Highlights
-
• People with multiple sclerosis (pwMS) and major depressive disorder (MDD) may experience improvements in depressive symptoms following intermittent theta burst stimulation (iTBS).
-
• iTBS treatment was safe and tolerable in pwMS and MDD; no serious adverse events were reported.
-
• pwMS who responded to iTBS also showed higher visuospatial memory and information processing post-treatment in comparison to non-responders.
Introduction
Depression is two to three times more common in people with multiple sclerosis (pwMS) than the general population. Reference Feinstein1 Treatment with antidepressant medication and cognitive behavior therapy offer modest benefits. Reference Feinstein1 Recent research has suggested that multiple sclerosis (MS) can be conceptualized as a network disorder. Reference Tranfa, Petracca and Moccia2 Neuroimaging analyses demonstrated that functional connectivity between MS lesion sites and a pre-established depression circuit correlates with the severity of depressive symptoms in MS, and this association demonstrates greater specificity for depression than for other MS-related neuropsychiatric symptoms. Reference Siddiqi, Kletenik and Anderson3 Furthermore, this MS depression circuit revealed a topographical similarity to an a priori identified depression circuit identified from stroke and penetrating head trauma lesions, Reference Padmanabhan, Cooke and Joutsa4 which aligned closely with the left dorsolateral prefrontal cortex (L-DLPFC)/subgenual cingulate cortex circuit implicated in the therapeutic effects of repetitive transcranial magnetic stimulation (rTMS) and deep brain stimulation for major depressive disorder (MDD). Reference Siddiqi, Schaper and Horn5 This convergence of data suggests that MS depression may be responsive to the same L-DLPFC rTMS targets identified for primary MDD; however, there is currently a lack of rigorous research in this area. Establishing evidence of the safety and effectiveness of rTMS to the L-DLPFC may expand the currently limited evidence-based therapeutic options available for patients with MS experiencing depressive symptoms. Reference Minden, Feinstein and Kalb6,Reference McIntosh, Liu, Allan and Grech7
The goals of this study are to investigate the safety and tolerability of intermittent theta burst stimulation (iTBS) rTMS to the L-DLPFC in improving depressive symptoms in pwMS and MDD. iTBS is an approved form of rTMS that involves brief protocols (∼3 minutes per day) that has been shown to be equal in efficacy to longer (∼37.5 minutes per day) high-frequency rTMS protocols to the L-DLPFC. Reference Blumberger, Vila-Rodriguez and Thorpe8 The antidepressant effects of L-DLPFC iTBS have been well established for MDD. The seminal non-inferiority THREE-D study reported a 49% antidepressant response rate and a 32% remission rate after 6 weeks of iTBS. Reference Blumberger, Vila-Rodriguez and Thorpe8 Meta-analyses of randomized control trials of L-DLPFC iTBS for MDD have subsequently confirmed the superiority of active versus sham stimulation in improving depressive symptoms. Reference Kishi, Ikuta and Sakuma9 Although iTBS rTMS is approved for use in MDD, there have been no safety and efficacy evaluations of this form of rTMS in MS patients with MDD. This study is the first report of iTBS rTMS to the L-DLPFC for depressive symptoms in MS.
It is hypothesized that L-DLPFC iTBS rTMS will be well tolerated by patients with MS and MDD and L-DLPFC iTBS will be associated with improvements in depressive symptoms. An exploratory hypothesis is that improvements in depressive symptoms with L-DLPFC iTBS rTMS will be associated with improvements in cognition function. This study contributes to the evaluation of the safety and tolerability of this potentially more rapid and accessible treatment for those suffering from MS and MDD.
Methods
Sample
Twenty-one pwMS and MDD participated in a single-center, open-label trial of iTBS rTMS delivered for just over three minutes to the L-DLPFC, five days a week for 4 weeks, for a total of 20 treatments. Inclusion and exclusion criteria appear in Table 1. MS and MDD diagnoses were verified by a neurologist and a psychiatrist, respectively.
