INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

BECAUSE OF THEIR CLOSE APPOSITION to the heart, the chest wall, the lungs, and the pericardium physically constrain the heart throughout life (7-11, 14, 19, 23). In doing so, these tissues limit ventricular preload and thus they limit ventricular output. The perinatal period may represent a time in life when cardiac function is particularly sensitive to ventricular constraint. In a previous study (8), we showed that a decrease in ventricular constraint might be a crucial adaptation at birth that allows cardiac output to increase to meet the metabolic demands of newborn life. Specifically, it is the reduction in the constraint applied to the heart by the lungs and chest wall with the beginning of air breathing that may be critical (8).

In the adult, ventricular constraint is amplified by decreasing chest wall and lung compliance, by increasing lung volume (17), and by the application of positive end-expiratory pressure (PEEP) (5, 14, 17). It is less certain how such changes in lung compliance, lung volume, and PEEP affect ventricular constraint in the newborn period. Nevertheless, clinical interventions, such as intermittent positive-pressure ventilation and PEEP, are widely used ventilatory strategies for babies with respiratory distress that may counteract the normal decrease in ventricular constraint that accompanies birth and thereby compromises cardiac function and vital organ perfusion.

Experimentally, ventricular constraint can be quantified by measuring pericardial pressure (P_PER) by using a liquid-containing balloon transducer positioned in between the pericardium and the heart (19). Clinically, it is seldom possible to directly measure pericardial pressure. However, a few intraoperative clinical studies (3, 24) support experimental studies (12, 20), which suggest that, at least in the adult, an estimate of P_PER can be obtained by measuring right atrial (RA) pressure (P_RA). We hypothesize that measurement of P_RA in the newborn is a useful estimate of ventricular constraint.

P_RA and thoracic inferior vena caval (IVC) pressure (P_IVC) are routinely measured in sick newborn infants and offer the potential to assess ventricular constraint. However, it is uncertain whether either of these pressures approximates pericardial pressure in the newborn. If P_RA or thoracic P_IVC reflect P_PER in the neonate, as they do in the adult, they may provide clinically useful tools for assessing ventricular constraint. Specifically, they would provide a means for assessing the effects that clinical interventions, such as volume therapy or mechanical ventilation, have on ventricular constraint. This may help predict their effect on cardiac function. In this study, we sought to determine whether either P_RA or thoracic P_IVC accurately reflects P_PER (and thus ventricular constraint) in the ventilated newborn (3-day-old) lamb. Because substantial cardiac remodeling occurs in the neonatal period, we also assessed whether P_RA remains an accurate measure of ventricular constraint throughout the neonatal period by assessing this relationship in neonatal (4-wk-old) lambs. In addition, because ventilation is frequently manipulated in neonatal intensive care units, we sought to determine whether the relationship between P_RA and P_PER is maintained during alterations in PEEP.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Instrumentation

Twelve 4-wk-old Merino × Border-Leicester lambs (25-27 days, 11.6 ± 0.6 kg; means ± SE) and six 3-day-old lambs (2-4 days, 6.4 ± 0.4 kg) were anesthetized (5 mg/kg of ketamine and 100 mg/kg of -chloralose for induction, followed by 25 mg · kg¹ · h¹ of -chloralose) and then intubated with a cuffed endotracheal tube. The lambs were placed supine and ventilated with a time-limited, pressure-controlled ventilator (model BP 200, Bourns; Riverside, CA). Ventilatory settings were adjusted to achieve normal blood gas and pH status. Body temperature (39.6°C ± 0.2) was monitored and maintained throughout the study with a warming blanket and a heating lamp.

