Introduction

Spinocerebellar degeneration (SCD) is a disorder primarily characterized by progressive cerebellar degeneration, resulting in impairments in coordination and balance [1]. SCD can be classified into hereditary (hSCD) and sporadic forms [2]. Among the sporadic forms, multiple system atrophy-cerebellar type (MSA-C) is noteworthy for its rapid progression compared with the hereditary forms of SCD (hSCD) [3]. Recent studies have indicated that patients with MSA-C experience faster disease progression and shorter survival. For example, the 10-year survival rate for patients with spinocerebellar ataxia type 6 (SCA6), a type of hSCD with relatively slow progression, is > 80%. However, the average lifespan of patients with MSA-C post-diagnosis is approximately 10 years [4, 5].

Notably, patients with hSCD and MSA-C present with cerebellar ataxia as a primary symptom; however, it is crucial to note that they are fundamentally different conditions. MSA-C is a rapidly progressive neurodegenerative disorder affecting multiple systems, including the autonomic nervous system, basal ganglia, and the cerebellum [3]. Therefore, the shorter survival time in patients with MSA-C is primarily due to these extensive systemic effects, particularly autonomic dysfunction, rather than cerebellar symptoms alone. However, many forms of hSCD, such as SCA6, primarily affect the cerebellum with limited involvement of other systems [6]. This fundamental difference in pathology may influence disease progression and response to rehabilitation interventions.

MSA-C and hSCD distinctly differ in disease progression; therefore, these variations may influence the effectiveness and sustainability of rehabilitation interventions. Rehabilitation programs, including vibration stimulation, aerobic exercises, and Tai Chi, have been widely studied and implemented as therapeutic approaches for patients with SCD [7, 8]. Intensive rehabilitation programs spanning 2–4 weeks have demonstrated significant benefits in several studies [9,10,11]. However, further research is required to compare the immediate and long-term effects of these interventions in patients with hSCD and MSA-C.

Considering the progressive nature of SCD, the long-term sustainability of rehabilitation programs is clinically important. Previous studies have reported mixed outcomes regarding the durability of rehabilitation effects across various SCD types. For example, Miyai et al. found that patients with SCA6 and idiopathic cerebellar ataxia who underwent intensive rehabilitation returned to baseline levels of function within 6 months [10]. Conversely, other studies have shown that continued rehabilitation can sustain benefits over an extended period, especially in specific types of hSCD, such as SCA7 and SCA2 [9, 12].

However, despite the documented benefits of rehabilitation, comprehensive studies that directly compare rehabilitation outcomes between patients with hSCD and those with MSA-C are lacking. This comparison is crucial for developing tailored intervention strategies and determining the appropriate follow-up frequency for each disease subtype.

Therefore, in this study, we aimed to investigate the effects and sustainability of intensive rehabilitation programs on balance function in patients with hSCD and MSA-C. We assessed the immediate and 6-month post-intervention outcomes using standardized measures of ataxia and balance function. By comparing hSCD and MSA-C, we aimed to elucidate how different cerebellar pathologies influence rehabilitation outcomes and their sustainability, considering varying disease progression rates and potential differences in neuroplasticity. This approach enables us to develop more effective, tailored rehabilitation strategies for each patient group and optimize intervention protocols. Therefore, in this comparative study, we seek to enhance our understanding of cerebellar ataxia rehabilitation, potentially improving patients’ functional outcomes and quality of life through more targeted and efficient approaches.

Patients and Methods

Participants

In this retrospective cohort study, we analyzed the electronic medical records of patients with SCD who participated in an intensive rehabilitation program at the National Center of Neurology and Psychiatry (Tokyo, Japan) between June 2015 and April 2021. The program at the center was limited to those who could walk independently, although the use of assistive devices was permitted. Therefore, participants could live independently or with minimal assistance at home. The study adhered to the principles of the Declaration of Helsinki and was approved by the Ethics Committee of the National Center of Neurology and Psychiatry (Approval No. A2018-104). Volunteers were given verbal explanations about the purpose, methods, safety considerations, and risks of the experiment and provided informed consent in advance.

