Anatomically and Physiologically Based Reference Level for Measurement of Intracardiac Pressures
Background Many reference levels have been proposed for the measurement of intracardiac pressures, but none have met with universal acceptance. In the first part of our study, we evaluated 10 cardiologists’ understanding of how hydrostatic pressure influences intracardiac pressures as measured with fluid-filled catheters. In the second part, we proposed and validated a new zero level (H): the uppermost blood level in the left ventricular (LV) chamber relative to the anterior chest wall for a patient in the supine position. A comparison was made of LV minimum diastolic pressure measured by reference to H versus measurements made with the zero level at midchest.
Methods and Results Using two-dimensional echocardiography, we determined H in the LVs of seven normal patients (five male, two female; age, 49±9 years) undergoing routine cardiac catheterization. H was determined from a left parasternal short-axis view and calculated as the average distance between end diastole and end systole of the endocardium of the uppermost segment of the LV anterior wall below the fourth or fifth intercostal space of the left sternal border on the anterior surface of the chest wall, with the patient in the supine position. A micromanometer/fluid-filled lumen catheter was then positioned in the LV, and we compared the micromanometer LV minimum pressure (LVPmin) obtained when the reference fluid-filled transducer was aligned at midchest with the LVPmin obtained when the reference fluid-filled transducer was aligned at H. LVPmin referenced to a midchest fluid-filled external transducer was measured as 5.1±1.6 mm Hg (range, 2.4 to 7.2 mm Hg) versus −0.6±0.6 mm Hg (range, −1.6 to 0.4 mm Hg) when referenced to H (P<.001). A significant linear relation was found to exist between patient anterior-posterior chest diameter and the magnitude of hydrostatic pressure influences related to pressure referenced at midchest (r=.88; P<.01).
Conclusions External fluid-filled transducers should be used with the goal of removing hydrostatic pressure and other influences so that the presence of subatmospheric pressure during diastole in any of the cardiac chambers is accurately measured. To achieve this goal, intracardiac pressure should be referenced to an external fluid-filled transducer aligned with the uppermost blood level in the chamber in which pressure is to be measured. The current practice of referencing the zero level of LV diastolic pressure to an external fluid-filled transducer positioned at the midchest level results in systematic overestimation due to hydrostatic effects and produces physiologically significant error in the measurement of diastolic intracardiac pressure.
It is widely accepted that ventricular equilibrium volume is a chamber property that defines the end-systolic volume below which the left ventricle stores elastic energy that subsequently contributes to the dynamics of diastolic rapid filling.1 2 3 4 5 6 7 8 9 10 Because the accurate measurement of left ventricular (LV) pressure is a crucial element in the measurement of this myocardial chamber property, it is critical that the methods by which LV diastolic pressure is presently measured be examined closely. We have previously demonstrated that the site at which pressure is measured within the LV significantly affects the measurement of LV diastolic pressure.11 These regional LV pressure differences have been shown to affect the calculation of LV equilibrium volume.12 The purpose of this report is to explore an additional consideration in the accurate measurement of LV diastolic pressure and to propose a new method for referencing LV diastolic pressure in the in vivo human heart. We believe a standardized approach to the measurement of LV diastolic pressure must be developed and universally implemented if meaningful interpretation and comparisons are to be made between diastolic pressures and pressure-derived diastolic indexes measured in different patients or between such measurements made in different laboratories.
In cardiovascular research, the standard for measurement of intracardiac pressure waveforms is the micromanometer catheter. Despite the ability of micromanometers to produce reasonably stable and accurate high-fidelity pressure waveform recordings free of the problems and artifacts associated with fluid-filled pressure measurement systems, micromanometer catheters are generally referenced to fluid-filled systems to correct for any drift in the micromanometer signal and to establish a common zero pressure reference level, ostensibly to remove errors due to hydrostatic pressure influences. The proper removal of hydrostatic pressure that allows for the meaningful and useful comparison of LV pressures, especially diastolic pressures, between patients is the subject of this report.
