Phys_Guide.fm

GE Healthcare

Physician’s Guide

to GE Stress Systems

2051167-002 ENG Revision C

2 Physician’s Guide Revision C

2051167-002 2014-01-07

CASE and MAC are trademarks owned by GE Medical Systems

Information Technologies GmbH, a General Electric Company going to

market as GE Healthcare.

All other trademarks contained herein are the property of their

respective owners.

Revision History

This manual is subject to the GE Medical Systems Information

Technologies change order service.

The revision code, a letter that follows the document part number,

changes with every manual update.

The initial version of the manual has the letter A.

Part No. Revision Code Date Comment

2009352-001 Rev A 2001-12 Initial Release

2009352-001 Rev B 2003-08 ECO 073369

2009352-001 Rev C 2005-06 ECO 080698

2009352-001 Rev D 2010-02 Replaced by 2051167-002 Rev A

2051167-002 Rev A 2010-02 New: Exercise Test

Interpretation, Audio

Assessment, four new exercise

measurements.

Updated: Interpreting/Correcting

TWA.

2051167-002 Rev B 2012-01 New: Exercise induced

arrhythmias.

Exercise Test Interpretation:

Statement texts are more

moderate and more descriptive.

Audio Assessment: Channels

exchanged

2051167-002 Rev C 2014-01 New: Exercise devices in

exercise test interpretation (XTI).

Graphical XTI output

Revision C Physician’s Guide 3

2051167-002

Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Computerized Stress ECG Analysis and Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Acquisition of the ECG Signal . . . . . . . . . . . . . . . . . . . . . . . . . 9

ECG Signal Acquisition with GE Stress Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Characteristics of the ECG Signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Digitization of the ECG Signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Environmental Noise Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3 Conditioning the ECG Signal . . . . . . . . . . . . . . . . . . . . . . . . 15

The Importance of Noise-Free ECGs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Noise Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Baseline Roll Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

50/60-Hz Line Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Cubic Spline Baseline Correction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

FRF (Finite impulse response Residual Filtering) Algorithm . . . . . . . . . . . . . . . . . . . . 19

4 Signal Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Mathematical Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Intelligent Lead Switch Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Detection of Cyclic Artifact Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

QRS Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Correlation and Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Incremental Updating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Relearn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Pace Enhance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4 Physician’s Guide Revision C

2051167-002

5 Measurement Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Heart Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

ST Segment Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

QRS Onset and Offset Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

ST Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

ST Slope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

ST Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

ST Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

R-Wave Amplitude . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

QRS Width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

ST/HR Slope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

ST/HR Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Recovery ST Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

ST/HR Hysteresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Arrhythmia Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Arrhythmia Detection Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Exercise-Induced Wide QRS Tachycardia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Exercise-Induced Supraventricular Tachycardia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Duke Treadmill Score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

METS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Actual Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

HR Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

FVE Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Chronotropic Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Blood Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Double Product / Rate Pressure Product (RPP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Heart Rate Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Dynamic ST Scan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

ST Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

PWC Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Target Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

6 T-Wave Alternans (TWA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Intended Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

The TWA Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Cubic Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

TWA Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

Revision C Physician’s Guide 5

2051167-002

Interpreting/Correcting TWA Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

TWA Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

7 Exercise Test Interpretation (XTI) . . . . . . . . . . . . . . . . . . . . . 67

Description of the Exercise Test Interpretation Program . . . . . . . . . . . . . . . . . . . . . . . 67

Limitations of the Exercise Test Interpretation (XTI) Program . . . . . . . . . . . . . . . . . . . 70

Examples for complete interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

Rules for risk assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

Rules for functional response assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

Rules for ischemia assessment (coronary artery disease) . . . . . . . . . . . . . . . . . . . . 76

Rules for overall statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Technical Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

Graphical XTI output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

8 Audio Assessment of Exercise Tests . . . . . . . . . . . . . . . . . . 83

Fast Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

9 Resting ECG Interpretation and Pre Test Risk

Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

AHA Coronary Heart Disease Risk and Stroke Risk Prediction . . . . . . . . . . . . . . . . . . 87

Coronary Heart Disease Risk Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Stroke Risk Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Definitions of Terms and Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

10 Exercise Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Role of Stress Testing in Reducing Cost of Healthcare . . . . . . . . . . . . . . . . . . . . . . . . 92

A Clinical Approach to Exercise Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

Concepts Useful in Interpreting Exercise Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Percentage of Predicted Maximal Heart Rate (MHR) Achieved . . . . . . . . . . . . . . . . 95

Total Exercise Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Total METS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Functional Aerobic Impairment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

Percent of Predicted O2 Consumption Achieved . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

Symptoms and Signs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

6 Physician’s Guide Revision C

2051167-002

Electrocardiographic Response to Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

Heart Rate and Blood Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

Exercise Test Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

The ST Segment Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

Problems in the Interpretation of ST Segment Responses to Exercise . . . . . . . . . 101

Sensitivity, Specificity, Predictive Value, and Pretest Likelihood . . . . . . . . . . . . . . . 102

Other Exercise Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Computer Processing of the Exercise ECG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

11 Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

In-Test and Final Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

12 Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Abbreviations and Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

13 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

Relevant Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

Revision C Physician’s Guide 7

2051167-002

1 Introduction

Computerized Stress ECG Analysis and

Interpretation

The first analysis of the ECG signal on a digital computer was performed

by Taback in 1959

. At that time he and his group reported the benefits

of this revolutionary technology as: an ability to make more precise

measurements, a reduction in the distortion of the ECG signal, and

accessibility to storage techniques.

Computerization of the stress ECG has evolved greatly since that

original attempt. There are four generally recognized benefits to the

computerization of the ECG signal:

1. The advanced signal-processing techniques make it possible to

present clean electrocardiograph tracings in the presence of severe

artifact and noise.

2. Computer-generated measurements are consistent and eliminate

the well documented interobserver variability.

3. Properly constructed algorithms have the ability to accurately

recognize and remove aberrant beats from the signal processing.

4. The application of computers is making it possible to generate more,

and better, criteria for positive exercise tests, for example, automatic

Exercise Test Interpretation, Duke Treadmill Score, and ST/HR

Criteria.

The Physician’s Guide to GE Stress Systems has been written to provide

the practitioner with a conceptual insight into GE’s innovative methods

of signal processing and exercise test interpretation. Our hope is that it

will enhance the application of computerized exercise

electrocardiography to your practice.

The software that is at the heart of GE stress systems is in its third

decade of continued development. While it encompasses many aspects of

advanced signal processing, data is presented in formats that preserve

continuity with traditional clinical methods.

The functionalities described in this guide are provided by the

GE software package HEART Exercise containing the XTI exercise

test interpretation feature. It is installed in different GE Stress Systems

and has a separate, device-independent 510K FDA approval. The

specifics presented in this guide refer to GE Stress Systems, e.g., CASE,

CardioSoft*, CS*, MAC 1200, and MAC 1600. They describe the

capabilities of GE stress products. Depending on whether the device is a

8 Physician’s Guide Revision C

2051167-002

Introduction: Computerized Stress ECG Analysis and Interpretation

high end or a low end product, or whether options have been purchased

or not, some of the functions described may not be available on your

device. Consult your operator’s manual for details.

Note

Please note that the products CardioSoft and CS may not be

available in some countries.

Revision C Physician’s Guide 9

2051167-002

2 Acquisition of the

ECG Signal

ECG Signal Acquisition with GE Stress Systems

GE stress systems simultaneously acquire all 12 leads of the

conventional ECG. Eight of the leads are acquired directly (I, II, and V1

through V6). The remaining four (III, aVR, aVL, and aVF) are derived

via Einthoven’s law.

Figure 2.1 Frontal plane leads shown with reference to Einthoven’s triangle

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. (This complies with American Heart Association

recommendations

Most report formats show only a portion of a 10-second ECG strip. For

example, the standard 12-lead presentation displays only 2.5 seconds

from each of the four lead groups. Regardless of the data seen on hard

copy reports, GE stress systems simultaneously acquire and process the

complete database of 12 leads for the duration of the complete procedure.

RA LA

LEAD I

LEAD II

LEAD III

10 Physician’s Guide Revision C

2051167-002

Acquisition of the ECG Signal: ECG Signal Acquisition with GE Stress Systems

Note

For those clinicians interested in testing with more or less than the

conventional 12-lead electrocardiogram, GE systems provide the

capability for simultaneous acquisition of 3, 6 or 15 leads. Consult

your operator’s manual for details.

GE systems with color capability indicate the signal quality of the

applied electrodes on-screen, using different colors:

green: the electrode is properly applied

yellow: poor signal quality

red: electrode disconnected, high impedance, or lead break

white: not used

Sometimes pacemaker spikes are not visible in the ECG signal. This can

occur when the spikes have a low amplitude, e.g., less than 2 mV, and a

short duration, e.g., less than 0.1 ms. GE stress systems are able to

detect small pacemaker pulses and provide the possibility to enhance

them, i.e., the user can enable the pace enhance function.

Revision C Physician’s Guide 11

2051167-002

Acquisition of the ECG Signal: Characteristics of the ECG Signal

Characteristics of the ECG Signal

The raw electrocardiograph signal is obtained in analog form. An analog

signal is defined as a continuous signal which varies in amplitude with

time. Figures 2.2 and 2.3 illustrate analog signals.

Figure 2.2 A 2-Hz analog signal

The frequency of an analog signal is defined as the number of complete

cycles that occur per second. The frequency of a signal is labeled in Hertz

(Hz). Figure 2.2 is an example of a low-frequency wave. Notice that it

does not repeat rapidly. The frequency is 2 Hz (2cycles per second).

Figure 2.3 10-Hz analog signal

A higher frequency wave is shown in Figure 2.3. It is compressed and

repeats more often. The frequency is 10 Hz.

Figure 2.4 presents a raw (analog) ECG signal. The typical ECG

representation of the cardiac cycle consists of various waveforms that

vary in frequency. The general frequency characteristics of ECG data

are presented in Table 2.1.

AMPLITUDE

TIME

1.0

.50

AMPLITUDE

TIME

1.0

.50

12 Physician’s Guide Revision C

2051167-002

Acquisition of the ECG Signal: Characteristics of the ECG Signal

Figure 2.4 The ECG waveform shown as a continuous analog signal with both amplitude

and time components

Table 2.1 General frequency characteristics of an ECG signal

ST Segment — 0.05 Hz to 5 Hz

Baseline Roll — < 1 Hz

P Wave — < 5 Hz

T Wave — < 5 Hz

QRS Complex — 10 to 40 Hz

Muscle Artifact — > 35 Hz

AMPLITUDE

TIME

Revision C Physician’s Guide 13

2051167-002

Acquisition of the ECG Signal: Digitization of the ECG Signal

Digitization of the ECG Signal

Digitizing an analog signal such as the ECG requires periodic sampling

at fixed time intervals.

Figure 2.5 shows the effects of analog-to-digital conversion. Note how

the sampling rate affects the resolution of the signal. In general, the

greater the sampling rate the better the resolution.

Figure 2.5 The effects of different sampling rates on the analog-to-digital conversion of the

ECG signal (Froelicher, V. F., Exercise and the Heart 3

edition

)

Incoming analog signals are digitized with GE’s unique patient

acquisition module. Data is acquired at a rate of more than 1,000

samples per second resulting in values every 0.05 mm at standard chart

writer speed. (GE systems comply with and exceed American Heart

Association recommendations.)

100 samples/second

10 samples/second

with phase shift

10 samples/second

14 Physician’s Guide Revision C

2051167-002

Acquisition of the ECG Signal: Environmental Noise Elimination

Environmental Noise Elimination

To acquire cardiac waveforms accurately, we have taken great care to

design our systems to exclude environmental noise. For the purposes of

this publication, we will define environmental noise as that artifact

which originates outside of the system. Several noise-exclusion

mechanisms are employed.

Let us first discuss noise generated by signals external to the patient’s

body; that is, in addition to the small voltages generated by the

myocardium, the ECG equipment receives signals coming from electrical

equipment in the environment. These signals are called common mode

because all of the leads on the body “see” them: they are common to all of

them. A common mode signal can be many times greater than the ECG.

Therefore, it is important to eliminate.

The ability of the electrocardiograph to reject this signal is called

common mode rejection. Due to practical limitations, it is not possible to

entirely eliminate the common mode signal. However, we are able to

greatly reduce it. The amount of reduction is called the common mode

rejection ratio.

With GE’s common mode rejection ratio only 1 part in a million is left.

For example, if the common mode signal is 100 volts, only 100 microvolts

would be left in the ECG recording.

In addition to the high quality rejection circuitry, there is another way to

minimize the deleterious effects of the common mode signal. Imagine we

could perfectly couple the system and patient together; both system and

patient would experience the same common mode signal. In absolute

terms, the common mode signal would still exist, but the acquisition

hardware would not see it in relation to the patient’s body.

GE achieves this by taking the acquisition function out of the system

hardware. Acquisition is performed within the patient cable. Since the

patient acquisition module is small and close to the patient’s body, it can

track the common mode signal of the patient. There is almost no voltage

difference between them. This, in addition to the use of the right leg

electrode, results in almost no common mode signal relative to the

patient acquisition module.

Finally, as a result of the digitization of the ECG signal within the

patient acquisition module, analog cable noise is eliminated. Thus,

regardless of cable movement or length, GE systems receive the cleanest

data possible for processing.

Revision C Physician’s Guide 15

2051167-002

3 Conditioning the ECG

Signal

The Importance of Noise-Free ECGs

A fundamental requirement of any system designed for exercise

electrocardiography is the ability to present noise-free ECGs without

distortion of the waveforms. In the absence of aggressive filtering, this

presents a formidable challenge as measurements of the smallest ECG

wave (the ST segment) in minute increments (tenths of millimeters)

under trying circumstances (strenuous exercise) are attempted.

The diagnostic value of stress testing is enhanced by computer-assisted

systems. GE systems aim at producing accurate measurements through

an improved signal-to-noise ratio in the ECG recorded during exercise.

Signal-to-noise improvements are achieved by the application of filters

and/or signal processing algorithms.

16 Physician’s Guide Revision C

2051167-002

Conditioning the ECG Signal: Filtering

Filtering

Filters, in the traditional sense, are mechanisms for removing certain

frequencies of an analog signal. Low-pass filters permit the passage of

frequencies below a specified value to be reflected in the waveform.

Conversely, high-pass filters permit only those frequencies above the

specified value to be included in the process.

The American Heart Association requires diagnostic ECG

instrumentation to be capable of recording waveforms with a fidelity of

0.05 to 150 Hz, or more challengingly, instead of 150 Hz, it must be

capable of processing triangle pulses of 1.5 mV (15 mm) amplitude and

20 ms width, according to ANSI/AAMI standard EC11

. GE stress

systems exceed the American Heart Association specifications. The

plurality of GE’s stress systems, such as CASE, have a non aggressive

low-frequency response of 0.01 Hz, in order to prevent artifactual ST

segment abnormalities

5,8

Noise Filters

The QRS complex represents the high-frequency component of the

electrocardiogram. Low-pass filters of 20, 40, and 100 Hz are user

selectable on GE stress systems for the purpose of attenuating muscle

noise from the baseline. Figure 3.1 illustrates the effects of successively

more aggressive low-pass filters on the ECG signal. These filters

significantly reduce muscle artifact and have essentially no effect on the

low-frequency components of the ST segment. R-wave amplitude,

however, will be attenuated with the 20-Hz and 40-Hz filters.

Figure 3.1 The effects of successively more aggressive low-pass filters on the ECG signal

(H. Blackburn (ed), Measurement in Exercise Electrocardiography: The Ernst

Simonson Conference. Charles C. Thomas Publisher, Springfield 1969

Froelicher V.F, Exercise and the Heart: 3

edition

)

100Hz

60 Hz

50Hz

40 Hz

30 Hz

20 Hz

10 Hz

5 Hz

1 Hz

Revision C Physician’s Guide 17

2051167-002

Conditioning the ECG Signal: Filtering

Baseline Roll Filters

Filters at the low end of the frequency spectrum of electrocardiographs

are called baseline roll filters. Those as aggressive as 0.25 Hz or even

0.5 Hz effectively reduce baseline roll due to respiration. However, some

potentially introduce a phase shift in the QRS complex resulting in

artificial changes in the region of the ST segment. Filters such as these

should never be employed in instruments designed to reproduce

physiologic ST segment changes. GE systems exceed AHA

recommendations by employing a non-aggressive high-pass filter at the

low end of the frequency range. Instead, baseline roll is controlled with

the implementation of a baseline correction algorithm. (See “Cubic

Spline Baseline Correction” on page 18 and “FRF (Finite impulse

response Residual Filtering) Algorithm” on page 19.)

GE’s incrementally updated median QRS complexes are presented at the

full diagnostic frequency response. Regardless of the noise and baseline

roll filters chosen for the ECG data, ST-segment and R-wave amplitude

measurements are made on unfiltered data. This provides the highest

level of accuracy possible with the greatest reproducibility.

50/60-Hz Line Filters

Despite all of the methods used to reject common mode signals, power

line interference (often referred to as 50/60 Hz “buzz”) will continue to be

part of the acquired signal. This is due to magnetic field induction of

differential signals in loops formed by a lead connection to the body.

Since digitization takes place at the patient, lead wire length is very

short, minimizing these signals. Nevertheless, GE uses a line frequency

filter, removing any remaining 50/60-Hz buzz.

This filter must know the line frequency, 50 or 60 Hz, at which the

system operates. The appropriate filter is preset at the factory prior to

shipment. Your system removes the line frequency noise by monitoring

for, locking in on, and subtracting the opposite sinusoidal wave. This

filter has a dramatic effect. Notice its operation on the following signal

below.

Figure 3.2 The elimination of AC line frequency interference with application of a 50/60-Hz

line filter.

18 Physician’s Guide Revision C

2051167-002

Conditioning the ECG Signal: Cubic Spline Baseline Correction

Cubic Spline Baseline Correction

GE’s stress systems feature a cubic spline algorithm for removing

baseline roll. This is employed as an alternative to aggressive baseline

filtering. The technical details of this method are described in

“Electrocardiogram Baseline Noise Estimation and Removal Using

Cubic Splines and State-Space Computation Techniques”, Computers

and Biomedical Research

The cubic spline algorithm requires that three consecutive isoelectric

points be reliably detected. The onset of the QRS is ideal for this

purpose. At low heart rates, a point in the isoelectric area between the T

and P waves is used in addition. The three consecutive points are used to

establish and estimate the baseline roll. This estimate is then subtracted

from the original ECG to yield a baseline-roll-reduced ECG with no

waveform phase distortion. Figure 3.3 demonstrates the capability of the

cubic spline.

Figure 3.3 Application of the cubic spline to an ECG signal that exhibits respiratory induced

baseline roll.

As three isoelectric points are required for the cubic polynomial to

operate, the ECG is delayed 2 seconds, for example rhythm reports

recorded with the spline engaged are delayed by these 2 seconds.

3 seconds

Revision C Physician’s Guide 19

2051167-002

Conditioning the ECG Signal: FRF (Finite impulse response Residual Filtering) Algorithm

FRF (Finite impulse response Residual Filtering)

Algorithm

Most GE stress systems feature the FRF algorithm for removing noise

and baseline roll. This is employed as an alternative to aggressive noise

and baseline filtering. Also it is an alternative to the Cubic Spline

Baseline Correction which only reduces the baseline roll. The method

was published in “Artifact Processing during Exercise Testing”, Journal

of Electrocardiology

The FRF algorithm reduces the artifacts in the ECG stream, but with

much less distortion of the QRS complexes. It consists of a block that

updates the median beat and a function that subtracts the median beat

from the ECG and then outputs a residual signal. The residual signal is

fed into a low-pass filter, a high-pass filter, and finally into a function

that adds the median beat (see Figure 3.4).

Figure 3.4 Block diagram of the FRF algorithm

The median beat is continuously updated only if the current QRS

complex correlates with the median beat. The chosen correlation limits

guarantee continuous updating. The subtraction function subtracts the

median beat only if there is a reasonable accordance between median

beat and current beat. If the current beat is a PVC, for example, no

subtraction is done. The median beat is subtracted from the QRS onset

to the T end. The P wave is not subtracted. In cases of PSVCs, atrial

fibrillation, atrial flutter, AV block II (Wenckebach, Mobitz), and AV

block III, for example, a subtraction of the P wave of the median beat

would be erroneous.

The result of the subtraction function is the residual signal. This signal

is filtered by a low-pass filter to reduce muscle noise and a high-pass

filter to reduce baseline wander. The cutoff frequencies of the filters are

set to values that avoid unacceptable distortion of the remaining P

waves and PVCs in the residual signal. The constant delay of the filters

enables the addition function to add the median beat to the filtered

residual signal at the exact position, i.e., the position where it was

subtracted before. The addition function does not add the P waves. If a

median beat is not available or in case of a pace ECG, the filters are

switched off.

20 Physician’s Guide Revision C

2051167-002

Conditioning the ECG Signal: FRF (Finite impulse response Residual Filtering) Algorithm

For your notes

Revision C Physician’s Guide 21

2051167-002

4 Signal Processing

Mathematical Algorithms

Signal processing employs mathematical algorithms to

 effectively detect beats,

 classify their morphology, and

 properly superimpose normally conducted beats.

The most prevalent method of signal processing takes an average of a

defined time window or block of consecutive beats. This process, termed

arithmetic averaging, produces a mean QRS complex.

GE uses a unique noise-rejection algorithm called incremental updating.

This mechanism, sometimes called incremental averaging, produces a

median QRS complex. The technical aspects of this program were

described in Trends in Computer-Processed ECG’s, North Holland Press,

1977.

In a perfect world where stress ECGs would not be affected by random

noise – muscle artifact in particular – one might be hard-pressed to label

a single approach to signal processing as superior. However, under

actual conditions where noise is a fact of life, the median QRS complex

produced by the incremental update process steps to the front.

Signal processing consists of five consecutively applied algorithms:

 First, the best leads for QRS detection must be selected.

 Second, the incoming beats must be detected, and heart rate updated.

 Third, the incoming beats are classified on the basis of RR intervals

and QRS morphology.

 Fourth, a correlation and alignment algorithm overlays the

incoming beat to the dominant template for best fit and determines

the degree of agreement between the incoming beat and the

template.

 Finally, successfully correlated beats incrementally update the

dominant template.

During exercise testing and especially in the high exercise phase, it

occurs very often that only a few of the applied leads deliver a

reasonable ECG quality.

It is obvious that by selecting only good leads, the QRS complex

detection quality can be improved. Instead of using an algorithm, which

combines leads without taking account of their quality, GE developed a

new algorithm, the Intelligent Lead Switch algorithm.

