Estimation

of Infarct Size From Serum

MB Creatine Phosphokinase Applications

BURTON

E. SOBEL,

and Limitations

MD,

Results of enzymatic estimates of infarct size have been verified under defined experimental conditions, and close correlations have been obtained between enzymatically and morphologically estimated infarct size in patients. Nevertheless, to provide a basis for improved enzymatic estimates we explored several aspects of the original model. The first order disappearance rate of creatine phosphokinase (CPK) was verified by observed high correlation coefficients of the logarithm of CPK versus time after myocardial infarction in patients or intravenous injection of purified myocardial CPK in dogs. Selected hemodynamic interventions simulating derangements accompanying myocardial infarction including acceleration of heart rate, diminution of cardiac output and reduction of renal or hepatic perfusion in conscious dogs did not markedly alter CPK disappearance. To exclude contributions from noncardiac CPK to enzymatic estimates we performed studies with the MB CPK isoenzyme. Under standard assay conditions, MB CPK was found virtually exclusively in myocardium. Serial serum MB CPK curves paralleled those of total CPK from patients with uncomplicated infarction. Similar MB curves were obtained even in patients whose noncardiac CPK values distorted the total CPK curve after intramuscular injections. The correlation coefficient between infarct size estimated from total CPK and MB CPK was 0.97 in 12 patients with hemodynamically uncomplicated infarction. Thus, hemodynamic perturbations associated with infarction are unlikely to affect CPK disappearance and hence should not lead to spurious enzymatic estimates of infarct size. Furthermore, improved enzymatic estimates can be obtained by quantitative assay of MB CPK, a more specific myocardial marker, avoiding spurious estimates due to contributions from noncardiac enzyme.

FACC

ROBERT ROBERTS, MD KENNETH B. LARSON, PhD St Louis, Missouri

From the Cardiovascular Division, Washington University School of Medicine, St. Louis, MO. This work was supported in part by NIH SCOR in lschemic Heart Disease 1 PI7 HL 17646-01 (Washington University, St. Louis) and Grant RR-00396 from the Division of Research Resources, National Institutes of Health, Bethesda, Md. Address for reprints: Burton E. Sobel. MD, Cardiovascular Division, Washington University School of Medicine, 660 South Euclid Ave., St. Louis, MO. 63110.

474

March

31, 1976

Activity:

The American

Journal

In 1971 we estimated infarct size enzymatically by analyzing serial changes in total serum creatine phosphokinase (CPK) activity.l CPK was utilized as the marker enzyme because unlike many other enzymes such as lactic dehydrogenase (LDH) and glutamic oxaloacetic transaminase (GOT), CPK in the heart is confined virtually exclusively to myocardium rather than to connective tissue or other nonmyocardial elements. We first examined the hypothesis that myocardial CPK depletion was related to infarct size in experimental animals and demonstrated that rabbits with coronary occlusion exhibited decreased myocardial CPK activity in proportion to infarct size estimated morphologically.2 Subsequently, in collaboration with Maroko, Braunwald and their co-workers,” we found that open chest dogs with coronary occlusion exhibited myocardial CPK depletion 24 hours later in proportion to the severity of the initial ischemic insult assessed electrophysiologically soon after coronary occlusion. Furthermore, CPK depletion 24 hours after coronary occlusion correlated with histologic and electron microscopic criteria of necrosis in the same sites. of CARDIOLOGY

Volume

37

INFARCT SIZE FROM MB CPK-SOBEL ET AL.

Having shown that CPK depletion reflected apparent infarct size under defined experimental conditions, we then examined relations between serum CPK changes and myocardial CPK depletion in conscious dogs subjected to experimental coronary occlusion.ls4 Conscious dogs were used to avoid spurious estimates of myocardial CPK release reflecting release of enzyme from noncardiac sources. CPK released from the heart into the circulation was estimated from serial serum CPK changes analyzed with a simple mathematical model* based on the concept that the rate of change of serum CPK activity reflected two competing phenomena: release of CPK from the heart and disappearance of CPK from the circulation with a first order fractional disappearance rate, kd. In these animals with hemodynamically uncomplicated infarction, CPK lost from the heart (measured by analysis of myocardium) correlated well with CPK released into the blood, estimated from serum CPK changes.1*4 On the basis of earlier observations by others,596 it was not surprising that marked serum enzyme elevations were associated with extensive infarction. Nevertheless, we were encouraged by the observed proportionality between quanitative estimates of CPK released from the heart and myocardial CPK depletion measured directly. After the initial formulation and evaluation of enzymatic estimation of infarct size, refinements and improvements in methodology proceeded in several directions.7-g This report points out some applications and limitations of current methods for enzymatic estimation of infarct size, describes some of the progress made in strengthening the approach and delineates some areas of future investigation likely to prove particularly useful. Potential

