REVIEW
Repetitive Transcranial Magnetic Stimulation of the Primary Motor Cortex in the Treatment of Motor Signs in Parkinson’s Disease: A Quantitative Review of the Literature Anosha Zanjani, BSc,1 Konstantine K. Zakzanis, PhD, CPsych,1* Zafiris J. Daskalakis, MD, PhD, FRCPC,2 and Robert Chen, MBBChir, MSc, FRCPC3 1
Department of Psychology, University of Toronto Scarborough, Toronto, Ontario, Canada Centre for Addiction and Mental Health, University of Toronto, Toronto, Ontario, Canada 3 Division of Neurology, Department of Medicine, Krembil Neuroscience Centre and Toronto Western Research Institute, University Health Network, University of Toronto, Toronto, Ontario, Canada 2
ABSTRACT:
Parkinson’s disease (PD) is a progressive disorder characterized by the emergence of motor deficits. In light of the voluminous and conflicting findings in the literature, the aim of the present quantitative review was to examine the effects of repetitive transcranial magnetic stimulation (rTMS) targeting the primary motor cortex (M1) in the treatment of motor signs in PD. Studies meeting inclusion criteria were analyzed using meta-analytic techniques and the Unified Parkinson’s Disease Rating Scale (UPDRS) sections II and III were used as outcome measures. In order to determine the treatment effects of rTMS, the UPDRS II and III scores obtained at baseline, same day, to 1 day post rTMS treatment (short-term follow-up) and 1month post stimulation (long-term follow-up) were compared between the active and sham rTMS groups. Additionally, the placebo effect was evaluated as the changes in UPDRS III scores in the sham rTMS groups. A placebo effect was not demonstrated, because sham
------------------------------------------------------------
*Correspondence to: Konstantine K. Zakzanis, Department of Psychology, University of Toronto Scarborough, 1265 Military Trail, Toronto, Ontario, M1C 1A4, Canada, E-mail: [email protected] Funding agencies: This study was supported by
Relevant conflicts of interest/financial disclosures: Anosha Zanjani— none; Konstantine K. Zakzanis—none; Zafiris J. Daskalakis—none; Robert Chen—Supported by the Canadian Institutes of Health Research (MOP15128) Full financial disclosures and author roles may be found in the online version of this article. Received: 5 January 2015; Revised: 5 February 2015; Accepted: 11 February 2015 Published online 00 Month 2015 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/mds.26206
rTMS did not improve motor signs as measured by UPDRS III. Compared with sham rTMS, active rTMS targeting the M1 significantly improved UPDRS III scores at the short-term follow-up (Cohen’s d of 0.27, UPDRS III score improvement of 3.8 points). When the longterm follow-up UPDRS III scores were compared with baseline scores, the standardized effect size between active and sham rTMS did not reach significance. However, this translated into a significant nonstandardized 6.3-point improvement on the UPDRS III. No significant improvement in the UPDRS II was found. rTMS over the M1 may improve motor signs. Further studies are C 2015 Interneeded to provide a definite conclusion. V national Parkinson and Movement Disorder Society
K e y W o r d s : Transcranial magnetic stimulation (rTMS); Parkinson’s disease; Meta-analysis; Review; Treatment efficacy
Parkinson’s disease (PD) is a progressive disorder affecting more than 6 million people worldwide.1 Progression of PD is characterized by the worsening of motor deficits.2 These motor disturbances are thought to be at least partly related to dysfunction in the primary motor cortex (M1).3 Transcranial magnetic stimulation (TMS) has been used in the treatment of motor symptoms in PD.4 Repetitive TMS (rTMS) can modulate cortical excitability and activate different brain regions.5,6 Consequently, rTMS can improve clinical symptoms in neurological conditions characterized by altered M1 functions such as PD.7 Because of its superficial location, M1 can be stimulated easily using TMS.8 The therapeutic role of rTMS targeting the M1 in patients with PD is not well defined.9 Although a
Movement Disorders, Vol. 00, No. 00, 2015
1
Z A N J A N I
E T
A L
number of studies have been undertaken, a marked variability in their protocols exists. This variability may account for conflicting results that characterize primary studies, making a clear conclusion difficult to discern. These inconsistencies have generated skepticism about the potential efficacy of rTMS in the treatment of motor signs in PD.9 No published systematic quantitative review has elucidated the efficacy of M1 stimulation using rTMS in the treatment of motor signs in PD. Previous reviews have pooled the results of studies that targeted different areas and not exclusively M1.10,11 The pooling of different stimulation sites, especially those unrelated to motor behavior, makes arriving at a reliable conclusion about the efficacy of rTMS in the treatment of motor signs is difficult. Moreover, since the publication of these reviews, several controlled clinical trials using M1 rTMS have been published.2,4,7-9 These primary studies have employed more strict inclusion criteria, which has allowed for the opportunity to examine stimulation of M1 exclusively. Furthermore, no reviews have addressed how rTMS treatment impacts functional outcome in terms activities of daily living (ADL) in patients with PD, which is important in establishing its clinical significance. A further shortcoming of previous syntheses of this literature was that the stability of symptom improvement after rTMS treatment compared with that of the 1-month follow-up was not evaluated. We posit that such findings may shed some further insight into the temporal effects of rTMS. Accordingly, we undertook a quantitative review of the research literature to investigate the potential efficacy of rTMS targeting M1 in the treatment of PD motor signs. By systematically examining effect sizes across primary studies, we set out to address the following questions: 1. What were the short-term benefits of rTMS targeting the M1 in patients with PD? 2. Were improvements exhibited at 1-month post stimulation? 3. How stable were the effects of rTMS on motor signs in multisessional rTMS protocols from the end of the last session to the 1-month follow-up? 4. Did placebo effects contribute to clinical improvement? 5. Did rTMS targeting M1 lead to improvement in ADL?
Methods Literature Search A systematic search of the literature to the end of September 2014 was undertaken. Two authors independently reviewed each article for inclusion criteria in terms of quality and validity. The following elec-
2
Movement Disorders, Vol. 00, No. 00, 2015
tronic databases were searched: PubMed, PsycInfo, MEDLINE, Web of Science, Scholars Portal Search, and Google Scholar. The following key words were used in varying combinations: Parkinson’s disease, Parkinsonism, transcranial magnetic stimulation, brain stimulation, noninvasive brain stimulation, and rTMS. This was supplemented by scanning the reference lists of retrieved studies and reviews to find additional studies for inclusion.
Inclusion Criteria I. rTMS was employed in single or multiple sessions. II. A randomized, sham controlled trial methodology was used, either parallel or crossover in design. III. The target population was idiopathic PD diagnosed according to validated clinical criteria. IV. Motor signs were assessed using Unified Parkinson Disease Rating Scale (UPDRS) part III. V. The paper was available in English. VI. The ADL was assessed using UPDRS part II. VII. The post-stimulation assessment was conducted on the same day or the day after the last stimulation session. These were considered short-term follow-up assessments. VIII. For a long-term follow-up assessment to be included, it must have been conducted at 1 month after the last stimulation session. IX. The M1 was the stimulation target. X. The study included data from which an effect size could be computed. If the same study used more than one active group (eg, two different rTMS frequencies), one group was randomly selected to be included in the meta-analysis. For the studies that met our criteria but did not report their findings in a way in which effect sizes could be computed (e.g. mean and standard deviation), the authors were contacted to provide usable data.
