What is the burden of respiratory distress syndrome (RDS) in the developing world?

In 2010, an estimated 3.1 million newborns, mainly in low-income countries, died in the first month of life.2 Complications of premature birth are now the second leading cause of childhood mortality worldwide, with respiratory distress syndrome (RDS) the single most important contributor to these deaths.2 Inexpensive and easy to implement therapies that effectively treat RDS are, therefore, urgently needed in the developing world.

What is the strength of evidence for the use of CPAP to treat RDS?

Three separate meta-analyses of trials conducted primarily in developed countries compare a policy of delivery room CPAP versus intubation for preterm infants.3–5 Together, these conclude that early CPAP results in a small but statistically significant reduction in the composite outcome of bronchopulmonary dysplasia and/or death.3–5 Additionally, these trials indicate that CPAP reduces the need for mechanical ventilation and shortens the duration of supplemental oxygen exposure.4 ,6

Martin et al demonstrate clear benefit with early use of CPAP in developing countries. In a multicentre randomised controlled trial (RCT), bubble CPAP compared to oxygen therapy with or without mechanical ventilation reduced the need for invasive support by approximately 40% (RR 0.59; 95% CI 0.43 to 0.81) in very low birthweight infants. A large observational study (n=1152) reported a 50% decrease in the need for mechanical ventilation in a tertiary unit after CPAP became available. Finally, three RCTs compared bubble to ventilator-generated CPAP and found that fewer infants treated with bubble CPAP failed non-invasive therapy. As ventilators are often not available in resource-limited settings, these are particularly relevant findings.7

While there are concerns for pneumothoraces with the use of CPAP, the risk is likely overestimated.8 Provider education can decrease the risk of pneumothorax and, when present, need for urgent treatment of pneumothoraces is rare.8

Are there relevant differences between studies conducted in developed versus developing countries?

Potential differences that may impact the comparability of studies conducted in developed and developing countries include the underlying pathophysiology of the disease, the characteristics of evaluated subjects, and the estimated efficacy and safety of the intervention. While differences in perinatal care may influence the epidemiology of RDS in resource-limited settings, it is unlikely that there are fundamental differences in the pathophysiology or patient characteristics. The majority of evaluated infants were low or very low birthweight infants who required moderate or higher amounts of supplemental oxygen by nasal cannula or head box. Most were cared for in tertiary neonatal intensive care units where mechanical ventilation was available. In both settings, failure of non-invasive support and need for mechanical ventilation are reduced with CPAP.

The safety profile of CPAP also appears to be similar. While Martin et al suggest that further studies are needed to determine the effect of CPAP on mortality in resource-limited settings, we posit that such studies are unnecessary. We agree that the studies conducted in developing countries were underpowered to detect differences in mortality. The available data, however, provide no reasons to expect that CPAP will increase mortality. In a multicentre RCT comparing early CPAP to oxygen therapy with or without mechanical ventilation in 12 South American centres, the risk of mortality, albeit not significantly, was lower in the CPAP-treated group (RR 0.72; 95% CI 0.41 to 1.25). In a Cochrane review, CPAP use was associated with a significant reduction in mortality (RR 0.52; 95% CI 0.32 to 0.87) and none of the included trials raised concern for increased mortality risk.6 Although the majority of the studies included in the Cochrane were conducted over 30 years ago in developed countries, no significant evidence to date suggests that CPAP use will increase mortality.

Is the use of CPAP feasible in resource-limited settings?

One important benefit of bubble CPAP is its low cost. As Martin et al highlight, simple CPAP set-ups can cost as little as US$10. Another important consideration is the impact that local beliefs and practices can have on the successful introduction of novel medical therapies. Martin et al note that in two studies, implementation of CPAP was adversely affected by both parental concerns regarding the effects of supplemental oxygen and provider unfamiliarity with the CPAP equipment. We are reassured, however, by several reviewed studies conducted in settings with no prior CPAP experience. In those, no major adverse events associated with poor CPAP management were reported. In the situation of CPAP for RDS, we argue that continuing to view the therapy as ‘experimental’ will only further engender these cultural barriers. Efforts should now, instead, turn to education and training.

To conclude, we must consider the potential cost of delaying implementation of safe and effective therapies. A parallel duplication of trials in the same disease illustrates this risk. Since the 1994 National Institutes of Health consensus statement supporting the use of antenatal corticosteroids, four RCTs, including over 600 subjects, evaluated steroid efficacy in low and middle-income countries.7 The treatment effect for prevention of mortality estimated from these four studies was greater than the treatment effect estimated from the studies conducted in developed countries, and is similar to that from the Cochrane review of CPAP.6 ,7 During the time those studies were ongoing, an estimated 500 000 neonatal deaths per year could have been prevented by widespread use of antenatal corticosteroids.7 It would be a tragedy to allow this fate to befall more infants by denying them access to CPAP while further studies are performed. In a commentary on the use of CPAP for RDS published in 1973, Chernick, in light of the already accumulating efficacy data, suggested that many more controlled studies ‘would be foolish’.9 We agree.