Background The chest wall, lungs, and pericardium limit diastolic filling of the left ventricle in the fetus, neonate, and adult. To determine the effect that these tissues have on the fetal right ventricle (RV), we studied six fetal lambs (142 days of gestation).
Methods and Results Pregnant ewes were anesthetized (ketamine and α-chloralose), and the fetuses were partially delivered by cesarean section. Fetuses were instrumented to record RV stroke volume, RV end-diastolic pressure (Prved), intrapericardial pressure (Pip), and pleural pressure. Prved was varied between 2 and 20 mm Hg under three conditions: initially with a closed chest and a closed pericardium (CCCP); subsequently with an open chest (chest wall and lungs retracted) and a closed pericardium (OCCP); and finally after the chest wall, lungs, and pericardium were retracted (OCOP). At equal Prved, stroke volume increased substantially when the chest wall and lungs were retracted from the heart and increased further on subsequent retraction of the pericardium (eg, at Prved of 9 mm Hg, stroke volume increased from 1.2±0.2 mL [mean±SEM] in the CCCP condition to 2.9±0.4 and 4.2±0.3 mL in the OCCP and OCOP conditions, respectively, P≤.05). The limitation of stroke volume in the CCCP and OCCP conditions occurred because Pip increased in an almost one-to-one fashion as Prved increased; as a consequence, RV preload (RV end-diastolic transmural pressure, Prved minus Pip) was relatively unchanged.
Conclusions The chest wall–lung combination and the pericardium each significantly constrain the fetal RV and together limit RV stroke volume.
Before birth, the ventricles of the heart function in parallel rather than in series as they do after birth.1 In this parallel arrangement, the RV dominates and pumps 1.5 to 2 times more blood than the LV.1 2 In addition, fetal combined ventricular output (normalized to body weight) exceeds that of the adult and increases even further immediately after birth.1 2 3 This increase in cardiac output at birth is surprising, given that both the fetal LV and RV appear to lack significant cardiac reserve. Although the Frank-Starling mechanism operates in the fetal heart,4 a lack of cardiac reserve is illustrated by the fact that neither LV nor RV output increases if ventricular filling pressure is increased beyond control levels.5 6 7 8 The structural immaturity of the cardiac myocytes,9 the reduced compliance of the fetal ventricles,10 and, particularly in the case of the RV, the sensitivity to increasing afterload6 7 11 12 have all been thought to contribute to this apparent lack of cardiac reserve. However, as we have demonstrated previously, the apparent lack of reserve in the LV of the fetus results largely from the constraint applied to the heart by the surrounding tissues.13 As ventricular filling pressure increases, the constraint applied to the heart by the chest wall–lung combination and the pericardium also increases. As a consequence, LV stroke volume is restricted by the imposed limitation on preload.
Although we have previously estimated that the chest wall–lung combination and the pericardium act in concert to limit fetal LV function,14 15 much as they do in the adult,16 it is not known which of these tissues specifically constrains the fetal heart. Nor is it known to what extent this constraint influences the fetal RV, the dominant ventricle of the fetus. Therefore, in this study we sought first to determine whether the tissues that surround the heart (chest wall, lungs, and pericardium) constrain the fetal RV and limit fetal RV function. Second, we sought to determine the relative contribution that the chest wall–lung combination and the pericardium make to the total constraint applied to the fetal heart.
Six pregnant ewes (Merino×Border Leicester cross, 142 days of gestation) were anesthetized (5 mg/kg ketamine and 100 mg/kg α-chloralose for induction, followed by 1 mg·kg−1·h−1 ketamine and 50 mg·kg−1·h−1 α-chloralose) and then intubated and ventilated while supine (2 to 5 cm H2O positive end-expiratory pressure, 100% O2). Each fetal lamb (4.8±0.3 kg) was partially delivered by cesarean section. We placed the head of the fetus into a saline-filled bag to prevent air breathing. The fetal trachea was then isolated, cannulated, and ligated to prevent the loss of lung liquid during the surgical preparation. The upper body of the fetus was delivered and positioned supine on the ewe's abdomen, taking care not to interrupt the umbilical circulation. The tracheal catheter was connected to a calibrated pressure transducer (0 to 50 mm Hg), and tracheal pressure was recorded.
