New aspects of airway mechanics in pre-term infants

Subjects

Pre-term infants from the Neonatal Unit at the Homerton University Hospital, London, UK, were eligible for recruitment if they were born at ≤36 completed weeks of gestation without major congenital abnormalities and required minimal ventilatory assistance (defined as continuous positive airway pressure (CPAP) and/or supplemental oxygen for <24 h after delivery). Gestational age was assessed from mothers’ date of last menstrual period and from obstetric ultrasound scans performed at or before 20 weeks of pregnancy. The pre-term infants were compared with a group of healthy full-term infants recruited antenatally at the Dept of Paediatrics, University Hospital of Berne, Switzerland. Infants were ineligible for recruitment if they had experienced any respiratory problems, including upper or lower respiratory illnesses prior to testing. The study was approved by the University of Berne, Berne, Switzerland and the East London and City Research Ethics Committees, London, UK. Informed written consent was obtained from the parents, who were usually present during the measurements.

Study design

All infants were studied unsedated in natural sleep, 30 min to 1 h after a feed. In both centres, infants were measured using an identical protocol and equipment 26, 29. Respiratory data were collected during consecutive periods of relatively quiet regular breathing in room air, with the infant settled in the supine position, while heart rate and oxygen saturation were monitored. Impedance measurements were performed prior to any other lung function measurements. A transparent Rendell–Baker face mask (size 0; Ambu International, Bath, UK) was held over the infant’s mouth and nose. A leak-free seal and reduction of dead space was created using therapeutic silicone putty (Carters, Bridgend, UK). The effective dead space volume of the face mask was ∼6 mL, as described previously 30, while that of the tube with the propeller valve was 7 mL. Between the short sets of measurements, the tube was detached to minimise dead space and CO₂ rebreathing. Measurements were repeated 10–25 times within each test occasion, provided the child remained undisturbed. Where possible the entire protocol was repeated on the same day after an interval of 3–8 h to assess repeatability of measurements.

High frequency impedance measurements Z(f)

The principles and technical details of the HIT have been described in detail previously 29. Identical custom-built equipment was used in both centres. Briefly, high-frequency respiratory input impedance was measured with a propeller valve that rapidly (within 1 ms) occluded the airway opening five times within a period of 0.15 s (duration of each period of closure and opening being 15.5 ms) during tidal breathing without disturbing the infant. The resulting pressure and flow oscillations were measured by the wave tube technique 31 using two piezo-electric transducers (EuroSensor; Model 33, London, UK). Spectral analysis was used to calculate respiratory input impedance from the pressure and flow signals 32. The f_ar,1, defined by a zero crossing in the imaginary part in the presence of a relative maximum in the real part of the impedance spectrum (Z_re(f_ar,1)) was extracted from the impedance spectrum, assessed between 32 and 512 Hz (fig. 1⇓). Data were not accepted for analysis if: 1) multiple peaks occurred; 2) the relative maximum of Z_re did not occur at the zero-crossing in the imaginary part; 3) the coherence was <0.9; or 4) oscillatory pressures changes were <0.15 kPa 26. After separate primary analysis at each centre (M. Henschen in London, UK; I. Brookes in Bern, Switzerland), all data were reviewed by the same person for acceptability (U. Frey, Bern, Switzerland). Within each test occasion, the first 10 technically acceptable manoeuvres were taken for further analysis.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

Example of one measurement of a high-frequency impedance spectrum from a single infant. The antiresonant frequency (f_ar,1) is defined by the relative maximum in the real part (Z_in,re(f_ar,1)) in the presence of a zero crossing in the imaginary part. Z_in: respiratory input impedance.

Statistics

Short-term repeatability of 10 technically acceptable measurements of f_ar,1 within each test occasion was expressed as the coefficient of variation (CV % = 100×sd/mean). Bland and Altman 33 analysis was used to assess within-subject between-occasion repeatability on the same day. Comparison of results between pre-term infants and healthy full-term infants were undertaken using the Wilcoxon two-sample test.