22 Physician’s Guide Revision C

2051167-002

Signal Processing: Intelligent Lead Switch Algorithm

Intelligent Lead Switch Algorithm

The advantage of the algorithm is a more reliable QRS detection, even in

the high exercise phases. The method was published in “Artifact

Processing during Exercise Testing”, Journal of Electrocardiology

Initially, after application of all relevant electrodes, the algorithm

selects the two best leads. During exercise testing, the selection process

is restarted when the best leads become disconnected or the signal

quality of the selected best leads becomes insufficient.

During exercise testing the algorithm switches between the two best

leads, always selecting the artifact-free lead. In the presence of special

arrhythmias the algorithm also switches to the lead that is better suited

for QRS complex detection, e.g., in the case of a bigeminy with very big

premature ventricular complexes and small normal complexes or

ventricular tachycardia with very small complexes.

During exercise testing the algorithm looks at the arrhythmia results of

the alternative lead and corrects the result of the current lead if

necessary. This would be the case when a premature ventricular

complex in the current lead is very small (then the algorithm recognizes

a pause), but displayed more clearly in the alternative lead.

The algorithm consists of:

 up to 15 independent and equivalent units for QRS detection, ECG

quality evaluation, and event classification

 a logical unit for starting/restarting the selection process of the two

best leads, for selecting the results of the best channel, correcting

the event classification of the best channel, and correcting the

trigger points (times where the QRS complexes are located) of the

best channel

The ECG quality is calculated on the basis of the QRS complex

amplitudes, the levels of middle and high frequency noise, and the

electrode status (e.g., connected or disconnected electrodes). Examples

for classified events are pauses, premature supraventricular complexes

and premature ventricular complexes. If a pause is detected in the best

channel, and the algorithm finds premature ventricular complexes in

other channels, it will correct the event classification and the trigger

points of the best channel.

Revision C Physician’s Guide 23

2051167-002

Signal Processing: Detection of Cyclic Artifact Algorithm

Detection of Cyclic Artifact Algorithm

During an exercise test the patient normally walks or runs on a

treadmill or rides a bicycle. In both cases the patient produces cyclic

artifacts. With increasing exercise, the artifacts increase as well.

The origin of those artifacts is muscular activity or changes in electrode

position, caused by the movement of the patient. Electrode position

changes produce artifacts whose frequency content is very often similar

to the QRS complexes. For this reason, the artifacts are very difficult to

detect. Furthermore, they are dangerous because they disturb the

detection of QRS complexes. This can lead to wrong heart rate values, to

wrong arrhythmia results, and to erroneous interpretation of the

exercise test.

Once a cyclic artifact rhythm is detected the information is used as an

input for the Intelligent Lead Switch Algorithm for selection of other

ECG channels with reduced artifact levels, for example, to ECGI and

ECGV6 in Figure 4.1. The method was published in “Novel Signal

Processing Methods for Exercise ECG”, International Journal of

Bioelectromagnetism

ECG and the cyclic artifacts are two independent rhythms. During

exercise both, the RR intervals and the intervals of the cyclic artifacts,

do not vary very much over a short time range. To identify and separate

the two independent rhythms an algorithm is used which tries to find

independent chains of intervals. Cyclic artifacts are detected when the

algorithm has found another independent chain in addition to the chain

of RR intervals. Leads containing two independent chains are of poor

quality.

Figure 4.1 ECG with cyclic artifacts. In leads ECGV2 to V5 both rhythms are visible, in

ECG leads I and V6 only the ECG rhythm is visible, and in ECG lead V1 only the

cyclic artifacts are visible.

24 Physician’s Guide Revision C

2051167-002

Signal Processing: QRS Detection

QRS Detection

In GE stress systems the QRS complexes are detected independently in

different leads. Then the information about the detected QRS complexes

of every lead is combined in a logical way depending on artifacts,

amplitudes, etc. (see preceding “Intelligent Lead Switch Algorithm” on

page 22). The combined information results in an accurate and reliable

QRS detection, even in the high exercise phase.

Detecting QRS complexes in each individual lead has the advantage of

not loosing any information about the QRS complex, T wave, and P

wave. Since the morphology of the ECG waves is not destroyed, the

discrimination of QRS complexes and artifacts is very high.

QRS detection algorithms used earlier and by the competition transform

the ECG leads and combine them to one signal for QRS detection. All

these transformations eliminate important information with the effect

that QRS detection will be more disturbed by artifacts or will be less

sensitive.

The new QRS detection algorithm adapts to the current QRS complexes,

to T waves, to P waves, and to the overshoots or undershoots of a

pacemaker ECG. The knowledge about the T waves and P waves is used

to avoid the misclassification of a T wave, or a P wave as a QRS complex,

for instance, and provides a higher tolerance to artifacts.

Revision C Physician’s Guide 25

2051167-002

Signal Processing: Correlation and Alignment

Correlation and Alignment

After QRS detection, the correlation and alignment algorithms form the

basis for the incremental update. The process begins by selecting an

incoming beat. This beat becomes the seed for the dominant median

template. The following beat is superimposed on the dominant template.

The beat is aligned horizontally and vertically until the maximum

possible correlation coefficient is attained. If the correlation coefficient is

considered to be insufficient, the incoming beat is classified as an ectopic

beat and is not considered for further processing. If the correlation

coefficient is high, the beat is used to update the dominant median

template. If the process begins by selecting an ectopic beat, the ensuing

normal beats will not correlate. In this case further complexes are

collected. Then the resulting ECG pattern is analyzed for selection of the

correct (normal) beats for the dominant template. This will occur when

ventricular ectopic beats, e.g., bigeminy and trigeminy, are found in the

ECG.

The template is held in a buffer of 1200 milliseconds. The QRS trigger

point is arbitrarily placed at 400 milliseconds into this buffer (see

Figure 4.2). This large buffer ensures that QT and PQ intervals, even

when they are extremely long, can be stored and processed in their

entirety.

Figure 4.2 Placement of the median template into a 1200-millisecond buffer

Figure 4.3 Incoming beat superimposed on the median template and shifted for maximum

alignment

1200 ms

buffer

400 ms

QRS detection

trigger

Incoming beat

Template

26 Physician’s Guide Revision C

2051167-002

Signal Processing: Incremental Updating

Incremental Updating

Whenever the incoming beat correlates with the median template, the

template is updated (positively or negatively) by a fixed increment or a

fraction of the difference between the median template and the incoming

beat, whichever is less. With this process, the median morphology tends

in the direction of consistent, non-random changes in amplitude, for

example, ST segment depression. The process assumes that noise is

random, and the erratic amplitude changes associated with noise are

limited to immeasurably small changes in the median beat.

Relearn

If the QRS morphology changes during stress testing in a normal way,

the median update algorithm will be able to follow these changes.

Occasionally patients present significant changes in QRS morphology.

GE stress systems recognize these changes and “relearn” the median

template. During this process, the ST segment measurement points are

also relearned. An automatic relearn also occurs after reapplying

disconnected electrodes. A manual relearn can be initiated by the user.

When relearning, an incoming beat is established as the dominant

template. If the next beats do not correlate with the first beat, the

relearning procedure starts again. In case of ventricular ectopia, e.g.,

bigeminy and trigeminy, the algorithm is able to select the normal beats

for creating the new dominant median beat.

During the relearn phase, a question mark (?) appears after the ST

measurements of the median complex on the screen. Also, any reports

run during this period will display question marks after the ST

measurements. This indicates that the values may be unreliable during

the relearn phase.

As the system monitors each beat for proper shape discrimination, a

record of the last beats is constantly updated to determine the number of

beats successfully updated. If 20 of the last preceding beats fail to align,

the warning “Median Update Ceased” will appear on the screen, and an

automatic relearn occurs. Prolonged, excessive artifacts may cause this

message to appear. The median beat and the ST measurements remain

on the screen. If within a one-minute period a new median beat cannot

be created, the median beat and the measurements are deleted, and the

automatic relearn function is repeated . The manual relearn may be

used at this time if visual inspection of the raw rhythm shows

discordance with the median.

Revision C Physician’s Guide 27

2051167-002

Signal Processing: Pace Enhance

Pace Enhance

Displaying pacemaker pulses is sometimes difficult. In some cases the

pacemaker pulses are narrow and cannot be displayed; in other cases

they are large and disturb the readability of the ECG. The pace enhance

function solves these problems.

When enabled, the pace enhance function will replace the pacemaker

pulse with a marker of the same polarity as the ECG signal in each lead.

If the pacemaker pulse is small (in the range of ±0.1 mV) a positive

marker is added to the pacemaker pulse.

This is what the pace enhance function does in detail:

It adds a marker (1.5 mV amplitude, 6 ms duration) of the same polarity

as the pacemaker pulse to the ECG electrode signal.

It limits the added sum of pacer pulse and marker to 0.5 mV in the ECG

lead signal.

28 Physician’s Guide Revision C

2051167-002

Signal Processing: Pace Enhance

For your notes

Revision C Physician’s Guide 29

2051167-002

5 Measurement Values

Heart Rate

The heart rate is calculated as a 16-beat sliding average and the display

is updated every second. This provides stability to the rate during sinus

arrhythmia and atrial fibrillation, yet allows rapid tracking of dramatic

rate changes as in paroxysmal tachycardia.

In the initial learn mode, the heart rate is calculated by averaging the

RR intervals of the available beats. The heart rate is displayed when at

least four beats are detected.

Evaluation of the accuracy of the heart rate is based on the results from

annotated databases consisting of exercise ECGs from treadmill tests

and exercise ECGs from bicycle tests.

Table 5.1 Heart rate performance (DOC0996283)

Note

On GE systems, the leads used for the QRS detection are selected

automatically, but they are also user selectable before and during

the stress test. Consult your operator’s manual for details.

The heart rate HR is one of the most important outcomes of an exercise

test. The HR course during exercise contains valuable information about

coronary stenosis

64,79

and cardiovascular mortality risk

33,39,41

. But it is an

indirect method and, therefore, other factors, such as the exercise device

used—treadmill or bicycle—also influence the HR values. Most patients

achieve higher peak HR values on treadmills than on bicycles

Database average mean error average RMS value

Bicycle ECGs - 0.14 bpm 0.76 bpm

Treadmill ECGs - 0.2 bpm 0.99 bpm

30 Physician’s Guide Revision C

2051167-002

Measurement Values: ST Segment Values

Figure 5.1 Mean values of "HR at peak exercise" of different exercise devices and

pharmacological Adenosine stress test, based on the GE-Healthcare

exercise/stress test database, containing more than 20.000 patients.

ST Segment Values

GE’s methodology involves locating an isoelectric reference (E point) and

the QRS offset (J point), and measuring the ST segment at a user-

defined or heart-rate related distance x past the J point. (See Figure 5.2)

It is identical in approach to the manner with which one would manually

measure the values.

Figure 5.2 GE stress systems measure ST segments by locating an isoelectric reference

(E point), the offset of the QRS complex (J point) and the post-J measurement

point (J+x point) for measuring the ST segment.

E point

J point

Revision C Physician’s Guide 31

2051167-002

Measurement Values: QRS Onset and Offset Detection

QRS Onset and Offset Detection

QRS onset precedes the first deflection in the used leads, and QRS offset

occurs where the last steep slopes of depolarization are replaced by the

more or less flat ST segment. The method employed by GE stress

systems uses up to 16 leads, determining the earliest onset and latest

offset in the leads. The isoelectric point (E point) is placed at

10 milliseconds prior to the earliest QRS onset. This avoids the Q and R

waves, yet stays well off the P wave even with decreasing or initially

short P-R intervals.

For stress testing two methods for calculation of the E and J points are

user selectable. The first method calculates the E and J points before the

actual exercise, i.e., in the pre-exercise phase (single calculation). Then,

during exercise, the median beat is aligned between these two points.

The second method continuously recalculates the E and J points with

every incoming beat (continuous calculation). The reasons for using the

first method are historical. The second method, which is the factory

default method, is the correct one, because it is able to adapt the E and J

points when the QRS width changes. In the majority of cases, however,

the QRS width will not change significantly during stress testing.

Waveform measurement is relatively straightforward once QRS onset

and offset are known. All measurements are taken from the median

complexes, providing immunity from noise while at the full diagnostic

frequency.

Note

Both the E and J points are user adjustable on most GE stress

systems. Consult your operator’s manual for details.

ST Level

ST levels are simply the amplitude difference between the isoelectric

reference point (E) and the post-J measurement point (J+x point). The

operator has the possibility to enter a fixed value, e.g., 60 milliseconds

for the distance x, or to select a method for heart-rate adaptive

calculation of the distance x. Two methods for heart-rate-adjusted

calculation are available:

 x = RR/16

RR = RR interval. The formula is also used in GE’s 12SL resting

program

 x = 3/16 (656/(1 + 0.001 × HR) – 90 ms)

HR = heart rate. The part of the formula in italics is derived from

P.M.Rautaharju et al., Estimation of qt prolongation, Journal of

Electrocardiology, 23:111-117, April 1990

Instead of using a heart rate adjusted method, the ST level

measurements may be adjusted by the user at any time. For example, an

ST measurement point of 80 milliseconds post J may be manually swept

forward as heart rates increase.

32 Physician’s Guide Revision C

2051167-002

Measurement Values: ST Slope

ST Slope

An ST slope is measured over an interval established as 1/8 of the

average RR interval to a maximum of 80 milliseconds. The interval

starts at the J point. This definition of slope takes into account the

correlation between heart rate and repolarization time. Effects of the

T wave on ST slope measurement are avoided.

In an algebraic sense, the slope of a given line is defined as the rise

divided by the run. Figure 5.3 illustrates the algebraic determination of

a slope on an ECG with ST-segment depression. Contrary to visual

examination, the slope would be reported as positive.

Figure 5.3 The algebraic determination of this ST slope would report it as positive despite

the presence of 2 millimeters of ST-segment depression

GE systems employ a least squares fit in order to provide a close match

with visual impressions. The actual slope measurement in millivolts per

second is the result of a least squares fit of a straight line in

40-millisecond subintervals in the ST interval from J to 1/8 of the

average RR interval. (See Figure 5.4) The most negative slope is

reported. This provides a close match with visual impressions.

Figure 5.4 By employing a least squares fit for ST slope determination, GE provides a

better match with the visual impression

80ms post JJ point

Baseline

2 mm ST

depression

1/8 RR interval

(up to 80 ms)

Revision C Physician’s Guide 33

2051167-002

Measurement Values: ST Integral

ST Integral

An ST integral is measured over an interval from the QRS offset to the

point where the ST segment crosses the isoelectric line, or to the point

corresponding to 7/16 of the total ST-T duration, whichever occurs first.

An ST integral below the isoelectric line greater than 10 µVs is

considered abnormal. Note: One square mm on electrocardiographic

paper at standard speed (25 mm/s) and calibration (10 mm/mV) is

4µVs.

ST Index

An ST index is "1" when the ST depression is 1.0 mm (0.1 mV) or

greater, or the sum of ST segment depression in mm plus the ST down

slope in mV/s is 1.0 or greater. If the conditions above are not fulfilled,

then the ST index is "0". The ST depression is measured in the post J

measurement point (J+x point).

R-Wave Amplitude

The R-wave amplitude is the difference in amplitude between QRS onset

and the first maximum positive value.

QRS Width

The QRS width is the difference between QRS offset (J point) and QRS

onset. During exercise testing, the QRS width is calculated continuously,

but two values are the most important ones: QRS width at the beginning

of the exercise phase and QRS width at peak exercise. These values are

used to detect exercise-induced bundle branch blocks.

Figure 5.5 Exercise-induced LBBB, induced in the exercise phase, lead V6

Exercise-induced bundle branch blocks are rare, but they have the

tendency to become permanent bundle branch blocks.

34 Physician’s Guide Revision C

2051167-002

Measurement Values: ST/HR Slope

Continuous calculation of QRS onset and offset is the precondition for

detection of exercise-induced bundle branch blocks. Please make sure

that the settings are correct (example: see Figure 5.6).

Figure 5.6 “Calculation (E, J point)” setting in CASE/CardioSoft/CS for proper detection of

the exercise-induced bundle branch block. J point is the QRS offset, E point is

calculated from the QRS onset by subtracting 10ms.

ST/HR Slope

The ST/HR slope was originally developed by Dr. R. J. Linden at the

University of Leeds, United Kingdom, in the late 1970s

. It is reported

to yield more accurate electrocardiographic determination of the

presence and severity of coronary heart disease. While originally applied

only to bicycle ergometry testing, it has recently been adapted for

treadmill exercise by Dr. Paul Kligfield and colleagues at Cornell

University.

15,17,18,19,20,21,22,23

The criteria are based on the near parallel increase of myocardial oxygen

demand and heart rate with increasing effort. Essentially, the exercise

electrocardiogram is evaluated by linear regression analysis of the rate-

related change in ST depression as measured at 60 milliseconds post-J

point. Originally, the 12 classical leads and bipolar lead CM5 were

evaluated. Changes in leads aVR, aVL, and V1 have been found to be

poorly specific and are currently ignored. The addition of lead CM5

improves sensitivity appreciably.

A plot of ST segment depression versus heart rate is drawn for all

measured leads. Linear regression analysis is used to best fit a line

beginning at peak exercise and extending backward through at least

three points until significance is obtained (p < 0.05). The points are

taken from the ST level and heart rate at the end of each 2-minute

exercise stage. The slope of this line is then determined and presented in

units of microvolts/beats per minute. The steepest slope of all the leads

is reported and graphed, if the ST segment depression is ≥50 µV.

The large increments in heart rate between stages of the Bruce protocol

were found to yield an inadequate number of points for proper slope

evaluation. A modification of the Bruce protocol (half work loads in

2-minute exercise stages) more closely approximate the roughly 10 beats

per minute increments found in ergometry testing. This protocol,

developed at Cornell University by Okin, et al.

, is reproduced here and

is strongly recommended for the highest predictive accuracy.

Revision C Physician’s Guide 35

2051167-002

Measurement Values: ST/HR Index

Table 5.2 Cornell Treadmill Protocol for ST/HR Slope

Former work

established the cut point for normalcy at 2.4µV/bpm.

Patients with three-vessel disease had slopes above 6.0µV/bpm.

ST/HR Index

The ST/HR index was developed by Dr. Paul Kligfield and colleagues at

Cornell University. The ST/HR index is an approximation to the ST/HR

slope and therefore comparable results can be achieved

Compared with standard test criteria, simple heart rate (HR)

adjustment of ST depression during exercise electrocardiography can

improve the identification and assessment of underlying coronary artery

disease. Since heart rate during exercise drives progressive ST segment

depression in the presence of coronary obstruction that limits flow

reserve, the ST/HR index controls for the increasing metabolic severity

of ischemia that accompanies exercise. Improvement of exercise test

sensitivity with the ST/HR index results from reclassification of

otherwise “equivocal” and even “negative” test responses, including

increased identification of one and two-vessel disease in men and in

women. In addition, in population studies of low and moderate risk

subjects, the ST/HR index can increase the prognostic value of the

exercise electrocardiogram for prediction of cardiac risk and mortality

In contradiction to ST/HR slope, ST/HR index is much more simpler, is

not restricted to special protocols, does not need a significant regression

line, and always provides a value. It is simply calculated by dividing the

change of the ST depression from the baseline value (exercise start) to

maximum exercise by the change in heart rate over the same time

period. The leads are scanned for the greatest ST/HR index.

Duration of Exercise Stage (min) Speed (mph) Grade (%)

21.70

21.75

21.710

22.111

22.512

23.013

23.414

23.815

24.216

24.617

25.018

25.519

36 Physician’s Guide Revision C

2051167-002

Measurement Values: Recovery ST Level

Figure 5.7 ST/HR index and ST/HR slope

ST/HR index is an approximation of the ST/HR slope

Recovery ST Level

The measurements above do not include the recovery phase. In addition

to the exercise phase, the recovery phase delivers valuable information.

The ST level 0.5 mm at 3 minutes of recovery, for example, is more

reliable and precise than the ST level of 1mm at peak exercise (see

Figure 5.9).

Revision C Physician’s Guide 37

2051167-002

Measurement Values: ST/HR Hysteresis

ST/HR Hysteresis

The ST/HR hysteresis was developed by Dr. Rami Lethinen et al.

Tampere University, Finland, as an extension of the ST/HR loop

. It is a

highly powerful measurement for identification of coronary artery

disease (CAD)

37,90

and prediction of acute myocardial infarction

. It is

comprehensive, because it takes both, the exercise and the recovery

phase in account. Precondition for a proper ST/HR hysteresis

measurement is an adequate recovery phase. Proposed are 3 minutes.

ST/HR hysteresis increases the sensitivity and specificity in both men

and women, and is a “more competent method in CAD detection in

women than ST-segment depression and ST/HR index”

ST/HR hysteresis is calculated by integrating the difference in the ST

segment depression between the exercise and recovery phase over the

HR (heart rate) for up to 3 min of the recovery phase (see Figure 5.8).

After integration, the integral is divided by the HR difference (peak

exercise HR – minimum HR during recovery)

37,40,41

Figure 5.8 Significant ST/HR hysteresis. The area between the lower curve (exercise

phase) and the upper curve (recovery phase) divided by the difference of the

peak exercise HR and the minimum HR during 3 minutes recovery is the ST/HR

hysteresis.

Please note: Negative level values, or ST depression values, are above the zero

line. A negative level expressed as a value above the zero line indicates a

positive exercise test.

ST/HR hysteresis is an advanced ST depression measurement, because

it is the average difference of ST depression between recovery and

exercise phase. Consequently the dimension of ST/HR hysteresis is mV

(millivolt) or mm (millimeter).

38 Physician’s Guide Revision C

2051167-002

Measurement Values: ST/HR Hysteresis

The following Figure 5.9 shows the ischemia detection quality

differences of different methods. At a threshold of 0.01 mV, ST/HR

hysteresis achieves a sensitivity of 80% and a specificity of more than

80%. The ST depression at peak exercise at the threshold of 0.1 mV

(1mm) achieves only a sensitivity of 55% with the same specificity.

Figure 5.9 Receiver-operating characteristic curves for the continuous diagnostic variables

and the operating point for the dichotomous heart rate (HR) recovery loop. The

curve symbols refer to the partition values of the variables. Some of these

values are specified, expressed in millivolts for the ST/HR hysteresis and ST

depressions, and in mV/beat/min for the ST/HR index. AUC = area under the

receiver-operating characteristic curve

Sometimes ST/HR hysteresis curves are disturbed. This occurs in

patients with pacemaker, atrial fibrillation or exercise induced bundle

branch blocks. Responsible for this phenomenon are abrupt changes in

the ST segment and/or discontinuous changes in the heart rate.

