INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The lung develops considerably during childhood: the relative content of "true" elastin increases markedly during the first year of life and the number of alveoli increases approximately 10 times from birth up to the age of 8 yr, after which lung size increases further owing to alveolar growth (1). These changes with age should be reflected in the pressure-volume (P-V) relations of the lungs. However, lung mechanics has only been reported in infants and in children older than 5 yr of age, probably because standardized pulmonary function tests are difficult to perform in awake preschool children. To circumvent this problem, we measured lung and chest wall mechanics during anesthesia and paralysis. When measuring lung compliance (C_lung), esophageal pressure (Pes) is commonly taken to represent pleural pressure. However, the Pes-lung volume relation may vary between different body positions. Knowles and coworkers and Ferris and coworkers found that in supine awake adults, the Pes-V curve deviated, below 50% of the vital capacity, toward high pressures in contrast to the corresponding curves obtained in the lateral, sitting, or prone positions (4, 5). On the other hand, van de Woestijne and coworkers were unable to confirm those findings, but instead found a deviation at high volumes in the sitting position (6). Thus, there is a need to learn more about how the Pes-V curve varies between body positions. In recent years, positional change has been increasingly used as a means of improving lung function in mechanically ventilated children and adults with respiratory distress (7, 8). During such maneuvers, Pes may be measured for evaluation of C_lung, which emphasizes the need to know the normal shape of the Pes-V curve in different positionsinformation that is particularly sparse in children.

The aims of the present study were first, to measure C_lung and chest wall mechanics from infancy to puberty and second, to investigate and interpret possible differences in Pes-V relations and respiratory mechanics resulting from body position. We performed the measurements with the child in two positions: supine and right lateral.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Two girls and fifteen boys (Table 1) 0.2 to 15.5 yr of age, median 2.3 yr, scheduled for urogenital or lower abdominal surgery were studied. None had a history of lung disease and physical examination indicated normal lung and heart function. Informed consent was obtained from the parents and from the child if old enough. The study was approved by the local Human Studies Ethics Committee.

Anesthesia and Ventilation

Anesthesia was induced with intravenously administered thiopental (n = 15) or with halothane inhalation (n = 2). After induction, succinylcholine was administered intravenously and the trachea was intubated with a cuffed endotracheal tube. The cuff was inflated during measurements to prevent leakage. Anesthesia and paralysis were maintained with 1% halothane in air/oxygen (fraction of inspired oxygen [FI_O2] = 0.6) and a nondepolarizing muscle relaxant: vecuronium (0.1 mg/kg; n = 13) or atracurium (0.5 mg/kg; n = 4), which was given shortly before the measurements. Electrocardiogram (ECG), end-tidal CO₂, and pulse oximetry (Sp_O2) were continuously monitored. Between and after the measurements the child was connected to a Servoventilator 900C (Siemens-Elema, Solna, Sweden) set at volume control, a constant inspiratory flow, a rate of 20 to 30 min¹, and a tidal volume that gave an end-tidal CO₂ pressure of 4 to 5 kPa.

Measurements

Two sizes of latex balloon catheters, depending on the age of the child, were used for Pes registration; the balloon was 50 or 100 mm long with a diameter of 18 or 36 mm, respectively (when inflated to 5 cm H₂O of pressure). The balloon had good static and dynamic accuracy when tested in its working range (9). After endotracheal intubation and while the child was supine and breathing spontaneously, i.e., before the nondepolarizing muscle relaxant was given, the catheter was passed via the mouth and esophagus into the stomach. The balloon was insufflated with 2 (5) ml air and thereafter exsufflated to a pressure of 5 cm H₂O. A volume of 0.3 (0.8) ml of air was then introduced via a three-way tap to bring the balloon within its working range, and the catheter was connected to a pressure transducer. The pressure in the balloon was registered on an ink-jet recorder (EM-81, Siemens-Elema).

The position of the balloon in the stomach was confirmed by recording a positive pressure during inspiration. The catheter was then slowly withdrawn until the pressure became negative during inspiration and approximately 3 cm further to clear the balloon of the cardiac sphincter. Correct positioning was confirmed by airway occlusion at end-expiration during spontaneous breathing. If the ratio between the change in Pes and the change in airway pressure (Paw) was in the range 0.94 to 1.00, the position was considered satisfactory, otherwise the position or volume of the balloon was adjusted (9). In addition, as an independent confirmation, the final depth of the catheter tip was checked against the calculated distance from the mouth to the level of the diaphragmatic domes. A modification of Zapletal's formula was used: distance (cm) = length (cm)/5.5 + L, where L = 6 or 9 cm for children below or above 1 yr of age, respectively (10).