Interpretation of the findings and model hypothesis

Investigation of factors determining airway function in immature infants is essential to further improve understanding of the impact of pre-term delivery on the subsequent development of airway structure and function. From a structural point of view, airways are relatively large in relation to lung volume during early life, but the maximal flows that can be conveyed through such airways may be somewhat less than anticipated, since such flows are a function of the ν of travelling pressures waves; hence, they are related not only to diameter, but also to airway wall compliance 19. One hypothesis is that the structural immaturity of the airway walls results in highly compliant airways, crucially determining flow limitation in immature airways of prematurely born infants. The high compliance of the airway walls in immature animals has been demonstrated in vitro 22, 34, but the situation might be different in vivo and/or in human infants, depending on the strength of airway–parenchyma attachments and the relationship between lung and chest wall compliance 24. This relationship will be further complicated by the tendency of young infants to dynamically elevate their end-expiratory volume, thereby further regulating airway patency and elastic recoil, and by changes in sleep state, which may affect both lung volume and upper airway tone 16, 23. A true picture of the role of these complex interacting structural and functional systems can therefore only be estimated in vivo. Since increased transmural pressures may affect airway compliance 20, 35, infants who had received mechanical ventilation were excluded from this study. Only one infant received CPAP therapy (20 h; f_ar,1 = 157 Hz), while two had brief supplemental oxygen (1 day, f_ar,1 = 195 Hz and 11 h, f_ar,1 = 177 Hz). Visual inspection of results indicated no bias with respect to data from these three infants, and identical conclusions were reached, whether or not they were excluded from the analysis.

The interpretation of these results in terms of their physiological meaning must be very cautious and can only be undertaken by reference to simplified models. At higher frequencies, pressure waves follow the physical laws of acoustics. In a large-diameter, simple, rigid straight tube, ν corresponds to the first harmonic acoustic antiresonant frequency, comparable to the sound pitch of a flute. The frequency of this resonating sound is dependent on wave speed, which in turn is mainly dependent on gas density and the length of the tube. In a branching network of compliant small tubes (such as the airways), the frequency at which this antiresonance occurs is still dependent on ν but, in such a system, ν does not correspond to the free field sound wave speed, i.e. the wave speed in a simple, rigid straight tube. Under such circumstances, ν is no longer dependent simply on length and gas density, but also on the airway wall mechanics (compliance) and, to a minor extent, by airway diameter in very small tubes. In the terminal airways of human adults where the diameter is <1 mm, ν is 62% of the free-field speed of sound 28. Thus, in very peripheral airways, ν is significantly reduced. A reduction in ν in these distal airways would cause them to resonate at a lower frequency. More important, however, is the influence of airway wall compliance. Airway wall compliance strongly influences ν and, thus, f_ar,1 in compliant airways. The relationship between airway path length, diameter, wall compliance, ν and f_ar,1 is highly complex and the influence of their components cannot easily be distinguished. Nevertheless, speculation is possible, based on the theory of simplified elastic tube models and animal models 36, 37.

A significantly reduced f_ar,1 was found among the pre-term group, who were not only more immature, but also younger and smaller than the full-term infants. Since mean airway path length must be shorter the smaller the infant, one might have expected f_ar,1 to be higher in this group. Indeed, assuming a certain proportionality between crown–heel length and mean airway path length 10, the mean airway path length would be expected to be ∼30% lower and, thus, f_ar,1 to be ∼30% higher in the pre-term group. Theoretically, airway diameter also has a certain influence on wave propagation in small tubes and, therefore, f_ar,1. However, based on published animal model estimations 36, 37, and assuming that airway resistance was ∼30% higher in pre-term than in full-term infants 17, a corresponding diameter scaling factor would only decrease f_ar,1 by <10% 36. Based purely on length and diameter, f_ar,1 would therefore be expected to be higher in younger, smaller infants. This suggests that increased airway wall compliance due to immaturity must have a significant influence on f_ar,1 and hence ν in pre-term infants. The latter would be consistent with structural findings 22, 34.