The fetal sternum was then split, and the ribcage and lungs were retracted. A small (2- to 3-cm) incision was made in the pericardium along the atrioventricular sulcus, and an ultrasonic flow probe was positioned on the main pulmonary artery (probe model S, flowmeter model T108, Transonic Systems Inc). The cable of the flow probe exited the pericardium through a 1-cm incision overlying the pulmonary artery. Recently, the flow recorded from these probes, although linearly related to true flow, has been shown to consistently underestimate pulmonary artery flow in the sheep by ≈30% to 45%.17 We determined a calibration factor for these probes using thermodilution (Edwards Laboratories Cardiac Output Computer model 9520, Baxter Healthcare Corp). In two lambs, we verified that the flow probe underestimated the thermodilution flow by 33±1% and corrected our recorded flows accordingly.
To record pulmonary artery pressure, we positioned a saline-filled catheter in the pulmonary artery distal to the flow probe and secured it with a purse-string suture. Pip was recorded with a small (2×2-cm internal dimensions), flat, liquid-containing balloon positioned within the pericardial space overlying the RV. The pericardial incisions were closed with interrupted sutures, with care taken not to reduce pericardial volume. The edges of the pericardial incisions were not overlapped, and no attempt was made to seal the incision because the liquid-containing balloon accurately records Pip under these conditions in the fetus, neonate, and adult.13 14 15 18 A second balloon transducer was positioned on the pericardium, overlying the intrapericardial balloon, to record Ppl. Both balloon transducers were calibrated before insertion and subsequently at the end of each study by the methods of McMahon et al.19
RV pressure was measured with a transducer-tipped catheter positioned through the RV free wall (Millar Mikro-tip model TC-510, Millar Instruments, Inc). The zero level for this transducer was adjusted to equal that recorded from the central fluid-filled lumen of the catheter. The chest was then closed and made airtight, and 2 to 4 cm H2O of continuous suction was applied to evacuate trapped air.
We also placed a fluid-filled catheter in the carotid artery, from which blood samples were withdrawn for blood gas and pH analysis (Radiometer ABL 500 blood gas analyzer, Radiometer Inc). Finally, a catheter was positioned in the jugular vein for blood withdrawal and infusion.
We connected the vascular catheters and balloon transducers to calibrated strain-gauge manometers (Cobe CDX III, Cobe Laboratories). All pressures were referenced to the midplane of the RV. The strain-gauge manometers, the transducer-tipped catheter, and the flow probe were connected to a signal conditioner (Cyberamp 380, Axon Instruments Inc) and low-pass filtered at 100 Hz. All physiological signals were recorded on a thermal chart recorder (model 7758A, Hewlett Packard) and on computer at a sampling rate of 200 Hz with an analog-to-digital converting board (ADAC 4801/16 A/D board, ADAC Corp), acquisition software (CVSOFT Data Acquisition and Analysis Software, Odessa Computer Systems Ltd), and a personal computer (486 DX/50).
The experiments began 15 to 30 minutes after the surgery was complete and when normal arterial blood gas and pH status was attained. Tracheal pressure was then measured and compared with the value recorded before the chest was opened. By adjusting the volume of fluid in the fetal lung, we ensured that the tracheal pressure after instrumentation equaled that before the chest was opened. Each fetus was then treated with atropine (0.2 mg/kg, supplemented at 15-minute intervals) and propranolol (1.0 mg/kg) to block autonomic compensations of heart rate to alterations in fetal blood pressure.7
Data for RV function curves were generated by rapid withdrawal and subsequent reinfusion of fetal blood, and if necessary maternal blood, in sufficient quantities to vary fetal Prved over a range of 2 to 20 mm Hg. We studied each fetal lamb under three conditions in sequence. Initially, RV function was assessed when both the chest and pericardium were closed (CCCP), a condition in which the chest wall–lung combination and the pericardium could constrain the heart. Subsequently, the chest wall and lungs were retracted from around the heart (OCCP), and RV function was again assessed. In this condition, only the pericardium constrained the heart. Finally, while continuing to retract the chest wall and lungs, we widely incised the pericardium (OCOP) and reassessed RV function. We did not randomize the order of the study because it was not feasible to adequately reconstruct the pericardium after widely incising it transversely and longitudinally.