Outcome Variables The effect size derived from each study reflects the difference between active and sham rTMS treatment on the UPDRS parts II and III scores. The UPDRS, a widely accepted scale that is valid, reliable, and high in internal consistency, was used as the outcome measure.12,13 Specifically, UPDRS III was used to evaluate motor function, and the UPDRS II was used to assess ADL.
Data Analysis Comprehensive Meta-analysis Version 2 was used for statistical analyses. The main statistics produced included standardized mean differences (Cohen’s d), nonstandardized mean differences, and their corresponding confidence intervals (95% CI), and P values.
R T M S
The method to pool continuous outcomes in a metaanalysis combining parallel and crossover trial designs has been established.14 Our analysis is based on a weighted average of the 2 design treatment estimates. All effect sizes and mean differences computed were pooled and weighted using a random effects model. The random effects model acknowledges the betweenstudy heterogeneity that may exist beyond sampling error and incorporates this additional variability in the standard error of the estimated effect size. Significant effect sizes were also transformed into a non-overlap percentage using Cohen’s15 idealized distributions, which can be further transformed into an overlap percentage to articulate the meaningfulness of an effect size.16,17 With respect to the interpretation of our effects, a positive effect size represents an improvement, and a negative effect size indicates a deterioration of motor function or ADL. In accordance with Cohen’s15 guidelines, Cohen’s d values from 0.2 to 0.4 were considered to show a small effect, whereas effect sizes from 0.4 to 0.7 were deemed to indicate moderate to substantial effect. Effect sizes below 0.2 were considered to indicate no significant effect.15 According to Shulman et al.’s18 study on clinically important differences (CID) in the UPDRS III, a minimal CID was considered to be a 2.3- to 2.7-point change on the UPDRS III, a moderate CID was thought to be a 4.5- to 6.7-point difference, and a large CID was considered to be a 10.7- to 10.8-point change. Publication bias was investigated visually using funnels plots and statistically using Egger’s regression intercept tests (Egger test). For all calculations, a Pvalue of less than 0.05 was considered significant.
Results Literature Search The literature search yielded 105 articles. Among these, 11 randomized, sham controlled trials met the inclusion criteria (Fig. 1). A parallel design was implemented in seven, and a crossover design was used in four of the included clinical trials. Additionally, five studies provided a 1-month follow up, all of which involved rTMS applied in multiple sessions.2,4,19-21 A total of 246 patients with PD (109 active only, 93 sham only, and 44 both active and sham) contributed to the statistical analysis. The characteristics of the included studies are shown in Table 1. The mean age of the PD patients was 62.9 years, with an average Hoehn and Yahr’s stage of 2.3. Mixed results were reported in the included clinical trials. Four studies reported no significant effect, and seven studies found significant improvement in motor signs after rTMS treatment.2-4,7-9,19-23
A N D
P A R K I N S O N ’ S
D I S E A S E
FIG. 1. Flow diagram of study selection and inclusion in the metaanalysis.
Furthermore, one crossover study3 included two active groups and one sham group. For this study, one active group (the low frequency rTMS group) was chosen at random to be included in the meta-analysis. Additionally, the authors of four of the studies were contacted to obtain additional data needed for effect size calculation that was not included in their original publication.3,8,9,21
The Placebo Effect Because patients with PD can experience significant placebo effects after rTMS,24 changes in UPDRS III score between baseline and post-sham rTMS treatment were evaluated. Figure 2 displays these findings. No evidence was seen of publication bias. The distributions of the funnel plots were fairly symmetrical, and the Egger test was not significant. No significant difference was found when comparing baseline with short-term post-sham rTMS UPDRS III scores. Figure 2A displays the forest plot for the 11 studies (137 patients) included in this analysis, exhibiting a pooled effect size of 0.21 (95%CI 5 20.031 to 0.45; df 5 10; P 5 0.088; P-value Egger test 5 0.35). The crossover studies analyzed separately (44 patients) yielded an effect size of 0.24 (95% CI 5 20.18 to 0.66; df 5 3; P 5 0.27; P-value Egger test 5 0.84). When parallel studies (93 patients) were evaluated independently, an effect size of 0.19 (95%CI 5 20.095 to 0.48; df 5 6; P 5 0.19; P-value Egger test 5 0.54) was obtained. Figure 2B and C illustrates the forest plots of the long-term follow-up assessment being compared with baseline (Fig. 2B) and with short-term follow-up (Fig. 2C) evaluations in the sham rTMS group. This included five parallel design studies with a total of 77 patients. When comparing baseline with the long-term follow-up evaluations (Fig. 2B), a pooled nonsignificant effect size of 20.035 (95%CI 5 20.47 to 0.40; df 5 4; P 5 0.88; P-value Egger test 5 0.089) was observed. Similarly, the comparison between the
Movement Disorders, Vol. 00, No. 00, 2015
3
Z A N J A N I
E T
A L
TABLE 1. Summary of included study characteristics
Study
Frequency (Hz) Pulse
Blinding
Design
Arias et al., 2010 Gonzalez-Garcia et al., 20118 Benninger et al., 20122 Bornke et al., 200423 Khedr et al., 200319
Double Double
Parallel Parallel
1 25
100 1000
90% RMT 80% RMT
C F8
10 15
65.50 64.40
18 17
2.50 2.40
On On
On On
No No
No No
Double
Parallel
50
600
80% AMT
C
8
66.50
26
2.50
On
On
Yes
Double Double
Crossover Parallel
10 50
1000 10% RMT 2000 120% RMT
F8 F8
2 10
55.92 57.65
12 36
2.25 n.a
Off Off
On Off
No No
Filipovic et al., 20107 Lefaucheur et al., 20043 Siebner et al., 200022 Lomarev et al., 200620
Single Double
Crossover Crossover
1 0.5
1800 600
90% RMT 80% RMT
F8 F8
8 1
64.50 64.00
10 12
3.30 3.40
On Off
Off Off
No No
Yes—1 month No Yes—1 month No No
Single Double
Crossover Parallel
5 25
2250 90% RMT 1200 100% RMT
F8 F8
1 1
57.00 64.50
10 18
2.00 n.a
Off On
Off On
No No
a
Double
Parallel
50/5
30
80% AMT
C
8
63.85
26
2.95
On
Off
Yes
Double
Parallel
0.2
100
110% AMT
C
8
67.20
61
3.00
On
On
No
9
Benninger et al., 20114 Okabe et al., 200321
Intensity
Medication Status Medication Number Number During Status of Mean of Mean rTMS During UPDRS FollowCoil Sessions Age Patients H&Y Treatment Assessment II up
No Yes—1 month Yes—1 month Yes—1 month
a
Intermittent theta burst stimulation study. Intensity: RMT 5resting motor threshold; AMT 5active motor threshold; Coil: C 5 circular coil; F8 5 figure 8 coil; H&Y 5 Hoehn and Yahr Stages; rTMS 5 repetitive transcranial magnetic stimulation; UPDRS 5 Unified Parkinson’s Disease Rating Scale.
short-term and the long-term follow-up (Fig. 2C) showed a nonsignificant pooled effect size of 20.23 (95%CI 5 20.71 to 0.25; df 5 4; P 5 0.34; P-value Egger test 5 0.064).
