Electrocardiography:
The electrocardiogram (ECG), as used today, is the product of a series of technological and physiological advances pioneered over the past two centuries. Early demonstrations of the heart's electrical activity reported during the last half of the 19th century, for example, by Marchand and others, were closely followed by direct recordings of cardiac potentials by Waller in 1887. Invention of the string galvanometer by Willem Einthoven in 1901 provided a reliable and direct method for registering electrical activity of the heart. By 1910, use of the string galvanometer had emerged from the research laboratory into the clinic. Subsequent achievements built on the limited, but very solid foundation supplied by the early electrocardiographers. The result has become a widely used and invaluable clinical tool for the detection and diagnosis of a broad range of cardiac conditions, as well as a technique that has contributed to the understanding and treatment of virtually every type of heart disease. Furthermore, the ECG is essential in the management of major metabolic abnormalities such as hyperkalemia and certain other electrolyte disorders, as well as assessing drug effects and toxicities such as caused by digitalis, antiarrhythmic agents, and tricyclic antidepressants. Moreover, it has remained the most direct method for assessing abnormalities of cardiac rhythm.




Fundamental principles:

Use of the ECG for any of these clinically important purposes is the final outcome of a complex series of physiological and technological processes. First, an extracellular cardiac electrical field is generated by ion fluxes across cell membranes and between adjacent cells. These ion 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 electrical field passes through numerous other structures, including the lungs, blood, and skeletal muscle, before reaching the body surface. These structures--known as transmission factors--differ in their electrical properties and perturb the cardiac electrical field as it passes through them. The potentials reaching the skin are then detected by electrodes placed in specific locations on the extremities and torso and configured to produce leads. The outputs of these leads are amplified, filtered, and displayed by a variety of electronic devices to construct an ECG recording. Finally, diagnostic criteria are applied to these recordings to produce an interpretation. The criteria have statistical characteristics that determine the clinical utility of the findings. Each of these steps influences the final product--the clinical ECG.

Electrocardiographic display systems:

Another group of factors that determines ECG waveforms includes the characteristics of the electronic systems used to amplify, filter, and digitize the sensed signals. ECG amplifiers are differential amplifiers, that is, they amplify the difference between two inputs. For bipolar leads, the differential output is the difference between the two active leads; for unipolar leads, the difference is between the exploring electrode and the reference electrode. This differential configuration significantly reduces the electrical noise that is sensed by both inputs and hence is canceled. The standard amplifier gain for routine electrocardiography is 1000 but may vary from 500 (half-standard) to 2000 (double standard).
Amplifiers respond differently to the range of signal frequencies included in an electrophysiological signal. The bandwidth of an amplifier defines the frequency range over which the amplifier accurately amplifies the input signals. Waveform components with frequencies above or below the bandwidth may be artifactually reduced or increased in amplitude. In addition, recording devices include high- and low-pass filters that intentionally reduce the amplitude of specific frequency ranges of the signal. Such reduction in amplitude may be done, for example, to reduce the effect of body motion or line voltage frequencies, that is, 60-Hz interference. For routine electrocardiography, the standards of the American Heart Association require a bandwidth of 0.05 to 100 Hz.

Amplifiers for routine electrocardiography include a capacitor stage between the input and the output terminals; that is, they are capacitor coupled. This configuration blocks direct-current (DC) voltage while permitting flow of alternating-current (AC) signals. Because the ECG waveform may be viewed as an AC signal (which accounts for the waveform shape) that is superimposed on a DC baseline (which determines the actual voltage levels of the recording), this coupling has significant effects on the recording process. First, unwanted DC potentials, such as those produced by the electrode interfaces, are eliminated. Second, elimination of the DC potential from the final product means that ECG potentials are not calibrated against an external reference level. ECG potentials must be measured in relation to an internal standard. Thus, amplitudes of waves are measured in millivolts or microvolts relative to another portion of the waveform. The TP segment, which begins at the end of the T wave of one cardiac cycle and ends with onset of the P wave of the next cycle, is usually the most appropriate internal ECG baseline.
An additional issue is the digitizing or sampling rate for computerized systems. Too low a sampling rate will miss brief signals such as notches in QRS complexes or brief bipolar spikes and reduce the precision and accuracy of waveform morphologies. Too fast a sampling rate may introduce artifacts, including high-frequency noise, and requires excessive digital storage capacity. In general, the sampling rate should be at least twice the frequency of the highest frequencies of interest in the signal being recorded. Standard electrocardiography is most commonly performed with a sampling rate of 500 Hz, with each sample representing a 2-millisecond period.
Cardiac potentials may be processed for display in numerous formats. The most common of these formats is the classic scalar ECG. Scalar recordings depict the potentials recorded from one lead as a function of time. For standard electrocardiography, amplitude is displayed on a scale of 1 mV to 10-mm vertical displacement and time as 200 msec/cm on the horizontal scale. Other display formats are used for ambulatory electrocardiographyAND for bedside ECG monitoring.
 