Pitfalls in Enzymatic Infarct Size

Estimation

of

Since results of enzymatic estimates of infarct size are directly dependent on values used in the calculations for the CPK disappearance rate (kd), the ratio of CPK released into the circulation compared with that depleted from myocardium (CPKn/CPKp) and the distribution volume of the enzyme (see Appendix), the variance of these measurements and their accurate estimation in individual experimental animals and patients requires critical evaluation.7 In our initial studies, distribution volume of CPK was estimated by the dilution principle after intravenous injection of partially purified canine myocardial CPK in conscious dogs. Although our initial estimate of the distribution volume was larger than that of plasma volume, more recent experiments (see Results) indicate that the enzyme is distributed in a volume equivalent to plasma volume. Our initially high estimate probably reflected inadequate protection of extracted myocardial CPK used in the earlier study, resulting in partial denaturation of injected enzyme. A description of calculations currently used in our laboratory is included in the Appendix. l

The modified estimate of distribution volume does not require alteration of previously reported estimates of infarct size because the proportionality constant between CPK released from the heart and CPK depleted (measured directly) was derived empirically in conscious dogs. The CPK&PKn ratio has remained quite constant in a variety of experimental circumstances involving conscious dogs subjected to coronary occlusion in which the ratio could be determined by direct measurement of myocardial CPK depletion at the conclusion of each experiment and at selected intervals during the course of experiments.1y4J0-12 With the mathematical model used to analyze serial serum CPK changes, results of calculations of infarct size would not be spuriously affected by variation in the rate of washout of enzyme from the heart as long as the CPKnICPKn ratio remained constant during the experiment and from animal to animal. Although we have observed some variation in this ratio in conscious dogs subjected to coronary occlusion and also subjected to interventions including administration of propranolol,i3 reperfusion of myocardium1° and acceleration of heart rate by ventricular pacing,i2 the ratio has consistently been observed to average 0.15 within a range of f0.016 (standard error of the mean). In our initial reports1 we noted that the fractional disappearance rate (kd) of CPK from the circulation of conscious dogs varied substantially from animal to animal. For convenience, in preliminary applications of enzymatic estimation of infarct size to patients, we used an average value of kd. Norris et al.* have suggested a modification of the method in which kd is estimated for each individual patient from analysis of the terminal portion of the serum CPK disappearance curve after enzyme release from the heart has presumably ceased. Additional evaluation of the variance of kd appeared to us to be needed (see Results) because of the direct dependence of estimated infarct size on the value used for kd. If kd varied substantially as a result of hemodynamic perturbations likely to occur in association with acute myocardial infarction, enzymatic estimates of infarct size would be distorted. Accordingly, we evaluated the effects of several hemodynamic interventions on the disappearance of CPK from the circulation in conscious dogs14 (see Results) and, in addition, with the use of radioactively labeled CPK, we examined the effect of infarction itself on CPK disappearance after injection of radioactively labeled enzyme. Differentiation

of Cardiac

from Noncardiac

CPK

Estimates of infarct size in the clinical setting might be influenced by CPK released from tissues outside the heart. Because the MB isoenzyme of CPK appears to be a relatively specific marker of heart muscle, we were eager to estimate infarct size on the basis of serial changes in MB CPK activity in serum rather than of changes in total CPK activity alone. We have previously described a quantitative assay for MB CPK activity not requiring scanning of elec-

March31, 1976

The American Journal of CARDIOLOGY

Volume 37

475

INFARCT s1.x

FROM ~6

CPK-SOEEL

ET AL

trophoretic media, but the procedure is somewhat laborious.15 Accordingly, a new method was developed utilizing separation of individual CPK isoenzymes by batch adsorption with glycophase DEAE glass beads (Corning Glassware, Medfield, Mass.).l’j After the assay had been developed and validated, we examined the distribution of the MB CPK isoenzyme in extracts of numerous human tissues including myocardium obtained at surgery. With this quantitative assay procedure, myocardium was found to be virtually the only human tissue surveyed containing appreciable amounts of the MB CPK isoenzyme.17 Accordingly, one would anticipate that estimates of infarct size based on analysis of serial changes in MB CPK activity in serum would be less prone to error than estimates based on total CPK. Therefore, in this study we analyzed serial changes in serum MB CPK activity in serum from patients with uncomplicated acute myocardial infarction, estimated infarct size from the MB CPK changes observed and compared results with estimates based on total CPK activity in the same patients. Patients selected for initial evaluation were those with uncomplicated infarction, a group in whom enzymatic estimates of infarct size based on total CPK activity were most likely to be valid. Prediction

