INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The anatomic makeup of the respiratory system changes dramatically during the first years of life. Developmental changes in the airways, chest wall, and parenchyma all occur, leading to rapid transformations of the respiratory systems mechanical properties. Previous studies have shown consistent increases in spirometric indices until early adulthood, followed by a slow decline in lung function with increasing age (1, 2). Standard spirometry is not suitable for children younger than 6 to 7 yr of age and hence cannot be used to describe mechanical changes in the respiratory system of infants or young children. The rapid thoracic compression (RTC) technique and more recently the raised volume rapid thoracic compression (RVRTC) technique have been developed to describe the changes in FEV, FVC, and forced flows in sedated infants with respect to age, after administration of bronchoactive agents and in disease (3). However, these techniques are unable to provide separate information on the changes in the airways and tissues due to growth. The interrupter technique has been used to partition the respiratory system resistance (Rrs) into airway (Raw) and viscoelastic tissue components and also to calculate quasistatic compliance (8). Lanteri and Sly (8) described changes in both airway and tissue mechanics with length; however, this study was conducted in intubated subjects and thus does not truly reflect the developing infant respiratory system.

A number of investigators have used the forced oscillation technique (FOT) to determine the mechanical properties of the airways in infants and young children (9, 10). These studies obtain values of the respiratory system impedance (Zrs) by superimposing the oscillatory signal (2 to 6 Hz) on the infant's tidal breathing. These investigators then derived regression equations for the relationship between airway mechanics and length. However, to adequately describe the mechanical properties of both the airways and the tissues simultaneously, the range of frequencies included in the oscillatory signal must extend to the spontaneous breathing rate (0.5 Hz) (11, 12). Sly and coworkers (13) recently described an adaptation of the RVRTC technique that invoked the Hering-Breuer reflex, inducing a pause in breathing in sedated infants and hence allowing an oscillatory signal encompassing the spontaneous breathing frequency (0.5 to 21 Hz) to be used. They demonstrated that this approach allowed reliable low frequency respiratory impedance spectra (Zrs) to be collected. In the present cross-sectional study the RVRTC (7) and the low-frequency forced oscillation techniques (LFOT) (13) were performed in a population of healthy infants with no history of respiratory disease. The aim of the study was to determine the relationships between forced expiratory volumes, airway and tissue mechanics and length in healthy infants in the first 2 yr of life.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

This was a cross-sectional study involving 37 infants (15 male, 22 female) of ages ranging from 7 wk to 2 yr, recruited from the general population. Information regarding parental smoking status (maternal, paternal, and both parents: n = 3, 5, and 5, respectively) and history of asthma in immediate family members (negative, positive, and unknown: n = 21, 14, and 2, respectively) were obtained. Twelve infants had their lung function measured on two separate occasions. Forced expiratory data were obtained on 28 infants, eight of whom had repeated lung function measurements (total observations = 36); LFOT data were collected on 28 infants, with six having repeated lung function (total observations = 34). In 23 infants, measures of both FEV_0.5 and Zrs were obtained. Infants were free of history of asthma, bronchiolitis, or bronchitis and had less than three occurrences of lower respiratory illnesses per year of life; infants with prematurity of birth were excluded from the study. Anthropometric data on the infants are outlined in Table 1. The infants were sedated with an oral dose of choral hydrate (70 to 100 mg/kg) and were laid in the supine position; the head was supported in the midline and the neck was slightly extended. Heart rate (HR) and oxygen saturation (Sa_O2) were monitored continuously throughout the protocol. The experimental protocol was approved by the Human Ethics Committee of the Princess Margaret Hospital for Children. Parents gave written informed consent and were generally present during the study.

Forced Oscillation Technique

Zrs spectra were measured using the LFOT during an apneic pause induced by the Hering-Breuer reflex as described in detail by Sly and coworkers (13). Briefly, a pseudorandom oscillatory signal containing 16 frequency components in the 0.5 to 20 Hz range was applied at the face mask by a loudspeaker-in-box system. Flow () was measured with a screen pneumotachograph (17 mm internal diameter [ID]) connected to an ICS 33NA002D differential pressure transducer (ICSensors Inc., Milpitas, CA). An identical pressure transducer was used to sense the oscillatory component (Prs) of the transrespiratory pressure (Ptr). The signals of Prs and were low-pass filtered at 25 Hz, sampled at 128/s with a 12-bit analog-digital converter, and stored on an IBM-compatible computer for later analysis.