Excessive noise can also disturb the curve.

Revision C Physician’s Guide 39

2051167-002

Measurement Values: Arrhythmia Detection

Arrhythmia Detection

Arrhythmia analysis distinguishes between single-beat events and

multi-beat events. Single-beat events are classified by linking

characteristics such as the current RR interval, the current QRS

morphology, the dominant RR interval, and the dominant QRS

morphology with a rule database. An example for a rule type leading to

the PVC classification is

the current RR interval is shortened

and

the next RR interval is a compensatory pause

and

the morphology is different from the dominant QRS morphology

then

the current beat is a PVC.

Multi-beat events result from multiple, consecutive single-beat events.

Two consecutive PVCs, for instance, are a couplet, an alternating

sequence of PVC and normal beat is classified as a ventricular bigeminy

(see Table 5.3).

In addition, three consecutive 4-second sliding windows are checked for

ventricular fibrillation/flutter, rapid ventricular tachycardia and

asystole.

40 Physician’s Guide Revision C

2051167-002

Measurement Values: Arrhythmia Detection

Table 5.3 Arrhythmic events with their acronyms. All events, except artifact and learn

phase, which are not true events, appear in the event window. An event involving

complex ectopy is highlighted. On CASE systems and CardioSoft/CS, for

example, these events are shown in red

When the arrhythmia analysis is started, the system needs

approximately 5seconds of ECG data to establish the thresholds for

QRS detection. The arrhythmia analysis program will analyze two leads

of ECG. These two leads are either selected manually by the user or they

are automatically determined by the program. In the latter case, the

program scans all available leads to find the two best leads. This takes

place at the same time as the determination of the QRS detection

thresholds and prolongs the initial process by 1 second. Then, after

10 beats have been collected, the dominant beat (learned QRS complex)

will be determined. On the full-disclosure ECG, each of these

10 complexes is labeled with an “L”. The first beat after the end of the

learn phase that matches the morphology of the dominant beat, is

displayed in the event window and labeled “QRSL”.

The GE arrhythmia analysis algorithm accepts noisy ECGs, allowing for

almost uninterrupted monitoring and for a high sensitivity in

identifying events. Extremely noisy beats or artifactual beats, however,

are rejected or not accepted for event classification. On the full-

disclosure ECG, these beats are labeled with an “A”. Ventricular

bigeminy is identified as an alternating sequence of normal beats and

PVCs. At least two PVCs must be detected. A ventricular escape beat is

identified when the current RR interval is prolonged and the morphology

differs from that of the dominant beat.

learned QRS complex QRSL

* asystole ASYSTO

* ventricular fibrillation/flutter VFIB

* ventricular tachycardia VTAC

* ventricular run (> 2 PVCs) RUN

* ventricular couplet (2 PVCs) CPLT

atrial fibrillation AFIB

pause of 2 missed beats PAU2

pause of 1 missed beat PAU1

ventricular bigeminy VBIG

paroxysmal supraventricular tachycardia PSVT

pacer error PERR

ventricular escape beat ESC

premature ventricular complex PVC

premature supraventricular complex PSVC

pacemaker capture PCAP

artifact A

learn phase L

* event of complex ectopy

Revision C Physician’s Guide 41

2051167-002

Measurement Values: Arrhythmia Detection

For pacemaker patients (this information should be entered on the

patient demographics screen) the arrhythmia analysis algorithm

analyzes the temporal relationship of pacer pulse and QRS complex.

“Pacemaker capture” is detected when the QRS complex occurs within

300 ms of the pacer pulse. If the QRS complex occurs later than 300 ms

after the pacer pulse or not at all, the event is labeled as “pacer error”.

Atrial fibrillation is detected on the basis of irregular RR intervals and

absence of P-waves. Ventricular fibrillation/flutter is identified by

frequency analysis in consecutive 4-second windows.

Detection of ventricular fibrillation/flutter or rapid ventricular

tachycardia is based on three consecutive 4-second sliding windows. If

the criteria for ventricular fibrillation/flutter or rapid ventricular

tachycardia are fulfilled in at least two of the windows, VFIB or VTAC is

classified. Asystole is detected in the last 4-second sliding window but

only if no fine ventricular fibrillation is detected.

Arrhythmia detection cannot be disabled by the user. Continuous

detection is necessary to prevent abnormals from entering into the

update process. If arrhythmia detection were disabled and enabled

during the stress test, it would be impossible for the system to provide a

complete and correct documentation of arrhythmia events.

During the stress test, the event window shows the most recent event

with its event label, i.e., the new event overwrites the previous one.

However, if events belong to the same category, the first of these events

remains displayed. This prevents the screen display from becoming

unsteady. All events that can be displayed are listed in Table 5.2.

Artifactual QRS complexes (artifact “A”) and QRS complexes from the

learn phase (“L”) do not appear in the event window.

Events with complex ectopy (see Table 5.3) such as a ventricular run,

will be highlighted in the event window.

The event window continuously displays the ventricular ectopics (VE)

per minute. This value is calculated as the sum of all PVCs and ESCs

detected in the past 60-second interval, including the PVCs of

ventricular tachycardias, runs, couplets, and bigeminy. A new PVC or

ESC will thus immediately update the VE/min (ventricular ectopic per

minute) value.

During the stress test, the system can save a full-disclosure ECG for

review of the arrhythmias at the end of the test procedure. The full-

disclosure ECG is sampled at a lower rate sufficient for evaluation of

arrhythmias. For more information, please consult your operator’s

manual.

42 Physician’s Guide Revision C

2051167-002

Measurement Values: Arrhythmia Detection Performance

Arrhythmia Detection Performance

Evaluation of the accuracy of the algorithms is based on the results from

annotated databases. Depending on the intended use of the algorithms,

namely for exercise testing, we compiled a set of annotated databases

with more than 1000 ECGs, consisting of exercise ECGs from treadmill

tests, exercise ECGs from bicycle tests, pacemaker ECGs, the MIT-BIH

database and the AHA database

To evaluate the QRS detection and the ventricular ectopic performance

we followed the standard ANSI/AAMI EC57, Testing and reporting

performance results of arrhythmia and ST segment measurement

algorithms

Table 5.4 QRS complex detection performance. (DOC0996283)

sensitivity = true positives / (true positives + false negatives)

pos.predictivity = true positives / (true positives + false positives)

Table 5.5 Ventricular Ectopic Detection Performance (DOC0996283)

sensitivity = true positives / (true positives + false negatives)

pos.predictivity = true positives / (true positives + false positive)

false positive rate = false positives /(correct negatives + false positives)

Database sensitivity pos. predictivity

MIT-BIH 99.8% 99.8%

AHA 99.6% 99.8%

Bicycle ECGs 99.9% 99.96%

Treadmill ECGs 99.9% 99.9%

Database sensitivity pos. predictivity false positive rate

MIT-BIH 92% 94% 0.43%

AHA 94% 97% 0.29%

Bicycle ECGs 80% 93% 0.03%

Treadmill ECGs 77% 84% 0.21%

Revision C Physician’s Guide 43

2051167-002

Measurement Values: Exercise-Induced Wide QRS Tachycardia

Exercise-Induced Wide QRS Tachycardia

Exercise-induced wide QRS tachycardias are rare. They are mostly

ventricular tachycardias. An exercise-induced wide QRS tachycardia is

detected when at least 10 consecutive premature wide complexes with a

heart rate (HR) > 140 bpm occur. The tachycardia is analyzed in both,

exercise and recovery phase.

Figure 5.10 Exercise-induced ventricular tachycardia (VT) occurred in recovery, V5

44 Physician’s Guide Revision C

2051167-002

Measurement Values: Exercise-Induced Supraventricular Tachycardia

Exercise-Induced Supraventricular Tachycardia

Only the recovery phase is analyzed. The exercise phase is not analyzed,

because, from a technical point of view, discriminating between heart

rate increase due to exercise versus supraventricular tachycardia is

difficult. But according to Maurer et. al

: "tests in patients with known

or suspected heart disease; 14 of the 22 cases of exercise-induced SVT

were observed during recovery". Thus, the majority of exercise-induced

SVT can be detected during recovery.

Exercise induced SVT are possibly predictive of atrial fibrillation. "Risk

of development of lone atrial fibrillation during long-term follow-up in

subjects with exercise-induced supraventricular tachycardia"

Figure 5.11 Exercise-induced supraventricular tachycardia in recovery phase

Revision C Physician’s Guide 45

2051167-002

Measurement Values: Duke Treadmill Score

Duke Treadmill Score

The Duke Treadmill Score (DTS) is well validated and clinically useful

for risk assessment in patients with either established or suspected

coronary artery disease for whom the desirability of additional testing

must be determined

9,11,12

. One of the major goals of risk assessment is to

identify the low-risk patients for whom no additional testing is required.

Whereas multiple valid strategies for accomplishing this exists, there is

no consensus on the optimal approach. Cost-containment pressures may

indicate the use of exercise testing as the preferred initial strategy in

patients who are able to exercise and have an interpretable

electrocardiogram.

The Duke Treadmill Score is calculated:

In the formula, exercise time is in minutes, deviation is mm, and angina index is 0,1, or 2.

The Duke Treadmill Score (DTS) is a predictor for mortality

31,39,41

. It is

calculated with the exercise duration in minutes, the maximum exercise-

induced ST deviation in mm, and the exercise angina index. The angina

index has a value of 0 if the patient experienced no angina during the

exercise, 1 if the patient experienced no exercise-limiting angina, and 2

if angina was the reason the patient stopped exercising.

The ST deviation is the amount of exercise-induced ST-segment

deviation observed (the largest elevation or depression after resting

changes have been subtracted). The ST segment deviation during

exercise is the horizontal or downsloping depression or elevation

Leads aVR, aVL and V1 are excluded. If the amount is less than 1mm,

the value is corrected to 0 (ACC/AHA 2002 Guideline Update for

Exercise Testing

The score normally has a range from -25 (indicating the highest risk) to

+15 (indicating the lowest risk). For classification the score is divided

into three groups: high risk (score < 10), moderate risk (-10 to +4), and

low risk (

≥ +5). The prognosis of 5 year survival (55–99%) and the

average annual mortality (0.2–9%) is calculated on the basis of the

Nomogram of the Prognostic Relations, described in “Prognostic value of

a treadmill exercise score in outpatients with suspected coronary artery

disease”, and on the data from treadmill exercise testing according to the

Bruce protocol (or the equivalent in multiples of resting oxygen

consumption METS from an alternative protocol).

DTS = exercise duration – 5 × ST deviation

– 4 × exercise angina index

46 Physician’s Guide Revision C

2051167-002

Measurement Values: METS

METS

Exercise capacity can be reported in estimated metabolic equivalents

(METS) of exercise. The translation of exercise duration or workload into

METS (oxygen uptake expressed in multiples of basal oxygen uptake,

3.5 O

ml/kg per minute) has the advantage of providing a common

measure of performance regardless of the type of exercise test or protocol

used.

MET level or exercise duration achieved on exercise testing is an

important predictor of adverse cardiac events after myocardial

infarction. This observation appears to hold true for tests performed on

the treadmill and the bicycle ergometer. Failure to achieve 5 METS

during treadmill exercise is associated with a worse prognosis

The METS calculation for treadmill is applied according to the revised

formula presented in the American College of Sports Medicine’s

Guidelines for Exercise Prescription and Testing, 3

edition, 1986

. The

equation for Metabolic Equivalents follows:

where grade is given in percent and speed is given in miles per hour.

If speed is given in km/h the formula is modified to:

The METS calculation for bicycle ergometers is applied according to the

formula presented in the European Heart Journal in 1994

The METS equation for bicycle ergometers follows:

where load is given in watts and weight is given in kilograms.

MET levels are extrapolated between stages of exercise. One minute of a

stage must be completed to obtain full stage MET values

47,51

. At any

point thereafter, full credit is given for the stage.

Depending on age and gender, limits for poor METS, or poor exercise

capacity, can be estimated according to see Table 5.6.

METS

speed 26.8 0.1 1.8 grade 100⁄×+()×× 3.5+

3.5

---------------------------------------------------------------------------------------------------------------------------------=

METS

speed 43.1 0.1 1.8 grade 100⁄×+()×× 3.5+

3.5

---------------------------------------------------------------------------------------------------------------------------------=

METS

12.3 load× 3.5 weight×+

3.5 weight×

-----------------------------------------------------------------------=

Revision C Physician’s Guide 47

2051167-002

Measurement Values: METS

Table 5.6 Estimated METS thresholds for poor exercise capacity

The exercise capacity METS is a powerful predictor of mortality. Myers

et. al wrote: “Exercise capacity is a more powerful predictor of mortality

among men than other established risk factors for cardiovascular

disease

”. Figure 5.12 shows that in case of a high METS value,

generally accepted risk factors like hypertension, COPD, diabetes,

smoking, BMI, and cholesterol do not cause a higher number of deaths,

or, in other words, do not increase the relative risk beyond 1.0.

Figure 5.12 Relative Risks of Death from any cause among subjects with various risk factors

who achieved a functional capacity of less than 5 METS or 5 to 8 METS, as

compared with subjects whose functional capacity was more than 8 METS.

Numbers in parentheses are 95 percent confidence intervals for the relative

risks. BMI denotes body-mass index, and COPD chronic obstructive pulmonary

disease, Myers et. al.

In a treadmill versus bicycle study, Rahimi et.al.

found that METS

values from treadmill exercise tests differ from those obtained in bicycle

exercise tests. Most patients achieve higher METS values on treadmills

than on bicycle ergometers.

Age Women Men

< 29 < 7.5 < 8

30–39 < 7 <7.5

40–49 < 6 < 7

50–59 < 5 < 6

60–69 < 4.5 < 5.5

70–79 < 3.5 < 4.5

> 79 < 2.5 < 3.5

48 Physician’s Guide Revision C

2051167-002

Measurement Values: Actual Load

Actual Load

The actual load is an exercise test outcome when bicycle ergometers are

used and it is very similar to bicycle ergometer METS. METS is simply

the actual load corrected by the patient’s weight.

During a bicycle exercise test the load starts with an initial value of

50 watts, for instance, and then increases gradually by 25 watts every

2 minutes, for instance. Immediately after the change to a new stage,

the patient’s actual load is lower than the load because the patient has

to adapt to the new load.

Actual loads are extrapolated between stages of exercise

. One minute

of a stage must be completed to obtain the full target load. At any point

thereafter, full credit is given for the stage

47,51

Revision C Physician’s Guide 49

2051167-002

Measurement Values: HR Recovery

HR Recovery

The HR recovery value is the decrease in heart rate in the first minute of

recovery (see Figure 5.13). The change to the recovery phase is initiated

by a user action. To avoid a measurement error due to a delayed user

action, the program goes back up to two minutes and uses the time point

of the maximal heart rate as the beginning of the recovery phase.

HR recovery reflects vagal activity. A low value, or a low decrease,

indicates a high risk for overall mortality.

33,39,41

Figure 5.13 Heart rate decrease of more than 25 bpm in the first minute of recovery

The lower the heart rate decrease in the first minute, the higher the risk

(see Figure 5.14).

Figure 5.14 Estimates of the Relative Risk of death within six years according to Heart Rate

Recovery one minute after peak exercise. Dashed lines represent the

95 percent confidence interval.

50 Physician’s Guide Revision C

2051167-002

Measurement Values: FVE Recovery

FVE Recovery

Frequent ventricular ectopics in the recovery phase of an exercise test

(FVE recovery) is a good predictor of mortality. Frequent ventricular

ectopy during recovery after exercise is a better predictor of an increased

risk of death than ventricular ectopy occurring only during exercise.

Frequent ventricular ectopics occur with frequent ventricular premature

beats, ventricular couplets, ventricular runs, ventricular bigeminy, and

ventricular tachycardia

Figure 5.15 Example of frequent ventricular ectopics

FVE recovery values become questionable in patients having frequent

ventricular ectopics in the pretest phase. In those cases the FVE

recovery value is followed by a question mark and is excluded from

mortality prediction

Note

Ventricular couplets, ventricular runs, ventricular bigeminy and

ventricular tachycardia consist of ventricular ectopic beats.

Revision C Physician’s Guide 51

2051167-002

Measurement Values: Chronotropic Response

Chronotropic Response

A patient’s chronotropic response, or percentage of "HR reserve used", is

calculated using the heart rate at rest (HR

rest

) and at peak exercise

(HR

peak

). The following equation calculates the percentage of HR reserve

used during the exercise test.

%HR

reserve used

= 100 x (HR

peak

- HR

rest

) / (220 - age - HR

rest

)

Chronotropic incompetence is identified, when the percentage of HR

reserve used is below 80% (or 62% for patients on ß-blockers).

“The HR response to exercise is related to several parameters including

age, resting HR, functional capacity, cardiac function, extent of coronary

artery disease, and the autonomic nervous system”

. HR reserve used is

related to the same parameters; however, it has been normalized with

respect to the patient's resting heart rate as well as the expected peak

heart rate for age. Results presented in figure 5.16 demonstrate a

correlation between HR reserve used and coronary artery occlusions

greater than 50%. Others have also found that a low HR reserve used

value is associated with “carotid atherosclerosis, independent of the

established risk factors in healthy men, which could contribute to the

high incidence of cardiovascular diseases in subjects with chronotropic

incompetence.”

Still others have found it to be a predictor of acute

myocardial infarction

as well as mortality.

35,38,39,41

Thus, there is

significant evidence that HR reserve used is a good indicator of the

extent of atherosclerotic disease. The exercise interpretive program uses

HR reserve used for improving the specificity for the detection of cardiac

ischemia.

Figure 5.16 HR reserve used and the relation to coronary artery stenosis >

50%. ECG data

together with angiographic reference data from FINCAVAS

study, Tampere,

Finland, 556 patients.

52 Physician’s Guide Revision C

2051167-002

Measurement Values: Blood Pressure

Blood Pressure

Blood pressure values are received from an external blood pressure

measurement device or must be entered manually via the key pad.

A blood pressure above 250 mmHg at maximum load is considered

abnormal.

Double Product / Rate Pressure Product (RPP)

The Double Product is the product of the heart rate and the systolic

blood pressure. Therefore it is also called the Rate Pressure Product

(RPP). For normal exercise response RPP should exceed

20,000 mmHg/min

. The Double Product correlates with the myocardial

oxygen consumption (“Exercise and the Heart”, Froelicher

)

Heart Rate Values

Maximal Predicted Heart Rate — Upon entering the patient’s age,

GE stress systems automatically calculate the Maximal Predicted Heart

Rate. The system offers two calculation methods:

1. WHO: Max. Predicted HR = 220 – age in years

2. AHA: Max. Predicted HR = 160 (age < 25 years) or

115 (age > 75 years) or

160 – (age – 25) × 45/50 (between 25

and 75 years)

In the event that your laboratory makes use of different predictive

criteria, manual entry of specific values is simply completed during the

entry of demographic data.

Target Heart Rate — GE systems determine the Target Heart Rate as

a user-defined percentage of the Maximal Predicted Heart Rate. While

the default value is set at system initialization, a user may elect to alter

it on a test-to-test basis.

Dynamic ST Scan

Dynamically scan all leads for the “worst case” ST segment depression.

When ST segment depression occurs in another lead that is more severe

than the lead currently shown, the display automatically changes to that

particular lead. The Leads aVR, aVL, and V1 are excluded.

Revision C Physician’s Guide 53

2051167-002

Measurement Values: ST Criteria

ST Criteria

The user has the possibility to enable the continuous monitoring of the

ST segment. When the selected ST criteria are reached, the program

alerts the user and informs of the leads involved. ST criteria can be

defined separately for ST depression and for ST elevation.

ST depression (measured in the post-J point J+x) – A combination of

level and slope values can be entered:

Level values: off, < -0.05, -0.1, -0.15, -0.2, -0.3 mV

Slope values: off, horizontal and down sloping (< 0.5 mV/s),

down sloping (< -1mV/s)

Leads aVR, aVL, and V1 are excluded from monitoring

ST elevation (measured at point J+20 ms) – Only level values can be

entered for ST elevation. Slopes are not considered.

Level values: off, > 0.05, 0.1, 0.15, 0.2, 0.3 mV

Lead aVR is excluded from monitoring

PWC Calculation

The Physical Working Capacity (PWC) is calculated by dividing the

current load (in Watts, on a bicycle ergometer) by the patient’s weight

(in kilogram). PWC 130, PWC 150, and PWC 170 are the PWC values at

a heart rate of 130, 150, and 170 bpm, respectively. A missing PWC

value will be calculated by linear extrapolation when the heart rate has

reached the approximate target heart rate (< 10 bpm). For example, at a

heart rate of 161 bpm the PWC 170 value is calculated by

extrapolation

Target Load

The Target Load (in Watts, for bicycle ergometers) is calculated on the

basis of the patient’s weight, height, gender and age

For manual calculation use the following procedure:

1. Obtain the patient’s height and weight.

2. In Figure 5.17, draw a line from the height value to the weight

value.

3. Read the value at the point where this line crosses the body surface

line.

54 Physician’s Guide Revision C

2051167-002

Measurement Values: Target Load

Figure 5.17 Nomogram for the calculation of the body area (according to Du Bois)

4. Together with the patient’s age and gender, use this value to read off

the target load in the appropriate table below (Table 5.7 for female

patients, Table 5.8 for male patients).

Revision C Physician’s Guide 55

2051167-002

Measurement Values: Target Load

Table 5.7 Average expectation values of maximal target load, bicycle, values in Watts,

female patients

Table 5.8 Average expectation values of maximal target load, bicycle, values in Watts,

male patients

According to: Wonisch M, Berent R, Klicpera M, Laimer H, Marko C,

Schwann H, Schmid P. Tabellen zur Berechnung der österreichischen

Solllastformel. Praxisleitlinien Ergometrie, Journal für Kardiologie

2008, 15 (Supplementum A). Extrapolated for ages > 64 years and higher

body surfaces.

