An Unusual Cause of Airway Obstruction during General Anesthesia with a Laryngeal Mask Airway

S. E. C. is a 54-yr-old woman who was scheduled to undergo open reduction and internal fixation of an isolated trimalleolar ankle fracture that she sustained after a fall. Her medical history was significant for an idiopathic, partial-complex seizure disorder of approximately 25 years’ duration. Her surgical history was significant for placement of a vagal nerve stimulator (VNS) approximately 2 yr previously to improve control of her seizures. Her current medications consisted of carbamazepine (1,200 mg/day), multivitamins, and calcium supplements.

The patient was premedicated with 1 mg intravenous midazolam and 50 μg intravenous fentanyl. General anesthesia was induced with 40 mg lidocaine-150 mg propofol, and a No. 4 laryngeal mask airway (LMA) was inserted without difficulty on the first attempt. Anesthesia was maintained with isoflurane (1–1.4%) in oxygen (50%, balance nitrogen).

Spontaneous respiration resumed almost immediately after LMA insertion. However, shortly thereafter, the patient was noted to have inspiratory stridor and sternal retractions suggesting partial obstruction. Continuous positive airway pressure was applied (approximately 20 cm H²O), resulting in markedly decreased stridor and retractions. Positive pressure was continued for approximately 1 min, after which the obstruction spontaneously resolved. Approximately 5 min later, partial obstruction and stridor recurred but were again relieved by continuous positive airway pressure. After approximately 1 min, the apparent obstruction resolved spontaneously, only to recur again approximately 5 min later.

This pattern of intermittent obstruction followed by an unobstructed interval recurred over approximately 40 min. It was then decided to perform fiberoptic laryngoscopy through the lumen of the LMA to identify the source of the obstruction. A fiberoptic bronchoscope swivel adapter was attached to the LMA to permit insertion of a fiberoptic bronchoscope without disconnecting the patient from the anesthetic circuit. Initial examination, during an unobstructed period, showed the LMA to be appropriately positioned and the cords to be fully open (fig. 1). The fiberoptic scope was left in place until the next obstructed episode, at which time the left arytenoid and aryepiglottic fold were seen to have been pulled across the midline to almost completely obstruct the glottic opening (fig. 2). Application of positive airway pressure moved the vocal fold part of the way back toward its normal position and improved ventilation; however, when positive airway pressure was released, the arytenoid would again move across the midline to obstruct the airway. This persisted for approximately 1 min, after which it spontaneously resolved.

Given the fixed periodicity of the patient’s obstructive episodes, we began to suspect that her VNS was the source of the intermittent abnormal positioning of her arytenoid. Review of her records revealed that her VNS was set to fire at 30 Hz for 1 min followed by a 5-min off-time. This pattern exactly matched the pattern of intermittent obstruction. We reasoned that stimulation of the left vagus nerve resulted in stimulation of the recurrent and superior laryngeal nerves, which produced tetanic contraction of the musculature on the left side of the larynx.

Vagal nerve stimulators were approved by the US Food and Drug Administration in 1997 to control intractable seizures. They consist of a pulse generator and a bipolar lead that encircles the left vagus nerve in the neck. The generator is programmable, and the controllable variables include on-time, off-time, stimulus amplitude, and stimulus frequency. The mechanism by which VNSs reduce seizure severity is not known but is presumed to result from retrograde conduction of electrical impulses via  the vagus nerve to stimulate brain structures known to influence seizures.

Although the VNS is effective in reducing seizure frequency, 1,2it also produces noticeable side effects. Most problems relate to the period of vagal nerve stimulation, and the most common problem (66% of patients) is that of alterations in the voice, such as hoarseness or breathiness. 3Other problems include dysphagia, dyspnea, aspiration, muscle pain, and paresthesias.

The VNS-mediated side effects associated with the voice are caused by stimulation of both the superior and the recurrent laryngeal nerve, which are branches of the vagus nerve. Together, these two nerves innervate all the muscles of the larynx. Several studies that have examined patients with voice abnormalities during vagal nerve stimulation have documented that the left vocal fold and arytenoids are displaced medially during the period of vagal stimulation. For example, Zalvan et al . 4reported two adult cases of voice change after vagal nerve stimulation. During videoendoscopy, both patients’ left vocal folds were noted to move to the midline and become immobile during vagal nerve stimulation. Interestingly, the immobility persisted for a brief period (5–7 s) after stimulation ceased. Similarly, Lundy et al . 5demonstrated movement of the vocal folds to the midline during vagal nerve stimulation at 40 Hz. Interestingly, these investigators found that the direction of vocal fold movement was dependent on the frequency of stimulation. At a frequency of only 20 Hz, vagal nerve stimulation produced abduction of the vocal fold. Rate-related differences in movement of the vocal folds are presumably related to differences in which pharyngeal muscles are being stimulated at a given frequency.

The current patient did not volunteer any complaints of airway problems related to her VNS, nor was she observed to have any voice alterations during the preoperative evaluation. After the fact, we contacted her attending neurologist, who confirmed that she has not had any airway-related problems (the new HIPPA law made direct contact with the patient to ask about airway problems inadvisable). Therefore, the patient seems to be a member of the large minority who do not experience noticeable airway side effects during VNS firing. Given the apparent absence of airway symptoms while awake, it is curious that she had such marked airway narrowing during anesthesia. It may be that while awake, she is able to subconsciously compensate for the effects of the VNS. Alternatively, it may be that general anesthesia unmasked the effect of the VNS on her airway. Regardless of the reason, her absence of airway symptoms suggests that this is not a guarantee that patients without noticeable airway problems while awake will not develop airway obstruction while anesthetized.

In summary, the VNS produces abnormal motion of the vocal folds during stimulation in many patients. In the current case, this abnormal motion was responsible for intermittent episodes of partial airway obstruction during general anesthesia with an LMA. When anesthetizing patients with a VNS, it may be prudent to consider turning the VNS off during anesthesia if an LMA is to be used. In addition, in those patients with a VNS in whom an endotracheal tube is indicated, one should consider the possibility of vocal fold trauma if the VNS intermittently adducts the vocal fold against the hard endotracheal tube. In these patients, it would also seem prudent to consider turning the VNS off during the duration of anesthesia.