The nondepolarizing muscle relaxant was given and the measurements started approximately 15 min after induction of anesthesia. In one child, only supine measurements were made. With the others, measurements were first made in the supine position in five children and first in the right lateral position in 11 children. When supine, the child had the arms along the side of the body and the head supported by a small pillow. In the lateral position, the frontal plane of the child was perpendicular to the operating table. The legs were flexed 90° both in the knee and in the hip joints. The upper (left) arm was flexed 90° in the elbow and shoulder and was supported by a pillow. The head rested on a pillow. The setup has previously been described (11) and includes a computer, a printer, two pressure transducers (SCX01DN; SenSym, Rugby, UK), a heated Fleisch pneumotachograph (Gould 2, Lausanne, Switzerland) connected to a differential pressure transducer (MP 45-1-871, range ± 2 cm H₂O; Validyne, Northridge, CA), a flow interrupter, and a water manometer. The flow interrupter consisted of an electromagnetic valve (closing time 30 ms) placed over a soft rubber connector.

The measurement sequence was as follows. The lungs were insufflated to a pressure of 30 to 40 cm H₂O, which was maintained for 2 to 3 s before the computer was activated. This resulted in closure of the interrupter for 1 s, after which the lungs were slowly deflated, either manually with a supersyringe or by connecting the system via a resistance to a vacuum reservoir. During deflation, the flow interrupter cyclically opened for 0.16 s and closed for 0.16 s, each cycle thus lasting 0.32 s. Deflation continued until Paw had reached zero, when the computer stopped the measurement sequence. The time for deflation was 13 to 25 s, which equaled 42 to 80 occlusion cycles. In order to avoid noise caused by the interrupter itself, to allow it time to close (0.03 s), and to obtain an acceptable pressure plateau, only the pressure signal between 0.08 and 0.12 s after the start of closing the interrupter was processed. The mean Paw value during this period was taken to represent the alveolar pressure during the occlusion. Pes was sampled and averaged over the same period. To obtain the lung volume decrement during each interrupter opening, the flow signal was integrated over 0.32 s from midocclusion to midocclusion. Flow and pressure signals were A/D-converted every 10 ms and processed by the computer. The series of volume decrements and corresponding pressures were used by the computer to construct three P-V curves: Paw versus V, Pes versus V, and Paw minus esophageal (transpulmonary) pressure versus V. In addition, P and V values were recorded continuously by the ink-jet recorder and the data were stored on computer disks. All measurements were made in duplicate.

Analysis of the Data and Statistics

The elastic equilibrium volume (EEV) was defined as the lung volume at 0 cm H₂O of Paw. Because of time restrictionsthe studies were not allowed to take more than 30 min of anesthesia timeEEV was not actually measured. In the present report the term is only used to designate the lower end-point of the P-V curves. Inspiratory capacity (IC) was defined as the volume difference on the airway P-V curve between 30 and 0 cm H₂O. The expression "at IC" refers to conditions at Paw = 30 cm H₂O, i.e., at the top of the P-V curves. The maximal compliance of the respiratory system (Crs) is the slope of the straight line between the upper and lower curvilinear segments of the airway P-V curve (Figure 1). The slope was found by linear regression, after the operator had defined the end-points of the line. C_lung was found in the same volume interval (Figure 1). In practice, C_lung obtained this way was close to maximal C_lung, i.e., to the C_lung obtained by looking directly for the maximal slope of the transpulmonary pressure-volume curve. Compliance of the chest wall is not reported. This was because the Pes-V curve was sometimes close to vertical, i.e., chest wall compliance would have been close to infinity in some patients, which would have made the interpretation of means and regression equations problematic. Instead, its inverse, chest wall elastance (Ecw), was obtained. In order to relate Ecw to total respiratory system elastance (ERS), the latter was obtained as 1/Crs. For the analysis of the effect of growth, the static properties of the respiratory system were correlated to a power of body length (12). In all such regression equations, length was expressed in centimeters.

View larger version (12K): [in this window] [in a new window]   Figure 1.   P-V diagrams in the lateral position obtained in a 1.5-yr-old, 80-cm-long boy. Each dot represents an individual occluder cycle. The intersections of the two horizontal lines with the Paw-V curve indicate the points between which the Paw-V curve is nearly linear. The maximal Crs (the slope of the straight line in the figure) was obtained by linear regression on the values between these points. Clung was obtained between the same volumes, from the transpulmonary P-V curve (not shown). Ecw was also obtained in this volume interval, from measurements on the Pes-V curve.