The normal ECG

The heart is activated with each cardiac cycle in a very characteristic manner determined by the anatomy and physiology of working cardiac muscle and the specialized cardiac conduction systems. P wave is generated by activation of the atria, the PR segment represents the duration of atrioventricular (AV) conduction, the QRS complex is produced by activation of both ventricles, and the ST-T wave reflects ventricular recovery. Table 5-2 includes normal values for the various intervals and waveforms of the ECG.

Atrial activation and the P-wave:

Atrial activation begins with impulse generation in the atrial pacemaker complex in or near the sinoatrial (SA) node. The rate of discharge of the SA node and, hence, the heart rate is dependent on parasympathetic and sympathetic tone, as well as the intrinsic properties of the SA node, extrinsic factors such as mechanical stretch, and various pharmacological effects.

Electrocardiograph
An electrocardiograph (ECG or EKG) records the electrical activity of the heart. Preceding each contraction of the heart muscle is an electrical impulse generated in the sinoatrial node; the waves displayed in an ECG trace the path of that impulse as it spreads through the heart. Irregularities in an ECG reflect disorders in the muscle, blood supply, or neural control of the heart.


The normal ECG

The heart is activated with each cardiac cycle in a very characteristic manner determined by the anatomy and physiology of working cardiac muscle and the specialized cardiac conduction systems. P wave is generated by activation of the atria, the PR segment represents the duration of atrioventricular (AV) conduction, the QRS complex is produced by activation of both ventricles, and the ST-T wave reflects ventricular recovery. Table 5-2 includes normal values for the various intervals and waveforms of the ECG.

Atrial activation and the P-wave:

Atrial activation begins with impulse generation in the atrial pacemaker complex in or near the sinoatrial (SA) node. The rate of discharge of the SA node and, hence, the heart rate is dependent on parasympathetic and sympathetic tone, as well as the intrinsic properties of the SA node, extrinsic factors such as mechanical stretch, and various pharmacological effects.


Increasing attention is being directed at the beat-to-beat changes in heart rate, termed heart rate variability, to gain insight into neuroautonomic control mechanisms and their perturbations with aging, disease, and drug effects. For example, high-frequency (0.15-0.5 Hz) fluctuations mediated by the vagus nerve occur phasically, with heart rate increasing during inspiration and decreasing during expiration. Attenuation of this respiratory sinus arrhythmia is a consistent marker of physiological aging and also occurs with diabetes mellitus, congestive heart failure, and a wide range of other cardiac and noncardiac conditions that alter autonomic tone. Lower-frequency (0.05-0.15 Hz) physiological oscillations in heart rate are associated with baroreflex activation and appear to be jointly regulated by sympathetic and parasympathetic interactions. A variety of complementary signal processing techniques are being developed to analyze heart rate variability, including the very low-frequency (<0.05 Hz) components and circadian rhythms. These methods include time-domain statistics, frequency domain techniques based on spectral (Fourier) methods, and tools derived from nonlinear dynamics, including chaos (complexity) theory and statistical physics.
Once the impulse leaves this pacemaker site, atrial activation spreads in several directions. First, propagation is rapid along the crista terminalis and moves anteriorly toward the lower portion of the right atrium. It also spreads across the anterior and posterior surfaces of the atria toward the left atrium. The last area to be activated is over the inferolateral aspect of the left atrium, which is activated by convergence of these anterior and posterior wave fronts moving from right to left. Although right atrial activation begins before activation of the left atrium, activation occurs simultaneously in both atria during much of the overall atrial activation time. At the same time, activation spreads through the interatrial septum, beginning high on the right side and moving around the fossa ovalis to the reach the top of the interventricular septum.

