Comparative Effectiveness of Repetitive Transcranial Magnetic Stimulation in Unipolar and Bipolar Depression

Introduction

In Canada, repetitive transcranial magnetic stimulation (rTMS)is approved for the treatment of major depressive disorder (MDD) and is considered a first-line treatment for major depressive episodes (MDE) after at least 1 failure with pharmacotherapy.¹ The evidence is much stronger in MDD¹ versus bipolar depression (BD).² Some trials have examined the efficacy of rTMS in BD specifically, and many of the foundational trials included unipolar depression and BD. The current evidence suggests that rTMS is efficacious in BD, though absolute rates of clinical response differ significantly across studies, with some reporting between 0% and 100% clinical response in both active and sham groups.³ Yet, there are limited real-world data on the relative effectiveness of rTMS in treating MDD and BD using clinician-rated measures of depressive symptoms.

Methods

We performed a retrospective chart review of outpatients with a DSM-IV-TR or DSM-5 diagnosis of acute MDE treated with rTMS between 2012and 2016 (hospital clinic’s public funding discontinued in 2016). Eligibility criteria included failure to respond to or tolerate first-line pharmacotherapy, and exclusion criteria included seizure disorder, intracranial ferromagnetic device, acute suicidality, psychosis, and pregnancy/lactation. Seventy-six patients received between 2 and 6 weeks, based on patient factors, of daily high-frequency rTMS (10 Hz, 3,000 pulses, 4 s trains and 26 s intertrain interval, 120% resting motor threshold) using a Magstim Super Rapid-2 device (Magstim Company Ltd., UK). Stimulation was delivered to the left dorsolateral prefrontal cortex (L-DLPFC) as determined using the 6 cm “rule” parasagittally anterior to the motor cortex. Patients remained on stable pharmacological regimens throughout their treatments.

Our primary outcome was clinical response defined as a reduction of ≥50% in clinician-rated depressive symptoms using the 21-item Hamilton Depression Rating Scale. Our secondary outcome was the change in clinician-rated depressive symptoms. Statistical analyses were performed in JMP14. Shapiro Wilkes’s tests of normality were performed before independent 2-tail t tests were used to test statistical significance between categories of depression forage, MDE, medications, baseline HRDS-21, rTMS treatments, and % reduction in HRDS-21. For gender, clinical response, and benzodiazepine, gabapentin/pregabalin, and other anticonvulsant usage, Pearson’s chi-square coefficient was used to test for difference between the groups ( Table 1 ). Across all tests, threshold for significance was set at α < 0.05.

Table 1.

Comparison of Unipolar and Bipolar Patients.

	Unipolar Depression (N = 63; mean ± SEM or n [%])	Bipolar Depression (N = 13; mean ± SEM or n [%])	Test Statistic	P Value
Age (years)	49.3 ± 1.6	54.0 ± 3.1	1.37	0.18
Gender F/M	34F/27M	5 (8)	1.28	0.26
Current MDE (months)	5.1 ± 0.82	5.3 ± 1.32	0.11	0.91
Psychotropic medications (#)	1.71 ± 0.18	2.50 ± 0.47	1.56	0.14
Benzodiazepine usage	21/61	2/13	1.81	0.18
Gabapentin/pregabalin usage	4/61	1/13	0.022	0.88
Other anticonvulsant usage	5/61	4/13	5.11	0.024
Baseline HRDS-21	19.4 ± 0.75	18.2 ± 1.39	0.75	0.46
rTMS treatments (#)	28.7 ± 0.66	26.2 ± 1.95	1.22	0.24
HRDS-21 reduction (%)	37.6 ± 3.3	25.5 ± 6.2	1.73	0.10
Clinical response (%)	24 (39.3%)	1 (7.7%)	4.80	0.029

Note. Total sample size n = 76. Data available at treatment end (n) were 61 for the unipolar and 13 from the bipolar groups. Numbers are reported as mean ± standard error of the mean unless otherwise specified. MDE = major depressive episode; HRDS-21 = Hamilton Depression Rating Scale; rTMS = repetitive transcranial magnetic stimulation.

Results

Baseline demographic and clinical characteristics of the sample are described in Table 1. Of 76 patients treated with rTMS, 7 had bipolar disorder type I and 6 had bipolar disorder type II. Two individuals with MDD discontinued treatment and clinician-rated depressive symptoms were not available at treatment end. Patients with MDD and BD did not differ on important demographic variables or number of rTMS treatments, and 89.5% of patients received ≥20 treatments.

At the conclusion of treatment, there was a significantly lower rate of clinical response among patients with BD than MDD (7.7% vs. 39.3%, χ² = 4.79, df = 1, P = 0.029; OR = 0.12; 95%CI, 0.01 to 1.05). This was paralleled in depressive symptoms (mean difference 12.08 ± 7.74(SD), t(72) = 1.73, P = 0.10) but did not reach statistical significance. No significant adverse effects occurred, including seizure or hypomanic/manic switching. There were no reported dropouts from side effects.

Discussion

Our naturalistic data with an adequately sized cohort of participants suggest that patients with BD are less likely to achieve clinical response than those with unipolar depression with high-frequency L-DLPFC rTMS. Importantly, rates of response in our MDD subsample align with the 40% to 55% response rates observed in large randomized controlled rTMS trials.¹ This suggests that differences between unipolar and BD participants are unlikely due to methodological error or bias.

While rTMS for BD appears to be efficacious, the randomized sham-controlled literature is much smaller than that in unipolar depression, with few dedicated trials.³ In cohort and open-label studies, BD has either been associated with comparable rates of response as unipolar depression using the same stimulation parameters in our study⁴ or marginally lower rate of response with heterogeneous stimulation parameters.⁵ This latter study, with a large subsample of 146 BD patients, had substantial numbers of participants receiving low-frequency right DLPFC rTMS and very few receiving L-DLPFC rTMS⁵ which may account for some difference in outcome compared to our study. Previous meta-analysis has shown that low-frequency R-DLPFC appear to have superior efficacy over sham rTMS, compared to the efficacy over sham when utilizing faster frequencies of L-DLPFC rTMS in those with BD.³ Moreover, both of these studies favored self-reported depressive symptoms rather than clinician-rated symptoms, and accordingly, the overall rates of response could exceed what large clinician-rated samples would suggest.

This study is subject to limitations. Foremost is its naturalistic and retrospective nature; however, this is tempered by clinician-rated measures at baseline and at treatment end. There are few patients with BD, and accordingly, small study bias may limit our findings external generalizability. Additional research should substantiate differential effectiveness of rTMS between those with MDD and BD, and the optimal stimulus parameters for BD.