Article Text

Efficacy and safety of bubble CPAP in neonatal care in low and middle income countries: a systematic review
  1. Simone Martin1,
  2. Trevor Duke2,
  3. Peter Davis3,4
  1. 1Neonatal Intensive Care Unit, Royal Children's Hospital, Parkville, Victoria, Australia
  2. 2Department of Paediatrics, Centre for International Child Health, University of Melbourne MCRI, Royal Children's Hospital, Parkville, Victoria, Australia
  3. 3Department of Neonatal Medicine, Royal Women's Hospital, Parkville, Victoria, Australia
  4. 4Department of Obstetrics and Gynaecology, University of Melbourne, Parkville, Victoria, Australia
  1. Correspondence to Dr Simone Martin, Neonatal Intensive Care Unit, Royal Children's Hospital, 50 Flemington Rd, Parkville, VIC 3052, Australia; simone.langford{at}


Introduction Forty per cent of global child deaths occur in the neonatal period. Low and middle income countries need effective and simple methods to improve hospital-based neonatal care. Bubble continuous positive airway pressure (CPAP) may have a role in improving the quality of respiratory support in hospitals in low and middle income countries.

Aim To examine the evidence for the efficacy and safety of bubble CPAP in neonates with respiratory distress in low and middle income settings.

Method A systematic search (1946–March 2014) was performed of Pubmed, Ovid MEDLINE, Web of Science, Google Scholar and the references of relevant articles. Articles meeting inclusion criteria (CPAP for respiratory distress in infants <28 days of age in hospitals in low and middle income countries) were assessed using Grading of Recommendations, Assessment, Development and Evaluation and Newcastle-Ottawa Quality Assessment Scale methodology. Outcomes included need for mechanical ventilation, complications and mortality.

Results In three studies, the initial use of bubble CPAP compared with oxygen therapy, followed by mechanical ventilation if required, reduced the need for mechanical ventilation by 30%–50%. In another three trials comparing bubble CPAP with ventilator CPAP, mortality and complication rates were similar, while meta-analysis of CPAP failure in these same trials showed a lower failure rate in the bubble CPAP groups (p <0.003).

Conclusions There is evidence that bubble CPAP is safe and reduces the need for mechanical ventilation. Further research into the efficacy of bubble CPAP in low-income and middle-income countries is needed.

  • Neonatology
  • Respiratory

Statistics from

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.

What is already known on this topic

  • Bubble continuous positive airway pressure (CPAP) is safe and effective for the management of newborns with respiratory distress in high-income countries.

What this study adds

  • Bubble CPAP reduces the need for mechanical ventilation in neonates with respiratory distress in tertiary referral hospitals in low and middle income countries.

  • Bubble CPAP provides a safe and effective alternative to ventilator CPAP for respiratory distress in low and middle income countries.

  • Bubble CPAP can be effectively applied by nurses and other health workers in low and middle income countries, and may improve neonatal survival and quality of neonatal care in these settings.


Worldwide, each year, about three million babies die before 28 days of life, representing 40% of global under-5 mortality. Neonatal survival has gained prominence with the recognition that Millennium Development Goal 4 (a two-thirds reduction in childhood mortality from 1990 to 2015) can only be achieved by a reduction in neonatal deaths.1 ,2 The global neonatal mortality rate (NMR) currently stands at 23 per 1000 live births, but there is a wide gradient between and within countries. Lower-income countries report an NMR as high as 52, while upper-income countries report NMRs of as few as 1 per 1000 live births.3 Within countries, the NMR can vary twofold to threefold between geographical, socioeconomic and ethnic groups.4