of Infarct Size

In previous reports, we projected serum CPK changes anticipated in individual conscious experimental animals and patients on the basis of changes observed during the first 5 to 7 hours after the onset of enzymatically detectable myocardial infarction.4 The principle underlying the initial approach was that serial serum CPK changes in general conformed to an arbitrarily selected mathematical function, the lognormal function. Projected enzyme curves, obtained by least squares approximation of the best-fit curve of this type in each case, gave rise to estimates of predicted infarct size. When these estimates were compared with estimates of observed infarct size, based on analysis of actual serum CPK values, reasonable agreement was obtained in conscious experimental animals and in patients with hemodynamically uncomplicated acute myocardial infarction. Results in a prospective series of patients studied recently at Washington University with this approach are shown in Table I. As can be seen, predicted and observed values for infarct size correlate fairly closely and the average ratio of observed to predicted infarct size approximates unity. However, substantial variation between estimates of predicted and observed infarct size occurred in several individual patients, in keeping with our previous observations in conscious dogs with experimental coronary occiusion and in patients. Thus, when the prediction technique is used to evaluate the effect of a therapeutic intervention on the evolution of myocardial necrosis, it must be recognized that no conclusion can be drawn from a single patient or a single experimental animal. On the other hand, with small groups of patients or small groups of experimental animals, comparison of ob476

March 31. 1976

The American

Journal

of CARDIOLOGY

served with predicted infarct size appears to provide a useful tool in evaluating therapeutic interventions objectively.” The empirical approach initially described for prediction of infarct size4 is somewhat analogous to projecting the growth of individual grade school children on the basis of an arbitrary assumption that children grow linearly with respect to time. One can then project the growth of any given child by measuring his height at several intervals over time, estimating the slope of the best-fit line relating the values obtained, and projecting growth from the line derived from this estimated slope. Hypothetically, a more accurate estimation of growth could be obtained by modeling the growth process, that is, by defining the specific variables, such as activity, nutrition, endocrine status and other physiologically significant factors, on which growth of individual children depends. If these variables can be identified and estimated in an individual child, more realistic projections should be obtainable. By analogy, efforts to improve predictions of infarct size in our laboratory have been directed toward developing a physiologically rather than an empirically based model of factors influencing the release of CPK from the heart and its disappearance from the examining and refining circulation,? experimentally the model in order to make it more closely conform to measurable physiologic variables, and projecting serum CPK changes based on the physiologic model rather than on an arbitrary mathematical function. Physiologically based models offer not only the promise of improved projections of serum enzyme changes, and hence better predictions of infarct size in individual patients, but also the likelihood that refined hypotheses will be developed, capable of experimental evaluation, bearing on the processes governing behavior in the circulation of CPK and other biochemical markers of myocardial infarction.

TABLE

I

Observed (ISO) and Predicted (ISPI infarct Size in Patients With Acute Myocardial Infarction Without Shock

Case no.

2 3 4 5 6 7

8 9 10 11 12 13 14 15 16

ISO”

ISP’

Overall SEt

Peak CPK (mlU/ml)

9

8 20 23 27 54 68 12 19 24 132 23 209 64 JO 31 13

21 52 81 30 146 59 29 24 29 181 105 225 113 165 75 28

109 315 393 327 672 622 126 7 74 207 1087 1044 1666 675 1012 382 193

30

21 27 43 73 16 21 27 113 75 233 50 80 35 12

Expressed as CPK-g-eq (see Table IV, footnote 21. +Overall standard error of the curve projected from data obtalned during the first 7 hours after the initial CPK elevation. l

Volume

37

INFARCT SIZE FROM MB CPK-SOBEL

Materials

and Methods

Assay of CPK isoenzyme activity: CPK activity was assayed fluorometrically with the Rosalki procedure.ls Individual CPK isoenzymes were separated with a microbatch procedure developed recently in our laboratory utilizing glycophase DEAE glass beads.16 Under selected conditions of low ionic strength, the glass beads adsorb the MB CPK isoenzyme preferentially. After the glass beads have been separated from supernatant fractions of the sample, the MB isoenzyme can be desorbed by increasing the ionic strength of the medium and the MB isoenzyme activity can be recovered quantitatively. In these studies, 50 ml polycarbonate tubes containing 300 mg of dry glycophase DEAE glass beads, mesh size 120 to 200, were preweighed. One ml of 100 millimolar Tris-hydrochloride, pH 8.0 and 3 millimolar dithiothreitol followed by 0.125 ml of sample were added to each tube. Equilibration for 3 minutes resulted in complete adsorption of MB CPK. The beads sedimented spontaneously, and the supernatant fraction was recovered quantitatively. The beads were washed twice with 45 ml of buffer, and MB CPK was then desorbed by adding 100 ~1 of 3.8 molar sodium chloride per gram of residual aqueous phase. Under these conditions and with serum samples constituted with human MB and MM CPK (isolated from myocardium obtained at necropsy) in a ratio of l:lO, over a range of 50 to 741 mlU/ml total activity, recovery of the MB CPK isoenzyme averaged 97 percent with a standard deviation of 3 percent. In samples from patients with acute myocardial infarction reproducibility of the assay exhibited a standard deviation of less than 5 percent (no. = 25). Tissue extracts of human skeletal muscle, myocardium, gastrointestinal tract, brain and other organs were obtained from patients at the time of surgery and prepared in sucrose, EGTA, mercaptoethanol as previously described17 prior to assay for CPK isoenzyme activity with the microbatch procedure. Implementation of hemodynamic perturbations in conscious dogs: To evaluate the effects of hemodynamic perturbations on the disappearance rate of intravenously injected canine myocardial CPK in conscious dogs, hemodynamic interventions were induced simulating those seen in patients with myocardial infarction. Animals were fitted with instruments at least 1 week before each study during surgical procedures performed under anesthesia with sodium thiamylal, 10 mg/kg bodyweight, and 0.05 percent halothanem. Depending on the specific experiment, an inflatable balloon cuff catheter was placed around the inferior vena cava or arterial cuffs were placed around renal, hepatic or celiac arteries for later use in diminishing renal or hepatic perfusion. Cardiac output was monitored with electromagnetic flow probes around the aorta, and heart rate was altered with the use of epicardial pacing wires sutured into the right ventricle. Cannulas, probes and pacing wires were exteriorized, and the animals were permitted to recover completely so that resting heart rate was less than 100 beats/min and serum CPK activity less than 50 mlU/ml. After the animals had recovered completely from preliminary surgery, partially purified canine myocardial CPK radioactively labeled with 14C-formaldehyde was injected intravenously 4 hours before implementation of selected interventions. Blood samples were obtained through polyethylene jugular venous catheters for analysis of CPK activity and radioactivity every 30 minutes for 4 hours before each intervention and subsequently for 4 hours to determine the effect of the intervention on the disappearance rate of radioactively labeled intravenously injected CPK. In addi-

