1. Introduction

Patients with neurogenic dysphagia typically have lesions involving the central nuclei and neural pathways of the glossopharyngeal and vagus nerves.¹ These lesions impair swallowing function, increase the risk of aspiration, and predispose patients to recurrent pulmonary infections. Accordingly, tracheostomy is frequently required in clinical practice to improve pulmonary ventilation–perfusion function.^2,3 Researchers have reported that up to 46.5% of patients with stroke require tracheostomy.⁴ However, tracheostomy induces multifaceted anatomical and physiological alterations, which may adversely affect the swallowing function. First, the tracheostomy tube may mechanically restrict laryngeal elevation and anterior rotation.^5–7 In addition, an inflated cuff may exert mechanical compression on the esophagus, thereby compromising the clearance of secretions and swallowing function.⁸ Second, patients with neurogenic dysphagia undergoing tracheostomy are ventilated via an artificial airway, which eliminates airflow through the upper respiratory tract and oropharynx. This reduces oropharyngeal mucosal sensitivity, attenuates afferent sensory input, and interferes with the central regulation of the swallowing reflex.⁹ Furthermore, long-term intubation may result in laryngopharyngeal mucosal edema, granulation tissue formation, and even vocal fold fixation, thereby further exacerbating dysphagia and aspiration risk.^10,11 Previous studies have demonstrated that 59%–70% of patients with tracheostomy exhibit aspiration.¹²

Decannulation is a complex process that requires multidimensional assessment based on the patient’s underlying disease, lesion location, swallowing function, and other clinical factors.¹³ Effective swallowing function is a key predictor of successful decannulation, whereas severe dysphagia has been associated with a decannulation failure rate of 47%.¹⁴ Clinical controversy remains regarding the optimal sequence of interventions—that is, whether clinicians should first address the tracheostomy tube or prioritize the management of dysphagia.¹⁵ In previous studies, researchers compared patients with an in situ tracheostomy tube to those in whom the tracheostomy tube was temporarily removed and the tracheostoma covered with gauze, thereby evaluating only the immediate effects of transient tube removal.^16,17 Such study designs do not adequately capture the dynamic changes in laryngeal and swallowing function before and after definitive decannulation.

Flexible endoscopic evaluation of swallowing (FEES) is an established gold-standard instrumental assessment of swallowing, with the advantages of repeatability and bedside applicability.¹⁸ FEES allows direct visualization of structural movement, accumulation of secretions, clearance ability, pharyngeal residue, and aspiration.¹⁹ In recent years, FEES has increasingly been used to guide decannulation decisions in patients with tracheostomy and has demonstrated significant value in assessing whether patients can achieve effective vocal fold closure, exhibit a cough reflex, and perform spontaneous throat-clearing maneuvers after cuff deflation.^10,20 Although prior studies have explored the application of FEES in tracheostomy management, there are still few data describing the dynamic evolution of FEES parameters in the same patient cohort before and after decannulation.^8,20 Notably, recovery of swallowing function in patients with neurogenic dysphagia undergoing tracheostomy is often nonlinear and influenced by multiple factors, including level of consciousness, fatigue, and cognitive engagement.¹³ Therefore, in this retrospective study, we analyzed paired FEES data obtained before and after decannulation to simulate longitudinal observation and address the limitations of single-time-point assessments.

This retrospective study aimed to comprehensively compare FEES-related characteristics before and after decannulation in the same group of patients with neurogenic dysphagia undergoing tracheostomy. Using FEES, we sought to determine whether the presence of a tracheostomy tube affects swallowing function and explore the specific features of dysphagia associated with tracheostomy. Our findings aim to provide objective evidence to inform the ongoing clinical controversy regarding intervention priorities (decannulation versus dysphagia treatment) and to offer an evidence-based perspective for the clinical management of patients with neurogenic dysphagia undergoing tracheostomy.

