Patients recovering from moderate-to-severe coro-navirus disease 2019 (COVID-19) reportedly have dys-phagia or difficulty swallowing¹. However, specific info-rmation related to the etiology and nature of post- COVID-19 dysphagia has not been confirmed². App-roximately 17-30% of patients with COVID-19 develop Acute Respiratory Distress Syndrome (ARDS), requiring intensive care unit (ICU) admission and invasive venti-lation in severe cases³. In addition, 3-62% of patients experience dysphagia after critical illness, which pro-longs hospitalization and is associated with the possi-bility of reintubation, pneumonia, and a high risk of death¹. Therefore, dysphagia is a major problem in patients with COVID-19.

Aspiration pneumonia is caused by the aspiration of different materials, such as oropharyngeal secretions or gastric contents, into the airway⁴. Dysphagia is associated with compromised patient outcomes including malnutrition, aspiration pneumonia, increased length of hospital stay, and higher mortality^5-7. Although COVID-19 has been associated with dysphagia, the degree or severity of dysphagia due to COVID-19 com-pared to aspiration pneumonia has not been evaluated, and factors affecting the severity of dysphagia have rarely been evaluated. The videofluoroscopic swallowing study (VFSS) is the gold standard for assessing dys-phagia, and it shows the functional anatomy and phy-siology of swallowing⁸. It assesses the oral and phary-ngeal phases of swallowing and detects intradeglu-titive silent aspiration as well^8,9. This study aimed to compare the differences in dysphagia characteristics and severity between patients with COVID-19 pneu-monia and aspiration pneumonia, and to identify factors affecting the difference in severity between the two groups.

This retrospective study was conducted at XX University Hospital. We reviewed the electronic medical records of 54 patients diagnosed with COVID-19 pneumonia who subsequently developed dysphagia, and were re-ferred for a VFSS because of signs and symptoms of dysphagia during admission. Referrals for VFSS were determined by a history suggestive of aspiration, or when dysphagia was suspected on bedside swallow test¹⁰. COVID-19 pneumonia was defined as the presence of respiratory symptoms or fever, polymerase chain rea-ction positivity for COVID-19, and chest computed tomography (CT) findings corresponding with alveolar-interstitial pneumonia. For comparison, we reviewed the electronic medical records of 44 patients with aspi-ration pneumonia diagnoses who were referred for a VFSS because of dysphagia symptoms. Aspiration pneu-monia was defined as respiratory symptoms or fever, a specific history of aspiration, or chest CT findings corresponding with alveolar-interstitial pneumonia. This study investigated the medical records of the enrolled patients between April 2020 and May 2022. Demographic characteristics, dysphagia-related comor-bidities and clinical variables, and results of the VFSS, including the modified videofluoroscopic dysphagia scale (mVDS), American Speech-Language-Hearing Ass-ociation National Outcomes Measurement System (ASHA-NOMS), and penetration aspiration scale (PAS), were extracted from the electronic medical records. The requirement for informed consent was waived because of the retrospective nature of the study, and the study was approved by the XX University Hospital Research Ethics Committee (IRB No: XX 2022-05-037).

1. Patients’ dysphagia-related comorbidities and dysphagia-related clinical variables

Patients’ with a history of dysphagia-related como-rbidities or risk factors for dysphagia, such as stroke, parkinsonism, dementia, head and neck cancer, muscle diseases such as myopathy and dystrophies, and chronic obstructive pulmonary disease (COPD), were investigated. Dysphagia-related clinical variables such as ICU admi-ssion, intubation, tracheostomy, ARDS diagnosis after admission, lengths of hospital stay and ICU stay were evaluated. In addition, the occurrence of stroke, ence-phalitis, Guillain-Barré syndrome, critical illness poly-neuropathy, or myopathy after admission, which could induce dysphagia, was evaluated by checking electronic medical records.