Rehabilitation Training Procedure

The participants underwent a 4- or 2-week intensive rehabilitation program conducted five times weekly. Each daily schedule comprised 1 h of individual training with a physical therapist, 1 h with an occupational therapist or speech therapist, and 1 h of self-balance training. The personal training with the physical therapist focused on movement balance and muscle-strengthening exercises performed on a mat, as well as standing and walking without special equipment. This component adhered to the rehabilitation program of the Research Committee for Ataxia Disease (research team under the Ministry of Health, Labor, and Welfare in Japan; http://ataxia.umin.ne.jp/rehabilitation/). Occupational therapy sessions included manual dexterity training, while speech therapy sessions focused on improving articulation for patients with speech disorders. The self-balance training was mainly conducted on a mat and featured exercises that could be safely performed based on the individual physical therapy sessions, including kneeling and crawling.

Outcomes

Outcomes were assessed at three-time points: baseline (pre-intervention), end of the intervention (post-intervention), and 6-month follow-up. The primary outcome measure was the scale for the assessment and rating of ataxia (SARA) score, which quantifies the severity of cerebellar ataxia symptoms on a scale ranging from 0 to 40, with higher scores indicating more severe symptoms [13]. Based on previous research [14], the minimal detectable change (MDC) for SARA was established at 1.0 point.

Balance impairment in patients with SCD is multifactorial [15, 16]. Therefore, the balance evaluation systems test (BESTest) was used as a secondary outcome measure to comprehensively assess the balance function. The BESTest evaluates balance function across six systems, with a total score ranging from 0 to 108, where higher scores represent better balance capabilities. The six subcategories assessed were biomechanical constraints, stability limits/verticality, anticipatory postural adjustments, postural responses, sensory orientation, and stability in gait. Each subcategory independently evaluates critical aspects of the balance function [17]. The reliability and validity of the BESTest in individuals with SCD have been established [18], and the MDC for the BESTest was calculated to be 8.7 points.

To minimize evaluator bias, the individuals administering the training were different from those conducting the assessments. Seven evaluators who had undergone thorough training in the assessment methods performed the outcome measurements. Furthermore, to ensure consistency, the same evaluator conducted each patient’s assessments throughout the study period.

Statistics

All statistical analyses were performed using R (version 4.0.2) [19]. We initially examined the differences in demographic data between the hSCD and MSA-C groups using Pearson’s chi-squared test for categorical variables and Welch’s t-test for continuous variables. Furthermore, a three-way (2 × 2 × 3) repeated-measures analysis of variance (ANOVA) was performed to observe the main and interaction effects of two between-patient factors (diagnosis: hSCD vs. MSA-C, and intervention term: 2 weeks vs. 4 weeks) and one within-patient factor (time: pre vs. post vs. follow-up). Greenhouse–Geisser epsilon corrections were used as needed to meet the sphericity assumption. Post-hoc comparisons were conducted using Bonferroni significant difference tests, with family-wise alpha levels set at 0.05. The effect size was determined using generalized eta squared (η2G), and percentage changes with 95% confidence intervals were calculated. The Anovakun function in R was used for the analysis. All statistical analyses were performed using complete case analysis without missing data imputation. Statistical significance was set at p < 0.05.

Results

Total Patient Population

Notably, 160 patients were enrolled during the designated study period. For individuals who participated in the program multiple times, only data from their initial participation were considered. At this stage, data from 74 participants remained viable for further evaluation. Of these, 49 patients with complete pre- and post-intervention outcome data and 6-month follow-up information were included in the study. One patient was excluded because he sustained a fracture while participating in the program.

Consequently, the final cohort for analysis comprised 49 patients (Fig. 1). This group included 30 patients with hSCD, specifically 15 with SCA6, 10 with SCA31, and 5 with SCA3. Additionally, 19 patients were diagnosed with MSA-C. The baseline characteristics of both groups are detailed in Table 1, which shows no significant differences in sex, age, baseline SARA scores, or baseline BESTest scores.

