Strategies of left ventricular unloading during VA-ECMO support: a network meta-analysis

Background Left ventricle (LV) unloading during VenoArterial ExtraCorporeal Membrane Oxygenation (VA-ECMO) reduces the risk of LV distention, stagnation and pulmonary congestion resulting from the increased afterload. Lacking direct comparisons between unloading strategies we used network meta-analysis to indirectly compare different unloading approaches. Methods A literature research was performed to include all studies on VA-ECMO reporting data on mechanical LV unloading. The pre-specified outcome was in-hospital death. Results Literature search identified 389 studies: 16 were included in the analysis (3930 patients). Two strategies of mechanical LV unloading were compared: afterload reduction (IABP) and preload reduction (Impella pump, right upper pulmonary/trans-septal catheters, LV surgical vents). Any LV unloading strategy was associated with mortality reduction with overall OR = 0.54; 95% CI 0.42–0.70; p < .001. Targeting afterload was associated with reduced mortality (OR = 0.61 95% CI 0.46–0.81; p < .001; I2 = 61%), as targeting preload (OR = 0.34 95% CI 0.21–0.55; p < .001; I2 = 0%). Significant between group difference was observed (p = .04): to further explore this we performed a network meta-analysis. Indirect comparisons between afterload and preload reduction were estimated. Any unloading technique was confirmed better than none but preload targeting resulted better than afterload targeting. Conclusion Any unloading strategy in VA-ECMO patients was associated with lower mortality as compared to no-unloading. Preload reduction strategies resulted superior to afterload reduction.