20 -24

25 -

30 -

35 -

40 -

45 -

50 -54

55 -

60 -

65 -

70 -

75 -

80 -

Age /

Body

surface

100 98 95 93 91 89 87 85 83 81 79 77 75

1.2 -

1.29

108 105 103 101 99 96 94 92 90 87 85 83 81

1.3 -

1.39

116 113 111 108 106 103 101 99 96 94 91 89 86

1.4 -

1.49

124 121 118 116 113 111 108 105 103 100 98 95 92

1.5 -

1.59

131 129 126 123 120 118 115 112 109 107 104 101 98

1.6 -

1.69

139 137 134 131 128 125 122 119 116 113 110 107 104

1.7 -

1.79

147 144 141 138 135 132 129 126 123 119 116 113 110

1.8 -

1.89

155 152 149 146 142 139 136 132 129 126 123 119 116

1.9 -

1.99

163 160 156 153 150 146 143 139 136 132 129 125 122

2.0 -

2.09

171 168 164 160 157 153 150 146 142 139 135 131 128

2.1 -

2.19

179 176 172 168 164 160 156 153 149 145 141 137 134

2.2 -

2.29

187 183 179 175 171 167 163 159 155 151 148 144 140

2.3 -

2.39

20 -24

25 -

30 -

35 -

40 -

45 -

50 -54

55 -

60 -

65 -

70 -

75 -

80 -

Age /

Body

surface

195 188 181 173 166 159 151 144 137 129 122 115 107

1.6 -

1.69

207 199 191 184 176 168 160 153 145 137 129 121 114

1.7 -

1.79

219 211 202 194 186 178 169 161 153 145 136 128 120

1.8 -

1.89

231 222 213 205 196 187 178 170 161 152 144 135 126

1.9 -

1.99

242 233 224 215 206 197 187 178 169 160 151 142 132

2.0 -

2.09

254 245 235 225 216 206 196 187 177 168 158 148 139

2.1 -

2.19

266 256 246 236 226 216 206 195 185 175 165 155 145

2.2 -

2.29

278 267 257 246 236 225 215 204 193 183 172 162 151

2.3 -

2.39

290 279 268 257 246 235 224 213 202 191 180 169 158

2.4 -

2.49

301 290 278 267 256 244 233 221 210 198 187 175 164

2.5 -

2.59

313 301 289 277 265 254 242 230 218 206 194 182 170

2.6 -

2.69

325 313 300 288 275 263 251 238 226 214 201 189 176

2.7 -

2.79

56 Physician’s Guide Revision C

2051167-002

Measurement Values: Target Load

For your notes

Revision C Physician’s Guide 57

2051167-002

6 T-Wave Alternans

(TWA)

Intended Use

The T-Wave alternans analysis is intended to provide the measurements

of the fluctuations of the ST-T-waves. The T-Wave alternans

measurements produced by the T-Wave alternans analysis are intended

to be used by qualified personnel in evaluating the patient in

conjunction with the patient's clinical history, symptoms, other

diagnostic tests, as well as the professional’s clinical judgment.

CAUTION

Results of the T-Wave alternans program must be

reviewed by a qualified physician, and should be used

only as an adjunct to clinical history, symptoms, and the

results of other non-invasive and/or invasive tests.

58 Physician’s Guide Revision C

2051167-002

T-Wave Alternans (TWA): Introduction

Introduction

There is a separate T-Wave Alternans Physician’s Guide

available,

which describes in more detail applications of TWA (Stress and Holter),

describes the clinical relevance of TWA analysis, and gives an overview

of relevant publications.

Two main methods exist for detection of T-Wave Alternans: the spectal

and the time domain method. A consensus guideline

on both methods

has been published. The time domain method is described here.

Electrical alternans affecting the ST-segment and T-wave is common

among patients at increased risk for ventricular arrhythmias. T-wave

alternans has been described in research literature as “a fluctuation in T

wave morphology occurring on an every-other-beat basis”

27,28,29,77

Although the electrical alternans affects both the ST-segment and the T-

wave, henceforward the alternans is referred to as “T-wave alternans”or

TWA. TWA on the electrocardiogram may serve as a noninvasive marker

of vulnerability to ventricular tachyarrhythmias, N Engl J Med 1994;

330:235-41

. TWA is known as a marker of cardiac electrical instability

with the potential for arrhythmia risk stratification. An example of this

clinical phenomenon is shown in figure 6.1.

The method for detection of electrical alternans of the ST-segment and

T-wave described here was adopted from Harvard Institutes of

Medicine, Boston. It was invented by Bruce D. Nearing and Richard L.

Verrier and is described in the patent “System and Method for

Quantifying Alternation in an Electrocardiogram Signal”

. The software

is capable of detecting morphology fluctuations in the ST-segment and

the T-wave using a method based on current standard computerized

ST-T waveform measurements. Furthermore, the method developed by

Dr. Nearing and Dr. Verrier directly measures the morphologic

fluctuation between odd and even beats in the time domain without

mathematical transformation, which allows visual confirmation of

measurements by an over-reader. The time domain measurement of

TWA is very similar to standard ST measurements, with which

clinicians are familiar.

The T-Wave Alternans (TWA) algorithm is to be used in a hospital,

doctor’s office, or clinic environment by competent health care

professionals for recording ST-T wave morphology fluctuations for

patients who are undergoing cardiovascular disease testing. The T-Wave

alternans analysis only provides the measurements of the fluctuations of

the ST-T-waves.

Revision C Physician’s Guide 59

2051167-002

T-Wave Alternans (TWA): Introduction

Figure 6.1 12-lead ECG from a bicycle exercise test. The amplitude of the ST-T segment

changes from beat to beat in an alternating pattern of higher and lower

ST levels. T-wave alternans of 50 µV is visible in lead V4 (ECGV4), 20 µV in

lead V5 (ECGV5).

60 Physician’s Guide Revision C

2051167-002

T-Wave Alternans (TWA): The TWA Algorithm

The TWA Algorithm

The basis for measuring TWA in the time domain is the formation of a

median PQRST complex for odd and even beats. The median beat

complex is formed in a way to minimize the effects random noise and to

allow comparison of the alternative beat T wave fluctuations by

comparing ST measurements in the ST-T wave on a lead-by-lead basis.

The key processes in this method are signal conditioning to reduce noise

effects, rejection of ectopic beats, and formation of a median PQRST

complex for odd and even beats.

First, the baseline shift of the current PQRST complex of the incoming

ECG is corrected using a cubic spline correction filter (see

Figure 6.2). The cubic spline is calculated on the basis of three points

taken from the isoelectric line preceding three consecutive QRS

complexes, which is an estimate of the baseline shift and is then

subtracted from the ECG. The cubic spline correction has the advantage

of not affecting the low frequency content of the signal, and has therefore

no negative effect on the T-wave alternans. The signal is then filtered

with a 40 Hz low pass filter to reduce high frequency noise produced

by muscle activities.

Additionally, every ECG lead is tested for noisy beats. Noisy beats are

detected/excluded by analyzing the high and middle frequency

content in the ST-T segment, and by analyzing the increase of the

variability of the ST-T segments in the odd beats or the increase of the

variability in the even beats. Noisy beats are excluded from further

processing. Beats excluded due to noise are always excluded in pairs

(e.g., noisy beat and an adjacent beat) to keep the even- odd- sequence.

The noise value is arbitrarily set to a value of 100µV when too many

beats have been excluded and the measurement is questionable.

Figure 6.2 Block diagram of the TWA algorithm

Subsequently, the algorithm separates odd and even incoming beats and

updates the odd median or the even median beat. The odd and even

median beats are updated by either a fixed increment, or a fraction of

the difference between the median template and the incoming beat,

Revision C Physician’s Guide 61

2051167-002

T-Wave Alternans (TWA): The TWA Algorithm

whichever is less. The default update factor is 1/8, but other factors (1/4,

1/16, 1/16 or 1/64) are selectable. These factors will affect the rate at

which the median templates track changes in the incoming signal. A

smaller factor (i.e.,1/64) will track changes slower, but will be more

resistant to noise. The update fraction to select will depend on the

incoming signal.

Then both the odd and even median beats are aligned to a cubic spline

(see section “Cubic Alignment” on page 62). This method has a

tremendous effect in reducing the low frequency noise within the PQRST

cycle, without disturbing the TWA measurement. It is a highly efficient

expansion of the cubic spline correction.

The TWA value is derived by calculating the maximal difference

between odd and even beats in the range of QRS end to T-wave end. A

nonlinear filter is applied to the TWA measurements to compute the

maximal difference in TWA measurements (see section “TWA

Calculation” on page 63). The amount of the maximal difference is the

TWA value. The noise value is an estimated RMS value, calculated from

the differences between even and odd median beat in the isoelectric area

before the QRS complex and after the T-wave.

The TWA value is annotated with a question mark if

 the heart rate is above the high heart rate limit (default 125 bpm),

 the noise value is greater than the high noise threshold (default

20 µV), and

 too many noisy beats have been excluded.

TWA values with a question mark are not used for calculation of a

maximum TWA value, for instance, and are represented by a gap or as a

dashed line in the trend curves.

Ectopic beats can have unpredictable effects on TWA measurements

. A

premature beat can affect the T-wave of the preceding complex;

therefore, premature complexes are detected and excluded together with

the preceding beat from incremental updating of odd and even median

beats.

Disconnected leads are also excluded from the TWA measurement

process.

62 Physician’s Guide Revision C

2051167-002

T-Wave Alternans (TWA): Cubic Alignment

Cubic Alignment

Correcting the baseline wander only with the cubic spline previously

described is not sufficient. Since only points in the isoelectric area

preceding the QRS complexes are used, the baseline wander can be

removed only to a limited extent. To further reduce baseline wander,

additional points which are located after the T-end are used for cubic

spline correction. This helps compensate for the differences between the

isoelectric areas in front of the QRS complex and the isoelectric area

between the T-wave and P-wave. The isoelectric area before the QRS

complex is influenced by atrial repolarization of the P-wave. Other

reasons for different amplitudes in both “isoelectric areas” could be a

short PR interval or a merging of P and T-waves. Applying the cubic

spline correction algorithm to points before the QRS complex and also to

points after the T-wave will further enhance removal of artificial

baseline wander, thus improving the measurement quality of T-wave

alternans values.

The cubic alignment algorithm aligns both the even and the odd median

beat to a cubic spline (see Figures 6.3 and 6.4). This cubic spline is

calculated on the basis of three points located exactly between the odd

and even median beats before the P-wave (1), before the QRS complex

(2), and after the T-wave (3). The algorithm is very tolerant to the

location of the points. Its yields very good results, even when atrial

repolarization or a short PR interval hides the isoelectric line before the

QRS complex, or a merging T-wave and P-wave in the other two points

hides the isoelectric T-P-interval.

Figure 6.3 Example of insufficient baseline wander correction. Odd and even median beats

(superimposed) after cubic spline correction. The value of 75 does not represent

T-wave alternans.

223

Revision C Physician’s Guide 63

2051167-002

T-Wave Alternans (TWA): TWA Calculation

Figure 6.4 Same odd and even median beats as in figure 6.3, but with alignment to a cubic

spline

TWA Calculation

After the odd and even median beats have been formed, the TWA value

is measured as the maximum difference between the ST-T wave

amplitudes in the two medians. The ST-T wave amplitude used for

comparison is the amplitude in the area between the end of the QRS and

the end of the T wave. When calculating the maximum difference

between odd and even beat in the area between QRS end and T-end,

there is still a possibility that high frequency noise can cause inaccurate

measurement of the TWA value. Therefore, a nonlinear filter is used to

further minimize the effects of high frequency noise. The filter uses a

20 ms window, one each in both the odd and the even median beats, both

starting at QRS end. The minimal difference between the windows is

selected and stored, and then the windows are moved one step towards

the end of the T-wave. Again, the minimal difference is selected and

stored and the windows are moved once more. The procedure is repeated

until the windows reach T-end. Then the maximum of the stored values

is considered the TWA value. The values are stored every 10 s and are,

therefore, available for the TWA trend, for instance.

64 Physician’s Guide Revision C

2051167-002

T-Wave Alternans (TWA): Interpreting/Correcting TWA Results

Interpreting/Correcting TWA Results

The higher the noise and artifact level of the ECG, the more difficult it is

to measure the TWA accurately. Noisy ECG can cause false positive

TWA values even in healthy patients. To assure maximum performance

of the algorithms, follow all the recommended guidelines for operation,

including skin preparation, electrode selection and placement.

Please note that electrode and leadwire movement can cause artifacts

similar to the frequency content of TWA. Securing the leadwires is

strongly recommended.

A positive T-wave alternans is defined when the TWA value reaches/

exceeds an amplitude threshold of 65 µV (for update factor 1/8)

Figure 6.5 TWA and Noise trend. The peak shortly after 8 minutes is a TWA episode.

Dashed curve segments identify unreliable TWA values. Noise values of 100 µV

indicate that too many beats have been excluded.

The user has the possibility to exclude leads with poor quality. In the

TWA quantitative study

it is recommended to use only the precordial

leads for TWA detection. Limb leads should be excluded because of their

high noise levels, which are usually much higher than the noise levels of

the precordial leads. But in the precordial leads only the horizontal

plane can be examined.

From the limb leads, lead II is the best and it is orthogonal to lead V2. It

has on average the highest QRS peak to valley amplitude of all limb

leads (II = 1 mV, I = 0.6 mV, III = 0.7 mV). Lead -aVR is close to II. With

including limb leads II and aVR also the vertical plane can be examined

sufficiently, without the disadvantage of increasing the noise level by

the leads aVL, I, aVF and III. General proposal: exclude leads aVL,

I, aVF and III.

The TWA algorithm does not exclude atrial fibrillation because it is

assumed that TWA in atrial fibrillation patients is highly predictive. But

atrial fibrillation waves can falsify TWA values. When disturbing

fibrillation waves exist, they are mostly apparent in lead V1. For those

cases it is proposed to exclude lead V1.

During treadmill testing it can happen that, due to the patient’s left arm

movement, the electrode V6 is touched continuously. This causes

artifacts, disturbing the TWA measurement. For those cases it is

proposed to exclude lead V6.

Revision C Physician’s Guide 65

2051167-002

T-Wave Alternans (TWA): Interpreting/Correcting TWA Results

The user interface, see Figure 6.6, helps the ECG reader to identify

incorrect TWA episodes. Since the TWA algorithm is a time domain

method, TWA can be visually confirmed, by clicking the TWA value of

interest in the trend curve, shortly before a peak. Therewith the

according odd and even TWA median beats (one upon the other) pop up

and the original ECG is shown for this time point. Based on this

information the reader is able to identify TWA values as erroneous, by

identifying confounding artifacts that are visible in the median beats

and/or original ECG on the same page. In addition the user interface

provides the necessary means for correcting erroneous TWA values.

Figure 6.6 User interface for TWA examination and correction. ECG from the Finnish

Cardiovascular Study (FINCAVAS)

, from Tampere University, Tampere,

Finland.

Note

Ischemia induced TWA

might be achieved at lower heart rates and,

therefore, at lower noise levels during supine bicycle exercise tests.

The assumption is based on the GE Healthcare stress database: The

same mean ST depression is achieved with upright bicycle exercise

tests at a heart rate of 139 bpm and with supine bicycle exercise

tests at a heart rate of 113 bpm. Besides, Wetherbee et.al.

compared treadmill upright and bicycle supine and found higher ST

depressions in supine bicycle exercise tests.

66 Physician’s Guide Revision C

2051167-002

T-Wave Alternans (TWA): TWA Settings

Figure 6.7 GE bicycle ergometer adjustable for supine exercise tests

TWA Settings

Even though the following settings can be changed, it is advised that you

leave them at their default factory settings.

 Factor for incremental updating. The factor will affect the rate at

which the median templates track changes in the incoming signal.

1/8 is the default factor, but other factors are available: 1/4, 1/8, 1/16,

and 1/64. With the factor 1/8, the algorithm will track TWA faster,

but will be more susceptible to the effects of noise. A smaller factor

(i.e., 1/64) will track changes slower, but will be more resistant to

noise. Factor 1/8 is the better choice, because therewith the transient

and short TWA episodes can be detected.

 Heart rate limit (default 125 bpm). Setting the heart rate limit to a

lower value can help avoid false measurements at high exercise, but

it also reduces the sensitivity of TWA detection.

 Noise limit (default 20 µV). Setting the noise limit to lower values

can help to avoid false TWA measurements due to muscle tremor, for

instance. Only high frequency noise can be controlled by this limit. A

lower limit also reduces the sensitivity of TWA detection.

Revision C Physician’s Guide 67

2051167-002

7 Exercise Test

Interpretation (XTI)

Description of the Exercise Test Interpretation

Program

An exercise test on a treadmill or a bicycle delivers a large number of

measurements that are valuable in predicting morbidity/mortality, in

detecting coronary artery disease and in describing the functional

exercise response of a patient. The ideal user would take all available

measurements, compare them with known thresholds, and come to a

complete assessment of the exercise test. However, it is very difficult to

have a comprehensive knowledge of all measurements and their

thresholds, especially of the new ones, namely HR recovery, FVE

recovery, ST/HR hysteresis, and HR reserve used.

The Exercise Test Interpretation (XTI) program compares the exercise

measurements against established thresholds and provides statements

and reasoning texts (explanation of the statements) when thresholds are

68 Physician’s Guide Revision C

2051167-002

Exercise Test Interpretation (XTI): Description of the Exercise Test Interpretation Program

exceeded.

Following is an example of a statement with the accompany-

ing reasoning text:

ST/T changes indicative of ischemia

because ST/HR hysteresis > 0.02 mV in [V5 V6] and

HR reserve used < 70%

The XTI program consists of rules and a rule interpreter. The rule

interpreter receives the input data, combines them with the rules, and

creates statements and reasoning texts.

The intention of the statements and reasoning texts is to provide a short,

clear, and accurate overview of the results of an exercise test. If the user

needs detailed information, GE Healthcare stress systems also offer

access to ECG strips, trend curves, median beats, etc.

Figure 7.1 Program structure of the exercise test interpretation. The exercise

measurements and other data are combined with the rules. Then statements

and reasoning texts are created.

The input data are exercise measurements (ST level, ST slope, ST/HR

hysteresis, ST/HR slope, ST/HR index, T-wave alternans value, heart

rate, ventricular ectopic beats per minute, METS, systolic blood

pressure, etc.) and other data (age, gender, medication, angina score,

lead descriptors, stress test device [treadmill or bicycle], etc.).

Measurements are available from the pretest phase, exercise phase,

recovery phase (1 min), recovery phase (3 min), and/or different leads.

ST levels are from all leads and all phases.

Based on the input data and the rules, the rule interpreter creates

 statements on risk prediction,

 statements on functional response,

 statements on ischemia (coronary artery disease), and

 technical statements.

Revision C Physician’s Guide 69

2051167-002

Exercise Test Interpretation (XTI): Description of the Exercise Test Interpretation Program

In addition, an overall statement is created, namely:

 normal exercise test response, or

 borderline exercise test response, or

 abnormal exercise test response

 undefined exercise test response

Apart from the overall statements, a statement is accompanied by a

reasoning text. In general, the reasoning texts conform to the rules.

Example:

Probably increased risk of malignant arrhythmias .. statement

because T-wave alternans ≥ 65 µV in [V3] .. reasoning text

Medication texts are additional reasoning texts. They appear only with

medication sensitive statements when the according medicament name

was entered.

Example:

Reduced heart rate response .. statement

because HR reserve used < 65% .. reasoning text

Possible cause: Medication beta-blocker .. medication text

A statement does not appear when it is not significant. This means, a

statement is created only if the corresponding rule is true. The

statement “Probably increased risk of malignant arrhythmias,” for

example, appears only when the T-wave alternans value is greater than

or equal to 65 µV.

Technical statements are created to alert the user when the standard

leads are incomplete, when the exercise test is too short, and when the

patient’s age is below 18 years.

The statements and reasoning texts are available in 19 languages:

English, French, German, Italian, Spanish, Portuguese, Swedish,

Hungarian, Polish, Norwegian, Danish, Dutch, Czech, Japanese,

Chinese, Russian, Korean, Finnish and Turkish. If a language is not

available, the texts appear in English.

70 Physician’s Guide Revision C

2051167-002

Exercise Test Interpretation (XTI): Limitations of the Exercise Test Interpretation (XTI) Program

Limitations of the Exercise Test Interpretation (XTI)

Program

XTI was developed for the main indications for exercise testing. These

indications are diagnosis of coronary artery disease, testing

vulnerability to arrhythmia during exercise and recovery, evaluating

exercise capacity, and assessment of a patient’s risk for mortality and

morbidity. The most appropriate test method for XTI is a standard

exercise test up to the patient’s maximal capacity during exercise with a

3-minute recovery phase at a low work load (see figure 7.2).

Figure 7.2 Standard exercise test

Submaximal exercise tests (for post MI patients), a patient’s physical

impairment or poor motivation may reduce sensitivity in ischemia

detection or risk assessment.

If a patient is not motivated, statements, such as “Reduced heart rate

response to exercise” or “Insufficient exercise capacity” might be

presented.

XTI was developed for treadmill and bicycle testing. When performing

pharmacological testing or pharmacological testing combined with

exercise, low METS values occur. A low METS value might suppress

statements on heart rate response and exercise capacity, for instance.

Warning

The results of the Exercise Test Interpretation must be

confirmed by a qualified physician or cardiologist and

should be used only as an adjunct to clinical history,

symptoms, and the results of other non-invasive or

invasive tests.

Revision C Physician’s Guide 71

2051167-002

Exercise Test Interpretation (XTI): Examples for complete interpretation

Examples for complete interpretation

Example 1

Probably increased risk of cardiovascular event

because HR recovery ≤ 12 per minute

Reduced heart rate response to exercise

because HR reserve used < 65%

Insufficient exercise capacity

because metabolic equivalent (METS) < 6

ST/T changes indicative for ischemia

because ST/HR hysteresis > 0.02 mV in [V5] and HR reserve used < 70%

Abnormal exercise test response

Example 2

ST/T changes may be clinically significant

because horizontal or downsloping ST ≤ -0.1 mV in [V5]

Borderline exercise test response

Example 3

Cannot rule out clinically significant ST/T changes

because horizontal or downsloping ST ≤ -0.05 mV in [V5]

Borderline exercise test response

Example 4

Probably normal exercise response

72 Physician’s Guide Revision C

2051167-002

Exercise Test Interpretation (XTI): Rules

Rules

The rules of the Exercise Test Interpretation determine when a certain

interpretive statement is displayed. The rules contain the threshold

values which, when exceeded, trigger the display of the respective

interpretive statement (identified in the following sections in bold).

Most rules combine different measurements. METS is often used for

such combinations. For example, a low METS value with a large TWA

value has been shown to be more predictive of malignant arrhythmias.