Differences between the supine and the lateral position were analyzed for statistical significance by the two-tailed paired t test. Analysis of variance was used to assess whether regression coefficients were significantly different from zero. Data are presented as mean ± SD unless otherwise indicated. A factor of 1.09 was used to convert volume and flow from ATPS to BTPS condition.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Shape of the Pes-V Curve

Figure 2 presents mean curves from the whole group of children, in order to convey an overview of the results. Some consistent features could be observed. First, the Pes-V curve was much steeper than the Paw-V curve. Thus, at IC, i.e., at Paw 30 cm H₂O, Pes was only 11.3 ± 2.6 cm H₂O in the lateral position and very similar in the supine position: 11.4 ± 2.6 cm H₂O. Second, Pes at EEV was always distinctly positive in the supine position (6.6 ± 2.2 cm H₂O), whereas on average it was close to zero (1.1 ± 1.9 cm H₂O) in the lateral position. This implies that the Pes-V curve was even steeper in the supine than in the lateral position, as illustrated in Figure 3A, which depicts curves obtained in a 1.2-yr-old boy. Figure 3B, with curves from a 2.5-mo-old boy, shows the same general pattern, i.e., supine and lateral esophageal pressures were similar at IC, whereas Pes at EEV was more positive in the supine than in the lateral position. Pes in this child even increased as lung volume decreased toward EEV. This was seen also in two other children. In three others, Pes did not decrease perceptibly with decreasing lung volume in the lower part of the Pes-V curve.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Esophageal and airway pressures versus volume (in percent of IC). Mean curves from the entire cohort of patients are shown.

View larger version (14K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Figure 3.   P-V curves from two children. Duplicate curves in the supine and lateral positions are shown. (A) Recording from a 1.2-yr-old, 80-cm-tall boy. (B) Recording from a 2.5-mo-old, 59-cm-long boy.

C_lung

As Figure 4 shows, there was a strong correlation between C_lung (ml/cm H₂O) and length of the child (cm). Thus, C_lung in the lateral position = 0.0017 × length^2.26; r² = 0.90, p < 0.001. The values for supine C_lung were similar to those seen in the lateral position (Figure 4, Table 2).

View larger version (14K): [in this window] [in a new window]   Figure 4.   Clung versus length in this and four other studies. The studies by Zapletal (25), Baran and coworkers (26), and von der Hardt and coworkers (27) were done in sitting awake children; the one by Nisbet and coworkers (28) in anesthetized supine children. In all studies, Clung was taken as the steepest slope of the expiratory transpulmonary pressure-volume curve, except for the study by Nisbet and coworkers, in which the slope was found between FRC and FRC + 0.1 × IC. Curves of best fit, relating Clung (ml/cm H2O) to a function of length of the child (cm), were given by the respective investigators except for Baran and coworkers and Nisbet and coworkers where the curves were constructed by us from their data. The equations were as follows. Baran and coworkers: Clung = 0.0086 × length1.909, r2 = 0.44. von der Hardt and coworkers: Clung = 0.081 × length1.454, r2 = 0.23. Nisbet and coworkers: Clung = 0.0021 × length2.1319, r2 = 0.71. Zapletal and coworkers: Clung = 0.0043 × length2.082, r2 = 0.61. Present study: Clung in the lateral position = 0.0017 × length2.26, r2 = 0.90.

Ecw

The supine Ecw values were always less than those obtained in the lateral position (Table 2). Ecw expressed as a fraction of total respiratory system elastance (Ecw/Ers) was 33 ± 12% in the lateral position and 12 ± 16% supine (p < 0.001). In the lateral position, the ratio correlated positively but weakly with length (r² = 0.26, p = 0.04). There was no significant correlation in the supine position.

Crs and IC

Crs was 23 ± 10% less in the lateral than in the supine position (p < 0.001) (Table 2). In both positions, it correlated strongly with growth (Figure 5A). Thus, Crs (ml/cm H₂O) in the lateral position = 0.004 × length^1.97 (cm); r² = 0.95, and Crs in the supine position = 0.0033 × length^2.08; r² = 0.93. 

View larger version (11K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   Figure 5.   Relation of maximal Crs (A) and IC (B) to length. The regression equations are given in the text. Comparison with a previous study (11) is made.

