Carbon dioxide elimination as a guide to venoarterial extracorporeal membrane oxygenation weaning: a prospective observational study.

Abstract

Background: Despite the increasing use of venoarterial extracorporeal membrane oxygenation (VA-ECMO) in cardiogenic shock (CS), reliable predictors of successful weaning remain poorly defined. This study investigated the role of carbon dioxide elimination parameters in VA-ECMO weaning.

Methods and results: To assess the potential role of end-tidal carbon dioxide (EtCO₂) in predicting successful VA-ECMO weaning, we conducted a prospective observational study at Anzhen Hospital between January 2023 and December 2024. The primary endpoint was the predictive performance of EtCO₂ and VCO₂NL ratio for successful VA-ECMO weaning. Real-time EtCO₂ monitoring was performed using infrared capnography in mechanically ventilated patients, and the ratio of native lung carbon dioxide elimination (VCO₂NL) to total carbon dioxide elimination (VCO₂TOT) was calculated as the VCO₂NL ratio. Dynamic changes in these parameters were analysed in relation to weaning outcomes. Among 294 patients receiving VA-ECMO for refractory CS during the study period, 91 were included, yielding 562 data points. Both EtCO₂ (odds ratio [OR] = 1.26, 95% confidence interval [CI] 1.17-1.37, p < 0.0001) and the VCO₂NL ratio (1.14, 95% CI 1.09-1.21, p < 0.0001) showed significant correlations with successful weaning. A VCO₂NL ratio > 79% and EtCO₂ > 34 mmHg showed strong predictive value for successful weaning (area under the receiver operating characteristics (ROC) curve values of 0.85 [95% CI 0.80-0.89] and 0.84 [95% CI 0.79-0.89], p < 0.0001).

Conclusions: EtCO₂ and the VCO₂NL ratio may be valuable indicators for predicting successful VA-ECMO weaning. Higher EtCO₂ and VCO₂NL ratio values are associated with a greater likelihood of successful weaning.