Patterns of Swallowing After Supraglottic Laryngectomy


  • John M. Schweinfurth MD,

    Corresponding author
    1. Division of Otolaryngology, Albany Medical College, Albany, New York
    • John M. Schweinfurth, MD, Hershey Medical Center, Department of Otolaryngology, MC H091, P.O. Box 850, 500 University Drive, Hershey, PA 17033-0850, U.S.A.
    Search for more papers by this author
  • Steven M. Silver MD

    1. Division of Otolaryngology, Albany Medical College, Albany, New York
    Search for more papers by this author

  • Presented at the Meeting of the Eastern Section of the American Laryngological, Rhinological and Otological Society, Inc., Providence, Rhode Island, January 29, 1999.


Objective To understand the anatomical and physiological basis for early recovery of swallowing function after supraglottic laryngectomy.

Study Design Retrospective review.

Methods The records of nine patients who had undergone supraglottic laryngectomy at the Stratton Veteran's Administration Hospital (Albany, NY) between 1994 and 1998 were reviewed. Videofluoroscopic swallowing studies were obtained on all patients as early as was safely possible and were reviewed by a multidisciplinary team of physicians, nurses, and speech pathologists with regard to anatomical and functional differences between successful and unsuccessful recovery of swallowing function.

Results Five of nine patients resumed regular diets including thin liquids within 1 year of surgery; three patients remained dependent on enteral support. Swallowing success was most closely associated with short oropharyngeal transit time and an anterosuperior position of the larynx. Laryngeal positioning, tongue base mobility, and placement and coordination of the bolus for maximum swallowing efficiency can be improved with time and speech therapy.

Conclusions Factors that placed patients at significantly higher risk for aspiration included low laryngeal position and delayed oropharyngeal transit time. Tight lingual-laryngeal closure did not completely prevent aspiration. At the time of the initial surgical procedure it may be important to position the laryngeal remnant as far superior and anterior under the tongue base as possible.


Since the introduction of supraglottic partial laryngectomy by Alonso 1 in 1947, it has become an accepted procedure in the treatment of selected carcinomas of the larynx. Cure rates approaching those of total laryngectomy can be obtained with preservation of the phonatory function of the larynx. 2 An important result of this procedure is that a large portion of the protective anatomy of the upper airway is disrupted, resulting in dysphagia. During the pharyngeal phase of swallowing, entrance of food into the upper airway is prevented by closure of the glottis, depression of the epiglottis over the laryngeal inlet, and anterior rotation and elevation of the larynx to oppose the base of tongue. 3 Standard supraglottic laryngectomy involves resection of the hyoid bone, the epiglottis, the valleculae, the aryepiglottic folds, the upper third of the thyroid cartilage, and the false vocal cords. The procedure may be extended to remove part of the tongue base, the pyriform sinus, and one or both of the arytenoid cartilages. The result is disruption of the normal swallowing mechanism by removing two of the three barriers of the airway and supraglottic sensation. 2,4 This loss of sensation may lead to improper timing of upper esophageal relaxation, which further hinders bolus entry, possibly leading to aspiration. To establish successful deglutition, the patient must accommodate to the alterations in upper airway anatomy after surgery.

Videofluoroscopy is the definitive technique for studying swallowing function because it defines physiology as well as anatomy and can be viewed at reduced speed. The procedure is based on a barium swallow test, which is modified to focus on the upper airway and use a variety of consistencies of swallowed material. During the study the fluoroscopy tube focuses on the lips, the posterior pharyngeal wall, the palate, and the bifurcation of the airway and esophagus inferiorly. The lateral view permits determination of oral and pharyngeal transit times, areas of bolus stasis, and retention as well as aspiration of contrast material. The anterior view can be used to evaluate symmetry as the bolus divides and pours into the pyriform sinuses before reforming at the opening of the esophagus. 5 The procedure may be performed with material of differing consistency, and the patient may be asked to perform a number of therapeutic techniques. After the procedure the examiner should have a definitive answer regarding the presence and timing of aspiration. 6 By reviewing each patient's disease at presentation, surgical treatment, videofluoroscopic studies, and rehabilitative course, we attempted to identify mechanisms used to achieve a successful swallow and the anatomical effects of long-term rehabilitation. Certainly, there are numerous conditions that affect the postoperative swallowing performance of the supraglottic laryngectomee. It is not the authors' intent to provide a comprehensive review of all the factors involved in each patient's swallowing success or failure, nor is it to compare patients as if they were identical cases. The study's purpose is to show that the majority of patients regained good swallowing performance and to attempt to define the mechanisms used to accomplish this, demonstrating similarities between the successful and the unsuccessful swallow. To our knowledge, no study in the North American literature has examined swallowing progress with time and postulated which patients may have a better swallow prognosis with therapy.


