Article Text


Randomised controlled trial of nasal continuous positive airways pressure (CPAP) in bronchiolitis
  1. Lena P Thia,
  2. Sheila A McKenzie,
  3. Tom P Blyth,
  4. Caro C Minasian,
  5. Wanda J Kozlowska,
  6. Siobhan B Carr
  1. Department of Paediatric Respiratory Medicine, Royal London Hospital, London, UK
  1. Lena P Thia, Department of Paediatric Respiratory Medicine, Royal London Hospital, Whitechapel Road, London, UK; lthia{at}


Aims: To compare continuous positive airways pressure (CPAP) with standard treatment (ST) in the management of bronchiolitis.

Methods: Children <1 year of age with bronchiolitis and capillary Pco2 >6 kPa were recruited and randomised to CPAP or ST and then crossed over to the alternative treatment after 12 h. ST was intravenous fluids and supplemental oxygen by nasal prongs or face mask. The change in Pco2 was compared between the groups after 12 and 24 h. Secondary outcomes were change in capillary pH, respiratory rate, pulse rate and the need for invasive ventilatory support.

Results: 29 of 31 children completed the study. Pco2 after 12 h fell by 0.92 kPa in children treated with CPAP compared with a rise of 0.04 kPa in those on ST (p<0.015). If CPAP was used first, there was a significantly better reduction in Pco2 than if it was used second. There were no differences in secondary outcome measures. CPAP was well tolerated with no complications identified.

Conclusions: This study suggests that CPAP compared with ST improves ventilation in children with bronchiolitis and hypercapnoea.

Statistics from

In England, 28 of every 1000 hospital admissions in children <1 year of age are due to respiratory synctial virus (RSV) bronchiolitis1 and 2–5% of these patients have respiratory failure.2 Continuous positive airways pressure (CPAP) is a treatment widely used for respiratory failure in neonates3 and is now used in several centres for the management of bronchiolitis. However, there are no randomised controlled studies to assess its value in this condition.

In an observational study of 23 infants managed over a 5-year period,4 the mean respiratory rate (RR), pulse rate (PR) and partial pressure of carbon dioxide in arterial blood (Pco2) fell significantly within the first 3 h of CPAP treatment. Similar benefit after 2 h was demonstrated in a further 10 children treated with nasal CPAP.5 These observational studies suggest that CPAP might be an effective intervention in bronchiolitis. It could be that the early use of CPAP may prevent the need for invasive ventilation. Because invasive ventilation is not usually required in bronchiolitis, to consider this as a primary end-point would need a multicentre trial with large numbers of subjects. Therefore, the purpose of this study was to compare CPAP with standard treatment (ST) in the management of moderately severe bronchiolitis to determine whether the treatment improved ventilation. Results could help inform a more ambitious study.



The study was conducted over three winters from October 2002 to March 2005. Children were <1 year of age with a clinical diagnosis of bronchiolitis and capillary Pco2 measurements >6 kPa. Children with congenital heart disease, neuromuscular disease, and mid-face dysmorphism prohibiting use of nasal prongs were excluded. Those requiring immediate invasive ventilation due to recurrent apnoeas, profound hypoxia, collapse or Pco2 >12 kPa were also excluded. Children with chronic lung disease and those who had been premature were not excluded.

Study protocol

The study lasted for 24 h. Children were randomised in blocks of four, blind to those who recruited the subjects. Eligible children were randomised to receive either (a) standard treatment plus nasal CPAP for 12 h followed by standard treatment alone for the next 12 h or (b) standard treatment alone for 12 h followed by standard treatment plus nasal CPAP for the next 12 h. Standard treatment was defined as minimal handling, intravenous fluids and oxygen by nasal prongs or face mask. Nasal CPAP was applied using the Infant Flow System (Viasys, Conshohocken, PA) with pressures of 5–6 cm H2O. Both groups had supplemental oxygen to achieve oxygen saturation above 92%, recorded on the foot or hand by pulse oximeter. Corticosteroids, bronchodilators or adrenaline were not used.

Demographic data were recorded. PR, RR, capillary Pco2 and capillary pH were recorded at 6-h intervals (ie, before treatment or change in treatment and midway through the 12 h).

After completion of the 24-h study, further care was provided according to medical need. Children could also be withdrawn at any time at the discretion of the medical team for the following reasons: recurrent apnoeas, worsening hypercapnia and profound hypoxia despite maximal oxygenation. Reasons were documented.

Written, informed consent was obtained from parents. Ethical approval for the study was obtained from the local research ethics committee.

Statistical analysis

The data were inspected after the first 12 patients, as we did not know how variable the response to either CPAP or ST was likely to be. It was then estimated that 28 patients would be needed to show a difference of a fall of 1 kPa in Pco2 with 80% power at 0.05 significance at 12 h. This period of time was chosen empirically to allow for any wash-out effect. Paired t tests were used for continuous variables and Mann-Whitney testing for other data. Analysis of variance was used to compare demographic data and presenting Pco2 in patients recruited, not approached and refusals.


During the study period, 53 children were eligible. Eleven children were not entered into the study because their families were not approached for consent. This was either because the parents or translators were not present or a member of the study team was not available to take informed consent. Eleven families refused consent. Thus 31 children were recruited; 16 were randomised to receive CPAP first (fig 1).

