Please see attachment of case report for full details.
Summary:
During peripheral ECMO therapy on a patient with a persistent VSD, ejection from the left ventricle was found to be impaired by transient hypocapnic responses. Loss of LV ejection was first identified during a routine membrane “sigh” where gas flow was increased to 10l/min for 30-60 seconds to blow condensed moisture from the gas phase of the membrane oxygenator.
During this “sigh” procedure the patient’s arterial systolic pressures dropped form 110mmHg down to 50mmHg. After full diagnostics of ECMO circuit the gas flow blender was found to be capable of delivering 30l/min to the oxygenator without the operator’s full knowledge. Although theorised; it is thought that delivery of massively hypocapnic blood to the arterial system resulted in peripheral vasoconstriction, failure of the LV to eject via the aortic valve and preferential offloading to the right ventricle.
See attached
This was a very unusual phenomena so a full examination of ECMO components was performed. During this examination we attached a gas flow meter to the Sechrist gas blender to check the meter was calibrated correctly. From gas flows 1L/min-10L/min the blender was delivering flows equal to that indicated by the bubble flow meter, above 10l/min however gas flows were essentially unregulated. During our testing when the bubble was raised just 0.5cm above the 10L/min line the actual delivered gas flow was 28-30L/min (what i assume to be wall pressure gas). As a result of such high gas flows being delivered to the oxygenator, the partial pressure of dissolved CO2in the blood was being lowered drastically, this in turn was leading to a short term spike in vasoconstriction in the peripheries (through peripheral chemoreceptors and endothelial NO responses to hypocapnia), which caused the LV to preferentially shunt to the RV through the VSD and a functional loss in ejection through the aortic valve (leading to the arterial pressure loss).
In response; warning tape was placed above the 10L/min line on the blender, Sechrist was contacted and nurses were instructed to sigh at no more than 9L/min.
Replacing the affected blender was not considered as all available blenders at GJNH were tested and found to have the same issue.
