Abstract
Objective: To investigate and identify factors associated with successful venoarterial extracorporeal membrane oxygenation (VA-ECMO) while initiating, during, and before VA-ECMO weaning.
Design: A single-center retrospective cohort study.
Setting: A 510-bed tertiary hospital.
Patients and participants: We consecutively enrolled 51 patients treated with VA-ECMO at our center between April 2010 and March 2016. Patients who received venovenous ECMO, backup VA-ECMO, and post-cardiotomy VA-ECMO were excluded.
Interventions: VA-ECMO for cardiopulmonary arrest (CPA) or cardiogenic shock. Successful VA-ECMO weaning was defined as survival for more than 24 hours after discontinuation without requiring reintroduction.
Measurements and results: Factors associated with VA-ECMO were collected at initiation, during, and before weaning. These were then compared between the successful and unsuccessful VA-ECMO weaning groups. We included 41 patients after 10 exclusions; 17 (41.5%) and 24 (58.5%) were weaned successfully and unsuccessfully, respectively. Among the factors measured at initiation and prior to weaning, higher blood pressure (BP) and pH and lower serum potassium and creatinine were associated with successful weaning. In addition, lower white blood cell counts at initiation, lower lactate levels, and higher pulse pressure before weaning were also associated with successful weaning. Moreover, among the factors analyzed at initiation, successful weaning was associated with CPA during catheterization, percutaneous coronary intervention (PCI), and a shorter CPA-ECMO time. Among the factors studied before weaning, patients successfully weaned from VA-ECMO had lower Kidney Disease: Improving Global Outcomes (KDIGO) urine output stages and intra-aortic balloon pump (IABP) use.
Conclusion: BP and anaerobic condition-related factors, such as pH or lactate, can be important predictors for successful VA-ECMO weaning. Other clinical factors may also influence VA-ECMO weaning, such as CPA during catheterization, PCI performance, IABP use, the CPA-ECMO time, and renal function markers such as the KDIGO urine output stage and serum creatine and potassium levels.
Masahiro Watanabe, Yoichi Ajiro, Shun Hasegawa, Kazuyuki Hamada, Kei Tsukamoto, Takashi Saito, Ryosuke Furuya, Tatsuji Komatsu, Fumiaki Mori