A great variety of reference levels for measurement of cardiac pressure have been proposed,13 14 none of which have met with universal acceptance. We believe that the persistent problem related to the choice of proper zero pressure reference level arises from two sources. First, there is a misunderstanding among many clinicians and researchers as to how hydrostatic pressure influences pressures measured with fluid-filled catheter systems. Second, to the best of our knowledge, no explicit logical position has been enunciated detailing the goal that the ideal zero pressure reference level should meet. Thus, in the first part of our study, we sought to assess the extent to which physicians trained in cardiology understand how hydrostatic pressure influences measurements made with fluid-filled catheter systems. In the second part of the study, we proposed an anatomically and physiologically appropriate reference level for the measurement of cardiac pressures, the objective being the accurate detection of subatmospheric pressure when it is present within any cardiac chamber. We then presented LV diastolic pressure measurements obtained by reference to our proposed new zero level and compared them with measurements obtained by reference to another commonly used level (midchest).
To assess the extent to which physicians trained in cardiology understand how hydrostatic pressure influences measurements made with fluid-filled catheter systems, we questioned 10 members of the clinical and research faculties of the cardiovascular division of the Washington University School of Medicine. Each individual was shown two diagrams portraying hypothetical pressure measurement situations and then asked a single question about each. The questions and diagrams are presented in Figs 1⇓ and 2⇓. Each individual was provided with a brief description of each figure. Fig 1⇓ was described as a fluid-filled catheter system in which the external transducer was positioned 10 cm below the surface of a cylinder of water. After the transducer was zeroed to atmosphere, a fluid-filled catheter was connected to the transducer, and its tip was submerged in the cylinder of water to the same depth below the surface as the external transducer (10 cm). The respondent was then asked to predict the pressure measured by such a system. After the individual’s pressure prediction was recorded, the individual was then shown a second configuration, detailed in Fig 2⇓. It was explained that the only alteration from the previous figure was that the tip of the fluid-filled catheter had now been plunged to the bottom of the cylinder of water, so that the tip was now located 40 cm below the surface of the water column, with the transducer remaining at a position 10 cm below the surface of the water column. Again, after this explanation, the individual was asked to predict the pressure that would be recorded by the system.
Determination of the Uppermost LV Blood Level
Using two-dimensional echocardiography, we determined the uppermost LV blood-endocardial interface (H) in the LVs of seven normal patients (five male, two female; age, 49±9 years) undergoing routine cardiac catheterization for the evaluation of chest pain. All patients had normal ventricular function and volume, without any sign of significant cardiac disease. H was calculated from a left parasternal short-axis view of the LV chamber during quiet respiration (Fig 3⇓). Because H varies slightly (≈1 cm) throughout the cardiac cycle, this measure was calculated as the average distance between end diastole and end systole of the endocardium of the uppermost segment of the LV anterior wall below the fourth or fifth intercostal space of the left sternal border on the anterior surface of the chest, with the patient in the supine position. With a yardstick equipped with a bubble-in-fluid level, each individual patient’s anterior-posterior chest diameter was measured as the distance equivalent to the perpendicular distance from the point on the anterior surface of the chest at which the echocardiographic view was obtained to the table.
Determination of LV Minimum Pressure Referenced at Midchest and H
A micromanometer-angiographic catheter (model SPC-474A, Millar Instruments) was then positioned in the LV of each patient by standard catheterization techniques. We then compared the LV minimum pressure (LVPmin) obtained when the reference fluid-filled transducer was aligned at midchest with the LVPmin obtained when the fluid-filled transducer was aligned at H. To do this, an LV pressure signal was obtained via the fluid-filled lumen of the micromanometric catheter with the external transducer positioned at midchest. The micromanometer pressure signal was then aligned during diastasis with this fluid-filled catheter signal (Fig 4⇓), and LVPmin was determined from the micromanometer signal during brief end-expiratory apnea. The external transducer was then repositioned at H, the micromanometer and fluid pressure signals were realigned, and LVPmin was determined again from the micromanometer signal.
A value of P<.05 was considered significant. Student’s t test was used for paired data.
