Central motor cortical repetitive transcranial magnetic stimulation at 5 Hz or 10 Hz combined with cerebellar 1 Hz stimulation for ataxia in multiple system atrophy of the cerebellar type: a retrospective cohort study of efficacy and predictors of response
Year: 2026
Type: preprint
Abstract: Objective To evaluate the efficacy of central motor cortical repetitive transcranial magnetic stimulation (rTMS) at different frequencies (5 Hz and 10 Hz) combined with cerebellar 1 Hz stimulation on the core domains of ataxia in patients with multiple system atrophy of the cerebellar type (MSA-C), and to explore the independent predictors of the response. Methods In this retrospective study, 48 patients with a confirmed diagnosis of MSA-C were included. All patients received bilateral cerebellar hemispheric 1 Hz rTMS as background therapy and were divided according to the frequency of central cortical stimulation into a control group (cerebellar 1 Hz stimulation only, n=15), a 5 Hz combination group (n=16) and a 10 Hz combination group (n=17). Ataxia severity before and after treatment was assessed with the International Cooperative Ataxia Rating Scale (ICARS); absolute improvements and improvement rates were calculated for each subscore and for the total score. Continuous variables were described as mean ± standard deviation (SD) or median (interquartile range) according to normality; within-group pre–post comparisons used the Wilcoxon signed-rank test, and between-group comparisons used one-way analysis of variance (ANOVA) or the Kruskal–Wallis H test with Bonferroni correction. Clinical response was defined by an absolute improvement threshold (ΔICARS total score ≥8 points), and independent influencing factors were analysed by Firth-penalised logistic regression; sensitivity analyses were performed with a total-score improvement rate ≥20% (and with ≥15% and ≥30% as alternative thresholds), and robustness results are reported for the ΔICARS ≥10-point criterion. Missing cognitive and mood scale scores were imputed by multiple imputation by chained equations (MICE; m=5; the mean of the five imputations constituted the analysis dataset) and then included in the analysis. Subgroup and interaction analyses were performed according to the median disease duration (2 years) and median age (57 years). Results After treatment, the ICARS total score in the overall cohort was significantly lower than at baseline (median difference 7.0 points, P<0.001), and posture and gait (ICARS1), limb kinetic function (ICARS2) and speech function (ICARS3) all improved significantly (P<0.001 for all), whereas oculomotor disorders (ICARS4) did not improve significantly (P=0.799). Efficacy differed significantly among the three groups: median total-score improvement rates were 10.7%, 20.3% and 21.9%, respectively (P=0.004); when a total-score improvement rate ≥20% was used to define clinically significant efficacy, the response rates of the three groups were 6.7%, 50.0% and 58.8% (P=0.006), and both the 10 Hz and 5 Hz groups were significantly better than the control group (Bonferroni-corrected P=0.006 and P=0.040), with no statistically significant difference between the two frequencies. Domain-specific comparison showed statistically significant between-group differences in both the absolute improvement and the improvement rate of limb kinetic function (P=0.013 and P=0.005), with the 10 Hz group significantly better than the control group. When clinical response was defined as ΔICARS ≥8 points, the response rates of the three groups were 20.0%, 56.2% and 64.7% (P=0.029); Firth-penalised multivariable regression showed that baseline ICARS total score was an independent predictor of clinical response (OR=1.076, 95% CI 1.013–1.143, P=0.017), whereas 5 Hz (OR=4.298, 95% CI 0.757–24.412, P=0.100) and 10 Hz (OR=5.847, 95% CI 0.885–38.640, P=0.067) stimulation did not reach independent statistical significance after adjustment. When response was defined as ΔICARS ≥10 points, the response rates were 0%, 43.8% and 47.1% (P=0.007), and there were no responders in the control group. Subgroup analysis showed a significant group×age interaction (P=0.020); the clinical response rate of patients aged ≥57 years receiving 10 Hz stimulation was 85.7% (exploratory finding; small subgroup sample). The group×duration interaction was not statistically significant (P=0.210). Conclusions Central 5 Hz and 10 Hz rTMS combined with cerebellar 1 Hz stimulation both significantly improved the symptoms of ataxia in patients with MSA-C, and the efficacy was domain-specific, with the greatest improvement in the limb kinetic domain; sensitivity analyses across multiple criteria (ΔICARS ≥8 points, improvement rate ≥20%, ΔICARS ≥10 points) consistently supported the efficacy benefit of high-frequency stimulation in the same direction, but stimulation frequency did not reach independent predictive statistical significance after adjustment, and baseline ICARS total score was significantly associated with absolute improvement. Age may have a modifying effect on the efficacy of stimulation frequency (exploratory finding; small sample). This study provides preliminary support for the efficacy of the central 10 Hz combined with cerebellar 1 Hz regimen; the optimal stimulation parameters and independent predictors remain to be verified in larger prospective studies.
Source: medRxiv
Institution: Beijing Sanbo Brain Hospital
Language: English
Subfield: Neurology
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
Funder: Capital Medical University
OpenAlex ID: W7236161611