The pattern of atrial activation noted above produces the normal P wave. Activation beginning high in the right atrium and proceeding simultaneously leftward toward the left atrium and inferiorly toward the AV node corresponds to a mean frontal plane P wave axis of approximately 60 degrees. Based on this orientation of the heart vector, normal atrial activation projects positive or upright P waves in leads I, II, aVl , and aVf . The pattern in lead III may be either upright or downward, depending on the exact orientation of the mean axis, that is, upright if the mean axis is more positive than +30 degrees and negative otherwise.
P wave patterns in the precordial leads correspond to the direction of atrial activation wave fronts in the horizontal plane. Atrial activation early in the P wave is oriented primarily anteriorly over the right atrium and later posteriorly over the left atrium. Thus, the P wave in the right precordial leads (V1 and, occasionally, V2 ) is commonly biphasic, with an initial positive deflection followed by a later negative one. In the more lateral leads, the P wave is upright and reflects right-to-left spread of the activation fronts.
P wave duration is normally under 120 milliseconds. The amplitude in the limb leads is normally under 250 muV, and the terminal negative deflection in the right precordial leads is normally under 100 muV in depth.
Atrial depolarization is followed by atrial repolarization. The potentials generated by atrial repolarization are not usually seen on the surface ECG because of their low amplitude (usually under 100 muV) and because they are superimposed on the much higher amplitude QRS complex.[22] They may be observed as a low-amplitude wave with a polarity opposite that of the P wave (the Ta wave) during heart block and may have special significance in influencing ECG patterns during exercise testing.[23] Deviation of the PR segment (corresponding to the atrial ST segment) is, as described below, also an important marker of acute pericarditis and, more rarely, atrial infarction.



AV node conduction and PR segment:
The PR segment is the isoelectric region beginning with the end of the P wave and ending with onset of the QRS complex. It forms part of the PR interval, which extends from onset of the P wave to onset of the QRS complex. The normal PR interval measures 120 to 200 milliseconds in duration.
The PR segment is the temporal bridge between atrial activation and ventricular activation. It is during this period that activation of the AV node, the bundle of His, the bundle branches, and the intraventricular specialized conduction system occurs. As noted above, atrial repolarization also occurs during this period. Most of the conduction delay during this segment is due to slow conduction within the AV node.
Upon exiting the AV node, the impulse traverses the bundle of His to enter the bundle branches and then travels through the specialized intraventricular conduction paths to finally activate ventricular myocardium. The PR segment appears isoelectric because the potentials generated by these structures are too small to produce detectable voltage on the body surface at the normal amplifier gains used in clinical electrocardiography. The standard ECG detects only activation and recovery of working myocardium, not the specialized conduction system. Signals from elements of the conduction system can be recorded from the body surface by using very high gains (over 25,000) and signal-averaging techniques or from intracardiac recording electrodes placed against the base of the interventricular septum near the bundle of His.



