Evaluation of Myocardial Function

Transcription

1 Evaluation of Myocardial Function Kenneth E. Jochim, Ph.D., and Douglas M. Behrendt, M.D. ABSTRACT In assessing myocardial contractility one may examine isolated heart muscle, the isolated whole heart in controlled circumstances, or the heart in an intact patient. In each situation a number of different indices of contractility may be recorded, each of which has merit but none of which is perfect. Some reflect events occurring during isometric contraction prior to ejection, such as rate of change in intraventricular pressure, maximum velocity of muscle shortening, and velocity of shortening of contractile elements. Others reflect ejection phenomena such as stroke volume, the Starling curve, and the ejection fraction. None of these indices is entirely independent of preload, afterload, and heart rate, and these factors must be controlled. In the whole heart, especially in a clinical setting, this may be difficult. Many clinical studies of myocardial function fail to recognize the need for controlling these variables and therefore are of limited validity. W hen assessing ventricular performance we are basically concerned with the performance of the myocardium, which may be looked at in three situations. First, we may consider a thin, isolated bundle of essentially homogeneous and parallel heart muscle fibers suspended in a bath of physiological salt solution, the physical and chemical characteristics of which can be kept constant. Uniform contractions can be maintained for many hours by electrically stimulating the muscle at a fixed frequency. Under these circumstances all important variables affecting muscle performance can be precisely controlled and measured. Alternatively, the intact heart can be isolated and studied in a controlled setting. In this instance the muscle fibers, arranged in a complex geometrical shape, are not all subjected to the same physical and chemical environmental conditions. Finally, the heart can be studied in a patient. In this situation many important physiological variables cannot be adequately controlled or measured. In the diseased heart the presence of ventricular hypertrophy or areas of infarction adds to the complexity because of the resulting gross myocardial nonhomogeneity. Isolated Muscle Preparations When response to stimulation is manifested by changes in length, tension, or both in controlled experiments with isolated muscle, four indices of performance that can be accurately recorded as functions of time are length (L), rate of change in length (dwdt), tension (T), and rate of change in tension (dt/dt). Muscle performance as reflected by these indices is determined by four variables: (1) preload, (2) afterload, (3) frequency of contraction, and (4) contractility, that From the Department of Physiology and the Department of Surgery, Section ofthoracic Surgery, University of Michigan Medical School, Ann Arbor, Mich. Supported in part by Michigan Heart Association Grant NO Address reprint requests to Dr. Behrendt, C-7175 University Hospital, Ann Arbor, Mich THE ANNALS OF THORACIC SURGERY

2 Evaluation of Myocardial Function inherent characteristic of the muscle which determines how it will perform when the other three variables are fixed. Preload is the resting tension in the muscle, determined by the extent to which the muscle is stretched in the relaxed state. The relationship between resting tension and resting length expresses the passive, elastic characteristics of the muscle. Afterload is the additional tension imposed upon the muscle after it begins to shorten (if it is allowed to shorten). It may be constant during the shortening, as when the muscle lifts a weight, or changing, as when it stretches a spring. Frequency of contraction, of course, is simply determined by the frequency of the electrical stimuli delivered to the muscle. Contractility is discussed in detail below. The influence of these four factors has been studied in both isometric and isotonic contractions of isolated muscle. For isometric contractions, both ends of the muscle are fixed to give the desired resting length, and the tension in the muscle before and during contraction is recorded. No shortening is possible with this arrangement. Beginning with a small resting tension, it is found that as preload (and therefore resting length) is increased, peak developed T increases and the time required to reach peak T remains the same (Fig. 1A). Both (dt/dt),,, and (dt/dt),,, during contraction increase; beyond an optimum preload these changes reverse. This improved performance with increasing preload is an expression of the well-known Frank-Starling mechanism. Afterload, of course, has no meaning in isometric contractions since there is no shortening. An increase in frequency of contractions over a limited range (with constant preload) increases peak T and decreases time to peak T. This Treppe effect is interpreted as indicating an increase in contractility; it is the same as that produced by known positive inotropic agents such as norepinephrine or increased calcium ion (Fig. 1B). For isotonic contractions, the preload is set at a desired value and the muscle is arranged so as to permit shortening by lifting a weight off a support; the weight represents a constant afterload. Upon contraction, muscle tension increases from its initial value until it equals the afterload; thereafter it remains constant at this value during the shortening phase and subsequent lengthening phase until the FIG. 1. Typical plots of the development of isometric tension by an isolated paptllary muscle. (A) An increase in initial tension results in a rise in peak tension but no change in the time to peak tension. (B) An augmentation in contractility results in both an increase in peak tension and a shortening of the time to peak tension. t -J.- I A Time I B Time VOL. 20, NO. 1, JULY,

