Marquette 12SL Physicians Guide Rev C March 2019

Publication Information The information in this manual only applies to 12SL version 23. It does not apply to earlier product versions. Due to continuing product innovation, specifications in this manual are subject to change without notice. Marquette, MUSE, MAC, Hookup Advisor, SEER, and 12SL are trademarks owned by GE Medical Systems Information Technologies, Inc., a General Electric Company going to market as GE Healthcare. All other trademarks contained herein are the property of their respective owners. The document part number and revision are on each page of the document. The revision identifies the document’s update level. The revision history of this document is summarized in the following table. Revision

Date

Comment

A

3 April 2015

Internal release.

B

13 July 2015

Customer release.

C

29 March 2019

Updated for new findings.

To access other GE Healthcare Diagnostic Cardiology documents, go to the Common Documentation Library (CDL), located at http://apps.gehealthcare.com/servlet/ClientServlet?REQ=Enter+Documentation+Library, and click Cardiology. To access Original Equipment Manufacturer OEM) documents, go to the device manufacturer's website.

Intended Audience This manual is intended for qualified health care professionals using the 12SL ECG Analysis Program. It may also be useful for those who are not responsible for interpreting 12-lead ECGs but want to learn more about the capabilities and limitations of this medical device. See Contents for more details.

Manual Purpose: Ancillary Documentation, Product Labeling GE’s Marquette 12SL ECG Analysis Program is a prescriptive class II medical device, cleared by the Food and Drug Administration (FDA). The 12SL program does not directly acquire the ECG signal. It is used as a component in devices such as electrocardiographs, which digitize the analog ECG, and in computer systems, which receive digital 12-lead ECGs from other sources so that an initial ECG interpretation can be generated by 12SL for review and correction by a physician. The International Electrotechnical Commission (IEC) requires manufacturers to disclose the performance of ECG analysis programs used in diagnostic electrocardiographs. “The intent is that this performance information be readily available to customers who want to know the information. The intent is not to require expansion of OPERATOR documentation to include this performance information. This information may be disclosed in one of the documents that are created and made generally available by the manufacturer of an ELECTROCARDIOGRAPH. Examples of these documents are physician’s guides and technical notes, in addition to the OPERATOR’s guide.”[1] This 12SL Physician’s Guide is not an operator manual. It is ancillary to the operator’s manual and is considered product labeling. What the FDA terms “product labeling” extends beyond what is printed on the medical device or in an operator manual. It is brochures or any material regarding the product. If a manufacturer discloses the accuracy of the program in its Physician’s Guide, it needs evidence to support it.

Intended Use of Computerized ECG Computerized electrocardiography has been in existence since the late 1950’s. [2, 3] Despite its widespread use,[4] there is little written that directly addresses the intent of computerized electrocardiography. The pioneers of this technology had motivations which ranged from the esoteric - like demonstrating that a computer could mimic human intelligence to the basic need of recording artifact free tracings.[5] Some of the favorable developments which resulted from the evolution of this technology were hardly imagined at its inception. For example, computerized ECG has been shown to reduce the cost of managing ECG services, e specially as the volume of ECGs that need to be interpreted increases.[6] A major reason for this is that it reduces “analysis time by up to 24% to 28% for experienced readers.” [7] It should be made clear that a computerized analysis of the ECG is not a substitute for human interpretation. Statements of accuracy need to be viewed from a statistical perspective. Although accuracy levels may be high, outliers can and will exist. A computer does not have the ability to include the entire clinical picture of the patient. A person with organic heart disease can exhibit an ECG within normal limits. In a study of 391,208 patients with acute myocardial infarction, the initial ECG obtained in the emergency department was normal in 30,759.[8] Conversely, a normal individual can have an abnormal appearing ECG.[9, 10] The ECG must always be reviewed by a physician in the context of the patient and acted upon with sound clinical judgement.

Intended Use of 12SL Program as Registered with FDA (as recorded in 510k# K141963, cleared July 2014) The 12SL ECG Analysis Program assists the physician in measuring and interpreting resting 12-lead ECGs for rhythm and contour information by providing an initial automated interpretation. The interpretation by the analysis program may then be confirmed, edited, or deleted by the physician. The analysis program is intended for use in the general population ranging from healthy subjects to patients with cardiac and/or non-cardiac abnormalities. The analysis program is intended for use in hospitals, outpatient clinics, emergency departments, and out-of-hospital sites such as ambulances and patients’ homes. The ACS Tool option is intended for adult patient populations who are suspected clinically to have acute coronary syndrome.

Bibliography for this section 1.

2.

International Standard IEC 60601-2-51:2003 Medical electrical equipment. Particular requirements for safety, including essential performance, of recording and analysing single channel and multichannel electrocardiographs, I.E.C. (IEC), Editor. 2003, International Electrotechnical Commission (IEC). p. 86. Burch, G.E. and N.P. De-Pasquale, A History of Electrocardiography. 1964, Chicago: Year Book Medical Publishers. 309.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

2

3. 4. 5. 6. 7. 8. 9. 10.

Pipberger, H.V., R.A. Dunn, and A.S. Berson, Computer methods in electrocardiography. Annu Rev Biophys Bioeng, 1975. 4(00): p. 15-42. Kligfield, P., et al., Comparison of automated interval measurements by widely used algorithms in digital electrocardiographs. American Heart Journal, 2018. Macfarlane, P.W., et al., Comprehensive Electrocardiology. 2010: Springer London. Carel, R.S., Cost-effectiveness analysis of computerized ECG interpretation system in an ambulatory health care organization. J Med Syst, 1982. 6(2): p. 121-30. Schläpfer, J. and H.J. Wellens, Computer-Interpreted Electrocardiograms: Benefits and Limitations. Journal of the American College of Cardiology, 2017. 70(9): p. 1183-1192. Welch, R.D., et al., Prognostic value of a normal or nonspecific initial electrocardiogram in acute myocardial infarction. Jama, 2001. 286(16): p. 1977-84. Pelliccia, A., et al., Clinical Significance of Abnormal Electrocardiographic Patterns in Trained Athletes. Circulation, 2000. 102(3): p. 278-284. Hill, A.C., et al., Accuracy of interpretation of preparticipation screening electrocardiograms. J Pediatr, 2011. 159(5): p. 783-8.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