Three deep inflations were accomplished by means of the RVRTC pump through the face mask to the infant. At the end of the third inflation, the airway was occluded at a Ptr of 20 cm H₂O to induce the Hering-Breuer reflex. In the resulting short apneic period, a low-frequency pseudorandom signal was driven into the infant's respiratory system by the loudspeaker. Measurements were 6 s in length and examined for leak or respiratory efforts, with corrupted recordings being excluded. Three to 10 impedance measurements were collected in each infant. After each series of measurements the impedance of the apparatus dead space (between the pneumotachograph and the airway opening) was determined. The impedance of the dead space was considered as a lumped shunt impedance in parallel with the respiratory system, and the corrected Zrs was calculated as described previously (13).

Analysis

Zrs data were computed from the cross-power spectra between the measured and Prs signals and the stored driving signal (11) by fast Fourier transformation using 2-s time windows and 95% overlapping. The individual Zrs data were then fitted to a linear model of the respiratory system (14) in the 0.5 to 15 Hz range by minimizing the root- mean-square difference between the measured and modeled impedance data. Frequencies coinciding with the measured HR or its harmonics were excluded from the model fit. The model consisted of an airway and tissue compartment. The airway compartment contained a frequency-independent resistance (Raw) and inertance (Iaw); the tissue compartment was characterized by coefficients of a constant-phase tissue damping (G) and elastance (H).

Tissue hysteresivity () was calculated as G/H (15). Scatters in the anthropometrical data and the respiratory mechanical parameters were expressed in SE values.

RVRTC Technique

Forced expiratory data were obtained over an extended volume range using the technique described by Hayden and coworkers (7). This technique uses a pump (Inflate-all; Coleman Co. Inc., Wicket, KS) to raise the infant's lung volume above the tidal range. The pump draws room air through a pneumotachograph (No. 3719; Hans Rudolf Inc, Kansas City, MO) and into the infant via a soft-rimmed face mask (King System, Noblesville, IN). and airway opening pressure (Pao) were measured, amplified (SC14C; RHT-INFODAT, Montreal, PQ, Canada) and recorded on a computer using BRATLAB data acquisition and analysis software (RHT-INFODAT). The infants were inflated three times to a Ptr of 20 cm H₂O with passive deflations between each inflation. Using a jacket connected to a positive pressure reservoir, a compression force was applied to the thorax and abdomen after the third inflation. The compression force was standardized to transmit a pressure of 20 cm H₂O to the airway at end-inflation resulting in an airway opening pressure of 40 cm H₂O (7). Forced expiratory flow was determined, integrated and volume-time curves were produced. FEV_0.5, FEV_0.75, and FEV₁ were calculated from the forced expiratory volume-time curves. Three to 10 technically correct measurements were obtained from each infant.

Statistics

Multilevel models were used to examine the relationships between lung function (natural logarithms of G, H, Raw, and FEV_0.5) and length (Lt). These models accounted for the nonindependence of the data owing to the repeated measurements for each child and repeated visits for some children. Initially, models with three levels (individual observations, within separate visits, and within children) were fitted, but variation at the second level (that is, covariation between visits for the same child) was not statistically significantly different from 0 after accounting for differences in Lt, so the model was collapsed into a two-level model. Lt (centered about the mean), Lt² and Lt³ (after centering) were potential covariates in all models. This allows a wide range of potential functional relationships between Lt and response, varying from a straight line (coefficients for Lt² and Lt³ both 0) to a cubic curve with two turning points. Although a wide range of other functional forms could have been considered, it is very unlikely in a data set of this size that a useful discrimination could have been made. Sex and family history of asthma were considered as covariates to examine their influence. Smoking status was not included owing to small numbers. Family history of asthma was treated as a binary variable.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figure 1 illustrates a representative impedance measurement and the corresponding model fit. The negative frequency dependence of both the resistance (Rrs) and reactance (Xrs) curves at low frequencies represents the frequency-dependent behavior of the respiratory tissues. The plateau in Rrs, observed at higher frequencies reflects the frequency-independent Raw. The zero crossing in Xrs represents the resonant frequency of the respiratory system, below and above which the elastic and inertive properties, respectively, dominate.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Individual respiratory impedance spectra (Zrs) and its corresponding model fit (solid line). The upper panel (Rrs) represents the resistive component of Zrs; the lower panel (Xrs) represents the reactive portion of Zrs. The inverse frequency dependence of the respiratory tissues can be seen at low frequencies (< 2 Hz) in both Rrs and Xrs. The plateau at higher frequencies in Rrs represents the frequency-independent Raw. The zero-crossing point in Xrs is the resonant frequency, with the inertive properties of the airways (Iaw) dominating frequencies above this point. Open symbols represent data points corrupted by cardiac noise and are excluded from the model fit.