2. VFSS, mVDS, PAS, and ASHA-NOMS

The VFSS was conducted according to the Logemann protocol by a rehabilitation medicine physician and an occupational therapist¹¹. The other physician, the researcher, simultaneously observed and scored the test. The researcher was blinded to the patient infor-mation. The VFSS was performed using a fluoroscopy device (GEMSS Spinel 3G) and an MM-190 Meditech monitor, and recorded as a video file (30 frames/s). Initially, the patients were seated on a chair in an up-right position. Subsequently, the study was performed with the patients on the lateral side. During the VFSS, patients swallowed food with a stepwise increase in consistency from liquid to solid material¹¹. The study was terminated when massive aspirations were observed. According to the World Health Organization recom-mendation, all studies were conducted with the exa-miner wearing four items of personal protective equi-pment (facial shield, mask, gown, and gloves)¹². The mVDS was developed to interpret the VFSS findings. It has nine categories that represent oral (lip closure, mastication, and oral transit time) and pharyngeal (pha-ryngeal triggering, epiglottis inversion, pharyngeal coa-ting, vallecular residues, pyriform sinus residues, and tracheal aspiration) function. The mVDS with 100 points was created according to the odds ratios of various prognostic factors to measure objective dysphagia quan-titative scores with VFSS results¹³. The PAS is an 8- point scale used to describe and assess both the seve-rity of airway invasion events and patients’ response during the VFSS⁹. The ASHA-NOMS determines the functional level of swallowing¹⁴. This scale assesses the supervision required for feeding and the appropriate type of diet, with scores ranging from 1 to 7. In the present study, the ASHA-NOMS level was determined after completion of the VFSS by a physiatrist.

3. Statistical analysis

A Chi-squared (χ²) test was used for categorical variables, and an independent t- or Mann–Whitney test was used for continuous variables to compare the demographics, comorbidities, and clinical variables related to dysphagia between the COVID-19 pneu-monia and aspiration pneumonia groups. We examined the associations between demographics, comorbidities, clinical variables, the presence of COVID-19, and tra-cheal aspiration scores of the mVDS using logistic regression. Pearson’s correlation was used to evaluate the correlation between clinical variables such as lengths of hospital stay, ICU stay and the total mVDS score. A chi-square (χ²) test was used to determine the proportion of PAS between the two groups. We consi-dered a P-value <0.05 as significant. All analyses were performed using SPSS 22.0 (SPSS Inc., Chicago, IL, USA).

In total, 98 patients were included in this study. Fifty- four patients were included in the COVID-19 pneu-monia group (mean age, 72.57±14.78 y; 40 identified as men and 14 as women), and 44 patients in the aspi-ration pneumonia group (mean age, 79.25±8.20 y; 30 identified as men and 14 as women).(Table 1) All patients underwent the VFSS because of their dysphagia symptoms. In the COVID-19 pneumonia group, the period from isolation release to the VFSS was 12.52± 10.89. The incidence of dysphagia-related comorbidities, such as stroke, parkinsonism, dementia, head and neck cancer, muscle diseases such as myopathy and dystrophies, and COPD, were not significantly different between the two groups.(Table 2) ICU admission, intu-bation, tracheostomy, and ARDS diagnosis after admi-ssion were more common in the COVID-19 pneumonia group (P<0.001 for other variables, and P=0.007 for tracheostomy) than in the aspiration pneumonia group. The lengths of hospital stay and ICU stay were significantly increased in the COVID-19 pneumonia group (P<0.001) than in the aspiration pneumonia group.(Table 2) In both groups, stroke, encephalitis, Guillain-Barré syndrome, and critical illness, polyneu-ropathy or myopathy did not occur after admission. The total mVDS score was significantly higher, indicating more severe dysphagia, in the COVID-19 pneumonia group (P=0.038) than in the aspiration pneumonia group. Among the mVDS subscores, tracheal aspiration was significantly higher (P<0.001) and oral transit time tended to be longer in the COVID-19 pneumonia group (P=0.069).(Table 3) In the logistic regression analysis, age (P=0.034), COVID-19 (P<0.001), ICU admission (P= 0.012), tracheostomy (P=0.029) and ARDS diagnosis after admission (P=0.036) were significantly associated with tracheal aspiration. After adjusting for age, sex, comorbidities, and clinical variables, COVID19 conti-nued to be significantly associated with worse tra-cheal aspiration scores (P=0.04).(Table 4) Functional dysphagia levels according to the ASHA-NOMS were similar in both groups (χ²=5.341, P=0.501).(Table 5) Among the clinical variables, only the length of stay was significantly correlated with the total mVDS score (r=0.241, P<0.05).(Table 6) When the proportion of PAS was compared between the two groups, the PAS score was significantly higher in the COVID-19 pneu-monia group (P<0.001) (Table 7) than in the aspira-tion pneumonia group.