3

INTRODUCTION ... 5 The Marquette™ 12SL™ Program: A Brief History ...5 ECG Analysis/12SL Timeline ...5 Hazard Information ...6 Prescription Device ...6 AN OVERVIEW OF 12SL IN TWO PARTS ... 8 PART I: CRITERIA AND METHODOLOGY ... 9 Digitization of the Analog ECG ...9 Hookup Advisor™ ...11 Pacemaker Detection and Annotation ...14 Signal Conditioning and Removal of Noise ...16 Detection and Measurement ...22 Criteria – Rules for Interpretation ...32 Screening Criteria: Suppressed Statements, Increased Specificity ...148 ECG Classification ...150 Decision Support for Acute Coronary Syndromes (ACS) ...151 Critical Values ...176 Serial Comparison ...179 PART II: STATEMENT OF VALIDATION AND ACCURACY ...182 Overall Impact of Computerized ECG ...182 Development and Validation Process...190 Program Structure: Measurements Before Interpretation ...193 Testing of 12SL™ Measurements via Standardized Database...201 Impact of Hookup Advisor™ on Accuracy of ECG Measurements ...203 Independent Evaluation of 12SL™ Measurements ...205 Predictive Value/Clinical Correlation of 12SL™ Measurements ...215 Accuracy of Interpretive Statements: Reported Results ...225 Interpretation of Rhythm: Reported Results ...227 Contour Interpretation: Reported Results ...237 Overall Classification: Reported Results ...259 Serial Comparison ...260 Conclusion ...260 APPENDICES: STATEMENT LIBRARY, PEDIATRIC TABLES, AND 12SL VERSIONS ...266 Appendix A: Statement Library Arranged by Statement Category ...267 Appendix B: Statement Library Arranged by Statement Number ...283 Appendix C: Pediatric Tables ...303 Appendix D: 12SL Version Identification ...315 BIBLIOGRAPHY FOR ENTIRE DOCUMENT ...318

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

4

Introduction The Marquette™ 12SL™ Program: A Brief History The Marquette™ 12SL™ program has been in existence since 1980. It was the first commercially available program to analyze all 12 leads simultaneously recorded for the entire 10-seconds of the diagnostic resting ECG. In 1982, the 12SL program was embedded into a computerized electrocardiograph known as the MAC-II. It was the first of its kind, generating a 12 lead interpretation at the bedside in less than 10 seconds.[11] Since then, GE Healthcare has continued to evolve the Marquette™ 12SL™ program. The Marquette™ 12SL™ program has been validated on a variety of platforms beyond the diagnostic electrocardiograph, including bedside monitors, stress-testing systems, pre-hospital defibrillators, Holter recorders, and PCbased systems.

ECG Analysis/12SL Timeline 1980 – 12SL™ program introduced on MUSE™ system[11] 1982 – Incorporated into a computerized electrocardiograph: MAC-II™[11, 12] 1984 – 12SL™ Serial Comparison program[13] 1986 – Automated testing of 12SL using non-ECG, gold-standard databases[14] 1987 – Pediatric analysis, based on Davignon tables, incorporated into 12SL[15] 1988 – Analysis of extra leads, generating vector loops at an electrocardiograph [16] 1989 – Recognition of ST-elevated acute myocardial infarction (MI) in prehospital setting[17] 1991 – 12SL™ in a pre-hospital defibrillator equipped with 12-lead ECG[18] 1992 – 500 samples per second analysis, compression and storage[19] 1993 – 12SL™ in a bedside monitor, equipped with 12-lead ECG[20] 1995 – ACI-TIPI integrated into 12SL for prediction of acute cardiac ischemia[21] 1997 – Automated QT dispersion and T-wave principal component analysis.[22] 1998 – ECG Research Workstations for systematic assessment of ECG measurements [23-25] 1999 – MacRhythm: 12SL™ incorporates asynchronous P wave detector based on QRS subtraction [26] 2000 – Gender specific acute MI criteria[27] 2001 – Improved pacemaker detection using ECG acquired at 4,000 samples per second (SPS) [28] 2002 – 12SL™ in a Holter recorder, equipped with 12 lead ECG[29-31] 2003 – New 12SL™ QT algorithm,[25] validated by core lab on more than 40,000 ECGs[32] 2004 – Pattern recognition of noise via Hook-up Advisor™ tied to interpretation performance[33] 2005 – 12SL™ cleared for measurement and trending of 12-lead ambulatory recordings[34] 2006 – Recognition of acute right ventricular infarction via analysis of V4R [35] 2010 – Detection of biventricular and low energy artificial pacing on data acquired at 75K SPS [36] 2011 – Acute coronary syndrome tool based on the use of a neural network [37] 2012 – T-wave morphology measures related to hERG channel block [38-46]

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

5

2014 – Detection of left ventricular hypertrophy (LVH) in accordance with ACC recommendation [47] 2015 – Detection of Brugada Type 1 pattern in accordance with ESC guideline [48] 2017 – Combined LVH criteria improves prediction of stroke, myocardial infarction, etc. [49]

Hazard Information The terms Danger, Warning, and Caution are used throughout this manual to point out hazards, and to designate a degree or level of seriousness. Familiarize yourself with their definitions and significance. Hazard is defined as a source of potential injury to a person. DANGER indicates an imminent hazard which, if not avoided, will result in death or serious injury. WARNING indicates a potential hazard or unsafe practice which, if not avoided, could result in death or serious injury. CAUTION indicates a potential hazard or unsafe practice which, if not avoided, could result in minor personal injury or product/property damage. NOTE provides application tips or other useful information to assure that you get the most from your equipment. Additional safety messages that provide appropriate safe operation information may be found throughout this manual. WARNING: INTERPRETATION HAZARD 12SL analyses should be used only as an adjunct to clinical history, symptoms, and the results of other non-invasive and or invasive tests. 12SL analyses must be reviewed by a qualified physician.

Prescription Device CAUTION: United States federal law restricts this device to sale by, or on the order of, a physician.

Bibliography for this section 11. 12. 13.

14.

15. 16. 17. 18. 19. 20. 21. 22.

