*Dr K. D. Kalache, Perinatology Research Branch, NICHD, NIH, Wayne State University/Hutzel Hospital, 4707 St Antoine Boulevard, Detroit, MI 48201, USA.
Objective To examine changes in intra-tracheal fluid flow parameters during fetal breathing movements throughout the second half of pregnancy in the normally developing human fetus.
Design Prospective cross-sectional study.
Setting Fetal medicine unit at the Charité University Hospital in Berlin.
Methods Assessment of tracheal fluid flow was attempted in 340 healthy fetuses (GA 20–40 weeks) in which fetal breathing movements were seen by B-mode scan. Colour Doppler was applied to visualise the tracheal fluid flow, followed by spectral Doppler to record the velocity waveforms. The records of 53 fetuses divided into five gestational age groups (20–23, 24–27, 28–31, 32–35 and 36–40 weeks of gestation) containing 40 or more continuous breathing cycles (inspiration and expiration) were considered for analysis. Only regular breathing phases were examined and the volume obtained by integration of the tracheal fluid flow displaced during fetal breathing movements was calculated.
Results The intra-tracheal flow volume moved during inspiration (Vi) and expiration (Ve) increased until 36 weeks of gestation after which there was a flattening until term. This suggests either a reduction of lung liquid production or a diminished lung liquid volume. The median difference between Vi and Ve was positive in the first four age groups and negative in the last one suggesting that, in mature fetuses, the effect of fetal breathing movements no longer results in an influx.
Conclusions Our data demonstrate a modification in fetal behaviour that manifests itself during the last four weeks before birth and has the potential to reduce lung liquid volume.
It is now known that fetal lungs develop as fluid filled organs and that the clearance of the liquid from the potential airspace must be achieved before the first breath. It is, however, wholly unknown whether lung liquid is lost before labour, during labour or postnatally. Experiences in the sheep fetus1 and epidemiological observational studies in humans2 have shown that there is increased morbidity in newborns delivered electively by caesarean section compared with those delivered spontaneously suggesting that a certain amount of lung liquid is removed during labour. There is also a growing evidence that respiratory problems can occur in the newborn when delivered, even spontaneously at 37 or 38 weeks, suggesting the presence of a second mechanism that may reduce lung liquid acting over the last two or three weeks of pregnancy3–5.
Studies in fetal lambs have demonstrated that fetal breathing movements are associated with intra-tracheal pressure deflections and small tidal movements of fluid reflecting the fetal respiratory efforts6,7. In recent years, major interest has focused on the observation of this phenomenon in human fetuses. Improvements in the diagnostic potential of ultrasound equipment has made possible the assessment of fetal breathing movements-related fetal nasal8,9 and tracheal10–12 fluid flow velocities. The present study was undertaken to use ultrasound to study changes in fetal breathing movements-related intra-tracheal fluid flow parameters and address the question of whether there is a reduction in lung liquid volume before birth.
Fetal breathing movements related to tracheal fluid flow was studied in 340 fetuses when rhythmic diaphragmatic movements were identified using real-time ultrasonography. Healthy mothers, between 20 and 40 weeks of pregnancy, had been referred to our centre for routine antenatal ultrasonography. All pregnancies were dated using the last menstrual period and fetal biometry in early pregnancy. Women included in the study were taking no medications and were not in labour at the time of examination. Each fetus was studied once. The results were analysed off-line and did not alter the management of pregnancy. The hospital ethical committee approved the study protocol, and informed consent was obtained from women who participated voluntarily in the study.
All ultrasonographic examinations were performed between 8am and 1pm before a meal. All women were placed in the supine position, with head and trunk slightly elevated. Most study periods lasted 20–40 minutes depending on success in obtaining representative Doppler curve signals. Recordings were made with a commercially available ultrasound system (HDI 3000, Advanced Technology Laboratories, Solingen, Germany) applying colour and spectral Doppler analysis using a 7–4 or a 4–2 MHz convex probe. Whenever fetal breathing movements were detected during real-time B-mode imaging, the transducer was moved towards the fetal neck. After obtaining a clear coronal view of the trachea, colour Doppler was switched on to visualise ‘streaming’ of the intra-tracheal fluid. This was colour-coded in red or blue depending on the flow direction (Fig. 1). At this phase of the examination the sample volume was placed over the coloured area at a level just beneath the infraglottic cavity. The length of the sample volume (3–5 mm) was chosen so as to cover the trachea. The angle of insonation in each case was less than or equal to 60°. The colour mode was then switched off to rule out displacement of the sample volume while the Doppler spectra were being collected (Fig. 2). The fetal breathing movements related to intra-tracheal velocity waveforms obtained from the infraglottic portion of the trachea consist of successive positive and negative Doppler signals corresponding to expiration and inspiration, respectively. The sweep speed was set to the lowest level to obtain a maximum simultaneous display of successive cycles on the screen. The maximal diameter of the trachea was measured at the level where the sample volume was positioned. All examinations were videotaped using an S-VHS recorder. All Doppler measurements were performed by the same observer (K.D.K).