Responses to diagrams. Of the 10 faculty members responding to the question regarding the hypothetical pressure measurement model presented in Fig 1⇑, 8 responded correctly, indicating that the fluid-filled catheter system would record a hydrostatic pressure of 10 cm H2O. However, in response to the configuration in Fig 2⇑, all 10 gave incorrect responses, 8 indicating that the configuration would produce a pressure measurement of 40 cm H2O and 2 indicating that the pressure measurement obtained would be 30 cm H2O. The correct answer to the question posed concerning Fig 2⇑ was, as in Fig 1⇑, 10 cm H2O. The only factor that determines the magnitude of the impact of hydrostatic pressure on pressure measurement with a fluid-filled catheter system is the relation of the height of the external transducer to the uppermost fluid level in the vessel or chamber in which the pressure is being measured.
Midchest and H pressure reference levels. The average anterior-posterior chest diameter, as measured from the table to the point on the anterior surface of the patient’s chest at which the echo transducer was positioned (fourth or fifth intercostal space at the left sternal border), was 24.9±4.0 cm (Table⇓). As shown in Fig 5⇓, this resulted in an average midchest value of 12.4±2.0 cm below the anterior surface of the chest. Measurement of H by echocardiography indicated that the uppermost LV blood-endocardial interface in the LVs of the seven patients was 5.3±0.9 cm below the anterior surface of the chest.
Determination of LVPmin by a micromanometer signal aligned to a fluid-filled transducer signal referenced at midchest or at H (Table⇑). LVPmin (diastolic) measured using midchest versus H as the zero reference level produced significantly different results. At the midchest reference level, LVPmin for the group was 5.1±1.6 mm Hg (range, 2.4 to 7.2 mm Hg). At the H transducer reference level, LVPmin for the group was −0.6±0.6 mm Hg (range, −1.6 to 0.4 mm Hg). This difference was significant (P<.001). With H as the reference level, five of seven patients (71%) displayed subatmospheric pressures during the early diastolic filling phase (Table⇑).
Hydrostatic pressure influences as a function of anterior-posterior chest diameter. Assuming LVPmin as measured at reference level H to be free of LV hydrostatic pressure influences, Fig 6⇓ represents the relation between pressure measurement error due to hydrostatic pressure influences and anterior-posterior chest diameter when LV pressure is referenced to an external fluid-filled transducer positioned at the midchest level (r=.88; P<.01).
Our results clearly indicate that there is a misunderstanding among some researchers and clinicians as to how hydrostatic pressure influences the measurement of pressure by use of fluid-filled catheter systems. The correct answer to both hypothetical pressure measurement configurations depicted in Figs 1⇑ and 2⇑ is 10 cm H2O. Thus, the only factor that contributes to measured hydrostatic pressure with a fluid-filled catheter system is the position of the transducer with respect to the uppermost fluid level in the container or chamber in which presssure is being measured. With this in mind, it is clear that the configuration presented in Fig 7⇓ would also result in a hydrostatic pressure measurement of 10 cm H2O.
Although such a result appears counterintuitive, examination of Fig 8⇓ clarifies the effect. Removal of the water cylinder from the configuration presented in Fig 2⇑ results in the catheter tip being positioned 30 cm below the transducer, producing a pressure reading of −30 cm H2O. Replacement of the cylinder would add a hydrostatic component of 40 cm H2O. Thus, the resulting pressure reading would be 10 cm H2O. In Fig 7⇑, by the same reasoning, a catheter positioned 9 cm above the transducer, contributing a pressure component of 9 cm H2O, plus 1 cm of hydrostatic pressure would also produce a fluid-filled transducer pressure reading of 10 cm H2O.
It follows that the proper level to position a fluid-filled transducer to remove all hydrostatic pressure influences would be even with the top of the fluid column in the chamber or vessel in which the pressure is being measured. In this way, pressure measured with a fluid-filled catheter at any level within the vessel would indicate a zero hydrostatic pressure component, as shown in Fig 9⇓. Thus, any additional pressure added by compression or removed by a mechanism generating suction would be measured accurately, essentially free of any hydrostatic pressure effects caused by the LV blood level. Removal of hydrostatic pressure influences, for example, would be a crucial first step for the accurate determination of certain properties of the LV such as equilibrium volume, which, by definition, is based on the precise measurement of the generation of subatmospheric pressure by the LV.15 This level would also provide a superior method for patient-to-patient comparison of LV filling pressures in the everyday measurement of LV pressures in the catheterization laboratory.