Rowlandson, I., Computerized electrocardiography. A historical perspective. Ann N Y Acad Sci, 1990. 601: p. 343-52. Rautaharju, P.M., Eyewitness to history: Landmarks in the development of computerized electrocardiography. Journal of Electrocardiology, 2016. 49(1): p. 1-6. Rowlandson, I., Strategy for Serial Comparison. Proceedings of the 1986 Engineering Foundation Conference, Computerized Interpretation of the Electrocardiogram XI. New York: Engineering Foundation., 1986: p. 106109. Rowlandson, I., New Techniques in Criteria Development. Proceedings of the 1985 Engineering Foundation Conference, Computerized Interpretation of the Electrocardiogram X, 1985(New York: Engineering Foundation 1985): p. 177-184. Drazen, E., et al., Survey of computer-assisted electrocardiography in the United States. J Electrocardiol, 1988. 21 Suppl: p. S98-104. Reddy, B.R., D.W. Christenson, and G.I. Rowlandson, High-resolution ECG on a standard ECG cart. J Electrocardiol, 1988. 21 Suppl: p. S74-9. Rowlandson, I., P.J. Kudenchuk, and P.P. Elko, Computerized recognition of acute infarction. Criteria advances and test results. J Electrocardiol, 1990. 23 Suppl: p. 1-5. K903644 SERIES 1500 12-LEAD ANALYSIS OPTION - 510K. 1991, U.S. Food and Drug Administration, Center for Devices and Radiological Health. Reddy, B.R.S., et al., Data compression for storage of resting ECGs digitized at 500 samples/second. Biomed Instrum Technol, 1992. 26(2): p. 133-49. K921669, MARQUETTE SL SERIES TRANSPORT REMOTE ACQUISITION - 510K, in U.S. Food and Drug Administration, Center for Devices and Radiological Health. 1993. Aufderheide, T.P., et al., Test of the acute cardiac ischemia time-insensitive predictive instrument (ACI-TIPI) for prehospital use. Ann Emerg Med, 1996. 27(2): p. 193-8. Aufderheide, T.P., et al., QT dispersion and principal component analysis in prehospital patients with chest pain. Computers in Cardiology, 1997: p. 665-668.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

6

23. 24. 25. 26. 27. 28.

29. 30. 31. 32. 33. 34. 35.

36. 37. 38.

39. 40. 41. 42. 43. 44. 45. 46. 47.

48.

49.

Xue, Q. and B.R. Reddy, Clinical research workstation, U.S.P. Office, Editor. 2002 GE Medical Systems Information Technologies, Inc. (Waukesha, WI): United States. K981024, QT DISPERSION AND T WAVE ANALYSIS PROGRAM (QT GUARD ANALYSIS SYSTEM), in U.S. Food and Drug Administration, Center for Devices and Radiological Health. 1998. Xue, Q. and S. Reddy, Algorithms for computerized QT analysis. J Electrocardiol, 1998. 30 Suppl: p. 181-6. Reddy, B.R., et al., Prospective evaluation of a microprocessor-assisted cardiac rhythm algorithm: results from one clinical center. J Electrocardiol, 1998. 30 Suppl: p. 28-33. Xue, J., et al., A new method to incorporate age and gender into the criteria for the detection of acute inferior myocardial infarction. J Electrocardiol, 2001. 34 Suppl: p. 229-34. Taha, B. and S. Reddy, Method and apparatus for automatically detecting and interpreting paced electrocardiograms, U.S.P. Office, Editor. 2001, GE Medical Systems Information Technologies, Inc. (Waukesha, WI): United States. Batchvarov, V., K. Hnatkova, and M. Malik, Assessment of noise in digital electrocardiograms. Pacing Clin Electrophysiol, 2002. 25(4 Pt 1): p. 499-503. Batchvarov, V., Short and Long Term Reproducibility of the QT/RR Relationship in Healthy Subjects. Journal of American College of Cardiology, 2001. 37(2_Supp_A): p. 1A-648A. Batchvarov, V., Bazett Formula is not Suitable for Assessment of the Circadian Variation of the Heart Rate Corrected QT Interval. Journal of American College of Cardiology, 2001. 37(2_Supp_A): p. 1A-648A. Hnatkova, K., et al., Precision of QT Interval Measurement by Advanced Electrocardiographic Equipment. Pacing Clin Electrophysiol, 2006. 29(11): p. 1277-84. Farrell, R. and B. Young, Effect of Lead Quality on Computerized ECG Interpretation. Computers in Cardiology, 2004. 31: p. 173-176. K042782, SEER MC, in U.S. Food and Drug Administration, Center for Devices and Radiological Health. 2005. K060833, 12SL ECG ANALYSIS PROGRAM; Computerized detection of right ventricular involvement in acute inferior myocardial infarction, in U.S. Food and Drug Administration, Center for Devices and Radiological Health. 2006. Ricke, A.D., et al., Improved pacemaker pulse detection: clinical evaluation of a new high-bandwidth electrocardiographic system. J Electrocardiol, 2011. 44(2): p. 265-74. Xue, J., et al., Added value of new acute coronary syndrome computer algorithm for interpretation of prehospital electrocardiograms. J Electrocardiol, 2004. 37 Suppl: p. 233-9. Graff, C., et al., Identifying drug-induced repolarization abnormalities from distinct ECG patterns in congenital long QT syndrome: a study of sotalol effects on T-wave morphology. Drug safety : an international journal of medical toxicology and drug experience, 2009. 32(7): p. 599-611. Graff, C., et al., T-wave morphology reveals greater effect of d,l-sotalol than QTc. Journal of Electrocardiology, 2008. 41(6): p. 644. Andersen, M.P., et al., A Robust Method for Quantification of IKr-Related T-Wave Morphology Abnormalities. Computers in Cardiology, 2007: p. 341-344. Andersen, M.P., et al., Repeatability of T-wave morphology measurements: superiority of a principal component analysis–based lead. Journal of Electrocardiology, 2007. 40(6): p. Pages S81-S81. Graff, C., et al., T-wave convexity measure outperforms notch criterion as diagnostic marker for the HERG genotype. Journal of Electrocardiology, 2007. 40(6S): p. S82-S83. Haarmark, C., et al., Independent novel T-wave descriptors of repolarization. Journal of Electrocardiology, 2007. 40: p. S131. Graff, C., et al., T-wave Morphology as a Covariate in Drug-induced QTc Prolongation. Comput. Cardiol, 2009. 36: p. 589-592. Porta-Sánchez, A., et al., T-Wave Morphology Analysis in Congenital Long QT Syndrome Discriminates Patients From Healthy Individuals. JACC: Clinical Electrophysiology, 2016. Xue, J., et al., Study of repolarization heterogeneity and electrocardiographic morphology with a modeling approach. Journal of Electrocardiology, 2008. 41(6): p. 581-587. Hancock, E.W., et al., AHA/ACCF/HRS Recommendations for the Standardization and Interpretation of the Electrocardiogram: Part V: Electrocardiogram Changes Associated With Cardiac Chamber Hypertrophy A Scientific Statement From the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society Endorsed by the International Society for Computerized Electrocardiology. Journal of the American College of Cardiology, 2009: p. j.jacc.2008.12.015. Priori, S.G., et al., 2015 ESC Guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death. The Task Force for the Management of Patients with Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death of the European Society of Cardiology (ESC)Endorsed by: Association for European Paediatric and Congenital Cardiology (AEPC), 2015. Okin, P.M., et al., Combining ECG Criteria for Left Ventricular Hypertrophy Improves Risk Prediction in Patients With Hypertension. Journal of the American Heart Association, 2017. 6(11).

