CELL FREE DNA-BASED PREDICTION OF PROGNOSIS FOR PATIENTS ON VENO-ARTERIAL EXTRACORPOREAL MEMBRANE OXYGENATION

Introduction : Elevated cell-free DNA (cfDNA) was observed in patients receiving venoarterial (VA) extracorporeal membrane oxygenation (ECMO), but the clinical relevance of cfDNA is still not clear. We aimed to establish a predictive model based on the cfDNA to predict the prognosis for patients on ECMO, and reveal the values of cfDNA for complications of limb ischemia and bleeding/thromboembolic events. Methods : Single-center, retrospective evaluation of patients with ECMO support from 2018 through 2023. The derivation cohort included 133 adults diagnosed with cardiogenic shock who received VA-ECMO for circulatory support. We developed three independent features and combined them with a logistic model to predict mortality. Predictive performance was assessed through Bootstrap analysis and validated by another cohort of 27 patients. The values of cfDNA for complications were analyzed by restricted cubic spline analysis, receiver-operating characteristic curves and multivariate regression analyses. Results : A total of 133 adults who underwent VA-ECMO for refractory cardiogenic shock were entered into the derivation cohort. The logistic model, consisted of cfDNA, the worst mean arterial pressure (MAP) before ECMO and the worst lactate within 24 h of VA-ECMO implantation was predictive and performed similarly for validation cohorts (area under the receiver operating characteristic curve: 0.768 vs. 0.747). Restricted cubic spline analysis revealed a positive linear relationship for the risk of limb ischemia (linear, P = 0.006; area under the receiver operating characteristic curve of 0.75 [95% CI, 0.656-0.848]), a U-shaped trend for bleeding events (nonlinear, P = 0.214), and a negative trend for thrombotic events (linear, P = 0.552). Conclusions: In addition to MAP and lactate levels, elevated cfDNA levels within 48 h of ECMO support were highly associated with mortality for patients. Additionally, cfDNA is predictive of limb ischemia.