Summary of the literature
| Study | Setting | Study design | Population | Intervention | Control | Outcome(s) | Result per outcome (intervention vs control) | Risk of bias per outcome# |
| Cunningham and McMurray, 2012 [15] | Scotland, paediatric ward | Observational cohort | <18 months with bronchiolitis | NA | Usual care (n=68) | Time to stable SpO2 above 90% and 94% | 22 h difference | NA |
| Schuh et al., 2014 [16] | Canada, tertiary care ED | Double-blind RCT | <12 months with bronchiolitis and SpO2 ≥88% | Altered SpO2 monitor by +3% (n=105) | Normal SpO2 monitor (n=108) | Admission rates | 25% vs 41%; p=0.005 | Low |
| Unscheduled visits | 14.3% vs 21.3%; p=0.18 | Low | ||||||
| Cunningham et al., 2015 [17] | UK, multicentre, paediatric wards | Double-blind RCT | <12 months with bronchiolitis | 90% threshold (n=307) | 94% threshold (n=308) | Length of disease | 1.0 day shorter, 95% CI −1–2 days | Low |
| Return to feeding | 2.7 days shorter, 95% CI −0.3–7 days; HR 1.22 (95% CI 1.04–1.44); p=0.015 | Low | ||||||
| Return to normal health | 1.0 day shorter, 95% CI 0–3 days; HR 1.19 (95% CI 1.01–1.41); p=0.043 | Low | ||||||
| Length of stay | 40.9 h vs 50.9 h; HR 1.28 (95% CI 1.09–1.50); p=0.003 | Low | ||||||
| Readmissions/re-attendance | 12/307 vs 26/308 (reported as nonsignificant, no p-value) | Low | ||||||
| Cost-effectiveness | £290, 95% CI −£657–£78 | Low | ||||||
| Peters et al., 2018 [18] | UK, multicentre, PICUs | Pilot RCT, open-label | <16 years with critical illness | 88–92% oxygenation target (n=53) | >94% oxygenation target (n=54) | Mortality | 7.4% vs 7.5%; relative risk 0.98; 95% CI 0.26–3.72 | Some |
| Length of stay | 1.0 day shorter; 95% CI −0.8–2.9 days; p=0.29 | Some | ||||||
| Patel et al., 2019 [19] | USA, EDs | Open-label RCT | 2–18 years with asthma exacerbation | Titrated oxygen, only if SpO2 <92%, during nebulisation (n=47) | High concentration oxygen, 100% 4 L·min−1, during nebulisation (n=49) | % of patients with PtcCO2 rise of >4 mmHg at 60 min Asthma score at 60 min | 10.6% vs 40.8%; p=0.001 Only reported in figure, 3.5 vs 4.5; p=0.0001 | Some High |
| Van Hasselt et al., 2020 [20] | UK, multicentre, EDs | Observational cohort | 6 weeks–12 months with bronchiolitis | Centres with 90% threshold (n=162) | Centres with 92% threshold (n=158) | SpO2 as reason for admission | 27% vs 37%; p=0.05 | Serious |
| Length of stay | 41 h vs 59 h; p=0.0074 | Serious | ||||||
| Maitland et al., 2021¶ [21] | Uganda and Kenya, multicentre, paediatric wards | Open-label RCT | <12 years with pneumonia and hypoxaemia (80–92%) | 80% threshold (n=727) | 92% threshold (n=729) | Mortality at 48 h | 1.4% vs 2.5%; p=not reported | Low |
| Mortality at 28 days | 3.9% vs 4.1%; p=not reported | Low | ||||||
| Treatment failure | 4.6% vs 2.3%; p=not reported | Low | ||||||
| Length of stay | 0.62 days shorter; 95% CI 0.53–1.59 days | High | ||||||
| Neurocognitive sequelae at 28 days | 2.3% vs 2.9%; p=not reported | Some |
NA: not applicable; SpO2: peripheral oxygen saturation; ED: emergency department; RCT: randomised controlled trial; HR: hazard ratio; PICU: paediatric intensive care unit; PtcCO2: transcutaneous carbon dioxide. #: risk of bias assessed as low, some or high by the Risk of Bias 2 tool, or as low, moderate, serious or critical by the Risk Of Bias In Non-Randomized Studies – of Interventions tool; ¶: the trial was stopped early when a local doctor started multiple court cases to stop the trial owing to safety concerns; although monitoring and ethics committees saw no safety issues in the trial and all court cases were won, inclusions slowed to a halt, which made the trial unfeasible (results of all included patients up to that point were analysed).