Continuous gas transfer monitoring during extracorporeal membrane oxygenation. (January 2017)

• Record Type: Journal Article
• Authors: Brendle, C.; Hackmack, K.-F.; Kühn, J.; Wardeh, M.N.; Janisch, T.; Kopp, R.; Rossaint, R.; Stollenwerk, A.; Kowalewski, S.; Misgeld, B.; Leonhardt, S.; Walter, M.
• Is Part Of: Biomedical signal processing and control. Volume 31(2017)
• Page Start: 321
• Page End: 330
• Publication Date: 2017-01
• Subjects: Biomedical system -- Intensive care -- Extended Kalman filter -- Non-linear -- Multi-input/multi-output; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2016.08.023
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract : Highlights: Continuous supervision of extracorporeal O2 and CO2 gas transfer rates is proposed. Only non-invasive sensors (e.g. in gas phase) are used (no blood gas analysis). An extended Kalman filter for estimating O2 transfer rate is proposed. Accuracy and performance were evaluated using in silico and in vivo data. Results suggest that application in closed-loop control is feasible. Abstract: Extracorporeal membrane oxygenation therapy can prevent hypoxia, and the related inherent danger of death, if mechanical ventilation does not maintain sufficient physiological gas exchange. This form of therapy provides supplemental extracorporeal blood oxygenation and decarboxylation. The most important values to be monitored for the supervision of extracorporeal gas exchange are blood gas measurements, which are normally obtained only intermittently. Continuous measurement during this therapy is rarely implemented because the devices used are costly, and the amount of supervision and maintenance required is complex and expensive (even on the relatively short term of several hours). We present an alternative approach to supervise extracorporeal gas exchange that avoids the use of blood gas analyzers. This is achieved using venous oxygen saturation and additional sensors within the gas phase of the extracorporeal setup. Finally, carbon dioxide gas transfer is measured directly and oxygen gas transfer is estimated using an extended Kalman filter based on an evaluated gray-box model. The proposed method is characterized in silico and evaluated using in vivo data (accuracy of oxygen gas transfer estimation 9.1 ± 49 mL/min). Results suggest that monitoring of gas transfer with adequate performance is possible and that clinical application using this measurement value in closed-loop control is also feasible.