An Investigation of Factors Related to Food Intake Ability and Swallowing Difficulty After Surgery for Thoracic Esophageal Cancer

Our principal objective in investigating swallowing difficulty after surgery for thoracic esophageal cancer was to determine whether food intake is possible during the early postoperative period. In addition to clarifying the incidence of swallowing difficulty, we investigated the effects of RLNP and other preoperative and intraoperative factors on absence of the food intake ability during this period. We also examined association between RLNP, and factors examined by VF during the pharyngeal phase of swallowing.

As noted above, laryngoscopic examination revealed RLNP in 65.6% of our patients. This incidence of RLNP may seem high in comparison with incidences previously reported, but the previously reported diagnoses of RNLP were based strictly on clinical symptoms such as hoarseness, and the time between surgery and diagnosis differed from the time between the two among our patients. Therefore, comparison between our data and data previously reported is difficult. For our study, RNLP was defined as either unilateral or bilateral VCP. In one reported study of the incidence of post-esophagectomy RNLP, the disorder was diagnosed by laryngoscopic examination, but RNLP was defined only as bilateral VCP, and the RNLP was documented in about 50% of patients 2 weeks after the surgery. According to reports to date of diagnosis of RLNP by laryngoscopic examination, the incidence of RLNP following surgery for esophageal cancer ranges from 36% to 75%. The incidence among our patients falls within this range. RNLP has been reported to improve by more than 50% during postoperative follow-up. We believe, then, that RNLP is present in about 65% of patients on POD 1, but that the percentage decreases gradually with time. We also believe that RNLP can be present in patients without clinical symptoms.

Meanwhile, 34.4% of our patients were without an inability to take food orally, and this differed from the 65.6% of patients who exhibited RLNP. Our data suggest that postoperative swallowing difficulty may be independent of the presence of RLNP. The incidence of postoperative swallowing difficulty did not differ between patients with and without RLNP, despite the fact that RLNP has been considered a major factor in postoperative swallowing difficulty, according to studies to date. Surprisingly, in this study, we found no relation between RLNP and the factors identified by VF, laryngeal invasion and aspiration. We know that various nerves and muscles work together to effect airway protection and swallowing during the pharyngeal phase. The glossopharyngeal nerve induces the pharyngeal phase reflex, and the airway is then closed by the vocal cords, the false vocal cords, and the epiglottis. This is followed by simultaneous elevation of the hyoid and opening of the esophageal orifice, after which the food bolus is moved into the esophagus by the squeezing motion of the pharyngeal constrictors. With the exception of the cricothyroid muscle, the laryngeal muscles are supplied by the RLN. Accordingly, vocal cord closure is impaired when RLNP occurs [14]. Thus, ineffective closure of the airway increases the risk of aspiration. However, airway protection is not established by vocal cord closure alone. It is due also to the position and movement of the false vocal cords and to the movement of the epiglottis [18]. This means that aspiration in patients is not necessarily due to the RLNP. Basic research performed in animal models has indicated that the larynx is mostly closed even after transection of one or both RLNs, and this does not affect swallowing [19]. Thus, we believe that it is important to consider the fact that postoperative swallowing difficulty cannot be explained by RLNP alone.

Our study yielded two points of major clinical significance. The first is the clear relation between aspiration caused by RLNP and postoperative swallowing difficulty, and this relation is reflected in the C–D and JCOG classification systems, which are generally used to grade postoperative complications. However, it is not possible to correctly evaluate complications, even when these tools are used [10, 11]. The C–D system considers treatment of complications to be the principal objective of classification. The C–D system does not specify RLNP as a specific item but instead includes it under the respiratory section. The JCOG system, which is based on the C–D system, considers either aspiration or RLNP as fulfilling a single criterion. We did not observe any relation between a swallowing difficulty and RLNP in our study patients, so this simple criterion might not allow for accurate evaluation of complications. The C–D system represents complications according to their severity, which is based on the degree of treatment needed for them. Under this system, RLNP is classified according to the degree of treatment needed for aspiration pneumonia resulting from the RLNP. However, when RLNP and swallowing difficulty can be directly evaluated by VF and video esophagography, it might be possible to prevent aspiration in advance. It is even possible that the VF evaluation prevented aspiration in some of our patients, which would have influenced our data. By extension, therefore, we believe postoperative evaluation by means of both VF and video esophagography is necessary for accurate evaluation of RLNP and swallowing dysfunction.