It is assumed that at end diastole, a static equilibrium exists across the RV free wall.18 20 At that moment, the pressure recorded within the ventricle, Prved, equals the sum of the pressure applied across the ventricular wall (Prvtm) and any additional forces applied to the ventricular wall by the surrounding tissues (Pip) (Fig 1⇓). Pip is the pressure recorded by the balloon transducer in the pericardial space and equals the sum of forces applied to the RV by the chest wall–lung combination (Ppl) and those applied purely by the pericardium (Ptp). When the chest wall–lung combination and the pericardium have been retracted from around the RV, surrounding pressure is atmospheric and Prved equals Prvtm. It is end-diastolic transmural pressure that most closely reflects ventricular volume and thus best reflects ventricular preload.18 20 In the CCCP or OCCP condition, Prvtm can be calculated by subtracting Pip from Prved.
RV function curves were constructed from a beat-by-beat analysis of the data recorded during each of the three conditions. For each lamb, and within each condition, these data were averaged in steps of 1 mm Hg. Prved, as an index of ventricular preload, was plotted against RV stroke volume to generate RV function curves. In addition, RV function curves were also generated by using Prvtm as the index of preload. An ANOVA for repeated measures was used to compare variables recorded over the range of end-diastolic pressures common to each lamb (5 to 15 mm Hg) and each of the three conditions of study. A Student-Newman-Keuls test was used to identify differences revealed by the ANOVA. A value of P≤.05 was assumed to be statistically significant for all comparisons. Data are presented as mean±SEM. All surgical and experimental procedures were approved by the Monash University Standing Committee on Ethics in Animal Experimentation.
Fig 2⇓ illustrates the measurements of Ppl, Pip, pulmonary artery flow, and Prved from one of the fetal lambs studied. At similar Prveds (13 mm Hg), pulmonary artery flow increased substantially when the chest wall–lung combination was retracted from the heart (OCCP) and increased further as the pericardium was subsequently retracted (OCOP).
Fig 3⇓ represents the data collected from all six lambs averaged and plotted as RV function curves. The form of the RV function curve generated in the CCCP state with Prved as the index of preload (Fig 3A⇓) was similar to those of previous studies.5 6 7 8 RV stroke volume increased as Prved was increased from a value of <5 mm Hg to ≈12 mm Hg. Beyond a Prved of ≈12 mm Hg, RV stroke volume plateaued and subsequent increases in Prved were not accompanied by further increases in RV stroke volume. After the chest wall–lung combination was retracted, the RV function curve was shifted upward and RV stroke volume was significantly greater at each Prved (Fig 3B⇓, Table 1⇓). The RV function curve was again shifted upward when the pericardium was retracted, with RV stroke volume being increased significantly over a range of Prveds of 5 to 12 mm Hg. Beyond a Prved of 12 mm Hg, RV stroke volume in the OCOP condition was not significantly different from the OCCP values (Fig 3C⇓, Table 1⇓).
RV function curves were also generated for both the CCCP and the OCCP conditions, with the calculated Prvtm used as the index of preload (Fig 3D and 3E⇑⇑, open symbols). In each case, RV stroke volume was linearly related to Prvtm, and a plateau was not observed in the RV function curves. The Prvtm–stroke volume relations closely approximated the data recorded in the OCOP condition (Fig 3F⇑), confirming the accuracy of our Pip measurements.
The development of Prvtm during volume loading was significantly limited by the chest wall–lung combination and the pericardium (Fig 4A⇓). During the CCCP condition, increases in Prved were accompanied by almost one-to-one increases in Pip (slope, 0.91±0.03). By contrast, Prvtm increased much less. As Prved increased to ≈9 mm Hg, Prvtm increased slightly (slope, 0.30±0.05). Prvtm increased much less beyond this Prved (slope, 0.09±.02). Even when Prved was increased to 20 mm Hg, Prvtm was limited to 2.6±0.6 mm Hg. In the CCCP condition, Pip was determined primarily by Ppl at low end-diastolic pressures and was determined equally by Ppl and Ptp at higher end-diastolic pressures (Fig 4B⇓).
