (ICG) signal although very different in morphology, also possesses a beat by beat nature caused by the rhythmic pumping of the blood through the heart, thereby causing a change in the impedance of the thorax. Comparing this to
There are many scientific publications relating to the use of the ICG in measuring cardiac output. All of these publications however fail to address the one main problem still to be solved before the ICG can be used in a practical setting. There are various methods of acquiring an ICG signal. They all involve an alternating signal being applied across the sensing electrodes. This signal is either designed to possess a fixed voltage or a fixed current. The effect that the object under test exhibits on the test signal is measured and the impedance for the object is then calculated. A component of this measurement is the impedance of the blood within the heart. As this blood volume changes during a cardiac cycle, the measured impedance changes. Unfortunately the impedance measured due to blood volume change within the heart is considerably less than the total impedance of the subjects torso. This means that the ICG is sensed as a small change in value of a much larger baseline impedance. Other factors such as volume change due to inspiration and expiration also affect the measurement. Even very slight muscle activity and motion can affect the electrode contact impedance and considerably affect measurement. All these factors result in the use of the ICG for measurement of cardiac output being restricted to laboratory or controlled conditions.
There are many techniques known to those skilled in the art which can reduce the effects of these factors. Some methods involve filtering, signal averaging[7], Fourier analysis and differential impedance measurement. There are also documented methods which analyse the ICG once it has been acquired. Typically these have involved the use of mathematical derivative and integrative processes attempting to glean measurement of cardiac output from the ICG.
According to the present invention there is provided a method of determining whether events in a heart's electrical activity result in corresponding blood flow events, the method comprising:
generating a signal indicative of the degree of concordance of the ICG with the ECG as a function of measurements made in respect of said detected waveform features.
The invention further provides an external defibrillator having circuitry arranged to implement the method specified above and to selectively advise a shock dependent upon the value of said signal indicative of the degree of blood flow.
An embodiment of the invention will now be described, by way of example, with reference to the accompanying drawings, in which:
As previously described the prior art has included many techniques that employ parameters determined from the ICG. The intent has been to measure the amount of cardiac output to subsequently determine whether or not to deliver therapy. Any practical and theoretical attempts to provide more accurate diagnosis by AED algorithm's have therefore used this determination in various ways to augment the determination made from the ECG.
However, the following presents a novel method which does not place the emphasis on measuring the amount of cardiac output or blood flow with any serious degree of accuracy. The basis of this may not immediately appear to be a directive that could possibly furnish a solution to the overall problem. It has however achieved exactly that.
The essence of the method involves “gating” the information supplied by the ICG with that obtained from the ECG. There is no direct attempt to quantitatively measure blood flow with any degree of accuracy. Rather the aim is to qualitatively determine whether or not there is any concordance between the cardiac events depicted by the ECG and those demonstrated by the ICG. This novel approach of ECG-ICG event gating has proved very effective in discriminating shockable from non-shockable VT's in both human and animal subjects, and it is likely that the approach is applicable to other conditions such as atrial flutter/fibrillation and VF. It must be appreciated that the gating can be applied to any rhythm and use a multitude of parameters measured from both ECG an ICG without departing from the scope of the invention.
The gating uses ECG and ICG signals which have first been processed using methods well know to those skilled in the art. In the case of the ICG these remove as much interference, contamination and variation as possible while leaving the fundamental ICG intact. Unlike the prior art however, this pre-processing can be performed much more aggressively. The reason for this is that the invention does not require an accurate quantitative measurement of the cardiac output. The pre-processing design has therefore a far larger degree of freedom to clean the signal, thereby almost totally removing interference due to movement, etc., a solution not previously possible using other approaches. The frequency bandwidth overlap of ECG content and interference has always been a troublesome dilemma for biomedical engineers. The electrical activity of the heart occurs much faster than the mechanical contraction, which means that the ICG content due to heart contraction and cardiac output results in a content bandwidth in the ICG that incorporates more lower frequencies. This means that the dilemma is much worse for the ICG than that of the ECG. With many of the pre-processing constraints removed however the signal conditioning that can be used for this invention provides a signal that has a far higher signal to noise ratio than can be used by the prior art. This means the ultimate result is more reliable especially in “out of hospital” emergency response situations.
The AED comprises conventional patient electrodes 10 which are applied to the patient to obtain both the ECG and ICG waveform signals simultaneously. Signal conditioning circuitry 12 filters each signal to a bandwidth of 3-20 Hz using analog filters with a Butterworth response. The circuitry 12 further differentiates the ICG to obtain dz/dt as the ICG signal for subsequent analysis, and individually but simultaneously digitises the ECG and differentiated ICG signals. The digitised ECG and ICG signals are passed to respective microprocessor-based feature extraction circuits 14, 16 respectively.