Inclusion and exclusion criteria
Table 1. Long description
Inclusion criteria for participants include: 1. Men and women ≥18 and ≤70 years of age, inclusive. 2. History of MS confirmed by a neurologist. 3. Patients who are able and willing to give consent and able to adhere to treatment schedule and attend study visits, as determined by study psychiatrist. 4. Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-V) diagnosis of Major Depressive Disorder (MDD). 5. Moderate severity with a Hamilton Depression Rating Scale (HAMD) at least 16. 6. Patients have had no increase or initiation of any psychotropic medication in the 4 weeks prior to screening. 7. Pass the TMS safety screening questionnaire. 8. Women of childbearing potential must agree to use a barrier contraception method throughout the study. Exclusion criteria for participants include: 1. Active substance abuse or dependence in the last three months, except nicotine. 2. Active suicidal intent. 3. Currently pregnant (as determined by history and serum human chorionic gonadotropin (HCG)) or lactating. 4. A diagnosis of Bipolar Disorder. 5. A history of past or current psychotic symptoms. 6. Diagnosis of obsessive-compulsive disorder, post-traumatic stress disorder (current or within the last year), assessed by a study investigator to be primary and causing greater impairment than MDD. 7. Having failed a course of electroconvulsive therapy (ECT) in the current episode or previous episode. 8. Previous trial of rTMS. 9. Personality disorder deemed to be primary pathology. 10. If participating in psychotherapy, must have been in stable treatment for at least 3 months prior to entry into the study, with no anticipation of change in the frequency of therapeutic sessions, or the therapeutic focus over the duration of the study. 11. Clinically significant laboratory abnormality, in the opinion of the investigator. 12. Unstable medical illness. 13. Contraindication to rTMS, e.g. presence of cardiac pacemaker, intracranial implant, or metal in the cranium. 14. Currently on more than 2 mg of lorazepam or equivalent. 15. History of seizures, or currently on anticonvulsant for seizures. 16. Concurrent use of a medication that may lower the seizure threshold, in the opinion of the investigator e.g. stimulant.
MS = multiple sclerosis; TMS = transcranial magnetic stimulation; rTMS = repetitive transcranial magnetic stimulation; DSM-V = Diagnostic and Statistical Manual of Mental Disorders, 5th Edition; ECT = electroconvulsive therapy; HCG = human chorionic gonadotropin.
Primary outcome measure
A reduction in depressive symptoms on the Hamilton Depression Rating Scale (HAMD-17) Reference Hamilton10 from day 0 of treatment to the end of the 4-week trial. This was measured as a categorical response rates of ≥ 50% reduction from baseline HAMD-17. Remission rate was defined as HAMD-17 ≤ 7.
Secondary outcomes
Cognition was evaluated with serial versions of the Brief International Cognitive Assessment for MS (BICAMS) Reference Langdon, Amato and Boringa11 containing the California Verbal Learning Test Second Edition Total Learning (CVLT-II-TL), the Brief Visuospatial Memory Test – Revised Total Learning (BVMT-R-TL) and the Symbol Digit Modalities Test (SDMT). Cognitive domains assessed include verbal memory, visuospatial memory and information processing, respectively. Participants were assessed at baseline and post-treatment.
Treatment
Resting motor threshold (RMT), defined as the minimum energy level delivered in single pulses to the left motor cortex which resulted in muscle activation of the thumb muscle (pollicis brevis) in the right hand in ≥ 50% of stimulations, Reference Blumberger, Vila-Rodriguez and Thorpe8 was empirically determined in each patient. RMT was determined through visual inspection, by consensus of the rTMS physician and the rTMS nurse. After determination of the RMT energy level and location, the coil was then advanced 5.5 cm anteriorly to target the L-DLPFC in a guideline-concordant manner. Reference Trapp, Purgianto and Taylor12 In line with established iTBS rTMS protocols, stimulation was provided in triplet 50 Hz bursts, 2 second trains and with an intertrain interval of 8 seconds, for a total of 20 trains and 600 pulses per session. Reference Blumberger, Vila-Rodriguez and Thorpe8 The energy was delivered at 100% of RMT during the first treatment and then titrated to 110% on treatment day 2 and subsequently to 120% RMT for the remainder of the treatments, as tolerated. Reference Blumberger, Vila-Rodriguez and Thorpe8
Side effects were determined by the Toronto Side-Effect Scale (TSES) Reference Vanderkooy, Kennedy and Bagby13 at weeks 2 and 4 after starting iTBS.
Statistical analysis
Between-group (depression responder versus non-responders) comparisons were undertaken with independent t-tests for continuous data and chi-square tests for categorical data. Paired t-tests were undertaken for longitudinal data (pre-and post-treatment). A between-group comparison of the Expanded Disability Status Scale (EDSS) was assessed with the Mann–Whitney U test. Cohen’s d and 95% confidence intervals were provided where appropriate. Z-scores were obtained from Canadian normative data Reference Walker, Marino, Berard, Feinstein, Morrow and Cousineau14 for demonstrating the results graphically. All analyses were conducted using SPSS v.29.
