**Department of Physiology
Mayo Medical School
Rochester, MN
We have demonstrated that there is a cyclic variation in the total volume of the heart (total contents of the pericardial sac). However, the intracardiac cycle change in total heart volume is considerably less than the combined right and left ventricular stroke volume. It is hypothesized that the maintenance of a near constant heart volume relationship is particularly important for the right heart which has a normally low work load and thus any increased work load required to move extra cardiac structures would be a significant increase over baseline. Evidence in shown supporting this hypothesis. The accompanying movie demonstrates our work using 2-D echo and MRI verifying a constant heart volume relationship in awake humans. (The video accompanying this article received the "1987 Outstanding Film Award" from the American College of Chest Physicians.)
The fetal heart functions in utero within a high inertia, noncompressible environment whereby both the lungs and uterus are fluid filled, and yet, in this high inertia noncompressible environment, this tiny heart manages to eject blood. One might be led to the hypothesis that it is a functional imperative for the heart to develop in such a way so as to pump blood while maintaining a constant total heart volume, where the total heart volume is defined as the total contents of the pericardial sac. In fact, the fetal heart environment is not dissimilar to the experimental environment set by Sass and colleagues1 whereby, in order to determine if breathing of liquid fluorocarbon could prevent the blood oxygen desaturation occurring during exposure to high gravitational forces, dogs were placed in rigid water-filled plexiglass cylinders, the lungs were filled with the liquid fluorocarbon, and the fluorocarbon was moved in and out of the lungs by moving the water in and out of the plexiglass cylinder. Indeed, this environment protected the dogs against oxygen desaturation during high G exposure, but in addition a surprising finding was the fact that the cardiac output under these conditions remained essentially normal. This would suggest that either the heart was performing a considerable amount of work shifting blood volumes in the presence of its completely liquid environment, or the heart was able to pump blood without changing its volume. In fact, nearly fifty years ago, Hamilton2 (of Stewart-Hamilton 3,4 dye-curve fame) conjectured that the heart functions at a constant volume. Hamilton's experiments utilized percutaneously-placed fish hooks in the myocardium at the apex and atrioventricular valve plane and monoplane fluoroscopy to demonstrate that the apical markers moved very little while the fish hooks at the atrioventricular valve plane moved vigorously in a base-to-apex direction. It is perhaps the inadequacies of the techniques used by Hamilton which have caused his hypothesis regarding the constantancy of volume to have been largely forgotten. The Dynamic Spatial Reconstructor5has now allowed for a detailed probing of Hamilton's hypothesis, that is, under normal physiologic conditions, the total volume of the heart, defined as the total contents of the pericardial sac, remains essentially constant throughout the cardiac cycle.
For total heart volume to remain constant, one of two things has to happen. Either there is a perfect reciprocal emptying and filling of the atria and ventricles, accomplished by a piston-like motion of the atrioventricular valve plane while the epicardial apex of the heart remains fixed in space, or any mismatch in atrial and ventricular emptying and filling has to be made up for by changes in volume of the myocardium. If one presumes the cardiac muscle to be noncompressible, then any change in myocardial volume must be attributable to a change in myocardial blood volume.
To test whether or not the potential fetal constant heart volume relationship is reflected in the adult, wherein the lungs are air filled, we have utilized a fast dynamic volumetric x-ray CT scanning device called the Dynamic Spatial Reconstructor (DSR)5. The DSR is composed of 14 x-ray guns mounted in a semicircle and 14 juxtaposed television cameras recording projection images of the thorax produced on the hemicylindrical fluorescent screen. Since U. S. television is composed of 240 usable lines, by utilizing the digitized gray scale image from the same line number obtained from each of the 14 cameras, we are able to reconstruct a cross sectional image of the thorax and thus obtain up to 240 spatially-contiguous 0.9 mm thick CT cross sections of the body. Since the gantry rotates at 15 rotations per minute, we can utilize sequential 14 camera image data sets to increase the number of angles of view used in the reconstruction process and thus increase the reconstructed image resolution and/or we can gate together data from multiple cardiac cycles.