IC was also less in the lateral than in the supine position (by 20 ± 9%; p = 0.001, see Figure 5B and Table 2). The equations for best-fit curves with length of the child were: IC in the lateral position (ml) = 0.0083 × length^2.48 (cm); r² = 0.96, and supine IC = 0.0105 × length^2.47; r² = 0.97.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It was found that in anesthetized and muscle-paralyzed infants and children: (1) Pes at the EEV was higher, and the Pes-V slope was steeper, in the supine position than in the lateral. At IC, Pes was the same in the two positions. (2) C_lung increased with growth and was similar in the supine and the lateral position. (3) IC and Crs increased with growth and were less in the lateral position than in the supine. (4) Ecw was approximately one-third of Ers in the lateral position, the ratio correlating weakly with length of the child. The ratio was less (approximately one-eighth) and was unaffected by body size in the supine position.

The cohort included only two girls (1 and 6 yr old) and 15 boys, reflecting the fact that we were studying children operated on in the caudal part of the body, including several boys undergoing hypospadias repair. It is possible that the results were influenced by the gender imbalance. However, in a previous study we could not find any difference in lung volumes and Crs between boys and girls when correction was made for height (11). Also, Lanteri and Sly, who studied 51 children between 3 wk and 15 yr of age, found no influence of sex on inspiratory Crs (13). These comments notwithstanding, there certainly exists a difference in lung volumes between (adult) men and women, even after correction for body size (14). However, we were mainly studying prepubertal subjects (Table 1).

The measurements were made during anesthesia and muscle relaxation. This entailed at least two potential problems. First, atelectasis forms within minutes of the induction of anesthesia (15, 16). This problem is fairly easily circumvented because the atelectasis can readily be reexpanded by a vital capacity maneuver (17). We therefore inflated the lungs to a pressure of 30 to 40 cm H₂O for 2 to 3 s before each measurement sequence. Second, Ecw during anesthesia and muscle paralysis will be different from that in the awake or lightly sedated state because of loss of intercostal and diaphragmatic muscle activity. Therefore, our results regarding chest wall compliance are only valid in muscle-paralyzed individuals.

We used an interrupter technique to achieve no-flow conditions during the pressure measurement. However, a truly static P-V curve cannot be obtained in vivo. When the interrupter is closed, the pressure equilibrates in the lungs and in the airways. In healthy individuals, the equilibration mainly reflects the viscoelastic properties of the lungs, i.e., the stress relaxation of the lung tissues and the alveolar lining (18). Jonson and coworkers found that the viscoelastic time constant in healthy adults was 0.82 ± 0.11 s (18). This implies that an occlusion time of more than 2 s is needed to obtain a "true" pressure plateau. On the other hand, if such long occlusion periods are used, the continuous gas exchange during the prolonged deflation will cause artifacts of the P-V curve (19). We used an occlusion time of 0.16 s, which may seem short. However, the time versus pressure curve was horizontal during the latter part of each occlusion, indicating that the occlusion time was probably adequate. Moreover, in a previous study of paralyzed healthy children, we used the same interrupter technique and in seven of these, we prolonged the occlusion time to 0.64 s (11). The pressure-time curves throughout the occlusions were nearly horizontal, also with the prolonged occlusions, and the Paw-V curves were very similar to those obtained with the shorter occlusion time.

Pes, measured by a balloon technique, has been considered to give accurate estimations of the pleural surface pressure in anesthetized adults and children (9, 20). The present study tallies with this finding with respect to the lateral position. Thus, Pes measured in this position decreased smoothly toward zero as lung volume decreased toward EEV. With the child supine, however, a markedly positive Pes at EEV was consistently seen (+7 ± 2 cm H₂O). In three children, part of the Pes-V curve even had a negative slope (an example is shown in Figure 3B). The findings in the supine position are thus difficult to reconcile with the hypothesis that Pes represented an overall pleural pressure. Three possibilities will be discussed, as follows.

Possibility I. The positive Pes at EEV was due to a mediastinal artifact. The lower esophagus is located behind the heart, and it is reasonable that the weight of the mediastinal content should affect the Pes in the supine position, particularly at low lung volumes. Indeed, Knowles and coworkers, and Ferris and coworkers, suggested that there were artifacts caused by the mediastinum on the Pes-V curve at volumes below 50% of vital capacity in the supine position, but not in the sitting, lateral, or prone positions (4, 5). On the other hand, van de Woestijne and coworkers, who studied healthy subjects in the sitting and supine positions, found that the mediastinum caused an artifact in the sitting position at high volumes but not in the supine position (6).