Results
Demographics and disease-related data
Average age was 45.53 (SD = 10.82) years; 76.5% were female; 94.1% had relapsing–remitting MS (RRMS); 94.1% were taking concurrent antidepressant medication; mean disease duration was 12.88 (SD = 7.64) years; mean years of education was 15.65 (SD = 2.21); median EDSS was 2.0 (range = 0–6.5). The mean time between treatment and assessment was 2.94 days (SD = 10.29). There were 10 (58.8%) treatment responders and seven (41.2%) non-responders (n = 17).
Demographic and disease-related comparisons between treatment responders and non-responders
There were no statistically significant between group baseline differences in age, sex, education, disease course and duration, EDSS and disease-modifying drug use (see Table 2). However, estimates of effects revealed that non-responders were older on average by 8.33 years (d = −0.81) and had MS for 7.97 years (d = −1.19) longer than responders.
Demographic and neurological comparisons between responders and non-responders
Table 2. Long description
Table 2 compares baseline demographic and neurological characteristics between responders (n = 10) and non-responders (n = 7). Responders were younger on average (42.1 vs. 50.4 years) and had shorter disease duration (9.6 vs. 17.6 years) than non-responders, although neither difference was statistically significant (p = .121 and p = .061, respectively). Responders also had slightly more years of education (16.1 vs. 15.0 years; p = .327). Median EDSS scores were similar between groups (2.5 vs. 2.0; p = .364). Six responders (60.0%) and all non-responders (100.0%) were female (p = .103). All responders had relapsing-remitting multiple sclerosis (RRMS), while one non-responder had primary progressive multiple sclerosis (PPMS) (p = .412). All responders and 71.4% of non-responders were using a disease-modifying therapy (DMT) (p = .154). Overall, no statistically significant baseline differences were observed between responders and non-responders.
a Fisher’s exact test. CI = confidence interval; SD= standard deviation; EDSS = Expanded Disability Status Scale; RRMS = relapsing–remitting multiple sclerosis; PPMS = primary progressive multiple sclerosis; DMT = disease-modifying therapy.
Baseline psychometric comparisons between treatment responders and non-responders
There were no statistically significant baseline differences in scores on the HAMD-17, CVLT-II-TL, BVMT-R-TL and SDMT (see Table 3).
Pyschometric comparisons between responders and non-responders at baseline and post-treatment
Table 3. Long description
Table 3 compares psychometric outcomes between responders (n = 10) and non-responders (n = 7) at pre- and post-treatment. At baseline, groups did not differ significantly on depression severity, verbal learning, visuospatial learning, or processing speed (all p > .05). The largest baseline difference was for processing speed, where responders scored higher than non-responders (51.7 vs. 40.3), but this was not statistically significant (p = .068). At post-treatment, responders showed lower depression scores than non-responders (6.5 vs. 19.6; p < .001). Responders also scored higher on visuospatial learning (25.9 vs. 19.3; p = .039) and processing speed (55.0 vs. 31.7; p = .003). No significant post-treatment difference was observed for verbal learning (53.3 vs. 45.1; p = .245). Overall, responders and non-responders showed similar baseline performance, but responders demonstrated significantly lower depression severity and better visuospatial memory and processing speed post-treatment.
CI = confidence interval; SD= standard deviation; HAMD-17 = Hamilton Depression Rating Scale; CVLT-II-TL = California Verbal Learning Test Second Edition Total Learning; BVMT-R-TL = Brief Visuospatial Memory Test – Revised Total Learning; SDMT = Symbol Digit Modalities Test.
Post-treatment psychometric comparisons between treatment responders and non-responders
Depression scores decreased significantly from baseline to post-treatment with a mean reduction of 10.24 points (d=1.55, 95% CI 0.83 to 2.25, P < .001). Responders had significantly lower HAMD-17 scores post-treatment (d = −3.01, P < .001). Seven participants (41.2%) remitted post-treatment. Scores on the HAMD-17 declined by 14.80 points for responders from pre-(M = 21.30, SD = 4.90) to post-treatment (M = 6.50, SD = 2.64, t [9] = 11.88, P < .001, d = 3.76) and by 3.71 points for non-responders from baseline (M = 23.29, SD = 4.39) to post-treatment (M = 19.57, SD = 6.05, t [6] = 3.36, P = .015, d = 1.27).
