Quantifying Oxygenator Performance in Extracorporeal Membrane Oxygenation: Validation and Transfer of the Hamilton Methodology to Long-Term Support

Abstract

Background Maintaining adequate oxygen delivery is critical in extracorporeal membrane oxygenation (ECMO), yet monitoring oxygenator performance remains largely dependent on reactive blood gas analysis (BGA) and subjective visual inspections. While Hamilton et al. developed a methodology to quantify gas exchange efficiency for short-term cardiopulmonary bypass (CPB), its applicability to long-term ECMO support has not been established. This study aimed to adapt and validate this model for real-time monitoring of ECMO oxygenator degradation. Methods In this monocentric, retrospective study, we analyzed 2,441 pre- and post-oxygenator BGA pairs from 164 patients treated with either the Quadrox PLS® (n=66) or A.L.ONE® (n=101) oxygenator. We adapted the Hamilton equation to calculate a corrected fraction of inspired oxygen FiO2 (cFiO2), normalizing performance to a target arterial partial pressure of oxygen (PaO2) of 150 mmHg. Statistical equivalence between measured and predicted oxygen transfer (VO2) was assessed using Two One-Sided Tests (TOST) and linear regression stratified by oxygenator runtime. Results Both oxygenator models demonstrated high correlation between VO2 and cFiO2. Performance degradation over time was successfully quantified through shifting regression slopes. TOST analysis confirmed statistical equivalence (α = 0.05) between the model’s predictions and clinical BGA data, indicating high predictive accuracy across various clinical states. Conclusion The Hamilton methodology can be effectively transferred to the ECMO setting, providing an objective, real-time tool to monitor membrane efficiency. By normalizing for patient-side variables, this mathematical framework allows for the early detection of oxygenator dysfunction, potentially facilitating proactive device exchange and improving patient safety.