Effects of combined repetitive transcranial magnetic stimulation and treadmill training on gait performance in Parkinson's disease - a randomised controlled trailc

Abstract

Background: Gait disturbance is a hallmark of Parkinson's disease (PD) and a common cause of disability. Treatment for Parkinsonian gait impairment remains challenging thus far, especially as the disease progresses. Novel interventions are warranted to maximise a patient's gait performance and delay disability. A previous study reported enhanced benefits from treadmill training when primed with high-frequency repetitive transcranial magnetic stimulation (rTMS), but its long-term effects and the mechanisms underlying improvement are unclear. While the enhanced effect may be due to a transient increase in cortical excitation by high-frequency rTMS, it is plausible that a cortical-inhibiting protocol would also augment the training effect by taking advantage of homeostatic metaplasticity. Low-frequency rTMS has not been considered as a method in priming motor training in the treatment of PD. Therefore, a comparison of high- and low-frequency rTMS may help to elucidate the mechanism underlying the beneficial priming effect observed with high-frequency rTMS. In this double-blind, randomised, placebo-controlled trial, we investigated whether 1 Hz and 25 Hz rTMS were more effective than sham rTMS in augmenting the effect of treadmill training in patients with PD in the long-term. We further investigated whether behavioural improvement produced by active rTMS was mediated by changes in corticomotor excitability. In addition, the relative efficacy between 1 Hz and 25 Hz rTMS in augmenting the training effect was explored. We hypothesised that if increased brain excitability is central to priming motor learning, 25 Hz rTMS will be more effective. Conversely, if expanding the range of possible plastic change is essential for priming motor learning, 1 Hz rTMS will be more effective. Method: Fifty-one participants with mild to moderate PD were randomly assigned (1:1:1) to receive 12 sessions of rTMS (25 Hz, 1 Hz or sham) followed by treadmill training over three weeks. The primary outcome measure was the fast walking speed. The secondary outcomes were the timed-up-and-go test (TUG), dual-task TUG (DT-TUG) test, two-minute walk test, mini Balance Evaluation Systems Test (mini-BESTest) and the motor section of the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS-III). Electrophysiological outcome measures included the recruitment curve (RC), cortical silent period (CSP) and short-interval intracortical inhibition (SICI). The participants were assessed at four timepoints, before intervention (baseline), one day (Post), one month (Post1m) and three months (Post3m) after the intervention ended. Data were analysed based on the intention-to-treat principle. Outcome measures were analysed using a two-way repeated measures ANOVA with assessment intervals (baseline, Post, Post1m, Post3m) as the within-factor variable and group (1Hz-TT, 25Hz-TT, sham-TT) as the between-factor variable. Where there was an interaction effect between group and time, a one-way repeated measure ANOVA and t-tests with Bonferroni adjustment for multiple comparisons was used. A between-group analysis of change from the baseline was performed using a one-way ANOVA with Tukey's adjustment for multiple comparisons. Pearson's correlation analysis was applied to establish the associations between the electrophysiological and behavioural measures. All statistical tests were two-tailed with a 5% level of statistical significance. Results: Fifty subjects completed the study and their data were used in the data analysis. The subjects in the three groups had comparable demographics and clinical characteristics. A marginal group-by-time interaction and significant time main effect was present for the primary outcome measure (i.e. fast walking speed). A significant improvement in fast walking speed was found in all subject groups at Post, Post1m and Post3m. Between-group differences revealed a trend toward greater improvement in the 1Hz-TT group compared to sham-TT at Post (p = 0.070) which then reached significance at Post3m (p = 0.012). The 25Hz-TT group trended toward greater improvement at Post (p = 0.051) and at Post3m (p = 0.051) compared to sham-TT. No difference was found when comparing 25Hz-TT with 1Hz-TT at any time point. Significant improvement of secondary outcome measures was observed in both the 1Hz-TT and 25Hz-TT groups for stride length, TUG performance, DT-TUG performance, MDS-UPDRS-III scores, the two-minute walk test, and mini-BESTest from Post to Post3m. Significant improvement was only evident in stride length, the two-minute walk test and mini-BESTest from Post to Post3m in sham-TT. No significant effect was found in DT-TUG and MDS-UPDRS-III in sham-TT while improvement in TUG performance returned to baseline level in the sham-TT group at Post1m assessment. A between-group comparison of the change from baseline showed a trend toward a greater improvement in TUG at Post in 25Hz-TT compared to sham-TT (p = 0.056), which reached significance at Post3m (p = 0.039). Improvement in DT-TUG in 25Hz-TT was greater than sham-TT at Post1m (p = 0.006) and Post3m (p = 0.048). Improvement in MDS-UPDRS-III was greater in 25Hz-TT at Post (p < 0.001) and at Post1m (p < 0.001) compared to sham-TT. Improvement in mini-BESTest was greater in 25Hz-TT at Post compared to sham-TT (p = 0.010). 1Hz-TT showed a greater improvement in MDS-UPDRS-III than sham-TT at Post (p = 0.048). The relative efficacy between active rTMS groups shows 25Hz-TT outperformed 1Hz-TT in TUG (p = 0.029) at Post1m and had a trend of greater improvement in MDS-UPDRS-III at Post (p = 0.056). No significant effect of any intervention was found for absolute electrophysiological outcome measures (RC, CSP and SICI) at any time points. Between-group change from baseline analysis showed a greater increase in CSP at Post in the 25Hz-TT group compared to the sham-TT group (p = 0.047). A trend toward a greater increase in SICI was found in the 25Hz-TT group compared with the sham-TT group at Post (p = 0.068), which reached significance at Post1m (p = 0.012). Correlation analyses revealed behavioural improvements were associated with increased intracortical inhibition in both real rTMS groups. For 25Hz-TT, greater increase in SICI at Post was associated with greater reduction in MDS-UPDRS-III at Post1m (r = 0.524, p = 0.045), and prolonged CSP duration was associated with reduced TUG time at Post1m (r = -0.506, p = 0.038). For 1Hz-TT, prolonged CSP was associated with reduced MDS-UPDRS-III at Post1m (r = -0.560, p = 0.020) and Post3m (r = -0.491, p = 0.045) as well as reduced TUG time at Post3m (r = -0.551, p = 0.022). Prolonged CSP at Post1m was associated with reduced TUG at Post3m (r = -0.504, p = 0.039). Conclusion: This is the first study to report on the long-term effects of combined rTMS and treadmill training on gait and motor performance in patients with PD. The present study also extends our understanding of the beneficial effects of rTMS when used to prime treadmill training in PD patients and sheds light on the possible neural mechanisms underlying behavioural improvement. Our data demonstrated that both 1 Hz and 25 Hz rTMS are superior to the sham treatment in augmenting improvement in trained behaviour when combined with treadmill training. Furthermore, the therapeutic effect of rTMS-primed treadmill training translated to long-term improvement of motor symptoms and dual-task walking. Given that there is an imbalance corticomotoneuron excitability toward a state of disinhibition in PD, motor improvements were associated with the restoration of cortical inhibition mechanism in active rTMS groups. Long-term motor improvement and a correlative normalisation of cortical excitability suggests an expanse of the homeostatic plasticity through the cumulative interaction of rTMS and treadmill training. The rebalancing of cortical excitability by rTMS appears critical for plasticity induction in PD.