Rules for risk assessment

Probably increased risk of cardiovascular event:

 HR recovery < 12 per minute

Excluded: PVCs in recovery > 3 per minute, age <

20 years

 Bicycle: Duke treadmill score (DTS) < -11

Treadmill: Duke treadmill score (DTS) < -10

Excluded: METs <

1.8, age < 40 years

Probably risk of cardiovascular event:

 PVCs in recovery > 7 per minute

Bicycle: excluded peak HR <

120

Probably increased risk of malignant arrhythmias:

 Incremental update factor 1/4:

T-wave alternans ≥ 85µV

 Incremental update factor 1/8:

T-wave alternans ≥ 65µV

 Incremental update factor 1/16:

T-wave alternans ≥ 32µV

 Incremental update factor 1/32:

T-wave alternans ≥ 16µV

 Incremental update factor 1/64:

T-wave alternans ≥ 10µV

Excluded: Leads I, III, aVF, aVL, age <

Probably increased risk of stroke/cardiovascular event:

 Atrial fibrillation in pretest or exercise phase

Excluded: Age <

Exercise-induced bundle branch block:

 QRS width at peak exercise > 130 ms and

QRS width at start exercise > 120 ms and

Difference of QRS widths > 30 ms

Revision C Physician’s Guide 73

2051167-002

Exercise Test Interpretation (XTI): Rules

Exercise-induced wide QRS tachycardia:

 Wide QRS tachycardia in exercise phase, but not in pretest phase

 Wide QRS tachycardia in recovery phase, but not in pretest and

exercise phase

Exercise-induced supraventricular tachycardia:

 Supraventricular tachycardia in recovery phase

Cannot compute risk:

 No rules above are true and METS < 1.8 and METS > 1.0

Rules for functional response assessment

Significantly reduced heart rate response to exercise:

 Treadmill: HR reserve used < 65%

 Bicycle upright: HR reserve used < 42%

 Bicycle semi supine: HR reserve used < 35%

 Bicycle supine: HR reserve used < 27%

 if beta-blocker, output info text

“Possible cause: medication beta-blocker”

Reduced heart rate response to exercise:

 Treadmill: HR reserve used < 80%

 Bicycle upright: HR reserve used < 60%

 Bicycle semi supine: HR reserve used < 50%

 Bicycle supine: HR reserve used < 37%

 HR reserve used < 65% and METS > 2

 if beta-blocker, output info text

“Possible cause: medication beta-blocker”

The heart rate response rules above are excluded when METS is

between <

1.8 and > 1.0.

Insufficient exercise capacity:

Treadmill:

Female:

 METS < 7.5, age ≤ 29 years

 METS < 7.0, 30 ≥ age ≤ 39 years

 METS < 6.0, 40 ≥ age ≤ 49 years

 METS < 5.0, 50 ≥ age ≤ 59 years

 METS < 4.5, 60 ≥ age ≤ 69 years

74 Physician’s Guide Revision C

2051167-002

Exercise Test Interpretation (XTI): Rules

 METS < 3.5, 70 ≥ age ≤ 79 years

 METS < 2.5, age > 79 years

Male:

 METS < 8.0, age ≤ 29 years

 METS < 7.5, 30 ≥ age ≤ 39 years

 METS < 7.0, 40 ≥ age ≤ 49 years

 METS < 6.0, 50 ≥ age ≤ 59 years

 METS < 5.5, 60 ≥ age ≤ 69 years

 METS < 4.5, 70 ≥ age ≤ 79 years

 METS < 3.5, age > 79 years

Bicycle:

Female:

 METS < 6.0, age ≤ 29 years

 METS < 5.6, 30 ≥ age ≤ 39 years

 METS < 4.8, 40 ≥ age ≤ 49 years

 METS < 4.0, 50 ≥ age ≤ 59 years

 METS < 3.6, 60 ≥ age ≤ 69 years

 METS < 2.8, 70 ≥ age ≤ 79 years

 METS < 2.0, age > 79 years

Male:

 METS < 6.4, age ≤ 29 years

 METS < 6.0, 30 ≥ age ≤ 39 years

 METS < 5.6, 40 ≥ age ≤ 49 years

 METS < 4.8, 50 ≥ age ≤ 59 years

 METS < 4.4, 60 ≥ age ≤ 69 years

 METS < 3.6, 70 ≥ age ≤ 79 years

 METS < 2.8, age > 79 years

Reduced exercise capacity:

Treadmill:

Female:

 METS < 10.0,age ≤ 29 years

 METS < 9.0, 30 ≥ age ≤ 39 years

 METS < 8.0, 40 ≥ age ≤ 49 years

 METS < 7.0, 50 ≥ age ≤ 59 years

 METS < 6.0, 60 ≥ age ≤ 69 years

 METS < 4.5, 70 ≥ age ≤ 79 years

 METS < 4, age > 79 years

Male:

 METS < 11.0, age ≤ 29 years

Revision C Physician’s Guide 75

2051167-002

Exercise Test Interpretation (XTI): Rules

 METS < 10.0,30 ≥ age ≤ 39 years

 METS < 8.5, 40 ≥ age ≤ 49 years

 METS < 8.0, 50 ≥ age ≤ 59 years

 METS < 7.0, 60 ≥ age ≤ 69 years

 METS < 5.5, 70 ≥ age ≤ 79 years

 METS < 4.5, age > 79 years

Bicycle:

Female:

 METS < 8.0, age ≤ 29 years

 METS < 7.2, 30 ≥ age ≤ 39 years

 METS < 6.4, 40 ≥ age ≤ 49 years

 METS < 5.6, 50 ≥ age ≤ 59 years

 METS < 4.8, 60 ≥ age ≤ 69 years

 METS < 3.6, 70 ≥ age ≤ 79 years

 METS < 3.2, age > 79 years

Male:

 METS < 8.8, age ≤ 29 years

 METS < 8.0, 30 ≥ age ≤ 39 years

 METS < 6.8, 40 ≥ age ≤ 49 years

 METS < 6.4, 50 ≥ age ≤ 59 years

 METS < 5.6, 60 ≥ age ≤ 69 years

 METS < 4.4, 70 ≥ age ≤ 79 years

 METS < 3.6, age > 79 years

The exercise capacity rules above are excluded in case of METS <

1.8.

Abnormal blood pressure response:

 Max.syst.blood pressure > 250 mmHg

 Max. syst. blood pressure > 33.3 kPa

Insufficient rate pressure response:

Bicycle:

 Max. Rate Pressure Product (RPP) < 16,000 mmHg per min

 Max. Rate Pressure Product (RPP) < 2,133 kPa per min

Treadmill:

 Max. Rate Pressure Product (RPP) < 20,000 mmHg per min

 Max. Rate Pressure Product (RPP) < 2,666 kPa per min

76 Physician’s Guide Revision C

2051167-002

Exercise Test Interpretation (XTI): Rules

The rate pressure rules above are excluded in case of METS <

2 and

METS >

1.0.

Cannot compute functional response:

 No functional response rules above are true and METS < 1.8 and

METS > 1.0.

Rules for ischemia assessment (coronary artery disease)

The following ischemia rules are applied to leads I, II, III, aVF, V2-V6.

Leads aVR, aVL and V1 are generally excluded. In case of peak exercise

LBBB and LBBB shaped pacemaker stimulation, the rules are also

excluded.

ST/T changes indicative of ischemia:

 ST/HR hysteresis > 0.05 mV

The following rules are suppressed when the preceding ischemia rules is

true.

Bicycle:

 ST/HR hysteresis > 0.02 mV and METSs > 2.0 and HR reserve used

< 70%

Treadmill:

 ST/HR hysteresis > 0.02 mV and METSs > 2.0 and HR reserve used

< 85%

Others:

 ST/HR hysteresis > 0.02 mV and METSs < 2.0 and HR reserve used

> 8%

Excluded for the above ischemia rules: Recovery time <

60 s; Leads V2,

V3, and V4 in case of RBBB at peak exercise or 3 min recovery.

The following rules are suppressed when at least one of the preceding

ischemia rules is true.

Bicycle:

 ST/HR index > 1.6 mV/bpm and HR reserve used < 70%

Treadmill:

 ST/HR index > 1.6 mV/bpm and HR reserve used < 85%

Excluded: Leads V2, V3, and V4 in case of RBBB at peak exercise.

Revision C Physician’s Guide 77

2051167-002

Exercise Test Interpretation (XTI): Rules

ST/T changes may be clinically significant:

ST = ST amplitude at peak exercise

The following rule is suppressed when at least one of the preceding

ischemia rules is true.

 ST level ≤ -0.1 mV and ST slope < 0.05 mV/s

Excluded: Leads V2, V3, and V4 in case of RBBB at peak exercise. All

leads in case of LBBB and LBBB shaped pacemaker stimulation at peak

exercise.

The following rule is suppressed when at least one of the preceding ST

segment changes rules is true.

 recovery ST ≤ -0.05 mV and recovery ST slope < 0.05 mV/s

Excluded: Leads V2, V3, and V4 in case of RBBB at 3 min recovery. All

leads in case of LBBB and LBBB shaped pacemaker stimulation at 3

min recovery.

Cannot rule out clinically significant ST/T changes:

ST = ST amplitude at peak exercise

The following rules are suppressed when at least one of the ischemia

rules above is true.

 ST ≤ -0.05 mV and ST slope < 0.05 mV/s (horizontal)

 ST < -0.15 mV and ST slope < 1 mV/s (slightly ascending)

Excluded: Leads V2, V3, and V4 in case of RBBB at peak exercise

The rules above are excluded in case of peak exercise LBBB and LBBB

shaped pacemaker stimulation

Rules for overall statements

Probably normal exercise response:

None of the preceding rules was true.

Borderline exercise response:

The “ST/T changes may be clinically significant”, the “Cannot rule out

clinically significant ST/T changes”, the “Reduced exercise capacity”,

and/or the “Reduced heart rate response to exercise” rules have been

true.

Abnormal exercise response:

At least one of the preceding rules was true, except the rules for the

borderline exercise response

Undefined exercise response:

None of the preceding rules was true and METS <

1.8.

78 Physician’s Guide Revision C

2051167-002

Exercise Test Interpretation (XTI): Rules

Technical Rules

No ECG

Inhibits all other rules

Standard leads incomplete:

One or more missed leads: I, II, III, aVR, aVL, aVF, V1 – V6

Warning! Results are questionable:

Age < 18 years,

Exercise_time < 1 min,

Recovery time < 1 min,

Undefined start exercise.

Caution: Age unknown (45 years assumed)

Age outside range of 0 ... 200 years.

Caution: Gender unknown (male assumed)

Not male or female.

Revision C Physician’s Guide 79

2051167-002

Exercise Test Interpretation (XTI): Graphical XTI output

Graphical XTI output

The graphical XTI output is an easy-to-understand and comprehensive

display of the results of an exercise test. It translates the XTI

statements into positions on a 4x4 matrix. The matrix consists of four

categories

 Risk

 Functional response

 Ischemia (CAD)

 Overall

and four exercise response columns, namely:

 undefined (grey)

 normal (green)

 borderline (yellow)

 abnormal (red)

Position -33 0 33 67 100%

Figure 7.3 Graphical display of XTI results. Categories are Risk, Functional response,

Ischemia (CAD) and Overall. Exercise test response columns are undefined,

normal, borderline and abnormal.

First the statements are assigned to the 4 categories “Risk”, “Functional

Response”, ”Ischemia (CAD)” and “Overall”. Then the positions of the

corresponding statements are calculated for each category to provide

marker positions. The positions are based on the quality of the

statements in a category. The quality is assessed by the specificity and

reliability of the measurements used in the statement (see Table 7.1). If

more than one statement is true in one category, the statement with the

highest quality defines the final position. If no statement in a category is

true, the marker is positioned in the middle of the green area in this

category or in the middle of grey in case of no ECG or undefined exercise

response (METs value too small).

The main contributor for “Risk” is HR recovery, for “Functional

response” it is METS, and for “Ischemia (CAD)” it is ST/HR hysteresis.

The marker for the “Overall” category row is normally copied from the

category above whose marker is farthest to the right.

80 Physician’s Guide Revision C

2051167-002

Exercise Test Interpretation (XTI): Graphical XTI output

Examples:

The statement “Increased risk of cardiovascular death” is translated

according to its CVD detection quality to the marker position at the

beginning of the red area of the “Risk” row.

The statement “Insufficient exercise capacity” is translated according to

its specificity for overall mortality detection to the marker position in the

middle of the red area of the “Functional response” row.

The statement “ST/T changes indicative of ischemia” with ST/HR

hysteresis > 0.05 mV is translated according to its CAD detection quality

to the marker position in the upper part of the red area of the “Ischemia

(CAD)” row.

Table 7.1 Statements and their positions

 By positioning the cursor on a marker, the corresponding statement

and reasoning text are displayed (Figure 7.4). By positioning the

Statement Position in %

Risk

Probably increased risk of cardiovascular event 86

Probably risk of cardiovascular event (FVE recovery) 45

Probably increased risk of malignant arrhythmias 75

Probably increased risk of stroke/cardiovascular event 80

Exercise induced bundle branch block 58

Exercise induced wide QRS tachycardia 55

Exercise induced atrial fibrillation 49

Exercise induced supraventricular tachycardia 52

Undefined risk -17

Probably normal 17

Functional response

Significantly reduced heart rate response to exercise 81

Reduced heart rate response to exercise 50

Insufficient exercise capacity 85

Reduced exercise capacity 55

Abnormal blood pressure response 77

Insufficient rate pressure response 75

Undefined functional response -17

Probably normal 17

Ischemia (CAD)

ST/T changes indicative of ischemia (ST/HR hysteresis) 96

ST/T changes indicative of ischemia (ST/HR hysteresis + HR reserve used) 90

ST/T changes indicative of ischemia (ST/HR index + HR reserve used) 88

ST/T changes may be clinically significant (ST peak ) 50

ST/T changes may be clinically significant (ST recovery) 45

Cannot rule out clinically significant ST/T changes 40

Probably normal 17

Others

Probably normal exercise response 17

Undefined exercise response -17

No ECG -20

Revision C Physician’s Guide 81

2051167-002

Exercise Test Interpretation (XTI): Graphical XTI output

cursor in the “Functional response” category, for instance, all

corresponding statements and reasoning texts of this category are

displayed (Figure 7.5). By positioning the cursor in the “Overall”

category, all statements and reasoning texts are displayed (Figure

7.6).

Figure 7.4 Example: Statement and reasoning texts when cursor is at marker position

"abnormal functional response".

Figure 7.5 Example: Statement and reasoning texts when cursor is in the "Functional

response" position.

Figure 7.6 Example: Statement and reasoning texts when cursor is in the "Overall"

position.

82 Physician’s Guide Revision C

2051167-002

Exercise Test Interpretation (XTI): Graphical XTI output

For your notes

Revision C Physician’s Guide 83

2051167-002

8 Audio Assessment of

Exercise Tests

An audio replay gives a fast overview of an exercise test andprovides a

highly sensibledemonstration of all kind of discontinuities, for instance,

arrhythmias like atrial fibrillation, intermitted supraventricular

tachycardias and intermitted branch bundle blocks. Audio assessment is

intended to be an addition to the visual assessment at the end of an

exercise test. It might help not to overlook important ECG phenomena.

Figure 8.1 Button on CardioSoft/CS for starting and stopping replay of an audio ECG.

During replay all other exercise results can also be examined visually

Fast Assessment

The ECG is reproduced 60 times faster. Therefore an exercise test of

10 min, has a duration of 10 sec, for instance.

The increasing of the frequency corresponds to the exercise phase,

relative to the increasing of the heart rate during exercise. The

subsequent decreasing of the heart rate corresponds to the recovery

phase.

The reproduction is in stereo. The right channel represents lead V2 and

the left channel lead V5.

Based on the recent knowledge about audible ECG, following

abnormalities/discontinuities are assessable:

 Absolute arrhythmia (atrial fibrillation)

 No HR response to new stage

 No steady increasing of HR during exercise

 No significant HR increase from rest to peak exercise

 No steady decreasing of HR during recovery

 No significant decrease of HR during recovery

 PVCs in recovery phase

 V2 is significantly louder than V5

84 Physician’s Guide Revision C

2051167-002

Audio Assessment of Exercise Tests: Fast Assessment

Examples

To listen to an audio ECG please click on the loudspeaker symbol.

The ECGs and clinical data, are from the Finnish Cardiovascular Study

(FINCAVAS)

, from Tampere University, Tampere, Finland.

ra0001 (A0001)

Normal exercise response (patient 67 y, f, 154 cm, 61 kg)

During exercise the heart rate increases normally. The changes of the

stages are clearly audible. A few supraventricular premature beats are

audible during exercise and recovery.

The loudness of V2 is similar to the loudness of V5.

ra1016 (D0018)

Atrial fibrillation. From the beginning there is a grumbling sound,

typical for atrial fibrillation. The sound continues during the whole

exercise test.

The patient (73 y, m, 186 cm, 86 kg) died 3 months after the exercise

test, because of artereosclerotic heart disease (WHO-ICD, I25.1

The loudness of V2 is higher than the one of V5, reason left bundle

branch block (LBBB).

Figure 8.2 Atrial fibrillation in pretest phase, lead V1

ra0232 (A0233)

Abnormal exercise response. The patient (67 y, m, 161 cm, 68 kg) died

3 months after the exercise test, because of an acute anterior myocardial

infarction (WHO-ICD, I21.0

During the recovery phase frequent ventricular ectopics and a sudden

increase of heart rate are audible.

The loudness of V2 is higher than the one of V5, due to an old anterior

myocardial infarct.

Revision C Physician’s Guide 85

2051167-002

Audio Assessment of Exercise Tests: Fast Assessment

Figure 8.3 Increase of the heart rate and ventricular ectopics(VE/min) in recovery phase.

For more details on audio assessment of exercise tests, see the Audio-

ECG Physician’s Guide.

Recovery

86 Physician’s Guide Revision C

2051167-002

Audio Assessment of Exercise Tests: Fast Assessment

For your notes

Revision C Physician’s Guide 87

2051167-002

9 Resting ECG

Interpretation and Pre

Test Risk Assessment

With their exercise test systems CASE and CardioSoft/CS, for instance,

GE Healthcare offer tools to support the planning of a patient’s exercise

test procedure. These tools are the 12 SL Resting ECG Interpretation

program, see Physician’s Guide

, and the AHA (American Heart

Association) Coronary Heart Disease Risk and Stroke Risk Prediction

program (see below).

AHA Coronary Heart Disease Risk and Stroke Risk

Prediction

A number of risk factors have been found to be associated with a

distinctly enhanced risk of coronary heart disease, including sudden

coronary death, myocardial infarction, and angina pectoris.

Framingham and other epidemiological studies have shown that a

prediction of the incidence rate of coronary heart disease can be made for

groups of persons well in advance of the appearance of symptoms.

Among the risk factors identified are: gender, age, cigarette smoking,

elevated blood pressure, high levels of serum cholesterol, low HDL-

cholesterol (High Density Lipoprotein cholesterol), diabetes, and ECG

abnormalities. These factors are not the only risk factors which might

logically be considered in assessing risk of coronary heart disease, but

they are a set of proven merit which can be readily measured by an office

nurse or a technician without hazard or trauma to the patient.

Coronary Heart Disease Risk Prediction

GE’s Coronary Heart Disease Risk Prediction program is an

implementation of an update of an earlier risk handbook (American

Heart Association. Coronary Risk Handbook. Estimation of Coronary

Heart Disease in Daily Practice, Dallas 1973). It is based on more recent

analysis of the Framingham Heart Study which adds HDL-cholesterol to

the risk profile (Anderson KM, Wilson PDF, Odell PM, Kannell WB. “An

updated Coronary Risk Profile”, Circulation. 1990). In general, the more

risk factors present or the greater the degree of abnormality of any

88 Physician’s Guide Revision C

2051167-002

Resting ECG Interpretation and Pre Test Risk Assessment: AHA Coronary Heart Disease Risk and Stroke Risk Prediction

factor, the greater the risk. The program combines factors and facilitates

assessment of risk for persons whose blood pressure or lipid values may

not reach some arbitrary value set as “abnormal”. By using the program,

fewer persons will be misclassified because they have borderline values

for blood pressure and lipids.

The program provides a synthesis of the information expressed as the

probability of a coronary event within five or ten years of risk factor

assessment. It allows pre-symptomatic assessment of coronary

vulnerability from a coronary risk profile. These results do not

necessarily apply to persons who already have coronary heart disease.

Stroke Risk Prediction

Probability of stroke risk is directly related to the level of risk factor

abnormality, in the case of age and systolic blood pressure, and to the

number of risk factors present, i.e., diabetes mellitus, cigarette smoking,

cardiovascular disease, or electrocardiographic abnormality – left

ventricular hypertrophy or atrial fibrillation. GE’s program combines

the risk factors and makes a quantitative assessment of risk even

though the systolic blood pressure is not high enough to be classified as

definitely in the hypertensive range. Since most hypertensives are in the

borderline category, and this level of blood pressure elevation is clearly

important in stroke incidence, the program takes such intermediate

levels into account.

Stroke risk is somewhat greater for men than women, and the risk

profiles are gender-specific. For men and women a certain systolic blood

pressure level on treatment has a higher risk than the untreated level;

consequently anti-hypertensive treatment is taken into account. For

women an additional interaction adjustment is needed. The relationship

between blood lipids and stroke is clearly different from their key role in

the pathogenesis of coronary heart disease. Total cholesterol and HDL-

cholesterol are important risk factors influencing coronary heart disease

but not stroke.

GE’s program provides a synthesis of the risk factor information for each

gender expressed as the probability of a stroke event in a specified ten

year period. This probability of stroke in a pre-symptomatic individual

does not necessarily apply to persons with prior cerebral infarction or

hemorrhage.

Revision C Physician’s Guide 89

2051167-002

Resting ECG Interpretation and Pre Test Risk Assessment: AHA Coronary Heart Disease Risk and Stroke Risk Prediction

Definitions of Terms and Measurements

Risk factors in the program and their method of measurement are

defined in the following list. If other measurement techniques are used

to determine any of the values, they should be adjusted to those defined.

HDL-Cholesterol High density lipoprotein cholesterol determined after heparin-

manganese precipitation.

Fasting Total

Cholesterol

The physician is advised that the cholesterol values in the

program are based on the Abell-Kendall method. If your

laboratory uses another method, you should determine from the

laboratory director the correction factor necessary to convert to

the Abell-Kendall values. The correction factor varies among

laboratories and must be determined by each laboratory. Most

direct and automated cholesterol determinations give values

five percent to 15 percent above those stated in the tables. Use

of corrected values will affect the precision of risk estimates

made from these tables. Plasma cholesterol measurements

used here are 3 to 4.7 percent lower than serum (Cloey T,

Bachorik PS, Becker D, Finney C, Lowry D, Sigmund W.

