Abstract
BACKGROUND: Prone positioning improves survival in patients with acute respiratory distress syndrome (ARDS) by reducing ventilator-induced lung injury and enhancing ventilation-perfusion matching. Whether these physiological benefits translate to patients supported with veno-venous extracorporeal membrane oxygenation (VV-ECMO) remains uncertain. METHODS: This study compared the effects of prone versus supine positioning on the wet-to-dry lung weight ratio, gas exchange, respiratory mechanics, electrical impedance tomography, histopathology, microbiology, hemodynamics, and extracorporeal circuit function in pigs with severe ARDS undergoing VV-ECMO. Pigs with severe ARDS were placed on VV-ECMO according to EOLIA criteria and then randomized to either prone or supine positioning for 48 h, while receiving an ultraprotective ventilation strategy. Futility analyses were performed at half of the planned sample size using conditional power calculations, with early termination criteria set at < 10% probability of achieving statistical significance. RESULTS: Eight pigs with severe ARDS were randomized after VV-ECMO initiation. Prone positioning, compared with the supine position, resulted in a similar wet-to-dry lung weight ratio (6.77 [6.13-8.17] vs. 6.70 [6.22-8.48]; p = 0.89), meeting the futility threshold. Compared with the supine position, prone positioning significantly reduced the proportion of ventilation in non-dependent lung regions (51 [47-61.25]% vs. 86 [68.50-90]%; p = 0.02), thereby indicating a redistribution of ventilation toward dependent areas. Consistent with this shift in ventilation distribution, prone positioning redistributed histopathological lung injury, with relative preservation of non-dependent regions and greater damage in dependent zones, but without a net global decrease. Shunt fraction approached 100% in both groups, with no significant differences. Pulmonary CO₂ elimination was 8.45 (1.45-26.13) mL/min in the prone group and 0 (0-0.70) mL/min in the supine group (p = 0.23). Lung compliance showed no intergroup differences (14.67 [12.75-19.91] vs. 16.68 [15.51-19.47] mL/cmH₂O; p = 0.88), with similarly elevated end-inspiratory transpulmonary pressures. No significant differences were observed in systemic or pulmonary hemodynamics, nor in VV-ECMO circuit function. CONCLUSIONS: Prone positioning did not decrease the overall severity of lung injury. Rather, it shifted the distribution of damage, with greater involvement of dependent regions and relative preservation of non-dependent areas.