In this review update, we included four studies (479 participants) assessing the efficacy of non-invasive delivery methods for the treatment of LRTI in children and 14 observational studies assessing the clinical sign indicators of hypoxaemia in children with LRTIs.


Three RCTs (399 participants) compared the effectiveness of nasal prongs or nasal cannula with nasopharyngeal catheter; one non- RCT (80 participants) compared head box, facemask, nasopharyngeal catheter and nasal cannula. The nasopharyngeal catheter was the control group. Treatment failure was defined as number of children failing to achieve adequate arterial oxygen saturation. All included studies had a high risk of bias because of allocation methods and lack of blinded outcome assessment. For nasal prongs versus nasopharyngeal catheter, the pooled effect estimate for RCTs showed a worrying trend towards no difference between the groups (two RCTs; 239 participants; RR 0.93, 95% CI 0.36 to 2.38). Similar results were shown in the one non-RCT (RR 1.0, 95% CI 0.44 to 2.27). The overall quality of this evidence is very low. Nasal obstruction due to severe mucus production was different between treatment groups (three RCTs, 338 participants; RR 0.20, 95% CI 0.09 to 0.44; I2 statistic = 0%). The quality of this evidence is low.


The use of a facemask showed a statistically significant lower risk of failure to achieve arterial oxygen > 60mmHg than the nasopharyngeal catheter (one non-RCT; 80 participants; odds ratio (OR) 0.20, 95% CI 0.05 to 0.88). The use of a head box showed a non-statistically significant trend towards a reduced risk of treatment failure compared to the nasopharyngeal catheter (one non-RCT; OR 0.40, 95% CI 0.13 to 1.12). The quality of this evidence is very low.


To determine the presence of hypoxaemia in children presenting with LRTI, we assessed the sensitivity and specificity of nine clinical signs reported by the included observational studies and used this information to calculate likelihood ratios. The results showed that

there is no single clinical sign or symptom that accurately identifies hypoxaemia.


Authors’ conclusions

It appears that oxygen therapy given early in the course of pneumonia via nasal prongs at a flow rate of 1 to 2 L/min does not prevent children with severe pneumonia from developing hypoxaemia. However, the applicability of this evidence is limited as it comes from a small pilot trial.


Nasal prongs and nasopharyngeal catheter are similar in effectiveness when used for children with LRTI. Nasal prongs are associated with fewer nasal obstruction problems. The use of a face mask and head box has been poorly studied and it is not superior to a nasopharyngeal catheter in terms of effectiveness or safety in children with LRTI. Studies assessing the effectiveness of oxygen therapy and oxygen delivery methods in children with different baseline risks are needed. There is no single clinical sign or symptom that accurately identifies hypoxaemia in children with LRTI. The summary of results presented here can help clinicians to identify children with more severe conditions.


This review is limited by the small number of trials assessing oxygen therapy and oxygen delivery methods as part of LRTI treatment. There is insufficient evidence to determine which non-invasive delivery methods should be used in children with LRTI and low levels of oxygen in their blood.


Please Sign in (or Register) to view further.