ET AL.

tional experiments, the effects of inactivation of the reticuloendothelial system on CPK disappearance was evaluated by inducing blockade with intravenous injection of zymosan, 10 mg/kg. Prediction of infarct size based on a physiologic rather than an empirical model of CPK release: Our initial approach to developing a physiologically based model of CPK release from myocardium undergoing infarction entails the following assumptions: 1. CPK is released within the heart and diffuses passively in the tissue toward permeable capillaries or lymphatics. 2. CPK enters the circulation by convection and is distributed rapidly and uniformly within the plasma volume. 3. The appearance of CPK in the circulation is rate-limited by diffusion (Step 1). 4. The disappearance of CPK from the circulation is first order and constant during the study (but not necessarily the same in different animals or patients). For purposes of the model, the zone of infarction is envisioned as containing n spheres each with CPK content m, radius a, and surface area A = 4 ra2. By solving the diffusion equation for this geometry, and on evaluating the total CPK flux through the surfaces of the hypothetical spheres, we obtain an expression for the rate of CPK appearance in the circulation with parameters to be determined. The parameters in the appearance rate expression are CPKR, the total amount of CPK released per ml of blood; T, a parameter incorporating the CPK diffusion constant (a function of its molecular weight); and average CPK diffusion path length, a. By combining the modeled CPK appearance function with a compartmental model of CPK disappearance,4,7 we construct a tracer conservation condition in the form of a first order differential equation describing changes in CPK activity per unit volume of blood. With the use of least squares curve fitting techniques implemented with a digital computer we then obtain values of these parameters that yield the best fit of the theoretical enzyme activity curve to observed serial serum CPK data. Operationally, in applying and evaluating this model experimentally, parameters of the modeled appearance function are fit to observed serum CPK changes. The extent to which the model describes the behavior of CPK within the circulation can be evaluated by comparing curves fit by the model to actual enzyme data in patients with acute myocardial infarction and by considering the extent to which parameter estimates resulting from the fitting procedure conform to physiologically realistic values. For example, if a parameter for length of diffusion path was estimated to be of the order of meters rather than millimeters, clearly the model would have little applicability to the biologic phenomenon being examined. In evaluating this preliminary approach to physiologically modeling CPK release in patients, we adopted as our criterion the goodness of fit of curves to actual serial CPK data. Using this criterion, we compared the curves derived from the diffusive transport model to curves obtained with the empirical algorithm used previously, the lognormal function.

Results Disappearance of MB and Total CPK From the Circulation Studies in experimental animals: In conscious dogs not subjected to hemodynamic or physiologic perturbations, enzyme activity and radioactivity of radioactively labeled canine myocardial CPK decreased in a generally parallel fashion suggesting that

March 31, 1976

The American

Journal

of CARDIOLOGY

Volume

37

477

INFARCT SIZE FROM MB CPK-SOBEL

ET AL.