3 JOCHIM AND BEHRENDT A After Load B After Load FIG. 2. Typical forcehelocity curves from a papillary muscle preparation. (A) Increasing the afterload from Po to Po shifts the curve to the right but Vma, is unchanged. (B) An increase in contractility shifts the curue to the right and also increases Vmax. weight is returned to its support. The customary recorded variables are T, L, and dudt. The velocity of muscle shortening is taken as the maximum value of dudt reached during the contraction phase. With preload and frequency of contractions constant, one can record a series of contractions at different afterloads. It is found that AL, the extent of shortening, and (dudt),,, both decrease toward zero as the afterload increases. When velocity of shortening is plotted as a function of afterload, a so-called force/velocity curve is obtained. The point of intersection of this curve with the abscissa (Po) represents the minimum load against which the muscle cannot shorten, i.e., the maximum tension it can develop. When the curve is extrapolated to zero afterload (which, of course, cannot be realized experimentally), the intersection with the ordinate represents the maximum velocity of shortening that would be obtained with no afterload (V,,,). When the preload is increased, the force/velocity curve is shifted in such a way that Po is increased but V, is unchanged (Fig. 2A). On the other hand, when the preload is kept constant but the frequency of contractions is increased, in general both Po and V,,, are increased, as is also the case when positive inotropic agents such as norepinephrine are added to the muscle bath (Fig. 2B). The velocity of muscle shortening determined during isotonic contractions is actually measured while muscle tension is not changing. If a muscle fiber is considered to consist of a passive, elastic component, the series elastic element (SE), in series with a contractile component, the contractile element (CE), then the velocity of shortening of the whole muscle fiber is actually the velocity of shortening of the CE when muscle tension is constant. Under these conditions the length of the SE is unchanging. V,,,, then, is the maximum velocity of shortening that the CE is capable of displaying under a given set of conditions when tension (afterload) is zero. As indicated, it is not influenced by preload, but it is augmented by an increase in frequency of contractions and by a variety of humoral agents; it is decreased by depressants such as anoxia and anesthetic agents. V, is commonly used as an index of myocardial contractility. Intact Heart Preparations In the complex situation presented by the intact, spontaneously beating heart, there is reason to believe that performance of the ventricular myocardium 32 THE ANNALS OF THORACIC SURGERY

