Implementation of bubble continuous positive airway pressure for children with severe pneumonia and hypoxemia in intensive care unit of Dhaka Hospital, Bangladesh-Effect on pneumonia mortality

Background: After the completion of a randomized trial at Dhaka Hospital in 2013, bubble continuous positive airway pressure (BCPAP) oxygen therapy was incorporated as the part of the standard treatment for children with severe pneumonia with hypoxemia in an intensive care unit at Dhaka Hospital in August 2013 instead of World Health Organization (WHO) standard low flow oxygen therapy.

Objective: To understand the long-term effectiveness of the introduction of bCPAP oxygen therapy by comparing pneumonia mortality in the post-trial period (August 2013 to December 2017) with the pre-trial (February 2009 to July 2011) and trial periods (August 2011 to July 2013).

Methods: It was a retrospective analysis of prospectively collected hospital data of all admissions. Mortality rates of all children with WHO-defined pneumonia, and the subset of children with severe pneumonia and hypoxemia (oxygen saturation <90%) were evaluated.

Results: The analysis covered 10,107 children with pneumonia: 2523 in the pre-trial (414 with severe pneumonia and hypoxemia; none of them received bCPAP), 2959 during the trial (376 with severe pneumonia and hypoxemia; 79 received bCPAP), and 4625 in the post-trial period (1208 with severe pneumonia and hypoxemia; 1125 had bCPAP). The risk of death from pneumonia in the post-trial period was lower than in pre-trial (adjusted risk ratio [RR] = 0.73, 95% confidence interval [CI] = 0.58-0.92; p = 0.007), among children with severe pneumonia and hypoxemia, the risk of death was lower in the post-trial period than in the pre-trial (adjusted RR = 0.46, 95% CI = 0.37-0.58, p < 0.001), and the trial period (adjusted RR = 0.70, 95% CI = 0.51-0.95; p = 0.023).

Conclusion: After the introduction of bCPAP oxygen therapy as part of the routine management of severe pneumonia and hypoxemia in the ICU of the Dhaka hospital, we observed significantly lower mortality, even after accounting for measurable confounding.