Table 1 . Demographics and treatments of patients with COVID-19 pneumonia.

	Total	COVID-19 pneumonia with dysphagia, N or mean (% or SD)	Aspiration pneumonia, N or mean (% or SD)	χ2/t (P-value)
Age (years)
<65	16	14 (25.9)	2 (4.50)	8.113 (0.004)
≥5	82	40 (74.1)	42 (95.5)
Sex
Male	70	40 (74.1)	30 (68.2)	0.412 (0.521)
Female	28	14 (25.9)	14 (31.8)
Steroid
Yes	45	45 (83.3)
No	9	9 (16.7)
Remdesivir
Yes	17	17 (31.5)
No	37	37 (68.5)
Sepsis
Yes	10	10 (18.5)
No	44	44 (81.5)

COVID-19: coronavirus disease 2019, SD: standard deviation.

Table 2 . Comorbidities and clinical characteristics of patients according to pneumonia type.

	Total	COVID-19 pneumonia with dysphagia, N or mean (% or SD)	Aspiration pneumonia, N or mean (% or SD)	χ2/t (P-value)
Comorbidities
Stroke
Yes	30	16 (29.6)	14 (31.8)	0.055 (0.815)
No	68	38 (70.4)	30 (68.2)
Parkinsonism
Yes	10	4 (7.40)	6 (13.6)	1.027 (0.311)
No	88	54 (92.6)	38 (86.4)
Dementia
Yes	22	11 (20.4)	11 (25.0)	0.298 (0.585)
No	76	43 (79.6)	33 (75.0)
Muscle diseases
Yes	2	1 (1.90)	1 (2.30)	0.021 (0.883)
No	96	53 (98.1)	43 (97.7)
Head & neck cancer
Yes	1	0 (0)	1 (2.30)	1.240 (0.265)
No	97	54 (100)	43 (97.7)
COPD
Yes	6	1 (1.9)	5 (11.4)	3.812 (0.051)
No	92	53 (98.1)	39 (88.6)
Clinical characteristics
ICU admission
Yes	16	14 (25.9)	2 (4.50)	8.113 (<0.001)
No	82	40 (74.1)	42 (95.5)
Intubation
Yes	40	31 (57.4)	9 (20.5)	13.705 (<0.001)
No	58	23 (42.6)	35 (79.5)
Tracheostomy
Yes	21	17 (31.5)	4 (9.10)	7.219 (0.007)
No	77	37 (68.5)	40 (90.9)
ARDS
Yes	29	25 (46.3)	4 (9.10)	16.108 (<0.001)
No	69	29 (53.7)	40 (90.9)
Period from isolation release to VFSS		12.52±10.89
Length of hospital day		58.07 (36.71)	31.97 (25.33)	<0.001
Length of ICU stay		23.67 (24.68)	6.77 (12.77)	<0.001

ARDS: acute respiratory distress syndrome, COPD: chronic obstructive pulmonary disease, COVID-19: coronavirus disease 2019, VFSS: videofluoroscopic swallowing study, ICU: intensive care unit, SD: standard deviation.

Table 3 . Modified videofluoroscopic dysphagia scale.

