Deviation from Clinical Routines Can Reveal Sources of Device Design Vulnerability

Ventilation is the movement of air or gas from the external environment into the alveoli of the lungs. In the critical care setting, the ability to mechanically ventilate a patient in acute distress is a lifesaving skill, as these patients often cannot adequately breathe on their own. As such, ventilating is almost always more important than intubation per se. In emergencies, initial ventilation typically is established using a simple face mask in conjunction with a bag that the clinician manually squeezes to generate positive pressure and gas flow through the patient's airway. These bags are commonly referred to as Ambu bags, a proprietary term that traces to a popular airway equipment brand. When a patient is breathing spontaneously, their inspiratory muscles, mainly the diaphragm, generate a pressure force that by convention is referred to as a “negative inspiratory force” or negative pressure that pulls outside air into the lungs. In contrast, when an inspiratory drive is absent, air or other gases can be “pushed” into the lungs by mechanical means (referred to clinically as positive pressure).

Face masks are designed to have a soft-contoured, air-filled, cushion-like cuff that lies directly on the patient's face, thereby allowing a seal to be formed over the mouth and nose (Figure 1). Achieving a proper seal is crucial to generating positive pressure. The cushion commonly consists of polyvinyl chloride plastisol due to its malleable properties.¹  These masks are used extensively in anesthesiology because general anesthesia, and particularly intravenous (IV) anesthetics, often impede a patient's ability to breathe on their own.

Further, a paralytic medication is typically used in the setting of general anesthesia in order to optimize intubation conditions before an endotracheal tube is placed to secure the airway. The paralytic medication will completely prevent all skeletal muscle movements, including those of the diaphragm, hence eliminating any remaining spontaneous breathing drive that the patient may still have. The typical clinical sequence of events on induction (i.e., initiation of the anesthesia) of general anesthesia is to administer the IV anesthetic first and, only after adequate mask ventilation is confirmed, to then follow with a paralytic medication.

In the event that mask ventilation fails (e.g., upper airway obstruction, equipment failure, or an unexpected airway pathology such as a tracheal fistula), the clinician may be able to backtrack to safety if a spontaneous respiratory drive is regained by the patient. Common IV induction agents such as propofol have the ideal pharmacokinetic property of a very short duration of action. Accordingly, their respiratory depressing effect can potentially be undone with the passage of time by opting to awaken the patient if ventilation cannot be achieved as expected² .

Before each procedure, the standard of care in anesthesiology is to check the circuit apparatus as part of the anesthesia machine check.³  This machine check includes a positive pressure test whereby the breathing circuit is checked for leaks. The overall steps to the positive pressure test are outlined in Figure 2. Some anesthesia machines automatically perform the test on their own with the press of a button by the clinician. A common design in anesthesia machines includes a blank metal knob to which the circuit can be connected in order to close off the circuit and allow for pressurization by the machine (Figure 3). Such design requires intentional physical removal of the mask from the circuit in order to occlude the apparatus. The implicit assumption in this design is that the mask is reliable enough to be removed and excluded from the pressure test. This contrasts with the broad recognition that the mask is a vital component of the breathing circuit and without which the circuit is almost useless.

We experienced failure of this type of mask and, at the time of the incident, were unable to identify its cause. After we induced general anesthesia, it was immediately evident that we could not generate positive pressure and failed to ventilate. After ruling out patient-specific causes that would interfere with ventilation, various external equipment-related reasons must be considered.

The most common culprits for such a scenario are probably the presence of a leak somewhere along the path of oxygen flow or insufficient oxygen supply to the machine. These leaks can occur in the connections of the breathing circuit or within the anesthesia machine itself. At the time of our experience, we could not identify an apparent leak and the patient's oxygen saturation rapidly desaturated as we were failing to ventilate adequately. Thankfully, intubation was successful, and ventilation was then established via the endotracheal tube.

Afterwards, a close examination of the equipment eventually revealed a tear in the cuff of the mask. We believe that the mask was torn during induction when general anesthesia was initiated. This elderly patient had advanced dementia and refused to let the clinical team establish IV access. The decision was made to perform a mask induction in which inhaled anesthetic gases are used via the face mask instead of IV induction agents. This practice is common in young pediatric patients, for which IV access is challenging to achieve before induction of anesthesia. Given that our patient of this instance was elderly, he was missing numerous teeth and his remaining teeth were larger than those of a pediatric patient (Figure 4). When mask induction was initiated, the patient thrashed his head aggressively from side to side and grabbed the mask with his hands to forcefully remove it from his face. This necessitated the clinician to firmly hold the mask to the patient's face. In this process, the sharp edges of the patient's teeth likely caught the cuff and tore it.

Before this incident, we did not suspect the mask itself to be the equipment piece responsible for the leak. Given its exclusion from the pressure test, it is likely that the engineers who designed the anesthesia machine also did not think of it as a culprit for a leak. A leak around the mask is a common etiology of failed ventilation, but this occurs due to difficult airway features, such as a beard, deformed or abnormal facial structure, or conditions requiring considerably higher pressures to be generated (e.g., an obese patient).

This patient's airway was clinically unremarkable on exam during the preoperative physical evaluation. A leak around the mask was therefore not thought of and was low on the differential of problem etiology. Our general approach to diagnosing real-time leaks that occur after a proper machine check with a satisfactory pressure test was to focus on any changes that may have occurred after the test and to listen for audible signs of a leak. Before this incident, we partitioned leaks into two groups based on the physical size of the opening: large versus small equipment deformities. Our impression was that a large opening (e.g., a marked circuit disconnect) would be relatively obvious and visually apparent, whereas a small opening (e.g., tear, hole, loose connection fitting) would be more easily evident by an audible hissing sound. Our thought process was that a minor insult or opening in the circuit would still allow for some level of pressure to be generated within the apparatus and that this pressure escapes or leaks with turbulence and is therefore audible.

As the above experience illustrates, the aforementioned dichotomy is a categorization not always maintained. The tear in the mask was tiny and on the inside portion of the mask (Figure 5), making it hard to visualize or hear. Nonetheless, because it was a tear in a cuff, it translated to a larger area of compromise that effectively prevented any positive pressure from being generated.

This example demonstrates that even though the face mask itself does not take part in the positive pressure leak test, it can still be an important source of a major leak. Moreover, it highlights that when a medical device is used in a fashion at variance with its usual use or under altered conditions, extra vigilance to new sources of malfunction is warranted.