Fig. 1
figure 1

Flowchart of the study population. NCNP, National Center of Neurology and Psychiatry; SCD, spinocerebellar degeneration

Full size image
Table 1 Comparison of group demographics
Full size table

We also calculated the prevalence of non-cerebellar symptoms in each group. In the MSA-C group, the prevalence of autonomic dysfunction and Parkinsonism was 32% and 21%, respectively. In the hSCD group, the prevalence of both autonomic dysfunction and Parkinsonism was 3%. The MSA-C group showed a higher prevalence of autonomic dysfunction than did the hSCD group. However, in all patients, the autonomic symptoms were mild and did not significantly impact daily activities. Similarly, regarding Parkinsonism, all affected individuals only exhibited mild rigidity upon examination, without gait freezing, bradykinesia, or akinesia severe enough to interfere with daily living.

The average disease duration from onset was 3 years and 10 years in the MSA-C and hSCD groups, respectively. The MSA-C group primarily consisted of patients in the early stages of the disease, predominantly exhibiting cerebellar symptoms.

During the 6-month follow-up period post-program completion, none of the 49 patients participated in alternative treatment programs or clinical trials of novel therapeutics.

Differences in SARA Score Changes between the hSCD and MSA-C Groups

The analysis indicated that SARA scores improved due to the intervention; however, the extent of improvement varied between disease types (hSCD or MSA-C). A three-way repeated ANOVA revealed a significant main effect of time (F(1, 89) = 15.012, p = 0.001, η²G = 0.039) (Table 2). This effect was moderated by a significant interaction between diagnosis and time (F(1, 89) = 7.266, p = 0.001, η²G = 0.019). Subsequent simple main effect tests showed a significant effect of disease type at the 6-month follow-up (F(1, 45) = 12.193, p = 0.001, η²G = 0.213), indicating differences in the temporal changes in SARA scores between the hSCD and MSA-C groups.

Table 2 ANOVA summary table for SARA and BESTest
Full size table

For the hSCD group (left panel of Fig. 2), post-hoc tests revealed a mean score improvement of 1.02 points after the intervention compared with pre-intervention (Bonferroni-corrected p-value = 0.024). However, at the 6-month follow-up, the mean score declined by 0.97 points from the post-intervention score (Bonferroni-corrected p-value = 0.028), resulting in a score that was essentially back to baseline levels, with a mean difference of only 0.04 points compared with pre-intervention (Bonferroni-corrected p-value = 0.921).

Fig. 2
figure 2

Time series boxplots of SARA and BESTest scores. Lower SARA scores indicate better patient condition, whereas higher BESTest scores denote improved balance function. Each dot represents individual patient data. Red lines indicate p values for the hSCD group after Bonferroni correction. Blue lines indicate p-values for the MSA-C group. SARA, scale for the assessment and rating of ataxia; BESTest, balance evaluation systems test; hSCD, hereditary spinocerebellar degeneration; MSA-C, multiple system atrophy-cerebellar type

Full size image

In contrast, the MSA-C group demonstrated a minor mean improvement of 0.53 points post-intervention compared with pre-intervention (Bonferroni-corrected p-value = 0.498). However, at the 6-month follow-up, their scores had declined by an average of 2.77 points from the post-intervention scores (Bonferroni-corrected p-value = 0.001) and by 3.3 points compared with the pre-intervention scores (Bonferroni-corrected p-value = 0.001). This result is illustrated in the right panel of Fig. 2.

Differences in BESTest Score Changes between the hSCD and MSA-C Groups

The analysis revealed that the intervention had different effects on the BESTest scores between the hSCD and MSA-C groups. A three-way repeated measures ANOVA showed a significant main effect of time (F(1, 71) = 33.982, p = 0.001, η²G = 0.0746) (Table 2). This effect was moderated by a significant interaction between diagnosis and time (F(1, 71) = 5.484, p = 0.01, η²G = 0.012). Additionally, a significant simple main effect of disease type was observed at the 6-month follow-up (F(1, 45) = 4.549, p = 0.038, η²G = 0.091), indicating differences in the temporal changes of BESTest scores between the hSCD and MSA-C groups.

