Abstract
BACKGROUND: In refractory respiratory failure (RF), extracorporeal membrane oxygenation (ECMO) is a salvage therapy that seeks to reduce lung injury induced by mechanical ventilation. The parameters of optimal mechanical ventilation in children during ECMO are not known. Pulmonary ventilatory management during this therapy may impact mortality. The objective of this study was to evaluate the association between ventilatory parameters in children during ECMO therapy and in-hospital mortality. METHODS: A systematic search of PubMed/MEDLINE, Embase, Cochrane, and Google Scholar from January 2013 until May 2022 (PROSPERO 450744), including studies in children with ECMO-supported RF assessing mechanical ventilation parameters, was conducted. Risk of bias was assessed using the Newcastle-Ottawa scale; heterogeneity, with absence <25% and high >75%, was assessed using I(2). Sensitivity and subgroup analyses using the Mantel-Haenszel random-effects model were performed to explore the impact of methodological quality on effect size. RESULTS: Six studies were included. The median age was 3.4 years (IQR: 3.2-4.2). Survival in the 28-day studies was 69%. Mechanical ventilation parameters associated with higher mortality were a very low tidal volume ventilation (<4 ml/kg; OR: 4.70; 95% CI: 2.91-7.59; p < 0.01; I(2): 38%), high plateau pressure (mean Dif: -0.70 95% CI: -0.18, -0.22; p < 0.01), and high driving pressure (mean Dif: -0.96 95% CI: -1.83, -0.09: p = 0.03). The inspired fraction of oxygen (p = 0.09) and end-expiratory pressure (p = 0.69) were not associated with higher mortality. Patients who survived had less multiple organ failure (p < 0.01). CONCLUSION: The mechanical ventilation variables associated with higher mortality in children with ECMO-supported respiratory failure are high plateau pressures, high driving pressure and very low tidal volume ventilation. No association between mortality and other parameters of the mechanical ventilator, such as the inspired fraction of oxygen or end-expiratory pressure, was found. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023450744, PROSPERO 2023 (CRD42023450744).