INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: infant; premature; lung function tests

In preterm infants, lung development normally taking place in utero will instead occur after birth and under very different conditions. They include active breathing with strain and relaxation of immature lung tissue, as well as lung perfusion with full cardiac output and with exposure to considerably higher oxygen tension than during fetal life. Although it is uncertain when alveolarization begins in the human lung, it is generally accepted that the structure of the terminal respiratory unit changes over the last trimester of gestation (1, 2). Mechanical factors, both static and dynamic, have been shown to affect fetal lung growth and development (3). It has also been demonstrated in animals that formation of alveoli can be impaired by mechanical ventilation (4). Delayed and disturbed formation of the terminal respiratory units is a typical phenomenon in bronchopulmonary dysplasia (BPD) (5).

These results emphasize that lung development is a vulnerable process in immature animals and humans. However, little is known about whether the maturation process of the lungs of healthy human preterm infants is affected by preterm birth. In some studies, prematurity has been considered a risk factor for impaired airway function in childhood (6) and for asthma in young adults (11), whereas others have failed to identify long-term consequences (12).

The aim of this work was to assess the consequences of preterm birth for functional development of the lungs over the period from premature birth until term. Our hypothesis was that an early transition from intrauterine to extrauterine environment changes pulmonary development and function. The study was designed to compare healthy preterm infants, without any history of lung disease, with healthy term infants at the same postmenstrual age (PMA) close to term. Different aspects of lung properties and function were studied: lung volume, static compliance of the respiratory system, indices of small and large airway function, and total and alveolar ventilation. In a subset of the preterm infants, lung function was also assessed at 32 to 37 wk PMA to monitor functional lung development over the last weeks up to term.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Thirty-two preterm infants, born at a mean gestational age of 29.5 wk (range, 25 to 33 wk) with a neonatal clinical course without signs of illness and without need of extra oxygen after 3 d of age, were studied at a mean age of 75 d (range, 39 to 103), corresponding to a mean PMA of 39.8 wk (range, 38 to 41). Sixteen were males and 16 females. Their mean birth weight was 1.42 kg (range, 0.84 to 2.50) and body weight at time of study was 3.12 kg (range, 2.41 to 4.25). Twenty-one of the mothers were treated with at least two doses of antenatal corticosteroids. Two of the mothers were smokers. Twenty of the preterm infants were also studied at an average PMA of 34.2 wk (range, 32 to 37).

The preterm infants were compared with 53 full-term infants, studied at an age of 28 to 72 h at a PMA of 39.9 wk (range, 37 to 42) and with a mean body weight of 3.55 kg (range, 2.66 to 4.54). Twenty-eight were males, 25 females. They were recruited from the normal nursery of the hospital. The selection criteria were gestational age at birth of 37 to 42 wk; appropriate weight and length for gestational age; no signs of respiratory distress, malformations, or disease; and availability for study beyond 24 h of age. No mother had smoked during pregnancy and none was treated with corticosteroids. All infants were studied after informed parental consent and in the presence of at least one of the parents. The study was approved by the ethical board of the medical faculty.

Procedures

Functional residual capacity (FRC) and indices of gas mixing efficiency were measured by the multiple breath nitrogen washout method, applied for newborn infants as described previously (15). In brief, the infants were studied supine after a meal during sleep with regular breathing. No sedation was used. A Rendell-Baker mask size 0 (Soucek Mask, London, UK) was gently fitted over the nose and mouth of the infant. The mask was connected to the side wall of a tube with a constant flow of breathing gas (approximately 150 ml/s). The equipment dead space was calculated to be below 1 ml. At the outlet, the bypass flow was led through a pneumotachograph (Fleisch 0; Hugo Sachs Elektronik, March-Hugstetten, Germany) with a linear response at actual flow rates. At the inlet, a valve was attached so that breathing gases could be switched instantaneously. When the mask was fitted, breathing flow rates were recorded as a modulation of the constant bypass flow through the pneumotachograph. The probe of a fast infrared nitrogen meter (Hewlett-Packard 47320A, Hewlett-Packard Company) was mounted at the mask outlet so that the nitrogen concentration of the gas passing into and out of the infant could be recorded continuously.