“Reevalution of serum-plasma differences in total cholestrol

concentration”. JAMA 1990; 253:2788-2789.).

Systolic Blood

Pressure, BP sys

(mmHg)

Casual pressure taken with the subject seated and resting for

five minutes. An average of two measurements is preferred.

Cigarette Smoking Refers to whether or not an individual is a cigarette smoker or

has quit only within the last 12 months. The program does not

take into account the intensity of the smoking habit and

contrasts only smokers and non-smokers.

Anti-Hypertensensive

Therapy

Currently on anti-hypertensive medication.

Cardiovascular Disease History of myocardial infarction, angina pectoris, coronary

insufficiency, intermittent claudication or congestive failure.

Atrial fibrillation/flutter History of atrial fibrillation/flutter.

Left Ventricular

Hypertrophy by ECG

LVH by ECG consists of finding tall R waves in leads reflecting

potentials from the left ventricle accompanied by ST segment

depression or T wave inversion.

90 Physician’s Guide Revision C

2051167-002

Resting ECG Interpretation and Pre Test Risk Assessment: AHA Coronary Heart Disease Risk and Stroke Risk Prediction

For your notes

Revision C Physician’s Guide 91

2051167-002

10 Exercise Testing

General

Typically an exercise test consists of three phases: the pretest, the

exercise and the recovery phase. Exercise and recovery phase are

important for calculation of the ST/HR hysteresis, for instance. An

exercise test should comprise a pretest phase of one minute or more,

started after successful applied electrodes, an exercise phase, and a

recovery phase of three or more minutes.

In their Exercise Test Systems, GE Healthcare offers tools to support the

planning of the test procedure for a patient. These tools are the 12 SL

Resting ECG Interpretation program, see Physician’s guide

, and the

AHA Coronary Heart Disease Risk and Stroke Risk Prediction program,

discussed in chapter 9.

Good electrode application and preparation is essential to receive good

results. A lack of diligence in applying the electrodes causes high noise

and multiple artifacts in the ECG and leads to difficulties in evaluating

an exercise test and their measurements, especially measurements

relating to ST segment and to T-wave alternans. Good skin preparation,

electrode quality, and good placement are important factors for

achieving reliable ECG measurements from the ST/T segment, for

instance, and a reliable Exercise Test Interpretation. To assure

maximum performance, follow all the recommended guidelines for

operation, including skin preparation, electrode selection, and

placement. Please note that electrode and leadwire movement can cause

artifacts similar to the frequency content of the ST/T segment of the

ECG. Securing the leadwires is strongly recommended.

92 Physician’s Guide Revision C

2051167-002

Exercise Testing: Role of Stress Testing in Reducing Cost of Healthcare

Role of Stress Testing in Reducing Cost of

Healthcare

Unnecessary Cath Lab interventions have been identified as one of the

most wasteful expenditures in all of healthcare, worldwide.

91,92

Even

Consumer Reports* has qualified it as being in the top 10.

Fortunately, the stress test is an effective tool that can be used to curb

this waste. Below are a few quotes from the scientific literature that

make this point:

 The appropriateness of PCI (percutaneous catheter intervention)

has been challenged because many patients undergoing PCI lack

documentation of ischemia (lack of oxygen) by non-invasive testing

(that is, stress testing) prior to the procedure…”

 On the basis of the evidence, PCI is considered formally appropriate

only if the patient has (1) ischemic symptoms that might be

improved by revascularization, (2) objective evidence of ischemia by

stress testing, and (3) failed a trial of optimal medical management

because the ischemia is intractable to maximally tolerated doses of

antiischemic medications.”

One of the root causes for this waste is the erroneous but, unfortunately,

widespread belief that plaque which significantly narrows a coronary

artery will lead to a heart attack. This belief has resulted in a

preoccupation with these obstructions,

causing clinicians to invasively

treat patients , even when they are asymptomatic, by mechanically

dilating these areas and inserting a stent to keep them dilated. This is

despite the preponderance of evidence that heart attacks are caused by

clots which result from plaque rupture (that is, where the plaque loses

its fibrous cap and leaks clot activating factors into the blood that flows

by it).

In fact, most heart attacks result from a rupture of a plaque that

is so small it does not elicit any symptoms

and is often invisible on the

X-ray images obtained via the Cath Lab.

96-98

Nevertheless, it took the results of the COURAGE trial to put care

providers on notice

that mechanical treatments of these obstructions,

even when the narrowing of the coronary artery exceeds 70%, is of little

benefit unless the patient is symptomatic and has been found to have

objective evidence of myocardial ischemia.

100

Given the controversy raised by the COURAGE trial, more investiga-

tions were launched. Meta-analysis of similar trials has also generated a

similar conclusion: “initial stent implantation for stable CAD shows no

evidence of benefit compared with initial medical therapy for prevention

of death, nonfatal MI, unplanned revascularization, or angina.”

101,102

The cost of this inappropriate treatment has been found to be

substantial. Indeed, without documented evidence of ischemia, “the

added cost of PCI was approximately $10,000, without significant gain

in life-years or quality-adjusted life-years. The incremental cost-

effectiveness ratio varied from just over $168,000 to just under $300,000

* Source: http://www.consumerreports.org/cro/2012/08/treatment-traps-

to-avoid/index.htm (Accessed February 13, 2013)

Revision C Physician’s Guide 93

2051167-002

Exercise Testing: Role of Stress Testing in Reducing Cost of Healthcare

per life-year or quality-adjusted life-year with PCI. A large minority of

the distributions found that medical therapy alone offered better

outcome at lower cost. The costs per patient for a significant

improvement in angina frequency, physical limitation, and quality of life

were $154,580, $112,876, and $124,233, respectively.”

103

Estimates of the number of inappropriate treatments are derived from

several sources, which need to be interpreted in relation to the different

reasons for performing PCI. For instance, PCI is also used for treating

heart attacks due to plaque rupture and clotting. However, emergency

Cath Lab procedures (which often include stents) are appropriate. In

fact, these emergency procedures have been found to be very effective

versus the other treatment for a heart attack, namely thrombolytic

therapy.

104-107

Therefore, it is important to separate out procedures

related to emergency PCI (also called primary, direct or immediate)

versus those related to an elective PCI due to stable plaque.

According to the most recent statistics available from the AHA,

it is

estimated that 1,133,000 PCI procedures are performed per year in the

United States. In addition, it is estimated that another 1,072,000

inpatient diagnostic cardiac catheterizations were performed.

Diagnostic CATH procedures are also relevant to this issue since it may

be done as an extra step to confirm the presence of a narrowing before

the PCI is actually performed. Nevertheless, it is not evident from these

AHA statistics which of these procedures are due to an elective versus

emergency condition. Instead, we need to turn to cohort studies to

estimate the level of inappropriate treatment.

For example, in the United States, “a retrospective, observational cohort

study using claims data from a 20% random sample of 2004 Medicare

fee-for-service beneficiaries aged 65 years or older who had an elective

PCI (N=23 887)”

108

found that only 44.5% (n=10 629) of these patients

underwent stress testing within the 90 days prior to the elective PCI.

Moreover, “there was wide regional variation among the hospital

referral regions with stress test rates ranging from 22.1% to 70.6%

(national mean, 44.5%; interquartile range, 39.0%-50.9%).”

108

“These

numbers are remarkably similar to those in the United Kingdom, where

43% of patients have stress testing before elective PCI. Furthermore,

revascularization rates also vary widely, with an 83% higher rate in

Florida than in Oregon. Revascularization rates depend on race (28%

variation) and cardiac catheterization rates (68% variation), which in

turn depends on hospital admission rates for CAD as well as the number

of cardiac surgeons and interventionalists in the local population.”

109

In a larger, “multicenter, prospective study of patients within the US

National Cardiovascular Data Registry undergoing PCI between July 1,

2009, and September 30, 2010, at 1,091 US hospitals”,

110

it was found

that of the “500,154 procedures classified”, 144,737 (28.9%) were elective

PCIs.

110

For emergency situations, almost all of the procedures were

appropriate. However, in the non-acute setting, “only 50% of procedures

were classified as appropriate, 38% as uncertain, and 12% as

inappropriate.” Moreover, the majority of these inappropriate

procedures were performed in patients with little to no angina or with

low-risk ischemia on stress testing. The study also found “substantial

hospital variation in the rate of inappropriate PCI”.

110

This led the

authors to conclude that “a better understanding of the clinical settings

in which inappropriate PCIs occur and reduction in their variation

across hospitals should be targets for quality improvement.”

110

94 Physician’s Guide Revision C

2051167-002

Exercise Testing: A Clinical Approach to Exercise Testing

A Clinical Approach to Exercise Testing

Note

This chapter is taken unchanged from the preceding physician’s

guides. It is not completely state of art, especially in risk assessment

and new exercise measurements, but it remains valuable in a lot of

aspects of exercise testing.

Note

The following material in this chapter 10 has been adapted from The

Clinical Approach to Exercise Testing by Stephen P. Glasser, M. D.,

F. A. C. P., F. A. C. C. and Pamela I. Clark.

Reprinted with permission from Lippincott/Harper & Row.

Multistage exercise testing allows the physician to observe a subject’s

physiological adaptation to exercise and, as such, is a valuable extension

of a standard history and physical examination. Exercise testing may be

used to evaluate symptoms such as chest pain, palpitations, dyspnea, or

easy fatigability; estimate the severity of coronary artery disease;

appraise the effects of therapeutic interventions (e.g., surgery, drugs,

physical training). In addition, exercise testing may be an aid in

objectively evaluating an individual’s functional capacity and tolerance

for stress and may help in choosing an appropriate program of physical

conditioning. Finally, recent work suggests that the exercise response is

of considerable predictive value.

56,57,58,59,60

An abnormal ST segment

response alone results in 10 to 15 fold greater likelihood of developing

some coronary event over the ensuing 3 to 5 years, when compared with

a normal ST response. (See Table 10.1)

Table 10.1 Applications of Exercise Testing

It is apparent that the most significant contributions to the safety of

exercise testing are (1) selecting which patients can be safely tested and

(2) deciding when a test should be terminated (that is, the point at which

sufficient diagnostic information has been obtained, but beyond which

further exercise might result in unnecessary risk to the patient). Both of

these require knowledge of the condition of the patient and of the

exercise procedure and is the reason that most clinical laboratories

require that a physician be present to supervise all exercise tests. Two

basic questions must be answered before exercise testing is initiated.

1. Will the results of testing change future medical management?

2. Do the benefits of testing outweigh the possible risks?

1. Evaluate symptoms

2. Estimate severity of disease

3. Appraise therapy

4. Determine exercise tolerance

5. Choose a program of physical conditioning

6. Predict future cardiac events

Revision C Physician’s Guide 95

2051167-002

Exercise Testing: Concepts Useful in Interpreting Exercise Tests

Concepts Useful in Interpreting Exercise Tests

Percentage of Predicted Maximal Heart Rate (MHR) Achieved

Figures for maximal predicted heart rate, age-adjusted, have been

compiled as ”target” heart rates towards which the subject works. (See

Table 10.2

) Because the exercise heart rate is a good indicator of heart

work, the percent of MHR at which symptoms of electrocardiographic

changes occur is helpful in assessing a person’s degree of disability. Also,

the percent of maximal predicted heart rate that a patient achieves at

peak exercise can provide an estimate of the efficacy of the test. If a

patient must terminate exercise because of limiting noncardiac factors,

the percent of maximum heart rate achieved may help the examiner to

decide if the cardiovascular system was sufficiently stressed for valid

interpretation of the electrocardiographic response to exercise. Thus, a

test in which the maximal heart rate achieved is less than 85 percent of

predicted, in which no abnormal signs or symptoms occur, cannot be

called a normal test but, instead, must be termed an inconclusive test.

Table 10.2 Predicted Maximal Heart Rates*

Total Exercise Time

The total time a subject spends on the treadmill can be used as an

indicator of functional capacity by comparing that performance with age

and gender-matched individuals. If the protocol is kept constant, serial

studies of an individual patient can be of value in detecting decreasing

tolerance to exercise or in evaluating the effects of treatment.

Total METS

Resting oxygen consumption is approximately 3.5 mm/kg/min, or

1 MET. If a subject is tested to his maximum physiological capacity, he

is presumed to have reached his maximum oxygen consumption

(VO2 max). Increasing physical work loads require increasing amounts

of oxygen or multiples of resting oxygen consumption (eg, 2 METS,

3 METS). Because it is known what number of METS are required for

each stage of a particular protocol, the total METS achieved can be an

indicator of work capacity without the use of more sophisticated

equipment to measure expired gases. The usual work capacities of

different clinical subsets are listed in table 10.3.

Age 20-29 30-39 40-49 50-59 60-69

Heart Rate 190 182 179 171 164

(Proceedings of the National Workshop on Exercise in the Prevention, in the Evaluation,

and in the Treatment of Heart Disease. S. C. Med. Assoc., 65:1, 1969)

* Average maximal rates published by 10 investigators.

96 Physician’s Guide Revision C

2051167-002

Exercise Testing: Concepts Useful in Interpreting Exercise Tests

Table 10.3 Work Capacities as they Relate to Clinical Subsets.

Functional Aerobic Impairment

Functional aerobic impairment (FAI) is the difference between the

estimated VO

max and that predicted for age and sex

and can be

calculated from the equation:

If the Bruce protocol is used and the subject is not allowed to bear any

weight on the handrails, his treadmill duration time can be plotted

against his age on a standard nomogram and his functional aerobic

impairment estimated without actually measuring oxygen uptake (see

Figure 10.1). It must be emphasized that the nomograms shown here are

valid only when used with the Bruce protocol, and the treadmill duration

time must be calculated beginning with stage 1 (not including the lesser

stages, zero or one-half).

Percent of Predicted O

Consumption Achieved

Maximum oxygen consumption (VO

max) is the highest level of oxygen

uptake that an exercising subject can achieve; that is, if physical work is

further increased, oxygen consumption will fail to increase, having

reached physiologic limits. Oxygen consumption is limited by cardiac

output and by extraction of O

by the peripheral tissues (arterio-venous

difference). Predicted VO

max for a healthy subject is influenced by age,

gender, body weight, and level of habitual physical activity. The percent

of predicted oxygen consumption that an individual achieves at peak

exercise can be estimated without complicated analysis of expired air.

Since the FAI is the percent reduction in predicted oxygen uptake, the

percent of normal predicted VO

max achieved can be estimated by

subtracting the FAI from 100 percent

. Thus, an individual with

30 percent FAI would have achieved approximately 70 percent of the

max predicted for his age, gender, and level of physical activity. A

person with 0 percent FAI is estimated to have reached 100 percent of

his predicted VO

max.

6 METS or less limited patients

7 to 11 METS asymptomatic patients

12 to 15 METS healthy, active men

16 to 20 METS endurance athletes

FAI

predicted VO

max observed VO

max–()100×

predicted VO

max

---------------------------------------------------------------------------------------------------------- -

Revision C Physician’s Guide 97

2051167-002

Exercise Testing: Concepts Useful in Interpreting Exercise Tests

Figure 10.1 Nomograms for Assessment of Physiologic Impairment

-10

-20

SEDENTARY

ACT

IVE

YEARS OF AGE

WOMEN

FUNCTIONAL AEROBIC IMP

AIRMENT %

MINUTES OF DURATION

(Multistage Treadmill Test)

-10

-20

SEDENTARY

ACTIVE

YEARS OF AGE

MEN

FUNCTIONAL AEROBIC IMPAIRMENT %

MINUTES OF DURATION

(Multistage Treadmill Test)

98 Physician’s Guide Revision C

2051167-002

Exercise Testing: Concepts Useful in Interpreting Exercise Tests

Symptoms and Signs

The healthy individual should be free of symptoms during submaximal

effort; however, at peak exercise, fatigue, leg weakness, or exhaustion is

expected and may be accompanied by mild dizziness or nausea. Chest

pain, claudication, and extreme dyspnea are abnormal responses.

During dynamic exercise, the working muscle groups generate heat

(muscle temperatures as high as 109 °F (43 °C) have been recorded),

which is dissipated through the skin. Therefore, the normal response to

exercise is increased skin temperature and diaphoresis. Healthy

individuals may develop cool, clammy skin at peak exercise, often

associated with a drop in blood pressure. A similar reaction at less than

maximal exercise is abnormal.

Electrocardiographic Response to Exercise

Peaking of the P wave, shortening of the PR segment, downward

displacement of the PR (PQ) junction, shortening of the QT interval,

decrease in R-wave amplitude, and downward displacement of the

J point are all normal responses to exercise. The duration of normal

J-point depression is generally about 40 milliseconds; significant ST

depression that persists beyond that is abnormal. Elevation of the ST

segment at rest owing to the early repolarization phenomenon that

returns to baseline with exercise is a normal response.

The accepted electrocardiogram baseline for use in the evaluation of

changes in the ST segment is the PR (or PQ) junction, which is

somewhat lower than the usual isoelectric line of a resting ECG tracing.

Heart Rate and Blood Pressure

The expected response of the heart rate to exercise is an increasing rate

with increasing workloads and leveling off at maximal or near maximal

performance. The average change in heart rate from resting to peak

exercise (chronotropic reserve) for the healthy subjects participating in

the Seattle Heart Watch was 108 ±14 beats per minute (bpm) for 2532

men and 94 ±15 bpm for 244 women

. Lesser responses of heart rate can

be found in highly trained athletes and infrequently in normal

individuals. In occasional subjects, anxiety at the onset of testing will

cause an initially rapid heart rate that should normalize within the first

3 minutes of exercise, and rise steadily with each increasing workload

thereafter.

Average changes in systolic blood pressure from rest to exercise (intropic

reverse) in the Seattle Heart Watch group was 61 ±19 mmHg for the

healthy men and 42 ±18 mmHg for the healthy women

. Some

individuals will show a slight drop in systolic pressure during the first

stage of exercise, again most likely in response to lessening of anxiety

that has produced hypertension at rest. At maximal effort many subjects

show a drop in systolic pressure, which comes up again when exercise is

terminated, then drops slowly toward resting levels during recovery.

Immediately after maximal or near maximal exercise, a drop in blood

pressure is almost uniformly seen and about 10 percent of subjects have

Revision C Physician’s Guide 99

2051167-002

Exercise Testing: Concepts Useful in Interpreting Exercise Tests

a significant drop with resultant dizziness. This is due to cutaneous

vasodilatation and venous pooling and to a decrease in cardiac output.

Following the immediate drop, blood pressure again rises and, then,

gradually declines toward baseline values.

In response to decreasing systemic vascular resistance with exercise, a

small decrease in diastolic blood pressure (further widening the pulse

pressure) is expected in healthy individuals. Elevation of diastolic blood

pressure greater than 10 mmHg during exercise is considered an

abnormal response.

The heart rate blood pressure product (“double product”) can also be

used as a measure of myocardial oxygen demand. It has been shown that

the “triple product” (heart rate x systolic blood pressure x left

ventricular ejection time) correlated closely with VO

max, which is not

surprising since these factors are the major determinants of myocardial

oxygen demand. By eliminating left ventricular ejection time, a fair

correlation with VO2 max still exists and eliminates the need for the

additional difficult measurement.

100 Physician’s Guide Revision C

2051167-002

Exercise Testing: Exercise Test Interpretation

Exercise Test Interpretation

When interpreting an exercise test, the ST segment response remains

the most important exercise variable. However, abnormalities in the

other exercise variables should be considered, as well. These include: the

presence or absence of chest pain during testing, the responses of heart

rate and blood pressure, changes in the R-wave amplitude, exercise

duration, exercise-induced cardiac murmurs, gallops, rales, etc, and

exercise arrhythmias. A detailed discussion of all these variables is,

obviously, beyond the scope of this reference guide, and the interested

reader is referred to The Clinical Approach to Exercise Testing

another text on exercise testing. However, several of these variables,

including the ST segment response to exercise, will be briefly discussed.

The ST Segment Response

Four types of ST segment responses have been associated with coronary

disease (see Figure 10.2). ST segment elevation, slowly upsloping ST

depression, horizontal ST depression, and downsloping ST depression.

For maximal (and most submaximal) exercise tests, ischemic ST

depression of 1.0 millimeter or more at 80 milliseconds after the J point,

in a horizontal or downsloping configuration, is required to call a test

abnormal.

Figure 10.2 The ST Segment Response to Exercise.

However, most authorities agree that the abnormal ST segment

response is actually a continuum. The least abnormal is the marked

J-junction depression, with the ST segment crossing the baseline at 60 to

80 milliseconds. The next is the slowly upsloping, with J-junction

depression and an upsloping ST segment which is still at least

1 millimeter below the baseline at 80 milliseconds. More abnormal still

is frank ST depression, greater than 1 millimeter depressed at

80 milliseconds with a horizontal configuration. And, most markedly

abnormal is the pattern of ST depression with downsloping configura-

tion. Each pattern has been associated with coronary disease to some

degree and, generally, the likelihood of disease can be estimated by the

severity of the pattern, the time of its onset, and the double product at

ABCDEF

A Minimal J-point depression, isoelectric at 80 milliseconds.

B More marked J-point depression, but also isoelectric at 80 milliseconds.

C Slowly upsloping ST depression, remaining depressed more than 1 millimeter at

80 milliseconds.

D Horizontal ST depression.

E Downsloping ST depression.

F ST elevation.

Revision C Physician’s Guide 101

2051167-002

Exercise Testing: Exercise Test Interpretation

which it occurs. A subject who develops slowly upsloping ST depression

at maximum exercise is less likely to have significant disease than is the

person who develops marked downsloping ST depression at a low work-

load. It has become the custom in many laboratories to not label a test as

positive or negative, but to judge it as normal or abnormal and to relate

the test results to the likelihood of the presence or absence of coronary

disease and the severity of that disease. A typical test interpretation

might read: abnormal test, manifest by 3 millimeters downsloping ST

depression at 65 percent of predicted maximal heart rate, associated

with typical angina and a new S3 gallop. The likelihood of coronary dis-

ease is high, and the likelihood of extensive disease is high.

It should be noted that upsloping ST depression may be seen in one lead,

while horizontal or downsloping depression is present in another lead, so

that multiple lead systems may be valuable in increasing the sensitivity

of this response.