	COVID-19 pneumonia with dysphagia,mean (SD)	Aspiration pneumonia, mean (SD)	P-value
Total mVDS	51.36 (17.32)	42.56 (24.02)	0.038
Lip closure	0.24 (1.15)	0.00 (0.00)	0.114
Mastication	4.70 (5.69)	4.18 (5.60)	0.579
Oral transition time	1.18 (1.85)	0.56 (1.37)	0.069
Pharyngeal triggering	3.67 (3.50)	3.82 (3.53)	0.856
Epiglottis inversion	8.93 (6.07)	7.68 (6.47)	0.299
Pharyngeal coating	8.93 (6.07)	8.57 (6.23)	0.926
Vallecular residues	3.81 (1.98)	3.81 (2.84)	0.845
Pyriform sinus residues	6.42 (4.59)	6.35 (5.34)	0.736
Tracheal aspiration	12.79 (6.27)	6.95 (6.68)	<0.001

COVID-19: coronavirus disease 2019, mVDS: modified Video-fluoroscopic Dysphagia Scale, SD: standard deviation.

Table 4 . Logistic regression analysis of the association between clinical variables and tracheal aspiration (normal vs abnormal).

	Crude			Adjusteda
	β (SE)	P-value		β (SE)	P-value
Sex
Male	1.39 (1.58)	0.332		1.89 (1.66)	0.206
Female	0.00			0.00
Age (years)
<65	0.00			0.00
≥65	−3.72 (1.90)	0.034		−0.003 (0.06)	0.254
Stroke
Yes	−2.40 (1.54)	0.239		−2.91 (1.56)	0.056
No	0.00			0.00
Parkinsonism
Yes	0.04 (2.37)	0.527		−0.08 (2.39)	0.602
No	0.00			0.00
Dementia
Yes	0.46 (1.72)	0.878		0.87 (1.82)	0.513
No	0.00			0.00
Muscle diseases
Yes	6.97 (5.02)	0.366		2.48 (6.99)	0.761
No	0.00			0.00
Head and neck cancer
Yes	6.90 (7.09)	0.525		7.36 (10.67)	0.994
No	0.00			0.00
COPD
Yes	−0.27 (2.99)	0.790		0.09 (3.16)	0.870
No	0.00			0.00
ICU admission
Yes	3.70 (1.43)	0.012		3.62 (2.00)	0.066
No	0.00			0.00
Intubation
Yes	2.21 (1.44)	0.119		−4.07 (2.38)	0.186
No	0.00			0.00
Tracheostomy
Yes	3.03 (1.72)	0.029		1.99 (2.07)	0.121
No	0.00			0.00
ARDS
Yes	3.46 (1.53)	0.036		0.16 (2.24)	0.895
No	0.00			0.00
COVID-19
Yes	5.83 (1.31)	0.001		5.42 (1.60)	0.042
No	0.00			0.00

ARDS: acute respiratory distress syndrome, COPD: chronic obstructive pulmonary disease, COVID-19: coronavirus disease 2019, ICU: intensive care unit, SE: standard error.

^aAdjusted for all demographic covariates.

Table 5 . Functional swallowing status and diet management according to the ASHA-NOMS.

ASHA-NOMS	COVID-19 pneumonia with dysphagia,N or (%)	Aspiration pneumonia, N or (%)
Level 1		1 (2.3)
Level 2	27 (50)	25 (56.8)
Level 3	1 (1.9)	1 (2.3)
Level 4	5 (9.3)	1 (2.3)
Level 5	8 (14.8)	4 (9.1)
Level 6	6 (11.1)	8 (18.2)
Level 7	3 (5.6)	4 (9.1)

COVID-19: coronavirus disease 2019, ASHA-NOMS: American Speech-Language-Hearing Association National Outcomes Measurement System.

Table 6 . Functional swallowing status and diet management acco-rding to the ASHA-NOMS.

		Length of hospital stay	Length of ICU stay
mVDS	Pearson correlation coefficient	0.241*	0.144

mVDS: modified Videofluoroscopic Dysphagia Scale, ICU: intensive care unit.