Heart rate was not significantly different between the three conditions of study over the range of Prveds studied, averaging 152±8 bpm. Mean pulmonary artery pressure (an indication of RV afterload) was not different between the CCCP and the OCCP conditions. However, mean pulmonary artery pressure was elevated in the OCOP condition relative to the CCCP condition (P≤.01) at all Prveds. In addition, mean pulmonary artery pressure was significantly greater in the OCOP condition relative to the OCCP condition over the range of end-diastolic pressures of 5 to 9 mm Hg (P≤.05). To account for this increase in RV afterload, we calculated the product of RV stroke volume and mean pulmonary artery pressure as an index of RV stroke work (Fig 5⇓). RV stroke work index was significantly increased (P≤.05) at Prveds of 6 to 15 mm Hg in the OCCP condition relative to when the chest and the pericardium were both closed. RV stroke work index increased further (P≤.05) in the OCOP condition relative to the OCCP condition over a range of Prveds of 5 to 12 mm Hg, a change that was substantially greater in magnitude than the corresponding change in stroke volume.
Contractility was assessed by differentiating the RV pressure signal and determining the maximum dP/dt. dP/dt was greater in the OCCP condition than in the CCCP condition and greater still in the OCOP condition when compared at equal Prveds (Fig 6⇓). However, when dP/dt was plotted as a function of Prvtms (equal preloads), no significant difference was observed (eg, at a Prvtm of 2 mm Hg, dP/dt averaged 1010±104, 1096±85, and 1019±152 mm Hg/s in the CCCP, OCCP, and OCOP conditions, respectively).
Table 2⇓ lists the blood gas and pH data of the fetuses studied. These values are consistent with previously reported values and indicate a stable, healthy preparation.
Our studies reveal that the tissues that surround the heart (chest wall, lungs, and pericardium) apply considerable constraining forces to the fetal RV. In terms of pump function, the effect of this constraint was to limit stroke volume to as little as one third of its potential value. Retraction of the thoracic tissues increased RV stroke volume by as much as threefold from control values. At low Prveds, the majority of this constraint arose from the combination of the chest wall and lung. At higher end-diastolic pressures, this constraint was equally attributable to both the chest wall–lung combination and the pericardium.
Others who have assessed fetal RV and LV function using cardiac function curves have invariably used intracavitary pressures as the index of preload.5 6 7 8 Plotted in this way, the cardiac function curve exhibits a plateau phase that has been interpreted to reflect a limitation in ventricular function that is inherent to the myocardium. Yet, as we have shown in this study and in our previous study of the fetal LV,13 no plateau exists in the cardiac function curve when transmural pressure is used as the index of preload (Fig 3D and 3E⇑⇑, open symbols). As discussed below, the functional limitations of the fetal heart that were revealed in early studies5 6 7 8 exist not because of inherent myocardial properties but instead because of the constraining effects of the chest wall, lungs, and pericardium. It is only when constraint is removed from the heart (OCOP, closed circles, Fig 3F⇑) that the inherent properties of the myocardium determine the limits of ventricular function and lead to a plateau in the OCOP cardiac function curve. Notably, this plateau is evident only at transmural pressures that substantially exceed the physiological range.
As early as 1956, Katz21 suggested that “even the use of end-diastolic pressure as an index of end-diastolic volume is not justified.... Furthermore, if the expansion of the heart is limited by the pericardium, changes in end-diastolic pressure lose much of their meaning in terms of changes in end-diastolic volume.” Glantz and Parmley22 further speculated on the difficulties of using an intracavitary pressure to indicate preload. They predicted that when elevations in end-diastolic pressure are accompanied by equal elevations in Pip, ventricular preload may change only negligibly because ventricular transmural pressure remains relatively unchanged. Our data confirm that these cautionary views are particularly relevant to the study of fetal cardiac function. Accordingly, we have not equated Prved with ventricular preload but rather have used the calculated Prvtm as a measure of ventricular preload. In verifying the predictions of Glantz and Parmley,22 we have shown, initially in the LV13 14 15 and now in the RV, that as ventricular end-diastolic pressure increases, Pip also increases. This increase in Pip prevents end-diastolic transmural pressure from increasing, prevents ventricular preload from increasing, and explains the plateau of the fetal LV and RV function curves.