The video-recorded curves were reviewed in slow motion and interesting sequences were digitised frame by frame using an oculus frame-grabber which processes both S-Video and RGB signals were transferred to a commercially available image archiving and communication system (ViewPoint Bildverarbeitung GmbH). The digitised frames were then loaded into Adobe Photoshop 4.0, and segments containing the Doppler curves were extracted and stored as 8-bit RGB flat TIFF files. These segments were consecutively loaded and assembled in Microsoft PowerPoint 7.0 such that we obtained a paper strip of assembled Doppler segments in which order, time and the velocity scale were strictly respected.
Only strips with more than 40 continuous breathing cycles in which each inspiration was followed by expiration without interruption by cessation of breathing or by loss of the ultrasonic signal were considered for analysis. Five regular breathing cycles were then transferred from video-tape to the digital disc of an off-line medical computer analysis system (Image Vue DCR, Nova Microsonics, Mahwah, New Jersey, USA). After calibrating both the time scale and velocity scale the average of five regular cycles from each case was determined for the fetal breathing movements-related intra-trachea flow-volume, which was calculated by using the formula:
where TVI is the time–velocity integral and d the diameter of the trachea.
Analysis revealed that the data were not normally distributed and the Mann–Whitney U test was used to assess differences between the different age groups. Differences were considered to be significant with values of P < 0.05.
Data from 53 fetuses were considered for final analysis. The data were divided up into five groups (20–23, 24–27, 28–31, 32–35 and 36–40 weeks of gestation) based on gestational age at the time of the study (Table 1). All 53 babies were born at term. Routine postnatal examinations confirmed the absence of anomalies and there was no evidence of respiratory insufficiency.
Table 1. Classification of the five age groups with number of breathing cycles (inspiration + expiration).
Gestational age (weeks)
Total no. of fetuses
Median (range) of recorded breathing cycles with range in parentheses
Before 36 weeks of gestation, the median intra-tracheal inspiratory flow volume increased and reached its maximum in 32 to 35 weeks of gestation [Group 1: 0.16 mL (range 0.06–0.62); Group 2: 0.39 mL (range 0.12–1.30); Group 3: 0.90 mL (range 0.33–1.81); Group 4: 1.98 mL (range 0.65–4.53)]. At 36 to 40 weeks of gestation, it was almost the same as at 32 to 35 weeks of gestation [Group 5: 1.82 mL (range, 0.20–4.67)] (Fig. 3).
There was no difference in the median expiratory intra-tracheal flow volume after 36 weeks of gestation [Group 5: 2.0 mL (range 0.52–4.85)]. Before this time, this volume increased and reached its maximum in 32 to 35 weeks of gestation [Group 1: 0.14 mL (range 0.06–0.53); Group 2: 0.36mL (range 0.12–1.09); Group 3: 0.96 mL (range 0.19–1.64); Group 4: 1.98 mL (range 0.72–4.28)] (Fig. 4).
Before 36 weeks of gestation, the median difference between the volume of tracheal fluid moved during inspiration and the volume of tracheal fluid moved during expiration was positive and reached its maximum at 32 to 35 weeks of gestation [Group 1: 0.02 mL (range 0.005–0.09); Group 2: 0.06 mL (range −0.08–0.27); Group 3: 0.06 mL (range −0.24–0.39); Group 4: 0.25 mL (range −0.24–0.80)]. At 36 to 40 weeks gestation, it significantly decreased to become negative [Group 5: −0.19 mL (range −0.55–0.48)] (Fig. 5).
In the present study fetal lung function during the second half of normal pregnancy was assessed by analysing the effect of fetal breathing movements on intra-tracheal fluid flow using the Doppler technique. The approach, which we have previously termed pulmonary Doppler ultrasonography13 is a major advance over other modalities to assess fetal breathing movements for the following reasons. First, it is sensitive enough to allow a differentiation between distinct fetal breathing patterns as early as the second trimester14. Hence, standardised sets of measurements can be obtained by considering exclusively the regular symmetric pattern and overcome bias due to the great variability of fetal breathing movements. Second, the knowledge gained in previously reported fetal breathing movements studies was limited to the incidence and the frequency of fetal breathing movements15–19 and later to timing components of the individual breath cycle8,20. However, these are mainly related to central nervous maturation and not necessarily to lung development, which requires an adequate lung growth along with normal maturity. A major advantage of pulmonary Doppler ultrasonography is that the intra-tracheal flow volume during each breathing cycle can be estimated and can be used to differentiate lethal and less severe lung hypoplasia in fetuses with congenital diaphragmatic hernia21. Pulmonary Doppler ultrasonography therefore offers an alternative form of fetal lung function assessment.