Even though the impact of the different components on f_ar,1 cannot be separated quantitatively, it can still be concluded that differences in f_ar,1 probably reflect differences in ν in the airways of pre-term compared with full-term infants. This has crucial implications. Maximal flows through a compliant tube are related to ν in the tube. ν is the speed at which a small disturbance travels in a compliant tube filled with gas. The maximal flow in a compliant tube (V'_max) is the product of velocity and tube area 19. Thus, f_ar,1, velocity and V'_max are related. These considerations can, however, only be qualitative and not quantitative, since the relationship between f_ar,1, airway path length, airway wall compliance and the frequency of the travelling pressure waves is highly nonlinear 26, 30, 36–40. Nevertheless, based on these findings, the present authors hypothesise that the ability of airways to carry large flows is very different in pre-term than in full-term infants.

Repeatability of f_ar,1

Despite the consistency of measurements within any one testing session, the within-subject variability of f_ar,1 when measurements were repeated after several hours limits the potential clinical usefulness of this technique. The reasons for such variability are manifold, but include the fact that airway wall compliance is part of a complex regulatory system maintaining balanced flow through the airways. Other parts of this regulatory system may be lung volume, elastic recoil and airway diameter. The degree of between-occasion variability observed in the pre-term infants in this study is similar to that published previously in 10 healthy unsedated full-term infants on two different days within the same week 30. There are very limited data describing between-occasion repeatability for lung function tests in infants with which to compare the current results. Due to the difficulties in undertaking such studies, most are based on very small numbers of subjects and all have used different approaches to reporting “repeatability” 41–43, which complicates comparisons. Nevertheless, most appear to reflect a similar degree of between-occasion variability for forced expiratory flows in infancy, as was found for high frequency input impedance in the current study.

Limitations to the study and concomitant factors

The main limitation of the current study was the fact that measurements were made through a face mask, and the smaller the child, the larger the contribution of the shunt compliance of the face mask 30. The present authors tried to overcome or at least partially compensate for this potential error by using the same face mask and filling it with silicon putty to reduce the dead space, but this could have contributed to the lower values of f_ar,1 among the pre-term infants since the “effective” dead space would have been relatively large in relation to body size in this group. Similarly, input impedance measurements include the upper airways, and their influence cannot be distinguished from the lower airways 25, 29, 40. On the other hand, study design was strengthened by the fact that measurements were undertaken during spontaneous unsedated sleep, thereby reflecting the complexities of the real dynamic situation.

The lower post-conceptional age of the pre-term infants at time of study was due both to shortened gestation and the fact that, for practical reasons, they were measured at an earlier post-natal age than their full-term counterparts. Differences in post-natal age could, therefore, have contributed to developmental differences of f_ar,1 between full- and pre-term infants. While a much larger, preferably longitudinally studied population would be required to investigate the separate effects of gestational versus post-natal immaturity, it should be noted that since the pre-term infants were studied before the expected date of delivery, the effects of pre-term birth per se are likely to have made an important contribution. A further potential limiting factor was the difference in ethnic background between the London and Bern groups, which partly reflected local population characteristics, but was greater than had been anticipated when planning the current study. In retrospect, given the characteristics of the Swiss population, data collection in London should have been limited to pre-term infants born to White mothers to avoid any confounding. In reality, given the available resources and the prolonged period of recruitment this would have entailed, this was not feasible. It has been shown that forced expiratory flows are higher in black than white pre-term infants during the first weeks of life 44, 45. While such differences could reflect differences in intrathoracic airway calibre, they are more likely to reflect transient differences in breathing pattern during the neonatal period among black babies, in whom a lower nasal resistance 17, 46 and increased expiratory braking during tidal breathing 44, 45 has also been noted. While statistical analysis of the effect of ethnicity in this study was precluded both by the sample size and by the heterogeneity of the pre-term group (with almost half the infants being of mixed ethnic origin (table 1⇑)), visual inspection of the data revealed complete overlap according to ethnicity.