UPDRS III Scores at Short-term Follow-up Figure 3A displays the UPDRS III forest plot for the effect size comparison between active and sham rTMS for the change from baseline to the short-term followup in 11 studies. Two hundred forty-six patients with PD were used in this computation, which included four crossover (44 patients) and seven parallel studies (202 patients). A low but significant pooled effect size of 0.27 (95%CI 5 0.034 to 0.50; df 5 10; overlap percent 5 80; P 5 0.025; P-value Egger test 5 0.67) was seen, with most studies showing improvement. These results translate to an average of 3.88 (95%CI 5 0.97 to 6.80; df 5 10; P 5 0.0090) point improvement in UPDRS III compared with sham stimulation. Crossover and parallel studies analyzed separately did not produce a significant effect size. However, their effect sizes were found to be similar in magnitude and direction. Crossover studies generated an effect size of 0.28 (95%CI 5 20.14 to 0.70; df 5 3; P 5 0.19; P-value Egger test 5 0.64) and the randomized double-blinded parallel studies yielded an effect size of 0.27 (95%CI 5 20.063 to 0.61; df 5 6; P 5 0.11; P-value Egger test 5 0.72). Similarly, when the nonstandardized mean differences in UPDRS III scores were compared, crossover studies produced a pooled mean difference of 3.08 (95%CI 5 21.95 to
4
Movement Disorders, Vol. 00, No. 00, 2015
8.12; df 5 3; P 5 0.23), and parallel studies yielded a pooled mean difference of 4.11 (95%CI 5 20.86 to 9.07; df 5 6; P 5 0.11). Furthermore, the distributions of the funnel plots were fairly symmetrical, revealing no significant asymmetry toward positive effect sizes (Fig. 4A), and the effect sizes yielded nonsignificant P values for the Egger tests (P > 0.05).
UPDRS III at 1-Month (Long-term) Follow-up Figure 3B and C display the UPDRS III forest plots for the effect size comparison between active and sham rTMS for the change from baseline to shortterm (Fig. 3B) and to long-term (Fig. 3C) follow-up. The effect sizes of five parallel studies with 167 patients (77 sham and 90 active) were pooled. The comparison between active and sham rTMS for the change from baseline to long-term follow up on the UPDRS III scores (Fig. 3B) showed a nonsignificant pooled effect size of 0.31 (95%CI 5 20.00014 to 0.63; df 5 4; P 5 0.050; P-value Egger test 5 0.26). However, this translated to a significant, nonstandardized improvement of 6.29 points (95%CI 5 0.50 to 12.07; df 5 4; P 5 0.033) on UPDRS III compared with sham stimulation. Figure 3C displays the UPDRS III forest plot comparing active and sham rTMS for the change from short-term to long-term follow-up. It shows a nonsignificant pooled effect size of 0.12 (95%CI 5 20.31 to 0.55; df 5 4; P 5 0.58; P-value Egger test 5 0.074). Additional tests such as funnel plot visual inspection (Fig. 4B, C) and the Eggers test did not suggest publication bias (P > 0.05).
R T M S
A N D
P A R K I N S O N ’ S
D I S E A S E
FIG. 2. Effect sizes (Cohen’s d) with corresponding 95%CI and P values for UPDRS III evaluations in PD patients in the sham rTMS groups. (A) Baseline vs. short-term post-sham rTMS. (B) Baseline vs. 1-month follow-up. (C) Short-term post-sham rTMS vs. 1-month follow-up.
UPDRS II Evaluations The pooled effect sizes comparing the effects of active and sham rTMS on UPDRS II scores at baseline, short-term, and long-term follow-up were computed. Only two parallel design studies with a total of 52 patients (26 active and 26 sham) were included in this analysis. The comparison between the effects of active and sham rTMS for the change from baseline to short-term follow-up assessment showed a pooled nonsignificant effect size of 20.059
(95%CI 5 20.60 to 0.49; df 5 1; P 5 0.83). Similarly, when comparing active and sham rTMS on the UPDRS II scores obtained between baseline and the long-term follow up, a nonsignificant pooled effect size of 0.15 (95%CI 5 20.39 to 0.70; df 5 1; P 5 0.58) was obtained. Furthermore, when active and sham rTMS were compared between the shortterm and the long-term follow-up, a nonsignificant effect size of 0.21 (95%CI 5 20.34 to 0.76; df 5 1; P 5 0.45) was found.
Movement Disorders, Vol. 00, No. 00, 2015
5
Z A N J A N I
E T
A L
FIG. 3. Effect sizes (Cohen’s d) with corresponding 95%CI and P values for UPDRS III evaluations comparing active with sham rTMS groups. (A) Baseline vs. short-term follow-up assessment. (B) Baseline vs. 1-month follow-up. (C) Short-term vs. 1-month follow-up.
Discussion In light of the voluminous and yet conflicting evidence regarding its effectiveness,9 we undertook a quantitative review of the research literature to determine the efficacy of rTMS targeting M1 in the treatment of PD motor signs. Our results suggest that that rTMS may be effective in improving motor signs in
6
Movement Disorders, Vol. 00, No. 00, 2015
patients with PD both short-term and long-term. However, because of small sample sizes and other limiting factors, definitive conclusions cannot yet be drawn. Repetitive transcranial magnetic stimulation is prone to placebo effects because the protocol evokes skin sensation and noise.25 To demonstrate that the effects of rTMS are attributable to activation of brain
R T M S
FIG. 4. Funnel plots of effect sizes (Cohen’s d) according to their standard errors using Unified Parkinson’s Disease Rating Scale III as the outcome variable. Each circle represents an individual study. (A) Baseline vs. short-term post-treatment assessment. (B) Baseline vs. 1-month follow-up. (C) Short-term vs. 1-month follow-up.