4 Evaluation.f Myocardial Function is influenced by the same factors governing performance of the isolated muscle; but even in contrived experimental situations they are difficult, and sometimes impossible, to measure or record with an acceptable level of confidence. Furthermore, because of the large mass of muscle, the complicated geometry of its arrangement, and the presence in disease of abnormal areas, these variables differ in various regions of the heart. As an index of ventricular preload, it would be desirable to measure the average resting tension of all the ventricular muscle fibers at the end of diastole; since this is not possible, the left ventricular end-diastolic pressure (LVEDP) is usually measured instead. It is obvious, however, that the relationship between these two values is not constant but depends on ventricular volume and geometry. In addition, the relationship between LVEDP and resting fiber length, which is the important factor in the Frank-Starling mechanism, depends upon ventricular compliance, a property that can be altered by various influences. Ventricular afterload is the tension developed in the muscle during shortening. It is continuously changing during ejection, and hence ventricular contraction is not isotonic. Again, one is constrained to measure intraventricular pressure instead of muscle tension directly during ejection as an index of afterload. For the left ventricle it is more common simply to use mean aortic pressure as a measure of afterload. The greatest difficulty is encountered in attempting to assess changes in myocardial contractility in the intact heart. Whatever criterion is selected, it should be one that reflects changes in contractility without being influenced by alterations in preload or afterload, because only rarely will these two factors be unchanged in the states in which contractility is to be compared. Further, since it is known that frequency of contraction - i.e., heart rate - alters contractility, it is important that this variable too be controlled when the inotropic effects of some agent or procedure are to be determined. Ifall three variables remain unchanged, however, then alterations in stroke volume will indicate changes in contractility. In isometric contractions of isolated muscle, an increase in contractility shortens the time to peak T (see Fig. 1B) while changes in preload do not (see Fig. 1A). It might be suggested, then, that this criterion could be applied to the isovolumic phase of ventricular systole, in which the rapidly rising intraventricular pressure might be used as an index of muscle tension. The criterion is useless, however, since isovolumic contraction is ended by the beginning of ejection long before peak tension can be reached. Further, in isometric contractions of isolated muscle, (dt/dt),,, is augmented by an increase in either contractility or preload. But if the latter can be kept constant, then during the preejection phase of ventricular systole the maximum rate of change in intraventricular pressure, (dp/dt),,,, could conceivably mirror inotropic changes. Unfortunately, it has been found that the maximum value of dp/dt in the left ventricle is usually reached at the very end of isovolumic contraction, so that its value depends on the level of aortic diastolic pressure. The force/velocity curves plotted from data obtained from isotonic contractions of isolated muscle represent the velocity of shortening of the CEs as a function of their afterload. These curves indicate that with a fixed preload, VOL. 20, NO. 1, JULY,

5 JOCHIM AND BEHRENDT (-dwdt),, at a given afterload rises with an increase in contractility. In the intact ventricle during the preejection phase, the CEs are shortening at the same rate as the SEs are lengthening; their afterload is the rapidly increasing tension. Provided that preload (end-diastolic tension) is unchanged, then (-dudt),, at some given level of afterload should be an index of contractility, represented in the following equation: By definition, dt/dl is the stiffness (S) of the SEs. S (-dwdt)c, = dt/dt It has been shown experimentally that S is linearly related to the tension (T) in SE by the equation: S = KT + C. The constant C is small and is usually ignored, and it is assumed that the constant K has approximately the same value in different mammalian species (cat, dog, man). We may then substitute KT for S and write: KT (-dwdt)ce = dt/dt or (-dudt),, = 1.K (T,) If we assume that we may use intraventricular pressure as an index of muscle wall fiber tension (a somewhat hazardous assumption!), then: (-dwdt),, = 1 (7) K This equation indicates that the quotient (dp/dt)/p is an index of contractility, but only if preload is kept constant and if the same value of P is used in the beats in which contractility is to be compared. It is thus apparent that this quotient, although frequently used, is no better than dp/dt alone as an index of contractility. Other criteria have been utilized. Some, such as ejection fraction and peak velocity or acceleration of aortic flow, depend on measurements during the ejection phase of systole. All, however, are affected to a greater or lesser extent by preload and afterload. Myocardial Function in Patients Assessment of myocardial function in patients poses even greater problems. The usual measurements available clinically - i.e., arterial and atrial pressures, heart rate, and cardiac output -reflect cardiac function, but this is not the same thing as myocardial contractility. Cardiac function is the net result of the interplay 34 THE ANNALS OF THORACIC SURGERY

