Continuous monitoring of membrane lung carbon dioxide removal during ECMO: experimental testing of a new volumetric capnometer

Background: Extracorporeal membrane oxygenation constitutes a complex support modality, and accurate monitoring is required. An ideal monitoring system should promptly detect ECMO malfunctions and provide real-time information to optimize the patient–machine interactions. We tested a new volumetric capnometer which enables continuous monitoring of membrane lung carbon dioxide removal (V′CO2ML), to help in estimating the oxygenator performance, in terms of CO2 removal and oxygenator dead space (VDsML). Methods: This study was conducted on nine pigs undergoing veno-arterial ECMO due to cardiogenic shock after induced acute myocardial infarction. The accuracy and reliability of the prototype of the volumetric capnometer (CO2RESET™, by Eurosets srl, Medolla, Italy) device was evaluated for V′CO2ML and VDsML measurements by comparing the obtained measurements from the new device to a control capnometer with the sweep gas values. Measurements were taken at five different levels of gas flow/blood flow ratio (0.5-1.5). Agreement between the corresponding measurements was taken with the two methods. We expected that 95% of differences were between d − 1.96s and d + 1.96s. Results: In all, 120 coupled measurements from each device were obtained for the V′CO2ML calculation and 40 for the VDsML. The new capnometer mean percentage bias (95% confidence interval limits of agreement) was 3.86% (12.07-4.35%) for V′CO2ML and 2.62% (8.96-14.20%) for VDsML. A negative proportional bias for V′CO2ML estimation with the new device was observed with a mean of 3.86% (12.07-4.35%). No correlations were found between differences in the coupled V′CO2ML and VDsML measurements and the gas flow/blood flow ratio or temperature. Coupled measurements for V′CO2ML showed strong correlation (rs = 0.991; p = 0.0005), as did VDsML calculations (rs = 0.973; p = 0.0005). Conclusion: The volumetric capnometer is reliable for continuous monitoring of CO2 removal by membrane lung and VDsML calculations. Further studies are necessary to confirm these data