Microbubble Activity during Extra Corporeal Life Support

Objectives: Hypovolemia in extracorporeal life support (ECLS), causing reduced venous backflow, suction of the venous cannula and decreased pump flow, is assumed to be a major reason for microbubble formation. In this study, we established an in-vitro model mimicking hypovolemia and aimed to optimize elimination of microbubbles.

Material and methods: Patient and ECLS system were simulated by serial connection of reservoirs, one centrifugal pump and one membrane oxygenator. The system was primed with a diluted 40% glycerin solution. A bubble counter was used for microbubble detection and volume determination. Measurements were performed under physiological conditions as well as for hypovolemia simulation by continuous or intermittent occlusion of the venous line determined by a roller pump. Revolution speeds were set to 1500, 2000, 2500, 3000 and 3500rpm. A dynamic bubble trap (DBT) and a vacuum connected oxygenator were evaluated for microbubble elimination.

Results: Microbubbles were successfully generated with the established setting, even at 2500rpm, a standard setting used in clinical routine. Revolution speed and volume status were found to have a significant correlation to microbubble formation in the venous as well as in the outlet line. Dynamic negative pressure, induced by intermittent occlusion of the venous line to simulate an aspiration of the venous cannula, resulted in the highest incidence of microbubble formation. In general, cannula aspiration in the lower range (-66 to -143 mm Hg) led to a significant higher amount of bubbles in the pump outlet in comparison to the venous line (p < 0.001). This phenomenon was diminished with decreasing pressures. However, a huge difference in bubble size was measured (venous = 0,28 µL/L, outlet = 52,50 µL/L). The installation of a DBT significantly reduced the incidence of microbubbles up to 96.9% and simultaneously, a reduction of the total bubble volume of 99.8% was achieved (p < 0.001). Filtration via the oxygenator proved not efficient for smaller microbubbles (< 40 µm). In case of high microbubble activity however, the oxygenator showed no significant effect on microbubble count and volume.

Conclusion: Centrifugal pumps are a major inductor of microbubbles. The incidence of microbubbles significantly correlated with revolution speed and volume status. A DBT offers the potential to reduce the incidence and gas volume of microbubbles during ECLS.