To obtain a washout recording, breathing gas was switched to 100% oxygen during an expiration and continued until end-expiratory nitrogen concentration was below 2% (1/40th of the concentration at the start of the washout). The pneumotachograph was calibrated with known gas flow after each test. The nitrogen meter was calibrated with room air.

Calculations were made by computer. All signals were digitized with 12 bits accuracy at a frequency of 250 Hz and stored on disk. For the analysis, the time scale of the nitrogen signal was corrected for the measured transit time of the meter (usually 50 ms). The total exhaled N₂ volume was calculated as the sum of the N₂ volumes exhaled in each breath during washout; this volume was obtained by summing the product of simultaneously measured flow, N₂ concentration signals, and the inverted sampling frequency. No correction was made for tissue-derived nitrogen. FRC was calculated as the total expired nitrogen volume during washout, divided by end-tidal nitrogen concentration preceding the washout. Efficiency of gas mixing was assessed by calculating moment ratios of the course of nitrogen elimination. The zero, first, and second moments of end-tidal nitrogen concentration with accumulated expired volume as the independent variable, normalized by FRC, were calculated and the ratios of the first (M₁/M₀) and the second moment (M₂/M₀) to the zero moment, respectively, were used as described previously (15). Lung clearance index (LCI) was calculated as the amount of ventilation, expressed as the number of FRC turnovers necessary to reduce alveolar nitrogen to 1/40th of its original concentration during washout (16). The washout test was performed two to three times in each infant with at least 5 min in between.

Tidal volume, breathing frequency, and minute ventilation were calculated from the recording of flow. Dead space was assessed from the dilution coefficient calculated by fitting an exponential function to the expired nitrogen volume curve during washout (17), and alveolar ventilation was calculated using these data together with breathing frequency and minute ventilation.

Compliance and resistance of the respiratory system were assessed by the single occlusion technique (18) with computer-assisted calculations. A Rendell-Baker mask, attached to a pneumotachograph (Fleisch 0), was used and occlusions were performed manually close to maximal inspiration. Pressure was measured inside the mask. Flow- volume and pressure plots of the occluded breaths were presented on a screen. Expirations with a pressure plateau of at least 0.2 s and a linear flow-volume plot after occlusion were used for automatic calculation of the time constant of the respiratory system. Compliance of the respiratory system was then calculated from the volume and pressure differences at occlusions and resistance and conductance from the time constant of the respiratory system and the compliance data. The result was calculated as the mean of 4 to 10 occlusions.

Statistics

Group sizes were calculated to give a 90% power to detect differences between group means of one standard deviation of each normally distributed variable, with an alpha value of 0.05. The number of preterm infants studied was extended to 32 to allow for possible non-normally distributed variables. The full-term infants were studied at the same period of time and by the same technicians as the preterm infants but as many as 53 infants were included, as they also were part of another study. Variables were tested for normality by graphical plots. Variables known to vary with body weight with regression lines passing close to zero were normalized by dividing with body weight. Mechanical parameters were normalized to FRC. Student's t test for unrelated samples and Mann Whitney's U test were used to compare group means, and t test for paired data to compare data over time in the preterm group. Differences in variance were tested by the F test. Ninety-five percent confidence intervals (CI) of differences of means between the groups were calculated.

The problem with multiple significance tests was dealt with by accepting a difference in a category of variables (e.g., ventilation, mechanics, small airway function) at the 5% level only if at least one difference of the set tested had a p value less than 0.05/n, where n was the number of tests within the set.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The two groups studied at term differed in body size. Mean body weight was 3.55 kg in the full-term infants and 3.12 kg in the preterm infants (p < 0.001). Body lengths were 50.8 versus 48.6 cm (p < 0.001). At term the groups differed in total ventilation, calculated per kilogram body weight, but not in alveolar ventilation (Table 1). This was mainly due to a larger dead space in the preterm infants, 5.1 versus 4.4 ml/kg (p < 0.01), who also had larger relative tidal volumes (Table 1). Mean FRC was lower in preterm infants but after correction for their lower body weights at time of study the difference was less apparent, although still significant (Table 2).