We have previously6demonstrated in six anesthetized dogs that the total volume of the heart remains within 5% of its end-diastolic value at end-systole. Because our data throughput has significantly improved due to software enhancements to an image analysis package, implemented on a Sun Microsystems workstation,7we have now been able to evaluate the total heart volume relationship throughout the cardiac cycle. For the data shown here, we have utilized 2/60 second scan apertures (ie: two sequential firings of the 14 x-ray gun array and subsequent signal detection by the juxtaposed television camera compliment) and have gated together four sequential cardiac cycles. Data from two dogs are demonstrated. In both cases, the dogs were anesthetized with Inovar (0.1 cc/kg administered every 30 minutes IM) and 2:1 Nitrous Oxide/Oxygen. Anesthesia was initiated with a loading dose of Inovar (1 cc/5 kg IM) and a single IV injection of 30 mg sodium pentobarbital. Atrial fibrillation which commonly results in dogs from this anesthetic regime is eliminated by a single administration of Xylocaine. The heart was paced with a bipolar tip pacing catheter placed in the coronary sinus with heart rate set at 112 bpm. A radiolucent open-ended liquid-filled catheter was placed in the left ventricle via the left common carotid artery for pressure measurements, and a #7 Rodriquez catheter was placed at the vena cava/right atrial junction for radiopaque contrast injection. In the first dog (representative of four dogs studied throughout the entire cardiac cycle) the scans were accomplished without contrast injection. In the second dog, 1.5 cc/kg Iohexol was injected into the vena cava/right atrial junction and scan data was gathered during the time when the right heart and then the left heart was opacified. Utilizing a 2/60 second scan aperture in conjunction with 112 bpm cardiac pacing, we obtained 16 reconstructed volumetric data sets spanning a single cardiac cycle.
To quantitate the total heart volume in both dogs and, in the case of the second dog, to quantitate the volume of the right and left heart chambers and the total myocardial volume, we utilized a thresholding and region-growing technique previously reported. 8
As demonstrated in figure 1, typifying the plots from the series of four dogs studied with no contrast injection, we confirmed our previous finding that the total heart volume remains within 5% of the end-diastolic value throughout the cardiac cycle. We have plotted the total volume of the heart along with the left ventricular pressure curve. The dogs studied in this way thus far have all demonstrated the slight rise in total heart volume just at the onset of ventricular depolarization, and a reduction in total heart volume, with a minimum reached during the left ventricular isovolumetric relaxation phase. Despite the consistent decrease in total heart volume between end-diastole and end-systole, this volume change is considerably less than the combined stroke volume of the right and left ventricles.
Fig 1. Total heart volume at 2/60 second scan apertures plotted throughout the cardiac cycle with time point one coinciding with end-diastole. Also plotted here is the left ventricular pressure curve. Note that the total volume of the heart remains within 5% of the end-diastolic volume throughout the cardiac cycle; yet, the smoothness of the curve depicting volume change suggests that the sensitivity of our measurement is close to +1% and that there is a real volume change throughout the cardiac cycle even though the change is considerably less than the combined right and left ventricular stroke volume. Also note the slight increase in total heart volume occurring coincident with the onset of ventricular depolarization.
Although the accuracy of our measurements have been shown in the past to be +5%,8 the smoothness of the curve depicting total heart volume change throughout the cardiac cycle indicates that the sensitivity of our measurements is considerably better than 5%. The greatest variation in the measurement comes from the selection of a gray scale value range when selecting the window defining the cardiac region. If one is careful to reconstruct all volume images in the cardiac cycle using identical parameters and to then segment the heart from within the reconstructed images using the same threshold range, the relative measurement of change in cardiac volume is apparently highly accurate. This speaks strongly for the use of some form of computer based decision process in determining the borders of the organ of interest as opposed to manual tracing.
We have previously hypothesized9that it might be more important for the right heart volume to remain constant than it is for the left since the right heart functions at a low pressure and thus a low work load (work being pressure x volume) and any increased work load on the right heart is a significant portion of its total work load. To begin testing this, we evaluated the change in total right heart chamber volume versus total left heart chamber volume throughout the cardiac cycle in one dog. Because of the uncertainty10of our ability to repeatably identify the mitral and tricuspid valve planes in each volume image throughout the cardiac cycle, we have limited our measurements to the total right and total left heart chamber volumes. Since our measurement of total heart volume is defined as the total contents of the pericardial sac, when we talk about right heart chamber volume, this includes portions of the pulmonary artery and vena cavae which reside within the boarders of the pericardial sac. Similarly, the total left heart chamber volume includes a portion of the aortic root residing within the pericardial sac.
Fig. 2. Left panel represents a shaded surface display of the total contents of the pericardial sac imaged at end-diastole. The right panel represents a plot of total heart volume along with a pressure record of left ventricular pressure. Here, the total volume of the heart changed by only 3% throughout the cardiac cycle.
