Jacobs JR, Logemann J, Pajak TF, Pauloski BR, Collins S, Casiano RR, Schuller DE. Failure of Cricopharyngeal Myotomy to Improve Dysphagia Following Head and Neck Cancer Surgery. Arch Otolaryngol Head Neck Surg. 1999;125(9):942–946. doi:10.1001/archotol.125.9.942
To determine whether cricopharyngeal myotomy can improve dysphagia associated with head and neck cancer surgery.
Prospective, randomized, multicenter trial.
Twelve clinical sites across the United States.
Between 1989 and 1994, 125 patients undergoing combined modality therapy for head and neck cancer, including resection of the tongue base or supraglottic larynx, were prospectively entered into the trial.
Cricopharyngeal myotomy on a randomized basis.
Main Outcome Measures
Videofluoroscopic examination to determine oropharyngeal swallowing efficiency, which is defined as the ratio of percentage of the bolus swallowed to total swallowing time using 3 different bolus consistencies.
No significant difference in oropharyngeal swallowing efficiency between myotomy vs no myotomy was seen at 6 months of follow-up regardless of bolus consistency.
In this prospective test of cricopharyngeal myotomy, the procedure fails to significantly improve dysphagia associated with head and neck cancer surgery. The efficacy of this surgical procedure in other disease entities should also be rigorously explored.
THE CRICOPHARYNGEAL muscle was first described as a distinct entity by Valsalva in 1770 according to Ellis1 and others.2,3 Relaxation of this muscle along with the upward and forward movement of the larynx is thought to be one of the prime contributors to upper esophageal sphincter opening.4,5 Jackson and Shallow6 first identified abnormal cricopharyngeal relaxation as a clinical entity and suggested a possible relationship to the development of a pulsion diverticulum. Dohlman and Mattsson7 and Holinger and Benjamin8 divided the cricopharyngeus muscle endoscopically for management of a diverticulum. Kaplan,9 in 1951, was the first to suggest cricopharyngeal myotomy for a condition other than a diverticulum. He sectioned the muscle in patients with poliomyelitis and obtained a good result.
The presence of swallowing problems in patients surgically treated for head and neck cancer has been well documented in the literature.10- 21 Mitchell and Arminini22 reviewed their 6 cases of patients undergoing myotomy and recommended that the procedure be performed in cases of vagal injury, glossectomy, oropharyngeal resections, and partial laryngectomies. This recommendation was made despite the fact that none of their 6 cases involved head and neck resections. Multiple other authors23- 26 have also recommended myotomy at the time of major head and neck resection based on the experiences of relatively few patients.
Other authors have been less convinced of the value of cricopharyngeal myotomy in terms of management of dysphagia following head and neck cancer surgery. Litton and Leonard27 studied aspiration following partial laryngeal surgery and were unable to demonstrate improvement with myotomy. Flores et al,28 in one of the largest series in the literature, prospectively compared myotomy with no myotomy in 51 patients undergoing partial laryngeal surgery and were unable to demonstrate any improvement.
In summary, cricopharyngeal myotomy is a widely advocated procedure in the absence of many data supporting efficacy, especially from controlled clinical trials. Since dysphagia is well documented in patients who have undergone surgery for head and neck cancer, patients undergoing surgical management of the base of tongue and supraglottic larynx were selected as the study population for this randomized prospective trial. The goal of this trial was to investigate whether cricopharyngeal myotomy can improve dysphagia associated with head and neck cancer surgery.
Twelve institutions participated in this prospective, randomized study. The study was carried out under the auspices of the Radiation Therapy Oncology Group (RTOG). The institutions were members of either the Southwest Oncology Group or the RTOG (Table 1). The inclusion criteria for the study involved patients older than 18 years with squamous cell carcinoma of the head and neck that would require resection of the base of the tongue or supraglottic larynx as part of their overall management. Surgery of these anatomic areas is well recognized to result in postoperative dysphagia. The magnitude of this dysphagia is generally related to the size of the operative defect, which is directly related to the size of the primary lesion. A history of primary neurologic disorder that could affect swallowing was grounds for exclusion. Informed consent was obtained from all subjects. The study protocol specified surgical criteria for the study. Both the oncologic portion of the procedure and the length of the myotomy to be performed on a randomized basis were defined.
Videofluoroscopic examinations were used to evaluate the oropharyngeal swallow. Each subject attempted to perform 2 trials each of 4 liquid volumes (1, 3, 5, and 10 mL), one-third teaspoon of paste (Esophatrast), and one fourth of a cookie coated with barium paste for contrast for a maximum potential number of 12 protocol swallows. These swallows were recorded in the lateral plane. The data were gathered preoperatively, at 3 days after healing (defined as removal of feeding tube), at 3 months after healing, and at 6 months after healing. The swallow studies were reviewed centrally at the Swallow Physiology Laboratory, Department of Communication Sciences and Disorders, Northwestern University, Evanston, Ill. Laboratory personnel were blinded to the treatment group to which the patient was randomized. The results of the analysis were sent on computer disk to the RTOG statistical office for inclusion in the study database.