Our observations and those of Thornburg and Morton7 are in agreement in that both studies have demonstrated a functional limitation in fetal RV function. However, our interpretations of the nature of the limitation and its expression as a plateau differ. Thornburg and Morton7 8 concluded that the limitation and its expression as a plateau were unlikely to result from pericardial restraint. The measurements on which they based this conclusion depend critically on the accuracy with which pericardial pressure was measured. In earlier studies, Smiseth et al18 established that “pericardial pressure” is a radial contact stress that cannot be accurately measured with a fluid-filled catheter unless the pericardium is sealed and the volume of pericardial fluid increased. Although Thornburg and Morton showed that a good correlation exists between pericardial pressure measured with a balloon transducer and that pressure measured with an fluid-filled catheter in a carefully closed and sealed pericardium,23 in their study of RV function7 the pericardium was left open for a length of 3 cm, creating a condition in which the fluid-filled catheter would significantly underestimate pericardial constraint. Underestimating pericardial pressure would have two effects. First, it would obscure the constraining effect of the thoracic tissues. Second, it would create a plateau in the cardiac function curve because preload (RV intracavitary pressure minus pericardial pressure) would be overestimated. Conversely, accurate measurement of Pip with the balloon transducer18 reveals the important limitation on fetal cardiac function imposed by surrounding tissues. Moreover, with accurate determination of Prvtm, no plateau exists in the cardiac function curve (Fig 3D and 3E⇑⇑, open symbols).
Our studies were conducted in fetal lambs that had been anesthetized and partially delivered from the uterus. We were successful in performing these studies without compromising the umbilical circulation, as indicated by the normal arterial blood gas and pH values we report. Our recording of values of RV stroke volume that were less than in unanesthetized fetal preparations5 6 7 8 probably reflects the depressant action of anesthesia on the myocardium.
It is unlikely that our results are artifacts arising from our methods of instrumentation. We took care not to compromise pericardial volume during surgical preparation by minimizing the instrumentation within the pericardium (balloon volume <1.5 mL) and by not attempting to seal the pericardial incisions. We showed previously that the presence of a similar balloon transducer within the pericardial space does not significantly reduce fetal LV stroke volume.13 Moreover, in the present study, we observed that at low Prveds, the constraint applied to the RV arose primarily from Ppl, ie, the chest wall and lungs, whereas the pericardium remained relatively slack. Ppl continued to provide one half of the total constraint up to a Prved of 20 mm Hg (Fig 4B⇑). Finally, our finding that RV stroke volume doubles with the retraction of the chest wall–lung combination clearly shows that pericardial instrumentation was not the primary limitation on fetal RV stroke volume in the CCCP condition.
That the fetal heart is structurally immature9 is also clearly insufficient to explain the normal limitation on fetal cardiac function. By simply reducing the constraint applied to the fetal RV, we observed substantial increases in stroke volume. Nor can the lower compliance of the fetal heart compared with the adult10 explain the normal limitations on fetal cardiac function. Our data show that in the intact fetus, the inherent compliance of the myocardium has no functional implications, since the effective compliance of the ventricle is determined by the constraining influence of the chest wall–lung combination and the pericardium. It is only in experimental settings in which all of the thoracic tissues have been retracted from around the ventricle that the myocardial compliance itself determines ventricular preload and thus ventricular function (Fig 3F⇑, closed circles).