The sternum was split and a 2-cm incision was made in the pericardium along the atrioventricular sulcus. To record P_RA, we positioned a 3.5-Fr saline-filled umbilical venous catheter (Argyle, Sherwood Medical; St. Louis, MO) in the atrium through the right appendage and secured it with a purse-string suture. P_PER was measured with the use of a calibrated (15), flat, liquid-containing Silastic balloon transducer (8, 19). This was positioned within the pericardial space overlying the free wall of the left ventricle (LV) in six of the 4-wk-old lambs and in all six of the 3-day-old lambs, and over the right ventricular (RV) free wall in the remaining six 4-wk-old lambs. The pericardial balloons were held in place with a single suture to the myocardium. The pericardial incision was loosely approximated with interrupted sutures. Care was taken not to artificially reduce pericardial volume. No effort was made to seal the pericardium because the liquid-containing balloon transducers are known to accurately record pericardial pressure under these conditions in fetal, newborn, neonatal, and adult animals (7, 8, 10, 11, 19). The chest was then closed in layers, it was made airtight, and the trapped air was removed with 2-4 cmH₂O of continuous negative pressure applied to bilateral chest drainage tubes.

A saline-filled catheter was introduced through the femoral artery and advanced into the ascending aorta to record arterial blood pressure. Blood samples were withdrawn from this catheter for blood gas and pH analysis (model ABL 500, Radiometer; Copenhagen, Denmark). We also positioned saline-filled catheters within the thoracic portion of the IVC (via the femoral vein) to record pressure (P_IVC) and within the jugular vein for blood withdrawal and infusion. Finally, LV pressure (P_LV) was measured with a transducer-tipped catheter (model SPC-460, Millar Instruments; Houston, TX) that was advanced into the LV via a carotid artery. The zero-pressure level of this transducer was set to equal the pressure measured in a separate saline-filled catheter positioned within the LV via the carotid artery. The zero reference for all of the pressures was set to the midplane of the heart.

We connected the saline-filled catheters and balloon transducers to calibrated strain-gauge manometers (Transpac IV, Abbott Critical Care Systems; Sligo, Ireland). A similar strain-gauge manometer was used to record tracheal pressure. The transducer-tipped catheter and the strain-gauge manometers were connected to an amplifier and signal conditioner (Cyberamp 380, Axon Instruments; Foster City, CA). All of the physiological signals were low pass filtered at 100 Hz and continuously recorded on a thermal chart recorder (model 7758A, Hewlett-Packard; Waltham, MA). The signals were simultaneously digitized on a computer with a sampling rate of 200 Hz, with the use of an analog-to-digital converting board (model 4801/16, ADAC; Woburn, MA) and data acquisition software (CVSOFT, Odessa Computer Systems; Calgary, Canada).

Protocol

Data Analysis

As shown in Fig. 1, P_LVED equals the sum of the pressure across the ventricular wall, transmural P_LVED, and any forces applied to the ventricular wall by the surrounding tissues (recorded as P_PER). P_PER equals the sum of the forces arising from the chest wall-lung combination (pleural pressure) and the forces arising from the pericardium (transpericardial pressure). Transmural P_LVED, an index of ventricular preload (11, 20, 21, 24), was calculated by subtracting P_PER from P_LVED. Transmural P_LVED was also calculated by subtracting P_RA from P_LVED. It was then compared with the transmural P_LVED that was calculated by using the balloon transducer.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   At end diastole (ED), left ventricular (LV) pressure (PLVED) equals the sum of the pressure across the ventricular wall, transmural PLVEDtm and any forces applied to the ventricular wall by the surrounding tissues (recorded as pericardial pressure, PPER). PPER equals the sum of the forces arising from the chest wall-lung combination (pleural pressure, PPL) and the forces arising from the pericardium (transpericardial pressure, PTP). Increasing ventricular filling pressure might not be accompanied by increases in ventricular preload (transmural PLVED) if PPER increases at the same time.

The relationship between P_PER and P_RA and its possible dependence on age, location of measurement of P_PER (LV vs. RV), and PEEP was investigated with the use of analysis of variance and analysis of covariance software (SPSS, version 6.1). This procedure involved comparing alternative regression models for fitting P_PER data. All of the models incorporated P_RA but the slopes and intercepts were allowed to vary according to age, location of measuring P_PER, animals, and PEEP.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Blood gas and pH values [pH 7.39 ± 0.01, arterial PO₂ 131 ± 3 mmHg, arterial O₂ saturation (Sa_O2) 98 ± 1%, arterial PCO₂ 39 ± 1 mmHg, hemoglobin 8.2 ± 0.6 g/dl, base excess 1.4 ± 0.6 mmol/l] indicated a stable physiological preparation and appropriate ventilation. LV filling pressure varied from of 0 ± 1 to 20 ± 1 mmHg by manipulating intravascular volume. In five of the lambs, P_LVED exceeded 20 mmHg, and in two of the lambs, P_LVED exceeded 25 mmHg. During each of these manipulations, P_PER, P_RA, and P_IVC closely tracked each other and, as such, a strong linear relationship existed between P_RA and P_PER, between P_IVC and P_PER, and between P_IVC and P_RA (Fig. 2).

