Introduction

Esophagectomy is the standard treatment for malignancy of the esophagus [1]. Morbidity and mortality rates after esophagectomy have improved over time but are still higher than that of other commonly performed thoracic surgical procedures [2]. Previous studies have shown that postoperative pulmonary complications such as pneumonia and acute respiratory distress syndrome are the most frequent complications following esophagectomy, occurring in about 17 to 67 % of all patients [3–5]. Respiratory complications are the most common cause of postoperative death after esophagectomy, with postoperative pneumonia associated with a mortality rate of 20 % [6]. Therefore, preventing pulmonary complications is the main target of postoperative care following esophagectomy.

Aspiration, which can result in postoperative pneumonia, is a common and potentially fatal complication after esophagectomy [7]. Clinical bedside swallowing tests and videofluoroscopic swallowing study (VFSS) are used for diagnosing dysphagia and identifying subglottic aspiration. Clinical bedside swallowing tests, which do not require complex equipment, are a convenient method of evaluating aspiration. They can be performed relatively easily at the patient’s “bedside,” soon after esophagectomy, allowing early-stage detection of aspiration by physicians. Clinical bedside swallowing tests, however, can miss silent aspiration, which occurs when a patient aspirates without any outward symptoms. In such cases, VFSS is useful for diagnosing subglottic aspiration. It can detect small amounts of material entering the subglottis and evaluate the movement of pharyngeal and laryngeal structures related to the swallowing function. VFSS is not routinely conducted in clinical practice because the examination is difficult to perform immediately after surgery and requires that the patient be able to consume foods and liquids of varying consistencies.

Subglottic aspiration after esophagectomy in esophageal cancer patients has not been thoroughly evaluated and is usually overlooked despite a previous study which shows that comprehensive evaluation for aspiration, including the bedside swallowing tests and routine VFSS, can reduce the incidence of postoperative pneumonia [7]. Routine contrast esophagraphy is frequently performed after esophagectomy in order to screen for anastomotic leakage. However, the detection of subtle swallowing abnormalities such as silent aspiration can be difficult. The advantages of VFSS over routine esophagography include the diversity of food texture, a diverse amount of single-swallowed material, and the application of compensated techniques such as supraglottic swallowing; none of these is feasible with routine esophagography [6]. If the clinical bedside swallowing assessment is sufficient to detect subglottic aspiration after esophagectomy in esophageal cancer patients, then VFSS can be omitted; if not, VFSS can be performed selectively in patients who exhibit the clinical risk factors related to aspiration.

The aims of this study were to assess the usefulness of clinical bedside swallowing tests for detecting subglottic aspiration after esophagectomy in esophageal cancer patients and to identify the clinical risk factors for selecting patients at high risk of subglottic aspiration who may need further evaluation by VFSS.

Materials and methods

This study was approved by the Institutional Review Board of Samsung Medical Center (SMC 2014-03-057). Patients who underwent esophagectomy for esophageal cancer at a single institution between January and August 2013 were included in the study. Patients were excluded if they (1) had received a tracheostomy, (2) had a previous surgical history of gastrointestinal resection, or (3) had esophageal cancer combined with other tumors of the gastrointestinal tract that required modified surgical techniques not commonly in use.

Clinical bedside swallowing tests and VFSS were performed on all the patients 7 to 10 days post-surgery. Evaluations were performed and analyzed by a physiatrist and a speech-language pathologist, both with more than 10 years of experience. The clinical bedside swallowing tests covered changes in voice quality, voluntary cough strength, initiation of lingual movement (normal, delayed), initiation of laryngeal elevation (normal, delayed), and range of laryngeal elevation (normal, decreased) during a volitional swallow. Coughing, throat clearing, and change in voice quality during minimal oral feeds with sips of thin liquids were also included in the clinical bedside swallowing tests.

VFSSs were performed on each patient in lateral and frontal erect positions. The patient was then administered 3-, 6-, and 9-ml aliquots of liquid barium (thin) and barium pudding (thick) and 1 tsp of a barium-coated cookie from a spoon. Two repetitions of each combination (volume and consistency) were obtained. Subglottic aspiration was defined as the entry of material into the airway below the true vocal folds. Silent aspiration was defined as the entry of material below the level of the true vocal cords, without cough or any outward sign of difficulty. Patients with evidence of subglottic aspiration were recommended nil per os with instructions on oral care and hygiene and started on laryngeal elevation/pharyngeal constriction swallowing exercises, with serial reevaluation conducted radiographically with VFSS as clinically indicated until the patient’s diet could be appropriately determined.