We anticipated that, as in the adult,16 ventricular constraint would depend on Ppl and thus lung volume. Accordingly, we ensured that the tracheal pressure at the end of instrumentation equaled that before the chest was opened. By this means, we aimed to ensure that lung volume and the constraint applied by the chest wall–lung combination were normal and similar to those present in utero. However, because constraints of maternal origin may further limit ventricular function of the fetus in utero, it is possible that by partially delivering the fetal lamb, we may have reduced the constraint applied to the fetal heart during our initial condition of study (CCCP).
Our results confirm that the fetal RV is sensitive to afterload (pulmonary artery pressure7 ). The increment in RV function between the OCCP and the OCOP was much greater when RV stroke work rather than RV stroke volume was considered. However, sensitivity to afterload itself does not account for the plateau in the fetal RV function curve. The plateau remained a prominent feature of the RV function curve when RV stroke work index was plotted against Prved, but it was absent when RV stroke work index was plotted against Prvtm (Fig 5⇑, open symbols). These results confirm that the tissues that surround the RV substantially limit RV function, an effect that is independent of changes in afterload.
Finally, the increments in ventricular function observed with retraction of the thoracic tissues were not accounted for by changes in ventricular contractility. At a given Prved, dP/dt increased after the chest wall–lung combination was retracted and increased further after the pericardium was retracted. We ascribe these changes to increases in ventricular preload associated with the relief of thoracic constraint and the preload-dependent nature of dP/dt.24 Just as with the ventricular function curves (Figs 3 and 5⇑⇑), the apparent upward shift in our dP/dt data when plotted as a function of end-diastolic pressure (Fig 6⇑) reflects the inappropriateness of the use of intracavitary pressures as preload rather than transmural pressure. When they were compared at similar preloads (equal Prvtms), no differences were observed.
Our study demonstrates that the constraints applied to the fetal heart by the chest wall–lung combination and the pericardium provide the predominant limitation to fetal cardiac function. In a previous study of LV dimensions,15 we demonstrated that this constraint decreases rapidly with the initiation of pulmonary ventilation and umbilical cord occlusion at birth. We speculated that this decrease in constraint results from the aeration of the liquid-filled lungs of the fetus and proposed that this reduction in constraint may play an important role in increasing LV function at birth by increasing LV preload. This speculation was supported by our observations that pleural constraint to the LV is eliminated with aeration of the liquid-filled lungs of the fetus at birth. Because constraint also limits fetal RV function, we now predict that a reduction in constraint at birth should also increase RV function.
The increase in RV function we observed after ventricular constraint was reduced exceeded the increase we observed in the fetal LV function under similar experimental conditions.13 At birth, RV output increases less than LV output, because the RV functions at a higher level in the fetus,1 2 and with the closure of shunts the outputs of the two ventricles become equal. It is possible that ventricular interactions play a role in determining the magnitude of the increase in cardiac output that occurs at birth. For example, others have speculated that the increase in pulmonary venous return that accompanies the onset of ventilation may, by increasing LV filling, limit the filling of the RV.25 26 Although further studies are required to determine the exact role of ventricular interactions at birth, we envisage that a decrease in thoracic constraint contributes not only to the overall increase in cardiac function but also to the relative changes in each of the ventricles.
In summary, the tissues that surround the fetal heart substantially limit Prvtm and, accordingly, limit RV preload. It is this constraint that accounts for the plateau in the fetal cardiac function curve, not a limitation that is inherent to the myocardium. The source of this constraint varies with the filling pressure of the ventricle, being determined primarily by Ppl at low end-diastolic pressures and by both Ppl and Ptp at higher filling pressures. Relief of this constraint at birth is likely to play a major role in the increase of fetal cardiac function that follows delivery and the onset of breathing.
This work was supported by the Alberta Heritage Foundation for Medical Research (Dr Grant), the Monash Research Foundation for Mothers and Babies (Dr Grant), and the National Health and Medical Research Council of Australia (Dr Walker). We acknowledge the technical assistance of J. Wild and V. Brodecky and the editorial assistance of Dr P.J. Berger.
