Purpose: Several studies have reported adopting prone positioning (PP) in non-intubated patients with COVID-19-related hypoxaemic respiratory failure. This systematic review and meta-analysis evaluated the impact of PP on oxygenation and clinical outcomes.Methods: We searched PubMed, Embase and the COVID-19 living systematic review from December 1, 2019 to July 23, 2020. We included studies that reported using PP in hypoxaemic, non-intubated adult patients with COVID-19. Primary outcome measureed was the weighted mean difference (MD) in oxygenation parameters (PaO2/FiO2, PaO2 or SpO2) pre and post-PP. Results: Fifteen single arm observational studies reporting PP in 449 patients were included. Substantial heterogeneity was noted in terms of, location within hospital where PP was instituted, respiratory supports during PP, and frequency and duration of PP. Significant improvement in oxygenation was reported post-PP: PaO2/FiO2 (MD 37.6, 95% CI 18.8-56.5); PaO2 (MD 30.4 mmHg, 95% CI 10.9 to 49.9); and SpO2 (MD 5.8%, 95% CI 3.7 to 7.9). Patients with a pre-PP PaO2/FiO2 ≤150 experienced greater oxygenation improvements compared with those with a pre-PP PaO2/FiO2 >150 (MD 40.5, 95% CI -3.5 to 84.6) vs. 37, 95% CI 17.1 to 56.9). Respiratory rate decreased post-PP (MD -2.9, 95% CI -5.4 to -0.4). Overall intubation and mortality rates were 21% (90/426) and 26% (101/390) respectively. There were no major adverse events reported. Conclusions: Despite the significant variability in frequency and duration of PP and respiratory supports applied, PP was associated with improvements in oxygenation parameters without any reported serious adverse events. The results are limited by lack of control arm and adjustment for confounders. Clinical trials are required to determine the effect of awake PP on patient-centred outcomes.Systematic review registration: Registration/protocol in PROSPERO (CRD42020194080).