The medical records of nine consecutive patients treated by supraglottic laryngectomy and followed postoperatively with video swallowing studies at the Samuel S. Stratton Veteran's Administration Hospital in Albany, New York, from 1994 to 1998 were reviewed. In all patients the hyoid bone was included with the resection and the arytenoid cartilages were spared. One patient required limited resection of the tongue base. All but three patients underwent neck dissection.

Tracheotomies were maintained in all patients until they were able to tolerate a smaller metal tracheotomy tube plugged for 7 days without the need to remove the plug because of dyspnea. Swallowing studies and diet advancement were not delayed because of the presence of a tracheotomy or lack of complete stomal closure. Whenever possible, swallowing studies and therapy were performed with the tracheotomy tube (if present) plugged. In addition, tracheotomies were maintained in the face of planned radiation therapy despite suitability for decannulation at the time of examination.

Patients were classified into two groups based on their ability to resume a regular diet including thin liquids. Information regarding patient age, number of sites involved by the primary tumor (epiglottis, tongue base, false vocal folds, etc.) TNM classification stage, specific surgical procedures, and reconstruction methods was obtained. Data on use of radiation therapy and incidence of recurrence, as well as survival data, were also obtained.

Emphasis was placed on the results of swallowing evaluation by video fluoroscopy and speech therapy consultation in which patients were assessed for the degree of tracheal aspiration and oropharyngeal handling of feedings. The same therapists performed all examinations and subsequent speech and swallowing therapy. Videofluoroscopic examination was performed in each patient as soon after surgery as was safe and prudent based on voice quality and cough efficiency. Patients who are already taking food by mouth were evaluated for their risk of aspiration using clinical signs such as a “wet” voice and pitch change, and by neck auscultation. Patients with clinical evidence of aspiration underwent video esophagram. Modified barium swallow tests using soft, pureed, and liquid diets were evaluated. Degree of penetration was determined based on the percentage of bolus that passed the vocal cords and was either aspirated or cleared with cough, roughly quantified as mild, moderate, and severe. All studies used for the purpose of comparisons were obtained before the completion of radiation therapy. Videofluoroscopic studies were examined in detail with regard to pharyngeal handling and behavior of the contrast material swallowed. This included tongue base–to–pharyngeal wall contact, bolus velocity, clinging of the bolus to the tongue base, splitting of the bolus on the arytenoids, anterior movement of the arytenoids to open the postcricoid space, and epiglottis-like function of the tongue base on the arytenoids to channel the bolus away from the laryngeal introitus. Laryngeal height was considered low if the larynx was below the vertebral body of the third cervical vertebra at rest. Oropharyngeal transit time greater than 2 seconds was considered delayed. Within the regular diet group, comparison was made to later swallowing studies. Studies selected were the earliest stable examination after which there was no significant change in the parameters studied. For example, if subsequent studies of the same individual at 8 and 12 weeks were similar, the earlier study was used for statistical comparison. The length of time to decannulation; development of safe, effective oral feeding with each consistency of food; and removal of gastrostomy tubes was recorded. Patients who were not able to swallow without aspiration were kept on nothing-by-mouth (NPO) status to prevent pneumonia. The results of flexible fiberoptic nasopharyngolaryngoscopic examinations were also recorded. Specifically evaluated were arytenoid size, laryngeal closure, and amount of perilaryngeal pooling of secretions, which were classified as mild, moderate, or severe.

Comparisons between the regular-diet group and the NPO group were carried out by using Student t test. Statistical significance was accepted at the P ≤ .05 level.