2. Methods

2.1. Participants and data collection

We conducted a retrospective review of the medical records of patients with tracheostomy who underwent successful decannulation. This retrospective study was reported in accordance with the STROBE guidelines (https://www.strobe-statement.org/). We deemed patients to be eligible if they met the following criteria: (1) tracheostomy indicated for a neurogenic disease, such as stroke, traumatic brain injury, brain tumor, or hypoxic–ischemic encephalopathy; (2) age between 18 and 80 years; and (3) completion of FEES (ATMOS MedizinTechnik GmbH & Co. KG, Lenzkirch, Germany) both before and after decannulation within a 3-month period. Exclusion criteria were as follows: (1) tracheostomy performed for a non-neurogenic condition (e.g., head and neck cancer or respiratory failure); (2) a pre-existing history of dysphagia before the index neurological event; (3) FEES performed > 3 months before or after decannulation; and (4) absence of dysphagia or only mild dysphagia (e.g., trace residue on FEES). This study was approved by the Institutional Review Board of the Third Affiliated Hospital of Sun Yat-sen University (Approval Number[2021]02-351-01).

We collected sociodemographic data (e.g., age and sex) and clinical data (e.g., primary diagnosis, date of decannulation, and dates of FEES examinations). The FEES assessment comprised two components. The first component was an anatomic–physiologic assessment, including (1) vocal fold mobility; (2) the presence of arytenoid cartilage swelling; (3) secretion accumulation, assessed using the Murray Secretion Severity Rating Scale; (4) laryngeal sensation; and (5) clearance ability for sputum and secretions. The second component focused on the evaluation of swallowing. We administered 3-mL boluses of thickened green-dyed liquid, prepared by mixing 100 mL of water with 3 g of thickener (Softia S, NUTRI Co., Ltd., Yokkaichi, Japan). Data collected during the swallowing evaluation included (1) the presence of premature spillage, (2) delayed swallowing reflex, (3) residue severity (none, trace, mild, moderate, or severe), and (4) penetration and aspiration, assessed using the Penetration−Aspiration Scale (PAS). All FEES examinations were performed by an experienced physician trained in standardized FEES protocols, and all examinations were video-recorded for subsequent analysis.

2.2. Statistical analyses

Statistical analyses were performed using the SPSS version 27.0 for Windows (IBM Corp., Armonk, NY, USA). For continuous variables, normality was assessed using the Shapiro–Wilk test. Normally distributed data were presented as means ± standard deviation, and pre–post differences were examined using paired-sample t-tests. Non-normally distributed data were presented as median (interquartile range), and comparisons before and after decannulation were performed using the Wilcoxon signed-rank test. Categorical variables were summarized as frequencies and percentages and were compared using the Chi-square test. A two-sided P value of < 0.05 was considered to be statistically significant.

3. Results

3.1. Baseline characteristics

As shown in Fig. 1, a total of 56 patients were enrolled from July 2020 to September 2023. Three patients did not undergo a swallowing evaluation after decannulation. Table 1 presents the demographic and clinical characteristics of the study cohort. The mean age of the patients was 55.86 ± 16.67 years, and 42 patients (75.00%) were male. The underlying diseases included stroke (47/56), traumatic brain injury (3/56), brain tumor (4/56), and hypoxic–ischemic encephalopathy (2/56). The median disease duration was 6.50 (3.25–9.00) months. Most participants had a stroke, and the key outcomes were generally consistent within the stroke subgroup. However, because of the limited sample size of the nonstroke subgroup, subgroup analyses according to disease type were not performed. The median interval from decannulation to FEES was 9.50 (5.00–14.75) days.

3.2. Anatomic–physiologic assessment

We analyzed changes in the following indicators before and after decannulation: (1) vocal fold mobility, classified as normal, unilateral vocal fold mobility impairment (VFMI), bilateral VFMI, tremor, or not assessable because of impaired visualization caused by granulation tissue or other structural abnormalities; (2) arytenoid cartilage status, classified as normal or swelling; (3) secretion accumulation, evaluated using the Murray Secretion Severity Rating Scale; (4) laryngeal sensation, categorized as normal, deficit, or absent; and (5) sputum and secretion clearance ability, categorized as normal, effective, partial response, or no response. The results are summarized in Table 2.