Problems in the Interpretation of ST Segment Responses

to Exercise

Many clinical variables affect the diagnostic reliability of exercise ST

segment responses. Some causes of “false positive” ST changes are:

female gender and left bundle branch block (both discussed below), pre-

excitation syndromes, mitral valve prolapse, digitalis, diuretics, and

some psychotropic drugs. Sometimes masking “true positive” ST

responses are: right bundle branch block (also discussed below), left axis

deviation, previous myocardial infarction, and, again, some psychotropic

drugs. The clinician must be familiar with these modifiers of the ST

segment before attempting to interpret the exercise response of any

individual patient.

Women have a higher frequency of “false positive” ST segment responses

for a number of reasons. Foremost is that in the younger age groups the

likelihood of coronary disease in females is low regardless of the

presenting clinical syndrome. This low prevalence of disease suggests

results in a high false-positive rate, in accordance with Bayes’ Theorem.

Then, too, mitral valve prolapse and vasoregulatory asthenia, both

causes of false-positive ST changes, are common in women. Finally, it

must be emphasized that most of the data thus far accumulated about

exercise variables has been gathered from male subjects, and much more

needs to be learned about the responses of women to exercise.

Most investigators agree that left bundle branch block on resting

electrocardiogram totally precludes the reliable assessment of ST

changes with exercise. Orzan et al

, for instance, evaluated 57

symptomatic subjects with left bundle branch block, who had multilead

treadmill testing and cardiac catheterization. Exercise-induced ST

changes occurred equally in subjects with and without coronary disease,

irrespective of the criteria chosen. In contrast, Tanaka et al

showed no

increase in false-positive ST segment responses in patients with right

bundle branch block, as long as analysis was confined to the lateral

precordial leads (V4, V5, and V6). If ST segment depression occurred in

V1, V2, or V3 only, however, the false-positive rate was high. On the

other hand, there may be a mild reduction in sensitivity when right

bundle branch is present. An additional technical problem may be

encountered when analyzing the ST segment response to exercise in

102 Physician’s Guide Revision C

2051167-002

Exercise Testing: Exercise Test Interpretation

subjects with right bundle branch block. The wide terminal S wave seen

in right bundle branch block may obscure the location of the J point,

rendering quantification of ST segment shifts difficult.

Sensitivity, Specificity, Predictive Value, and Pretest Likelihood

In order to obtain the full benefit of each exercise test, one must have a

thorough knowledge of what is meant by the concepts of sensitivity,

specificity, predictive accuracy (value), and pretest likelihood. The term

“sensitivity” relates to the question: “Given a population with a

particular disease, what percent of patients will have an abnormal test

response?” Specificity relates to the question: “Given a population free

from a particular disease, what percent of patients will have a normal

test response?” Four possibilities exist when a test is performed,

dependent upon whether the individual tested has or does not have the

disease for which he is being tested:

The formulas for calculating the sensitivity and specificity of exercise-

induced depression of the ST segment in subjects with and without

angiographic coronary disease are:

Although useful concepts, these terms are limited clinically by the fact

that they approach the problem in reverse; thus, if we already know that

disease exists, there is little need to perform a diagnostic test.

The real clinical question to be asked is: “Given an ST segment response,

what is the probability that an individual has coronary artery disease?”

This question is addressed by the predictive accuracy (or predictive

value) of a test:

An estimation of the probability of coronary artery disease prior to

exercise stress testing in patients with different chest pain syndromes is

listed in Table 10.4. It shows that the likelihood of obstructive coronary

artery disease in someone with atypical angina is about 50 percent. This

figure is of course subject to modification by the gender, age, associated

risk factors, and the degree to which the angina is atypical. Table 10.4

also shows that a normal exercise test in a patient with atypical chest

Abnormal Test Normal Test

Disease present true positive false negative

Disease absent false positive true negative

sensitivity

true positive

true positive false negative+

---------------------------------------------------------------

specificity

true negative

true negative false positive+

---------------------------------------------------------------

predicted value of an abnormal test

true positive

true positive false positive+

-------------------------------------------------------------

predicted value of a normal test

true negative

true negative false negative+

-----------------------------------------------------------------

Revision C Physician’s Guide 103

2051167-002

Exercise Testing: Exercise Test Interpretation

pain results in a post-test likelihood of coronary disease of 25 percent.

Conversely, an abnormal ST segment response to exercise in such a

patient results in a post-test probability of coronary disease of

88 percent. The calculations in table 10.5 use only the ST segment

response to exercise testing. The occurrence of chest pain during

exercise—even though atypical in nature—or any other abnormal signs

and symptoms would, of course, influence the figures. These figures are

derived from Bayes’ Theorem, which permits calculation of the post-test

probability as it relates to the sensitivity and specificity of the particular

test. Thus, the theorem is useful in assessing the possibility that the test

result represents a false-negative or false-positive finding.

Table 10.4 Post-Test Odds and Probabilities of Disease as Related to Pretest Odds and

Probabilities

Other Exercise Variables

Of the many exercise variables, one of the most important responses to

monitor is blood pressure. The normal response of both systolic and

diastolic pressure has been discussed briefly in a previous section.

A small but significant number of patients undergoing exercise develop

hypotension either as an isolated abnormal finding or associated with

other exercise abnormalities. Recognition of this response is important

not only for the safe conduct of the test, but because of the value of the

sign as a predictor of critical narrowing of the coronary arteries. Morris

and McHenry

reported exercise-related hypotension in 21 of 272

patients with coronary artery disease. Eighty-eight of the 272 patients

had single vessel disease, none of whom had exercise hypotension.

However, 6 of 96 with double vessel disease and 15 of 88 with triple

vessel disease had the abnormal response. Thus, exercise hypotension

becomes more common as coronary disease increases in severity.

A number of investigators began studying R-wave amplitude changes

during treadmill testing after Bonoris et al

reported that a group of

patients with decreased R-wave amplitude after exercising had less

severe coronary artery disease and fewer wall motion abnormalities

when compared to a group of patients who had an increase or no change

in R-wave amplitude. Preliminary studies are conflicting, and it remains

to be seen whether R-wave amplitude changes will prove to be of value

in improving the sensitivity, specificity, or predictive value of the

exercise ECG.

Finally, signs of exercise left ventricular dysfunction (S3 gallop, rales) or

the appearance of ischemic events other than ST segment changes (new

mitral regurgitation murmur, anginal chest pain) during or post-

exercise are valuable in assessing the presence and extent of coronary

disease, and should be sought with each exercise test.

Clinical

Presentation

Pretest

Odds/Prob

Positive Test

Post-Test

Odds/Prob

Negative Test

Post-Test

Odds/Prob

Typical angina 9:1/90% 63:1/98% 9:3/75%

Atypical angina 1:1/50% 7:1/88% 1:3/25%

Asymptomatic 1:9/10% 7:9/44% 1:27/4%

104 Physician’s Guide Revision C

2051167-002

Exercise Testing: Computer Processing of the Exercise ECG

Computer Processing of the Exercise ECG

The interpretation of an exercise test has expanded beyond the

categorization of analysis of the ST segment response using a positive-

negative classification. Despite this fact, the ST segment remains the

most important exercise variable. There are several reasons why

computer-assisted analysis of the exercise ECG is becoming more

important in stress testing. Distortion of the electrocardiographic signal

during maximal exercise due to muscle and respiratory artifact

frequently prevents accurate interpretation of the ST segment response.

The potential that computer processing has for the removal of such

distortions is great (see Figure 10.3). Also, inter- and intra-observer

variations in the interpretation of the ST segment response have been

documented, and computer interpretation removes observer bias from

analysis. Important also is the degree of precision that can be achieved

by computer analysis of the ST segment slope. Classification of normal

and abnormal ECG responses during exercise depend on the accuracy of

the measurements used (that is, ST or J amplitude, ST slope or ST area).

Simoons et al

have noted that the reliability of such measurements

depends on the algorithms for location of the baseline and of the ST

segment itself, usually accomplished by utilizing the QRS complex in

some manner as a fiducial (or reference) point. Comparisons of the

results of different algorithms for detecting the onset and end of the QRS

complex, however, have shown that various programs differ

considerably. Therefore, one should always check the computer

determined QRS onset and offset points before reviewing the ST

segment measurements.

Real time Computer processed

Figure 10.3 Exercise ECG Tracings of a Patient at Maximal Exertion

Simoons et al

have also recently presented a set of recommendations

for the application of computers to diagnostic electrocardiography. (Refer

to Table 10.5) They suggest that the adoption of these guidelines can

result in an increase in sensitivity of 10 to 45 percent with only a small

loss of specificity (3 to 10 percent). Finally, future computer applications

will allow the assessment of other exercise variables and pretest and

post-test probabilities with a greater reliability than is now available.

Revision C Physician’s Guide 105

2051167-002

Exercise Testing: Computer Processing of the Exercise ECG

Table 10.5 Recommendations for Computer-Assisted Interpretation of the Exercise

Electrocardiogram

Hardware

 Adequate frequency response and safety

 Sample frequency 250 Hz or greater

 User interface in the exercise laboratory simple to operate

Signal conditioning

 Computation of averaged or median representative beats after rejection of premature

complexes and other beats with an abnormal configuration

 Warning messages when signal quality is poor

Measurements

 Definitions of onset and end of QRS complex, preferably in multiple, simultaneously

recorded leads

 Measurements of both ST depression and ST elevation

Diagnostic classification

 Special measures should be included to prevent errors in classification due to

excessive noise or baseline shift, as well as certain conditions that may interfere with

the measurement program, such as intraventricular conduction delay

 The whole system should be tested in a series of patients with independently

documented presence or absence of disease

Report

 Graphic representation of representative ECG complexes should be included with

marks that indicate fiducial points from which the measurements have been taken

 Graphic or numerical output of the measurements should be available in addition to

diagnostic statements

106 Physician’s Guide Revision C

2051167-002

Exercise Testing: Computer Processing of the Exercise ECG

For your notes

Revision C Physician’s Guide 107

2051167-002

11 Reports

In-Test and Final Reports

GE systems offer a multitude of reports during (in-test reports) and at

the end (final reports) of the stress test. These reports are the rhythm

report, the comparative medians report, the 4 x 2.5 format 12-lead

report, etc. For descriptions of the reports, consult your operator’s

manual. For two of these reports, some brief additional information

follows:

Linked Medians report – consists of joined or “linked” median

complexes for each lead. The medians are linked to match the current

heart rate. The result is a noise-free and clean ECG strip, readable as

usual. Since the ECG is artificially sequenced with median beats,

rhythm changes, and ectopic beats, for example, are not visible.

Therefore a real-time, one-channel rhythm strip is also provided.

Vector Loops report – Report showing X,Y,Z lead plots and three-

plane vector loops of horizontal, frontal, and sagittal planes using the

Frank X, Y, Z lead set. If the Frank lead set is not applied, GE’s stress

systems derive the X,Y,Z leads from the standard leads using the

Inverse Dower Matrix described in “Vectorcardiogram synthesized from

a 12-lead ECG: Superiority of the inverse dower matrix”.

108 Physician’s Guide Revision C

2051167-002

Reports: In-Test and Final Reports

For your notes

Revision C Physician’s Guide 109

2051167-002

12 Glossary

Abbreviations and Definitions

12SL

An acronym for GE’s 12 simultaneous lead resting ECG analysis

program.

Acquisition module

The unit which serves as the interface between the patient and the

stress system. Electrodes from the patient connect to the acquisition

module, and a cable connects the unit to the system.

Actual load

The patient’s effort to achieve the target load.

AFIB

Atrial fibrillation.

Algorithm

A step-by-step set of instructions for processing data.

Alignment

The purpose of alignment is to adjust the incoming beat to fit with the

median template with respect to the slope characteristics of the

waveform.

Arithmetic averaging

A signal-processing technique that aligns a set number of consecutive

beats and divides them by the number of beats included in the

alignment array yielding a mean QRS.

Artifact

Locally restricted noise or artificial complex not originated by the heart

activity.

Audio assessment

Possibility of GE exercise systems to replay ECG for audible assessment

of exercise tests.

110 Physician’s Guide Revision C

2051167-002

Glossary: Abbreviations and Definitions

Auto relearn

The process whereby GE stress systems detect a significant change in

QRS morphology and relearn both the median complex and the ST

segment measurement points.

BBB

Bundle branch block

Band pass filter

A filter that permits the passage of a specific frequency range (band).

For example, a 5 to 30-Hz band-pass filter would eliminate all

frequencies below 5 Hz and greater than 30 Hz.

Baseline roll filter

Filters on the low end of the ECG frequency spectrum (less than 1 Hz)

whose purpose is to eliminate extraneous baseline drift. AHA

requirements for diagnostic ECG equipment dictate that this filter be no

more aggressive than a 0.05 Hz high-pass filter.

Here, systolic blood pressure.

Bpm

Beats per minute.

Common mode noise

ECG interference generated from environmental equipment that is

common to all leads.

Common mode rejection ratio

A proportion indicating the system’s ability to ward off the effects of

environmental electrical noise.

Correlation

Correlation matches the shape (slope characteristics) of the incoming

beat to the median template. Where there is a sufficient correlation, the

incremental update of the median template proceeds.

Chronotropic response

See “HR reserve used”.

Cubic spline

A third-order polynomial technique for control of baseline drift employed

as an alternative to aggressive, potentially distorting filterings.

Cyclic artifacts

Artifacts caused by footfalls during treadmill testing or pedaling during

bicycle testing.

Revision C Physician’s Guide 111

2051167-002

Glossary: Abbreviations and Definitions

Double Product

The product of heart rate and systolic blood pressure, also called Rate

Pressure Product (RPP).

DTS

Duke Treadmill Score, calculated with test time, ST deviation, and

angina. Risk predictor.

Dominant template

See “Median complex”.

Dynamic ST scan

The leads are scanned dynamically for the worst case ST depression.

E point

The isoelectric reference point set in the PR segment.

ESC

Ventricular escape beat.

Exercise capacity

Ability of a patient to exercise.

Exercise test

Application of physical stress to a patient using treadmill, bicycle, drugs,

etc.

Exercise Test Interpretation

Part of HEART-Exercise software package. Translates measurements,

thresholds, and patient data to interpretation and reasoning texts.

Fiducial point

A landmark or reference point.

Filtering

With respect to electrocardiography, filtering refers to the extraction

and elimination of certain frequencies from the raw signal.

FRF algorithm

Algorithm for reducing the artifacts in the ECG with much less

distortion of the QRS complexes.

FVE recovery

Ventricular ectopics per minute in recovery phase.

Graphical XTI output

Graphical expression of exercise test results.

112 Physician’s Guide Revision C

2051167-002

Glossary: Abbreviations and Definitions

Hertz (Hz)

The labeling unit of a frequency in terms of cycles per second. Also

abbreviated as cps.

HEART-Exercise

The name for GE’s 12 simultaneous lead exercise ECG analysis

program. Contains the functionality described in this physician’s guide.

HR reserve used

Patient’s heart rate response to exercise calculated as

(HR

peak

- HR

resting

)/(220 - age - HR

resting

HR recovery

Decrease of the heart rate in first minute of recovery.

High-pass filter

A filter that permits the passage of frequencies above a specific value.

For example, a 0.01 Hz high-pass filter allows the passage of all

frequencies above 0.01 Hz while removing all those below.

Incremental updating

Signal-processing technique used by GE stress systems resulting in a

median complex. Successfully correlated incoming beats update the

median template by the smaller fixed increment or a fraction of the

difference between the template and the incoming beat.

Intelligent Lead Switch

Automatic selection of the best leads for QRS detection.

J point

The end of the QRS complex as delineated by the location where the last

steep slopes of depolarization are replaced by the more or less flat ST

segment.

LBBB

Left Bundle Branch Block.

Low-end filter

The high-pass filter that works on the lower end of the frequency

spectrum.

Low-pass filter

A filter that permits the passage of frequencies below a specific value.

For example, a 40-Hz low-pass filter allows the passage of all

frequencies below 40 Hz while removing all those above.

Mean complex

The resultant complex from the arithmetic averaging technique of signal

processing.

Revision C Physician’s Guide 113

2051167-002

Glossary: Abbreviations and Definitions

Median complex

The resultant complex from the incremental updating technique of

signal processing.

Median template

See “Median complex”.

Median update

See “Incremental updating”.

MET

Metabolic equivalent or exercise capacity. METS is the plural.

Multi-beat event

Consecutive single-beat events. For example, a bigeminy, consisting of

pairs of PVCs and normal beats.

Noise

All kinds of signal, disturbing the ECG, mostly caused by muscle

activities.

Noise filter

A user-selectable low-pass filter that can be set at either 20, 40 or

100 Hz.

Pace enhance

Improved display of pace pulses.

Post-J measurement point (J+x)

Point for measurement of the ST level.

PSVC

Premature Supraventricular Complex.

PVC

Premature ventricular complex

PWC

Physical working capacity. Load in watts devided by patient’s weight.

PWC 130, PWC 150, PWC 170

Physical working capacity at heart rate 130 bpm, 150 bpm, or 170 bpm.

QRS detection

The purpose of this step is to detect electrical impulses associated with a

heart beat of sinus or ectopic origin.

QRS offset

See “J point”.

114 Physician’s Guide Revision C

2051167-002

Glossary: Abbreviations and Definitions

QRS onset

First deflection of the QRS complex of all leads.

Relearn

If GE stress systems detect a significant change in QRS morphology or a

specific user command, then the incoming complex is established as the

dominant template and the ST segment measurement points are

recalculated.

Risk prediction

Prediction of morbidity, overall mortality, cardiovascular death, acute

myocardial infarction and/or malignant arrhythmias, based on data

obtained by exercise testing.

RPP (Rate Pressure Product)

See “Double Product”.

RR interval

Time distance to precedent beat.

Rule

Provides an interpretation by combining terms, consisting of

measurements and thresholds.

Single-beat event

Events dedicated to single complexes. Such events are PVC or SVPC, for

instance.

Signal acquisition

The process of obtaining the analog ECG signal from the patient and

converting it into a digital format. In GE systems this process is

performed in the acquisition module.

ST depression or ST segment depression

ST segment which is below the isoelectric line.

SVT

Supraventricular tachycardia

Signal conditioning

An enhancement of the ECG signal whose purpose is to present clean

waveform data by improving the signal-to-noise ratio. See also

“Filtering” and “Signal processing”.

Signal processing

The employment of mathematical algorithms to improve the ECG

presentation. See also “Incremental updating” and “Arithmetic

averaging”.

Revision C Physician’s Guide 115

2051167-002

Glossary: Abbreviations and Definitions

Signal-to-noise ratio

The amplitude of the ECG signal in proportion to the amplitude of the

noise in the signal.

Stress test

See “Exercise test”.

ST criteria

Possibility to set ST criteria for continuous monitoring of the

ST segment.

ST/HR hysteresis

Value derived from ST levels from exercise and recovery phase.

ST/HR index

Calculated by dividing the change of the ST depression from baseline to

maximum exercise by the change in heart rate over the same time

period.

ST/HR slope

Slope of a regression line through ST/HR points, beginning at peak

exercise and extending backward through at least three points until

significance is obtained.

ST Index

Value calculated with ST depression and ST down slope.

ST Integral

Area of ST segment.

ST level

Difference between an amplitude in the ST segment and the amplitude

before the QRS onset.

ST segment

ECG segment from QRS offset to T-wave onset.

ST slope

Slope of the ST segment.

Template

See “Median complex”.

Target load

Load in Watts the patient should achieve on bicycle.

TWA

T-wave alternans. Alternating shapes of the ST/T segments.

116 Physician’s Guide Revision C

2051167-002

Glossary: Abbreviations and Definitions

Ventricular Ectopic (PVC and ESC).

Ventricular tachycardia

XTI (eXercise Test Interpretation)

See “Exercise Test Interpretation”.

Revision C Physician’s Guide 117

2051167-002

13 Bibliography

Relevant Literature

1 American Heart Associations Committee on Electrocardiography: Recommendations

for standardization of leads and specifications for instruments in ECG/VCG.

Circulation 52:11, 1975.

2 Froelicher VF, Myers J, Follansbee WP, Labovitz AJ: Exercise and the Heart –

Edition.1993 by Mosby – Year Book, Inc.

3 Taback, L., et al: Digital recording of electrocardiographic data for analysis by digital

computer. Medical Electronics. 6:167, 1959.

4 Blackburn H (ed): Measurement in Exercise Electrocardiography: The Ernst

Simonson Conference. Charles C. Thomas Publisher, Springfield 1969.

5 Berson AS, Pipberger: The low-frequency response of electrocardiographs, a

frequent source of recording errors. American Heart Journal. 71:779, 1966.

6 Ellestad M.H.: Stress Testing: Principles and Practice – 4

Edition . F.A. Davis

Company, Philadelphia, 1996.

7 Meyer CR, Keiser HN: Electrocardiogram baseline noise estimation and removal

using cubic-splines and state-space computation techniques. Computers and

Biomedical Research 10:459, 1977.

8 Tayler, D.I., Vincent, R: Artifactual ST segment abnormalities due to

electrocardiographic design. British Heart Journal, 554:121, 1985.

9 Daniel B.Mark. An overview of risk assessment in coronary artery disease. The

American Journal of Cardiology, 73:19B–25B, March 1994.

10 P.M.Rautaharju, J.W.Warren, and H.P.Calhoun, Estimation of qt prolongation,

Journal of Electrocardiology, 23:111–117, April 1990.

11 Daniel.B.Mark, Linda Shaw, F.E.Harrell, M.A.Hlatky, K.L.Lee,J.R.Bengton,

C.B.McCants, R.M.Califf, and D.B.Pryor. Prognostic value of a treadmill exercise

score in outpatients with suspected coronaryartery disease. The New England

Journal of Medicine,325(12):849–853, Sept. 1991.

12 Daniel.B.Mark, M.A.Hlatky, F.E.Harrell, K.L.Lee, J.R.Bengton, R.M.Califf, and

D.B.Pryor. Exercise treadmill score for predicting prognosis in coronary artery

disease. Annals of Internal Medicine, 106(6):793–800, June 1987.

13 Willi.Kaiser and Martin Findeis. Artifact Processing during Exercise Testing, Journal

of Electrocardiology, Vol 32, Supplement,1999

14 Willi Kaiser and Martin Findeis, Novel Signal Processing Methods for Exercise ECG,

International Journal of Bioelectromagnetism, No.1, Vol.2, 2000.

15 Ameisen O, Okin PM, Devereux RB, Hochreiter C, Miller DH, Zullo MA, Borer JS,

Kligfield P: Predictive value and limitations of ST/HR slope. Br Heart J 1985; 53:547–

551.

16 Elamin MS, Mary DASG, Smith DR, Linden RJ: Prediction of severity of coronary

artery disease using slope of submaximal ST segment/heart rate relationship.