Data from children not recruited for the study were examined. There were no statistically significant differences in Pco2 at presentation or in days of preceding illness between the group recruited, the group not approached and the group whose parents refused. Although those not in the study were significantly younger (median age: 1.5 vs 2.63 months, p = 0.03), their bronchiolitis was of similar severity to those in the study.

A total of 29 children completed the study. There were no significant differences in baseline demographic characteristics (table 1). Two children were withdrawn from the study. The first child was withdrawn at 10 min because of profound hypoxia; the child was electively treated with CPAP and improved. The second child was withdrawn with a Pco2 of >12 kPa at 9 h for invasive ventilation. Both had been randomised to standard treatment first.

Table 1 Baseline characteristics

The results in the two parts of the study differed. There was a significantly greater change in Pco2 following CPAP when given first than when given second (table 2). The changes in Pco2 between the beginning and end of each 12 h treatment arm for each patient are shown in fig 2.

Figure 2 CO2 (kPa) measurements for individuals during the study.
Table 2 Change in Pco2 (kPa) during CPAP and ST periods

No significant differences were seen in PR or RR. Children receiving early CPAP did not show any difference in length of stay compared with those who received it later (mean (SD) 6.3 (2.3) vs 5.6 (1.5) days; p = 0.16).

What is already known on this topic

  • Management of bronchiolitis in infants is mainly supportive.

  • There is no evidence to suggest major benefits from treatment with drugs.

What this study adds

  • CPAP is effective in reducing capillary Pco2 and if used earlier rather than later, is more effective in reducing Pco2 in moderately severe bronchiolitis.

  • CPAP is well tolerated by infants with no complications seen.

CPAP was tolerated well by all children. No child’s condition gave concern that there could be a significant air leak. Nine children (three in the first CPAP group and six in the first ST group) required one dose of triclofos as sedation to tolerate CPAP. There was no significant difference in Pco2 in the patients who had triclofos and those who did not at the end of the treatment periods.


This study is, to our knowledge, the first randomised controlled trial investigating the value of CPAP in moderately severe bronchiolitis. It has shown that the use of CPAP in addition to ST results in a significant reduction in Pco2 by 1.35 kPa over 12 h compared with ST alone. This change is similar to that demonstrated in previous observational studies.4 5 The difference between early and late use is possibly because early treatment prevents the development of additional airway collapse and further tiring. Thus, it appears that early use slows and reverses the natural progression of the effects of the disease. This would support its early use in clinical practice. There is evidence in neonates that the early use of CPAP lowers intubation rates.6 Adults with acute hypercapnoeic respiratory failure also benefit from CPAP.7 Unsurprisingly, this study was not large enough to confidently demonstrate any difference in length of hospital stay. We chose a crossover design to increase the power of the study to detect differences in the primary outcome measure. No child in this study had clinical features which could have been related to CPAP causing an air leak. However, the study was too small to say whether CPAP is safe in children with bronchiolitis.

Capillary CO2 was measured using the same method to collect blood from the heel as was used in a study which has shown that the measurement correlates closely with arterial CO2 in children treated in a paediatric intensive care unit (r2 = 0.98).8 We chose change in capillary CO2 rather than oxygen requirement as this is difficult to measure precisely, especially in mouth breathers, because of the method of delivery by nasal prongs. We did not compare apnoeas because we were unable to store respiratory traces with the monitors we used.

The preferred outcome for the effect of CPAP in bronchiolitis would be the need for invasive ventilatory support. However, very few children with bronchiolitis require invasive ventilation and studies using this as the primary outcome would require a multi-centre trial enrolling many more children. We did not have the resources to embark on such a study. However, although we have not shown clinical improvement, the logical implication is that the physiological improvement may prevent the need for more invasive ventilatory support. Our results could help encourage and inform the design of a multi-centre study to address the clinical questions.

Although the children who were not recruited to the study appeared to have similar illness severity, they were of a younger age. This means that the usefulness of CPAP cannot be extrapolated to very young children. However, although the disorders it is used to treat are different, it is known that CPAP is a safe and useful therapy in neonates.9 The difference in age was only a month in this study, so it may be worth considering CPAP at any age in children with bronchiolitis.

The mode of action of CPAP in bronchiolitis is likely to be multi-factorial. CPAP probably reduces airway resistance, functional residual capacity and gas trapping in hyperinflated lungs.10 11 Ventilation of collapsed areas of lung is improved and underventilated areas recruited.7 In this way, the work of breathing is reduced and ventilation and perfusion matching improved. The effect on apnoeas in this age group with bronchiolitis is not known, although premature infants with apnoea benefit.12

Our study suggests that children with bronchiolitis with capillary Pco2 >6 kPa who do not require immediate invasive ventilatory support should be considered for nasal CPAP. A large multi-centre study would be needed to investigate whether CPAP is of value in reducing the need for invasive ventilation or hospital stay.


Acknowledgements: The authors are grateful to the nursing staff on paediatric HDU and the medical staff at the Royal London Hospital over the last 3 years of this study. We are grateful to Professor Michael Healy for advice about statistical analyses.


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.