No statistically significant difference in vocal fold mobility was observed before and after decannulation. In contrast, significant improvements were observed following decannulation in arytenoid cartilage swelling (P < 0.001), secretion accumulation (P < 0.001), laryngeal sensation (P < 0.001), and clearance ability (P < 0.001).

Arytenoid cartilage swelling was observed in 71.43% (40/56) of patients before decannulation and in 37.50% (21/56) after decannulation (χ² = 12.996, df = 1, P < 0.001), as illustrated in Fig. 2A. As shown in Fig. 2B, the median Murray Secretion Severity Rating Scale score decreased from 3 (2–3) before decannulation to 1 (1–2) after decannulation (P < 0.001). Normal laryngeal sensation was present in 8.93% (5/56) of patients before decannulation and in 26.79% (15/56) after decannulation; laryngeal sensation deficit was observed in 50.00% (28/56) of patients before decannulation and in 58.93% (33/56) after decannulation; laryngeal sensation was absent in 41.07% (23/56) of patients before decannulation and in 4.28% (8/56) after decannulation (P < 0.001), as illustrated in Fig. 2C. Normal clearance ability was absent before decannulation (0/56) but was present in 7.14% (4/56) of patients after decannulation. Effective clearance ability increased from 12.50% (7/56) before decannulation to 39.29% (22/56) after decannulation. Partial response occurred in 37.50% (21/56) of patients before decannulation and in 46.43% (26/56) after decannulation. The proportion of patients with no response decreased from 50.00% (28/56) before decannulation to 7.14% (4/56) after decannulation (P < 0.001), as demonstrated in Fig. 2D.

3.3. Swallowing evaluation

Three patients declined the FEES-based swallowing evaluation after decannulation. As the anatomic–physiologic assessment and swallowing evaluation represent distinct but complementary components that assess laryngopharyngeal anatomy/function and real-time bolus transit and swallowing safety, respectively, these three patients were excluded only from the swallowing analysis and remained eligible for the structural assessment. Consequently, the swallowing function was compared before and after decannulation in 53 patients. Overall, swallowing function improved significantly after decannulation. As illustrated in Fig. 3, significant reductions were observed in pyriform sinus residue (P < 0.001), vallecula residue (P < 0.001), and PAS scores (P < 0.001). The median pyriform sinus residue score decreased from 3 (2–4) before decannulation to 2 (1–3) after decannulation (P < 0.001). Similarly, the median vallecula residue score declined from 3 (2–4) before decannulation to 2 (1–3) after decannulation (P < 0.001). The median PAS score improved from 3 (1–7) before decannulation to 1 (1–3) after decannulation (P < 0.001).

4. Discussion

In patients with neurogenic dysphagia who have undergone tracheostomy, whether to prioritize dysphagia management or pursue early decannulation remains a subject of debate. Clinical experience often suggests that tracheostomy has a substantial impact on swallowing function, whereas some studies argue that swallowing disorders in these patients primarily reflect the underlying neurologic condition.¹⁵ In this retrospective study, we analyzed dynamic changes in swallowing function before and after decannulation in patients with neurogenic dysphagia who have undergone tracheostomy. Our findings indicate that tracheostomy tubes have detrimental effects on swallowing function, many of which appear to be at least partially reversible after decannulation. We observed significant improvements after decannulation, including arytenoid cartilage swelling, reductions in pharyngeal secretions, enhanced laryngeal sensation, and improved safety and effectiveness of swallowing. These results suggest that the presence of a tracheostomy tube may impair swallowing through mechanical compression and sensory suppression, whereas tube removal may facilitate functional recovery. Overall, these findings provide objective evidence to inform the optimization of airway management strategies in patients with neurogenic dysphagia and are broadly consistent with key principles outlined in the current clinical guidelines.