Cardiovasc Res 1980; 14:681–691.

118 Physician’s Guide Revision C

2051167-002

Bibliography: Relevant Literature

17 Elamin MS, Boyle R, Kardash MM, Smith DR, Stoker JB, Whitaker W, Mary DASG,

Linden RJ: Accurate detection of coronary heart disease by new exercise test. Br

Heart J 1982; 48:311–320.

18 Kligfield P, Ameisen O, Okin PM: Heart Rate Adjustment of ST Segment Depression

for Improved Detection of Coronary Artery Disease. Circulation 1989; 79:245–255.

19 Kligfield P, Okin PM, Ameisen O, Wallis J, Borer JS: Correlation of the exercise

ST/HR slope with anatomic and radionuclide cineangiographic findings in stable

angina pectoris. Am J Cardiol 1985: 56:418–421.

20 Kligfield P, Okin PM, Ameisen O, Borer JS: Evaluation of coronary artery disease by

an improved method of exercise electrocardiography: The ST segment/heart rate

slope. Am Heart J 1986; 112:589–598.

21 Okin PM, Kligfield P, Ameisen O, Goldberg HL, Borer JS: Improved accuracy of the

exercise electrocardiogram: Identification of three-vessel coronary disease in stable

angina pectoris by analysis of peak rate related changes in ST segments. Am J

Cardiol 1985; 55:271–276.

22 Okin PM, Ameisen O, Kligfield P: A modified treadmill exercise protocol for computer-

assisted analysis of the ST segment/heart rate slope: Methods and reproducibility. J

Electrocardiol 1986; 19:311–318.

23 Okin PM, Kligfield P, Ameisen O, Goldberg HL, Borer JS: Identification of

anatomically extensive coronary artery disease by the exercise ST segment/heart

rate slope. Am Heart J 1988; 115:1002–1012.

24 Paul Kligfield, Principles of simple heart rate adjustment

of ST segment depression during exercise electrocardiography, Cardiology Journal

2008, Vol. 15, No. 2, pp. 194–200.

25 Okin PM, Ameisen O, and Kligfield P. Recovery-Phase Patterns of ST Segment

Depression in the Heart Rate Domain. Identification of Coronary Artery Disease by

the Rate-Recovery Loop.Circulation Vol 80, No 3, September 1989

26 Bruce D.Nearing and Richard L.Verrier. System and method for quantifying

alternation in an electrocardiogram signal. Patent No.: US 6,169,919 B1, Jan. 2,

2001.

27 Antonis A. Armoundas and Richard J.Cohen. Clinical utility of t-wave

alternans.Cardiac Electrophysiology Review, 3:390–394, 1997.

28 Bruce D.Nearing and Richard L.Verrier. Modified moving average analysis of T-wave

alternans to predict ventricular fibrillation with high accuracy. J. Appl Physiol. Vol 92,

541–549, 200.

29 Paul Albrecht, Jeff Arnold, Srivat Krishnamachari, Richard Cohen. Exercise

Recordings for the Detection of T Wave Alternans. J. Electrocardiography, Vol 29,

Supplement, 46–51.

30 Daniel M. Bloomfield, Stefan H. Hohnloser, Richard Cohen. Interpretation and

Classification of Microvolt T Wave Alternans Tests. J. of Cardiovascular

Electrophysiology, Vol 13, No 5, 502–514.

31 D.B.Mark, L. Shaw, F.E.Harrell, M.A.Hlatky, K.L.Lee, J.R.Bengton, C.B.McCants,

R.M.Califf, and D.B.Pryor. Prognostic value of a treadmill exercise score in

outpatients with suspected coronary artery disease. N Engl J Med 1991; 325:849-

53.

General Electric Company.

33 C.R.Cole, E.H.Blackstone, F.J.Pashkow, C.E.Snader, and M.S.Lauer, Heart-Rate

Recovery Immediately After Exercise As a Predictor of Mortality, N Engl J Med 1999;

341:1351-7.

34 J.P.Frolkis, C.E.Pothier, E.H.Blackstone, and M.S.Lauer, Frequent Ventricular

Ectopy after Exercise as a Predictor of Death, N Engl J Med 2003; 348:781-90.

Revision C Physician’s Guide 119

2051167-002

Bibliography: Relevant Literature

35 M.S.Lauer, Exercise Testing, Part 1: Looking Beyond the ST Segment. Cardiology

Rounds, Brigham and Women’s Hospital, Boston, Massachusetts, May 2002 Volume

6, Issue 5.

36 M.S.Lauer, Exercise Testing, Part 2: The Value of Heart Rate Recovery. Cardiology

Rounds, Brigham and Women’s Hospital, Boston, Massachusetts, May 2002 Volume

6, Issue 6.

37 R.Lehtinen, H.Sievänen, J.Viik, V.Turjanmaa, K Niemelä and J.Malmivuo. Accurate

Detection of Coronary Artery Disease by Integrated Analysis of the ST-Segment

Depression/Heart Rate Patterns During Exercise and Recovery Phases of the

Exercise Electrocardiography Test. Am J Cardiol 1996; 78:1002–1006.

38 Lauer MS, Okin PM, Larson MG, Evans JC, Levy D. Impaired heart rate response to

graded exercise. Prognostic implications of chronotropic incompetence in the

Framingham Heart Study. Circulation 1996; 93:1520-6.

39 Gibbons RJ, Balady GJ, Bricker JT, Chaitman BR, Fletcher GF, Froelicher VF, Mark

DB, McCallister BD, Mooss AN, O’Reilly MG, Winters WL Jr. ACC/AHA 2002

guideline update for exercise testing: a report of the American College of

Cardiology/American Heart Association Task Force on Practice Guidelines

(Committee on Exercise Testing). 2002.

40 Rami Lehtinen, Jari Viik, Willi Kaiser, Mari Niemi, Tiit Kööbi, Kari Niemelä, Väinö

Turjanmaa: Accurate Detection of Coronary Artery Disease by ST-segment/Heart

Rate Hysteresis of the Exercise Electrocardiography Test. Journal of the American

College of Cardiology, vol 45, issue 3, suppl 1, page A158.

41 Paul Kligfield and Michael S. Lauer: Exercise Electrocardiogram Testing: Beyond the

ST Segment. Circulation 2006; 114; 2070–2082.

42 Tuomo Nieminen, Terho Lehtimäki, Jari Viik, Rami Lehtinen, Kjell Nikus, Tiit Kööbi,

Kari Niemelä, Väinö Turjanmaa, Willi Kaiser, Heini Huhtala, Richard L. Verrier, Heikki

Huikuri, Mika Kähönen: T-wave alternans predicts mortality in a population

undergoing clinically indicated exercise test, European Heart Journal,

doi:10.1093/eurheartj/ehm271, July 25, 2007.

43 David A. Driggers; David Marchant, Maximizing the exercise stress test, Critical

factors that enhance its validity, VOL 105 / NO 5 / MAY 1, 1999 / POSTGRADUATE

MEDICINE Table 3.

44 Svart K, et al, Exercise electrocardiography detection of coronary artery disease by

ST-segment depression/heart rate hysteresis in women: The Finnish Cardiovascular

Study, Int J Cardiol (2008), doi:10.1016/j.ijcard.2008.11.038.

45 Mikko Minkkinen, et. al., Enhanced Predictive Power of Quantitative TWA during

Routine Exercise Testing in the Finnish Cardiovascular Study, J Cardiovasc

Electrophysiol, Vol. 20, pp. 1–8.

46 Minkkinen, Mikko, Nieminen, Tuomo, Verrier, Richard L., Leino, Johanna, Lehtimäki,

Terho, Viik, Jari, Lehtinen,Rami, Nikus, Kjell, Kööbi, Tiit, Turjanmaa, Väinö and

Kähönen, Mika (2009) ‘Impaired exercise capacity predicts sudden cardiac death in a

low-risk population: enhanced specificity with heightened T-wave alternans’, Annals

of Medicin: DOI: 10.1080/07853890902802971

47 Wassermann et. al., Principles of Exercise Testing and Interpretation, DSc.

Philadelphia, PA: Lippincott Williams & Wilkins, 2005.

48 Testing and reporting performance results of cardiac rhythm and ST segment

measurement algorithms. ANSI/AAMI EC57:1998/(R)2003.

49 Guidelines for Exercise Prescription and Testing, 3rd edition, American College of

Sports Medicine’s 1986.

50 D.J.Mertens et.al, A simple formula for the estimation of maximal oxygen intake

during cycle ergometry. European Heart Journal (1994) 15, 1247–1251.

120 Physician’s Guide Revision C

2051167-002

Bibliography: Relevant Literature

51 Whipp Bj, Mahler M. Dynamics of pulmonary gas exchange during exercise. In:

West, J.B. (ed.) Pulmonary Gas Exchange. I. Ventilation, Bloodflow and Diffusion.

II. Organism and Environment. New York: Academic Press, 1980; 2:33–96

52 Wonisch M, Berent R, Klicpera M, Laimer H, Marko C, Schwann H, Schmid P,

Praxisleitlinien Ergometrie, Journal für Kardiologie 2008; 15, (Supplementum A –

Praxisleitlinien Ergometrie), 3-17.

53 GE Healthcare, Marquette™ 12SL™ ECG Analysis Program Physician’s Guide,

416791-004 Revision E.

54 International Classification of Diseases (ICD),

http://www.who.int/classifications/icd/en/.

55 Die Fahrradergometrie in der Praxis, Rost R, Hech H, Hollmann, Institut für

Sportmedizin und Kreislaufforschung der Deutschen Sporthochschule Köln, Bayer.

56 Arnow WS, Cassidy J: Five year follow-up of double Master’s test, maximal treadmill

stress test, and resting and postexercise apexcardiogram in asymptomatic persons.

Circulation 52:616, 1975.

57 Bruce RA, DeRouen TA: Exercise testing as a predictor of the heart disease and

sudden death. Hosp Pract 69, Sept., 1978.

58 Bruce RA, McDonough Jr: Stress testing in screening for cardiovascular disease. Bull

NY Acad Med 45:1288, 1969.

59 Cumming CR, Samm J, Borysyk, L, et al: Electrocardiographic changes during

exercise in asymptomatic men: 3 year follow up. Can Med Assoc J 112:578, 1975.

60 Froelicher VF Jr, Thomas MM, Pillow C, et al: Epidemiologic study of asymptomatic

men screened by maximal treadmill testing for latent coronary artery disease. Am J

Cardiol 34:770, 1974.

61 Proceedings of the National Workshop on Exercise in the Prevention, in the

Evaluation, and in the Treatment of Heart Disease. Myrtle Beach, SC, 1969. J Carol

Med Assoc (Suppl) 65 (1 to No. 12):1, 1969.

62 Fox SM, Naughton JP, Gorman PA: Physical activity and cardiovascular health. III.

The exercise prescription: frequency and type of activity. Mod Concepts Cardiovasc

Dis 41:25, 1972.

63 Bruce RA: Progress in exercise cardiology. In Yu PN, Goodwin JF (eds): Progress in

Cardiology, Vol 3. Philadelphia, Lea & Febiger, 1974, p 128.

64 Irving JB, Bruce RA, DeRouen TA: Variations in and significance of systolic pressure

during maximal exercise (treadmill) testing. Am J Cardiol 39:841, 1977.

65 Glasser SP, Clark PL: The Clinical Approach to Exercise Testing. Hagerstown,

Harper & Row, 1980.

66 Orzan F, Garcia E, Mather VS, et al: Is the treadmill exercise useful for evaluating

coronary artery disease in patients with complete left bundle branch block? Am J

Cardiol 42:36, 1978.

67 Tanaka T, Friedman MJ, Okada RD, et al: Diagnostic value of exercise-induced S-T

segment depression in patients with right bundle branch block. Am J Cardiol 41:670,

1978.

68 Morris SN, McHenry PL: The incidence and significance of exercise induced

hypotension (abstr). Am J Cardiol 39:289, 1977.

69 Bonoris PE, Greenberg PS, Castellanet MJ, et al: Significance of changes in R wave

amplitude during treadmill stress testing; angiographic correlation. Am J Cardiol

41:846, 1978.

70 Simoons ML, Hugenholtz PG: Estimation of the probability of exercise-induced

ischemia by quantitative ECG analysis.

Circulation 56:552, 1977.

71 Simoons ML, Hugenholtz PG, Ascoop CA, et al: Quantitation of exercise

electrocardiography. Circulation 63(3):471, 1981.

Revision C Physician’s Guide 121

2051167-002

Bibliography: Relevant Literature

72 Ashley E, Myers J, New Insights into the Clinical Exercise Test, ACSM’s Certified

News, Vol. 13, No. 4, October–December 2003

73 Håkan Kronander, Niklas Hammar, Werner Fischer-Colbrie, Jacek Nowak, Lars-Åke

Brodin, Håkan Elmqvist. Analysis of ST/HR hysteresis improves long-term prognostic

value of exercise ECG test. International Journal of Cardiology 148(2011)64–69

74 Diagnostic electrocardiographic devices, ANSI/AAMI EC11:1991/(R)2001/(R)2007

75 Kai P. Savonen, Timo A. Lakka, Jari A. Laukkanen, Tuomas H. Rauramaa, Jukka T.

Salonen, and Rainer Rauramaa, Usefulness of Chronotropic Incompetence in

Response to Exercise as a Predictor of Myocardial Infarction in Middle-Aged Men

Without Cardiovascular Disease, Am J Cardiol 2008; 101:992–998.

76 David S.Rosenbaum, Lance E.Jackson, Joseph M.Smith, Hasan Garan, Jeremy

Ruskin, and Richard J.Cohen. Electrical alternans and vulnerability to ventricular

arrhythmias. The New England Journal of Medicine, 330(4):235–241, January 27

1994.

77 Richard L. Verrier. Physiology of electrical alternans. Cardiac Electrophysiology

Review, 3:381 -386, 1997.

78 Vectorcardiogram synthesized from a 12-lead ECG: Superiority of the inverse dower

matrix. Lars Edenbrandt and Olle Pahlm., Journal of Electrocardiology, 4(21):361–

367, 1988.

79 Sutter JD et al. Chronotropic incompetence: are the carotid arteries to blame?

European Heart Journal (2006) 27, 897–898.

80 Heinsimer JA, Irwin JM, Basnight LL. Influence of underlying coronary artery disease

on the natural history and prognosis of exercise-induced left bundle branch block. Am

J Cardiol. 1987 Nov 1;60(13):1065-7.

81 Mathew S. Maurer, Elliot A. Shefrin and Jerome L. Fleg, Prevalence and prognostic

significance of exercise-induced supraventricular tachycardia in apparently healthy

volunteers. Am J Cardiol 1995;75:788-792.

82 Tuomo Nieminen, Rami Lehtinen, Jari Viik, Terho Lehtimäki, Kari Niemelä, Kjell

Nikus, Mari Niemi, Janne Kallio, Tiit Kööbi, Väinö Turjanmaa, and Mika Kähönen.

The Finnish cardiovascular study (fincavas): characterising patients with high risk of

cardiovascular morbidity and mortality. BMC Cardiovascular Disorders, 6:2-8, 2006.

83 Richard L. Verrier, Thomas Klingenheben, Marek Malik, Nabil El-Sherif, Derek V.

Exner, Stefan H. Hohnloser, Takanori Ikeda, Juan Pablo Martinez, Sanjiv M.

Narayan, Tuomo Nieminen, David S. Rosenbaum. Microvolt T-Wave Alternans

Physiological Basis, Methods of Measurement, and Clinical Utility—Consensus

Guideline by International Society for Holter and Noninvasive Electrocardiology.

J Am Coll Cardiol 2011;58:1309–24.

84 Sae Young Jae, Bo Fernhall, Kevin S. Heffernan, Mira Kang, Moon-Kyu Lee, Yoon-

Ho Choi, and Won Hah Park. Chronotropic response to exercise testing is associated

with carotid atherosclerosis in healthy middle-aged men. European Heart Journal

(2006) 27, 954–959.

85 Willi Kaiser, Martin Findeis, Rami Lehtinen, Terho Lehtimäki, Jari Viik. Exercise Test

Interpretation. Computing in Cardiology 2010;37:769−772.

86 2062179-002 Audio-ECG Physician’s guide to Audio-ECG Assessment of Stress

Tests.

87 K Rahimia, A Thomas, M Adam, B Hayerizadeh, G Schuler and M Secknus.

Implications of exercise test modality on modern prognostic markers in patients with

known or suspected coronary artery disease: Treadmill versus bicycle. European

Journal of Cardiovascular Prevention and Rehabilitation 2006, 13:45–50.

88 Jule N.Wetherbee, Virinderjit S. Bamrah, Michael J. Ptacin, John H. Kalbfleisch.

Comparison of ST Segment Depression in Upright Treadmill and Supine Exercise

Testing. JACC Vol.11 No2 February 1988.

122 Physician’s Guide Revision C

2051167-002

Bibliography: Relevant Literature

89 Myers, J., et al., Exercise capacity and mortality among men referred for exercise

testing. The New England Journal of Medicine, 2002. 346(11): p. 793-801.

90 Zimarino M, Barnabei L, Madonna R, Palmieri G, Radico F, Tatasciore A, Bellisarii FI,

Perrucci GM, Corazzini A, De Caterina R. Institute of Cardiology and Center of

Excellence on Aging, "G. d'Annunzio" University-Chieti Italy, Italy. A comparison of

the diagnostic performance of the ST/HR hysteresis with cardiopulmonary stress

testing parameters in detecting exercise-induced myocardial ischemia. Int J Cardiol.

2012 Dec 20.

91 Diamond, G.A. and S. Kaul, Evidence-based financial incentives for healthcare

reform: putting it together. Circulation. Cardiovascular quality and outcomes, 2009.

2(2): p. 134-40.

92 Simoons, M.L. and S. Windecker, Controversies in cardiovascular medicine: Chronic

stable coronary artery disease: drugs vs. revascularization. European Heart Journal,

2010. 31(5): p. 530-41.

93 Topol, E.J. and S.E. Nissen, Our preoccupation with coronary luminology. The

dissociation between clinical and angiographic findings in ischemic heart disease.

Circulation, 1995. 92(8): p. 2333-42.

94 Libby, P., P.M. Ridker, and A. Maseri, Inflammation and atherosclerosis. Circulation,

2002. 105(9): p. 1135-43.

95 Roger, V.L., et al., Heart disease and stroke statistics--2012 update: a report from the

american heart association. Circulation, 2012. 125(1): p. e2-e220.

96 Badimon, L., T. Padró, and G. Vilahur, Atherosclerosis, platelets and thrombosis in

acute ischaemic heart disease. European Heart Journal: Acute Cardiovascular Care,

2012. 1(1): p. 60-74.

97 Maseri, A. and V. Fuster, Is there a vulnerable plaque? Circulation, 2003. 107(16): p.

2068-71.

98 Matter, C.M., M. Stuber, and M. Nahrendorf, Imaging of the unstable plaque: how far

have we got? European Heart Journal, 2009. 30(21): p. 2566-74.

99 Atwater, B.D., J. Oujiri, and M.R. Wolff, The immediate impact of the Clinical

Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE)

trial on the management of stable angina. Clinical Cardiology, 2009. 32(8): p. E1-3.

100 Kereiakes, D.J., Ischemia is the critical determinant of revascularization benefit: an

interventionalist's perspective of the COURAGE trial. Reviews in cardiovascular

medicine, 2009. 10 Suppl 2: p. S45-52.

101 Stergiopoulos K, B.D.L., Initial coronary stent implantation with medical therapy vs

medical therapy alone for stable coronary artery disease: Meta-analysis of

randomized controlled trials. Archives of Internal Medicine, 2012. 172(4): p. 312-319.

102 Pursnani, S., et al., Percutaneous coronary intervention versus optimal medical

therapy in stable coronary artery disease: a systematic review and meta-analysis of

randomized clinical trials. Circulation. Cardiovascular interventions, 2012. 5(4): p.

476-90.

103 Weintraub, W.S., et al., Cost-effectiveness of percutaneous coronary intervention in

optimally treated stable coronary patients. Circulation. Cardiovascular quality and

outcomes, 2008. 1(1): p. 12-20.

104 Denktas, A.E., et al., Reduced-dose fibrinolytic acceleration of ST-segment elevation

myocardial infarction treatment coupled with urgent percutaneous coronary

intervention compared to primary percutaneous coronary intervention alone results of

the AMICO (Alliance for Myocardial Infarction Care Optimization) Registry. JACC

Cardiovasc Interv, 2008. 1(5): p. 504-10.

105 Kent, D.M., J. Lau, and H.P. Selker, Balancing the benefits of primary angioplasty

against the benefits of thrombolytic therapy for acute myocardial infarction: the

importance of timing. Eff Clin Pract, 2001. 4(5): p. 214-20.

Revision C Physician’s Guide 123

2051167-002

Bibliography: Relevant Literature

106 Tarantini, G., et al., Acceptable reperfusion delay to prefer primary angioplasty over

fibrin-specific thrombolytic therapy is affected (mainly) by the patient's mortality risk:

1 h does not fit all. Eur Heart J, 2010. 31(6): p. 676-83.

107 Terkelsen, C.J., et al., Primary PCI as the preferred reperfusion therapy in STEMI: it

is a matter of time. Heart, 2009. 95(5): p. 362-9.

108 Lin, G.A., et al., Frequency of stress testing to document ischemia prior to elective

percutaneous coronary intervention. JAMA: the journal of the American Medical

Association, 2008. 300(15): p. 1765-73.

109 Cassar, A., et al., Chronic coronary artery disease: diagnosis and management.

Mayo Clinic proceedings. Mayo Clinic, 2009. 84(12): p. 1130-46.

110 Chan, P.S., et al., Appropriateness of percutaneous coronary intervention. JAMA: the

journal of the American Medical Association, 2011. 306(1): p. 53-61.

GE Medical Systems

Information Technologies, Inc.

8200 West Tower Avenue

Milwaukee, WI 53223 USA

Tel: +1 414 355 5000

1 800 558 7044 (US only)

1 800 668 0732 (Canada only)

Fax: +1 414 355 3790

www.gehealthcare.com

GE Medical Systems

Information Technologies GmbH

Munzinger Straße 5

79111 Freiburg

GERMANY

Tel: +49 761 4543-0

Fax: +49 761 4543-233

Asia Headquarters

GE (China) Co.,Ltd.

No1 Huatuo Road,

Zhangjiang Hi-Tech Park Pudong,

Shanghai, P.R.China 201203

Tel: +86 21 38777888

Fax: +86 21 38777402