Thus, external transducer reference position is crucial to the assessment of LV filling pressures. To the best of our knowledge, no proper goal for referencing intracardiac pressure measurement to the atmosphere has ever been suggested. We therefore propose that external fluid-filled transducers should, optimally, be positioned with the goal of removing hydostatic pressure from the measurement so that subatmospheric pressure generation during diastole by any cardiac chamber is free of this confounding influence. The average error of up to 5.7 mm Hg for the measurement of LVPmin produced in our study by transducer alignment at midchest level is clearly unacceptable for meaningful comparison of diastolic pressures and diastolic pressure–derived indexes between patients.
In both clinical and animal research involving measurement of intracardiac hemodynamic pressure with micromanometers, two methods appear to be most widely used for referencing these pressure measurements to the atmosphere. A typical example of the first method can be found in an experiment conducted by Sabbah et al16 in which minimum LV diastolic pressure was measured in 15 patients with mitral stenosis. In that study, micromanometer pressure measurement was referenced to the atmosphere by holding the transducer in the air and adjusting its zero baseline to atmospheric pressure before placing the micromanometer in the LV. Using this approach, the investigators found an average minimum pressure in the LVs of these patients of −2±4 mm Hg. Subatmospheric pressures were demonstrated in 10 (67%) of 15 patients studied. From these results, the authors concluded that ventricular suction plays an important role in the dynamics of LV filling. Similarly, Paulus et al17 measured LV minimum pressure in 23 patients with mitral stenosis. However, that group referenced their micromanometer pressure measurements against lumen pressure referenced to an external fluid-filled transducer zeroed to atmosphere at the level of the midchest. The results of their study indicated that of 23 patients with mitral stenosis, only 3 (13%) were found to have LV subatmospheric pressures. The average LVPmin for their group of patients was found to be 2±3 mm Hg.
It is clear that in each of these studies, significant differences in hydrostatic pressure influences resulted from the techniques used for setting the zero reference pressure. This methodological difference probably contributed importantly to the differences in results reported by the two groups of investigators. Assuming that the results reported by Paulus et al17 included a systematic error of ≈5 mm Hg due to hydrostatic pressure influences, their corrected data would result in an average LVPmin of −3 mm Hg for the group, and correcting the individual data points by this same factor would imply generation of subatmospheric pressure in 19 of 23 patients (83%). Assuming a conservative component of 1 mm Hg due to submersion of the micromanometer under the blood in the LV cavity in the study by Sabbah et al,16 correction by this amount would result in a group average LVPmin of −3 mm Hg, with generation of subatmospheric pressure in 13 of 15 patients (87%).
Summary and Implications for Clinical Pressure Measurement and Research Studies of Diastolic Function
Appropriate external transducer reference position is pivotal in the assessment of LV filling pressures. In addition, our study indicates that a misunderstanding exists among some researchers and clinicians as to how hydrostatic pressure influences the measurement of pressure by use of fluid-filled catheter systems. We conclude that the practice of referencing diastolic pressure to external fluid-filled transducers positioned at midchest level14 18 merits reconsideration. The midchest level has no consistent relation to the position of the heart within the chest, and, as can be seen in Fig 6⇑, the magnitude of hydrostatic pressure influences is related directly to anterior-posterior chest thickness. For example, from the equation expressed in Fig 6⇑, the error incurred by measurement of pressure at midchest level due to hydrostatic pressure in a patient with an anterior-posterior chest diameter of 30 cm would be 7.5 mm Hg.