There were considerable differences in compliance and resistance of the respiratory system between the groups, with lower compliance and higher resistance in the preterm group (Table 3). The difference in resistance could, however, be explained by differences in lung size, as specific conductance was not significantly different. On the other hand, there was a remarkable difference in compliance also after differences in FRC were taken into account. Specific compliance of the preterm infants was only 73% of that of the full-term group. Also, the time constant was significantly shorter in preterm infants. We also observed that standard deviation of compliance and time constant was significantly higher in the infants born at term. Compliance, but not specific compliance, tended also to have a skewed distribution but the nonparametric test gave the same result as the t test. The indices of gas mixing efficiency also showed significantly impaired gas mixing in the preterm group in comparison to full-terms (Table 4).

Twenty of the preterm infants were also examined at 32 to 37 wk PMA as well as at term (Table 5). FRC and compliance increased significantly from the first to the second measurement and gas mixing efficiency in terms of M₁/M₀ and LCI improved. When adjusted for growth in terms of body weight, FRC, however, was constant over the period. When the mechanical parameters were adjusted for lung size, both compliance and conductance were found to decrease significantly from the first to the second study.

The preterm infants whose mothers were treated with at least two doses of corticosteroids to prevent preterm birth (n = 21) were compared with the other preterm infants (n = 11). No significant differences were found in FRC (95% CI of the difference was 2.0 to 3.2 ml/kg), ratio of compliance to FRC (Crs/FRC; 0.0082 to 0.0096 cm H₂O¹), resistance (12 to 33 cm H₂O/L/s), or M₁/M₀ (0.31 to 0.10).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study we have tested the hypothesis that premature birth per se affects functional development of the lungs of preterm infants. Our results indicate that this is the case and that it has evident implications for lung volume, mechanics of breathing, gas mixing efficiency, and ventilation at term. We have considered two potential confounders of the results. The majority of the preterm infants had been exposed to at least two doses of antenatal steroids. When their results were compared with those of the rest of the preterm group, no significant differences were found. Although the possibility of a type II error must be considered, the results indicate neither a major impact of prenatal steroid treatment on functional lung development at term in the human infant nor that the differences found between the study groups were confounded by the difference in prenatal steroid exposure.

Another potential confounder was the fact that body size of the infants born at term was larger than that of the preterm infants at the time of the study, although PMA was the same. However, the regression line of the ventilatory variables on body weight passed close to zero, making adjustment by division by body weight the most simple tool to adjust for the differences in body size. The mechanical parameters were adjusted by means of measured FRC. The moment ratios and LCI are not sensitive to body size (15).

Our data show that premature infants have a moderately reduced FRC and a more pronounced impairment of gas mixing efficiency when compared with infants born at term. The gas mixing efficiency is strongly determined by gas flow in peripheral airways and by volume and arrangement of terminal airspaces (19). It is of interest that reduced FRC and gas mixing efficiency have been found in infants with BPD when compared with healthy preterm infants (20). In animal models of BPD (21), as well as in human infants with the disease (6, 22), the major changes of lung histology are fewer and larger terminal respiratory units. We propose that the reduced FRC and impaired gas mixing efficiency in the group of healthy preterm infants also are caused by impaired development of the terminal respiratory units. We suggest that premature breathing leads to disruption of developmental pathways essential for normal alveolarization.