Selected Abbreviations and Acronyms
|CCCP||=||closed chest, closed pericardium|
|OCCP||=||open chest (chest wall and lungs retracted), closed pericardium|
|OCOP||=||open chest, open pericardium (chest wall, lungs, and pericardium retracted)|
|Prved||=||RV end-diastolic pressure|
|Prvtm||=||RV end-diastolic transmural pressure|
|RV||=||right ventricle, right ventricular|
Reprint requests to Dr D.A. Grant, Neonatal Physiology Group, Institute of Reproduction and Development, Monash Medical Centre, 246 Clayton Rd, Clayton, Melbourne, Victoria 3168, Australia. E-mail firstname.lastname@example.org
- Received November 7, 1995.
- Revision received January 11, 1996.
- Accepted January 22, 1996.
- Copyright © 1996 by American Heart Association
Walker AM. Circulatory transitions at birth and the control of the neonatal circulation. In: Hanson MA, Spencer JAD, Rodeck CH, eds. Fetus and Neonate: Physiology and Clinical Applications, Vol 1: The Circulation. Cambridge, UK: Cambridge University Press; 1993:160-196.
Smolich JJ, Soust M, Berger PJ, Walker AM. Indirect relation between rises in oxygen consumption and left ventricular output at birth in lambs. Circ Res. 1992;71:443-450.
Kirkpatrick SE, Pitlick PT, Naliboff J, Friedman WF. Frank-Starling relationship as an important determinant of fetal cardiac output. Am J Physiol. 1976;231:495-500.
Gilbert RD. Control of fetal cardiac output during changes in blood volume. Am J Physiol. 1980;238:H80-H86.
Thornburg KL, Morton MJ. Filling and arterial pressures as determinants of RV stroke volume in the sheep fetus. Am J Physiol. 1983;244:H656-H663.
Thornburg KL, Morton MJ. Filling and arterial pressures as determinants of left ventricular stroke volume in fetal lambs. Am J Physiol. 1986;251:H961-H968.
Smolich JJ, Walker AM, Campbell GR, Adamson TM. Left and right ventricular myocardial morphometry in fetal, neonatal, and adult sheep. Am J Physiol. 1989;257:H1-H9.
Romero T, Covell J, Friedman WF. A comparison of pressure-volume relations of the fetal, newborn, and adult heart. Am J Physiol. 1972;222:1285-1290.
Hawkins J, Van Hare GF, Schmidt KG, Rudolph AM. Effects of increasing afterload on left ventricular output in fetal lambs. Circ Res. 1989;65:127-134.
Grant DA, Kondo CS, Maloney JE, Tyberg JV. Pulmonary and pericardial limitations to diastolic filling of the left ventricle of the lamb. Am J Physiol. 1994;266:H2327-H2333.
Grant DA, Kondo CS, Maloney JE, Walker AM, Tyberg JV. Changes in pericardial pressure during the perinatal period. Circulation. 1992;86:1615-1621.
Bednarik JA, May CN. Evaluation of a transit-time system for the chronic measurement of blood flow in conscious sheep. J Appl Physiol. 1995;78:524-530.
Smiseth OA, Frais MA, Kingma I, Smith ER, Tyberg JV. Assessment of pericardial constraint in dogs. Circulation. 1985;71:158-164.
McMahon SM, Permutt S, Proctor DF. A model to evaluate pleural surface pressure measuring devices. J Appl Physiol. 1969;27:886-891.
Tyberg JV. Ventricular interaction and the pericardium. In: Levine HJ, Gaasch WH, eds. The Ventricle: Basic and Clinical Aspects. Boston, Mass: Martinus Nijhoff Publishing; 1985:171-184.
Katz LN. Analysis of the several factors regulating the performance of the heart. Physiol Rev. 1955;35:91-106.
Glantz SA, Parmley WW. Factors which affect the diastolic pressure-volume curve. Circ Res. 1978;42:171-180.
Parmley WW, Chuck L, Yeatman L. Comparative evaluation of the specificity and sensitivity of isometric indices of contractility. Am J Physiol. 1975;228:506-510.
Thornburg KL, Morton MJ. Development of the cardiovascular system. In: Thorburn GD, Harding R, eds. Textbook of Fetal Physiology. Oxford, UK: Oxford University Press; 1994:95-130.