System for prediction of cardiac arrhythmias

• US 6,308,094 B1
• Filed: 08/17/1999
• Issued: 10/23/2001
• Est. Priority Date: 04/28/1998
• Status: Expired due to Fees

• Current Assignee: University of Pittsburgh of The Commonwealth System of Higher Education
• Original Assignee: University of Pittsburgh of The Commonwealth System of Higher Education
• Inventors: Shusterman, Vladimir, Aysin, Benhur, Grave, Ilan, Chaparro, Luis F., Anderson, Kelley P.
• Primary Examiner(s): Kamm, William E.
• Application Number:
  US09/376,027
• Time in Patent Office: 798 Days
• Field of Search: 600/515-518
• US Class Current: 600/516
• CPC Class Codes: A61B 5/363 Detecting tachycardia or bradycardia
• CPC Class Codes: A61B 5/7232 involving compression of the physiological signal, e.g. to extend the signal recording period
• CPC Class Codes: A61B 5/726 using Wavelet transforms

Abstract

A medical device and method for predicting cardiac arrhythmias, by gathering electrocardiographic data such as intervals between heart beats (RR-series) or other signal, mathematically decomposing or compressing the signal into several elements or components that contain the most significant information and tracking the changes in the several elements. The signal may be divided into time windows, and the signals decomposed into a plurality of coefficients or components such as Karhunen Loeve Transformation (KLT) coefficients that are predictive of the occurrence of a cardiac arrhythmia. The electrocardiographic data may be generated real-time, on-line, or be prerecorded data.

Claims

1. A medical device for predicting cardiac arrhythmias comprising:

means for receiving data from a patient of electrocardiographic activity over a period of time (hours);

means for dividing said data into a plurality of time segments;

means for decomposing said data from each time segment into a plurality of component signals that denote the magnitude and complexity of said electrocardiographic activity in each time segment; and

means for providing an output signal corresponding to each of said component signals for comparison with the component signals from the other time segments to facilitate predicting the likelihood of the occurrence of a cardiac arrhythmia.

2. A medical device as set forth in claim 1 in which said data from each segment is divided into consecutive time segments of several minutes each.

3. A medical device as set forth in claim 1 in which said means for monitoring data monitors a patient'"'"'s cardiovascular data on-line, real time.

4. A medical device as set forth in claim 3 in which said cardiovascular data comprises a patient'"'"'s RR-series.

5. A medical device as set forth in claim 1 in which said means for decomposing said data decomposes the data into at least two to ten component signals.

6. A medical device as set forth in claim 1 including an action unit for noting the number of component signals that exhibit substantially simultaneous changes and for causing action designed to prevent the occurrence of a cardiac arrhythmia.

7. A medical device as set forth in claim 1 which monitors predefined parameters that are used for the detection of changes in said time segments.

8. A medical device as set forth in claim 1 in which said means for decomposing said data includes means for calculating Karhunen-Loeve Transform (KLT) coefficients.

9. A medical device as set forth in claim 1 which includes means for comparing the component signals from said time segments for pattern recognition of changes that are predictive of a life threatening cardiac arrhythmia.

10. A cardiac rhythm monitoring system for predicting probability of a patient having a cardiac arrhythmia comprising:

means for monitoring a patient'"'"'s RR-series of the intervals between heart beats;

means for characterizing said RR-series using Karhunen-Loeve Transform (KLT) and generate KLT-coefficients indicative of both linear and nonlinear changes in the RR-series; and

means for determining the magnitude of said changes in the RR-series by using the time varying mean and variance of said KLT-coefficients and determining the complexity of said changes by calculating the number of KLT-coefficients that exhibit substantially simultaneous changes.

11. A system as set forth in claim 10 which includes means for displaying said KLT-coefficients to be read by medical personnel.

12. A system as set forth in claim 10 that includes means for dividing said RR-series during monitoring of a patient into consecutive time windows, and said changes in the RR-series provide a comparison of the signals in said time windows.

13. A system as set forth in claim 12 which includes means for calculating of KLT-eigenvectors based on said RR-series, spectral analysis of said eigenvectors, multiplication of their frequency representations by the corresponding KLT-coefficients in each time window, summing the products of said multiplication to obtain an approximate frequency representation (AFR) of the signal in each time window, and subtracting the AFR from the preceding time window to provide frequency characteristics of change in said RR-series.

14. A system as set forth in claim 10 in which KLT-coefficients are calculated by the following formula:

${\hat{C}}_x=\frac{1}{M}<br><br>\underset{i=1}{\overset{N}{\sum <br><br>}}<br><br>\text{\hspace{1em}<br><br>}<br><br>\left(x_i-{\hat{m}}_x\right)<br><br>{\left(x_i-{\hat{m}}_x\right)}^T$${\hat{m}}_x=\frac{1}{M}<br><br>\underset{i=1}{\overset{M}{\sum <br><br>}}<br><br>x_i$$y=A<br><br>\left(x-m_x\right)$

in which

C_x=covariance matrix

M=number of vectors in the pattern

N=length of vector

x_i=vectors

m_x=the mean vector

y_i=vector of KLT-coefficients in the window.

15. A medical device for monitoring a patient'"'"'s electrocardiographic data for predictive signs of increased risk of the occurrence of cardiac arrhythmia comprising:

cardiac rhythm monitoring means for monitoring an RR-series of the intervals between a patient'"'"'s heart beats;

computer means for dividing said RR-series into consecutive, nonoverlapping time windows of preselected duration and converting said RR-series into Karhunen-Loeve Transformation (KLT) coefficients that vary as a function of time; and

means for displaying said KLT-coefficients.

16. A medical device as set forth in claim 15 that includes an action unit for causing appropriate preventative actions.

17. A medical device as set forth in claim 15 which includes means for calculating KLT-eigenvectors based on said RR-series, spectral analysis of said eigenvectors, multiplication of their frequency representations by the corresponding KLT-coefficients in each time window, summing the products of said multiplication to obtain an approximate frequency representation (AFR) of the signal in each time window and subtracting the AFR from the proceeding time window to provide frequency characteristics of changes in said RR-series.

18. A medical device as set forth in claim 15 that is in the form of a cardiac rhythm monitoring device for hospital use.

19. A medical device as set forth in claim 15 that is suitable for out-of-hospital use.

20. A medical device as set forth in claim 15 that comprises an automatic external defibrillator.

21. A medical device as set forth in claim 15 that comprises a pacemaker.

22. A medical device as set forth in claim 15 that comprises a system for diagnostic ECG processing and analysis.

23. A medical device for predicting LTCA comprising means for decomposing a signal of electrocardiographic data into a plurality of elements, which contain the most significant information about the signal and tracking the changes in said elements.

24. A medical device as set forth in claim 23 that includes means for causing appropriate action to prevent a LTCA.

25. A method for predicting the risk of a patient having a cardiac arrhythmia comprising:

continuously monitoring a patient'"'"'s cardiac rhythm;

generating an RR-series of cardiac rhythm, said RR-series comprising the intervals between heart beats in consecutive time windows;

calculating KLT-coefficients from said RR-series, which KLT-coefficients vary as a function of time to show the magnitude and complexity of changes in said RR-series between time windows;

displaying said KLT-coefficients; and

reading said KLT-coefficients to predict the probability of the patient having a cardiac arrhythmia.