Substantial reductions in NMRs can be achieved with community-based interventions.5 ,6 However, there is evidence that when NMR falls below 25–30 per 1000 live births across all sectors of community, parallel improvements in hospital-based prenatal and neonatal care are required for further reductions.7 ,8 In such circumstances, interventions that improve care for the three main causes of neonatal deaths: prematurity, sepsis and birth asphyxia (together comprising 85% of neonatal mortality) are of key interest.1 ,3 Broad-based interventions for reducing these common causes of neonatal mortality include early newborn care, neonatal resuscitation, judicious use of antibiotics and intravenous fluids, promotion of breast feeding, prevention of apnoea and safe use of oxygen therapy.5 ,9 Respiratory distress syndrome (RDS) is a common complication of prematurity,10 ,11 while pneumonia is the focus in up to 36% of cases of neonatal sepsis.12 Because respiratory compromise is common to all three of the major causes of neonatal deaths, an effective means of managing respiratory distress could have a major impact. Mechanical ventilation is an expensive and complex respiratory support method requiring a high level of technical skill for use and maintenance at an adequate standard. Many countries do not have the funds, infrastructure or expertise to provide such technology to all patients or to a high standard.13–16 Bubble continuous positive airway pressure (CPAP) is a simpler intervention for respiratory distress, which is becoming cheaper and more accessible in low and middle income countries.

Bubble CPAP

CPAP entails the application of a constant distending pressure to the airways which prevents alveolar collapse and reduces the work of breathing in spontaneously breathing infants with respiratory distress.17 The benefits of CPAP in neonates with severe respiratory distress were first described in 1971.18 In high-income countries, CPAP has reduced the requirement for and duration of mechanical ventilation in neonates, and is increasingly the initial method of respiratory support for premature infants.11 ,19–21 A number of researchers around the world have recommended increased use of CPAP, suggesting the technique has the potential to substantially improve the rates of survival of infants with RDS in low and middle income countries.21–23

‘Bubble CPAP’ is based on a simple system where expiration occurs against a pressure generated by bubbling gas underwater. The airway distending pressure is determined by the depth at which the expiratory limb is placed, that is, 5 cm of water pressure is produced by placing the expiratory tubing of the breathing circuit 5 cm under water. The components of a CPAP circuit are: a gas flow source (oxygen and air), a mechanism to control oxygen concentrations, a humidifier, nasal prongs or mask as the patient interface, connecting tubing and a vessel of water into which the expiratory tubing is submerged.16 ,24 The efficacy of bubble CPAP has been established in high-income countries and it is used widely for the management of infants with respiratory distress from many causes.19 ,25 However, there is limited evidence of its effectiveness in low and middle income countries.

Safety concerns include the risk of pneumothorax, nasal septal damage and risks of excessive oxygen administration.24 ,26 In high-income countries, the safety profile of bubble CPAP appears similar to that of CPAP produced by other means (eg, ventilator-driven CPAP), and the majority of reported complications can be mitigated or prevented by adequate nursing care.16 ,21 ,24

CPAP in low–middle income countries

The key aim of this review is to examine the evidence for safety and efficacy of CPAP (particularly bubble CPAP) as a primary respiratory support to improve survival, compared with standard oxygen therapy, in neonates <28 days old with respiratory distress in low and middle income countries. Secondary aims include examination of the evidence for reduction in need for mechanical ventilation with the use of CPAP, CPAP-associated morbidity and equivalence in safety and efficacy of bubble CPAP compared with ventilator or other forms of CPAP.


A search strategy based on a Population, Intervention, Comparison and Outcome format was employed (see table 1). Ovid MEDLINE, PubMed and Web Of Science were searched using broad search terms: neonat*, newborn, infant, CPAP, ‘continuous distending pressure’, bubble, ‘third world’, ‘developing countries’, ‘low income’, ‘resource limited’, combined with Boolean search terms ‘AND’ and ‘OR’ as appropriate. Searches were performed up to 21 March 2014, with no limits applied to language or year of publication. Countries of interest were defined according to the World Bank classification system which lists countries with a gross national income per capita of <US$12 615 in 2012 as low and middle income or ‘developing’ countries.27 Studies addressing the use of CPAP in older children (>28 days of age), to enable weaning from mechanical ventilation, or focused on use of surfactant with CPAP were excluded. No studies were excluded on the basis of study design. All causes of respiratory distress were considered, although congenital anomalies and structural pathologies were excluded, where ascertainable.