LFOT

The relationship between natural logarithm (Ln) of the oscillatory parameters and Lt was best described by a quadratic equation; further variables (Lt³) were not significant and hence were not included in the model. (Table 2). Mean individual values for Raw, G, and H, the fitted regression lines and 95% confidence intervals (CI) are plotted in Figures 2, 3, and 4 for Raw, G, and H, respectively. A number of infants underwent repeated measurements (n = 6); the longitudinal data of these infants are also shown (Figures 2-4). It can be seen that the majority of infants show evidence of tracking with increasing length, although the numbers are too small to draw definite conclusions on this point. The mean coefficients of variation (CV) within subjects were similar for all oscillatory respiratory parameters (11.3 ± 0.8%, 22.4 ± 2.6%, 13.9 ± 1.2%, and 9.5 ± 0.7%, for Raw, Iaw, G, and H respectively: mean ± SEM).  was found to have no significant systematic variation with length (0.27 ± 0.01) and was not modeled. As Iaw is primarily influenced by the extrathoracic airway, it was not analyzed further.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Raw plotted against length. The solid line represents the fitted regression line; the long dashed line represents the 95% CI for the individual data. Mean measurements for each child are also shown (). The solid lines connecting individual data points represent those infants with repeated measurements showing tracking of airway mechanics.

View larger version (14K): [in this window] [in a new window]   Figure 3.   G plotted against length. The solid line represents the fitted regression line and the long dashed line the 95% CI for the individual data. Mean measurements for each child are also shown (closed circles). Individual data connected by solid lines are included, indicating those infants with repeated measurements showing tracking of tissue mechanics.

View larger version (14K): [in this window] [in a new window]   Figure 4.   H versus length. The solid line represents the fitted regression line and the long dashed line the 95% CI for the individual data. Mean measurements for each child are also shown (closed circles). Individual data connected by solid lines are included, indicating those infants with repeated measurements showing tracking of tissue elastance.

RVRTC

In contrast to the oscillatory parameters, the relationship between the Ln(FEV_0.5) and Lt was found to have a significant cubic length term (Table 2). Figure 5 shows mean individual FEV_0.5 values against length, the fitted regression line, and 95% CI. Infants with repeated measurements (n = 8) are also illustrated (dotted lines). The mean intrasubject CV was 7.6 ± 0.7%.

View larger version (13K): [in this window] [in a new window]   Figure 5.   FEV0.5 versus length. The solid line represents the fitted regression line and the long dashed line shows the 95% CI for the individual data. Mean measurements for each child are also shown (closed circles). In those infants for whom repeat data were obtained, mean measurements are connected by solid lines, allowing tracking of lung function to be illustrated.

Factors Influencing Lung Function Parameters

Sex was not a significant covariate within the present study. Approximately one-third of infants within the present study had a history of asthma within the immediate family. This was found to be a significant covariate for Raw, H, and FEV_0.5 (Table 3: Raw: p < 0.02, H: p < 0.01, and FEV_0.5: p < 0.005) and caused a worsening of lung function in these parameters.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrates the changes in the mechanical properties of the respiratory system with length over the first two years of life. The low-frequency forced oscillation and the RVRTC techniques were used to obtain values of airway and respiratory tissue mechanics and FEV_0.5, respectively. These parameters (Raw, Iaw, G, H, and FEV_0.5) were then included in multilevel models to examine the relationships between lung function and length; additional covariates (sex and family history of asthma) were also analyzed.

FOT

Previous data examining the changes with age in respiratory compliance (Crs) and Raw have been determined using passive flow-volume maneuvers, multiple occlusion technique, multiple linear regression during mechanical ventilation, and plethysmography (8, 16). Table 4 shows Crs, Rrs, and Raw values obtained from these investigations in comparison with our values of Crs and Rrs calculated from the 0.5 Hz impedance data, and the Raw determined by the model fitting.

In all cases, the values for Crs reported by previous investigators for a similar study population were higher than those ascertained in the present study. One possible explanation for these differences could be the higher average lung volume applied in our current measurements. Indeed Petak and coworkers (20) reported a significant decrease in H when the Ptr decreased from 20 cm H₂O to 10 cm H₂O. Further decrease in Ptr, however, caused significant elevations in H. Therefore, it is more likely that the differences in techniques used in the current study and those applied previously may explain these differences, because small-amplitude oscillations result in a significantly smaller Crs, particularly because of the amplitude dependence of the chest wall component (14, 21).