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

7

An Overview of 12SL in Two Parts To show an overview of the Marquette 12SL ECG Analysis Program, we can follow the same steps the software uses to analyze an ECG. Start with acquisition, followed by detection, measurement, etc. Often, a physician wants to know the clinical evidence regarding the performance of the program, not the steps it took to analyze the ECG. Consequently, this manual is divided into two parts: Part I: Criteria and Methodology Part II: Statement of Validation and Accuracy A side-effect of approaching this from both perspectives is that portions of the document will appear redundant. Whenever Part-II happens to cover the same topic as Part-I, the emphasis will not be on the “how”, but rather, “how well” the program performed the task. Given there are over a hundred peerreviewed scientific articles about 12SL, there are well over a hundred pages of performance metrics that must be presented in a series of tables using a format defined by the IEC. The approach of dividing the 12SL Physician’s Guide into two parts allows the guide to be used as a reference manual. Instead of having to read the 12SL Physician’s Guide from cover-to-cover, you should be able to find the information you need using the table of contents and the hyperlinks provided throughout the document.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

8

Part I: Criteria and Methodology Digitization of the Analog ECG Simultaneous 12-Lead Acquisition In 1979, GE Healthcare introduced simultaneous recording of 12 leads so that the computer could use all signals from all 12 leads to properly detect and classify each QRS complex.[11, 12] The Common Standards for Electrocardiography independently verified the advantage of this technique: “Conclusion: The simultaneous recording and analysis of all 12 standard leads...is certainly an improvement over the conventional recording of three leads at a time. Similarly...multilead programs proved to be more stable than those obtained by conventional programs analyzing three leads at a time...”[50] Although the 12SL program can be used in a variety of ECG devices, the 12SL program only analyzes data simultaneous recorded for 10 seconds from at least 12 leads . Eight of the leads are acquired directly (I, II, and

V1 through V6). The remaining four are derived via Einthoven's law (III) and Goldberger’s equations (aVR, aVL, and aVF): •

III = II - I

•

aVR = -(I + II)/2

•

aVL = I - II/2

•

aVF = II - I/2

Because of the inherent relationship of the standard limb leads to each other, Einthoven stated that at any given instant during the cardiac cycle, the sum of the potentials of leads I and III equals the potential of lead II. Similarly, Goldberger said that the sum of the three augmented leads at any instant in time equals zero (aVR + aVL + aVF = 0).

Most formats show only a portion of the 12-lead, 10-second data. An example of this is the standard 12-lead presentation which displays only 2.5 seconds from each of the 4 lead groups. Regardless of the data that you see, the complete data is always acquired. This is used by the 12SL analysis program for precise waveform measurement. It also allows you to choose from a multiple set of formats for accurate rhythm and contour diagnosis. This ability to acquire all leads simultaneously complies with the American Heart Association recommendations.[51]

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

9

Sampling Rate All resting electrocardiographs currently sold by GE Healthcare, analyze the waveform at a minimum of 500 samples per second (SPS). In some GE Healthcare resting electrocardiographs, the ECG is sampled at a much higher rate, such as 4,000 SPS. This is referred to as over-sampling and it used by the device to generate an average, cleaner signal at 500 SPS. Specifications for electrocardiographs, across the industry, often cite the raw sample rate (e.g. 4K SPS or higher) without clarifying that the ECG analysis and measurement software executes on data with a lower sample rate. Current guidelines for resting ECG analysis cite 500 SPS,[52] which is the minimum sample rate executed by 12SL. In some GE Healthcare electrocardiographs, the 12SL program can be configured to analyze the ECG at 1,000 SPS. Before the physiological data is sampled, analog filtering is applied. These filters attenuate high-frequency electrical noise that is not part of the physiological signal. If these analog filters were not present in the device, high-frequency signals could be digitized by the device and appear as low frequency noise, intermixed with the physiological cardiac signal. To eliminate this possible source of contamination, GE Healthcare applies an analog filter, known as an anti-aliasing filter. To detect high-frequency artifacts generated by electronic cardiac implants, GE Healthcare developed a patented[53, 54] high-bandwidth acquisition system that runs in parallel with the acquisition system for the physiological signal.[55-57] In some systems available from GE Healthcare there are two parallel digital data streams for analysis: one at 2K SPS (for the physiological signal from 0.04 to 250Hz), the other at 75K SPS (for pacemaker detection from 22 to 11KHz). The pacemaker channel is analyzed at 75K SPS.

Bibliography for this section 11. 12. 50. 51.

52. 53. 54. 55.

56.

57.

