Early weaning from oxygen therapy in African children with severe pneumonia.

Background: In Africa, severe pneumonia remains the major cause of paediatric hospitalisation, resulting in high requirements for oxygen therapy. Adequate supplies of oxygen are key challenges for many low-resource hospitals. The World Health Organization manual for oxygen therapy advises 2-3 days of oxygen therapy for pneumonia and recommends against early weaning, even in the absence of hypoxaemia. Few data support this recommendation. We describe the oxygen use and timing of weaning in the COAST trial of oxygen therapy (ISRCTN15622505).

Methods: Children aged 28 days to 12 years presenting to 6 hospitals in Uganda and Kenya with severe pneumonia and hypoxaemia (saturations < 92% on pulse oximetry (SpO₂) were eligible for the trial. Children in two strata (a) severe hypoxaemia (SpO₂ < 80%) and (b) moderate hypoxaemia (SpO₂ 80-91%) were allocated to receive high flow nasal therapy (HFNT), low flow oxygen delivery (LFO) or control (no immediate oxygen (moderate hypoxaemia stratum only)). Children were closely monitored over 48 h by pulse oximetry and weaned off oxygen once SpO₂ > 92%. We describe the oxygen use and proportion requiring respiratory support over time by intervention strategy.

Results: Of the 1842 children enroled the majority, 1454 (79%) had moderate hypoxaemia. In this stratum, by 2 and 8 h, 148 (41%) and 200/360 (55.6%) in the LFO arm had been weaned; in the HFNT arm, 213/362 (59%) were receiving respiratory support at 2 h in room alone, and by 8 h, 164/362 (45%) had been weaned. At 48 h, in the respective strata, 77-80% and 53-63% still had respiratory distress but without hypoxaemia and were thus not receiving oxygen. Median oxygen use at 48 h in the moderate hypoxaemia group was highest in LFO am 480L (IQR 236.2, 2132.2) compared to 113.4 L (IQR 0.0, 1453.9) in the HFNT and 0 L (IQR 0.0) in the control arms. Children requiring oxygen beyond 48 h, 17/33 (51.1%) and 9/46 (19.5%) in the respective strata, had additional cardiac conditions.

Conclusions: Closely monitoring SpO₂ resulted in early weaning and reduced the use of and exposure to oxygen. Where oxygen supplies are at a premium, this approach may improve equitable access for children with severe pneumonia.