Central reviews of the surgery and videofluoroscopic studies were performed. The surgical review was performed by the study surgical chairman and members of the RTOG surgical committee with the assistance of RTOG headquarters staff. For each surgical procedure, the operative and pathology reports were reviewed and scored relative to protocol prescription. The surgery was scored as follows: as per protocol, minor variations acceptable, major variations acceptable, and major variations unacceptable. The videofluoroscopic studies were reviewed for adequacy of the image for analysis at the Northwestern University facility. There was a formal data monitoring committee in place to oversee the trial's progress. The study protocol was approved by the National Institutes of Health and the review boards of the RTOG and all the participating institutions.
The randomization plan for the trial stratified the patients by participating institution such that treatment assignments were balanced after every 4 patients within each institution. On average, this would produce 2 treatment groups with similar overall prognosis and pretreatment characteristics.
The principal end point for the study was oropharyngeal swallowing efficiency (OPSE), a global measure that describes the interaction of speed of movement of the bolus and the safety and efficiency of the mechanism in clearing material from the oropharynx while preventing aspiration.29,30 The OPSE is the approximate percentage of the bolus swallowed into the esophagus divided by total transit time. The approximate percentage of the bolus swallowed is 100% minus the approximate percentage of oral residue, pharyngeal residue, aspiration before the swallow, and aspiration during the swallow. Total transit time is the sum of oral transit time and pharyngeal transit time. This index is a convenient measure for comparing various clinical populations that may have different impairments in the oropharyngeal swallow.30 Swallow efficiencies are presented as mean and SE for each treatment arm and bolus consistency (liquid, paste, cookie). The comparison of OPSE between the treatment arms was analyzed using nested repeated-measures analysis of variance at each postoperative evaluation.31 The model included 3 parameters—one each for the consistency effect, the treatment effect, and the interaction effect between consistency and treatment. Each derived parameter was tested against being different from zero. A 5V option of software (BMDP; Statistical Solutions Ltd, Cork, Ireland) was used in generating these models. The derived estimates for the main effect due to treatment in the model are reported along with their 95% confidence intervals. Positive values favor myotomy, whereas negative values do not.
A total of 125 patients with cancer of the head and neck who were to undergo resection of the tongue base or supraglottic larynx as part of their disease management were prospectively entered and randomized into the trial between March 1989 and March 1994. Twenty-two patients were excluded from statistical analyses, because of the absence of a preoperative videofluoroscopic study (n=14), because they received a total laryngectomy (n=5), or for various other reasons (n = 3), leaving 103 patients available for analysis. Fifty-four of these patients were randomized to receive a cricopharyngeal myotomy, and 49 did not receive myotomy. Distribution of pretreatment characteristics by treatment group is given in Table 2. Significance tests indicate that the 2 treatment groups were comparable in their pretreatment characteristics. Although patients may have had a primary tumor site in the oral cavity or hypopharynx, they were still eligible for this study because their surgical resection included either the tongue base or supraglottic larynx.
Assuming that myotomy would only be of value if the improvement was long-term, the 6-month postsurgical data were examined first. These data are derived from 48 patients who completed the 6-month videofluoroscopic study (25 with myotomy and 23 without myotomy). The 2 treatment groups had almost the same percentage of protocol swallows not performed (20% for the myotomy group vs 21% for the no myotomy group), indicating that the 2 groups had a similar degree of difficulty with the study protocol. One patient did not have OPSE data and was excluded from the statistical analyses. Mean OPSE is presented by treatment group and bolus consistency in Table 3. There was no significant consistency by treatment interaction (P=.70), indicating that treatment differences were similar for each consistency. No significant difference was observed between the 2 treatment groups (P=.95). The estimate for overall treatment effect was −0.39, with a 95% confidence interval of −5.75 to 4.96.
The findings for the other 2 postoperative evaluations (3 days after healing and 3 months after healing) were essentially the same as the 6-month evaluation. The data for the evaluation 3 days after healing are presented in Table 4. Seventy-two patients (38 with myotomy and 34 without myotomy) had videofluoroscopic studies performed at 3 days after healing. Patients assigned to the myotomy treatment had a slightly lower percentage of total protocol swallows performed (35% for myotomy vs 41% for no myotomy), indicating that they had greater difficulty with the study protocol. Six patients did not have OPSE data and were eliminated from statistical analyses. Mean OPSE is presented by treatment group and bolus consistency in Table 4. The consistency by treatment interaction was not significant (P=.68), and there were no significant differences observed between treatment groups (P=.81). The estimate for overall treatment effect was −0.82, with a 95% confidence interval of −7.47 to 5.82.