Post-hoc tests for the hSCD group demonstrated a mean improvement of 9.17 points in BESTest scores after the intervention compared with pre-intervention (Bonferroni-corrected p-value = 0.001). However, at the 6-month follow-up, the scores declined by an average of 6.74 points from the post-intervention scores (Bonferroni-corrected p-value = 0.001). Despite this decline, the scores at the 6-month follow-up remained an average of 2.43 points higher than the pre-intervention scores (Bonferroni-corrected p-value = 0.05). This result is illustrated in the right panel of Fig. 2.

In contrast, the MSA-C group exhibited a mean improvement of 7.56 points in BESTest scores after the intervention compared with pre-intervention (Bonferroni-corrected p-value = 0.001). However, at the 6-month follow-up, the scores deteriorated by an average of 13.26 points from the post-intervention scores (Bonferroni-corrected p-value = 0.001) and by 5.71 points compared with pre-intervention scores, although this latter difference was not statistically significant (Bonferroni-corrected p-value = 0.114). This result is illustrated in the right panel of Fig. 2.

Subcategory-specific Intervention Outcomes in the BESTest Scores

The subcategory analysis of BESTest scores revealed distinct outcomes across different domains (Table 3). A significant difference was observed in the “Biomechanical Constraints” domain based on diagnosis (F(1, 45) = 11.592, p = 0.001, η² = 0.134). The hSCD group showed mean scores of 13.1 ± 1.9 (pre), 13.9 ± 1.4 (post), and 13.6 ± 1.6 (follow-up), whereas the MSA-C group had mean scores of 12.4 ± 1.5 (pre), 12.8 ± 1.3 (post), and 11.9 ± 2.2 (follow-up).

Table 3 ANOVA summary table for the BESTest subcategory
Full size table

A significant effect of time was found in the “Stability Limits/Verticality” domain (F(1, 71) = 6.298, p = 0.005, η² = 0.06). Scores improved by 0.96 points from pre- to post-intervention (Bonferroni-corrected p-value = 0.002). However, they declined by 1.1 points at the 6-month follow-up (Bonferroni-corrected p-value = 0.002), with no significant change from pre-intervention to the 6-month follow-up (Bonferroni-corrected p-value = 0.738).

In the “Anticipatory Postural Adjustment” domain, significant effects were observed for diagnosis (F(1, 45) = 6.524, p = 0.014, η² = 0.106) and time (F(1, 90) = 10.796, p = 0.001, η² = 0.04). The hSCD group scores were 11.1 ± 2.3 (pre), 12.4 ± 2.0 (post), and 11.2 ± 2.7 (follow-up), whereas the MSA-C group scores were 9.5 ± 3.2 (pre), 10.5 ± 2.3 (post), and 9.1 ± 2.9 (follow-up). Scores improved by 1.24 points from pre- to post-intervention (Bonferroni-corrected p-value = 0.003). However, they declined by 1.23 points at the 6-month follow-up (Bonferroni-corrected p-value = 0.002), with no significant change from pre-intervention to the 6-month follow-up (Bonferroni-corrected p-value = 0.984).

Furthermore, a significant effect of time was observed for the “Postural Response” domain (F(1, 79) = 12.783, p = 0.001, η² = 0.069). Scores improved by 2.16 points from pre- to post-intervention (Bonferroni-corrected p-value = 0.002). They declined by 3.07 points from post-intervention to the 6-month follow-up (Bonferroni-corrected p-value = 0.002), with no significant change from pre-intervention to the 6-month follow-up (Bonferroni-corrected p-value = 0.180).