structures, an effective sham rTMS typically mimics cutaneous sensations such as the twitching of scalp and facial muscles, and the sound of the coil discharge.21 The sham stimulation method varied across the studies included in this review. Previous studies have shown, however, that different methods of sham rTMS do not appear to influence the outcome of rTMS studies and their placebo responses.2,10 Our results suggest no significant improvements in PD motor signs both at the short-term assessment and at the 1-month assessment after sham rTMS. The study design (parallel or crossover) did not influence the placebo effect. Similarly, the lack of a placebo effect was observed regardless of whether rTMS was found to have beneficial effect on motor signs. Consequently, our results did not detect a significant placebo effect of rTMS to the M1. The rTMS targeting the M1 has been shown to increase dopamine release in the putamen in patients with PD.26,27 Our analysis suggests that active rTMS over M1 may result in significant short-term improve-
A N D
P A R K I N S O N ’ S
D I S E A S E
ments in motor signs compared with sham rTMS. These improvements translated to a 3.83 average point reduction in the UPDRS III, which corresponds to a mild to moderate effect size.18 Thus, our results suggest that the modulation of M1 with rTMS may have a significant positive effect on motor performance in PD. When parallel and crossover studies were examined separately, no significant effect was found. This is likely attributable to the low power of the analysis because of small sample size. However, the effect sizes were similar, suggesting that the findings for the two types of studies were qualitatively consistent. Although a nonsignificant pooled effect size (P 5 0.05) was observed between active and sham rTMS when comparing baseline and 1-month followup scores of the UPDRS III, a significant reduction was found when the nonstandardized scores were analyzed. This discrepancy in significance may be attributable to the effect size being calculated as a standardized mean difference, whereas the mean difference is without standardization. Therefore, rTMS over M1 may potentially show long-term (1-month post treatment) improvements on UPDRS III scores in multi-sessional protocols, but further studies are needed. When follow-up scores were compared with baseline scores between active and sham groups, an average 6.29-point reduction in UPDRS III was found, suggesting moderate improvement.18 The results of the individual studies varied widely, and because of small sample sizes in the current study, the results should be interpreted with caution. When active and sham groups were compared on post-rTMS UPDRS III scores and their 1-month follow-up scores, no significant change was observed. This may indicate a possible stability in motor sign improvement up to 1 month after the last rTMS session. These findings suggest that the biological changes that occur with multiple sessions of rTMS treatment may play a role in its long-term effects.20 The potential therapeutic effects of rTMS are likely mediated by persistent functional modulation of cortical neuronal circuits beyond the time of magnetic stimulation itself.28,29 The assessment of how a treatment affects daily life is of critical importance in establishing its clinical relevance. Therefore, UPDRS II was also examined in this meta-analysis.30 No significant improvement in the UPRDS II was observed after active rTMS treatment when compared with sham rTMS. This analysis was, however, conducted with only two multisession studies that included 52 patients. Therefore, these results should be interpreted cautiously. More studies addressing ADL after rTMS treatment are needed to obtain a better understanding of how rTMS can mediate a patient’s ability to engage in their ADLs. Although rTMS has been previously shown to improve UPDRS III scores in PD, the degree and type
Movement Disorders, Vol. 00, No. 00, 2015
7
Z A N J A N I
E T
A L
of motor signs that improve varied among different studies. For example, Lomarev et al.20 found a significant improvement in walking speed, bradykinesia, and gait. Siebner et al.22 found significant improvement in bradykinesia, rigidity, and tremors. However, Khedr et al.19 found improvement in walking speed and mobility but no significant changes in tremors. Therefore, significant improvement on UPDRS III scores does not necessarily translate into a consistent pattern of improvements in all motor signs. Further studies are needed that report different subscores of the UPDRS III to establish whether rTMS improves specific PD motor symptoms. Each study included in this meta-analysis used a different combination of parameters, making determining whether the effects observed were attributable to one factor alone or to moderating factors affecting the parameter under investigation difficult. For example, high- and low-frequency rTMS have been shown to have different effects. Frequencies of 5 Hz and higher have been shown to enhance motor cortex excitability,28,31 whereas lower frequencies rTMS (1 Hz and lower) have been shown to briefly depress cortical excitability.29 Both regular rTMS and theta burst stimulation (a type of rTMS protocol) were included in this systematic review, yet whether this new pattern of stimulation differs behaviorally or biologically from the effects of regular rTMS in patients with PD is unknown. Different stimulus intensities, number of pulses per session, the type of coil used, as well as the number of rTMS sessions applied may also have different effects in PD motor signs. The interactions between various parameters were not considered in this meta-analysis. Additionally, we pooled studies from different countries (Germany, Egypt, United Kingdom, United States of America, Switzerland, Spain, Canada, France, and Japan) that may use different approaches in the treatment of PD. For example, the study of Khedr et al.19 from Egypt involves unmedicated, advanced PD patients, who are rarely seen in more developed counties. These variations in methodology and patient selection likely account for some of the discrepancies in the literature. Other factors that may account for the variations include whether treatment was administered in the on or off medication states,32 genetic polymorphisms of the subjects studied,33 the time of the day the study was conducted, and the different types of sham stimulation employed that may vary in their effectiveness in masking the study assignment. Furthermore, we did not report heterogeneity analysis because of the small number of included studies and consequently its lack of usefulness in this investigation. A nonsignificant result may not be taken as evidence of no heterogeneity because of the low statistical power. The nonsignificant Eggers tests also may
8
Movement Disorders, Vol. 00, No. 00, 2015
be misleading for the same reasons and must be interpreted with caution. Therefore, the existence of publication bias cannot be excluded. Additionally, this meta-analysis did not follow the stringency of the Cochran Review criteria in its study selection. Furthermore, some of the positive results of earlier studies have not been confirmed. Although our findings suggest that rTMS may have an adjunctive role in the treatment of PD, because of low statistical power and other limiting factors discussed earlier, a definite conclusion cannot be drawn with regard to its efficacy. Large randomized placebocontrolled trials assessing the impact of rTMS on motor functions and activities of daily living are needed to establish the long-term and short-term benefits of rTMS in PD.
References 1.
Schapira AH. Science, medicine, and the future: Parkinson’s disease. BMJ 1999;318:311-314.
2.
Benninger DH, Iseki K, Kranick S, Luckenbaugh DA, Houdayer E, Hallett M. Controlled study of 50-Hz repetitive transcranial magnetic stimulation for the treatment of Parkinson disease. Neurorehabil Neural Repair 2012;26:1096-1105.
3.
Lefaucheur JP, Drouot X, Von Raison F, Menard-Lefaucheur I, Cesaro P, Nguyen JP. Improvement of motor performance and modulation of cortical excitability by repetitive transcranial magnetic stimulation of the motor cortex in Parkinson’s disease. Clin Neurophysiol 2004;115:2530-2541.
4.
Benninger DH, Berman BD, Houdayer E, et al. Intermittent thetaburst transcranial magnetic stimulation for treatment of Parkinson disease. Neurology 2011;76:601-609.
5.
Maertens de Noordhout A. General principles for clinical use of repetitive transcranial magnetic stimulation (rTMS). Neurophysiol Clin 2006;36:97-103.
6.
Lefaucheur JP. Principles of therapeutic use of transcranial and epidural cortical stimulation. Clin Neurophysiol 2008;119:2179-2184.
7.
Filipovic SR, Rothwell JC, Bhatia K. Low-frequency repetitive transcranial magnetic stimulation and off-phase motor symptoms in Parkinson’s disease. J Neurol Sci 2010;291:1-4.
8.
Gonzalez-Garcia N, Armony JL, Soto J, Trejo D, Alegria MA, Drucker-Colin R. Effects of rTMS on Parkinson’s disease: a longitudinal fMRI study. J Neurol 2011;258:1268-1280.
9.
Arias P, Vivas J, Grieve KL, Cudeiro J. Controlled trial on the effect of 10 days low-frequency repetitive transcranial magnetic stimulation (rTMS) on motor signs in Parkinson’s disease. Mov Disord 2010;25:1830-1838.
10.
Elahi B, Chen R. Effect of transcranial magnetic stimulation on Parkinson motor function: systematic review of controlled clinical trials. Mov Disord 2009;24:357-363.
11.
Fregni F, Simon DK, Wu A, Pascual-Leone A. Non-invasive brain stimulation for Parkinson’s disease: a systematic review and metaanalysis of the literature. J Neurol Neurosurg Psychiatry 2005;76: 1614-1623.
12.
Movement Disorder Society Task Force on Rating Scales for Parkinson’s D. The Unified Parkinson’s Disease Rating Scale (UPDRS): status and recommendations. Mov Disord 2003;18:738750.
13.
Martinez-Martin P, Gil-Nagel A, Gracia LM, Gomez JB, Martinez-Sarries J, Bermejo F. Unified Parkinson’s Disease Rating Scale characteristics and structure. The Cooperative Multicentric Group. Mov Disord 1994;9:76-83.
14.
Curtin F, Altman DG, Elbourne D. Meta-analysis combining parallel and cross-over clinical trials. I: Continuous outcomes. Stat Med 2002;21:2131-2144.
15.
Cohn J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates, 1988.
R T M S
16.