Of nine patients undergoing supraglottic laryngectomy between 1994 and 1998, six were advanced to a regular diet, of whom one could not tolerate thin liquids, while three were kept on NPO status secondary to aspiration. The average patient age in the two groups was similar at 66 years for the regular-diet group and 64.6 years for the NPO group. In both groups the average stage of disease was stage 3; the regular-diet group had 1.8 sites involved and the NPO group had 2.6 (P = .34). The primary tumors of each patient are described in Table I. On average, the NPO group required a tracheotomy for airway maintenance for 18 weeks, while the regular-diet group required a tracheotomy for airway maintenance for only 6 weeks. Exactly 66% of each group underwent neck dissection. Two thirds of the NPO group and half of the regular-diet group subsequently underwent external-beam radiation therapy (Table I). All patients in the NPO group developed a recurrence (P = .02); two recurrences were local and one was regional. At the time of writing, two patients are dead of disease and one is alive with no current evidence of disease. One patient in the regular-diet group developed a local recurrence requiring a total laryngectomy 1 year after the original procedure and is alive without evidence of disease at the time of writing; another patient in the regular-diet group developed an unresectable neck recurrence and is alive with disease (Table I).

Table Table 1.. Patient Profile.
original image

BOT = base of tongue; FVC = false vocal cord; AE = aryepiglottic fold; NPO = non-oral diet.

Each patient was followed by flexible fiberoptic laryngoscopic examination, and all were noted to have good laryngeal closure at 4 weeks. All three patients in the NPO group and three in the regular-diet group were noted to have persistent moderate to severe arytenoid edema for a minimum of 4 weeks. The other three patients in the regular-diet group had mild to moderate edema at 4 weeks, which resolved in two patients at 8 and 32 weeks, respectively. This difference was not significant.

With regard to swallowing performance as recorded by video fluoroscopy, there were considerable differences in the handling of swallowed materials in each of the two groups, as listed in Table II. Studies were performed, on average, 17.8 weeks after surgery for the NPO group and 6.5 weeks after surgery for the regular-diet group. The major differences noted were low laryngeal height (P = .008) and slow bolus transit speed (P = .01) in the NPO group. Other parameters studied were penetration (P = .08), splitting of the bolus on the arytenoids (so that part enters the airway with resultant aspiration [P = .2]), and longer tracheotomy tube dependence (P = .3) in the NPO group. In all three patients in the NPO group aspiration was direct and occurred at the termination of the pharyngeal phase of swallowing as a result of splitting of the bolus on the arytenoid mass. Differences in direct contact between the tongue base and the arytenoids allowing the bolus to pass over the laryngeal introitus (“pseudoepiglottis”), and coordinated anterior motion of the arytenoids opening the esophageal introitus in time for bolus entrance were not statistically significant.

Table Table 2.. Comparison of Videofluoroscopic Swallowing Studies Between Groups.
original image

Pseudoepiglottis = closure of tongue against arytenoids.

Within the regular-diet group only, comparisons were made between initial and later studies to identify trends with time and speech therapy (Table III). Initial studies were performed, on average, at 4 weeks after surgery, and later studies at 10.6 weeks. Four patients remained tracheotomy dependent at the time of their initial swallowing study, but by the later studies all had had decannulation. Average times to tolerating a pureed, liquid, and solid diet were 10.5, 17.8, and 19.4 weeks, respectively. Aspiration in the regular-diet group was direct and occurred at the termination of the pharyngeal phase of swallowing in two patients. Penetration occurred well after the pharyngeal phase of swallowing because of pooling in three other patients. Clinical evidence of aspiration ceased, on average, at 21 weeks, with the exception of one patient who continued to aspirate thin liquids only. Average time to gastrostomy tube removal was 24 weeks.

Table Table 3.. Changes With Time and Rehabilitation in Observed Swallowing Study Parameters in Regular Diet Group.
original image

On videofluoroscopic examination, no changes with time during the study period were statistically significant. Factors that became more prominent were anterior transition of the larynx under the tongue base, appearance of “pseudoepiglottis” function with closure of the tongue base against the anterior motion of the arytenoids (P = .06), disappearance of perilaryngeal pooling, and an increase in laryngeal height. Factors that did not improve or change in the interval between swallowing studies included penetration and aspiration, transit time, tongue base–to–pharyngeal wall contact, and “splitting” of the bolus on the arytenoid mass, causing direct aspiration. Specifically, of two patients with evidence of aspiration on early studies, one had only very mild aspiration and the other aspirated liquids only on follow-up examination during the study period. Ultimately, only one patient in the regular-diet group continued to aspirate thin liquids only.