Ventricular activation and QRS complex
Ventricular excitation is the product of two temporally overlapping functions--endocardial activation and transmural activation. Endocardial activation is guided by the anatomical distribution and physiology of the His-Purkinje system. The broadly dispersed ramifications of this treelike or fractal system and the rapid conduction within it result in depolarization of most of the endocardial surfaces of both ventricles within several milliseconds and the simultaneous activation of multiple endocardial sites.
Earliest activity begins in three sites: (1) the anterior paraseptal wall of the left ventricle, (2) the posterior paraseptal wall of the left ventricle, and (3) the center of the left side of the septum. These loci generally correspond to the sites of insertion of the three branches of the left bundle branch. Wave fronts spread from these sites in anterior and superior directions to activate the anterior and lateral walls of the left ventricle. The posterobasal areas of the left ventricle are the last to be activated. Septal activation begins in the middle third of the left side and spreads across the septum from left to right and from apex to base.
Excitation of the right ventricle begins near the insertion point of the right bundle branch close to the base of the anterior papillary muscle and spreads to the free wall. The final areas to be involved are the pulmonary conus and posterobasal areas. Thus, in both ventricles, the overall endocardial excitation pattern begins on septal surfaces and sweeps down and around the anterior free walls to the posterior and basal regions in an apex-to-base direction.
The activation fronts then move from endocardium to epicardium. Excitation of the endocardium begins at sites of Purkinje-ventricular muscle junctions and proceeds by muscle cell-to-muscle cell conduction in an oblique direction toward the epicardium.
This sequence of endocardial and transmural activation results in the characteristic waveforms of the QRS complex. QRS patterns are described by the sequence of waves constituting the complex. An initial negative deflection is called the Q wave, the first positive wave is the R wave, and the first negative wave after a positive wave is the S wave. A second upright wave following an S wave is an R
wave. Tall waves are denoted by capital letters and smaller ones by lowercase letters. A monophasic negative complex is referred to as a QS complex. Thus, for example, the overall QRS complex may be described as qRS if it consists of an initial small negative wave (the q wave) followed by a tall upright one (the R wave) and a deep negative one (an S wave). In an RSr complex, initial R and S waves are followed by a small positive wave (the r wave).
The complex pattern of activation described above can be simplified into two vectors representing septal and left ventricular free wall activation. Initial activation of the interventricular septum corresponds to a vector oriented from left to right in the frontal plane and anteriorly in the horizontal plane, as determined by the anatomical position of the septum within the chest. This arrangement produces an initial positive wave in leads with axes directed to the right (lead aVr ) or anteriorly (lead V1 ). Leads with axes directed to the left (leads I, aVl , V5 , and V6 ) will register initial negative waves (septal q waves). These initial forces are normally of low amplitude and are brief (less than 40 milliseconds). Absence of these septal q waves is associated with septal infarction or fibrosis and commonly correlates with other ECG evidence of myocardial infarction and left ventricular mechanical dysfunction.
Subsequent parts of the QRS complex reflect activation of the free walls of the left and right ventricles. Because right ventricular muscle mass is considerably smaller than that of the left ventricle, it contributes little to QRS complexes recorded in the standard ECG. Thus, the second phase of the normal QRS can be considered to represent only left ventricular activity with relatively little oversimplification.
Once free wall activation begins, leads with leftward axes (leads I, aVl , V5 , and V6 ) show positive deflections following the initial septal q waves. Leads with axes oriented to the right (including lead aVr ) record negative potentials. As activation proceeds, the height of the R waves and the depth of the S waves progressively increase. Thus, leads I, aVl , V5 , and V6 typically show qR patterns and lead aVr registers rS waveforms.
The forms of the QRS complex frontal plane leads are variable and reflect differences in the mean QRS electrical axis. The normal mean QRS axis in adults lies between -30 degrees and +90 degrees. Mean QRS axes more positive than +90 degrees represent right axis deviation, and those more negative than -30 degrees represent left axis deviation. Mean axes lying between -90 and -180 degrees (or equivalently between +180 and +270 degrees) are referred to as extreme axis deviations. The designation indeterminate axis is applied when all six extremity leads show biphasic (QR or RS) patterns; this finding can occur as a normal variant or may be seen in a variety of pathological conditions.
The wide span of the normal axis results in a range of QRS patterns, especially in the inferior leads. This characteristic can be understood by referring to the hexaxial reference system in Figure 5-8 . If the mean axis is near 90 degrees, the QRS complex in leads II, III, and aVf will be predominantly upright with qR complexes; lead I will record an isoelectric qRS pattern because the heart vector lies perpendicular to the lead axis. This configuration is commonly referred to as a vertical heart position, although it probably has little to do with the anatomical position of the heart within the chest.[13] If the mean axis is nearer 0 degrees, the patterns will be reversed; lead I (and aVl ) will register a predominantly upright qR pattern, and leads II, III, and aVf will show rS or RS patterns, a configuration often referred to as a horizontal heart pattern.


