There was a three part failure: a) heart lung machine (hlm) blood level sensor erroneously detected low blood level after the sensor became unstuck from the reservoir, b) hlm did not make any audible alarm to alert clinical staff to inappropriately sensed low level, and c) hlm low level alarm intervention was incorrectly set to "stop" (turning the pump off) instead of "coast" (turning the pump to 2000 rpm).The patient was then carefully weaned from cardiopulmonary bypass.There was a point at which an alarm sensor on the cardiopulmonary bypass circuit caused a shutdown of the perfusion circuit while the aortic root and left ventricular vents were activated.This entrained air from the surgical field around the aortic cannula insert.It should be noted that the surgical field was insufflated with carbon dioxide gas for the major portion of the surgical procedure.Intraoperative transesophageal echocardiography failed to verify significant intracardiac air.The patient was carefully weaned from cardiopulmonary bypass.However, the patient subsequently developed an episode of ventricular fibrillation.Before scrubbing out, the md asked us to turn off the root vent.About 10 minutes after the cross clamp was removed at 1602, the lower level sensor came unstuck enough from the reservoir to no longer sense fluid (a common problem, which never should have the ability to cause harm).Immediately the low level sensor inappropriately turned the pump to stop, instead of coast at 2000 rpm.There was no audio alert as is common with the alarm, so we were not immediately alerted to the change.I noticed the increase in volume in the reservoir and then saw the red alert on the screen.I reacted by pushing the level sensor against the reservoir, to reattach it.Normally this would resume flow and pull the pump out of coast mode.However, since the pump went to stop, it required me to reset the knob to zero, acknowledge the alert, and resume flow.At the time the lv vent was on about 300 cc/min.As i was reacting to the change in flow, the md noticed at the table and commented that the aorta was very flat, and then said there was air in the arterial line.At that time, i was resetting the pump knob to restart normal flow.During the stop mode, the screen said -0.4 l/min of backflow.I turned the root vent on, slowly resumed normal flow and the patient was put into deep trendelenburg to avoid any air going to the brain.The cerebral saturations dipped on both sides during the reduction of flow.Prior to the event, the right side was significantly depressed in the 20s while the left side was in the 30s.With the resumption of flow the stats increased to levels similar to before the event.We stayed in deep trendelenburg for about 30 minutes while keeping the maps in the 80s.Doctors discussed potentially starting retrograde perfusion to flush any potential air embolism from the brain.They decided it would only introduce more potential for problems during an already complex and long case.I do not know the length of time that there was back flow but i estimate it was about 30 seconds given the amount of volume that was in the reservoir.After the surgery there was no evidence of any neurologic injury in relation to air embolism, confirmed by multiple ct scans.The settings on the pump were changed without perfusionist knowledge, but when the event occurred it was quickly and appropriately managed.
|