The 0.5 Hz Rrs data reported in the present study are comparable with those reported by previous investigators using different techniques (Table 4). The slight differences can be explained primarily by the higher lung volume applied in the present study, which causes a marked decrease in the airway component of Rrs (20). The values for Raw reported in the literature vary greatly depending on the technique used. Using the occlusion technique Lanteri and Sly (8) reported values of Raw at 25%, 50%, and 75% of expired volume in intubated, mechanically ventilated children. The Raw reported in the present study is higher than that of the Raw at 25% of expired volume reported in that study (8). The current values for Raw include the upper airway and nasal resistance whereas these components were excluded in their study because of the subtraction of the resistance of the endotracheal tube from their estimate of Raw. Plethysmographic, Raw data, reported in a consensus statement (19) are 1.5 to 2 times greater than those obtained in the present study. Besides the obvious effects of the elevated lung volume on Raw (20), the presence of a flow-dependent component in the plethysmographic Raw caused by the higher respiratory flows may also explain these differences. The mean coefficients of variation reported in the present study are comparable to those described in the American Thoracic Society/European Respiratory Society consensus statement (19).

RVRTC

The number of previous studies detailing forced expiratory parameter values in normal infants are limited. A small number of studies have reported predictive values for maximal flow at FRC (maxFRC) generated by the RTC technique (4, 16, 22). A previous study from our laboratory generated predictive values of FEV₁ for normal infants (6), using a prototype of the current equipment. The relationship between FEV_0.5 and length determined in the present study is comparable with that published by Turner and coworkers (6) (Table 4). The group mean intrasubject CV obtained for FEV_0.5 were also comparable with those described previously (6, 7) (Table 5).

Factors Influencing Lung Function Parameters

Sex was found to have no effect on lung mechanics in the present study, whereas family history of asthma was a significant covariate having a detrimental effect on Raw, H, and FEV_0.5. Recently Hanrahan and coworkers (16) reported mixed sex effects in a longitudinal study on 541 infants. Female infants were found to have significantly lower initial levels of Rrs than males, however, the decrease in Rrs with increasing length was slower in female infants. In contrast, although females had lower initial levels of Crs and a slower decline in Crs than males, these changes were not significant. In addition, Tepper and Reister (4) reported higher forced expiratory flows in females than males, although the difference was not significant. Overall, it is the smaller numbers in this cross sectional study that would account for the finding of a nonsignificant effect of sex.

Family history of asthma was found to have a significant detrimental effect on all lung function parameters, with the exception of tissue damping and Iaw. This is in marked contrast to the study of Tepper and Reister (4), who found no effect of family history on maxFRC or FRC. Young and coworkers (23) demonstrated an increased level of airway responsiveness to histamine in infants with a family history of asthma, although baseline lung function levels were not different between infants with or without a family history of asthma. Martinez and coworkers (24) reported an effect of maternal history of asthma of infants with late-onset wheezing (> 3 yr of age at onset) when compared with infants with no wheeze up to the age of 6. In addition, there was no significant trend for those children with late-onset wheeze to have lower maxFRC at age < 1 yr. The present study only examines lung function up to the age of 2 yr and hence cannot differentiate between those children who will be free of wheeze later in life and those who may develop late-onset wheeze. It is possible that the infants in the present study with a family history of asthma and decreased lung function may represent the group described by Martinez and coworkers (24) as late-onset wheezers.

Limitations of the Present Study

The present study has described the relationship between various lung function parameters and length in a cross-sectional population of healthy infants. Infants were defined as being healthy if they presented with less than three episodes of lower respiratory tract infection per year of life, and had no history of respiratory disease (asthma, bronchiolitis, and bronchitis). Within this population of so-called healthy infants, approximately one-third had a primary family member with a positive history of asthma, and hence could be presumed to be at risk of developing allergic disease in the future (24). Although the present study aims to characterize the relationship between lung function parameters and length, the results can not be considered as normative data or reference equations for future investigations. Longitudinal data exist on a small number of infants within this study for both the FOT (n = 6) and RVRTC (n = 8) techniques. However, these numbers are too small to draw definitive conclusions on changes in respiratory mechanics with lung growth. It can be seen, however, that these infants do exhibit tracking of lung function within the population as a whole, both with the low-frequency FOT (Figures 2-4) and the RVRTC (Figure 5). Larger studies with longitudinal measurements are required to accurately depict changes with lung growth. The present study has attempted to account for the differing lung volumes within the study by examining the relationship between lung function and length. The volume dependence of respiratory mechanics is well known and the measurement of lung volumes, most commonly FRC in this age group, to allow presentation of specific lung mechanics is needed to precisely describe changes in lung growth patterns. The determination of lung volume would also allow more direct comparisons between the low-frequency FOT, which is determined at raised lung volumes, and more conventional techniques.

In summary, the present study describes changes in oscillatory and forced expiratory mechanics with length in normal infants using the techniques of low-frequency forced oscillation and RVRTC in the first 2 yr of life. This information has been used to establish regression equations for the relationships between respiratory mechanics and length in sedated normal infants. Further studies, including longitudinal measurements of airway and respiratory tissue mechanics are needed to allow the development of normative data within this age group.