QRS duration

The upper normal value for QRS duration is traditionally given as <120 milliseconds. In a survey of 1224 men with normal QRS morphology and frontal plane axis, the 98 percent upper bound of QRS duration was 116 milliseconds. Women, on average, have somewhat smaller (about 5 to 8 milliseconds) QRS durations than men do.
An additional feature of the QRS complex is the intrinsicoid deflection. An electrode overlying the ventricular free wall will record a rising R wave as transmural activation of the underlying ventricular free wall proceeds. Once the activation front reaches the epicardium, the full thickness of the wall under the electrode will be in an active state with no propagating electrical activity. At that moment, the electrode will register negative potentials from remote cardiac areas still undergoing activation. The sudden reversal of potential with a sharp downslope is the intrinsicoid deflection and marks the timing of activation of the epicardium under the electrode.


Ventricular recovery and ST-T wave

The normal ST-T wave begins as a low-amplitude, slowly changing wave (the ST segment) that gradually leads to a larger wave, the T wave. Onset of the ST-T wave is the junction or J point and is normally at or near the isoelectric baseline of the ECG.
The polarity of the ST-T wave is generally the same as the net polarity of the preceding QRS complex. Thus, T waves are upright in leads I, II, aVl , aVf , and the lateral precordial leads. They are negative in leads aVr and variable in leads III and V1 through V3 .
Like activation, recovery in the ventricles occurs in a characteristic geometrical pattern. Differences in recovery timing occur both across the ventricular wall and between regions of the left ventricle. Transmural differences in recovery times are the net result of two effects--differences in action potential duration across the ventricular wall and the relatively slow spread of activation across the wall.[33] As activation moves from endocardium to epicardium, sites further away from the endocardium are activated later and later in sequence. However, action potential durations are longest near the endocardium and shortest near the epicardium, which produces a transmural gradient in recovery periods. Differences in action potential duration are greater than differences in activation times, so recovery is completed near the epicardium before it is completed near the endocardium. For example, one endocardial site may be excited 10 milliseconds earlier than the overlying epicardium (that is, transmural activation may require 10 milliseconds), and the action potential duration at the endocardium may be 22 milliseconds longer than on the epicardium. As a result, recovery will be completed 12 milliseconds earlier in the epicardium than in the endocardium.
The resulting recovery dipole will then be directed from sites of less recovery (that is, the endocardium) toward sites of greater recovery (that is, near the epicardium). The orientation of this dipole is in the same direction as transmural activation dipoles and is contrary to the expected direction as described earlier in this chapter; this difference is due to the presence of nonuniform recovery properties across the wall. If recovery times were uniform across the wall (or if differences in recovery times were less than differences in transmural activation times), the recovery dipole would have been directed toward the endocardium, that is, in the direction opposite the activation dipole. The result is, in normal persons, concordant QRS and ST-T wave patterns.
Regional differences in recovery properties likewise exist. Under normal conditions, it is the transmural gradients that predominantly determine ST patterns. However, as will be described, these regional differences account for the discordant ST-T patterns observed with intraventricular conduction defects.




QRST angle

This concordance between orientation of the QRS complex and ST-T wave can also be expressed vectorially. An angle can be visualized between the vector representing the mean QRS force and that representing the mean ST-T force--the QRST angle. This angle in the frontal plane is normally less than 60 degrees and usually under 30 degrees. Abnormalities of the QRST angle reflect abnormal relationships between the properties of activation and recovery.

The ventricular gradient

If the two vectors representing mean activation and mean recovery forces are added, a third vector known as the ventricular gradient is created. The concept of the ventricular gradient was developed to assess differences in the properties of ventricular activation and recovery. According to this concept, the more variability that exists in regional repolarization properties, the greater the difference from zero of the sum of the QRS and ST-T areas will be. In other words, the net QRST area, as measured by the ventricular gradient, correlates with the magnitude of regional differences in recovery properties. In addition, because changes in activation patterns produced by, for example, bundle branch block also cause corresponding changes in recovery patterns (see below), the ventricular gradient should allow a measure of regional recovery properties that is independent of the activation pattern. This measurement has possible relevance to the genesis of reentrant arrhythmias that are due, in part, to regional variations in refractory periods.

The U Wave

The T wave may be followed by an additional low-amplitude wave known as the U wave. It is usually under 100 muV in amplitude, normally has the same polarity as the preceding T wave, and is largest in the midprecordial leads and at slow heart rates. Its basis in cardiac electrophysiology is uncertain, but it may be caused by repolarization of the Purkinje fibers or by delayed repolarization in areas of the ventricle that undergo late mechanical relaxation.