TABLE

II

CPK Disa ppearance

Intervention None Heart rate accelerated from F

1500

g

1000

ET AL.

q

MM MB

CPK /&j CPK q

BB

CPK I

heart

E 0 500 n

intestine

IL____

fashion typical of that seen in patients with uncomplicated infarction (Fig. 5). The observed distribution of CPK isoenzymes in human tissues surveyed and the lack of effect of intramuscular injections on serial serum MB changes in patients with infarction suggested that estimation of infarct size based on MB CPK activity would be useful clinically. In our initial studies of total CPK, we compared myocardial CPK depletion (measured directly) with CPK appearance in the circulation in conscious dogs. However, as illustrated in Figure 6, canine myocar-

:s;:;

liver

spleen kidney

lung

dium contains only a small amount (less than 2 percent) of MB CPK detectable with the micro-batch assay procedure used in our study, and thus it was impossible to perform analogous experiments using canine CPK MB as the myocardial marker. Comparison of estimates from total CPK and MB CPK: To evaluate estimates of infarct size based on changes in serum MB CPK, studies were performed in patients with hemodynamically uncomplicated myocardial infarction in whom estimates based on total CPK could be used for comparison. The two

200

160

80

FIGURE 7. Relation between infarct size estimated enzymatically from serial changes in MB CPK activity and from serial changes in total serum CPK activity in patients with hemodynamically uncomplicated acute myocardial infarction. The correlation between these two enzymatic estimates was close despite the use of different values for the fractional disappearance rate for MB CPK compared with total CPK and the use of different values for the amount of MB compared with total CPK in normal and ischemic myocardium.

I-

4c l-

0

INFARCT

March

31, 1976

40 SIZE,

from

The American

80 serial

Journal

120 160 200 Total CPK changes (CPK-g-eq)

of CARDIOLOGY

Volume

37

461

INFARCT SIZE FROM MB CPK-SOBEL

ET AL.

500 l

ExperImental

---Lognormal

CPK

Values

Fit

-Diffusion-Model

F It

kd = .00197 ISO= Tau=1423

TIME

(Hours)

FIGURE 8. Relation between observed serum CPK values, the best-fit lognormal curve conforming to these values, and the best-fit curve conforming to the observed data derived from the physiologically based diffusive transport model. In this case, the best-fit lognormal curve was quite similar to the curve obtained with the physiologically based algorithm. However, the terminal portion of the lognormal curve deviated substantially from the late observed serum CPK values. In this figure and in figure 9 the value for k,, used is the value obtained from the diffusive transport model. IS0 = infarct size estimated enzymatically from observed serum CPK changes; tau = a parameter in the diffusive transport model obtained from the fitting procedure, as described in the text and Table IV.

1500 I l

Experimental

---Lognormal

CPK

Values

Fit

-Diffusion-Model

Fit

kd = .00156 ISO= Tau= 1396

TIME

(Hours)

of infarct size differed only in the following fashion (see Appendix): Individualized values for kd for both total CPK and the MB CPK isoenzyme were utilized as indicated in Table III, and calculations of infarct size took into account the measured differences in myocardial CPK and MB CPK activity in human myocardium obtained at surgery (Fig. 3). As can be seen in Figure 7, estimates of infarct size based on MB and total CPK changes correlated closely (r = 0.97) despite considerable differences in the apparent disappearance rates of MB and total CPK in individual patients.21 estimates

Evaluation Release

of a Physiologically

Based Model for CPK

As indicated under Methods, improved predictions of infarct size may result from utilization of physio-

482

March 31, 1978

The American

Journal

FIGURE 9. Relations between observed serum CPK values, the best-fit lognormal curve and the best-fit diffusive transport model curve obtained from data in a patient with acute myocardial infarction. In this case, the diffusive-model curve coincided substantially more closely with the observed data than with the best-fit lognormal curve. Abbreviations as in Figure 8.

of CARDIOLOGY

logically rather than empirically based algorithms. The physiologically based diffusive transport model used in this study fit actual serum CPK changes better than the previously used empirical lognormal function best-fitting the same data (Fig. 8 and 9). The improved fit was most evident in the ascending and terminal portions of the serial serum CPK curves. Thus, although the improved fit would not necessarily make a substantial difference in infarct size calculated from the fit curve, conformity to the actual behavior of serum CPK changes was closer with the physiologically based model. Estimates of infarct size based on actual serum CPK values correlated closely with estimates based on best-fit values obtained with the physiologically based model (Table IV). In addition, estimates of parameters were consistent with the thickness of adult myocardium.