Zakzanis KK. Brain is related to behavior (p
Repetitive Transcranial Magnetic Stimulation of the Primary Motor Cortex in the Treatment of Motor Signs in Parkinson’s Disease: A Quantitative Review of the Literature Anosha Zanjani, BSc,1 Konstantine K. Zakzanis, PhD, CPsych,1* Zafiris J. Daskalakis, MD, PhD, FRCPC,2 and Robert Chen, MBBChir, MSc, FRCPC3 1
Department of Psychology, University of Toronto Scarborough, Toronto, Ontario, Canada Centre for Addiction and Mental Health, University of Toronto, Toronto, Ontario, Canada 3 Division of Neurology, Department of Medicine, Krembil Neuroscience Centre and Toronto Western Research Institute, University Health Network, University of Toronto, Toronto, Ontario, Canada 2
ABSTRACT:
Parkinson’s disease (PD) is a progressive disorder characterized by the emergence of motor deficits. In light of the voluminous and conflicting findings in the literature, the aim of the present quantitative review was to examine the effects of repetitive transcranial magnetic stimulation (rTMS) targeting the primary motor cortex (M1) in the treatment of motor signs in PD. Studies meeting inclusion criteria were analyzed using meta-analytic techniques and the Unified Parkinson’s Disease Rating Scale (UPDRS) sections II and III were used as outcome measures. In order to determine the treatment effects of rTMS, the UPDRS II and III scores obtained at baseline, same day, to 1 day post rTMS treatment (short-term follow-up) and 1month post stimulation (long-term follow-up) were compared between the active and sham rTMS groups. Additionally, the placebo effect was evaluated as the changes in UPDRS III scores in the sham rTMS groups. A placebo effect was not demonstrated, because sham
------------------------------------------------------------
*Correspondence to: Konstantine K. Zakzanis, Department of Psychology, University of Toronto Scarborough, 1265 Military Trail, Toronto, Ontario, M1C 1A4, Canada, E-mail: [email protected] Funding agencies: This study was supported by
Relevant conflicts of interest/financial disclosures: Anosha Zanjani— none; Konstantine K. Zakzanis—none; Zafiris J. Daskalakis—none; Robert Chen—Supported by the Canadian Institutes of Health Research (MOP15128) Full financial disclosures and author roles may be found in the online version of this article. Received: 5 January 2015; Revised: 5 February 2015; Accepted: 11 February 2015 Published online 00 Month 2015 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/mds.26206
rTMS did not improve motor signs as measured by UPDRS III. Compared with sham rTMS, active rTMS targeting the M1 significantly improved UPDRS III scores at the short-term follow-up (Cohen’s d of 0.27, UPDRS III score improvement of 3.8 points). When the longterm follow-up UPDRS III scores were compared with baseline scores, the standardized effect size between active and sham rTMS did not reach significance. However, this translated into a significant nonstandardized 6.3-point improvement on the UPDRS III. No significant improvement in the UPDRS II was found. rTMS over the M1 may improve motor signs. Further studies are C 2015 Interneeded to provide a definite conclusion. V national Parkinson and Movement Disorder Society
K e y W o r d s : Transcranial magnetic stimulation (rTMS); Parkinson’s disease; Meta-analysis; Review; Treatment efficacy
Parkinson’s disease (PD) is a progressive disorder affecting more than 6 million people worldwide.1 Progression of PD is characterized by the worsening of motor deficits.2 These motor disturbances are thought to be at least partly related to dysfunction in the primary motor cortex (M1).3 Transcranial magnetic stimulation (TMS) has been used in the treatment of motor symptoms in PD.4 Repetitive TMS (rTMS) can modulate cortical excitability and activate different brain regions.5,6 Consequently, rTMS can improve clinical symptoms in neurological conditions characterized by altered M1 functions such as PD.7 Because of its superficial location, M1 can be stimulated easily using TMS.8 The therapeutic role of rTMS targeting the M1 in patients with PD is not well defined.9 Although a
Movement Disorders, Vol. 00, No. 00, 2015
1
Z A N J A N I
E T
A L
number of studies have been undertaken, a marked variability in their protocols exists. This variability may account for conflicting results that characterize primary studies, making a clear conclusion difficult to discern. These inconsistencies have generated skepticism about the potential efficacy of rTMS in the treatment of motor signs in PD.9 No published systematic quantitative review has elucidated the efficacy of M1 stimulation using rTMS in the treatment of motor signs in PD. Previous reviews have pooled the results of studies that targeted different areas and not exclusively M1.10,11 The pooling of different stimulation sites, especially those unrelated to motor behavior, makes arriving at a reliable conclusion about the efficacy of rTMS in the treatment of motor signs is difficult. Moreover, since the publication of these reviews, several controlled clinical trials using M1 rTMS have been published.2,4,7-9 These primary studies have employed more strict inclusion criteria, which has allowed for the opportunity to examine stimulation of M1 exclusively. Furthermore, no reviews have addressed how rTMS treatment impacts functional outcome in terms activities of daily living (ADL) in patients with PD, which is important in establishing its clinical significance. A further shortcoming of previous syntheses of this literature was that the stability of symptom improvement after rTMS treatment compared with that of the 1-month follow-up was not evaluated. We posit that such findings may shed some further insight into the temporal effects of rTMS. Accordingly, we undertook a quantitative review of the research literature to investigate the potential efficacy of rTMS targeting M1 in the treatment of PD motor signs. By systematically examining effect sizes across primary studies, we set out to address the following questions: 1. What were the short-term benefits of rTMS targeting the M1 in patients with PD? 2. Were improvements exhibited at 1-month post stimulation? 3. How stable were the effects of rTMS on motor signs in multisessional rTMS protocols from the end of the last session to the 1-month follow-up? 4. Did placebo effects contribute to clinical improvement? 5. Did rTMS targeting M1 lead to improvement in ADL?
Methods Literature Search A systematic search of the literature to the end of September 2014 was undertaken. Two authors independently reviewed each article for inclusion criteria in terms of quality and validity. The following elec-
2
Movement Disorders, Vol. 00, No. 00, 2015
tronic databases were searched: PubMed, PsycInfo, MEDLINE, Web of Science, Scholars Portal Search, and Google Scholar. The following key words were used in varying combinations: Parkinson’s disease, Parkinsonism, transcranial magnetic stimulation, brain stimulation, noninvasive brain stimulation, and rTMS. This was supplemented by scanning the reference lists of retrieved studies and reviews to find additional studies for inclusion.
Inclusion Criteria I. rTMS was employed in single or multiple sessions. II. A randomized, sham controlled trial methodology was used, either parallel or crossover in design. III. The target population was idiopathic PD diagnosed according to validated clinical criteria. IV. Motor signs were assessed using Unified Parkinson Disease Rating Scale (UPDRS) part III. V. The paper was available in English. VI. The ADL was assessed using UPDRS part II. VII. The post-stimulation assessment was conducted on the same day or the day after the last stimulation session. These were considered short-term follow-up assessments. VIII. For a long-term follow-up assessment to be included, it must have been conducted at 1 month after the last stimulation session. IX. The M1 was the stimulation target. X. The study included data from which an effect size could be computed. If the same study used more than one active group (eg, two different rTMS frequencies), one group was randomly selected to be included in the meta-analysis. For the studies that met our criteria but did not report their findings in a way in which effect sizes could be computed (e.g. mean and standard deviation), the authors were contacted to provide usable data.