6 Evaluation.f Myocardial Function of many factors, such as the peripheral resistance and heart rate as well as the state of myocardial contractility. This important distinction is frequently ignored in clinical studies. For example, data on cardiac output versus left atrial pressure before and after a cardiac operation may be useful in evaluating changes in overall circulatory function, but they do not necessarily permit evaluation of myocardial contractility. In the patient, as in the papillary muscle preparation, two phases of myocardial performance can be studied (Fig. 3). Thepreejectionphase - i.e., the period of left ventricular contraction prior to aortic valve opening - is analogous to the isometric contraction of a muscle preparation. Tension (pressure) is developed, but no change in length (volume) occurs. The left ventricular (dp/dt),,, and variables derived from it, such as (dp/dt)/p and time to (dp/dt),,,, are all measures of the isovolumic (isometric) phase of contraction as pressure is developed without flow. V,,, and VCE, the velocity of shortening of contractile elements, can be calculated from the initial portion of the left ventricular pressure curve. On the other hand, the ejection phase of the contraction may be compared to the isotonic shortening of an isolated muscle. This is of obvious importance to the clinician, since it bears directly on the events that occur after aortic valve opening when the heart pumps blood. The classic means of measuring this is the Starling FIG. 3. The variables measured during isometric contraction in the pafnllary muscle (top left) and preejection in the whole heart (bottom left) or during isotonic contraction of the muscle (top right) and the ejection phase in the whole heart (bottom right). The solid segments are the ones of interest in the respective pressure and flow curves (bottom panels). (See Appendix for symbols and abbreviations.) ISOMETRIC CONTRACTION Peak developed T dt/dt Time to peak T tension tronsducer ISOTONIC CONTRACTION 0 tension transducer prelwd AL dl/dt Force -Velocity Curve Vmax Po PRE-EJECTION PHASE EJECTION PHASE l- a 9 I W 8 B - max dp/dt Time to max dp/dt 'CE dp/dt/p /etc. Stroin pow signal Cardoc Output Stroke Volume Starling Curve sw vs. EDV or EDP FmX Max df/dt Eject. Froct. Vmax,V~ VOL. 20, NO. 1, JULY,

7 JOCHIM AND BEHRENDT t LVEW FIG. 4. Lgt ventricular ejection fraction (above) and left ventricular (dpldt),,,,, (below) measured in a dog while left ventricular end-diastolic pressure (LVEDP) was rapidly varied as heart rate, afterload, and contractility were kept constant. The marked dependence of ejection fraction and dpldt on p-eload are evident. curve, which relates the heart s capacity for performing work to its preload. The ejection fraction is another measure of muscle shortening. V,,, and VCE can be derived from analysis of wall motion, but this is a complex calculation involving many assumptions. In clinical application all these indicators of myocardial function are difficult to measure with precision. For example, the ejection fraction is usually determined by estimations of systolic and diastolic volumes from angiocardiograms or, more recently, echocardiograms. These calculations necessitate assumptions regarding the shape of the ventricle that are particularly questionable in disease states in which contraction is often not uniform. There is, however, a more fundamental concern with the way in which these measurements are used clinically. None of them is independent of the preload, afterload, and heart rate - any more than are measurements made with isolated papillary muscle -and this fact is often ignored. For example, ejection fraction is commonly used as an index of myocardial function before and after coronary bypass operation. This is invalid unless the variables that affect cardiac function are comparable in the two study periods. Figure 4 illustrates the clear dependence of ejection fraction upon preload in an isolated left ventricle preparation with constant heart rate and afterload. In the clinical situation it is difficult to control all these variables, but at least those which cannot be controlled must be recorded and their effects estimated when comparisons are made. Conclusions In summary, none of these indicators of myocardial contractility is perfect, although each is of some value. Each depends to some extent on preload, afterload, or heart rate. With whole animal preparations in our laboratory we measure several, since each bears on a different aspect of myocardial function. We record (dp/dt),,, as a measure of the preejection phase, and the Starling curve and ejection fraction (by thermodilution) as measures of the ejection phase. For the 36 THE ANNALS OF THORACIC SURGERY

