Article Text


Nasal continuous positive airway pressure: does bubbling improve gas exchange?
  1. C J Morley,
  2. R Lau,
  3. A De Paoli,
  4. P G Davis
  1. Neonatal Services, Royal Women’s Hospital, Melbourne, Victoria 3053, Australia
  1. Correspondence to:
    Professor Morley
    Neonatal Services, The Royal Women’s Hospital, 132 Grattan Street, Carlton, Melbourne, Victoria 3053, Australia;


In a randomised crossover trial, 26 babies, treated with Hudson prong continuous positive airway pressure (CPAP) from a bubbling bottle, received vigorous, high amplitude, or slow bubbling for 30 minutes. Pulse oximetry, transcutaneous carbon dioxide, and respiratory rate were recorded. The bubbling rates had no effect on carbon dioxide, oxygenation, or respiratory rate.

  • CPAP, continuous positive airway pressure
  • Paco2, arterial carbon dioxide
  • Spo2, oxygen saturation
  • Tcco2, transcutaneous carbon dioxide concentration
  • TcPo2, transcutaneous oxygen
  • continuous positive airway pressure
  • CPAP
  • carbon dioxide
  • ventilation

Statistics from

Nasal continuous positive airway pressure (CPAP) is an effective mode of respiratory support for neonates used in many nurseries.1 It can be delivered in several different ways. One technique uses short binasal prongs—for example, Hudson prongs—where the pressure in the device is generated by a continuous flow of gas past the nasal prongs with the distal end placed a set depth under water.2 As the gas flows through the system, it bubbles out underwater (fig 1). It has been suggested that oscillations in the pressure, due to bubbling, contribute to gas exchange by delivering low amplitude, high frequency oscillations to the lungs.3 If true, this simple, inexpensive system would be a useful method of improving the effectiveness of nasal CPAP.

Figure 1

 Diagram of “Bubbly Bottle” nasal continuous positive airway pressure (CPAP) system.

This study aimed to determine whether the pressure oscillations caused by bubbling affect transcutaneous carbon dioxide (Tcco2), transcutaneous oxygen (TcPo2), oxygen saturation (Spo2), heart rate, and respiratory rate.


This study is based on two concepts. Firstly, the vigorousness of underwater bubbling changes as flow through the CPAP system is altered. With high flow, bubbling is very vigorous with a high pressure amplitude. Flow can be reduced to a level where bubbling almost stops, but pressure is maintained with the meniscus still at the bottom of the underwater tube. Secondly, high frequency oscillation improves the removal of carbon dioxide from lungs and blood.4 Tcco2 is easily measured and is closely related to arterial carbon dioxide (Paco2) over short time periods.5

We decided to exploit these concepts to investigate whether bubbling that was vigorous, with a high amplitude, or slow influenced a baby’s gas exchange.

A convenience sample of stable preterm babies treated with bubble nasal CPAP (Fisher & Paykel Healthcare, Auckland, New Zealand) using Hudson prongs (Hudson Respiratory Care Inc, Temecula, California, USA) was studied. They were randomised, using sealed envelopes, to start on either vigorous, high amplitude, or slow bubbling. Slow bubbling was achieved by lowering the gas flow to the point when the bubbling “just” occurred so that the pressure was maintained with the gas/water meniscus at the end of the underwater tube. The vigorous, high amplitude bubbling was obtained by increasing the gas flow through the system by 3 litres/min above the flow required to obtain the lowest possible bubbling. Babies were studied for 30 minutes, and then crossed over to the alternative level of bubbling. During this time, the inspired oxygen and gas flow rate were not changed and the baby was not handled. The Royal Women’s Hospital Research and Ethics Committees approved the study.

The nasal CPAP was measured from a side port on the prongs using a low range transducer (Sensym; Sensortechnics, Puchheim, Germany; range 0–13 cm H2O). A transcutaneous monitor measured carbon dioxide and oxygen, and a pulse oximeter measured Spo2 and heart rate. Respiration rate was measured using the Graseby monitor (Graseby Medical Ltd, Watford, Hertfordshire, UK). Signals from these devices were recorded at 100 Hz using the Spectra physiological recording system (Grove Medical Ltd, Hampton, TW12 2EG, UK). The median value for each signal was calculated over each 10 seconds and recorded as a trend during the study. The last 15 minutes of each 30 minute recording was analysed to provide the most stable signal and allow washout of any effects from the previous flow rate.4

The study of 26 babies was sufficient to detect a difference of 3.0 mm Hg (0.39 kPa) in Paco2 with 80% power, if the SD of the difference was 5.0 mm Hg (0.65 kPa), derived from previous data.


Twenty six babies were studied. Their median gestational age was 27 weeks (range 24–32), birth weight 1033 g (range 604–1980). The nasal CPAP was 6 cm H2O (range 5–9). The baseline gas flow rate was 6 litres/min (range 5–9), and the inspired oxygen 21% (range 21–30).

Table 1 shows the results obtained during vigorous, high amplitude bubbling and slow bubbling. There was no effect of bubbling rate on Tcco2, TcPo2, Spo2, heart rate, or respiratory rate. The correlation coefficient for Tcco2 between the two bubbling rates was 0.936.

Table 1

 Effect of bubbling rate on Tcco2, TcPo2, Spo2, heart rate, and respiratory rate

Despite the fact that we were very careful to not change the CPAP pressure between the different bubbling regimens, there was a slightly lower pressure with slower bubbling (table 1). Figure 2 shows an example of the pressure amplitude at the device during vigorous, high amplitude and slow bubbling. For all episodes, the median (interquartile range) was 2.7 cm H2O (2.5–4.0) for slow and 6 cm H2O (4.6 to 7.1) for vigorous, high amplitude bubbling.

Figure 2

 A recording, at 100Hz and displayed at 1 cm/s, of the pressure recorded at the nasal continuous positive airway pressure device during slow bubbling and vigorous, high amplitude bubbling. The y axis shows pressure in cm H2O.


This study has shown that, after changing from vigorous, high amplitude to slow bubbling for 30 minutes there was no difference between Tcco2, Spo2, heart rate, or respiratory rate. The lack of effect on Tcco2 and Spo2 is similar to the results of Lee et al,3 who did a crossover study of bubble CPAP, compared with CPAP from a ventilator, through an endotracheal tube. Although they saw a small reduction in respiratory rate, and minute volume, which we could not measure with nasal CPAP. This may be because the oscillation amplitude at the device was relatively small, at about 5 cm H2O even with vigorous, high amplitude bubbling. This is about 10% of the pressure amplitude applied to an endotracheal tube during high frequency ventilation. Once transmitted to the alveoli, this pressure difference is unlikely to have much effect on Paco2.

There was a small but significant difference in the CPAP prong pressure between vigorous and low bubbling. This had no effect on the Tcco2 or oxygenation. We realised this was due to the effect of the bubble (just over 1 cm in diameter) on the end of the underwater tube. With vigorous bubbling there was a bubble present all the time, but with slow bubbling it was intermittent. In summary, we found no evidence to support the suggestion that pressure oscillations during Bubble Bottle CPAP improve gas exchange.


PD is supported by an NHMRC Practitioner Fellowship.


View Abstract


  • Competing interests: none declared

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.