The circuit 14 examines successive heartbeat cycles (periods) of the ECG to detect successive occurrences of each of a plurality of periodic waveform features E1 to EN indicative of electrical activity of the heart. One such feature is the R wave, shown at the top of
The circuit 16 examines successive heartbeat cycles (periods) of the ICG to detect successive occurrences of each of a plurality of periodic waveform features I1 to IN indicative of blood flow in the heart. However, the ICG waveform features I1 to IN are not arbitrarily chosen, but each is paired with one of the ECG waveform features E1 to EN such that each pair En, In (0<n<N+1) has an electromechanical relationship. By this it is meant that the particular ECG waveform feature En results (in a healthy heart) in the corresponding ICG waveform feature In. For the R wave in the ECG, the paired waveform feature in the ICG is the C wave, shown at the bottom of
Next, the detected features are passed to a temporal comparator 18. This defines a period of time (search period) following the detection of each ECG waveform feature En and examines the ICG waveform for the presence of the paired feature In during that search period. The duration of the search period is calculated from heart-rate dependant measurements made on the ECG and is designed such that the paired feature (if present) should appear within that time period following the detected ECG feature. The search period duration will therefore depend on the heart rate and the particular waveform features selected. In this particular embodiment the search period was calculated as twice the period of the QRS interval in the ECG.
Where a valid pair is detected, i.e. where the ICG feature In is found within the search period following the related ECG feature En, a measurement (hereinafter referred to as a parameter) is made in respect of each feature of the pair En, In. The parameters are not chosen arbitrarily, but are equivalent for the feature En and the feature In. Thus if the parameter for En is the width of the R wave, the equivalent parameter for In is the width of the C wave. Other parameter pairs are the height of R wave in the ECG and the height of C wave in the ICG, the ratio of the areas under the R and Q waves in the ECG and the ratio of the areas under the C and X waves in the ICG, and other parameter pairs which would be known to those skilled in the art. In the example shown in
This process, applied to all the features E1 to EN in each period of the ECG waveform, generates for each waveform period a respective set of parameter pairs X1src to XNsrc where src denotes the source of the parameter ECG or ICG. Thus X1ECG represents the value of parameter number one for the ECG and X1ICG the value of its ICG pair. For the case where a paired feature was not found, and hence a corresponding ICG parameter could not be measured, a large error value was substituted.
Next, the parameter pairs X1src to XNsrc are passed to a concordance estimator 20. In this embodiment the estimator generates the linear sum of all the comparisons of all the parameter pairs:—.
For the case particular parameters shown in
∂X
1
└X
n
ECG┘R-R−└XnICG┘C-C=∂R1−∂C (3)
where X1 is the identifier for that parameter pair.
Since in this particular example the parameter pairs are chosen such that they should have very little difference during concordance, this means that the estimate Y furnished a very low value when there was concordance between the ECG and ICG and a very large value when there was bad correlation or during disassociation when a large error value contributed. Thus it will be recognised that Y is indicative of the lack of concordance of the ICG with the ECG, the larger Y the higher the lack of concordance (it will be appreciated that a high degree of concordance indicates a rhythmic blood flow which is life-supporting and the patient does not need a shock).
Finally the estimate Y was selectively used by a diagnostic algorithm 22 which ignored its value for various rhythms but compared its value to a predetermined threshold when it had already classified the rhythm as VT in addition to several other criteria outside the scope of this specification. If the concordance estimate was below the threshold for the given VT then no shock was advised. Therefore, as a final qualification a shock was advised if:
Y≧Eth (4)
Thus the shock circuitry 24 of the AED gave a shock or no shock advisory according to the value of Y.
This technique proves very effective in reducing false positives for the algorithm 22. The result was also consistent when signals were analyzed which contained a large degree of impedance interference. The use of the technique is thought to be useful for application to rhythms other than VT, and it must be appreciated that this will not depart from the scope of the invention.
The technique described above provides a novel qualitative approach applied after the ICG has been acquired and conditioned. This technique overcomes the problems described in relation to the prior art and provides a method whereby the ICG can be employed in a practical environment to reliably discern shockable from non-shockable cardiac rhythms.
The invention is not limited to the embodiment described herein which may be modified or varied without departing from the scope of the invention.
Number | Date | Country | Kind |
---|---|---|---|
S2004/0468 | Jul 2004 | IE | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/EP05/07454 | 7/7/2005 | WO | 00 | 1/8/2007 |