Rowlandson, I., Computerized electrocardiography. A historical perspective. Ann N Y Acad Sci, 1990. 601: p. 34352. Rautaharju, P.M., Eyewitness to history: Landmarks in the development of computerized electrocardiography. Journal of Electrocardiology, 2016. 49(1): p. 1-6. Willems, J.L., et al., A reference data base for multilead electrocardiographic computer measurement programs. J Am Coll Cardiol, 1987. 10(6): p. 1313-21. Kligfield, P., et al., Recommendations for the Standardization and Interpretation of the Electrocardiogram. Part I: The Electrocardiogram and Its Technology. A Scientific Statement From the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society. Endorsed by the International Society for Computerized Electrocardiology. Circulation, 2007. ANSI/AAMI, Diagnostic electrocardiographic devices, EC11:1991/(R)2001, Editor. 1992. p. 39. Ricke, A. and G.I. Rowlandson, System and method of detecting and diagnosing pacing system malfunctions US Patent: 7,970,472. 2011, General Electric Company (Schenectady, NY) United States. Ricke, A. and G.I. Rowlandson, Module and device for discerning therapeutic signals from noise in physiological data US Patent: 8,170,655. 2012, General Electric Company (Schenectady, NY) United States. Ricke, A.D., et al., The relationship between programmed pacemaker pulse amplitude and the surface electrocardiogram recorded amplitude: application of a new high-bandwidth electrocardiogram system. Journal of Electrocardiology, 2008. 41(6): p. 526-530. Petrutiu, S., et al., High Resolution Electrocardiography Optimised for Recording Pulses from Electronic Pacemakers: Evaluation of a New Pacemaker Sensing System, in Computers in Cardiology. 2007, IEEE: Durham, North Carolina, USA. Ricke, A., et al., Advanced pacemaker detection. Journal of Electrocardiology, 2007. 40(6, Supplement 1): p. S33S33.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

10

Hookup Advisor™ Operator manuals exist for all GE Healthcare devices that acquire ECGs. These manuals specify proper electrode positions and patient preparation for obtaining a quality ECG. The following serves as a reminder to physicians and administrators the importance of quality control. This is especially true given how ECG services are often dispersed throughout the hospital, diminishing the role of the “Heart Station” in setting acceptable standards. GE Healthcare’s Hookup Advisor™ scores the 12-lead ECG for signal quality and encapsulates this information with the ECG before it is sent to the MUSE system. From there, analytical tools available on the MUSE system can be used to determine the origin of poor-quality ECGs so that corrective action can be taken. Hookup Advisor provides real-time feedback to the person acquiring the ECG. Hookup Advisor statements appear only on the screen during ECG acquisition on cardiographs that have the Hookup Advisor turned on. These statements never appear in an original interpretation.

Methodology Based on Pattern Recognition, Not Skin Impedance As opposed to measuring skin impedance, which has been found to be poorly correlated with signal quality,[58] Hookup Advisor uses pattern recognition on ECGs manually scored by cardiologists for acceptable quality. After such training, automated quality scores generated by Hookup Advisor have been found to be predictive of the accuracy of automated interval measurements as well as rhythm interpretations.[33, 59] See graphs of reported performance metrics for Hookup Advisor in Part II.

Proper Electrode Placement for Diagnostic Resting ECG In addition to artifacts, incorrect placement of electrodes can have a negative impact on the diagnostic value of the ECG. Although a limb-lead reversal has the most pronounced effect, a study of 150 subjects found that moving limb electrodes onto the torso shifted the P/QRS/T axis rightward and eliminated approximately 50% of significant Q-waves in cases of an old inferior infarction.[60] Although less obvious than a limb lead reversal, swapping chest electrodes is a common cause of poor Rwave progression and false positive interpretations of anterior-septal infarction.[61] In a study of 60 patients with known cardiac disease, ECG morphology changes became evident when chest electrodes were moved beyond 2 cm from their proper location, with V2 being the most sensitive to displacement errors.[62]

Electrode Placement: Continuous Monitoring versus Diagnostic 12-lead ECG In some monitoring environments, all 6 chest leads are applied with the result being that continuous 12-lead ECGs can be acquired by the monitor. Under this circumstance, the limb leads are usually put back on the torso using the Mason-Likar or Lund positions.[63] As already stated, this results in QRS axis changes and, in some case, the elimination of significant Q-waves in inferior leads. This practice is done to reduce noise and the tangling of lead wires since in the monitored patient has been shown to be beneficial for capturing transient ST/T wave changes due to acute ischemia,[64, 65] T-wave alternans,[66] complex drug-induced T-wave changes[67, 68] or transient arrhythmias that need to be effectively localized for ablation.[69] To reduce the confusion resulting from leaving the limb-leads on the torso, some institutions use the following techniques: •

Identify 12-leads coming from bedside monitors using a torso configuration for the limb-leads.

•

Do not continuously send 12-leads to the MUSE system - instead, sequester the 12-leads in the monitoring environment that come from continuous acquisition.

•

Send only 12-leads from the monitor that reflect an important change or only send those when the frontal plane configuration of the electrodes has been returned to the limbs.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

11

Consistency of Serial ECGs Even when the skin is marked as to where to place chest electrodes for repeat ECGs for serial analysis, normal day-to-day variation is considerable, especially with respect to QRS voltage measurement in the precordial leads.[70-72] Nevertheless, an undisciplined approach to recording ECGs increases the variability of ECG measurements and interpretive findings by both the computer and physician.[73, 74] This is unfortunate, given the growing evidence that serial ECG measurements can be predictive of heart failure or other serious clinical conditions.[75-78] In conclusion, studies have shown a significant incidence of limb lead reversal and wide variability in chest electrode placement, even among experienced ECG technicians.[79, 80] Training in proper lead positioning has demonstrated a reduction in these errors.[81] Periodic retraining should be routine for all personnel who are responsible for the recording of ECGs as recommended in clinical guidelines.[82]

Bibliography for this section 33. 58. 59. 60. 61. 62. 63.

64. 65. 66.

67.

68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80.