A significant time effect was observed in the “Sensory Orientation” domain (F(1, 74) = 9.163, p = 0.0006, η² = 0.036). Scores improved by 1.41 points from pre- to post-intervention (Bonferroni-corrected p-value = 0.001) and worsened by 1.35 points from post-intervention to the 6-month follow-up (Bonferroni-corrected p-value = 0.001), with no significant change from pre-intervention to the 6-month follow-up (Bonferroni-corrected p-value = 0.880).

Finally, a significant effect of time was observed in the “Stability in Gait” domain (F(1, 78) = 12.510, p = 0.0001, η² = 0.031). Scores improved by 1.79 points from pre- to post-intervention (Bonferroni-corrected p-value = 0.0001). They declined by 2.40 points from post-intervention to the 6-month follow-up (Bonferroni-corrected p-value = 0.0001), with no significant change from pre-intervention to the 6-month follow-up (Bonferroni-corrected p-value = 0.259).

Discussion

To our knowledge, this is the first study to investigate the effects of intensive rehabilitation on hSCD and MSA-C using the SARA and BESTest scores, which are typically used outcome measures in previous research. The results reveal that the effects of rehabilitation became distinctly apparent during the follow-up period. A detailed discussion of these findings is presented below.

Differential Intervention Outcomes in the hSCD and MSA-C Groups

The hSCD group demonstrated an improvement of 1.02 points in SARA scores after the intervention (Fig. 2), whereas the MSA-C group showed a smaller improvement of 0.53 points. Only the hSCD group showed a statistically significant improvement. Additionally, the change in the hSCD group exceeded the MDC. Regarding the BESTest results, the hSCD group improved by 9.17 points, compared with a 7.56-point improvement in the MSA-C group. Notably, both groups showed statistically significant improvements; however, only the change in the hSCD group exceeded the MDC. These results indicate that intensive rehabilitation significantly improved ataxia symptoms and balance function in the hSCD group.

The MSA-C group showed no significant improvement in SARA scores or changes beyond the MDC in the BESTest scores. These results suggest potential differences in the learning effects of rehabilitation interventions for cerebellar ataxia symptoms between patients with hSCD and those with MSA-C. A study investigating the relationship between motor learning effects using upper limb reaching tasks and changes in brain gray matter volume in patients with hSCD found an increase in premotor cortex volume after training [20]. This change is considered compensatory remodeling associated with dysfunction in the cerebello-cerebral motor system. However, a study using transcranial magnetic stimulation suggested impaired plasticity in the primary motor cortex in patients with MSA [21]. This indicates that patients with MSA may have reduced compensatory brain function compared with that in patients with hSCD.

The MSA-C group showed significant improvement in BESTest scores, although not reaching MDC, highlighting the test’s comprehensive evaluation of balance function beyond the cerebellar ataxia symptoms assessed using SARA [22]. In this study, we investigated the intervention effects across BESTest subdomains and observed improvements in all areas except “biomechanical constraints” (Fig. 3). These findings suggest that intensive rehabilitation in patients with MSA-C and hSCD may enhance balance function through factors beyond cerebellar ataxia, including postural stability and vestibular function [15, 23, 24]. The sensitivity of the BESTest to these changes elucidates its potential value in assessing rehabilitation outcomes for these disorders, particularly in capturing improvements in extracerebellar symptoms that contribute to overall balance function.

Fig. 3
figure 3

Time series boxplots of the BESTest subcategory. Higher BESTest scores denote improved balance function. Each dot represents the individual dataset. Significant main effects of diagnosis were observed in the “biomechanical constraints” and “anticipatory postural adjustments” domains. Significant main effects of time were observed in all domains except “biomechanical constraints.” p-values are shown after Bonferroni correction. BESTest, balance evaluation systems test; hSCD, hereditary spinocerebellar degeneration; MSA-C, multiple system atrophy-cerebellar type

Full size image

Divergent Long-term Effects in the SCD and MSA-C Groups at the 6-month Follow-up

The results of this study clearly demonstrated differences in outcome changes at the 6-month follow-up between the hSCD and MSA-C groups. For individuals with hSCD, the SARA score at the 6-month follow-up remained similar to or better than the baseline scores. In contrast, those with MSA-C exhibited apparent deterioration compared with baseline (Fig. 2). This difference can be attributed to the varying speeds of disease progression; MSA-C progresses much more rapidly than do other SCDs. The annual increase in SARA scores was reported to be 1.5 points for SCA3, 0.8 points for SCA6, and 0.8 points for SCA31, a pure cerebellar SCD unique to Japan [1, 25, 26].