Rather, intracardiac pressures should be referenced, at minimum, to an external fluid-filled transducer aligned with the uppermost level of blood in the chamber in which pressure is to be measured. We emphasize “at minimum” because, in the case of the LV, additional hydrostatic and other pressure influences, such as those resulting from the presence of the right ventricle (RV) and pericardial fluid lying above the LV, may transmit additional hydrostatic pressure influences through the interventricular septum. Additional research will be needed to determine how and under what circumstances such potential sources of hydrostatic force can and should be removed from the measurement of LV diastolic pressure. Such research will have to take into account that this form of “ventricular interaction” is dependent to some extent on the passive elastic properties of the ventricular septum.19 RV hydrostatic influences can probably be estimated by the inferior vena caval occlusion technique to measure the direct ventricular interaction, as described by Slinker and Glantz.20 In addition, the influences exerted by pericardial forces and intrathoracic pressures need to be accounted for in the in vivo state if, as would be the ultimate goal for the calculation of indexes of passive myocardial chamber properties, an accurate estimation of LV transmural pressure is to be obtained.21 22
Thus, uncontrolled hydrostatic pressure influences can result in a misrepresentation of the nature of LV filling (ie, presence or absence of ventricular recoil), which would, for example, result in potentially large errors in the determination of LV equilibrium volume, since the pressure-volume relation during early LV filling is known to closely parallel the volume axis.23 Additionally, in open-chest animal preparations, external transducers should be aligned to fit the particulars of the experimental situation.
In the setting of a busy cardiac care unit (CCU) or cardiac catheterization laboratory, it it also important that proper external transducer position be determined readily and implemented easily. Proper external transducer position could be determined by one of three approaches: (1) Make this measurement a routine part of the normal diagnostic echocardiographic procedure. However, not all patients undergo this procedure and, in this cost-saving era, an available echocardiograph and technician are unlikely practical additions to any CCU or catheterization laboratory. (2) Use lateral fluoroscopy in conjunction with a catheter tip directed to contact the anterior wall to estimate H. Although a potentially accurate method of determining H, this approach would be time-consuming in the catheterization laboratory and impractical in the CCU. (3) Estimate LV H level based on a large database. We believe this to be the most practical solution. For example, in a large patient population, H could be carefully estimated by echocardiography for the LV, and possibly for the other three cardiac chambers, by parasternal short-axis and long-axis views. Based on these data, either a single standardized mean value and SD for the H levels of the LV and other chambers might be established for the population as a whole or individual regression equations describing the relation of H to variables such as sex and body surface area might be constructed. At present, based on our small sample of patients with LV H level estimated from a parasternal short-axis view, we recommend that in the routine clinical setting external transducers be positioned approximately 5 cm below the left sternal border at the fourth intercostal space. Since the H level of the left atrium (LA) tends to be ≈1 cm below the H level of the LV and the H level of the RV tends to be ≈1 cm above the H level of the LV, errors due to hydrostatic influences in measurement of LA and RV chamber pressures with the transducer positioned at the LV H level would only be on the order of +0.8 or −0.8 mm Hg, respectively. Since the H level of the RA is essentially equal to the LV H level, hydrostatic error influences during pressure measurement in this chamber with the external transducer positioned at the LV H level should be minimal. Thus, at present, we believe that the H level of the LV, with its narrow range of variation as estimated in our patient sample (SD, ±0.9 cm), is probably the most practical position and subject to the smallest overall errors due to hydrostatic influences for the routine measurement of pressure in any of the cardiac chambers.
In the conduct of research, we suggest that micromanometer catheters alone should not be used for direct recordings of LV diastolic pressure because of the influence of hydrostatic pressure. Unless the catheter can be manipulated within the LV so that the micromanometer is positioned at the most anterior “surface” of the blood within the LV and held there throughout diastolic filling, errors due to hydrostatic influences will result. Micromanometer catheters can be corrected for hydrostatic pressure influences only by simultaneous use of LV cineangiography in the lateral position or echocardiography to determine the precise depth of the transducer below the LV blood pool. In this way, dynamic hydrostatic pressure influences could be accounted for on a moment-to-moment basis throughout diastolic filling. Such a method, although onerous, may be the optimal way to measure LV diastolic pressure for the purposes of research. Alternatively, fluid-filled pressure measurement systems can provide an effective reference pressure to adequately remove most of the hydrostatic pressure influences due to blood within the LV chamber. In the present study, this technique resulted in substantially different measured LV diastolic pressures compared with midchest. Placement of an external fluid-filled transducer should be done with the goal of removing hydrostatic pressure so that subatmospheric pressure during diastole in any of the cardiac chambers can be measured accurately.