The data for the 3-month posthealing evaluation revealed that 54 patients (28 with myotomy and 26 without myotomy) had videofluoroscopic studies performed. The 2 treatment groups each had 14% of the total protocol swallows not performed, indicating that the 2 groups had a similar degree of difficulty with the study protocol. Two patients did not have OPSE data and were excluded from statistical analyses. Mean OPSE is presented in Table 5. There was no significant interaction (P=.85) and no significant treatment difference (P=.29). The estimate for overall treatment effect was −3.63, with a 95% confidence interval of −10.37 to 3.11.
At all evaluation points, there was a significant difference in mean OPSE across consistencies (Table 3, Table 4, and Table 5). Paste and cookie OPSEs were significantly lower than the OPSE for liquids.
The surgical complications did not differ statistically between the myotomy and no myotomy groups (Table 6). The number of surgical quality control reviews completed and not completed is listed in Table 7. The most common variation identified was in the myotomy group. This was characterized as failure to define on the operative dictation the length of the myotomy performed. This variation was considered minor and acceptable.
In this prospective randomized study of patients with head and neck cancer, the performance of the myotomy did not have an impact on swallow efficiency. This result is consistent with those of other studies27,28 that investigated the swallow function of patients with head and neck cancer who had undergone myotomy. Despite the variety of primary sites of tumor for the patients in this study, resection of the tongue base or supraglottic larynx was involved in each patient's treatment. If the aim of cricopharyngeal myotomy is to assist in the opening of the cricopharyngeus during the swallow, then patients with resection of the tongue base or supraglottic larynx theoretically should benefit most from a myotomy, since resection of these structures is likely to disrupt normal pressure generation for bolus transit into and through the pharynx. In this study, however, cricopharyngeal myotomy did not have an impact on swallow function.
Deglutition is a complex neurophysiologic action. Functioning of the upper esophageal sphincter is an integral part of a successful swallow. There is no doubt that disruption of the actions of the upper esophageal sphincter can result in life-threatening dysphagia. Disruptions of the upper esophageal sphincter may involve disorders of contraction and relaxation of pharyngeal musculature. However, elevation of the larynx or issues of timing of the food bolus are also critical to the successful swallow. It would be reasonable to assume that a myotomy will only improve those cases where abnormal contraction and relaxation of cricopharyngeus muscle is the principal cause of the disorder. This surgical procedure has been advocated for a wide variety of disorders over time. Many of the studies that advocate the procedure's performance are retrospective, with small sample sizes and inadequate definition of the patient's swallowing physiology. Unfortunately, use of the current technology of videofluoroscopic examinations or manometric studies is unable to discriminate patients who could potentially benefit from the procedure. Combined videofluoroscopy and manometry or electromyography is needed.
The lack of statistical significance for the treatment differences observed in this study could be due to 1 of 2 reasons. Either there really is no difference in treatment or the sample size was too small to detect a real difference. All 3-point estimates for overall treatment effect were in favor of no myotomy. Using the upper 95% confidence limit, the largest probable difference in favor of myotomy was a small OPSE difference of 5.82, indicating that the lack of significance is likely due to a lack of a true difference rather than an insufficient sample size. The mean OPSEs in both the group undergoing myotomy and the group not undergoing myotomy indicate that, on average, the patients in this study had functional swallows after treatment as defined by their OPSE scores compared with healthy patients. In their investigation of normal swallowers, Rademaker et al32 found that the fifth percentiles of OPSE for liquid, paste, and masticated boluses were 56, 30, and 25, respectively. Given the functional status of the swallow of these patients who have undergone surgery, it may not be surprising that myotomy did not result in further improvement in swallow function. Patients with head and neck cancer and more severe swallowing disorders may still benefit from cricopharyngeal myotomy. The use of cricopharyngeal myotomy in patients with confirmed dysphagia who have undergone surgery for head and neck cancer may yield different results from those observed in this study. We would advocate that the efficacy of this surgical procedure be rigorously explored in other disease entities through prospective randomized trials.
Accepted for publication May 14, 1999.
This work was supported by grant 5R01CA46764 from the National Institutes of Health and grant P01 CA40007 from the National Cancer Institute, Bethesda, Md.
Presented at the American Society of Head and Neck Surgery Annual Meeting, Scottsdale, Ariz, May 11, 1997.
Corresponding author: John R. Jacobs, MD, Department of Otolaryngology, Wayne State University, UHC-5E, 540 E Canfield, Detroit, MI 48201.