8 Evaluation of Myocardial Function Starling curves stroke work is plotted against end-diastolic volume instead of end-diastolic pressure to eliminate the possible effects of changing compliance. In all recordings heart rate, afterload, and preload are controlled. The ideal index of contractility - which is easy to measure; independent of preload, afterload, and heart rate; and known to reflect significantly positive inotropic interventions - has not been found. Discussion (Drs. Jochim, Behrendt, Gay, Morrow, Michaelis, Brody, Levitsky, Buckberg, and Feinberg) In clinical practice, left ventricular contractility is often estimated by calculating the ejection fraction from cineangiograms, m,easuring the LVEDP, or calculating V, by extrapolation from left ventricular pressure tracings. Echocardiograms offer another means of measuring ejection fraction or wall motion. Each of these is subject to errors in measurement, and some involve highly questionable assumptions in the calculations. More importantly, each depends upon heart rate, preload, and afterload, and these are generally ignored in the clinical setting. Thus it is not surprising that one variable may appear to show reduced contractility while another does not, and that none of these correlate very well with the clinical outcome. Another problem is that ventricular function may not be homogeneous, particularly in diseased hearts, so that application of techniques designed for isolated muscle strips is not necessarily appropriate. Arrhythmias, abnormal sequence of electrical activation, hypertrophy, and areas of hypokinesia make questionable the assignment of a composite number such as ejection fraction or V,. Thus, in the same heart, V, was found to be 50 to 75% of normal during effective premature ventricular contractions (Brody). Obviously, this is the same heart, just a different scheme of conduction. Also, there is no reason to assume that normal heart muscle must have precisely the same level of contractility at every instant during the cardiac cycle. It has been assumed thatk, used in calculating V,,,, has a constant value, not only from one cat to one man but also from one man to another man, from one heart to the next, and from one portion of the heart to another. This is hard to accept in view of the varying temperatures, drug influences, edema, and myocardial fibrosis or hypertrophy that occur in clinical situations. Thus, in practice, simple eyeball inspection of the left ventriculogram is probably as good at present as any of the more sophisticated measures of ventricular function. Calculation of V, depends upon the left ventricular pressure curve with a long linear or nonlinear extrapolation to zero load. Use of maximum measured velocity avoids the extrapolation but not the dependence upon a good pressure recording (Levitsky). In the animal laboratory one can spend hours touching up the tracings, but this cannot be done in the operating room. Catheter transducers placed into the left ventricle through the superior pulmonary vein can be made to yield almost any V, one wants because they may hit the wall or papillary muscles VOL. 20, NO. 1, JULY,

9 JOCHIM AND BEHRENDT or the baseline may change. These mathematical manipulations of the pressure curve depend on the accuracy of the primary data input, which is good enough at present for rough estimates but not good enough to reveal subtle differences. By looking at several of these indicators of ventricular function simultaneously while controlling as many variables as possible, one is less likely to be deceived by any one of them. In the next few years many methods of myocardial protection are going to generate interest. In spite of the difficulties involved, their eventual effectiveness will have to be assessed by making long- as well as short-term measurements. Coronary artery perfusion pressure may influence contractility. These changes may be great if pressure is allowed to vary between 20 and 80 mm Hg. Between 90 and 100 mm Hg the changes are much smaller (Feinberg). Radioautographs have demonstrated that some areas of the heart may not be perfused (Buckberg). When perfusion pressure is raised contractility sc metimes increases, suggesting that new areas previously unperfused are recruited as contractile units. Thus, when studies are compared in the long term, perfusion pressure must be in the optimal range and must be specified for each study. We must remember also that coronary perfusion pressure and flow can vary independently. Available data suggest that unless coronary flow or perfusion pressure is reduced significantly below the normal range, there is little effect on contractility (Jochim). Another approach that could be considered is perturbation analysis (Feinberg). Instead of looking at static entities, a factor such as isoproterenol, exercise, or pacing can be added that perturbs the system. Looking at the magnitude of the responses evoked by standardized doses of these factors would provide another dimension in evaluating function at different points in time. Appendix Symbols used in text and figures:., CE Fmax (df/dt)max sw EDV EDP SE CE LVEDP Maximum velocity of muscle shortening at zero afterload Minimum afterload at which muscle shortening ceases Muscle tension Rate of change in muscle tension Change in muscle length Rate of change in muscle length Maximum rate of change in intraventricular pressure Instantaneous left ventricular pressure Velocity of shortening of contractile elements Maximum velocity of aortic flow Maximum acceleration of aortic flow Stroke work End-diastolic volume End-diastolic pressure Series elastic element Contractile element Left ventricular end-diastolic pressure 38 THE ANNALS OF THORACIC SURGERY