Volume

37

INFARCT SIZE FROM MB CPK-SOBEL

TABLE

Discussion

Diffusion

Clinical Applications The formulation for enzymatic estimation of infarct size presented several years ago has proved useful in evaluating patients with acute myocardial infarction. Recently, Bleifeld et a1.g demonstrated that infarct size estimated enzymatically in patients correlated closely (r = 0.98) with infarct size assessed morphologically in autopsy studies of patients dying relatively soon after infarction. The same group22 demonstrated that the alterations in ventricular compliance soon after infarction were related to the apparent extent of ventricular damage estimated enzymatically. Norris et a1.8 found that estimates of infarct size based on serial changes in serum CPK activity correlated with independent indexes of the severity of myocardial infarction these workers had previously utilized. Maley et a1.23utilized the prediction technique we initially described using empirical algorithms4 to evaluate potentially therapeutic interventions in patients with myocardial infarction; as in our experience, they observed a close correlation between observed and predicted infarct size in patients with infarction not subjected to therapeutic interventions. We have shown that infarct size estimated enzymatically is correlated with the severity of ventricular dysrhythmia seen soon after the onset of myocardial infarction,24 with prognosis25 and with ventricular performance assessed with radionuclides.26 In addition, we have evaluated the effect of several therapeutic interventions on infarct size by comparing their effects with observed versus predicted infarct size estimated enzymatically.ll In several series of patients, we have found the general relation between observed and predicted infarct size to be close when small groups of patients were compared.4J1,27T28 Thus, despite the need for continued refinement and evolution of enzymatic estimation techniques, clinical applications have already proved useful. Moreover, independent indexes of the severity of infarction and of its extent indicate that enzymatic estimates of infarct size, despite their imperfections, reflect the extent of infarction. Clinical implications of Enzymatic Study Data presented in this report are addressed to three aspects of enzymatic estimation: (1) further characterization of the disappearance of CPK from the circulation, particularly in the presence of hemodynamic perturbations such as those associated with acute myocardial infarction; (2) estimation of infarct size by analysis of serial changes in serum MB CPK isoenzyme activity, a more specific marker of myocardial damage than total CPK; and (3) evaluation of a physiologically based in contrast to an empirical algorithm for describing CPK appearance in the circulation in order to gain increased understanding of physiologic processes governing changes in serum

ET AL.

IV Model

Parameter

Estimates* Estimated Infarct Size (CPK-g-eq)+

Case no. 1 2 3 4 z : 9 10

(CPK)P (mlU/ml)

7 (min)

krr fmin-‘)

From Model

1186 3463 1507 1114 829 545 2918 2629 119 440

1423 1396 1185 752 633 722 2980 2592 422 638

0.0020 0.0016 0.0008 0.0012 0.0009 0.0011 0.0025 0.0019 0.0011 0.0011

42 174 87 49 31 26 116 117 19 16

From Observed Serum CPK Changes 49 192 88 5% 37 30 115 111 23 21

* Parameters estimated by least squares approximation including: (CPK), = CPK released per unit volume of blood; r = a parameter exoressed as the interval durina which diffusion of CPK occurs within ‘the tissue; kd = the fractional disappearance rate of CPK from the circulation. r is a function of the average diffusion path length and of Dr-PK. the diffusion coefficient of CPK in myocardium; its dimensions are minutes. Estimates of r when combined with a DCPK value calculated from the molecular weight of CPK result in an estimate of the average diffusion path length for CPK within myocardium. This value averaged approximately 3 mm. t One CPK-gram-equivalent = that amount of myocardium undergoing infarction liberating an amount of CPK into the circulation equivalent to the amount released from 1 g of homogeneously necrotic myocardium.

CPK and to improve prediction of infarct size based on projected serum enzyme values. Hemodynamic disturbances and CPK activity: Results indicate that hemodynamic perturbations, even when they are profound, do not greatly influence the disappearance rate of CPK from the circulation. Their relative lack of influence appears to reflect the important role of the reticuloendothelial system in the clearance of enzyme from the circulation.21 Furthermore, although previously recognized variation in the fractional disappearance rate of CPK activity from patient to patient is a prominent finding, the disappearance of the MB CPK isoenzyme and of total CPK, estimated from descending portions of serum enzyme curves, conforms closely to a monoexponential relation with a high correlation Accordingly, although the numerical coefficient. value for enzymatically estimated infarct size is of course directly dependent on the value used for kd, the assumptions that CPK disappearance is monoexponential and that the disappearance rate remains constant during the course of a study appear to be justified. Coupled with the suggestion of Norris et a1.s for individualizing kd, these results suggest that valid estimates of infarct size can be obtained enzymatically in spite of substantial variations in kd from patient to patient. MB CPK versus total CPK activity in estimation of infarct size: In evaluating the distribution of the MB CPK isoenzyme in human tissues, we foundi that only myocardium contains substantial amounts of this moiety. Furthermore, in patients with hemodynamically uncomplicated myocardial infarction,

March 31, 1976

The American Journal of CARDIOLOGY

Volume 37

483

INFARCT

SIZE

FROM

MB

CPK-SOBEL

ET

AL

estimates of infarct size based on changes in MB CPK activity and those based on total CPK correlated closely even though the disappearance rate parameters used in the two estimates differed substantially. Since interventions such as intramuscular injections, cardiac catheterization, noncardiac surgery and electrical cardioversion2g increase total serum CPK without generally increasing MB serum CPK activity, it appears likely that estimates of infarct size based on analysis of serial changes in the MB CPK activity will be particularly useful in patients with myocardial infarction accompanied by endogenous or exogenous trauma to tissue other than the heart. Projected MB CPK and prediction of infarct size: The results of our study are pertinent to prediction of infarct size on the basis of early serum enzyme changes. Regardless of which algorithm is used for projecting enzyme changes, prediction of infarct size will be facilitated when contributions to elevated serum enzyme activity are primarily or exclusively from the heart. The close correlation between estimates of infarct size based on MB and total CPK activity in patients with uncomplicated infarction suggests that the two are measuring the same phenomenon, that is, infarct size, when hemodynamic sequelae or iatrogenic trauma do not lead to release of noncardiac CPK into the circulation. Considerable experimental work is required to verify the extent to which valid predictions of infarct size based on projected MB CPK values can be obtained. However, it appears likely that predictions based on MB CPK