Responders had significantly higher BVMT-R-TL (d = 1.11, P = .039) and SDMT scores (d = 1.74, P = .003) than non-responders post-treatment. The CVLT-II-TL scores for the responders improved by 4.30 points from pre-(M = 49.00, SD = 10.60) to post-treatment (M = 53.30, SD = 9.81, t [9] = −1.62, P = .139, d = −0.51). Responders showed higher BVMT-R-TL scores by 1.30 points from baseline (M = 24.60, SD = 4.48) to post-treatment (M = 25.90, SD = 5.30, t [9] = −0.95, P = .366, d = −0.30). Responders significantly improved scores on the SDMT by 3.30 points from pre-(M = 51.70, SD = 12.19) to post-treatment (M = 55.00, SD = 13.29, t [9] = −3.01, P = .015, d = −0.95).
Non-responders showed decreased scores on the CVLT-II-TL by 2.00 points from baseline (M = 47.14, SD = 5.84) to post-treatment (M = 45.14, SD = 17.99, t [6] = 0.40, P = .701 d = 0.15). Non-responders’ scores also decreased on the BVMT-R-TL by 3.71 points from pre-(M = 23.00, SD = 3.92) to post-treatment (M = 19.29, SD = 6.78, t [6] = 1.56, P = .169 d = 0.59). Lastly, non-responders scored lower on the SDMT by 8.57 points from baseline (M = 40.29, SD = 11.13) to post-treatment (M = 31.71, SD = 13.49, t [6] = 1.65, P = .149 d = 0.63).
Serial CVLT-II-TL, BVMT-R-TL and SDMT z-score data are shown in Figure 1.
Mean (a) CVLT-II-TL, (b) BVMT-R-TL and (c) SDMT z-scores pre-and post-treatment by clinical response group. Error bars represent standard deviations. CVLT-II-TL = California Verbal Learning Test Second Edition Total Learning; BVMT-R-TL = Brief Visuospatial Memory Test – Revised Total Learning; SDMT = Symbol Digit Modalities Test.
Figure 1. Long description
Panel A displays mean CVLT-II-TL z-scores at pre-treatment and post-treatment for two groups: responders (solid blue line) and non-responders (dashed orange line). The vertical axis shows z-scores, with higher values indicating better performance. At pre-treatment, both groups have similar mean scores of approximately -1.1. After treatment, the responder group shows improvement, with the mean increasing to about -0.6. In contrast, the non-responder group remains essentially unchanged at approximately -1.1. Error bars are shown for each time point and indicate substantial variability within both groups and overlap. Overall, the graph suggests improved verbal learning performance over time among responders, with little change among non-responders. Panel B displays mean BVMT-R-TL z-scores at pre-treatment and post-treatment for two groups: responders (solid blue line) and non-responders (dashed orange line). The vertical axis shows z-scores, with higher values indicating better performance. At pre-treatment, both groups have similar mean scores near -0.5. Following treatment, the responder group’s mean score rises slightly to approximately -0.3. The non-responder group’s mean score declines to approximately -1.1, indicating worsening performance. Error bars indicate variability in scores and overlap considerably across time points. Overall, the pattern suggests divergent trajectories, with responders showing slight improvement and non-responders showing decline in visuospatial learning performance. Panel C displays mean SDMT z-scores at pre-treatment and post-treatment for responders (solid blue line) and non-responders (dashed orange line). The y-axis shows z-scores, with higher values reflecting better processing speed performance. At pre-treatment, responders have a mean score of approximately -1.0, while non-responders have a lower mean score of approximately -1.9. After treatment, responders improve to roughly -0.5, whereas non-responders decline further to approximately -2.4. Error bars indicate variability within both groups and are relatively large. Overall, the graph shows improved processing speed among responders and worsening performance among non-responders over the treatment period.
Two participants dropped out of the study: one due to discomfort and withdrawing consent, and the other due to the inability to establish RMT because of concurrent benzo usage. Both participants received another rTMS protocol outside of the study. Two participants were lost to follow-up. There were no serious adverse events.
Discussion
Current research has increasingly conceptualized psychiatric symptoms in MS as arising from disruptions in underlying neural networks. Evidence suggests that functional connectivity between MS lesion sites and a priori depression circuit derived from patients with focal brain lesions and other neuropsychiatric disorders is specific to depression symptomatology versus other MS-related symptoms. Reference Siddiqi, Kletenik and Anderson3 This putative MS depression circuit includes the L-DLPFC, the most commonly used brain target of rTMS for the treatment of MDD. Reference Ng, Wang, Nestor, Lipsman and Giacobbe15
Our results revealed that a little over half (58.8%) the participants achieved an antidepressant response after 4 weeks of L-DLPFC iTBS. While a placebo effect is possible, our findings are consistent with the known trans-diagnostic effects of iTBS in treating depressive symptoms across a variety of neuropsychiatric disorders. Reference Blumberger, Vila-Rodriguez and Thorpe8 The tolerability of iTBS in treating depressive symptoms in our study adds to the safety profile of iTBS in pwMS. Previous MS research has reported the use of the iTBS to the motor cortex to treat spasticity, also with only limited side effects. Reference Boutière, Rey and Zaaraoui16 Taken together, these results suggest the iTBS rTMS is a promising and safe treatment strategy for depressive symptoms in pwMS.