*Correlation significant at P=0.05.

Table 7 . Proportion of PAS in the two groups.

	Total	COVID-19 pneumonia with dysphagia, N (%)	Aspiration pneumonia, N (%)	χ2/t (P-value)
PAS				17.830 (0.000)
None	24	6 (11.1%)	18 (40.9%)
Penetration	28	13 (24.1%)	15 (34.1%)
Aspiration	46	35 (64.8%)	11 (25.0%)

PAS: Penetration Aspiration Scale, COVID-19: coronavirus disease 2019.

This is the first study to compare the degree of dysphagia and related clinical factors between COVID-19 pneumonia and aspiration pneumonia. Our study shows that COVID-19 pneumonia with dysphagia has a more severe clinical presentation with more ICU admissions, intubations, tracheostomies, and incidence of ARDS than other types of aspiration pneumonia. The duration of hospital stay was longer in patients with COVID-19 pneumonia than in those with aspi-ration pneumonia. Dysphagia was more severe in pa-tients with COVID-19 pneumonia than in those with aspiration pneumonia, which was confirmed by the VFSS and quantitative mVDS scores. Among the mVDS subscores, tracheal aspiration was significantly worse in the COVID-19 pneumonia group than in the aspira-tion pneumonia group.

This is the first study to objectively compare dys-phagia in patients with COVID-19 pneumonia and aspiration pneumonia using a quantitative mVDS score. Webler et al.¹⁵ reported increased tracheal aspiration, which was identified on VFSS in the COVID-19 group, and Lagier et al.¹⁶ showed a high prevalence of silent tracheal aspiration (80%) in patients with COVID-19 who recovered from ARDS. Although studies have shown that COVID-19 influences dysphagia, only these two studies confirmed tracheal aspiration using an objective method such as VFSS. The results of our study are consistent with those of previous studies. Tracheal aspiration is a critical factor for aspiration pneumonia and other complications. Interestingly, after adjusting for age, sex, medical comorbidities such as stroke, parkinsonism, and dementia, and clinical varia-bles such as ICU admission, intubation, tracheostomy, and ARDS, on its own, COVID-19 significantly influe-nced tracheal aspiration. Among the mVDS subscores, only tracheal aspiration showed a difference between the two groups. The reason could be that other factors not covered by the mVDS can cause aspiration. Unlike the videofluoroscopic dysphagia scale (VDS), the mVDS does not deal with bolus formation, tongue-to- palate contact, and premature bolus loss¹⁷. These factors are related to the function of the oral phase or the pre-swallowing aspiration. The mVDS does not address these factors, making it difficult to assess the various causes of aspiration. Therefore, this could a reason why only tracheal aspiration to showed a difference be-tween the two groups in the analysis using the mVDS.

The mechanism of dysphagia in patients with COVID-19 is still under investigation. However, COVID-19 is suspected to induce more severe inflammation, thro-mbosis, and cytokine storm than other respiratory infections^18-20. This could result in multiorgan failure and peripheral nerve- or organ-specific changes in the swallowing system. In patients with COVID-19, where the lungs are severely injured, identifying dyspha-gia to prevent disease progression is especially cri-tical^1,16,21,22. According to our study and previous studies^16,23, a dysphagia assessment tool that can eva-luate tracheal aspiration is very important. We reco-mmend referring patients at high risk of dysphagia for formal instrumental swallowing tests to assess the risk of aspiration, but only while following all safety precautions, and when the patient is no longer con-tagious²⁴.