In the current study, there was also an unexpected preponderance of males, but as the proportion was similar between the two groups, this should not have introduced any bias (fig. 2⇓). Numerous studies have shown that after correction for body size, the sex of an infant has a marked effect on airway function, though not on lung volumes, with lower maximal expiratory flows at low lung volumes being observed in males compared with females at any given height during infancy 45, 47–49 and later childhood. To date, nothing is known about the influence of factors such as ethnic background, sex or body size on high frequency Z(f) measurements, and a very large population study (∼200 infants) would probably be required to determine such effects 50.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

Plot of f_ar,1 against crown–heel length comparing female (○) and male (▴) infants.

A further concomitant factor may be tobacco exposure. Elliot et al. 51 suggested that airway wall mechanics and airway tissue coupling is altered in tobacco-exposed newborn animals. Therefore, it is likely that airway wall mechanics and, thus, f_ar,1 could be altered in tobacco-exposed infants. The current study was not sufficiently powerful to undertake subgroup analysis, and specific details regarding prior environmental tobacco smoke exposure were not recorded, although all the pre-term infants were studied in hospital and thus prior to any post-natal exposures.

Conclusions and clinical implications of findings

The HIT is applicable in unsedated pre-term infants. In common with other infant lung function tests, the f_ar,1 has a good short-term variability within a single test occasion, but shows considerable within-subject variation when tests are repeated some hours later. Despite the presence of shorter airways, the f_ar,1 of the high frequency impedance spectrum was significantly lower in pre-term than in full-term infants, suggesting that differences in f_ar,1 probably reflect differences in ν in the airways of pre-term infants when compared with full-term infants. Due to the complex nature of wave physics in compliant tubes and the limits of the study, it was not possible to quantitatively separate the impact of the different components on f_ar,1. Based on qualitative considerations, however, the present findings suggest that developmental differences in airway wall mechanics and airway–parenchyma coupling influencing airway wall compliance may play a critical role in determining f_ar,1. The latter would be consistent with morphological data in published animal models, showing higher compliance in immature airways 37.

This has important physiological, clinical and research implications. Since flow limitation is determined by ν and airway cross-sectional area 19, the present authors hypothesise that the physical ability of the airways to carry large flows is fundamentally different in pre-term and full-term infants, and that this probably cannot be accounted for simply by the absolute reduction in airway dimensions found in such infants. Interestingly, wave propagation and mechanical properties of the airway walls determine when the airway walls start to resonate (wheezing), since it is then that energy is transversally dissipated into the walls. Based on the present findings, it is likely that these wheezing phenomena in pre-term infants occur in a different frequency range than in older children and adults.

Based on these findings, it is likely that flow limitation in pre-term infants is not only determined by airway diameter and airway obstruction, but additionally by the mechanical properties of the airway walls. Clinically, this means that ventilation strategies using positive end-expiratory pressure, which affects end-expiratory level, elastic recoil and thus airway wall elasticity, will have a large impact on airflow through the immature airways in these very young infants. In simple clinical terms, if their airway walls are stabilised with increasing end-expiratory level, this will subsequently facilitate flow through the airways. Furthermore, it implies that flow limitation in respiratory distress with high intrathoracic pressure during active expiration might limit the airways more dramatically in pre-term than in healthy full-term infants. Drugs such as bronchodilators, which not only cause airway dilation but also increase airway wall compliance 52, may therefore reduce the flow in these collapsible immature airways. Bronchodilators thus need to be used with care, as they may potentially cause adverse effects in such infants 53, 54. This study has emphasised the importance of airway wall mechanics in pre-term infants; future studies are required to investigate whether factors such as post-inflammatory airway wall remodelling, sex, ethnic group, post-natal age or tobacco exposure have additional effects on airway wall mechanics in these immature infants.