15 Electrocardiographic Monitoring
As late as 1970, the electrocardiographic monitor was not considered an integral part of monitoring strategies in the perioperative period. As a matter of fact, luminaries in the specialty regarded the use of ECG monitoring as “questionable value because of possible iatrogenic problems.” Concern also was expressed that anesthesiologists’ attention would be diverted from the patient.1 Currently, monitoring the ECG is a fundamental standard of monitoring of the American Society of Anesthesiologists.2 Despite the introduction of more sophisticated cardiovascular monitors such as the pulmonary artery catheter and echocardiography, the electrocardiogram (ECG; coupled with blood pressure measurement) serves as the foundation for guiding cardiovascular therapeutic interventions in the majority of anesthetics.3 It is indispensable for diagnosing arrhythmias, acute coronary syndromes, electrolyte abnormalities, (particularly of serum potassium and calcium), and some forms of genetically mediated electrical or structural cardiac abnormalities (e.g., Brugada syndrome; Table 15-1).4
TABLE 15-1 Basic Clinical Information Available from Electrocardiography
| Anatomy or Morphology |
| Physiology |
One of the most important changes in electrocardiography that has occurred recently is the widespread use of computerized systems for recording ECGs. Bedside units are capable of recording diagnostic quality 12-lead ECGs that can be transmitted over a hospital network, for storage and retrieval. Most of the ECGs in the United States are recorded by digital, automated devices, equipped with software, which can measure ECG intervals and amplitudes and provide a virtually instantaneous interpretation. However, different automated systems may have different technical specifications that can result in significant differences in the measurement of amplitudes, intervals, and diagnostic statements.5,6 The diagnostic specificity and sensitivity of the ECG to diagnose a particular abnormality also are not consistent. For example, finite limits are defined by the relation between sensitivity and specificity (usually inversely related) for detecting obstructive coronary artery disease (CAD). During exercise testing, the 12-lead ECG has a mean sensitivity of only 68% and a specificity of 77%.7,8 The resting 12-lead ECG is even less sensitive and specific.8 More complex ECG modalities are likely to improve its utility in the future (high-frequency QRS signal averaging). In this chapter, the theory and the operating characteristics of ECG hardware used in the perioperative period are presented to facilitate proper use and interpretation of monitoring data.
Historical perspective
An extensive review of the history of electrocardiography is beyond the scope of this chapter. However, several excellent reviews were published in honor of the centennial of the first recording of the human ECG.9–14 Willem Einthoven is universally considered the father of electrocardiography (for which he won the 1924 Nobel Prize for Medicine/Physiology). Many of the basic clinical abnormalities in electrocardio-graphy were first described using the string galvanometer (e.g., bundle branch block, delta waves, ST-T changes with angina). It was used until the 1930s, when it was replaced by a system using vacuum tube amplifiers and a cathode ray oscilloscope. With advancements in electrical engineering technology, the devices became more compact, portable, and user friendly. In the 1950s, a portable direct-writing ECG cart was introduced. The first analog-to-digital (A/D) conversion systems for the ECG were introduced in the early 1960s, although their off-line use was impractical and restricted until the late 1970s. In the 1980s, microcomputer technology became widely available and is now standard for all diagnostic and monitoring systems. Further improvements in hardware and software design led to the development of automated ST-analysis algorithms and their use in routine clinical practice.
Basic electrophysiology and electrical anatomy of the heart
The ECG is the final result of a complex series of physiologic and technologic processes.15 Physiologically, the ECG reflects differences in transmembrane voltages in myocardial cells that occur during depolarization and repolarization within each cycle. Ionic currents are generated because of ionic fluxes across cell membranes in myocardial cells during depolarization and repolarization. The cardiac cells are con tiguous and electrically connected by ion channels (gap junctions), which allows the ion current to pass through the cells and spread depolarization.16 Thus, the membrane potential changes in the heart can be considered a single depolarization that propagates through the whole heart, assuming different forms along the way.16 The pattern and sequence of depolarization that occur in the heart are depicted in Figure 15-1. There are many different types and subtypes of ion channels that are involved in the synchronized generation of electrical activity in the heart. Of note are the sodium, potassium, calcium, and chloride channels.15–17 Detailed discussion of these channels is beyond the scope of this chapter.