Table 1

PICO search strategy

Titles were screened and if clearly unrelated to the topic, the articles were rejected. The remaining abstracts were read to confirm appropriateness, followed by perusal of the full text and further exclusion of irrelevant papers. The reference lists of included articles were also searched. A review of the first 200 most relevant results using the same search terms entered into Google Scholar was also performed. SM performed the literature search, and all authors made decisions regarding study inclusion and exclusion. Where disagreement arose, the final decision was made by consensus. Papers were evaluated using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) and Newcastle-Ottawa Quality Assessment Scales to determine quality.28 ,29 RevMan5.2 software30 was used for meta-analysis of randomised control trial data.


The search strategy and inclusion and exclusion criteria are presented in a Transparent Reporting of Systemic Reviews and Meta-Analyses flow chart (figure 1), with 14 studies being found through systematic search of databases. Five additional studies were identified by searching the references of included studies and Google Scholar.31–35

Figure 1

Transparent Reporting of Systemic Reviews and Meta-Analyses (PRISMA) flow chart depicting literature search results.

Studies were categorised as either addressing the primary outcome of interest (mortality)14 ,33 ,36 or other outcomes. Several studies addressed secondary outcomes (reduction in need for mechanical ventilation via an endotracheal tube,37–39 and safety and efficacy of bubble CPAP compared with ventilator-driven CPAP35 ,40 ,41), while the largest number were case-series describing practical experience with CPAP, both bubble-based or via other systems.31 ,32 ,34 ,42–48 All studies described the use of CPAP in hospitals, with no studies describing home CPAP. The included studies are summarised in tables 24. Excluded papers are presented in online supplementary table S5 13 49–55

Table 2

Case control and cohort studies

Table 3

Randomised trials 

Table 4

Case series and case reports

Meta-analysis was performed on three randomised control trials found comparing bubble CPAP with ventilator CPAP.35 ,40 ,41 It was not possible to perform meta-analyses on other studies due to the significant variation in methodology and quality between them.

Searching without language limitations revealed three papers written in languages other than English in the standard databases. One was an abstract of a study performed in France, written in French.56 The other two were case series from Brazil in Portuguese and a cohort study from Nicaragua in Italian, of which only the Nicaraguan study was of sufficient quality and relevance to be included.39 ,50

The studies (see tables 24) are grouped according to study design and quality assessment methodology, such that case control and cohort studies are tabulated together and assessed according to the Newcastle Ottawa Quality Assessment Scale, while randomised control trials are assessed together using GRADE criteria. Case series and reports are described according to Newcastle Ottawa measures.

Does the use of CPAP in newborns with respiratory distress reduce mortality compared with standard oxygen therapy?

No randomised or non-randomised studies comparing mortality rates associated with supplemental oxygen therapy with those of CPAP were found.

Does the use of CPAP in newborns have an impact on mortality rates?

Three non-randomised studies examined the relationship between use of CPAP and mortality.14 ,33 Two studies were conducted in South Africa and enrolled very or extremely low birth weight babies. One of these two did not specify the type of CPAP used, and had a case-control study design, constituting low-quality evidence, despite moderate sample size.14 This study undertook multivariate logistic regression analysis and concluded that the use of CPAP was associated with survival. However, it is unclear whether the effect of CPAP on survival was independent of birth weight, and other confounding factors, and thus, there is insufficient evidence to reliably determine an effect on mortality. The second South African study indicating improved survival with CPAP (45% survived compared with 20%) used jet CPAP, and was a modified quasirandomised control trial that only included 21 infants, providing a moderate level of evidence but low power.33 The third trial was another quasirandomised trial comparing use of bubble CPAP with standard nasal prong oxygen therapy in Malawi. Based on the availability of equipment, newborns were allocated to the bubble CPAP group or nasal oxygen. The mortality rate of babies initially allocated to treatment with bubble CPAP was 30%, compared with 47% for those allocated to oxygen alone, a relative risk of 0.64 (95% CI 0.37 to 1.1). The difference in survival rates was not statistically significant.36

Does the use of bubble CPAP reduce the need for mechanical ventilation?