4.1. Anatomic–physiologic changes related to swallowing

Our study showed that the frequency of arytenoid cartilage swelling decreased and laryngeal sensation improved after decannulation. Laryngeal edema is a common complication in patients undergoing oral endotracheal intubation.²¹ and is frequently caused by direct laryngeal injury.²² However, the etiology of arytenoid edema in patients with neurogenic dysphagia who have undergone tracheostomy is more complex. In addition to direct trauma, edema may be associated with pooled secretions in the laryngopharynx, compression from a nasogastric tube, laryngopharyngeal reflux, and other contributing factors.^10,23 Langmore et al.¹⁰ suggested that edema may contribute to diminished laryngeal sensation. As laryngeal sensation helps prevent bolus entry into the airway, it is essential for swallowing safety.²⁴ When laryngeal sensation is compromised, the risk of aspiration increases, thus predisposing patients to pneumonia.²⁴

Despite these changes, we did not observe significant differences in vocal fold mobility before or after decannulation. This may be related to the characteristics of our study population. Because this was a retrospective analysis of FEES findings before and after decannulation, all included patients had already achieved successful decannulation. Patients who are successfully decannulated typically have adequate airway patency,^14,21 which may explain why a substantial proportion of the cohort (41.07%, 23/56) already demonstrated normal vocal fold mobility before decannulation. To reduce selection bias and better clarify the effects of tracheostomy on vocal fold function, future studies should adopt a prospective design and include patients who fail decannulation.

4.2. Secretion management and swallowing safety

Patients in this study exhibited reduced accumulation of secretions and improved clearance after decannulation. These changes were likely associated with restoration of subglottic air pressure, enhanced cough effectiveness, and improved laryngeal elevation.^25,26 The significant reduction in PAS scores further indicates improved swallowing safety after decannulation. Previous research has reported that in patients who have undergone tracheostomy, modifying subglottic air pressure using a Passy–Muir speaking valve can markedly reduce aspiration.²⁵ Accordingly, we propose that decannulation may remodel pharyngeal swallowing biomechanics to more effectively prevent bolus entry into the airway and facilitate complete epiglottic inversion. In addition, enhanced laryngeal sensation may contribute to more effective clearance, thereby further improving swallowing safety. Although some studies have suggested that the presence of a tracheostomy tube does not significantly alter movement of the hyoid bone or laryngeal excursion,^27,28 additional clinical trials are required to more comprehensively delineate the effects of tracheostomy on aspiration risk.

4.3. Changes in swallowing function

Regarding swallowing effectiveness, we found that decannulation substantially reduced pyriform sinus and vallecula residues. On the one hand, arytenoid cartilage swelling may mechanically impede epiglottic inversion and limit relaxation of the cricopharyngeal muscle. Resolution of edema after decannulation may therefore alleviate these mechanical obstructions.^6,29,30 On the other hand, improvements in laryngeal sensation may also contribute to reduced pharyngeal residue.^20,31

4.4. Clinical implications and guideline alignment

The findings of this study are consistent with the current clinical guidelines that emphasize the importance of objective instrumental assessment for decannulation decision-making, particularly FEES, as recommended in the Chinese Expert Consensus on Adult Tracheostomy Decannulation^32,33 and standardized endoscopic swallowing evaluation protocols (SESETD) validated in neurocritical care populations.^20,34 Our observation that decannulation enhances swallowing safety, as reflected by the lower PAS scores, and swallowing effectiveness, as reflected by the reduced pharyngeal residue, supports guideline-directed strategies to minimize tracheostomy duration, thereby reducing the risk of infection and optimizing airway protection.^34,35 These results corroborate recommendations advocating for early decannulation once objective criteria—such as the resolution of arytenoid edema and improvement in laryngeal sensation—are fulfilled, as outlined in the SESETD framework.^{20, 34}