changes should be more reliable for several reasons including: (1) MB CPK is a more specific marker of myocardial damage than total CPK, and (2) serum MB activity in normal subjects is so low and constant that uncertainties due to background correction are minimized. Despite their shortcomings, predictions of infarct size based on empirical phenomenologic models utilizing best-fit lognormal curves derived from early serial serum CPK changes have already proved useful in evaluating the effects of physiologic and pharmacologic interventions in ischemic injury in conscious experimental animals and patients with myocardial infarction. The inherent limitations in the precision of these empirical approaches have necessitated evaluation of groups of patients to characterize interventions. Our current efforts are designed to develop physiologically based models, to assess them experimentally and to apply them to estimation and prediction of infarct size based on analysis of serial changes in MB as well as total serum CPK activity. Such models should be useful not only in broadening the applicability of enzymatic prediction of infarct size but also in highlighting fruitful areas of investigation concerned with identifying processes governing the release of enzymes and other biochemical markers from the ischemic heart and their behavior in the circulation. Acknowledgment We thank Dr. Jerome R. Cox, Jr. and Ms. Joanne Markham, who devised and implemented the computer programs for estimating the model parameters.

References 1. Shell WE, Kjekshus JK, Sobel BE: Quantitative assessment of the extent of myocardial infarction in the conscious dog by means of analysis of serial changes in serum creatine phosphokinase (CPK) activity. J Clin Invest 50:2614-2625, 1971 2. Kjekshus JK, Sobel BE: Depressed myocardial creatine phosphokinase activity following experimental myocardial infarction in the rabbit. Circ Res 27:403-414, 1970 3. Maroko PR, Kjekshus JK, Sobel BE, et al: Factors influencing infarct size following experimental coronary artery occlusion. Circulation 43:67-82. 1971 4. Shell WE, Lavelle JF, Covell JW, et al: Early estimation of myocardial damage in conscious dogs and patients with evolving acute myocardial infarction. J Clin Invest 52:2579-2590, 1973 5. Killen DA, Tinsley EA: Serum enzymes in experimental myocardial infarcts. Arch Surg 92:418-422, 1966 6. Kibe 0, Nilsson NG: Observations on the diagnostic and prognostic value of some enzyme tests in myocardial infarction. Acta Med Stand 182:597-610, 1967 7. Sobel BE, Larson KB, Markham J, et al: Empirical and physiological models of enzyme release from ischemic myocardium. In, Computers in Cardiology. Long Beach, Calif, IEEE Computer Society. 1975, p 189 8. Norris RM, Whitlock RML, Barratt-Boyes C, et al: Clinical measurement of myocardial infarct size: modification of a method for the estimation of total creatine phosphokinase release after myocardial infarction. Circulation 51:614-620, 1975 9. Bleifeld W, Mathey D, Hanrath P, et al: Synopsis of intravitally determined infarct size and left ventricular hemodynamics in 50 patients with acute myocardial infarction. Circulation. in press 10. Bresnahan GF, Roberts R, Shell WE, et al: Deleterious effects due to hemorrhage after myocardial reperfusion. Am J Cardiol

404

March

31, 1976

The American

Journal

Of CARDIOLOGY

33:82-86, 1974 Shell WE, Sobel BE: Protection of jeopardized ischemic myocardium by reduction of ventricular afterload. N Engl J Med 291:481-486, 1974 12. Shell WE, Sobel BE: Deleterious effects of increased heart rate on infarct size in the conscious dog. Am J Cardiol 31:474-479, 1973 13. Shell WE, Sobel BE: Changes in infarct size following administration of propranolol in the conscious dog (abstr). Am J Cardiol 31:157, 1973 14. Roberts R, Sobel BE: Factors affecting disappearance of cre11.

15.

16.

17.

18. 19.

20.

21.

Volume

atine phosphokinase (CPK) from the circulation (abstr). Clin Res 23:205. 1975 Roberts R, Henry PD, Witteveen SAGJ, et al: Quantification of serum creatine phosphokinase (CPK) isoenzyme activity. Am J Cardiol 33:650-654, 1974 Henry PD, Roberts R, Sobel BE: Rapid separation of serum creatine phosphokinase isoenzymes by batch adsorption with glass beads. Clin Chem 21:844-849, 1975 Roberts R, Gowda K, Ludbrook P, et al: Specificity of elevated serum MB creatine phosphokinase activity in the diagnosis of acute myocardial infarction. Am J Cardiol 36:433-437. 1975 Rosalki SB: An improved procedure for serum creatine phosphokinase determination. J Lab Clin Med 69:696-705, 1967 Klein MS, Shell WE, Sobel BE: Serum creatine phosphokinase (CPK) isoenzymes following intramuscular injections, surgery, and myocardial infarction: experimental and clinical studies. Cardiovasc Res 7~412-418, 1973 Gowda KS, Roberts R, Sobel BE: Detection of myocardial infarction with serum CPK isoenzymes in surgical patients (abstr). Circulation 5O:Suppl lll:lll-109, 1974 Roberts R, Henry PD, Sobel BE: An improved basis for enzy-

37

INFARCT SIZE FROM MB CPK-SOBEL

22.