Some degree of aspiration is experienced in up to 67% of patients after supraglottic laryngectomy. 7,8 In our series, the combination of penetration and aspiration on initial examination was a poor prognostic indicator for later swallowing function, because 60% of patients who had immediate penetration on videoflouroscopy never returned to normal swallowing function. McConnel et al. 9 analyzed swallowing function in patients after supraglottic laryngectomy using manoflourography and found three factors predisposing to aspiration: decreased and delayed laryngeal elevation and poor lingual-laryngeal approximation. Freeman et al. 10 reported that prevention of aspiration depends on tight closure of the vocal cords, elevation of the larynx beneath the tongue base, and wide esophageal patency. In addition, learning the “supraglottic swallow” or postdeglutition cough minimizes the amount of aspiration in patients with aspiration secondary to pooling who can coordinate this maneuver successfully.

The swallowing mechanism is dynamic, almost fluid; therefore anything that tethers the larynx (i.e., tracheotomy, scarring, edema, and radiation fibrosis) will negatively affect swallowing performance. Steiniger et al. 11 have previously found an increased incidence of gastrostomy tube dependence in patients receiving postoperative radiation therapy after supraglottic laryngectomy. Not surprisingly, using videoflouroscopy we noted that the constellation of findings in early postoperative, postradiation therapy, and NPO patients was consistent with a slow, stiff, uncoordinated system, and that these findings largely improved with time and speech and swallowing therapy. Superior and anterior translation of the larynx improved tongue base–to–arytenoid approximation and esophageal inlet opening with channeling of the bolus past the airway. On direct videoflouroscopic observation, when the larynx does not elevate, the arytenoids fail to pull away from the posterior pharyngeal wall, leading to “splitting” of the bolus on the arytenoid mucosa reflecting part into the airway. This leads to direct aspiration in the late pharyngeal phase of swallowing. McConnel et al. 9 also found that tongue base occlusion of the laryngeal inlet is critical and that, with time, patients learned to coordinate laryngeal elevation with bolus delivery.

Goals of intervention in the postoperative phase for chronic aspiration are aimed at improving lingual-laryngeal contact. McConnel et al. 9 recommended posterior advancement of the tongue base with anterior resuspension of the larynx, which was successful in four patients. Cricopharyngeal myotomy, if not performed initially, may improve hypopharyngeal emptying and channeling of the bolus into the esophageal inlet, which also prevents pooling and secondary aspiration.

Hirano et al. 12 reviewed 38 patients who underwent supraglottic laryngectomy with modifications to prevent postoperative aspiration including approximation of the larynx to the base of the tongue and cricopharyngeal myotomy. Twenty-two specific variables were determined for each patient, and a statistical analysis was performed to determine their relationship to swallowing function, which was quantified by the duration of nasogastric intubation. Their analysis determined that only the extent of removal of the arytenoid cartilage and the symmetry in removal of the false vocal folds showed statistically significant relationships to the duration of nasogastric intubation. Of the remaining parameters, those with the closest association were the extent of removal of the thyroid cartilage, the false vocal folds, and the aryepiglottic folds. Performance of cricopharyngeal myotomy was not associated with shortened nasogastric intubation. In a similar study, Flores et al. 13 found a trend toward improved deglutition after cricopharyngeal myotomy but noted that this was not statistically significant. Also controversial is the management of the hyoid bone in supraglottic laryngectomy. Flores et al. 13 found that 72% of patients in whom the hyoid was completely resected recovered function compared with 42% in whom it was preserved. They concluded that preserving the hyoid does not improve swallowing recovery postoperatively.

Although six of nine patients resumed independent oral feeding, the average time to gastrostomy tube removal was 25 weeks. A nasogastric tube can cause pressure and irritation, which can interfere with the patient's recovery of independent swallowing function and possibly lead to complications such as septal necrosis and sinusitis. 14,15 Therefore we recommend either percutaneous endoscopic or open gastrostomy tube placement at the time of the original procedure to ensure adequate postoperative nutrition.


Supraglottic laryngectomy results in severe disturbance to the swallowing mechanism by removal of protective layers and sensation. Through the use of videofluoroscopy, we were able to perform a detailed examination of the residual mechanism in these patients and follow them with time. Factors that placed patients at significantly higher risk for aspiration included low laryngeal position and delayed oropharyngeal transit time. Although tight lingual-laryngeal closure is important in efficient swallowing function and can be learned with rehabilitation, it does not completely prevent aspiration. Therefore, at the time of the initial surgical procedure it is important to position the laryngeal remnant as far superior and anterior under the tongue base as possible. Despite a good prognosis for swallowing function in the majority of individuals, rehabilitation after supraglottic laryngectomy can be a slow process; therefore a reliable conduit for long-term enteral feeding is essential.