At present, relatively few studies have investigated the relationship between tracheostomy and dysphagia, and most available evidence has focused on endotracheal intubation. Dysphagia secondary to endotracheal intubation is well documented.³⁶ Prior investigations have demonstrated that endotracheal intubation may cause laryngeal injury associated with aspiration.³⁷ Consequently, the role of endotracheal intubation in the development of dysphagia is generally accepted. In contrast, the impact of tracheostomy on dysphagia remains controversial. Our results indicate that swallowing function improved after tracheostomy tube removal, which is consistent with clinical observations. Nevertheless, several studies have reported no association between tracheostomy and dysphagia.¹⁵ These studies often focused on patients with head and neck cancer or other conditions without pre-existing dysphagia. Conversely, other investigations have supported the influence of tracheostomy on dysphagia, including the study of acute respiratory failure conducted by Langmore et al.¹⁰

Dysphagia after tracheostomy can be attributed to a variety of factors, including abnormalities in laryngeal movement, upper airway edema, and pharyngeal weakness,¹⁰ some of which are consistent with our findings. The relationship between tracheostomy and dysphagia may vary depending on the underlying disease etiology. All patients included in our study had neurogenic disorders and presented with established dysphagia at baseline. Following decannulation, we observed a significant improvement in swallowing function. To mitigate the potential impact of rehabilitation treatment on dysphagia, we reduced the interval between the two FEES examinations. Meanwhile, we contend that a marked improvement in swallowing function is challenging to achieve within such a brief period through routine swallowing rehabilitation. Moreover, spontaneous recovery is unlikely in patients with a prolonged disease course.

However, our findings differ somewhat from studies reporting no direct association between tracheostomy and dysphagia.¹⁵ This discrepancy is likely related to differences in patient populations. We focused on patients with neurogenic dysphagia with pre-existing swallowing deficits, in whom removal of the tracheostomy tube may relieve secondary anatomical and physiological barriers (e.g., edema and secretion retention) to recovery. Clinically, our results highlight the need for individualized protocols that balance adherence to guideline-based criteria with patient-specific factors, such as neurologic prognosis, while prioritizing FEES-guided decision-making to avoid unnecessary prolongation of tracheostomy. Future implementation should integrate these findings into multidisciplinary care pathways, consistent with previous scoping reviews³⁸ that advocate combining instrumental assessments with targeted rehabilitation to optimize swallowing outcomes.

5. Limitations

There were several limitations to this study. First, although we attempted to minimize the interval between the two FEES assessments, it was not sufficiently short to completely rule out the potential contribution of rehabilitation therapy to the observed improvements in dysphagia. Second, due to the absence of a parallel control group, our pre–post self-controlled design cannot fully rule out spontaneous recovery as a partial explanation for the observed findings. We are currently conducting a prospective study to obtain more robust evidence. Third, all included patients had neurogenic disease with established dysphagia, and our data suggest that the presence of a tracheostomy tube aggravated their swallowing impairment. Because of the limited sample size in the nonstroke subgroup, we were unable to perform adequately powered subgroup analyses comparing stroke and nonstroke populations, although the key outcomes were consistent in the stroke subgroup. It remains uncertain whether the effect of tracheostomy on dysphagia differs across specific disease etiologies; therefore, to clarify disease-specific effects, we plan to expand future research to include additional conditions, such as head and neck cancers.

6. Conclusion

In patients with neurogenic dysphagia undergoing tracheostomy, the presence of a tracheostomy tube can reduce laryngeal sensation, impair clearance, and increase the burden of secretions. Moreover, it can exacerbate dysphagia, resulting in higher levels of pharyngeal residue and aspiration. In such patients, clinicians should consider decannulation, when clinically feasible and guided by objective criteria, as a priority in the overall management strategy rather than delay the removal of the tracheostomy tube until after dysphagia treatment.