CPAP treatment for children with pneumonia in low-resource settings

Pneumonia is the leading cause of mortality in children aged 1–59 months¹ and most of these deaths occur in sub-Saharan Africa and south Asia.¹ In addition to preventive measures, effective treatments targeting high-risk children are needed to lower pneumonia mortality globally.^2–5 In this Comment we aim to briefly review the physiology, current evidence, and evidence gaps for non-invasive continuous positive airway pressure (CPAP) treatment of children with pneumonia in low-resource settings.

CPAP delivers constant positive end expiratory pressure (PEEP) to a child who is spontaneously breathing.⁶ Traditionally, a mask or nasal prongs sealed against the nostrils are connected to a pressure generator and an airflow source.⁷ In conventional CPAP, a mechanical ventilator provides airflow and PEEP.⁷ In bubble CPAP (bCPAP), an oxygen concentrator or cylinder provides airflow, and the depth of expiratory tubing within a fluid reservoir generates PEEP (figure).^7,8 In resource-rich settings, conventional CPAP is used for children with severe respiratory illnesses and bCPAP is used for neonates with respiratory distress syndrome.^6,7 bCPAP has also been used as a neonatal treatment in low-resource settings.⁹

Figure. Bubble CPAP treatment for a child with WHO severe pneumonia.

Reproduced from McCollum et al,¹⁰ by permission of the International Union Against Tuberculosis and Lung Disease. Copyright © The Union. CPAP=continuous positive airway pressure.

Two randomised clinical trials^10,11 from low-resource settings investigated whether CPAP reduces pneumonia mortality outside the neonatal period. In Bangladesh, a trial¹⁰ from an urban tertiary referral hospital intensive care unit recruited children aged younger than 5 years with WHO-defined pneumonia and hypoxaemia. Children were randomly assigned to standard low-flow oxygen (2 L/min), high-flow oxygen (2–12 L/min), or bCPAP.¹⁰ At interim analysis, children had a 75% lower risk of pneumonia mortality after bCPAP, compared with low-flow oxygen (relative risk 0.25, p=0.02). However, the trial was terminated early because of safety concerns, which has raised questions about validity.¹² Additionally, the generalisability of the data for low-resource settings is questionable because the study was done in an intensive care unit. Most children who are admitted to hospital with pneumonia receive care at limited-resource district hospitals. Intensive care resources, including higher nurse-to-patient ratios, might be as important for effective CPAP delivery as the equipment itself.

Another study¹¹ compared conventional CPAP with standard care, among children aged 1–59 months, at two rural district hospitals in Ghana. The trial was a cluster crossover design, eligible children had tachypnoea and respiratory distress, and the primary outcome was all-cause mortality 2 weeks after enrolment. Children who were severely ill were ineligible. No significant differences in all-cause mortality were identified between the CPAP and standard care groups (relative risk 0.67, p=0.11). However, in adjusted secondary analyses, the authors reported a lower odds ratio for mortality among children aged younger than 1 year who were treated with CPAP compared with standard care. Notably, the study population was relatively low-risk; the overall trial mortality was 3.2%, few children had hypoxaemia, or were severely malnourished, and few had HIV or had been exposed to the virus.

Although CPAP would be expected to provide a survival benefit, the evidence available is inconclusive. Before scale-up is considered, evidence gaps must be addressed. Most importantly, it is unclear whether CPAP benefits high-risk children at the district hospital level, where most children receive care for pneumonia. CPAP is unlikely to substantially reduce overall child pneumonia mortality if it is not effective in these patients. The CPAP Improving Mortality for Pneumonia in African Children Trial¹³ aims to address this question by assessing whether bCPAP—compared with low-flow oxygen—improves hospital outcomes among children aged 1–59 months with WHO-defined severe pneumonia and either HIV infection or exposure, severe malnutrition, or hypoxaemia.

Understanding the effect of CPAP on high-risk cases in a setting that is representative of where most children receive care will provide additional evidence to inform policy. Without this information, a CPAP implementation strategy applied to all paediatric patients who are admitted to hospital with pneumonia might quickly overwhelm available resources, leaving health-care providers with the difficult task of deciding which children should be selected to receive CPAP.

Unanswered questions also exist regarding practical feasibility. CPAP equipment, infrastructure, and maintenance costs, as well as uninterrupted power, might not be possible in most hospitals. Technical innovations are likely to be necessary. Research evaluating the economics of CPAP scale-up, treatment, and cost-effectiveness is needed. Whether human resources that are already stretched can effectively incorporate CPAP into existing workloads is also unclear. Another consideration is how to appropriately balance oxygen and CPAP treatments so that treating one child with CPAP doesn’t prevent another child receiving oxygen. One oxygen concentrator provides low-flow oxygen for five to ten children simultaneously but can only serve one child if used for bCPAP.

Few treatments currently in development are as promising as CPAP, and we are optimistic about its future prospects, but key questions must first be answered before CPAP can be broadly implemented for children with pneumonia in low-resource settings.