An additional potential benefit to successfully treating depressive symptoms with rTMS may be improved cognition. Prevalence rates of cognitve dysfuction in pwMS range from 45% in RRMS to 70%–80% in people with progressive disease. Reference Feinstein1 Cognitive rehabilition, considered the treatment of choice, may produce moderate benefits, most notably for memory and processing speed impairment. Reference Lampit, Heine and Finke17 Our pilot data suggest similar indices improving with rTMS treatment alongside depressive symptom lifting. In particular, statistically significant differences in visual memory and processing speed were seen between responders and non-responders. What is noteworthy about the serial processing speed data is that significant improvement occurred in the group that had a higher baseline SDMT, albeit not statistically. This turned out to be the group whose depressive symptoms responded to rTMS. Reasons for this are unclear and cannot be gleaned from pilot data, but it may be that individuals with MS and a higher cognitive reserve who become depressed are more responsive to treatment with rTMS. Future research is need to confirm or refute this intriguing possibility.
While studies have confirmed the cognitive safety of rTMS, Reference Derakhshan, Lipsman and Feinstein18 the salutary effect of rTMS on cognition has not been fully elucidated and remains an area of active study. A systematic review and meta-analysis examining the effects of rTMS to the L-DLPFC in a diagnostically heterogeneous sample of 174 studies encompassing 7905 patients with both psychiatric and neurological disorders found a small effect size for improvements in global cognition, declarative memory and working memory in the active versus sham rTMS groups; however, there were no effects on attention and language. Reference Kan, Padberg and Giron19 Another systematic review of the neurocognitive effects of theta burst stimulation for treatment-resistant depression in adults revealed cognitive improvement across specific domains including executive functioning, attention, working memory and verbal learning. Reference Hutnyan and McClintock20 In contrast, other systematic reviews and meta-analyses which were restricted to comparing the effects of active versus sham rTMS on cognitive domains in those with MDD found no significant improvement in any cognitive domain. Reference Fu, Ren, Lei, Zhang and Zhang21,Reference Xia, Huang, Chen, Lei and Yue22 The inconclusive findings for the effects of rTMS on cognitive systems may be the result of heterogeneity in subjects studied and stimulation protocols. As a result, it remains uncertain whether improvements noted with rTMS represent a direct effect of the stimulation on cognition, or whether improvements result from the indirect effect of improvement in mood with rTMS delivered to the L-DLPFC. Reference Derakhshan, Lipsman and Feinstein18,Reference Toffanin, Folesani and Ferrara23
This pilot study has some limitations. Our sample was limited in size and was primarily comprised of individuals with RRMS and low EDSS scores.
Conclusion
In summary, 4 weeks of L-DLPFC iTBS can improve depressive symptoms in pwMS and with it, aspects of cognition too. Treatment was generally well tolerated. In order to address the potential placebo response, these pilot data point the way forward to a more substantial randomized controlled trial that could have considerable clinical significance.
Data availability statement
The data that support the findings of this study are available upon reasonable request.
Acknowledgements
This project was funded by the Harquail Centre for Neuromodulation.
Author contribution
AF and PG were responsible for study design, subject recruitment, data analysis and manuscript preparation. AB, SN, DD, CH and NL were responsible for study design and manuscript preparation.
Funding statement
None.
Competing interests
Authors AB, SN and DD declare no potential conflicts of interest. Author PG reports salary support from the Kerr Chair in Interventional Psychiatry and the University of Toronto, Department of Psychiatry, Academic Scholars Award. Author CH reports consulting fees from Abbott and Epiminder. Author NL reports consulting fees as Chair of the Scientific Advisory Board for Grey Matter Neurosciences. Author AF reports grants from MS Canada, book royalties from Johns Hopkins University Press, Cambridge University Press, Bloomsbury Publishing and G Editions, and speaker honoraria from Novartis, Merck and Roche.
Ethical standard
All subjects provided informed consent prior to participation. The study was approved by the research ethics board at Sunnybrook Health Sciences Centre.
Open access