Experiments in the fetal lamb have shown that the lung liquid volume rises until the end of gestation, reaching a level of 50 mL/kg of fetal body weight on the day before labour22,23. According to these authors, the entire volume of fetal lung liquid (∼250 mL) in ovine fetuses is cleared during normal labour24. In contrast to these studies, others have suggested that the volume25,26 and production rate27 of fetal lung liquid decrease up to 10 days before the onset of labour. Furthermore, it has been found that Na+ reabsorption, which is the only mechanism yet identified that could decrease lung liquid volume during labour, is able to remove only about 20 mL of liquid from the lungs (∼10% of the entire volume of fetal lung liquid)28. This leaves open the possibility that the remaining volume of fetal lung liquid (∼90%) must be cleared by mechanical compression in the pelvis and is still in the lungs of newborns delivered by caesarean section. This is in contrast with the slight increase in morbidity revealed in animal experiments and epidemiological studies in newly born babies delivered by caesarean section. Thus, it is likely that the vast bulk of the liquid present in the fetal lung is cleared several days before labour by a second mechanism.
Our study shows that both the inspiratory and expiratory intra-tracheal flow volume increased with advancing gestation until 35 weeks of gestation. In human pregnancies, this phase of lung development (20–35 weeks of gestation) is characterised by an important increase of lung growth29, tracheal growth13,30, lung compliance31 as well as lung liquid volume and secretion rate23. Which of these factors mainly influence fetal breathing movements related to upper airways dynamic during this stage of lung development remains speculative. After 36 weeks of gestation we observed a flattening of the inspiratory and expiratory intra-tracheal flow volume until term despite the fact that this phase is, similarly to the previous one, associated with an increase of lung growth, airway growth, and lung compliance. Thus, our findings support previous studies that have shown a decrease in the volume of lung liquid decreases before the initiation of labour. Our results also suggest that, during this phase of lung development, intra-tracheal flow volume is mainly influenced by lung liquid volume and secretion rate rather than by lung and tracheal growth.
Berger et al.1 showed that the bulk (>75%) of the liquid that fills the lungs of the fetal lamb at 140 days (which corresponds to 38 weeks of gestation in human) of gestation is cleared at some time before normal term birth. They demonstrated that a small amount (∼10%) is shifted into the lung parenchyma, that 25% remained in the airspace of the lungs at the end of labour and estimated that the remaining volume of lung liquid (∼66%) left the lung between 140 days and the beginning of labour. Cessation of fetal breathing movements over the three days preceding spontaneous delivery together with tonic and coughing activity were proposed as possible mechanisms that have the potential to reduce this volume, assuming that the upper airways of the fetus act as a ‘one-way valve’. However, animal experiments32 and Doppler studies in human fetuses11 have provided strong evidence that fetal breathing movements can result in movement of fluid from the amniotic cavity into fetal upper airways.
Our study demonstrates that after 36 weeks, during regular fetal breathing movements, there is a net efflux of fluid from the lung to the amniotic cavity. This suggests that mature fetuses have a tendency to ‘expire’ more than ‘inspire’ during regular breathing movements. Another explanation is that by 36 weeks of gestation, the rate of lung liquid formation has decreased such that the effect of fetal breathing movements no longer results in an influx. Nevertheless, this is an important observation as it provides evidence for a reduction in lung liquid before the onset of labour. Although some overlap between the five age groups was seen, statistical analysis confirmed this trend. It has been believed for a long time that the fetal lung after 36 weeks of gestation is morphologically mature and that an increase of exchanging surfaces takes place which is of no relevance to postnatal lung function. However, there is now ample evidence that the last weeks before birth are of some importance for postnatal lung function and that elective caesarean section before the due date is associated with a higher perinatal morbidity2,5,33,34. These problems may be related to a deficient surfactant system but also to insufficient lung liquid clearance.
In conclusion, our findings support a possible explanation as to why late gestation is beneficial to the fetus. We have shown a modification in tracheal lung liquid which may relate to a reduction in lung liquid before labour.
This study was supported by the Federal Ministry for Education and Research-Grant ‘Perinatale Lunge’ (No. 01ZZ9511).