Outcome Variables The effect size derived from each study reflects the difference between active and sham rTMS treatment on the UPDRS parts II and III scores. The UPDRS, a widely accepted scale that is valid, reliable, and high in internal consistency, was used as the outcome measure.12,13 Specifically, UPDRS III was used to evaluate motor function, and the UPDRS II was used to assess ADL.
Data Analysis Comprehensive Meta-analysis Version 2 was used for statistical analyses. The main statistics produced included standardized mean differences (Cohen’s d), nonstandardized mean differences, and their corresponding confidence intervals (95% CI), and P values.
R T M S
The method to pool continuous outcomes in a metaanalysis combining parallel and crossover trial designs has been established.14 Our analysis is based on a weighted average of the 2 design treatment estimates. All effect sizes and mean differences computed were pooled and weighted using a random effects model. The random effects model acknowledges the betweenstudy heterogeneity that may exist beyond sampling error and incorporates this additional variability in the standard error of the estimated effect size. Significant effect sizes were also transformed into a non-overlap percentage using Cohen’s15 idealized distributions, which can be further transformed into an overlap percentage to articulate the meaningfulness of an effect size.16,17 With respect to the interpretation of our effects, a positive effect size represents an improvement, and a negative effect size indicates a deterioration of motor function or ADL. In accordance with Cohen’s15 guidelines, Cohen’s d values from 0.2 to 0.4 were considered to show a small effect, whereas effect sizes from 0.4 to 0.7 were deemed to indicate moderate to substantial effect. Effect sizes below 0.2 were considered to indicate no significant effect.15 According to Shulman et al.’s18 study on clinically important differences (CID) in the UPDRS III, a minimal CID was considered to be a 2.3- to 2.7-point change on the UPDRS III, a moderate CID was thought to be a 4.5- to 6.7-point difference, and a large CID was considered to be a 10.7- to 10.8-point change. Publication bias was investigated visually using funnels plots and statistically using Egger’s regression intercept tests (Egger test). For all calculations, a Pvalue of less than 0.05 was considered significant.
Results Literature Search The literature search yielded 105 articles. Among these, 11 randomized, sham controlled trials met the inclusion criteria (Fig. 1). A parallel design was implemented in seven, and a crossover design was used in four of the included clinical trials. Additionally, five studies provided a 1-month follow up, all of which involved rTMS applied in multiple sessions.2,4,19-21 A total of 246 patients with PD (109 active only, 93 sham only, and 44 both active and sham) contributed to the statistical analysis. The characteristics of the included studies are shown in Table 1. The mean age of the PD patients was 62.9 years, with an average Hoehn and Yahr’s stage of 2.3. Mixed results were reported in the included clinical trials. Four studies reported no significant effect, and seven studies found significant improvement in motor signs after rTMS treatment.2-4,7-9,19-23
A N D
P A R K I N S O N ’ S
D I S E A S E
FIG. 1. Flow diagram of study selection and inclusion in the metaanalysis.
Furthermore, one crossover study3 included two active groups and one sham group. For this study, one active group (the low frequency rTMS group) was chosen at random to be included in the meta-analysis. Additionally, the authors of four of the studies were contacted to obtain additional data needed for effect size calculation that was not included in their original publication.3,8,9,21
The Placebo Effect Because patients with PD can experience significant placebo effects after rTMS,24 changes in UPDRS III score between baseline and post-sham rTMS treatment were evaluated. Figure 2 displays these findings. No evidence was seen of publication bias. The distributions of the funnel plots were fairly symmetrical, and the Egger test was not significant. No significant difference was found when comparing baseline with short-term post-sham rTMS UPDRS III scores. Figure 2A displays the forest plot for the 11 studies (137 patients) included in this analysis, exhibiting a pooled effect size of 0.21 (95%CI 5 20.031 to 0.45; df 5 10; P 5 0.088; P-value Egger test 5 0.35). The crossover studies analyzed separately (44 patients) yielded an effect size of 0.24 (95% CI 5 20.18 to 0.66; df 5 3; P 5 0.27; P-value Egger test 5 0.84). When parallel studies (93 patients) were evaluated independently, an effect size of 0.19 (95%CI 5 20.095 to 0.48; df 5 6; P 5 0.19; P-value Egger test 5 0.54) was obtained. Figure 2B and C illustrates the forest plots of the long-term follow-up assessment being compared with baseline (Fig. 2B) and with short-term follow-up (Fig. 2C) evaluations in the sham rTMS group. This included five parallel design studies with a total of 77 patients. When comparing baseline with the long-term follow-up evaluations (Fig. 2B), a pooled nonsignificant effect size of 20.035 (95%CI 5 20.47 to 0.40; df 5 4; P 5 0.88; P-value Egger test 5 0.089) was observed. Similarly, the comparison between the
Movement Disorders, Vol. 00, No. 00, 2015
3
Z A N J A N I
E T
A L
TABLE 1. Summary of included study characteristics
Study
Frequency (Hz) Pulse
Blinding
Design
Arias et al., 2010 Gonzalez-Garcia et al., 20118 Benninger et al., 20122 Bornke et al., 200423 Khedr et al., 200319
Double Double
Parallel Parallel
1 25
100 1000
90% RMT 80% RMT
C F8
10 15
65.50 64.40
18 17
2.50 2.40
On On
On On
No No
No No
Double
Parallel
50
600
80% AMT
C
8
66.50
26
2.50
On
On
Yes
Double Double
Crossover Parallel
10 50
1000 10% RMT 2000 120% RMT
F8 F8
2 10
55.92 57.65
12 36
2.25 n.a
Off Off
On Off
No No
Filipovic et al., 20107 Lefaucheur et al., 20043 Siebner et al., 200022 Lomarev et al., 200620
Single Double
Crossover Crossover
1 0.5
1800 600
90% RMT 80% RMT
F8 F8
8 1
64.50 64.00
10 12
3.30 3.40
On Off
Off Off
No No
Yes—1 month No Yes—1 month No No
Single Double
Crossover Parallel
5 25
2250 90% RMT 1200 100% RMT
F8 F8
1 1
57.00 64.50
10 18
2.00 n.a
Off On
Off On
No No
a
Double
Parallel
50/5
30
80% AMT
C
8
63.85
26
2.95
On
Off
Yes
Double
Parallel
0.2
100
110% AMT
C
8
67.20
61
3.00
On
On
No
9
Benninger et al., 20114 Okabe et al., 200321
Intensity
Medication Status Medication Number Number During Status of Mean of Mean rTMS During UPDRS FollowCoil Sessions Age Patients H&Y Treatment Assessment II up
No Yes—1 month Yes—1 month Yes—1 month
a
Intermittent theta burst stimulation study. Intensity: RMT 5resting motor threshold; AMT 5active motor threshold; Coil: C 5 circular coil; F8 5 figure 8 coil; H&Y 5 Hoehn and Yahr Stages; rTMS 5 repetitive transcranial magnetic stimulation; UPDRS 5 Unified Parkinson’s Disease Rating Scale.
short-term and the long-term follow-up (Fig. 2C) showed a nonsignificant pooled effect size of 20.23 (95%CI 5 20.71 to 0.25; df 5 4; P 5 0.34; P-value Egger test 5 0.064).