23.

24.

25.

matic estimation of infarct size. Circulation 52:743-754, 1975 Mathey D, Bleifeld W, Hanrath P, et al: Attempt to quantitate relation between cardiac function and infarct size in acute myocardial infarction. Br Heart J 36:271-279, 1974 Maley T, Gulotta S, Morrison J: Effect of methylprednisolone on predicted myocardial infarct size in man (abstr). Clin Res 22:288A, 1972 Sobel BE, Roberts R, Ambos HD, et al: The influence of infarct size on ventricular dysrhythmia (abstr). Circulation 5O:Suppl III: 111-IIO. 1974 Sobel BE, Bresnahan GF, Shell WE, et al: Estimation of infarct size in man and its relation to prognosis. Circulation 46:640648, 1972

26

27.

28.

29.

ET AL.

Kostuk WJ, Ehsani AA, Karliner JS, et al: C efi ventricular performance after myocardial infarction assessed by radioisotope angiocardiography. Circulation 47:242-249, 1973 deMello V, Roberts R, Sobel BE: Deleterious effects of methylprednisolone in patients with evolving myocardial infarction (abstr). Circulation 52:Suppl ll:ll-100. 1975 Gowda KS, Roberts R, Ambos HD, et al: Salutary effects of external counterpulsation in patients with acute myocardial infarction (abstr). Am J Cardiol 35:140, 1975 Ehsani AA, Ewy GA, Sobel SE: CPK isoenzyme elevations after electrical countershock (abstr). Circulation 48:Suppl IV: IV-129, 1973

APPENDIX Calculation

1. E(t) f(t)

kd

= activity of CPK in blood

CPK,

3. K

DV PCPK

where

(III/ml).

At, = ti+l -

= rate of change of CPK activity due to enzyme being released by heart (IU/(ml min)). = fractional rate of disappearance from blood (min-‘) dE = f(t) dt

2. CPK,

of Infarct Size from Serial Changes in MB or Total CPK (expressed as CPK-gram- equivalents)

E, = E(ti+l) + E(ti) 9 1 2

of CPK and

kdE.

f- = 1

= cumulative activity of CPK released by heart up to time T (IU/ml). =

= proportionality

Values currently

= distribution volume/unit body weight (ml/kg). = proportion of CPK released into blood compared with CPK depleted from the heart (IU/CPK-g-eq)/(IU/g).

kd(MB CPK) = obtained from serum curve = 44 ml/kg DV* = 0.15 PMB CPKt = 96 IU/g$ MB CPKN = 25 IU/g# MB CPKI = 4.1 K Infarct Size Based on Total CPK kd(CPK) DV*

CPKI

section of

FP’K”t

= CPK activity in a homogeneous infarcted myocardium (IU/g).

CPK; K

DV

K=

PcPK(CPKN

-

CPKI)

= infarct size (CPK-g-eq). = body weight (kg). IS = (K)(BW)(CPK,)

5. Given N observed values of CPK activity, E(ti), i= 1,2,. . . , N, CPK, can be estimated from (see 2 above) N-l

CPKrZC

fiAti= i=l

N-l

Ati= E(tN) + kd C EiAti, 1

i=l

March

used for constants follow:

Infarct Size Based on MB CPK

constant

section of

i=l

+ f(k).

2

(Note that E(t, = 0 is assumed.)

E(t)dt. f(t)dt = E(T) + kd s s (hate that CPK,,is a f&nction of T )

“c’(z+k&)

f(ti+l)

T

T

CPKN = CPK activity in a homogeneous normal myocardium (IU/g).

4. IS BW

ti,

AEi = E(ti+l) - Efti),

= = = = = =

obtained from serum curve 44 ml/kg 0.15 680 IU/gZ 180 IU/g 1 5.9 x 10-1

* Estimated as plasma volume for man (Nachman HM, James GW Ill, Moore JW, et al: Comparative study of red cell volumes in human subjects with radioactive phosphorous tagged red cells and T-1824 dye. J Clin Invest 29:258-264. 1950) + Calculated by analogy based on the empirical relation between CPK released calculated from serum changes and myocardial CPK depletion measured directly in conscious dogs with coronary occlusion.*.s t Measured directly as described in text. * Calculated as 14 percent of the estimated total amount of CPK remaining in a region of infarction n Calculated by analogy from percent of myocardial CPK depleted measured directly in conscious dogs 48 hours after coronary occlusion.*

31, 1976

The American

Journal

of CARDIOLOGY

Volume

37

465