Three studies, including 2002 neonates, investigated whether bubble CPAP reduces the need for mechanical ventilation. Two studies were of moderate to high quality,37 ,38 while the third was of low quality.39 Two studies studied two cohorts before and after the introduction of bubble CPAP in neonatal units in Fiji and Nicaragua, while the other was a multicentre, randomised control trial in South America.37 ,38 ,39 The Fiji and Nicaragua studies included predominately term or late preterm infants, while the study in South America was limited to infants of 800–1500 g birth weight. They found a reduction in mechanical ventilation of 30%–50% with the use of bubble CPAP, with no increase in mortality.

Is bubble CPAP as safe or effective as ventilator-driven CPAP?

Three randomised trials, including a total of 194 neonates, compared safety and efficacy of bubble CPAP with ventilator CPAP.35 ,40 ,41 These included a pilot study and its corresponding completed trial from the same research group in India. There was no overlap in participants. The final Indian trial had a sample size of 114, and reported that the mortality rates were not statistically significantly different between the two groups, although the limited sample size and low mortality rate (9/114, 8%) meant the study was underpowered to detect a difference. A similar trial from Iran had a sample size of 50 and also reported a higher success rate in the bubble CPAP group. Due to some major limitations, including structured allocation, lack of clear data on surfactant exposure and mortality, this study was classified as low quality.35 An abstract examining this question was excluded as it was not detailed enough to perform a quality assessment on the study methodology (see online supplementary table S5).53

A pooled analysis of the three randomised control trials was performed. There was a lower failure rate (need for intubation) on bubble CPAP than ventilator-driven CPAP (OR 0.32, 95% CI 0.16 to 0.67, p <0.003) (see figure 2). Meta-analysis of mortality rates was not possible due to a lack of data clarity. Complication rates were not compared because of differences in CPAP delivery interfaces.

Figure 2

Forest plot showing continuous positive airway pressure (CPAP) failure rates among neonates treated with bubble CPAP as compared with ventilator CPAP.

Uncontrolled observational studies

Several case series have described the use of a variety of CPAP delivery devices among heterogeneous neonatal populations in a diverse range of countries, with and without mechanical ventilation access. The mortality rates across the nine case series ranged from 9.5% to 26.6%. Pooling of the results, where possible, allows review of some basic data. Among 571 neonates treated with CPAP of any kind with the potential for mechanical ventilation access (either in the same hospital or via transfer to nearby referral hospitals), the mortality rate was 83/571 (14.5%).31 ,32 ,34 ,42–44 46–48 Of the 343 neonates for whom CPAP failure data are available, 110 (32.1%) failed CPAP therapy and required mechanical ventilation.31 ,32 ,34 ,42 ,43 ,47 Due to small numbers, it was difficult to determine the effect of birth weight on failure rate and mortality. Only 3 of 155 infants with reliable data (1.9%) sustained pneumothoraces.32 ,34 ,42 ,43

Safety of implementation

Six studies from three countries described implementation of CPAP in settings with few or no intensive care nursing staff.31 ,33 ,36 ,38 ,45 ,48 Despite the lack of previous CPAP experience, in all cases (total 252 neonates) there were no major adverse events directly attributable to poor management of CPAP administration. This suggests that nursing staff in low and middle income countries can be trained to implement CPAP safely. Only one of these studies was not in an urban tertiary or referral hospital.31

Application in subgroups of patients

Three case series allowed insight into the application of CPAP to different subgroups. Although only small numbers and heterogeneous contexts, the results were suggestive of better outcomes in neonates with birth weight (BW) >1000 g (CPAP failure rate 24/106 (22.6%)) than neonates <1000 g (CPAP failure: 11/24 (45.8%)).31 ,32 ,42 Outcomes also appeared to be better in those with mild to moderate respiratory distress (CPAP failure: 16/109 (14.5%)) compared to neonates with more severe disease (CPAP failure: 25/41 (60.9%)).32 ,42 ,43