Farrell, R. and B. Young, Effect of Lead Quality on Computerized ECG Interpretation. Computers in Cardiology, 2004. 31: p. 173-176. Wiese, S.R., et al., Electrocardiographic motion artifact versus electrode impedance. IEEE Trans Biomed Eng, 2005. 52(1): p. 136-9. Farrell, R.M. and G.I. Rowlandson, The effects of noise on computerized electrocardiogram measurements. J Electrocardiol, 2006. 39(4 Suppl): p. S165-73. Farrell, R.M., et al., Effects of limb electrode placement on the 12- and 16-lead electrocardiogram. J Electrocardiol, 2008. 41(6): p. 536-545. Zema, M.J., et al., Electrocardiographic poor R wave progression III. The normal variant. J Electrocardiol, 1980. 13(2): p. 135-42. Kania, M., et al., The effect of precordial lead displacement on ECG morphology. Medical & Biological Engineering & Computing, 2014. 52(2): p. 109-119. Welinder, A., et al., Differences in QRS Axis Measurements, Classification of Inferior Myocardial Infarction, and Noise Tolerance for 12-Lead Electrocardiograms Acquired From Monitoring Electrode Positions Compared to Standard Locations. Am J Cardiol, 2010. 106(4): p. 581-586. Drew, B.J., M.M. Pelter, and M.G. Adams, Frequency, characteristics, and clinical significance of transient ST segment elevation in patients with acute coronary syndromes. Eur Heart J, 2002. 23(12): p. 941-7. Fesmire, F.M., Which chest pain patients potentially benefit from continuous 12-lead ST-segment monitoring with automated serial ECG? Am J Emerg Med, 2000. 18(7): p. 773-8. Takasugi, N., et al., Continuous T-wave alternans monitoring to predict impending life-threatening cardiac arrhythmias during emergent coronary reperfusion therapy in patients with acute coronary syndrome. Europace, 2011. Zairis, M.N., et al., Continuous 12-lead electrocardiographic ST monitoring adds prognostic information to the thrombolysis in myocardial infarction risk score in patients with non-ST-elevation acute coronary syndromes. Clin Cardiol, 2005. 28(4): p. 189-92. Sarapa, N., et al., Electrocardiographic identification of drug-induced QT prolongation: assessment by different recording and measurement methods. Ann Noninvasive Electrocardiol, 2004. 9(1): p. 48-57. Nof, E., W.G. Stevenson, and R.M. John, Catheter Ablation for Ventricular Arrhythmias. Arrhythmia & Electrophysiology Review, 2013. 2(1): p. 45-52. Van Den Hoogen, J.P., et al., Reproducibility of electrocardiographic criteria for left ventricular hypertrophy in hypertensive patients in general practice. Eur Heart J, 1992. 13(12): p. 1606-10. Farb, A., R.B. Devereux, and P. Kligfield, Day-to-day variability of voltage measurements used in electrocardiographic criteria for left ventricular hypertrophy. J Am Coll Cardiol, 1990. 15(3): p. 618-23. Willems, J.L., P.F. Poblete, and H.V. Pipberger, Day-to-day variation of the normal orthogonal electrocardiogram and vectorcardiogram. Circulation, 1972. 45(5): p. 1057-64. Schijvenaars, B.J.A., et al., Effect of electrode positioning on ECG interpretation by computer. Journal of Electrocardiology, 1997. 30(3): p. 247-256. Schijvenaars, B.J., G. van Herpen, and J.A. Kors, Intraindividual variability in electrocardiograms. J Electrocardiol, 2008. 41(3): p. 190-6. Okin, P.M., et al., All-cause and cardiovascular mortality in relation to changing heart rate during treatment of hypertensive patients with electrocardiographic left ventricular hypertrophy. Eur Heart J, 2010. Okin, P.M., et al., Incidence of heart failure in relation to QRS duration during antihypertensive therapy: the LIFE study. J Hypertens, 2009. Jongh, M.C.d., et al. Serial ECG analysis after myocardial infarction: When heart failure develops, the ECG becomes increasingly discordant. in 2016 Computing in Cardiology Conference (CinC). 2016. Shamim, W., et al., Incremental changes in QRS duration in serial ECGs over time identify high risk elderly patients with heart failure. Heart, 2002. 88(1): p. 47-51. Wenger, W. and P. Kligfield, Variability of precordial electrode placement during routine electrocardiography. J Electrocardiol, 1996. 29(3): p. 179-84. Garcia-Niebla, J., et al., Technical mistakes during the acquisition of the electrocardiogram. Ann Noninvasive Electrocardiol, 2009. 14(4): p. 389-403.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

12

81. 82.

Thaler, T., et al., The frequency of electrocardiographic errors due to electrode cable switches: a before and after study. Journal of Electrocardiology, 2010. 43(6): p. 676-681. Campbell, B., et al., Clinical Guidelines by Consensus: Recording a standard 12-lead electrocardiogram. An approved method by the Society for Cardiological Science and Technology (SCST) 2017. , in Resting ECG, S.f.C.S.T. (SCST), Editor. 2017, SCST.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

13

Pacemaker Detection and Annotation Over the last decade, there has been a significant increase in permanent pacemaker implantation as well as advancements in pacemaker technology. A worldwide survey found that virtually all countries showed increases in the number of pacemaker implants.[83] More specifically, the United States “had the largest number of cardiac pacemaker implants (225,567) and Germany the highest new implants per million population (927).” [83] With regards to technological advancements of artificial pacing, consider the following: •

Artificial stimulation used to be confined to a single location: the apex of the right ventricle. Fast forward to the advent of cardiac resynchronization therapy (CRT) and artificial stimulations occurring in the right atrium as well as the right and left ventricle. Now optimum resynchronization therapy is being explored via multipoint pacing of the left ventricle.[84]

•

Lead wires used to only support unipolar pacing. Now even bipolar pacing is being replaced by leadless pacing.[85]

•

Pacemaker pulses observed at the body surface were so large, standards had to be developed to make monitoring manufacturers avoid falsely detecting them as QRS complexes. [86] Now, they “are often too small to be recognized on the standard ECG.”[51]

•

The minimum timing intervals between artificial stimuli was relatively fixed and certainly greater than 100ms. Now these are configurable and the interval so small, multiple pulses can appear as single artifact on the surface ECG.[87] Examples of Artificial Pacing, Then and Now Older Pacing Technology

Contemporary Artificial Pacing

All of this makes the interpretation of the paced ECG difficult and given it has been reported that roughly 10% of in-hospital ECGs are paced,[88] it has become a significant challenge for both the computer[89] and human reader.[85, 90] To combat this, GE Healthcare (Milwaukee, WI) developed a new high-bandwidth acquisition system to detect artificial stimuli that runs in parallel with the acquisition system for the physiological signal. There are two parallel digital data streams for analysis: one at 2K SPS (the physiological signal from 0.04 to 250Hz), the other at 75K SPS (the pacemaker detector channel from 22 to 11KHz). The pacemaker channel is analyzed at 75K SPS. By sampling the high-frequency spectrum, the challenge is to discriminate the electrical stimuli generated by the artificial pacemaker versus other high-frequency noise unrelated to pacing the heart, such as a left ventricular assist device (LVAD), pacemaker programmer or electro-static discharge. In 2010, this system was prospectively evaluated on patients with implanted pacemakers (different vendors at different settings) and challenged with differing levels of noise.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

14

The sensitivity for the detection of artificial pacing exceeded 99% while the positive predictive value remained at 100% regardless of the level of noise. [36] This system can detect pulses as small as .5mV and 0.2ms, which is several times more sensitive than the AAMI standard of 2mV and 0.5ms, and provides pacemaker annotations, including indications of biventricular pacing. In accordance with AHA/ACC/HRS recommendations, these annotations are supplied separately from the waveform in a “single row of the standard output tracing.”[51]

Bibliography for this section 36. 51.