Another important result was a considerably reduced compliance and specific compliance of the respiratory system in the group born preterm. Compliance also varied significantly more in the term group. Although the period of immediate adaptation to breathing had passed (17), it is still possible that great individual differences in lung water processing may be the reason for this variation. As FRC was only moderately reduced in the preterm group, the result can be interpreted as a gain of elastic recoil at lung volumes above FRC in the preterms. The change can be attributed to the lungs, to the chest wall, or to both. There are no data at hand suggesting that the soft thorax of preterm infants would stiffen considerably over the first months of postnatal life. One might speculate that the reduced compliance instead reflects an increase of elastic elements in the preterm lung, possibly similar to BPD (5), or to abnormal surfactant function.

There are few previous reports of direct comparisons of respiratory compliance in term and preterm infants near term. However, Chu and coworkers (23) already in 1964 reported specific lung compliance to be reduced in preterm infants. More recently, Merth and coworkers (24) found no difference between preterm and term infants in static compliance of the respiratory system, related to crown-heel length. Neither did they find any difference in gas mixing efficiency. However, the variation in PMA and body size was considerable in that study, which reduced the possibilities to detect differences. Although several studies have suggested an impaired airway function at school age in infants born preterm without any neonatal disease (6), only one has reported compliance measurements at that age. Parat and coworkers (9) showed reduced dynamic lung compliance in 7-yr-old children born preterm. Although more studies are needed before conclusions can be drawn, it might be possible that the reduced compliance demonstrated in the present study persists during childhood.

The infants born preterm breathed with larger tidal volumes, corrected for body weight, but not with higher frequency than full-term infants did when studied at the same PMA. However, their relative alveolar ventilation was similar and the difference was explained by a larger dead space in the preterm group. The localization of the extra dead space is unknown. It may well be that it is the result of the differences in gas mixing properties. The developmental pattern between 34 and 40 wk PMA in the preterm infants showed an increase of FRC in harmony with body weight but no signs of catch-up. Gas mixing efficiency, on the other hand, improved significantly, which can be explained by structural maturation of the terminal respiratory units. However, it did not reach the level of the infants born at term. Specific compliance in premature infants at 34 wk was close to that in full-term infants, but decreased significantly thereafter. This is consistent with a significant increase in elastic recoil of the respiratory system over this period. Whether this reflects a developmental pattern particular for premature infants or if it is present also in lungs developing in utero is not known. In any case, the result of this development for the preterm group was a considerably lower specific compliance at 40 wk PMA than the infants born at term.

Specific conductance did not differ between the groups at 40 wk, which suggests a similar development of large airways. A dramatic decrease in specific conductance occurred in preterm infants between 34 and 40 wk of PMA, similar to a previous report by Stocks and Godfrey (25). Resistance did not change significantly over the same period. The change in specific conductance may be interpreted in terms of disproportional lung growth with a major increase in the size of terminal airspace, as reflected by the expanding FRC, but essentially unchanged conducting airways.

The lung function abnormalities in healthy preterm infants shown in this study reflect disturbances in lung development that may have started before or after birth. Several postnatal conditions may adversely affect lung development in preterm infants during a period corresponding to the last trimester of pregnancy. They include mechanical forces related to breathing, exposure of epithelial and endothelial cells to higher oxygen tensions than during fetal life, influence of air flow on airway epithelium, absence of the distending effect of lung fluid, change of lung perfusion and blood volume, and absence of maternal or placental factors, but also nutritional or endocrine differences between the fetal and the extrauterine environment. Lung development may also be affected prenatally by the pathophysiologic process leading to preterm birth. For example, oligohydramnios is related to preterm labor (26) as well as to abnormal lung maturation (27). Chorioamnionitis has been related to preterm birth (28), to a fetal systemic inflammatory response (29), and to chronic lung disease (30). Cytokines are known to affect lung development, and induced expression of interleukin-1 (IL-1) in the lungs of transgenic mice during late gestation and postnatal period was recently reported to affect the terminal respiratory units with large alveoli with few septa (31). Also, in utero endotoxin exposure in lambs caused dilation of terminal airways (32).

We conclude that although the mechanisms are still unknown, preterm birth per se changes the normal development of lung function.