The second point of clinical significance is the dissociation between swallowing disorders and symptoms that suggest RLNP. Physical findings that suggest the presence of aspiration or RLNP, such as hoarseness or coughing, are commonly used in clinical settings to determine when to initiate food intake. However, our study did not show any relation between hoarseness and a postoperative swallowing difficulty. In terms of the dissociation between swallowing function and clinical symptoms, a previous study has included many patients with silent aspiration who did not present with hoarseness, indicating that it may not be possible to predict swallowing difficulty on the basis of physical findings alone [7]. We believe that the results of our study support this conclusion.

We investigated the relation not only between postoperative swallowing difficulty and the presence of RLNP but also between postoperative swallowing difficulty and various preoperative and intraoperative variables. Among these items, we noted a significant association for the item “cervical lymph node dissection.” Moreover, as mentioned above, we found no significant relation between swallowing difficulty and RLN injury. Cervical esophageal anastomosis is a standard procedure at our hospital, so all of our study patients underwent a cervical procedure on the left side during surgery for esophageal cancer. Furthermore, adding cervical lymph node dissection expanded the extent of the cervical procedure bilaterally. Among recently reported studies in which swallowing difficulty has been evaluated by means of VF after surgery for esophageal cancer, two have shown inadequate laryngeal elevation in patients with a postoperative swallowing difficulty [18, 20]. Results of these studies suggested that swallowing difficulty might be caused by inflammation and scarring that result from the cervical procedure. Results of one study are particularly interesting: that swallowing function appeared to be preserved to a similar extent between patients in whom cervical lymph node dissection was not performed and patients in whom the infrahyoid muscles were transected bilaterally to facilitate laryngeal elevation after cervical lymph node dissection [12]. In considering these results comprehensively, we believe that inflammation and scarring of the infrahyoid muscles, as well as incomplete relaxation due to muscle injury, are possible causes of postoperative swallowing difficulty. We believe that these factors also contributed to our study results. The fact that the incidence of swallowing difficulty increased to a greater extent when cervical lymph node dissection, rather than cervical anastomosis, was performed suggests that the risk of these disorders increases as the extent of the cervical procedure is expanded.

Our study was not without limitations. The first was the small number of patients included. This is important with respect to the analyses of T-factors, N-factors, and disease stages. It was also not possible for us to incorporate tumor location, and the extent of cervical lymph node dissection into our analysis because the number of patients in each category was small. Going forward, we will need to accumulate additional cases. The second limitation was that it was not possible to examine the relation between details of the RLNP and swallowing difficulty. In this study, RLNP was diagnosed regardless of whether the VCP was unilateral or bilateral. Factors such as unilateral vs. bilateral VCP, the degree of paralysis, degree of vocal cord atrophy, and reflex time were not documented laryngoscopically. In addition, laryngoscopy was performed only in the acute phase on POD 1, and it was difficult for patients to move to the examination room or ingest food. Therefore, we were limited to briefly observing the presence or absence of RLNP at bedside. However, because there have been many studies evaluating swallowing difficulty in relation to unilateral VCP, we believe that RLNP as we defined it was a valid study endpoint [18, 21]. The third limitation was that the study was a single center study. A standardized surgical method was used, making it easy for us to perform a comparison to invest factors related to swallowing dysfunction in patients who have undergone surgery, including lymphadenectomy, for thoracic esophageal cancer. However, there are reports of differences in the incidence of swallowing dysfunction based on the anastomotic method used [6], so we hope, going forward, to evaluate postoperative swallowing function by means of VF jointly with other institutions that make use of different surgical techniques. The fourth limitation was the fact that assessment of food intake ability by means of VF is somewhat subjective. Three criteria for absence of food intake ability were established and included in our study, and these were agreed upon based on the opinions of the speech therapist, neurologist, and gastroenterologist. VF has been considered a subjective means of evaluation [14]; in recent years, there have been some objective VF-based assessments that have incorporated measurable variables, such as laryngeal elevation (reported as a percentage) and pharyngeal transit time [18, 22]. We therefore believe that assessment of postoperative food intake will become more objective if these methods are pursued and reference values are established.

In conclusion, postoperative swallowing difficulty after surgery for thoracic esophageal cancer is affected very little by the presence of RLNP but greatly by inclusion of cervical lymph node dissection in the operative procedure. Thus, it is difficult to determine when to initiate postoperative food intake simply on the basis of the presence of RLNP and its associated symptoms, as we have done to date, and we believe that VF is a powerful tool that can be used for dynamic assessment. The C–D system can, to some degree, be relied upon for classification of RLNP and swallowing difficulty. We believe that the standards need to be reestablished to take VF and video esophagography findings into consideration.