Respiratory support for children with severe pneumonia and hypoxaemia in a developing country: a randomized trial of bubble CPAP, high flow nasal cannula therapy and standard flow oxygen

© 2104 Dr. Mohammod Jobayer ChistiBACKGROUND: Mortality rate from very severe pneumonia in many hospitals is more than 10% even with appropriate antibiotics, standard low flow (LF) oxygen by nasal prongs as recommended by the World Health Organization (WHO) and good supportive care. Bubble CPAP (BCPAP) and humidified high flow nasal cannula (HFNC) oxygen are increasingly used in neonatal care for reducing respiratory distress and hypoxemia. However, there were no published controlled trials (RCT) of the use of BCPAP in children beyond the newborn period with severe pneumonia and hypoxemia in developing countries. AIMS: Our aim was to evaluate the efficacy of BCPAP compared to LF or LF and HFNC therapy in managing children with severe pneumonia and hypoxemia. REGISTRATION: Clinicaltrials.gov identifier: NCT01396759 METHODS: We conducted an open RCT in the Dhaka hospital of the International Centre for Diarrheal Disease Research, Bangladesh (icddr,b) between August 2011 and July 2013. Children under five years of age who had severe pneumonia and hypoxemia as defined by WHO were randomly allocated to BCPAP, LF or HFNC oxygen therapies. The primary outcome was to determine if there were differences in treatment failure, based on objective clinical criteria. Mortality was our secondary outcome. We established a DSMB which reviewed adverse events and reviewed the planned interim analyses. RESULTS: During the study period 376 children were admitted with study criteria and 225 children were enrolled. There were two interim analyses. Two years into the study 79 children had been allocated to BCPAP, 79 to HFNC and 67 to the LF arms. Baseline and other characteristics were comparable between the 3 intervention groups, and all patients were available for analysis. At the second interim analysis the DSMB recommended stopping the study because of a significant difference in mortality. A total of 31 (14%) children fulfilled the criteria for treatment failure: among them 5 (6%) in BCPAP, 16 (24%) in LF and 10 (13%) in HFNC therapy. There were 23 deaths (10%): 3 (4%) in BCPAP therapy, 10 (15%) in LF therapy, 10 (13%) in HFNC therapy. Children who received BCPAP therapy had significantly lower treatment failure compared to those who received LF therapy (6% vs. 24%; p < 0.003) or LF and HFNC therapy (6% vs. 18%; p = 0.017). Mortality was significantly less in children who received BCPAP therapy compared to those who received LF therapy (4% vs. 15%; p = 0.040) or LF and HFNC therapy (4% vs. 13%; p = 0.035). In multiple logistic regression analysis, after adjusting for potential confounders, children who received BCPAP therapy had significantly less risk (OR = 0.29, 95% confidence interval = 0.09 – 0.88, p = 0.028) of developing treatment failure compared to other intervention groups and BCPAP therapy remained an independent predictor of survival (OR = 0.20, 95% confidence interval = 0.05 – 0.86, p = 0.031) in the management of children with severe pneumonia and hypoxemia. CONCLUSION: In Bangladeshi children under five with severe pneumonia and hypoxemia, compared to standard LF oxygen therapy, BCPAP oxygen therapy was associated with a significant lower probability of treatment failure and mortality. BCPAP should be considered an important option for treating children under five with severe pneumonia and hypoxemia