Continuous Oxygen Insufflation Using a Speaking Tracheostomy Tube Is Effective in Preventing Aspiration during Feeding

A 65-yr-old man received a spinal cord injury due to fracture dislocation of the 4th and 5th cervical vertebra. Initial neurologic examination revealed bilateral anesthesia below the C4 levels and quadriplegia. Twenty-four hours after injury, tracheostomy was performed. Three days after injury, respiratory failure due to pulmonary congestion developed, and ventilation was mechanically supported. Twelve days after injury, posterior fusion of the 4th and 5th vertebra and iliac bone graft were performed to stabilize the cervical spine. The postoperative course was uneventful and the neurologic deficit was unchanged.

Thirty-seven days after injury, rehabilitation using oral feeding was begun. Deglutition was practiced carefully for 2 weeks using liquid and semisolid feeding mixtures. Swallowing appeared well coordinated. Although no signs of aspiration were observed during eating, tracheal suction revealed significant and frequent aspiration associated with feeding. Increasing pressure within the cuff of the tracheostomy tube during feeding did not prevent aspiration.

The glottic closure reflex was examined in detail using fiberoptic laryngoscopy. A 6.0 mm-ID armored endotracheal tube was inserted through the tracheostomy stoma to ventilate the lungs. The cuff was inflated using the minimal leak method to seal the trachea, and ventilation was supported manually using a Jackson-Rees system. A 3.0 mm-OD fiberoptic laryngoscope was inserted upward though the stoma to visualize glottic movement. A disruption of the glottic closure reflex during deglutition was observed. As shown in Figure 1, downward ballooning and incomplete closure of the true vocal cord was observed during the pharyngeal phase of deglutition. Premature opening of the glottis was followed by massive penetration of water into the trachea (Figure 2). The cuff of the endotracheal tube was deflated by 50% to allow air flow into the larynx, and the glottic closure reflex was reassessed. When continuous positive subglottic airway pressure was applied manually during deglutition, downward ballooning of the true vocal cord disappeared, and the glottis remained closed. No significant penetration of water was observed.

To apply continuous positive subglottic airway pressure during feeding, a speaking tracheostomy tube (Argyle Aspiraid, Nihon Sherwood, Tokyo, Japan) was inserted. Through the narrow-gauge pilot tube ending above the cuff, [4]humidified oxygen was insufflated into the subglottic airway continuously during feeding. Oxygen insufflation at the rate of 1 l/min, optimal for phonation in this patient, abolished aspiration. Suction through the same tube demonstrated absence of food particles or liquid residue in the subglottic airway. Five days after the initial fiberoptic examination, a second fiberoptic examination was performed to assess any unexpected effects of oxygen insufflation. No significant abnormality was observed in subglottic and tracheal mucosa. Seventy-one days after injury, separation from mechanical ventilation was accomplished successfully.

It is well known that tracheostomy causes swallowing dysfunction and may promote pulmonary aspiration. [5,6]Cameron et al. reported that the presence of an inflated tracheal cuff does not reduce the incidence of aspiration. [1]In the current case, increasing tracheal cuff pressure during and after feeding did not effectively prevent aspiration. A videofluoroscopic study demonstrated that impaired laryngeal closure during deglutition is responsible for penetration of food into the airway below the true vocal folds. [2]In patients with impaired laryngeal closure, clearance of food particles and liquid residue from the larynx before inspiration is crucial to prevent aspiration. [7]However, insertion of a cuffed soft tracheostomy tube abolishes expiratory air flow (coughing) through the larynx. Stasis of secretions and food between the true vocal folds and the tracheal cuff eventually will induce aspiration into the lower airway. [3]Using a speaking tracheostomy tube similar to ours, Shahvari et al. demonstrated that removal of fluid from the subglottic airway was useful in preventing aspiration. [4]However, continuous suction of the subglottic airway during deglutition may promote penetration of food through the glottic opening. Moreover, it is difficult to remove food particles completely through the narrow-gauge pilot tube. Therefore, restoration and enforcement of laryngeal closure during deglutition is of primary importance in preventing aspiration in these patients.

The glottic closure reflex plays a central role in laryngeal closure during deglutition. In the current case, incomplete closure and early opening of the glottis were observed, accompanied by massive penetration of water into the subglottic airway. Sasaki et al. demonstrated in a series of animal experiments that tracheostomy led to disruption of the glottic closure reflex. [8]They suggested that neurophysiologic change after tracheostomy, as in the current case, may cause diminished after-discharge activity of the thyroarytenoid muscle and result in a weakened and unsustained adductor.

Fiberoptic examination clearly demonstrated that manually increased subglottic airway pressure was effective in enforcing the glottic closure reflex, restoring competent laryngeal closure in this patient. Continuous oxygen insufflation into the subglottic airway using a speaking tracheostomy tube was effective in preventing aspiration. It is suggested that increased subglottic airway pressure by continuous oxygen insufflation may be effective in enforcing laryngeal closure during deglutition and preventing penetration of food into the subglottic space. An insufflation flow rate of 1 l/min was selected by the patient for comfortable phonation. This flow rate proved effective in enforcing laryngeal closure.

Although the current case demonstrates that continuous oxygen insufflation into the subglottic airway is useful in preventing aspiration during feeding, non-physiologic flow of poorly humidified gas for a long period might be harmful to the airway mucosa. In this patient, no acute change of the glottic and tracheal mucosa was observed by fiberoptic examination. Further study of this issue is necessary to assess the clinical validity of this simple method.

The authors thank Dr. Dovie R. Wylie, for help in preparing this manuscript.