We identified 127 patients with PSP admitted to the Department of Neurology at Fukuoka University Hospital and NHO Omuta Hospital between January 2009 and April 2023. Of these, we selected 56 (mean age at the disease onset, 67.6 ± 6.4 years old; 35 men, 21 women) for inclusion in the study based on the Movement Disorder Society (MDS)-PSP criteria (17) which subdivided them into PSP with Richardson’s syndrome (PSP-RS) and other PSP phenotypes (Table 1). Patients with cerebral infarction, arteriosclerotic Parkinsonism, Lewy body disease, multiple system atrophy, and corticobasal degeneration were excluded. Patients receiving end-of-life treatment for cancer, respiratory disease, or gastrointestinal disease were excluded.

Based on the medical records of both hospitals, information about the basic attributes of each patient, including the presence or absence of dementia, food intake status, VFSS results, timing of nasogastric tube transition, and the time of gastrostomy were extracted. We defined the onset age as the age when the first symptom attributable to PSP was reported (18) (Figure 1). Regarding endpoints, the primary end point was set as the time of the definitive diagnosis, and the terminal endpoint was set as the time until discontinuation of oral intake; patients who were transferred to other hospitals or who did not see a doctor during the study period were censored at the date of their final visit. We determined the presence of cognitive decline at the diagnosis using the Mini Mental State Examination (MMSE) or Revised Hasegawa’s Dementia Scale (HDS-R). A score of < 27 on the MMSE and < 20 on the HDS-R was considered to indicate cognitive decline. The presence or absence of dysarthria at the diagnosis was determined based on their physician’s auditory evaluation in terms of articulation, fluency, and pronunciation.

This study was approved by the Ethics Committee of Fukuoka University Hospital (approval number: H21-10-001). Informed consent was obtained in the form of an opt-out provision on the study website.

For the VFSS, patients were asked to adopt the posture they used when eating meals, and fluoroscopy was performed from the side. A two-fold dilution of 110% (w/v) liquid barium was injected into the patient’s oral cavity, after which patients were signaled to start swallowing. The swallowing movements were then recorded on a DVD at 30 frames/s, and the presence of aspiration was evaluated by an assessor after the test. Patients in whom dysphagia had already been noted were first tested with 5 mL of thickened liquid and then with 5 mL of thin liquid if no aspiration was noted. Aspiration was defined as present when the liquid from the VFSS penetrated past a patient’s vocal cords into the trachea (A [+]). For patients without aspiration (A [−]), those in whom the liquid penetrated the larynx but did not reach the trachea were categorized as having laryngeal penetration (P [+]), while those without laryngeal penetration were categorized as (P [−]). A follow-up study was performed after the initial VFSS through April 2023.

Each VFSS image frame was saved in bitmap format and assessed visually. The angle changes of C1-C3 were measured in the pharyngeal phase (Figure 2), as the angles of the cervical vertebrae C1–C7 were flexed in that phase (19). Angle changes under 0.5° were taken to indicate dystonic rigidity of the neck.

The diet type was classified by a multidisciplinary team into seven levels based on the Functional Oral Intake Scale (FOIS) (20). Level 0 was tube-dependent, Level 1 could receive no oral intake, Level 2 was tube-dependent with minimal/inconsistent oral intake, Level 3 could receive consistent oral intake with some tube supplementation, Level 4 could receive total oral intake of a single consistency, Level 5 could receive total oral intake of multiple consistencies, requiring special preparation, Level 6 could receive total oral intake with no special preparation but with specific foods or liquid items avoided, and Level 7 could receive total oral intake with no restrictions.

We compared clinical details, including the age at the disease onset and at death as well as the disease duration, between patient subgroups. Only patients in whom an event occurred were included in determinations of the mean results and comparisons for each milestone of advanced disease. We performed univariable analyses with Fisher’s exact test for categorical variables and a two-tailed t-test or the Mann–Whitney U test, as appropriate, for continuous variables. The interval from the onset of the first symptom to each clinical milestone or death was graphically presented using Kaplan–Meier curves, with the curves from each patient subgroup compared using a log rank test. A Cox proportional-hazards model was used to estimate the hazard ratio and 95% confidence interval.

All statistical analyses were performed using the IBM SPSS Statistics software program for Windows (version 27.0; IBM Japan, Tokyo, Japan). For all tests, p values of <0.05 (two-tailed) were considered statistically significant.