83. 84. 85.

Ricke, A.D., et al., Improved pacemaker pulse detection: clinical evaluation of a new high-bandwidth electrocardiographic system. J Electrocardiol, 2011. 44(2): p. 265-74. Kligfield, P., et al., Recommendations for the Standardization and Interpretation of the Electrocardiogram. Part I: The Electrocardiogram and Its Technology. A Scientific Statement From the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society. Endorsed by the International Society for Computerized Electrocardiology. Circulation, 2007. Mond, H.G. and A. Proclemer, The 11th world survey of cardiac pacing and implantable cardioverter-defibrillators: calendar year 2009--a World Society of Arrhythmia's project. Pacing Clin Electrophysiol, 2011. 34(8): p. 1013-27. Zanon, F., et al., Multipoint pacing by a left ventricular quadripolar lead improves the acute hemodynamic response to CRT compared with conventional biventricular pacing at any site. Heart Rhythm, 2015. 12(5): p. 975-81. Mulpuru, S.K., et al., Cardiac Pacemakers: Function, Troubleshooting, and Management. Part 1 of a 2-Part Series, 2017. 69(2): p. 189-210.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

15

86.

AAMI/ANSI/ISO 60601-2-27:2011, Medical electrical equipment - Part 2-27: Particular requirements for the basic safety and essential performance of electrocardiographic monitoring equipment., I.E.C. (IEC), Editor. 2011, International Electrotechnical Commission (IEC). Chien, C.H.S., et al., A software algorithm for detection of biventricular pacemaker pulses in the surface ECG. Journal of Electrocardiology, 2013. 46(6): p. 622-623. Poon, K., P.M. Okin, and P. Kligfield, Diagnostic performance of a computer-based ECG rhythm algorithm. J Electrocardiol, 2005. 38(3): p. 235-8. Guglin, M.E. and N. Datwani, Electrocardiograms with pacemakers: accuracy of computer reading. Journal of Electrocardiology, 2007. 40(2): p. 144-146. Kenny, T., The Nuts and bolts of Paced ECG Interpretation. 2011: Wiley.

87. 88. 89. 90.

Signal Conditioning and Removal of Noise In the presence of noise, both physicians and computers make frequent mistakes.[91] If there is a way to remove noise from the signal without reducing the clinical value of the ECG, it should be pursued. This is known as signal conditioning, done by removing signals that exhibit characteristics which could not possibly be generated by the heart. Fortunately, the characteristics of the P-QRS-T have been well studied and there is plenty of documentation as to the limitation of which frequencies can be generated by the heart.[1, 52, 86, 92, 93] Frequencies below 0.67Hz cannot be generated by the heart. If the recorded signal exhibits a waveform below 0.67Hz, it is not of cardiac origin.[94] Frequencies above 20Hz only occur during a QRS complex and only for a brief period of time (<20ms), such as during the onset, peak or notch of a QRS.[95] The duration of a QRS complex resulting from an intact conduction system can be no more than 140ms long. [96] As a result, a complex that is longer than 200ms which contains frequent, high-frequency components (>20Hz) cannot come from the heart; it is more likely due to electrode motion artifact. The signal conditioning and removal of noise performed in conjunction with 12SL, includes the following topics: •

Removal of AC interference

•

Removal of baseline wander

•

Removal of high frequency artifact

•

Upper cut-off frequency

Removal of AC Interference A good example of a signal that has a characteristic which could not be generated by the heart is the signal that results from the radiated or conducted energy of wires or devices powered by alternating current (AC). As opposed to the QRS complex, AC interference is continuous and sinusoidal.

GE Healthcare electrocardiographs have a configurable setting for the removal of AC interference. The setting - either 50 or 60 cycles per second (Hz) - should match the line/mains frequency of the power grid where the electrocardiograph is operating. This allows the system to select a filter that specifically targets that frequency. A filter that attempts to eliminate a single frequency from the spectrum of frequencies is often referred to as a “notch filter”. Notch filter used in GE Healthcare’s electrocardiographs does more than simply attenuate either 50 or 60Hz. It locks onto the artifact and measures its amplitude as well as shape. Instead of

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

16

eliminating either 50 or 60Hz, the system forms a model of the noise and then subtracts it from the raw waveform.[97] There a couple of advantages to an adaptive AC interference notch filter. First, the filter attenuates the signal at just the right amount – no more, no less. Secondly, because the filter is synched up with the AC interference, it does not attenuate the naturally occurring transient 50/60 Hz frequencies that reside within the QRS complex. This is because the model of the artifact only adapts to a continuous 50/60Hz signal, not a transient signal.

Removal of Baseline Wander Baseline wander can be due to respiration, perspiration, body movements, loose electrodes, dry electrodes or the lack of using Ag/AgCl electrodes versus other electrode designs. [98] Measuring the ECG can be challenging in the presence of such artifact. In fact, if the baseline is wandering so much the signal does not remain on the page or saturates an amplifier, measurement of the ECG is not possible. Even in mild cases of baseline wander, the assessment of the ST-segment deviation from the raw ECG will be compromised since its amplitude should be measured in relation to QRS onset. A representative complex generated from this data is not immune to the problem. It will incorporate the amplitude variation occurring across each QRS complex. This will be particularly noticeable in the ST- segment of the representative complex. The slope of the ST segment will be a composite of the wandering baseline immediately following each QRS.