Demographic information and details from the esophagectomy operation were obtained to identify risk factors for subglottic aspiration. The candidate risk factors were age, sex, previous history of stroke, operation time, neoadjuvant combined chemotherapy and radiation therapy (CCRT), vocal cord paralysis, anastomotic leakage, three-field lymph node dissection, and cervical anastomosis. If operation time was determined to be one of the risk factors for aspiration, the specific cutoff point of operation hours that showed the highest odds ratio of subglottic aspiration by chi-square test was acquired.

Chi-square tests were used to determine the individual association of items from the clinical bedside swallowing evaluation and risk factors with the aspiration event. The sensitivity, specificity, positive predictive value, and negative predictive value for each item of the clinical bedside swallowing evaluation were calculated using VFSS as the gold standard. For variable selection, aspiration was regressed on each candidate risk factor using a univariate logistic regression model. Predictors with a two-sided univariate P value of <0.05 were selected for multivariate logistic regression analysis. The final model included only those risk factors with a multivariate two-sided P value of <0.05. The SPSS 21.0 software package (SPSS, Inc., and IBM Company, Chicago, IL, USA) was used for statistical analysis.

Results

A total of 124 patients were evaluated as possible study participants. Six patients were excluded from the study for the following reasons: two patients also had stomach cancer, two patients had a previous history of gastrointestinal resection, and two patients had a tracheostomy. Thus, a total of 118 patients were included in the study.

Preoperative demographic data, comorbid conditions, postoperative complications, and technical details are shown in Table 1. Overall, using VFSS, 38 patients (32.2 %) showed evidence of postoperative aspiration. Among those who had subglottic aspiration, 17 (44.7 %) were found to have undergone silent aspiration; i.e., they did not show cough or any outward sign of difficulty during aspiration.

Table 1 Preoperative demographic, comorbid conditions, postoperative complications, and technical details of patients undergoing esophagectomy
Full size table

In the clinical bedside swallowing assessment, a change in voice quality had a sensitivity of 68.4 %, specificity of 70.0 %, positive predictive value of 52.0 %, and negative predictive value of 82.4 %. Cough after swallowing was the second best test for screening subglottic aspiration, with a sensitivity of 55.3 % and specificity of 67.5 %. It had a positive predictive value of 44.7 % and a negative predictive value of 76.1 % (odds ratio (OR), 5.570, P = 0.018) (Table 2).

Table 2 Sensitivities, specificities, and predictive values of clinical bedside swallowing evaluation for the detection of the aspiration
Full size table

Prolonged operation time (OR, 0.651 per hour, P = 0.017) and vocal cord paralysis (OR, 9.1; P < 0.001) were found to be significant clinical risk factors related to subglottic aspiration, based on multivariate logistic regression analysis (Table 3). Operation time greater than or equal to 6 h had the highest sensitivity and specificity. Subglottic aspiration was seen in 29 of 69 patients (42.0 %) whose operation time was greater than or equal to 6 h (P = 0.007). Age, sex, previous history of stroke, neoadjuvant CCRT, anastomotic leakage, three-field lymph node dissection, and anastomosis site were not significantly associated with aspiration.

Table 3 Multiple logistic regression analysis of risk factors for aspiration
Full size table

The patients were classified into two groups according to length of operation time: high-risk group (presence of vocal cord paralysis or operation time more than 6 h) and low-risk group (absence of vocal cord paralysis and operation time less than 6 h). The high-risk group showed 42.3 % prevalence of subglottic aspiration while the low-risk group showed 17.0 % prevalence of subglottic aspiration. In patients with both vocal cord paralysis and operation time more than 6 h, the prevalence of subglottic aspiration was 84.6 %. Overall, among patients with subglottic aspiration, 41.7 % showed spontaneous recovery in follow-up VFSS performed 1 month after esophagectomy.

Aspiration pneumonia occurred in 13.2 % of patients that had aspirated, according to VFSS findings. Aspiration pneumonia did not occur in patients that, according to VFSS findings, had not aspirated.

Discussion

We found that clinical bedside swallowing tests had a sensitivity of 68.4 % and a specificity of 70.0 %, which suggests that without VFSS confirmation, these tests are not sufficient to detect subglottic aspiration after esophagectomy in esophageal cancer patients. Vocal cord paralysis and prolonged operation time were risk factors related to subglottic aspiration, and if a patient had one of these risk factors, the probability of subglottic aspiration increased to 42.3 %, which makes VFSS a necessary procedure.