Our results showed that tube feeding was initiated earlier in patients with cognitive decline at the diagnosis than in those without such a decline. Cognitive impairment, including attention deficit, delayed information processing, difficulty in problem solving, and impaired decision making, is found at a high rate in PSP patients (21), typically progressing along with other symptoms of PSP. Cognitive decline can hamper controlling one’s intake of food and liquids, thereby increasing the risk of dysphagia (5). Patients may find it difficult to make appropriate food choices, properly prepare meals, and adjust the amount of food they consume, all of which can lead to the development of pneumonia and prevent continuation of oral intake.

Our results further showed that tube feeding was also initiated earlier in patients who showed aspiration on the initial VFSS than in those without aspiration. The function of muscles involved in swallowing will deteriorate and lead to dysphagia with the progression of PSP (8). This makes it difficult to continue taking food and liquids by mouth, increasing the risk of choking and aspiration. Reduced reflexes and coordination in swallowing result in longer mealtimes and a decrease in the amount of intake, which makes nutritional management difficult (8). The rate of disease progression varies among PSP subtypes, but tube feeding typically becomes necessary within roughly 28 months after the initial VFSS or significantly earlier in patients with aspiration.

The present study showed that the early initiation of tube feeding was associated with axial rigidity at the diagnosis. With the progression of PSP, the function of muscles involved in posture control deteriorates, leading to increased tension in the central axis (trunk) of the body. This makes it difficult to maintain an upright posture, resulting in forward or backward flexion (22). Progression of axial rigidity will cause stiffness and reduced control of the neck muscles. Additionally, there is a correlation between axial rigidity and retrocollis. The progression to retrocollis impairs the accuracy and efficiency of swallowing, leading to dysphagia, and makes it difficult for boluses to pass through the pharynx, in spite of any compensatory approach, modification of the consistency of food, fluids, and meal posture (23).

While axial rigidity is a major criterion for diagnosing PSP, being reported in more than half of such patients, retrocollis is not necessarily a common feature in PSP (24). Indeed, neck stiffness has consistently been reported at a might greater prevalence than a feeling of limb stiffness (16). The degree of neck rigidity was not associated with trunk and limb rigidity; however, the present study showed an association between retrocollis and axial rigidity, which leads to the discontinuation of oral intake.

A number of potential survival predictors have been proposed (25), including age at the onset (1, 18, 26, 27), early falls (26, 28), dementia (1, 28), bulbar symptoms (26), degree of disability (29), clinical phenotypes, and a short time to disability (29). Seven milestones of disease progression have been proposed: frequent falling, cognitive disability, unintelligible speech, severe dysphagia, dependence on wheelchair for mobility, use of urinary catheters, and placement in residential or nursing home care (30). However, the need for feeding tube placement is uncommon as the first event and should not be a major focus of monitoring (31). The symptoms most frequently reported at baseline were found to be motor, cognitive/behavioral, systemic and bulbar, and sleep disturbance; furthermore, neck stiffness was highly prevalent at the baseline, whereas gait impairment was relatively infrequent (16). While there does seem to be some association between certain symptoms, including dysphagia, and an increased risk of death in PSP patients, the precise symptoms that predict fatal outcomes remain controversial. As the modified PSP rating scale showed as well as this study, neck rigidity or dystonia, which can be a cause of swallowing impairment, are cited as critical factors determining prognosis (32).

This present study was associated with some limitations. First, we could only determine the PSP subtype in 33 of the 56 participants because of the unfamiliar diagnostic criteria at the start of the study. However, these subtypes have a huge impact on the initial symptoms; for example, in PSP-P, axial rigidity is rarer than limb rigidity. The sample sizes were too small to perform an analysis according to the subtypes. Further follow-up studies with a larger population with categorization by subtype are needed. Second, we used MMSE and HDS-R to determine the presence of cognitive decline at the diagnosis. However, these two tools are merely screening methods for evaluating cognitive dysfunction and are not always suitable for evaluating cognitive dysfunction in PSP, especially to differentiate PSP-CBS from PSP-RS. As a practical matter, they are widely accepted by physicians in daily clinical practice for PSP as a screening tool to recognize moderate cognitive impairment despite variations in sensitivity and specificity (33, 34); we therefore adopted the test results. Third, in this study, we could not track some participants until the terminal phase because of difficulties keeping in contact with them or their family due to changing hospitals or switching to residential or nursing home care. The number of aborted cases during follow-up could influence the presentation of the Kaplan–Meier curves. In the future, a new tracking system should be introduced for patients in the terminal phase of PSP. Despite these limitations, these results will contribute to the understanding of the disease and the improvement of treatment strategies. Understanding the predictive factors and complications associated with dysphagia progression will aid in managing and caring for patients.