Key points are not available for this paper at this time.
Refractory hypoxemia during veno-venous (V-V) extracorporeal membrane oxygenation (ECMO) may require an additional cannula (VV-V ECMO) to improve oxygenation. This intervention includes risk of recirculation and other various adverse events (AEs) such as injury to the lung, cannula malpositioning, bleeding, circuit or cannula thrombosis requiring intervention (i.e., clot), or cerebral injury. During the study period, 23 of 142 V-V ECMO patients were converted to VV-V utilizing two separate cannulas for bi-caval drainage with an additional upper extremity cannula placed for return. Of those, 21 had COVID-19. In the first 24 h after conversion, ECMO flow rates were higher (5.96 vs 5.24 L/min, p = .002) with no significant change in pump speed (3764 vs 3630 revolutions per minute RPMs, p = .42). Arterial oxygenation (PaO 2 ) increased (87 vs 64 mmHg, p < .0001) with comparable pre-oxygenator venous saturation (61 vs 53.3, p = .12). By day 5, flows were similar to pre-conversion values at lower pump speed but with improved PaO 2 . Unadjusted survival was similar in those converted to VV-V ECMO compared to V-V ECMO alone (70% 16/23 vs 66.4% 79/119, p = .77). In a mixed effect regression model, any incidence of AEs, demonstrated a negative impact on PaO 2 in the first 48 h but not at day 5. VV-V ECMO improved oxygenation with increasing flows without a significant difference in AEs or pump speed. AEs transiently impacted oxygenation. VV-V ECMO is effective and feasible strategy for refractory hypoxemia on VV-ECMO allowing for higher flow rate and unchanged pump speed.
Dave et al. (Fri,) studied this question.