Clinical bedside swallowing assessment has been a useful screening tool for dysphagia, especially in patients with acute stroke [8, 9]. In this study, change in voice quality sensitivity was observed in 68.4 % of the cases of aspiration and change in cough after swallowing was observed in 55.3 %. Neither voice quality sensitivity nor cough after swallowing could be used for screening because 31.6 and 44.7 % of the cases of aspirations, respectively, were missed by voice quality and cough assessments. Furthermore, silent aspiration was found in 44.7 % of the patients that had been confirmed as having aspiration with VFSS. These findings indicate that aspiration after esophagectomy frequently occurs without clinically overt symptoms, making it difficult to screen using just clinical bedside swallowing assessment. Therefore, for early detection of silent aspiration during the acute postoperative period, a swallowing evaluation such as VFSS or fiber optic endoscopic evaluation of swallowing is required. On the other hand, we noticed that when follow-up VFSS was conducted on patients who showed aspiration in the initial VFSS 1 month after esophagectomy, 41.7 % of these patients no longer showed aspiration. This finding indicates that subglottic aspiration after esophagectomy occurs primarily during the acute period. Therefore, early detection and intervention are important for preventing secondary complications such as aspiration pneumonia.

Vocal cord paralysis is a common complication resulting from injury to the recurrent laryngeal nerve during esophagectomy. In our study, the incidence of vocal cord paralysis was 12.7 %, which is in accordance with the findings of previous reports from Western countries [10–12]. Recurrent laryngeal nerve injury incidence increased to 45.3 % in patients with three-field lymph node dissection [13]. However, only 22.9 % of our study subjects underwent three-field lymph node dissections, which may explain the low incidence of recurrent laryngeal nerve injury. Although mostly transient, vocal cord paralysis is known to be significantly associated with increased incidence of postoperative pneumonia [10, 11, 14]. There are several mechanisms that explain aspiration in vocal cord paralysis; the most likely mechanism is insufficient glottic closure and impaired laryngeal sensation [15]. Change in voice quality is the most significant complication identified by bedside swallowing assessment. It can be understood in the same context as vocal cord paralysis because the most common cause of change in voice quality is vocal cord paralysis. However, the number of patients with aspiration is greater than the number of patients with vocal cord paralysis. This difference indicates that vocal cord paralysis is not the only cause of aspiration. Some episodes of aspiration may be caused by temporary damage and postoperative edema of laryngeal structures related to swallowing after esophagectomy [7, 16–18].

The multivariate analysis identified prolonged operation time as a significant risk factor for postoperative aspiration (Table 3). To our knowledge, this is the first study to verify prolonged operation time as an independent risk factor for aspiration after esophagectomy. It should be noted that among patients who underwent operations longer than 6 h, 42 % showed aspiration after VFSS evaluation. Three-field lymph node dissection and cervical anastomosis failed to show significance in the multivariate analysis. However, they were statistically significant in the univariate analysis (Table 3) and have been demonstrated to be potential risk factors for aspiration after esophagectomy in previous studies [7, 19, 20]. Clearly, three-field lymph node dissection requires more time than the two-field lymph node dissection. Similarly, cervical anastomosis is a complicated procedure that, in our institution, is usually performed along with the three-field lymph node dissection. It is likely that three-field lymph node dissection and cervical anastomosis contribute to prolonged operation time. These predictors were not statistically significant in multiple regression analysis.

It is generally accepted that aspiration during or after swallowing is likely to cause aspiration pneumonia, especially in postoperative patients. However, in our study, there were fewer cases of aspiration pneumonia than cases of aspiration seen in VFSS. Langmore et al. have reported that aspiration is necessary but insufficient to cause pneumonia unless other risk factors are present [21]. Other than dysphagia, the risk factors were dependent on feeding, oral hygiene, number of decayed teeth, tube feeding, presence of more than one medical diagnosis, number of medications, and smoking. Because most of the patients in our study did not have any of these risk factors, there were fewer cases of aspiration pneumonia than there were patients with aspiration. Moreover, antibiotics were prophylactically administered to patients who experienced an aspiration event or had aspiration pneumonia, even without a definite abnormality in the chest radiograph. This may have reduced the occurrence of aspiration pneumonia among the patients.

In conclusion, aspiration is a common phenomenon during the acute postoperative period after esophagectomy for esophageal cancer patients. Although clinical bedside swallowing assessment has considerable value, it is not sufficient to reliably exclude subglottic aspiration after esophagectomy. VFSS should be required for early detection of aspiration, especially in patients with vocal cord paralysis or prolonged operation time. Additional studies are needed to fully explore the causes and interactions of these phenomena.