According to recent publications, approximately 5% of patients with COVID-19 require ICU treatment, with a high proportion of patients requiring prolonged mechanical ventilation owing to ARDS²¹. In addition to these conditions, which have been identified as key risk factors for developing critical illness, poly-neuropathy and myopathy, other neurological compli-cations such as encephalitis, stroke, and Guillain- Barré syndrome have also been reported in COVID- 19²⁵. All of the above-mentioned complications of COVID-19 can induce dysphagia. In this study, none of the patients in either group developed these com-plications after admission. In addition, a previous history of stroke, dementia, Parkinsonism, muscle dise-ases, COPD, and head and neck cancer can influence dysphagia. However, the incidence of these diseases was similar in both groups.(Table 2)

Patients with COVID-19 who develop ARDS must be transferred to the ICU and intubated. ICU admission is also a risk factor for dysphagia²⁶. Intubation and extended mechanical ventilation are the most impor-tant independent risk factors for dysphagia in patients with COVID-19^1,27. Mucosal injuries that occur due to endotracheal intubation, respiratory and swallowing muscle atrophy, and sensory loss cause silent micro-aspiration. Studies conducted in ICUs report that 83%-94% of laryngeal injuries are observed after an average of 1 week of intubation and that half of the patients develop dysphagia^28,29. Dysphagia is present in >50% of patients intubated for >48 h²². Moderate-to- severe dysphagia was observed in patients intubated for >7 days²². Prolonged intubation may reduce lary-ngeal sensation, resulting in silent aspiration³⁰. Similarly, tracheostomy increases the risk of dysphagia and aspiration and leads to laryngeal and pharyngeal dys-function during swallowing³¹. In addition, Goff et al.³¹ reported that ICU admission, intubation, ICU admission length, tracheostomy, and the incidence of ARDS were higher in patients with COVID19 pneumonia than in those with aspiration pneumonia. Therefore, all of the aforementioned factors could act simultaneously, resulting in more severe dysphagia in the COVID-19 pneumonia group.

Presbyphagia is a change in the swallowing mecha-nism in healthy older adults during normal aging²³. In this study, the COVID-19 pneumonia group had younger patients than the aspiration pneumonia group but showed more severe dysphagia. Although aspiration increases with age, COVID-19 can cause tracheal aspi-ration even in younger patients.

After the VFSS, the type of diet and degree of supe-rvision were determined using the ASHA-NOMS. Both groups were most commonly classified as level 2, which means that oral feeding was not possible, and only oral trials by the therapist in the treatment room were possible. Although tracheal aspiration was more severe in the COVID-19 pneumonia group, the percentage of each ASHA-NOMS level was not significantly diffe-rent between groups. This is because diet is not solely determined by the presence of aspiration. Physiatrists who decide on a diet should consider the host defense mechanism, the patient’s health condition, cognition, respiratory function, oral hygiene, and coughing abi-lity³².

This study has some limitations. First, it was descri-ptive and based on a retrospective design; owing to this method, there may be some bias in patient recruit-ment. It was difficult to evaluate COVID-19 pneumo-nia patients who were not referred. Some of them may have dysphagia, but were excluded from study. However, we attempted to select an appropriate com-parator group and applied stringent selection criteria. We also reviewed previous and in-hospital medical histories that could influence the patients’ dysphagia symptoms and compared them between the groups. Second, although factors that may affect dysphagia have been identified through medical records, electrical diagnosis of diseases such as myopathy, critical illness polyneuropathy, and Guillain-Barré syndrome that may occur after COVID 19 infection was not condu-cted. Therefore, it is difficult to conclude that these diseases were completely ruled out. Third, the number of patients was not very high; however, the patient profiles were homogeneous, similar to the VFSS findings. Fourth, a follow-up study was not conducted after the first VFSS. The prognosis of both groups needs to be evaluated in future studies to better understand dysp-hagia characteristics.

Patients with COVID-19 pneumonia with dysphagia symptoms or signs present with more severe clinical characteristics such as ICU admission, intubation, tra-cheostomy, ARDS, prolonged hospital stay, and ICU stay. They also showed more severe dysphagia, particularly tracheal aspiration confirmed by the VFSS, than those with aspiration pneumonia. Aspiration can worsen the already reduced lung function. Therefore, it seems better to evaluate dysphagia using VFSS, which could objec-tively reveal tracheal aspiration in patients with COVID- 19 with dysphagia.

This study was supported by the Inha University research grant.