In comparison, the annual increase in SARA scores for MSA is approximately 4 points [27]. Differences were also observed in the rate of decline in mobility. Previous studies have shown that individuals with MSA typically require a walking aid approximately 3 years after onset [28]. Conversely, those with SCA6 generally need a walking aid 11.2 years after onset and a wheelchair after a median of 24 years [29, 30]. Furthermore, individuals with SCA31 typically require a wheelchair 20 years after onset [26]. However, SCA3, a form of hSCD known for its relatively rapid progression, necessitates the use of a walking aid 6 years after onset and a wheelchair 7.1 years after onset [30]. Our results showed a clear difference in score changes at the 6-month follow-up between the hSCD and MSA-C groups. The MSA-C group exhibited worse scores at the 6-month follow-up compared with baseline scores.

However, the BESTest scores for both groups did not show a significant decline at the 6-month follow-up compared with baseline despite improvements after the intervention (Fig. 2). The hSCD group scored 2.43 points higher than the baseline scores, whereas the MSA-C group maintained scores comparable to baseline scores.

SCD is a progressive disease [30]; therefore, its natural course may result in declining outcome scores over time [31]. Maintaining baseline scores over a long-term follow-up period can indicate an intervention’s efficacy [32]. Patients with MSA-C show a faster rate of cerebellar ataxia symptom deterioration than do those with hSCD [33]. The 6-month follow-up SARA scores in this study reflect these differences in the natural progression of the disease. From these results, it can be inferred that SARA is sensitive to changes in cerebellar ataxia, which is central to SCD pathology. In contrast, BESTest scores were maintained at or above baseline in both groups. Studies using voxel-based morphometry to investigate brain volume changes before and after balance training in patients with SCD have found that the SCD group showed an increase in the volume of compensatory brain regions, such as the supplementary motor area, rather than the cerebellum [34]. The retention of compensatory improvements in balance function beyond cerebellar ataxia symptoms may explain the stable BESTest results.

Limitations

The MSA group experienced a decline in function at the 6-month follow-up compared with that at baseline. In contrast, the hSCD group maintained their baseline function levels. Maintaining these functions in progressive diseases is essential for effective interventions. Nevertheless, the duration for which these functions can be maintained at baseline levels remains unclear. Therefore, this survey is significant for patients with SCD who require continuous rehabilitation. In addition, the hSCD group was analyzed as a single entity. However, within hSCD, diseases such as SCA6, which is a pure cerebellar form, may have different prognoses compared with those for multisystem atrophy types such as SCA3. Therefore, further investigations with larger sample sizes are necessary. Furthermore, the composition of our hSCD group, which primarily included patients with SCA6 and SCA31, limits our ability to generalize these findings to all forms of hSCD. Future studies should aim to include a broader range of SCA subtypes, particularly those with significant extra-cerebellar involvement, to provide a more comprehensive comparison with MSA-C and better represent the full spectrum of hereditary SCAs.

Conclusions

The results of this study revealed differences in the effects of intensive rehabilitation programs between the hSCD and MSA-C groups. The hSCD group showed improvements in ataxia symptoms and balance, whereas the MSA-C group exhibited improvements only in balance function. Notably, these differences were still evident at the 6-month follow-up. The SARA scores in the hSCD group returned to approximately baseline levels, whereas the MSA-C group showed scores lower than baseline scores. However, the BESTest scores were maintained at or above baseline scores in both groups. These differences are likely due to variations in the rate of disease progression and the presence of symptoms other than cerebellar ataxia. Therefore, our results provide critical insights for planning rehabilitation programs for specific disease types.