The graph on the right of figure 2 shows left ventricular pressure and change in the total heart volume through the cardiac cycle. Note that in this case the maximum volume change between end-diastole and end-systole and back to end-diastole represents only a 3% reduction in total heart volume.
For the purpose of quantitation, we have next separated out the right heart chambers from the left heart chambers, and displayed these volumes separately in figures 3 and 4. In figure 3, we have plotted left ventricular pressure along with the variation of left heart chamber volume. The total left heart chamber volume relative to itself changed by 15-18%. However, left heart volume change relative to the end-diastolic total heart volume is 3%. In figure 4 we have produced a similar plot of total right heart chamber volume, and consistent with our hypothesis, right heart volume through the cardiac cycle is more stable than left heart volume. Right heart volume changed in this case by approximately 8% between end-diastole and end-systole or approximately 1.8% relative to the total heart volume.
Fig 3 Total left heart chamber volume. This is defined as the total blood volume associated with the left heart chambers contained within the pericardial sac and thus includes a portion of the aortic root.
Fig 4 Surface display of both imaged right heart chambers and a portion of the pulmonary trunk and vena cavae also contained within the pericardial sac. Changes in these total left and right heart chamber volumes are plotted along with left ventricular pressure in the right side of each figure. See text for further description.
It is of interest to note that left heart volume changed 3-4%, and right heart volume changed by 1-2% relative to the total heart volume, summing to a 5-6% change; yet, the total heart volume changed by only 2%. This suggests that there may have been a change in myocardial blood volume to make up this difference. As shown in figure 5, we are able to visualize the myocardium and thus obtain a measure of its volume, and by definition, we do measure an increase in total myocardial volume occurring at end-systole (isovolumetric relaxation phase of the cardiac cycle by our definition, defining end-systole as minimum left ventricular volume). There is room for error here in our measurement. To the extent that we define the borders based upon thresholded values, the trabeculations within the chambers could be interposed with opacified blood at end-diastole and thus we would define the chamber borders as being the point of departure of the trabeculations from the main portion of the heart wall. Perhaps at end-systole, the trabeculations might come close enough together that opacified blood between the trabeculations is ignored by the thresholding algorithm and we now pick the surfaces of the ridges formed by the trabeculae as the endocardial surface, thus leading to the calculation that the total myocardial volume increased. To the extent that this measurement is erroneous, the actual volume of the chambers is in fact less variable throughout the cardiac cycle than our measurements represent. This uncertainty will have to await further study. However, an independent measure of the total myocardial blood volume utilizing blush in the myocardial region following aortic root injections of radiopaque contrast agent11have also found that the blood volume of the myocardium increases at end-systole.
Fig. 5 Total myocardial volume depicted at 8 selected time points from 16 time points spanning a single cardiac cycle. Each image was generated from a volumetric reconstruction representing a 0.03 second scan aperture. Numbers represent temporal location of image.
Regardless of these uncertainties, we have demonstrated that the total heart volume throughout the cardiac cycle remains essentially constant with a small yet consistent decrease in volume occurring between end-diastole and end-systole. As is seen in the accompanying video, as one watches the graphic depiction of the geometric changes which occur between the various components of the total heart volume, the epicardial borders are seen to be doing very little while large changes are occurring at the endocardial surfaces. The epicardial apex remains essentially fixed in space while the atrio-ventricular valve plane moves in a piston- like motion, moving towards the apex in systole and towards the base in diastole.
As is depicted in the accompanying video, in a series of studies utilizing 2-D echocardiography, we have demonstrated that in humans, as in the dog, the epicardial apex remains fixed in space and there is a piston-like motion of the atrio-ventricular valve plane. This relationship is consistent with a configuration conducive to the maintenance of a constant heart volume relationship. As also depicted in the accompanying video, we are in the process of completing a series of studies in normal human volunteers utilizing "dynamic" multisliced magnetic resonance imaging. These studies in four successfully-completed scans have demonstrated that again similar to the dog, the total heart volume at end-systole remains within 4% of its end-diastolic value.12
In a series of animal studies seeking to describe the important parameters serving to maintain a constant heart volume relationship in terms of the cardiac environment, we have found that the constant heart volume relationship does not require an intact pericardial sac.13 Furthermore, it does not require an intact negative intrathoracic pressure environment.13 This constant heart volume relationship is apparently dependent on the anatomic configuration of the heart itself. Recently, Beyar and colleagues14, modeling the heart based upon a time-varying elastance concept also predicts a constant heart volume relationship based upon the anatomic configuration of the heart alone. However, we have demonstrated that by stiffening the environment of the heart, that is, by inflating the lung with positive pressure, the small 2-5% reduction in total heart volume between end-diastole and end-systole is completely eliminated.6 On the other hand, by stiffening the atrial muscle, such as during atrial fibrillation,15 we have found that the reduction in total heart volume at end-systole is potentiated, bringing total heart volume reduction to nearly 9-10%. Thus, this may offer a significant increase in the work load, particularly of the right heart.