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Linear relationships (r) existed between right atrial (RA) pressure (PRA) and PPER (left, PRA = 1.22 PPER  0.6, r = 0.99), between inferior vena caval (IVC) pressure (PIVC) and PPER (middle, PIVC = 1.21 PPER  1.1, r = 0.99) and between PIVC and PRA (right, PIVC = 1.00 PRA  0.6, r = 1.0) for an experiment in a 4-wk-old lamb.

The simplest regression model we considered was the one obtained by pooling the data from all lambs and plotting P_PER against P_RA. This explained 83.6% of the variation over the total of 3,320 (1 s) averaged data points. Accounting for age did not significantly improve the fit (F_2,30 = 0.9, NS; not significant) explaining 84.4% of the total variation. Nor was the fit significantly improved by accounting for the location of the balloon transducer (LV or RV, 84.6% of the variation; F_2,30 = 0.2, NS). Residual variation between animals accounted for 13.3% of the overall variation; thus the 18-regression line model (i.e., one line for each lamb) explained 97.9% of the total variation. Because each lamb was studied at two levels of PEEP, a 36-regression line model (i.e., separate regression lines for each animal and each level of PEEP) was also considered to assess the effect of increasing PEEP. The 36-regression line model accounted for an additional 1.32% of the total variation (P < 0.001). Finally, there was an average decrease in the slope of 0.06 and an average increase in the intercept of 1.0 between the 2.5 and 15 cmH₂O PEEP data. When incorporated into the 18-regression line model, this accounted for an additional 0.52% of the overall variation (P < 0.001), i.e., accounting for differences in slope and intercept associated with differences in PEEP accounted for ~40% of the improvement attained through the use of the 36-regression line model (39.4% = 100 × 0.52/1.32, F_2,34 = 11.1; P < 0.001), nevertheless, PEEP only accounted for 1.3% of the total variation observed.

Because a linear relationship existed between P_RA and P_PER in all 12 of the 4-wk-old lambs we studied, regardless of whether the balloon transducer was positioned over the LV (Fig. 3A), or the RV (see Fig. 3B), and regardless of age, all of the data were combined to determine the average relationships (P_RA = 1.19 P_PER + 0.0, r = 0.99, P_IVC = 1.24 P_PER + 0.1, r = 0.99 and P_IVC = 1.0 P_RA  0.2, r = 1.00).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Relationship between PRA and PPER from each of the 4-wk-old lambs (n = 12). There was a strong linear relationship between PRA and PPER in all lambs. Note that positioning the pericardial balloon transducer over the LV (A) or the right ventricle (RV) (B) did not significantly alter this relationship.

The relationship between P_RA and P_PER largely remained constant regardless of the level of PEEP (P_RA = 1.19 P_PER + 0.0, r = 0.99 at 2.5 cmH₂O, and P_RA = 1.29 P_PER  1.2, r = 0.98 at 15 cmH₂O). The same held true for the relationship between P_IVC and P_PER (P_IVC = 1.24 P_PER + 0.1, r = 0.99 at 2.5 cmH₂O, and P_IVC = 1.32 P_PER  1.2, r = 0.97 at 15 cmH₂O), and between P_IVC and P_RA (P_IVC = 1.0 P_RA  0.2, r = 1 at 2.5 cmH₂O, and P_IVC = 1.0 P_RA  0.2, r = 1 at 15 cmH₂O).

