THE anesthesia machine is the most integral piece of equipment used by anesthesia providers. When semiclosed circle systems are used, unidirectional valves become integral components that direct carbon dioxide containing gas across the absorbent and along the inspiratory limb, mixing with fresh gas as it flows into the patient. Unfortunately, human error is known to be the most common source of machine malfunction, particularly of the breathing circuit.1,2Such malfunctions should be detected during the application of the American Society of Anesthesiologists’ 2008 Recommendations for Pre-Anesthesia Checkout. In that situation, the self-inflating manual ventilation device (SIMVD) is suggested for first-line backup ventilation.‡Responses to such critical incidents are taught in simulation courses, which are required by the American Board of Anesthesiology as a component of its Maintenance of Certification in Anesthesiology program.2,3
We present a unique human-induced failure of the breathing circuit in a patient who was turned 180° away from the machine during major ear, nose, and throat surgery. Sequential responses to ventilate the patient were successfully applied by the resident and attending anesthesiologist based upon their experiences in simulation workshops. We describe the functional aspects of these effective methods for ventilating the patient during this machine failure.
A 53-yr-old, American Society of Anesthesiologists Class 3, 130-kg man was about to undergo parotidectomy and neck dissection in a position 180° away from the conventional General Electric Aestiva/5 anesthesia machine (GE Healthcare, Waukesha, WI). Sufentanil infusion was chosen for its akinetic effect to enable motor nerve monitoring. The patient was draped with adhesive paper sheets, in part stuck to the endotracheal tube, which were tucked and taped to the lateral side of the head of the operating table. After surgical incision, the resident anesthesiologist detected 8 mmHg inspired CO2. He increased total fresh gas flow to 6 l/min without an adequate reduction in inspired carbon dioxide. After further increasing the flow of 100% O2and minute ventilation, he next began to exchange both Medisorb®(GE Healthcare) soda lime absorbent cartridges in the Aestiva/5 absorbent canister. The lower cartridge was easily and quickly exchanged and replaced into the canister. However, the top cartridge was inadvertently placed upside-down and it lodged halfway into the canister, rendering it impossible to dislodge. The patient remained apneic with residual neuromuscular blockade and narcosis during this 3 min period as the SpO2began to fall from 100% to 92%. The attending was called as the resident slammed the cartridge onto the floor and backward into the canister, such that the assembly did partially fit back into the absorbent housing. Oxygen flow was increased to the maximum, the machine was switched to manual ventilation, the adjustable pressure limiting valve was fully closed, and the flush button was continuously depressed as the resident squeezed the reservoir bag. SpO2was 92% with suboptimal ventilation secondary to a large circuit leak, as the attending entered the room.
Observing this inadequacy of ventilation, the attending immediately disconnected the breathing hose expiratory limb from the Aestiva/5 flow sensor and momentarily initiated retrograde mouth to circuit ventilation (fig. 1A), as the SIMVD was connected to auxiliary oxygen and then connected to the same distal expiratory limb. As the bag was squeezed, there was substantial and effortless chest rise with increasing oxygen saturation (fig. 1B). The patient remained hemodynamically stable on sufentanil infusion as additional help arrived. Another absorbent canister was soon delivered from the operating room supply room and normal machine ventilation was quickly reestablished. The remainder of the case was uneventful and the patient had no intraoperative awareness. The duration of complete apnea was estimated at 3 min, whereas the duration of backup SIMVD ventilation was estimated to be 6 min.
The unidirectional mechanism of gas flow in the breathing hose requires unidirectional inspiratory and expiratory valves. Inspired carbon dioxide detection at the Y-piece thus allows for monitoring of the adequacy of soda-lime chemical absorption and will indicate incompetence of the unidirectional valves.
Conventional semiclosed circle systems (i.e. , General Electric Aestiva/5) are not designed to seal the circuit during soda-lime exchange. The human error in this case was the attempt to change the absorbent in a paralyzed patient who required mechanical ventilation. Furthermore, the airway was difficult to access with SIMVD. The absorbent exchange should not have been initiated because high fresh gas flow could have overcome the spent absorbent. However, with modern self-sealing options on some machines (Drägersorb CLIC®[Drägerwerk AG & Co., Lübeck, Germany] and General Electric EZchange®[GE Healthcare, Helsinki, Finland]), it is possible to exchange soda-lime intraoperatively.
Ventilatory emergencies secondary to machine failure would normally require intrusion into the sterile and/or operative field, for example, in neurosurgical cases. Our unique ventilation technique (see Supplemental Digital Content 1, http://links.lww.com/ALN/A908, a video showing a simulation of the intraoperative events) used the expiratory limb of the breathing circuit for retrograde flow of air, oxygen, and some anesthetic agent back to the patient. We do not recommend mouth-to-circuit ventilation at all, even though some circuits have a distal bacterial filter. Ventilation of the inspiratory limb is not effective because the gas would flow directly through an expiratory valve that opens. During the expiratory limb ventilation, it is possible that fresh machine anesthetic gas and oxygen would flow antegrade during exhalation, to fill the expiratory hose, or it might flow retrograde out of the open absorbent canister. We suspect the path of least resistance would be retrograde because the inspiratory valve would have to be lifted first for antegrade flow. This would predict no enrichment of oxygen in the expiratory limb, and significant rebreathing of patient gas. Therefore, a higher delivery of oxygen, although without anesthetic gas, would be achieved through use of the SIMVD connected to oxygen. This provides retrograde delivery of pure oxygen up the expiratory hose, mixing with a significant (measured 800 ml) volume of dead space gas. Reciprocal exhalation then flows antegrade down the expiratory hose and out through the SIMVD expiratory valve. There is no mixing of exhaled gas with fresh oxygen within the SIMVD itself, as might occur with a Mapleson D or Jackson-Rees system. This method was successfully used and it increased the oxygen saturation for approximately 6 min until the canister was replaced. We unfortunately did not observe the level of inspired carbon dioxide, oxygen, or anesthetic agent during this crisis, but this could be measured prospectively in a separate volunteer study. We are unaware of such measurements, and, in fact, are unaware of any published descriptions of this emergency ventilation technique.
In summary, we used an escalating technique of (1) high oxygen flush to overcome a major circuit leak, (2) momentary mouth-to-tube ventilation of the expiratory limb, and (3) SIMVD to expiratory limb ventilation to successfully ventilate and oxygenate this patient for approximately 6 min, without threatening the integrity of the surgical field.
We also question the need for a tapered design of the absorbent cartridge that allowed this human error to jam a cartridge into a canister (fig. 2). We speculate that the cartridge is tapered to prevent inversion once the proximal end turns purple from spent absorbent, and we welcome the manufacturer’s comment on this point.