UPDRS III Scores at Short-term Follow-up Figure 3A displays the UPDRS III forest plot for the effect size comparison between active and sham rTMS for the change from baseline to the short-term followup in 11 studies. Two hundred forty-six patients with PD were used in this computation, which included four crossover (44 patients) and seven parallel studies (202 patients). A low but significant pooled effect size of 0.27 (95%CI 5 0.034 to 0.50; df 5 10; overlap percent 5 80; P 5 0.025; P-value Egger test 5 0.67) was seen, with most studies showing improvement. These results translate to an average of 3.88 (95%CI 5 0.97 to 6.80; df 5 10; P 5 0.0090) point improvement in UPDRS III compared with sham stimulation. Crossover and parallel studies analyzed separately did not produce a significant effect size. However, their effect sizes were found to be similar in magnitude and direction. Crossover studies generated an effect size of 0.28 (95%CI 5 20.14 to 0.70; df 5 3; P 5 0.19; P-value Egger test 5 0.64) and the randomized double-blinded parallel studies yielded an effect size of 0.27 (95%CI 5 20.063 to 0.61; df 5 6; P 5 0.11; P-value Egger test 5 0.72). Similarly, when the nonstandardized mean differences in UPDRS III scores were compared, crossover studies produced a pooled mean difference of 3.08 (95%CI 5 21.95 to
4
Movement Disorders, Vol. 00, No. 00, 2015
8.12; df 5 3; P 5 0.23), and parallel studies yielded a pooled mean difference of 4.11 (95%CI 5 20.86 to 9.07; df 5 6; P 5 0.11). Furthermore, the distributions of the funnel plots were fairly symmetrical, revealing no significant asymmetry toward positive effect sizes (Fig. 4A), and the effect sizes yielded nonsignificant P values for the Egger tests (P > 0.05).
UPDRS III at 1-Month (Long-term) Follow-up Figure 3B and C display the UPDRS III forest plots for the effect size comparison between active and sham rTMS for the change from baseline to shortterm (Fig. 3B) and to long-term (Fig. 3C) follow-up. The effect sizes of five parallel studies with 167 patients (77 sham and 90 active) were pooled. The comparison between active and sham rTMS for the change from baseline to long-term follow up on the UPDRS III scores (Fig. 3B) showed a nonsignificant pooled effect size of 0.31 (95%CI 5 20.00014 to 0.63; df 5 4; P 5 0.050; P-value Egger test 5 0.26). However, this translated to a significant, nonstandardized improvement of 6.29 points (95%CI 5 0.50 to 12.07; df 5 4; P 5 0.033) on UPDRS III compared with sham stimulation. Figure 3C displays the UPDRS III forest plot comparing active and sham rTMS for the change from short-term to long-term follow-up. It shows a nonsignificant pooled effect size of 0.12 (95%CI 5 20.31 to 0.55; df 5 4; P 5 0.58; P-value Egger test 5 0.074). Additional tests such as funnel plot visual inspection (Fig. 4B, C) and the Eggers test did not suggest publication bias (P > 0.05).
R T M S
A N D
P A R K I N S O N ’ S
D I S E A S E
FIG. 2. Effect sizes (Cohen’s d) with corresponding 95%CI and P values for UPDRS III evaluations in PD patients in the sham rTMS groups. (A) Baseline vs. short-term post-sham rTMS. (B) Baseline vs. 1-month follow-up. (C) Short-term post-sham rTMS vs. 1-month follow-up.
UPDRS II Evaluations The pooled effect sizes comparing the effects of active and sham rTMS on UPDRS II scores at baseline, short-term, and long-term follow-up were computed. Only two parallel design studies with a total of 52 patients (26 active and 26 sham) were included in this analysis. The comparison between the effects of active and sham rTMS for the change from baseline to short-term follow-up assessment showed a pooled nonsignificant effect size of 20.059
(95%CI 5 20.60 to 0.49; df 5 1; P 5 0.83). Similarly, when comparing active and sham rTMS on the UPDRS II scores obtained between baseline and the long-term follow up, a nonsignificant pooled effect size of 0.15 (95%CI 5 20.39 to 0.70; df 5 1; P 5 0.58) was obtained. Furthermore, when active and sham rTMS were compared between the shortterm and the long-term follow-up, a nonsignificant effect size of 0.21 (95%CI 5 20.34 to 0.76; df 5 1; P 5 0.45) was found.
Movement Disorders, Vol. 00, No. 00, 2015
5
Z A N J A N I
E T
A L
FIG. 3. Effect sizes (Cohen’s d) with corresponding 95%CI and P values for UPDRS III evaluations comparing active with sham rTMS groups. (A) Baseline vs. short-term follow-up assessment. (B) Baseline vs. 1-month follow-up. (C) Short-term vs. 1-month follow-up.
Discussion In light of the voluminous and yet conflicting evidence regarding its effectiveness,9 we undertook a quantitative review of the research literature to determine the efficacy of rTMS targeting M1 in the treatment of PD motor signs. Our results suggest that that rTMS may be effective in improving motor signs in
6
Movement Disorders, Vol. 00, No. 00, 2015
patients with PD both short-term and long-term. However, because of small sample sizes and other limiting factors, definitive conclusions cannot yet be drawn. Repetitive transcranial magnetic stimulation is prone to placebo effects because the protocol evokes skin sensation and noise.25 To demonstrate that the effects of rTMS are attributable to activation of brain
R T M S
FIG. 4. Funnel plots of effect sizes (Cohen’s d) according to their standard errors using Unified Parkinson’s Disease Rating Scale III as the outcome variable. Each circle represents an individual study. (A) Baseline vs. short-term post-treatment assessment. (B) Baseline vs. 1-month follow-up. (C) Short-term vs. 1-month follow-up.