Elevations of P_LVED were accompanied by substantial increases in P_PER (P_PER = 0.71 P_LVED  2.1, r = 1.0) and in P_RA (P_RA = 0.66 P_LVED  1.2, r = 1.0). As a result, increases in transmural P_LVED were limited by the accompanying increases in ventricular constraint both when calculated as the difference between P_LVED and P_PER (transmural P_LVED = 0.29 P_LVED + 2.6, r = 1.0) or P_LVED and P_RA (transmural P_LVED = 0.34 P_LVED + 1.6, r = 1; Fig. 4). No significant difference existed between the transmural P_LVED calculated with the use of these two measures of ventricular constraint (P_PER or P_RA) over the range of P_LVED common to all lambs (3-12 mmHg, Student's paired t-test; P > 0.05).

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Transmural PLVED over the range of PLVED. Note that transmural PLVED is limited by ventricular constraint in 4-wk-old lambs, as it is not incremented in a one-to-one fashion as PLVED increases. Data are means ± SE from the 6 lambs, in which PPER was recorded over the LV, calculated as transmural PLVED = PLVED  PPER () and transmural PLVED = PLVED  PRA ().

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study in ventilated newborn lambs demonstrates that P_RA and thoracic P_IVC accurately reflect P_PER and thus ventricular constraint over a large range of LV filling pressures. Moreover, our results show that the relationship between P_RA and P_PER is maintained during development from newborn to neonate and during the application of PEEP. The strong linear relationship that we observed between P_RA and P_PER suggests that in the clinical setting of the neonatal intensive care unit, measurements of P_RA can be used to monitor how clinical interventions affect ventricular constraint and therefore ventricular preload.

The relationship we observed between P_RA and P_PER is in keeping with earlier studies conducted in adult animals (12, 20) and adult patients (3, 24). Our validation of P_RA as a measure of ventricular constraint in the newborn is of particular importance, given that in the adult, P_RA can exceed P_PER in the presence of ventricular hypertrophy (3, 24), and given that the fetal RV is relatively hypertrophied compared with that of the adult (18). This validation is also important given the substantial cardiac remodeling that occurs in the neonatal period. Our studies show that these changes do not influence the relationship between P_RA and P_PER in the newborn and neonatal period and that P_RA can be used to assess ventricular constraint early in life. The range of P_RA observed in our study span the range observed in adult human disease states. For example, our data exceed the range of P_RA and P_PER reported by Tyberg et al. (24) in adult patients undergoing coronary artery bypass and aortic valve replacement. Our data also span the range of pressures reported in patients with coronary artery disease, aortic stenosis, pulmonary hypertension, and cardiomegaly (3). Our observations indicated that the relation between P_RA and P_PER remained constant over a large range of P_LVED. Taken together, the results from adult studies and our current studies suggest that P_RA may provide a useful measure of ventricular constraint in neonates and newborns with abnormal cardiovascular function. Additional studies will be needed to confirm this suggestion.

Acute changes in pulmonary artery pressure and pulmonary vascular resistance often occur in the sick newborn and it is important to consider how such changes may affect the P_RA-to-P_PER relationship. On the basis of theoretical considerations, we predict that an acute increase in pulmonary artery pressure is unlikely to substantially alter the relationship between P_RA and P_PER. When pulmonary artery pressure increases, and blood flow from the RV is impaired, P_RA should also increase. As a result of an increase in RV volume and through the mechanism of ventricular interactions (22), P_PER should increase an equal amount provided right atrial compliance remains relatively constant. Although experimental evidence suggests that this is the case, because acute changes in pulmonary artery pressure associated with pulmonary embolism in the adult do not alter the relationship between P_RA and P_PER (1, 2), future studies should assess this prediction in the newborn.

Umbilical catheterization of the sick premature neonate in intensive care is a routine clinical practice that would permit a simple assessment of ventricular constraint. Although positioning these catheters into the RA can be difficult, the relationship observed between P_PER and P_IVC reveals that both P_RA and P_IVC provide an accurate measure of ventricular constraint because an equally strong relationship existed between P_IVC and P_PER as that between P_RA and P_PER. An ultrasound evaluation should be used to ensure that the catheter is truly placed within the RA or thoracic IVC because hepatic or umbilical venous pressure may not reflect P_RA.