At any point in time, the electrical activity of the heart is composed of differently directed electrical forces. However, these currents are synchronized by cardiac activation and recovery sequences to generate a cardiac electrical field in and around the heart that varies with time during the cardiac cycle. This cardiac electrical field passes through various internal structures such as lungs, blood, and skeletal muscles. The currents reaching the skin are then detected by the electrodes that are placed in specific locations on the body and uniquely configured to produce different ECG patterns or waveforms. Direction and strength of a lead vector depend on the geometry of the body and on the varying electric impedances of the tissues in the torso.18,19 As expected, placement of electrodes on the torso is distinct from direct placement on the heart because the localized signal strength that occurs with direct electrode contact is markedly attenuated and altered by torso inhomogeneities, which include thoracic tissue boundaries and variations in impedance. The standard 12-lead ECG records potential differences (represented as change of voltage over time) between prescribed sites on the body surface that vary during the cardiac cycle.4
The first deflection noted on the ECG is caused by atrial depolarization and is called the P wave. Although the depolarization of the sinoatrial node precedes the atrial depolarization (see Figure 15-1), the potential from these pacemaker cells are too small to be detected on the surface ECG. The width of the P wave reflects the time taken for the wave of depolarization to spread over both the right and left atria. In comparison with the ventricular action potential, the atrial action potential is narrower and has a less prominent plateau. The duration of atrial contraction is, thus, shorter, which permits another action potential to occur sooner and makes atria prone to a very high rate (atrial flutter). The repolarization atrial wave rarely is seen in a normal ECG because it is buried in the much larger QRS wave.
The ECG returns to its baseline between the end of atrial depolarization and the commencement of the QRS complex, which is the start of the QRS ventricular depolarization. This interval is called the PR interval. Though this period may seem electrically silent, it is a time of significant electrical activity. During this period, the wave of depolarization that started in the sinoatrial node is propagated through the AV node, the AV bundle, right and left bundle branches, and Purkinje fibers (see Figure 15-1).
The QRS wave is followed by a period when the ECG returns to the baseline and is called the ST segment. It is a time when the ventricle is completely depolarized and is represented by phase 2 of the action potential (see Figure 15-1). Even though the ventricles are depolarized, the ECG does not record any positive or negative waveforms because the whole ventricles are depolarized and there is no potential difference between sites. The ECG does not measure absolute levels of membrane potential but only records the potential differences.15 The same explanation also holds true for the T-P segment, which represents a time when the ventricles are fully repolarized; hence no significant potential difference is recorded on a surface ECG.
Sometimes small undulations can be seen after T waves but before P waves. These are called U waves and are thought to be generated by M cells, which are specialized midmyocardial cells with prolonged action potentials.15
Technical aspects of the electrocardiogram
Most clinicians assume that the ECG is a relatively simple technical device. However, an extensive amount of advanced electrical theory underlies both the recording and display of the ECG signal. Digital signal processing (DSP) is now used universally, and the average ECG unit incorporates several microprocessors. Anesthesiologists should familiarize themselves with the theory behind ECG acquisition to maximize rational clinical application and appreciate its clinical limitations. In this section, the basics of electrocardiography are presented, briefly considering the major components that are involved in the faithful rendition of the surface ECG, working from the skin and electrodes progressively to the final output on the screen. The reader is referred to a number of technical reviews for more detail.4,5,12,20–23
Processing of the ECG occurs in a series of steps as shown in Figure 15-2.4 These steps include (1) signal acquisition, including filtering; (2) data transformation, or rendition of data for further processing, including finding the complexes, classification of the complexes into “dominant” and “nondominant” (ectopic) types, and formation of an average or median complex for each lead; (3) waveform recognition, which is the process for identification of the onset and offset of the diagnostic waves; (4) feature extraction, which is the measurement of intervals and amplitudes; and (5) for the bedside 12-lead ECG machines, diagnostic classification.4 Diagnostic classification may be heuristic (i.e., deterministic, or based on experience-based rules) or statistical in approach.24
Figure 15-2 Schematic representation of the processes resulting in recording of the electrocardiogram.