Possibility II. Pes reflected a local pleural pressure. In a study in dogs, Gillespie and coworkers simultaneously measured Pes and the pressure in the pleural space (21). They found that changes in static Pes-V curves with posture (supine, prone, and right lateral) were not caused by mediastinal artifacts. Instead, Pes corresponded to a local pleural pressure obtained in the vicinity, i.e., to pleural pressure at a dorsal site near the diaphragm.

The positive Pes at EEV in our study might be caused by the abdominal pressure, which has a considerable influence on the pleural pressure near the diaphragm (22). Agostino and Hyatt analyzed the effects of gravity on the pleural pressure distribution in the lateral position at FRC in awake humans and found that pleural surface pressure at the upper region of the diaphragm is markedly negative because of the caudal pull of the subatmospheric pressure of the upper, i.e., nondependent abdomen (23). Further down, the abdominal pressure becomes positive owing to the weight of the abdominal contents. This pushes the diaphragm cranially, causing the pressure in the pleura adjoining the diaphragm to be close to zero. If a similar analysis is done of the pressure distribution in the supine position, in anesthetized and paralyzed patients, the pleural surface pressure at EEV may be expected to be slightly negative in the ventral part but increase and become positive in the dorsal part, owing to the cranial pushing effect on the paralyzed diaphragm caused by the positive pressure in the dependent part of the abdomen. This is approximately 15 cm H₂O in a normal adult (22). Because the lower esophagus lies in the dependent region of the thorax in the supine position and midway between the two sides in the lateral position, the measured Pes at EEV should theoretically be positive in the supine position and approximately neutral in the lateral. The positive Pes at EEV may thus well represent a similarly positive pleural pressure in the dorsal caudal thorax and, hence, a locally negative transpulmonary pressure favoring airway closure and atelectasis (24). In fact, the most common location for atelectasis during anesthesia is in the caudal dependent region of the lung (16, 17).

Possibility III. P-V curves obtained in the supine position did convey meaningful information about global chest wall and lung mechanics. As stated, it is unlikely that Pes in the supine position directly represented an overall pleural pressure, at least not at all lung volumes. This would imply, for example, that transpulmonary pressure at EEV was negative in most parts of the lungs. In spite of this, it is conceivable that the obtained values for Ecw and C_lung in our study were, in fact, valid even in the supine position. Thus, they were measured at lung volumes well above EEV, and related changes in P and V to each othernot absolute values. The fact that C_lung was the same in the two positions gives some extra support for the thought that the values for supine Ecw and C_lung might be valid, and this will be assumed in the following discussion.

To our knowledge, this is the first study reporting in vivo static C_lung and Ecw in preschool children. C_lung in the infants and in the older children agrees with previous studies (25) (Figure 4). There was a growth-related increase in the ratio between Ecw and Ers in the lateral position. This might be due in part to an increase in elastance of the abdominal-diaphragmatic component of the chest wall. This component is greatly influenced by the effect of gravity on the abdomen (23). The vertical and lateral distances of the abdomen increase with growth and, because of gravity, this will increase the mean abdominal pressure and thus the contribution to Ecw from the abdomen-diaphragm. However, the increase in the Ecw/Ers ratio may also partly be explained by an increase in elastance of the rib cage caused by the progressive mineralization of the ribs with growth and the increase in the ratio of bone to cartilage (29). The findings are consistent with those by Reynolds and Etsten who found that elastance of the chest wall in relation to total is less in neonates than in adults (30).

Compliance of the total respiratory system was less in the lateral than in the supine position, which tallies with the higher elastance of the chest wall in the lateral position. This might, in turn, have been a consequence of gravity causing a change in the shape of the chest wall. Another line of reasoning also supports our observation of a lower Crs in the lateral position. From studies in adults, it is known that Crs in the middle volume range is higher in the supine than in the upright position (23). The exact mechanism has not been clarified. However, since FRC has been found to be higher in both the lateral and upright positions than in the supine (24, 31, 32), whereas total lung capacity is similar in all these positions (23), this will imply a reduced IC in the lateral and upright positions. Indeed, in the present study, inspiratory capacity was 20% less in the lateral position than in the supine. Because IC can be regarded as a measure of complianceit represents the volume change between 30 and 0 cm H₂Oit is reasonable that Crs should also be reduced.

In conclusion, we found that C_lung, Crs, and IC increased with growth in a regular fashion and that the two latter measures, but not C_lung, differed importantly between the supine and lateral positions. Likewise, there was a substantial difference in the shape of the Pes-V curve between the two body positions.