The lungs have been described as the good hands that hold the heart, whereas at the same time it is recognized that they do much more than simply cradle the heart (4). Because of their close apposition to the heart, they act to limit diastolic filling. Although this constraint is amplified by mechanical ventilation, it is present in the fetus, newborn, and the adult even when PEEP equals atmospheric pressure (i.e., when intrapleural pressure is negative), and accounts for between 30 and 50% of the total constraint applied to the heart (7-9, 11, 14). This constraint reflects the fact that the lungs and the heart compete for space within the chest cavity and is a measure of the force per unit area with which the heart deforms the lungs or vice versa.

Although PEEP is known to alter ventricular constraint and limit cardiac function, in our study, as PEEP was increased, both P_RA and P_PER increased, and, as a result, the relationship we observed between P_RA and P_PER was largely maintained. Although accounting for the level of PEEP in our analysis significantly improved the fit of our regressions, the magnitude of this effect was small, accounting for 1.3% of the total variation. Thus the difference between the relationship of P_RA and P_PER at 2.5 and 15 cmH₂O PEEP was very small, i.e., in the 4-wk-old lambs, P_RA = 1.14 P_PER + 0.0 at 2.5 cmH₂O and P_RA = 1.19 P_PER  0.4 at 1.5 cmH₂O and, as such, relating P_RA to ventricular constraint remained valid during manipulations of airway pressure. This is of importance in the neonate, where ventilatory therapy can undergo many alterations to attain optimal gas exchange. Moreover, monitoring how P_RA changes in response to ventilatory manipulations may prove useful in optimizing mechanical ventilation, while minimizing ventricular constraint.

Transmural P_LVED was calculated on the basis of the balance of forces that exist in the LV at end diastole (Fig. 1) (8, 19, 21). Intravascular volume expansion increased both P_LVED and P_PER. As a result, the increase in transmural P_LVED that was observed was substantially limited by the increase in pericardial pressure (Fig. 4). The values of transmural P_LVED are in keeping with previously reported levels of transmural P_LVED in the newborn and neonatal lamb (7).

It has long been recognized that P_LVED is not a reliable measure of ventricular preload (6, 11, 13, 20). In settings where ventricular constraint increases, such as during the application of PEEP (5, 14, 17), or when chest wall and lung compliance decrease (17), ventricular end-diastolic pressure overestimates ventricular preload. The alterations in the pressure-volume relationship of the heart that accompany the use of PEEP may complicate the interpretation of ventricular function in the neonate (6). For example, if P_LVED were utilized to generate cardiac function curves before and after increasing PEEP and a downward shift in the cardiac function curve was observed, it would be interpreted as a decrease in contractility. However, if transmural P_LVED is used as the index of preload, the decrease in cardiac function is clearly explained by a decrease in preload acting via the Frank-Starling mechanism. Thus transmural P_LVED is a more reliable measure of preload than P_LVED. In clinical settings, P_RA is often used as an index of ventricular preload. Just as P_LVED is limited as a measure of ventricular preload, so too is P_RA. Elevated ventricular constraint may explain why cardiac function fails to improve in response to volume therapy in some sick neonates, although P_RA is increased. Appreciation of the magnitude of ventricular constraint in these situations would aid the clinician in selecting inotropic support rather than volume therapy. We have shown that in neonates transmural P_LVED can be accurately calculated as the difference between P_LVED and P_RA. This validation provides the clinician with the means to directly assess ventricular preload in settings where P_RA and P_LV are recorded.

In conclusion, our study shows that measuring P_RA or thoracic P_IVC provides a simple and accurate method of estimating pericardial pressure and assessing ventricular constraint in the newborn. Thus the information obtainable from variables routinely measured in the sick neonate has the potential to provide the clinician with a greater understanding of cardiac dysfunction in newborns undergoing intensive care. By measuring how P_RA changes with alterations in airway pressure, it may be possible to quantitate the effect of ventilatory therapies on cardiac constraint and help in optimizing ventilatory support, while minimizing cardiovascular dysfunction.