The QT interval
A final interval of the ECG waveform is the QT interval, which is measured from the beginning of the QRS complex to the end of the T wave. It includes the total duration of ventricular activation and recovery and, in a general sense, corresponds to the duration of the ventricular action potential.
The normal QT interval is defined by its duration, measured in milliseconds. Like the ventricular action potential duration, the duration of the QT interval decreases as heart rate increases. Thus, the normal range for the QT interval is rate dependent. One formula for relating QT interval duration to heart rate was developed by Bazett in 1920. The result is computation of a corrected QT interval, or QTc , by using the equation QTc =QT/(R - R)½ where the QT and RR intervals are measured in seconds. The normal QTc is generally accepted to be less than or equal to 440 milliseconds. Some studies suggest that it may be 20 milliseconds longer, and it is slightly longer, on average, in women. Because the end of the T wave can overlap with the beginning of a U wave, the QT interval is sometimes referred to as the QT(U) interval; this designation is particularly appropriate when considering the ECG effects of certain metabolic abnormalities that alter the duration of repolarization and the amplitude of the U wave (see below).
The Bazett formula, while widely used to adjust the QT interval duration for the effects of heart rate, has limited accuracy in predicting the effects of heart rate on the QT interval. Many other formulas and methods for correcting the QT interval for the effects of heart rate, including logarithmic, hyperbolic, and exponential functions, have been developed and tested, but they also have limitations. These limitations result from both physiological and computational problems. On one hand, the Bazett formula predicts an ever-increasing increment in the QT interval as the heart rate slows and an ever-decreasing increment as the rate rises, both of which are physiologically improbable. In addition, all these formulas do not account for the effects of autonomic tone on the QT interval independent of the effects on rate. They also do not account for the relatively slow adaptation of repolarization to changes in rate; for example, several minutes may be required for the QT interval to reach a new steady state after an abrupt change in heart rate.
A second property of the QT interval is that its duration is lead dependent, that is, the duration of the QT interval varies from lead to lead. In normal persons, the QT interval varies between leads by up to 50 milliseconds and is longest in the midprecordial leads V2 and V3 . This range of intervals, referred to as QT interval dispersion, may be related to electrical instability and the risk of ventricular arrhythmogenesis.


Normal Variants

These descriptions of the waveforms of the normal ECG represent patterns most often observed in normal adults. It is important to understand the limitations of assigning and interpreting normal ranges to ECG measurement. Values for many of the intervals and amplitudes to be described vary widely within the population as a function of age, race, gender, and body habitus and within individuals as a function of autonomic tone and activity level. Thus, what is normal in one condition may be abnormal in another. Tables of values for these and other measures for different population subgroups have been published. Some variations have been described above, including, for example, variations in rate, QRS axis, and QT intervals.
Other common variations occur in patterns of the ST segment and T wave. These variations are important to recognize because they may be mistaken for significant abnormalities. First, ST-T patterns are affected by maneuvers that change autonomic tone. For example, changing body position, hyperventilating, drinking cold water, and performing the Valsalva maneuver can produce modest ST segment depression and T wave inversion in as many as one-third of subjects.
T waves can be inverted in the right precordial leads . In adults, this inversion reflects the uncommon, but not necessarily abnormal persistence of patterns commonly seen in infants and children. T waves can be inverted in all precordial leads at birth and usually become more limited to the right side of the chest as time passes; by the age of 10 years, T wave inversion is generally limited to leads V1 and V2 . A persistent juvenile pattern is more common in women than in men and among the black population than among other racial or ethnic groups.
Second, the ST segment can be significantly elevated, especially in the midprecordial leads. The elevation begins from an elevated J point, is usually concave in form, is commonly associated with notching of the downstroke of the QRS complex, and can reach 500 muV in amplitude. This pattern is more common at slow than at rapid heart rates and in men than in women and is most often seen among black men. Although this physiological ST segment elevation pattern is commonly referred to as early repolarization, clinical studies have failed to demonstrate an earlier than normal onset of ventricular recovery.