Baseline wander is a low frequency signal. Filters that remove low frequencies are referred to as high-pass filters since they pass along higher frequencies yet leave behind lower frequencies. To deploy a high-pass filter, it is important to know the lowest possible frequency generated by the heart so that it will remain untouched by the filter. This can be determined via the heart rate. If the heart rate is 60 beats per minute (bpm), the lowest possible frequency is 1 cycle per second or 1Hz. “Heart rates below 40 bpm (0.67 Hz) are uncommon in practice.”[51] The 2007 ACC/AHA recommendations for standardization of the ECG stipulate that frequencies below 0.67Hz can safely be removed from the ECG. [51] Not all high pass filters are alike. Some not only attenuate low frequencies but shift them in time versus the high frequency components of the signal. This is known as phase distortion. Following is an example of the use of a high pass filter that exhibits phase distortion. [99] As the filter setting progressively goes beyond 0.05Hz, the ST segment becomes so distorted it appears to be an ST-elevated acute myocardial infarction (STEMI). While using a high pass filter with phase distortion, the only way to preserve the ST segment is to use a less aggressive filter setting (≤.05Hz.). This comes at the expense of not correcting the baseline.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

17

Example of ST-Segment Distortion Due to High-Pass Filter from the Literature.[99]

With the advent of digital sampling and storage of the ECG, high-pass filters can be designed so that they have zero phase distortion (ZPD). The use of a ZPD high-pass filter at a setting of 0.67Hz can “correct baseline drift while preserving the fidelity of the ST-segment.”[51] Since the introduction of the 12SL program, a ZPD high pass-filter (≤0.32Hz) has been used to remove baseline sway. Baseline wander is aggressively removed from the 12-lead report without ST-segment distortion. The representative complex generated by 12SL reveals a ST-segment not contaminated by baseline wander. Until recently, real-time rhythm strips were another matter. Not until the advent of the MAC VU360 or MAC 2000 has it been possible to use a ZPD high pass filter when acquiring and printing continuous rhythm strips. Via these newer products, the ST segment on a continuous rhythm strip will be the same as the ST-segment on the 12-lead report, even at the most aggressive filter setting, which on the MAC VU360 is 0.56Hz. This means anyone printing a rhythm strip does not have to be trained to properly contend with the tradeoff of selecting the appropriate filter setting to either preserve the ST-segment with a lower setting (≤0.05Hz), or remove the baseline sway via a higher setting (>0.05Hz). With these newer products, the high-pass filter setting can be set once and behave the same way for both the rhythm and 12-lead report without ST-segment distortion. ZPD is an important advancement now available when printing a rhythm strip. Consider that a study conducted in an emergency department found, that as opposed to the 12-lead ECG reports, 93% of rhythm tracings had clinically significant alterations that could be construed as an acute coronary syndrome (ACS) due to the use of a baseline roll filter without ZPD at a setting of 0.5Hz.[100] ZPD on both the rhythm and 12lead report eliminates this confusion. The following four diagrams are useful for describing what changed in GE Healthcare’s high-pass filter design. The first diagram portrays a high-pass filter that continuously corrects the baseline in real-time as the signal is acquired. The second diagram shows the ST-segment distortion that results when such a filter encounters a QRS complex that is either tall or long. The third diagram shows how 12SL can do the same operation without ST-segment distortion, because the system can correct the entire recording in both directions. The final diagram shows a high pass filter that utilizes digital rhythm. It digitizes the rhythm for a couple of seconds before correcting the waveform. In the latter case, the system can use the samples before and after the point where it corrects the baseline. More precisely, the first diagram shows the computer estimating the baseline sway and then subtracting it from the incoming signal. In real-time, the amplitude of each sample is measured relative to the middle of the channel. The estimate of the baseline sway is determined by having a running tally of a fraction of these amplitudes. That fraction becomes larger as the high-pass filter setting increases.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

18

Unfortunately, this filter cannot discriminate between a large or wide QRS complex versus baseline sway. See figure below. The filter can be unduly influenced by the QRS, resulting in an overshoot immediately after the QRS. This effect is apparent in the ECG presented above. Note that those leads with the largest QRS complex (such as V2) have the greatest distortion while those leads exhibiting a smaller QRS complex (such as aVL) have almost no ST-segment distortion. When using a filter of this type, the distortion from the ST segment can only be eliminated using a filter setting of 0.05Hz, but the baseline wander will not be removed.

The 12SL program has always been able to remove low frequencies (i.e., < 0.32Hz) without ST-segment distortion. It does this by running this same filter, forwards and backwards, over the entire 10 seconds. In this way, both sides of the QRS are similarly impacted and the ST is no longer depressed in relation to QRS onset.

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

19

The MAC VU 360 and MAC 2000 have obsoleted this approach. Instead, the correction of each sample point is based on a weighted average of the samples before and after it.

To accomplish this on a scrolling rhythm strip, the signal must be digitized and buffered for a couple seconds before it is displayed. This 2 second delay enables the filter to aggressively correct the baseline without STsegment distortion. To keep abreast of the ACC/AHA recommendations, which relaxed the high-pass filter setting from 0.05Hz to 0.67Hz for filters capable of ZPD, IEC/AAMI issued new performance standards for the low frequency response of diagnostic electrocardiographs.[92] This includes a simple test which can be performed by a biomedical engineer to evaluate the low-frequency response of any electrocardiograph and determine whether it uses a high-pass filter with ZPD. A 3mV, 100ms square wave fed into an electrocardiograph should not result in an artifact that exceeds 100µV; otherwise the user must select a lower filter setting (0.05Hz) to preserve the ST segment.

Removal of High-frequency Artifact Electrocardiographs have various low-pass filter settings, including 40Hz, 100Hz, or 150Hz. The lower the filter setting, the more aggressively the filter removes high frequency signals, which includes noise due to muscle tremor, electrode-motion artifact, etc. These low-pass filters also operate on the entire ECG signal and attenuate all high-frequency elements of the ECG signal, such as the QRS complex and pacemaker artifacts. To consistently measure the resting ECG and capture the proper QRS amplitude, the 12SL program always analyzes the ECG at the AHA / AAMI recommended full bandwidth of 150Hz, [52, 93] regardless of the low-pass filter setting. These settings are sometimes referred to as “writer settings”, since they do not affect the ECG interpretation. It should be noted, that all filter settings travel with the ECG. That is, the MUSE system can be configured to either portray the ECG signal as it was acquired at the electrocardiograph or at another specified filter

2056246-002C

Marquette™ 12SL™ ECG Analysis Program

20