This study was supported in part by SCOR in Ischemic Heart Disease, grant HL-17646, National Institutes of Health, Bethesda, Md. We thank Mark E. Shelton, MD, and Joseph L. Kenzora, MD, for their technical assistance.
Presented in part at the Sixth International Conference of the Cardiovascular Systems Dynamics Society, San Francisco, Calif, November 19, 1994.
- Received January 18, 1995.
- Revision received March 21, 1995.
- Accepted April 1, 1995.
- Copyright © 1995 by American Heart Association
Shintani H, Glantz SA. The left ventricular diastolic pressure volume relation, relaxation, and filling. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:57-88.
Yellin EL, Nikolic SD. Diastolic suction and the dynamics of left ventricular filling. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:89-102.
Courtois M, Ludbrook PA. Intraventricular pressure transients during relaxation and filling. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:150-166.
Udelson JE, Bonow RO. Radionuclide angiographic evaluation of left ventricular diastolic function. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:167-191.
Zile MR, Nishimura RA, Gaasch WH. Hemodynamic loads and left ventricular diastolic function: factors affecting the indexes of isovolumetric and auxotonic relaxation. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:219-242.
Gaasch WH, Blaustein AS, Lewinter MM. Heart failure and clinical disorders of left ventricular diastolic function. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:245-258.
Carroll JD, Carroll EP. Exercise-induced angina pectoris. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:259-273.
Nakamura Y, Sasayama S. Diastolic dysfunction during pacing-induced angina pectoris. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:274-285.
Wigle ED. Diastolic dysfunction in hypertrophic cardiomyopathy. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:373-389.
Snider AR, Meliones JN, Minich LL. Doppler echocardiographic evaluation of diastolic dysfunction in children. In: Gaasch WH, Lewinter MM, eds. Left Ventricular Diastolic Dysfunction and Heart Failure. Philadelphia, Pa: Lea & Febiger; 1994:408-426.
Courtois M, Kovács SJ Jr, Ludbrook PA. The transmitral pressure-flow velocity relationship: the importance of regional pressure gradients in the left ventricle. Circulation. 1988;78:661-671.
Nikolic SD, Feneley MP, Pajaro OE, Rankin JS, Yellin EL. Origin of regional pressure gradients in the left ventricle during early diastole. Am J Physiol. 1995;268:H550-H557.
Pederson A, Husby J. Venous pressure measurement: choice of zero level. Acta Med Scand. 1951;141:186-194.
Yang SS, Bentivoglio LG, Maranhao V, Goldberg H. From Cardiac Catheterization Data to Hemodynamic Parameters. Philadelphia, Pa: FA Davis Co; 1988.
Nikolic SD, Yellin EL, Tamura K, Vetter H, Tamura T, Meisner JS, Frater RWM. Passive properties of canine left ventricle: diastolic stiffness and restoring forces. Circ Res. 1988;62:1210-1222.
Grossman W, Baim DS, eds. Cardiac Catheterization, Angiography, and Intervention. Philadelphia, Pa: Lea & Febiger; 1991.
Slinker BK, Chagas ACP, Glantz SA. The importance of direct ventricular interaction decreases in chronic pressure overload hypertrophy in the dog. Am J Physiol. 1991;253:H158-H168.
Slinker BK, Glantz SA. End-systolic and end-diastolic ventricular interaction. Am J Physiol. 1986;251:H1062-H1075.
Tyberg JV, Taichman GC, Smith ER, Douglas NWS, Smiseth OA, Keon WJ. The relationship between pericardial pressure and right atrial pressure: an interoperative study. Circulation. 1986;73:428-432.
Tyberg JV, Misbach GA, Parmley WW, Glantz SA. Effects of the pericardium on left ventricular performance. In: Baan J, Arntzenius AC, Yellin EL, eds. Cardiac Dynamics. Boston, Mass: Martinus Nijhoff; 1987.
Nikolic SD, Yellin EL, Tamura K, Tamura T, Frater RWM. Effect of early diastolic loading on myocardial relaxation in the intact canine left ventricle. Circ Res. 1990;66:1217-1226.