Signal Acquisition and Power Spectrum of the Electrocardiogram
It is relevant to consider an electrocardiographic signal in terms of its amplitude (or voltage) and its frequency components (generally called its phase) to appreciate ECG signal acquisition. Voltage considerations differ depending on the signal source. Surface recording involves amplification of smaller voltages (on the order of 1 mV) than recording sites closer to the heart beneath the electrically resistant layers of the skin (e.g., endocardial, esophageal, and intratracheal leads). The “power spectrum” of the ECG (Figure 15-3) is derived by Fourier transformation, in which a periodic waveform is mathematically decomposed to its harmonic components (sine waves of various amplitudes and frequencies). The fundamental frequency for the QRS complex at the body surface is approximately 10 Hz, and most of the diagnostic information is contained below 100 Hz in adults. Spectra representing some of the major sources of artifact must be eliminated during the processing and amplification of the QRS complex.22 The frequency of each of these components can be equated to the slope of the component signal.6 The R wave with its steep slope is a high-frequency component (100 Hz), whereas P and T waves have lesser slopes and are lower in frequency (1 to 2 Hz). The ST segment has the lowest frequency, not much different from the “underlying” electrical (i.e., isoelectric) baseline of the ECG. Before the introduction of DSP, accurately displaying the ST segment presented significant technical problems, particularly in operating room and intensive care unit bedside monitoring units. Although the overall frequency spectrum of the QRS complex does not appear to exceed 40 Hz, many components of the QRS complex, particularly the R wave, can exceed 100 Hz. The American Heart Association (AHA) recommends a bandwidth of 0.05 to 100 Hz for monitoring and detection of myocardial ischemia.4 Very-high-frequency signals of particular clinical significance are pacemaker spikes. Their short duration and high amplitude present technical challenges for proper recognition and rejection to allow accurate determination of the heart rate. The frequencies of greatest importance for optimal ECG processing are presented in Table 15-2.5
TABLE 15-2 Range of Signal Frequencies Included in Different Phases of Processing in an Electrocardiographic Monitor
| Processing | Frequency Range |
|---|---|
| Display | 0.5 (or 0.05)–40 Hz |
| QRS detection | 5–30 Hz |
| Arrhythmia detection | 0.05–60 Hz |
| ST-segment monitoring | 0.05–60 Hz |
| Pacemaker detection | 1.5–5 kHz |
Digital Signal Processing of the Electrocardiogram
Computerized ECG processing has been adapted to all major clinical applications of the ECG. The earliest application of A/D signal processing occurred during exercise tolerance testing, when significant motion artifact and electromyographic noise make acquisition of a “clean” ECG signal difficult. Outside of the exercise treadmill laboratory, computer processing allows automated analysis of the diagnostic 12-lead ECG.25 The reader is referred elsewhere for more detailed discussions of this technology, and to the reports of the scientific council of the AHA on standardization and specifications for automated ECG and bedside monitors.4,25–28
Processing of the ECG signal by a digital electrocardiograph involves initial sampling of the signal from electrodes on the body surface. Nearly all current-generation ECG machines convert the analog ECG signal to digital form before further processing. The foundation of DSP is the A/D converter, which samples the incoming “continuous” analog signal (characterized by variable amplitude or voltage over time) at a very rapid rate, converting the sampled voltage into binary numbers, each of which has a precise time index or sequence. Greater sampling rates (≥ 10,000 to 15,000/sec), which are typically less than 0.5 millisecond in duration help detect pacemaker output reliably. Several technical recommendations regarding low-frequency filtering and high-frequency filtering recently have been published by the AHA.4
Formation of a Representative Single-Lead Complex
QRS waveform amplitudes and durations are subject to beat-to-beat variability and to respiratory variability between beats. Digital ECGs can adjust for respiratory variability and decrease beat-to-beat noise to improve the measurement precision in individual leads by forming a representative complex for each lead. Signal averaging is a critical component of this process. Noise is reduced using this technique proportionate by the square root of the number of beats averaged.4 Thus, a 10-fold reduction in noise is accomplished by averaging only 100 beats. Automated measurements are made from these representative templates, not from measurement of individual complexes. Average complex templates are formed from the average amplitude of each digital sampling point for selected complexes. Median complex templates are formed from the median amplitude at each digital sampling point. As a result, measurement accuracy is strongly dependent on the fidelity with which representative templates are formed. Because of the proprietary nature of this technology (the specific algorithms used are patented), the method used may vary by manufacturer. Consequently, the processed QRS complexes may vary in the “quality” of representation (i.e., if noise or aberrant beats are averaged into the complex, it will vary from the raw analog complex). The averaging process involves comparison of the voltages at a particular time point between the incoming complex and the template. Although the easiest method is to use the mean difference between voltages to update the “template,” the most accurate method is to use the median (because it is less affected by outliers, such as aberrant beats or other signals that have escaped QRS matching)4 (Figure 15-4).