The significance of the constant heart volume relationship is multifold. Just to name a few: 1) If atrial compliance alters the normal descent of the atrioventricular valve plane,we may need to alter our definition of ventricular ejection filling mechanics; 2) if atrial wall strain is influenced by ventricular events, then ventricular function may modulate the release of cardiac hormones such as atrial natriuretic peptides; 3) alteration in atrioventricular excitation delay may change the constant heart volume relationship and then increase wasted myocardial work load; and 4) if left ventricular hypertrophy reduces atrioventricular valve plane motion, as has recently been suggested by R.R. Nimma Gadda, A.A. Bove and colleagues (personal communication), perhaps subsequent right ventricular hypertrophy could result from the disruption of the constant heart volume relationship.
The authors would like to acknowledge the strong support staff of the Biodynamics Research Unit at the Mayo Clinic for continuing to make possible the studies utilizing the Dynamic Spatial Reconstructor. This project was supported in part by NIH grants HL-29886, HL-04664, and RR-02540. Dr. Hoffman is currently an Established investigator of the American Heart Association.
2. Hamilton WF, Rompf JH. Movements of the base of the ventricle and the relative constancy of the cardiac volume. Am J Physiol 1932; 102: 559-65.
3. Stewart GN. Researches on the circulation time and on the influences which affect it. IV. The output of the heart. J Physiol 1897; 22: 150-83.
4. Hamilton WF, Moore JW, Kinsman JM, Spurling RG. Simultaneous determination of pulmonary and systemic circulation times in man and of figure related to cardiac output. Am J Physiol 1928; 84: 338-44.
5. Ritman El, Robb RA, Harris LD. Imaging physiological functions: Experience with the DSR. Philadelphia: Praeger, 1985.
6.Hoffman EA, Ritman EL: Invariant total heart volume in the intact throrax. Am J of Physiol: Heart and Circulatory Physiology 1985; 249: H883-90 .
7. Robb RA, Heffernan PB, Camp JJ Hanson DP. A workstation for interactive display and quantitative analysis of 3-D and 4-D biomedical images. Proc Tenth Ann Symp Computer Applications in Medical Care, IEEE Cat No. 86CH2341-6 1986; 240-56.
8. Hoffman EA, Ritman EL. Shape and dimensions of cardiac chambers via computed tomography: role of imaged slice thickness and orientation. Radiology 1985; 155: 739-44.
9. Hoffman EA, Ritman EL. Heart-lung interaction: effect on regional lung air content and total heart volume. Annals of Biomed Engin 1987; 15: 241-57.
10. Hoffman EA, Heffernan PB. Investigation of the intrathoracic determinants of cardiac geometry aided by an improved interactive approach to the manipulation of surfaces. Computer Graphics '86 (NCGA)1986; III: 151-61.
11. Iwasaki T, Ritman EL. Intramyocardial blood volume dynamics in the cardiac cycle (abstr). Fed Proc 1984; 43: 422.
12. Hoffman EA, Ehman RL, Sinak LJ, Felmlee JP, Chandrasekaran K, Julsrud PR, and Ritman EL: Law of constant heart volume in humans: A non-invasive assessment via x-ray CT, MRI and echo (abstr). J Am Coll Cardiol 1987; 9 : 38A.
13. Hoffman EA, Ritman EL. Law of constant heart volume: Comparative role of intrathoracic environment and pericardium in dog (abstr). Proceed Intern Union Physiol Sci 1986; 16: 412 .
14. Beyar R, Goldstein Y. Model studies of the effects of the thoracic pressure on the circulation. Annals Biomed Engin 1987; 15: 373-383.
15. Hoffman EA, Ritman EL. Law of constant heart volume disrupted by atrial fibrillation (abstr). Fed Proc 1986; 45: 776.