structures, an effective sham rTMS typically mimics cutaneous sensations such as the twitching of scalp and facial muscles, and the sound of the coil discharge.21 The sham stimulation method varied across the studies included in this review. Previous studies have shown, however, that different methods of sham rTMS do not appear to influence the outcome of rTMS studies and their placebo responses.2,10 Our results suggest no significant improvements in PD motor signs both at the short-term assessment and at the 1-month assessment after sham rTMS. The study design (parallel or crossover) did not influence the placebo effect. Similarly, the lack of a placebo effect was observed regardless of whether rTMS was found to have beneficial effect on motor signs. Consequently, our results did not detect a significant placebo effect of rTMS to the M1. The rTMS targeting the M1 has been shown to increase dopamine release in the putamen in patients with PD.26,27 Our analysis suggests that active rTMS over M1 may result in significant short-term improve-
A N D
P A R K I N S O N ’ S
D I S E A S E
ments in motor signs compared with sham rTMS. These improvements translated to a 3.83 average point reduction in the UPDRS III, which corresponds to a mild to moderate effect size.18 Thus, our results suggest that the modulation of M1 with rTMS may have a significant positive effect on motor performance in PD. When parallel and crossover studies were examined separately, no significant effect was found. This is likely attributable to the low power of the analysis because of small sample size. However, the effect sizes were similar, suggesting that the findings for the two types of studies were qualitatively consistent. Although a nonsignificant pooled effect size (P 5 0.05) was observed between active and sham rTMS when comparing baseline and 1-month followup scores of the UPDRS III, a significant reduction was found when the nonstandardized scores were analyzed. This discrepancy in significance may be attributable to the effect size being calculated as a standardized mean difference, whereas the mean difference is without standardization. Therefore, rTMS over M1 may potentially show long-term (1-month post treatment) improvements on UPDRS III scores in multi-sessional protocols, but further studies are needed. When follow-up scores were compared with baseline scores between active and sham groups, an average 6.29-point reduction in UPDRS III was found, suggesting moderate improvement.18 The results of the individual studies varied widely, and because of small sample sizes in the current study, the results should be interpreted with caution. When active and sham groups were compared on post-rTMS UPDRS III scores and their 1-month follow-up scores, no significant change was observed. This may indicate a possible stability in motor sign improvement up to 1 month after the last rTMS session. These findings suggest that the biological changes that occur with multiple sessions of rTMS treatment may play a role in its long-term effects.20 The potential therapeutic effects of rTMS are likely mediated by persistent functional modulation of cortical neuronal circuits beyond the time of magnetic stimulation itself.28,29 The assessment of how a treatment affects daily life is of critical importance in establishing its clinical relevance. Therefore, UPDRS II was also examined in this meta-analysis.30 No significant improvement in the UPRDS II was observed after active rTMS treatment when compared with sham rTMS. This analysis was, however, conducted with only two multisession studies that included 52 patients. Therefore, these results should be interpreted cautiously. More studies addressing ADL after rTMS treatment are needed to obtain a better understanding of how rTMS can mediate a patient’s ability to engage in their ADLs. Although rTMS has been previously shown to improve UPDRS III scores in PD, the degree and type
Movement Disorders, Vol. 00, No. 00, 2015
7
Z A N J A N I
E T
A L
of motor signs that improve varied among different studies. For example, Lomarev et al.20 found a significant improvement in walking speed, bradykinesia, and gait. Siebner et al.22 found significant improvement in bradykinesia, rigidity, and tremors. However, Khedr et al.19 found improvement in walking speed and mobility but no significant changes in tremors. Therefore, significant improvement on UPDRS III scores does not necessarily translate into a consistent pattern of improvements in all motor signs. Further studies are needed that report different subscores of the UPDRS III to establish whether rTMS improves specific PD motor symptoms. Each study included in this meta-analysis used a different combination of parameters, making determining whether the effects observed were attributable to one factor alone or to moderating factors affecting the parameter under investigation difficult. For example, high- and low-frequency rTMS have been shown to have different effects. Frequencies of 5 Hz and higher have been shown to enhance motor cortex excitability,28,31 whereas lower frequencies rTMS (1 Hz and lower) have been shown to briefly depress cortical excitability.29 Both regular rTMS and theta burst stimulation (a type of rTMS protocol) were included in this systematic review, yet whether this new pattern of stimulation differs behaviorally or biologically from the effects of regular rTMS in patients with PD is unknown. Different stimulus intensities, number of pulses per session, the type of coil used, as well as the number of rTMS sessions applied may also have different effects in PD motor signs. The interactions between various parameters were not considered in this meta-analysis. Additionally, we pooled studies from different countries (Germany, Egypt, United Kingdom, United States of America, Switzerland, Spain, Canada, France, and Japan) that may use different approaches in the treatment of PD. For example, the study of Khedr et al.19 from Egypt involves unmedicated, advanced PD patients, who are rarely seen in more developed counties. These variations in methodology and patient selection likely account for some of the discrepancies in the literature. Other factors that may account for the variations include whether treatment was administered in the on or off medication states,32 genetic polymorphisms of the subjects studied,33 the time of the day the study was conducted, and the different types of sham stimulation employed that may vary in their effectiveness in masking the study assignment. Furthermore, we did not report heterogeneity analysis because of the small number of included studies and consequently its lack of usefulness in this investigation. A nonsignificant result may not be taken as evidence of no heterogeneity because of the low statistical power. The nonsignificant Eggers tests also may
8
Movement Disorders, Vol. 00, No. 00, 2015
be misleading for the same reasons and must be interpreted with caution. Therefore, the existence of publication bias cannot be excluded. Additionally, this meta-analysis did not follow the stringency of the Cochran Review criteria in its study selection. Furthermore, some of the positive results of earlier studies have not been confirmed. Although our findings suggest that rTMS may have an adjunctive role in the treatment of PD, because of low statistical power and other limiting factors discussed earlier, a definite conclusion cannot be drawn with regard to its efficacy. Large randomized placebocontrolled trials assessing the impact of rTMS on motor functions and activities of daily living are needed to establish the long-term and short-term benefits of rTMS in PD.
References 1.
Schapira AH. Science, medicine, and the future: Parkinson’s disease. BMJ 1999;318:311-314.
2.
Benninger DH, Iseki K, Kranick S, Luckenbaugh DA, Houdayer E, Hallett M. Controlled study of 50-Hz repetitive transcranial magnetic stimulation for the treatment of Parkinson disease. Neurorehabil Neural Repair 2012;26:1096-1105.
3.
Lefaucheur JP, Drouot X, Von Raison F, Menard-Lefaucheur I, Cesaro P, Nguyen JP. Improvement of motor performance and modulation of cortical excitability by repetitive transcranial magnetic stimulation of the motor cortex in Parkinson’s disease. Clin Neurophysiol 2004;115:2530-2541.
4.
Benninger DH, Berman BD, Houdayer E, et al. Intermittent thetaburst transcranial magnetic stimulation for treatment of Parkinson disease. Neurology 2011;76:601-609.
5.
Maertens de Noordhout A. General principles for clinical use of repetitive transcranial magnetic stimulation (rTMS). Neurophysiol Clin 2006;36:97-103.
6.
Lefaucheur JP. Principles of therapeutic use of transcranial and epidural cortical stimulation. Clin Neurophysiol 2008;119:2179-2184.
7.
Filipovic SR, Rothwell JC, Bhatia K. Low-frequency repetitive transcranial magnetic stimulation and off-phase motor symptoms in Parkinson’s disease. J Neurol Sci 2010;291:1-4.
8.
Gonzalez-Garcia N, Armony JL, Soto J, Trejo D, Alegria MA, Drucker-Colin R. Effects of rTMS on Parkinson’s disease: a longitudinal fMRI study. J Neurol 2011;258:1268-1280.
9.
Arias P, Vivas J, Grieve KL, Cudeiro J. Controlled trial on the effect of 10 days low-frequency repetitive transcranial magnetic stimulation (rTMS) on motor signs in Parkinson’s disease. Mov Disord 2010;25:1830-1838.
10.
Elahi B, Chen R. Effect of transcranial magnetic stimulation on Parkinson motor function: systematic review of controlled clinical trials. Mov Disord 2009;24:357-363.
11.
Fregni F, Simon DK, Wu A, Pascual-Leone A. Non-invasive brain stimulation for Parkinson’s disease: a systematic review and metaanalysis of the literature. J Neurol Neurosurg Psychiatry 2005;76: 1614-1623.
12.
Movement Disorder Society Task Force on Rating Scales for Parkinson’s D. The Unified Parkinson’s Disease Rating Scale (UPDRS): status and recommendations. Mov Disord 2003;18:738750.
13.
Martinez-Martin P, Gil-Nagel A, Gracia LM, Gomez JB, Martinez-Sarries J, Bermejo F. Unified Parkinson’s Disease Rating Scale characteristics and structure. The Cooperative Multicentric Group. Mov Disord 1994;9:76-83.
14.
Curtin F, Altman DG, Elbourne D. Meta-analysis combining parallel and cross-over clinical trials. I: Continuous outcomes. Stat Med 2002;21:2131-2144.
15.
Cohn J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates, 1988.
R T M S
16.
Zakzanis KK. Brain is related to behavior (p