A feature incorporated into most monitors is a visual trend line from which deviations in the position of the ST segment can be rapidly detected, which can aid online detection of ischemia. In addition, nearly all monitors display on-screen numerical values for the position of the ST segment used for ischemia detection (generally 60 to 80 milliseconds after the J point), although the specific fiducial point (based on heart rate) used usually can be adjusted by the clinician (Figure 15-5).
History and Description of the 12-Lead System
Where and how ECG electrodes are placed on the body are critical determinants of the morphology of the ECG signal. Lead systems have been developed based on theoretical considerations and references to anatomic landmarks that facilitate consistency between individuals (e.g., standard 12-lead system). Einthoven established electrocardiography using three extremities as references: the left arm, right arm, and left leg. He recorded the difference in potential between the left arm and right arm (lead I), between the left leg and right arm (lead II), and between the left leg and left arm (lead III) (Figure 15-6). Because the signals recorded were differences between two electrodes, these leads were called bipolar. The right leg served only as a reference electrode. Because Kirchoff’s loop equation states that the sum of the three voltage differential pairs must equal zero, the sum of leads I and III must equal lead II.21 The positive or negative polarity of each of the limbs was chosen by Einthoven to result in positive deflections of most of the waveforms and has no innate physiologic significance. He postulated that the three limbs defined an imaginary equilateral triangle with the heart at its center. Wilson refined and introduced the precordial leads into clinical practice. To implement these leads, he postulated a mechanism whereby the absolute level of electrical potential could be measured at the site of the exploring precordial electrode (the positive electrode). A negative pole with zero potential was formed by joining the three limb electrodes in a resistive network in which equally weighted signals cancel each other out. He called this the “central terminal,” and in a fashion similar to Einthoven’s vector concepts, he postulated it was located at the electrical center of the heart, representing the mean electrical potential of the body throughout the cardiac cycle. He described three additional limb leads (aVL, aVR, and aVF; Figure 15-7). These leads measured new vectors of activation, and in this way, the hexaxial reference system for determination of electrical axis was established. He subsequently introduced the six unipolar precordial V leads in 1935 (see Figure 15-6).29 Six electrodes are placed on the chest in the following locations: V1, fourth intercostal space at the right sternal border; V2, fourth intercostal space at the left sternal border; V3, midway between V2 and V4; V4, fifth intercostal space in the midclavicular line; V5, in the horizontal plane of V4 at the anterior axillary line, or if the anterior axillary line is ambiguous, midway between V4 and V6; and V6, in the horizontal plane of V4 at the midaxillary line4 (see Figure 15-6).
Clinical application of the unipolar limb leads was limited because of their significantly smaller amplitude relative to the bipolar limb leads from which they were derived. They were not clinically applied until Goldberger augmented their amplitude (by a factor of 1.5) by severing the connection between the central terminal and the lead extremity being studied (which he called “augmented limb leads”) in 1942. The limb leads, the precordial leads, and the augmented unipolar limb leads form what was accepted by the AHA as the conventional 12-lead ECG system.30 Einthoven’s law indicates that any one of the standard limb leads can be mathematically derived from the other two limb leads. Therefore, the “standard” 12-lead ECG actually contains eight independent pieces of information: two measured potential differences from which the four remaining limb leads can be calculated and the six independent precordial leads.4