Barriers to implementation

Local cultural and institution-based barriers are important in the introduction of a novel intervention. Two research groups working in Malawi noted local concerns that negatively impacted the implementation of bubble CPAP.45 ,48 In one case, community beliefs that oxygen use was associated with poor outcome resulted in parental reluctance to enrol in a trial of bubble CPAP, while in another case, staff shortages, particularly of those familiar with the equipment resulted in minimal use beyond the initial trial period.45 ,48


Currently, WHO provides recommendations about the use of oxygen therapy in neonatal care.57 These guidelines recommend the use of nasal prongs as a mechanism of delivery, and a flow rate of 0.5–1 L/min, increasing to 2 L/min in severe respiratory distress to achieve an oxygen saturation of >90%. Studies in high-income countries in the 1970s suggested a survival advantage for preterm newborns with severe respiratory distress managed with CPAP compared with standard oxygen therapy (relative risk 0.52, 95% CI 0.32 to 0.87).21 Since then, the use of CPAP has increased and become a standard of care for preterm newborns with moderate to severe respiratory distress, in tertiary and many non-tertiary hospitals. This review, including studies mostly from tertiary or referral hospitals in low and middle income countries, suggests that bubble CPAP also has a role in these environments and is associated with a reduced need for mechanical ventilation without an increase in mortality. Only three studies showed an association between CPAP use and survival, none of which were conclusive.14 ,33 ,36 Any real mortality effect may be influenced by other accessible resources including rescue therapies, such as mechanical ventilation and surfactant.14

Studies of the role of bubble CPAP in developing countries are limited by study design and quality, and the context in which studies were performed. We could find no good comparative studies evaluating whether bubble CPAP was associated with lower mortality for neonates with respiratory distress in low and middle income countries in non-tertiary settings, or where the only alternative respiratory support is supplemental oxygen therapy.

Various commercial bubble CPAP circuits are available, and there are ‘home-made’ versions described in the literature. Costs range from US$10 for simple devices to US$6000 for commercial units6 ,13 ,15 ,45 ,58 ,59 Due to its technical simplicity, bubble CPAP is cheaper, less invasive, more accessible and requires less technical skill for application than current commercial ventilators.6 ,22 Bubble CPAP is also compatible with kangaroo mother care which is a low-cost method of maintaining neonatal thermoregulation.5 ,6 ,24

It is likely that CPAP is in more widespread use in developing countries than is reflected in the literature. This may be due to the fact that where clinical resources are few, resources to devote to audit or research are also limited. Two of the published studies were performed in the context of introducing a new device where the drive to publish for commercial reasons may introduce an increased risk of publication bias.36 ,45 The systematic collection of the outcomes of infants managed with bubble CPAP in a database would be useful in determining the role of this intervention in low and middle income countries.

Some useful information can be gleaned about the acceptability and feasibility of implementing bubble CPAP from the published literature. Evidence from a limited number of sites suggests that bubble CPAP is a therapy that can be applied by nursing staff in a tertiary referral-level hospital. This may be of use to researchers who wish to conduct high-quality trials.


In countries faced with the challenge of improving neonatal care, there is a need for simple, cost-effective methods of respiratory support. Bubble CPAP is a promising intervention, which may reduce the need for high-cost mechanical ventilators, and can be applied by appropriately trained nursing staff. However, there is a need for more research into the impact of bubble CPAP compared with standard oxygen therapy on neonatal mortality and into effective implementation methods, especially in non-tertiary hospitals.


The authors thank Ms Poh Chua for her advice and assistance with the literature search, and Dr Roberto Chiletti for translating one of the full-text articles from Italian to English.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors SM generated the original concept, performed the literature search under the guidance of Poh Chua (see acknowledgements), put together the tables and figures and wrote the first draft.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles