Optimally Combining Transcranial Magnetic Stimulation with Antidepressants In Major Depressive Disorder: A Systematic Review and Meta-Analysis

Abstract

There is a critical knowledge gap in optimally combining transcranial magnetic stimulation (TMS) and antidepressants to treat patients with major depressive disorder (MDD). TMS is effective in treating MDD in patients who have failed at least one antidepressant trial, with accelerated protocols showing faster remission in treatment-resistant depression (TRD). Although clinicians routinely augment antidepressants with TMS, there is a knowledge gap in stopping versus continuing antidepressants or the dosing strategies when starting or tapering TMS. These considerations are important when considering maintenance TMS (delivered alone or in combination with suitable antidepressants) to maintain remission in MDD after the index course of TMS. As the first step towards filling this knowledge gap, we reviewed randomized controlled trials (RCTs) and open-label trials from 2 databases (PubMed/Medline and EMBASE) that compared active TMS combined with a pre-specified antidepressant dosed in the same manner for adults with MDD versus sham TMS combined with the same antidepressant as in the active arm. All studies were published between January 1, 2000, and December 31, 2023. We excluded case reports, case series, and clinical studies that augmented TMS with antidepressants and vice versa. We found 10 RCTs (n=654 participants) and performed a meta-analysis. This showed active TMS combined with pre-specified antidepressants had greater efficacy for MDD treatment than sham TMS combined with the same antidepressants as in the active arm (Hedge’s g = 1; 95% CI [0.27, 1.73]). The review and meta-analysis indicate greater short-term efficacy in combining antidepressants with TMS from the get-go in MDD. Given the increasing role of accelerated TMS protocols in expediting remission in MDD and the results of our meta-analysis, we advocate for RCTs examining the short-term and long-term effects of various antidepressant classes on these TMS protocols in MDD. This can also optimize and individualize maintenance TMS protocols to prevent relapse in MDD.

Introduction

Major depressive disorder (MDD) imposes significant disability and economic burdens on patients (Friedrich, 2017). Antidepressants are more effective compared to placebo in the treatment of MDD, with effect sizes ranging from 1.15–1.55 (Cipriani et al., 2018). Although there is a variety of definitions for treatment-resistant depression (TRD), the consensus definition encompasses a lack of response to adequate trials of one or more antidepressants from the same or different antidepressant classes (McIntyre et al., 2023). TRD has an estimated 30–55% prevalence in the general population (McIntyre et al., 2023; Zhdanava et al., 2021).

Transcranial magnetic stimulation (TMS) is a treatment option in patients with MDD who have failed 1–2 antidepressant trials (McClintock et al., 2018; McIntyre et al., 2023). With the advent of accelerated protocols like Stanford Neuromodulation Treatment (SNT) for MDD, TMS is increasingly becoming a treatment option for TRD (McIntyre et al., 2023). TMS does not have adverse cognitive effects and, consequently, has a greater patient preference (Berlim et al., 2011; Berlim et al., 2013; Magnezi et al., 2016). TMS is generally safe and well-tolerated (Perera et al., 2016).

There is a significant knowledge gap in clinical decision-making when initiating or stopping TMS in patients on antidepressants for TRD (McClintock et al., 2018). Possible options include 1) tapering the antidepressant(s) when initiating TMS and reinitiating them after the index TMS course, 2) continuing the antidepressant(s) with TMS, which avoids the need to re-initiate them after the index TMS course, or 3) augmenting the patient’s current antidepressant regimen with an SNRI or second generation antipsychotic (based on the previous trial history and degree of treatment resistance) while receiving TMS to expedite response and remission. Except for avoiding antidepressants that can lower seizure thresholds (bupropion and imipramine), there are no consensus guidelines on strategies to combine antidepressants with TMS for MDD (McClintock et al., 2018). Hence, it is critical to think about strategies to leverage the benefits of both TMS and antidepressants in MDD.

Previous studies that combined electroconvulsive therapy (ECT) with antidepressants have been inconclusive in the combination strategy’s benefits compared to ECT alone. While no differences were seen in the short term from a meta-analysis (Pluijms et al., 2021), one notable study showed greater benefits with nortriptyline than venlafaxine or placebo (Sackeim et al., 2009). Current evidence for the strategy’s long-term benefits comes from retrospective chart reviews, which have also been inconclusive (Kurimoto et al., 2021; Song et al., 2015). ECT differs remarkably from TMS paradigms regarding delivery methods, adverse effects, and patient acceptance, so extending the data from ECT to TMS is difficult.

The mechanisms of antidepressants and TMS in the treatment of psychiatric disorders are at least partially based on changes in neuroplasticity (Paulzen et al., 2014). Although antidepressants are known to be modulators of neurotransmitters such as serotonin, dopamine, and noradrenaline (Stahl, 2013), other proposed explanations about the mechanism of action are changes in network connectivity and neuroplasticity. For example, antidepressants have been shown to have synaptogenic, neurogenic, and neuroprotective effects in their treatment of psychiatric disorders (Paulzen et al., 2014). When used as an investigative tool, various single-pulse TMS paradigms have demonstrated that antidepressants alter cortical excitability and neuronal long-term potentiation (LTP) and long-term depression (LTD) (Minzenberg and Leuchter, 2019).

TMS affects neuronal plasticity by modulating long and short-term potentiation of neurons (Klomjai et al., 2015). Because both TMS and antidepressants affect neuroplasticity, perhaps their interactions in treating MDD may plausibly have an additive or synergistic effect. Two previous systematic reviews examined whether active TMS differentiated from sham TMS as an augmentation agent with antidepressants in decreasing HAMD scores in MDD (Brunoni et al., 2009; Liu et al., 2014).

Although both reviews showed the benefits of augmenting antidepressants with TMS, they fell short of examining the relative contributions of TMS and pre-specified antidepressants when started from the get-go in patients with MDD (Brunoni et al., 2009; Liu et al., 2014). Therefore, examining the effect of the combination on response and remission in MDD needs blinded RCTs wherein both treatment options (TMS and antidepressants) are administered simultaneously, dosed uniformly, and compared to sham TMS and placebo. A futuristic goal of this approach would also be to examine whether antidepressant continuation after the index TMS course modulates the need for maintenance TMS sessions.

The primary objective of this review is to summarize evidence from studies that rigorously combined TMS and antidepressants in the treatment of MDD. Hence, we hypothesized that antidepressants combined with active TMS would show a greater reduction in depressive symptoms than sham TMS combined with antidepressants.

Methods

PubMed/Medline and EMBASE were searched using the terms “transcranial magnetic stimulation” or “theta burst stimulation” along with terms denoting psychiatric drug classes (e.g., “SSRI,” “antipsychotic,” etc.) as well as individual psychiatric drug names (e.g., “fluoxetine,” “clozapine,” etc.). Drug classes that have utility in the treatment of MDD (antidepressants, antipsychotics, mood stabilizers, and benzodiazepines) were included in our search terms. Search terms used are listed in the supplementary material. PRISMA guidelines were followed to guide inclusion criteria for the review (Moher et al., 2009). All studies were published between January 1, 2000, and December 31, 2023. All studies were published in English and screened by title, abstract, and full text by two authors (PC and GR) before a decision was made to include or exclude the study.

We reviewed studies that compared active TMS combined with antidepressants dosed in the same manner for all trial participants versus sham TMS combined with the same antidepressants as in the active arm. We included studies that met the following inclusion criteria: 1) Administered TMS concurrently with antidepressant medication titration with a comparator placebo arm that combined sham TMS with the same antidepressant medication as in the active arm; 2) Participants in the study met criteria for major depressive disorder (MDD) as defined by any edition of the Diagnostic Statistical Manual of Disorders (DSM) or International Classification of Diseases (ICD); 3) The study used quantitative measures that assess symptom improvement in MDD (e.g., Hamilton Depression Rating Scale); 4) The study involved adult participants ≥ 18 years of age; and 5) The study needed to be an RCT. We excluded studies that augmented antidepressants with TMS or used TMS as an adjunctive intervention because antidepressant effectiveness was suboptimal in the treatment of MDD. We excluded open-label trials, case reports, and case series. We also excluded studies that administered either TMS or antidepressants as maintenance treatment for MDD in a non-concurrent fashion.

We performed a meta-analysis with the RCTs we included in the review. In the meta-analysis, the primary outcome measure studied was a reduction of symptom severity as measured by the Hamilton Depression Rating Scale (HDRS). A random effects model was used because high heterogeneity between studies was expected. The effect size was calculated for the RCTs as described by Morris (Morris, 2008), wherein the pooled pretest standard deviation was used to weight the differences between pre-post-means. Meta-analysis and calculations of heterogeneity (I²) were performed using RevMan 5.4 ((RevMan), 2020). Publication bias was assessed using Egger’s test for funnel plot asymmetry (Egger et al., 1997) (using ‘metabias’ from meta package in Rstudio).

Results

Of the 750 unique articles screened, 10 studies were included in the qualitative synthesis. Figure 1 shows the PRISMA diagram of how selected studies were included in the review and meta-analysis. Studies we excluded comprised:- 1) experimental human TMS studies measuring cortical excitability in response to psychotropics, 2) clinical trials combining TMS and antidepressants in conditions other than MDD [such as obsessive-compulsive disorder (OCD), tinnitus, migraine], 3) previous systematic reviews on utility of TMS for various disorders, 4) TMS combined with psychotropics other than antidepressants, including antipsychotics and benzodiazepines, 5) animal studies utilizing electroconvulsive stimulation (ECS), and 6) studies that augmented antidepressants with TMS or used TMS as an adjunctive intervention to antidepressants in treatment of MDD. We assessed the full texts of 10 articles for inclusion. Table 1 in the main manuscript summarizes the key characteristics of these studies included in the meta-analysis. Egger’s test for publication bias was not significant, indicating a lack of publication bias for studies included in the meta-analysis [t(9)=1.92, p=0.09] (See Figure 2) (Egger et al., 1997).

Figure 1.

PRISMA diagram of the study selection process.

Table 1.

Summary of included studies in Group 1.

Study	TMS Freq	Number of TMS sessions	MT	TMS target	Drug(s) studied (dose in mg)	Age & M: Fa	Population studied and age of onset	Drop out	n	Results
Randomized Controlled Trials (RCTs) that administered antidepressants and TMS concomitantly for major depressive disorder (MDD)
SSRI
Garcia-Toro et al., 2001	20 Hz (1200 pulses per session) or sham TMS	10 consecutive working days	90 %	Left dorsolateral prefrontal cortex (dlPFC) (5 cm anterior to motor cortex hotspot for contralateral abductor pollicis brevis)	Sertraline (50 mg/ d adjusted as needed). Mean daily dose not available.	43.2 (13.1), 5:6 45 (18.3), 5:6	Patients with major depression without a trial of sertraline for current depressive episode	6	22	Active TMS + sertraline did not differ from sham TMS + sertraline in score changes on Hamilton depression rating scale (HDRS), Beck Depression Inventory (BDI) and global clinical inventory (GCI).
Poulet et al., 2004	10 Hz or sham TMS. Number of pulses not available	10 consecutive working days	80 %	Left dlPFC (5 cm anterior to motor cortex according to EEG 10–20 system)	Paroxetine (20 mg/ d), dosing was fixed and not adjusted.	NA	Patients with non- resistant major depression	NR	19	Active TMS + paroxetine did not differ from sham TMS + paroxetine in HDRS score changes.
Bretlau et al. 2007	8 Hz (1289 pulses per session) or sham TMS	15 consecutive working days	90 %	Left dlFPC	Escitalopram (10 mg/d in week 1, 20 mg/d from weeks 2–12); dosing was fixed and not adjusted.	53.1 (10.1), 7:15 57.8 (10), 10:13	Patients with medication-resistant major depression	6	45	Active TMS + escitalopram was significantly superior to sham TMS + escitalopram at four time points - weeks 2 and 3, and at follow up during weeks 5 and 8.
Huang et al., 2012	10 Hz (800 pulses per session) or sham TMS	10 (5 sessions per week for 2 weeks)	90 %	Left dlFPC	Citalopram (20 mg/ d, adjusted as needed). Mean daily dose was 32.1 mg in active TMS group and 34.3 mg in sham TMS group. After two weeks, dosing was 40 mg if HDRS reduction was <50 % in the first two weeks.	32.8 (7.3), 9:19 31.4 (7.3), 8:20	Patients with firstepisode major depression, The age of onset was 31.8 years (7.4) in the active TMS group and 30.5 years (7.5) in the sham TMS group.	4	56	Active TMS + citalopram was significantly superior to sham TMS + citalopram at weeks 2 and 4.
Wang et al. 2017	10 Hz (800 pulses per session) or sham TMS	20 (5 sessions per week for 4 weeks)	80 %	Left dlFPC (5 cm anterior to the motor cortex hotspot for right abductor pollicis muscle)	Paroxetine for 8 weeks. Dose of 10 mg during week 1 and increased to 20–30 mg from day 8 till four weeks. Participants were also offered paroxetine treatment alone for an additional 4 weeks. Average daily dose in the active group was close to 30 mg, in the sham group it was between 30 and 35 mg.	28.8 (8.5), 12:10 30.1 (9.5), 11:10	Patient with first episode MDD. The study was divided into phases 1 and 2. In Phase 1, the participants were randomly assigned in a 1:1 ratio to either active rTMS (n = 22) or sham rTMS (n = 21) five times per week combined with paroxetine for 4 consecutive weeks.	5	43	Active TMS + paroxetine resulted in a significant decrease in HDRS scores compared to sham TMS + paroxetine (both at end of week 1 and week 4). At end of week 8, the difference in HDRS scores was not significant.
Pan et al., 2020	10 Hz (6000 pulses per session) or sham TMS	7 consecutive days	100 %	Left dlFPC (Brodmann area 46 based upon individual’s MRI image and MRI generated 3D curvilinear reconstruction of brain)	Escitalopram (10 mg/d), dosing was fixed and not adjusted.	18.1 (3.9), 2:19 21.4 (6.8)5:16	Patients with treatment-naïve unipolar major depressive disorder	8	42	Active TMS + escitalopram showed significantly greater antidepressant effect compared to sham TMS + escitalopram.
SNRI
Rossini et al., 2005	15 Hz (900 pulses per session) or sham TMS.	10 consecutive working days (over two weeks)	100 %	Left dlPFC (5 cm anterior to motor cortex hotspot for contralateral abductor pollicis brevis)	Six groups of treatment assignment. Combination of TMS or sham with medications escitalopram, sertraline or venlafaxine titrated to doses of 15 mg, 150 mg, and 225 mg respectively over two weeks. TMS lasted two weeks and medications continued for five weeks.	48.4 (13.7) 11:39 46.4 (12.1) 9:40	Patients with MDD (DSM-IV), unipolar with no history of mania or seizures. All patients had failed one antidepressant trial. The age of onset was 36.8 (14.8) in the active TMS group and 35.3 (13.3) in the sham TMS group.	3	99	Venlafaxine combined with active TMS had the greatest remission rates (80 %) at end of week 5 compared to sertraline (71.4 %) and escitalopram (68.7 %). In the sham TMS arm, sertraline had the greatest remission rates (66.5 %) compared to venlafaxine (50 %) and escitalopram (46.7 %).
Brunelin et al., 2014	1 Hz (360 pulses per session) or sham TMS	5 consecutive working days per week for 2–6 weeks (until remission)	120 %	Right dlPFC (6 cm anterior to motor cortex hotspot for contralateral left thenar muscle activation)	Venlafaxine (150 mg/d adjusted to 225 mg/d if needed) Mean daily dose was 179 mg across three groups.	53.3 (11.3) 20:34 56.2 (9.9) 16:35 54.2 (11.9) 16:34	Patients with a single episode or recurrent unipolar non- psychotic major depression	15	155	Active TMS + venlafaxine, active TMS + placebo and sham TMS + venlafaxine showed no significant differences in antidepressant effect.
Herwig et al., 2007	10 Hz (2000 pulses per session) or sham TMS	15 consecutive working days	110 %	Left dlPFC (F3 position according to EEG 10–20 system)	Choice of venlafaxine (75 mg/ d adjusted as needed) or mirtazapine (15 mg/ d adjusted as needed).	50 (15) 18:44 49 (13) 33:32	Patients experiencing a moderate to severe depressive episode, including patients with bipolar depression. The age of onset was 38 (16) in active TMS group and sham TMS group.	15	127	Active TMS + antidepressant did not differ from sham TMS + antidepressant in score changes on HDRS, BDI and Montgomery Asberg Depression Rating Scale (MADRS).
TCA
Rumi et al., 2005	5 Hz (1250 pulses per session) or sham TMS	20 (5 per week for 4 weeks)	120 %	Left dlPFC (5 cm anterior to motor cortex hotspot for contralateral abductor pollicis brevis)	Amitriptyline (intended dose of 150 mg after titration, however adjusted as tolerated and mean dose was 110.2 mg in active TMS group and 109.4 mg in sham TMS group)	39.3 (12.8) 3:19 38.9 (8.8) 4:20	Patients with non- psychotic major depression	NR	46	Active TMS + amitriptyline was significantly superior to sham TMS + amitriptyline at all four time points (week 1, week 2, week 3 and week 4).

TMS Freq – Frequency of TMS; MT – resting motor threshold; n – sample size; NA – Not applicable as the study was a retrospective one; NR – Not reported; DLPFC –

dorsolateral prefrontal cortex; iTBS – Intermittent theta burst stimulation.

^aMean age in years (SD) and the ratio of males to females in active TMS plus medication group followed by mean age in years (SD) and the ratio of males to females in sham TMS plus medication group. For Brunelin et al., 2014, the first set of values belongs to the active TMS plus placebo group, the second set of values belongs to the sham TMS plus venlafaxine group and the last set of values belongs to the active TMS plus venlafaxine group.

Figure 2.

Funnel Plot of Standard Error by Standardized Mean Difference

We included 10 RCTs in the meta-analysis (Table 1) encompassing 654 participants across all studies. All these studies compared two arms encompassing active TMS administered concurrently with antidepressant titration with a comparator placebo arm that combined sham TMS with the same antidepressant as in the active arm (Bretlau et al., 2008; Brunelin et al., 2014; Garcia-Toro et al., 2001; Herwig et al., 2007; Huang et al., 2012; Pan et al., 2020; Poulet et al., 2004; Rumi et al., 2005; Wang et al., 2017b). In all these studies, both arms received the same antidepressants titrated and dosed in the same manner, with active and sham TMS, respectively. In addition, the antidepressants were initiated as part of the RCT. Specifically, six of them combined selective serotonin reuptake inhibitors (SSRIs) (encompassing citalopram, paroxetine, and escitalopram) with TMS (Bretlau et al., 2008; Garcia-Toro et al., 2001; Huang et al., 2012; Pan et al., 2020; Poulet et al., 2004; Wang et al., 2017b), three combined a serotonin-norepinephrine reuptake inhibitors (SNRI) (venlafaxine) with TMS (Brunelin et al., 2014; Herwig et al., 2007; Rossini et al., 2005) and one combined amitriptyline with TMS (Rumi et al., 2005). Three RCTs used fixed doses of antidepressants with no adjustments during the study (Bretlau et al., 2008; Pan et al., 2020; Poulet et al., 2004). Scales used to measure outcomes included Hamilton Depression Rating Scale (HDRS), Montgomery-Asberg Depression Rating Scale (MADRS), and Beck Depression Inventory (BDI). Six out of ten RCTs reporting the antidepressant efficacy of the two groups found the combined use of antidepressants with active TMS resulted in significantly lower scores on depression scales and faster onset of antidepressant effect compared to antidepressants with sham TMS.

Amongst these studies, there was significant variability in TMS parameters, with two RCTs using low frequency TMS (1 Hz and 5 Hz) (Brunelin et al., 2014; Rumi et al., 2005) and others using high frequency TMS (10–20 Hz) (Bretlau et al., 2008; Garcia-Toro et al., 2001; Herwig et al., 2007; Huang et al., 2012; Pan et al., 2020; Poulet et al., 2004; Rossini et al., 2005; Wang et al., 2017b). The change in HDRS was used as the outcome measure for the meta-analysis because it was the most used across studies. Poulet et al (Poulet et al., 2004) and Rumi et al (Rumi et al., 2005) did not report HDRS scores, but they did report MADRS scores. Therefore, to incorporate these studies into the meta-analysis, MADRS score means and standard deviations were converted to equivalent HDRS score means and standard deviations using the chart provided by Leucht et al (Leucht et al., 2018). Because Poulet et al. (Poulet et al., 2004) only reported changes in depressive symptoms as percent changes relative to baseline MADRS scores, all values used in the meta-analysis were converted to percent changes relative to baseline HDRS scores to allow for incorporation of Poulet et al (Poulet et al., 2004) into the meta-analysis. Figure 3 shows the results of the meta-analysis.

Figure 3.

Meta-analysis of antidepressants combined with active versus sham TMS

The results of the random effects meta-analysis suggest that active TMS combined with antidepressants had greater efficacy for MDD treatment than sham TMS combined with the same antidepressants as in the active arm (Hedge’s g = 1; 95% CI [0.27, 1.73]). As expected, the studies had high heterogeneity (I² = 94%; p<0.00001). We assessed publication bias for included studies and found none [t(8) = 1.92, p=0.09).

One RCT was notable in examining head-to-head comparisons between active TMS combined with antidepressants from two different classes (escitalopram and sertraline, which are SSRIs, and venlafaxine, which is an SNRI) versus sham TMS combined with the same doses and titration, regimens of these three antidepressants (Rossini et al., 2005). The RCT had six arms (n=17 in active TMS and escitalopram, n=16 in active TMS and sertraline, n=17 in active TMS and venlafaxine, n=17 in sham TMS and escitalopram, n=16 in sham TMS and sertraline, n=16 in sham TMS and venlafaxine). The study showed a significant difference in HAMD score reduction in all active TMS arms compared to sham TMS until the end of four weeks. There was no significant difference among various antidepressants in the active arm. What was notable was that venlafaxine combined with active TMS had the greatest remission rates (80%) at the end of week five compared to sertraline (71.4%) and escitalopram (68.7%). In the sham TMS arm, sertraline had the greatest remission rates (66.5%) compared to venlafaxine (50%) and escitalopram (46.7%) (Rossini et al., 2005).

Discussion

The current systematic review and meta-analysis reviewed the literature on studies that compared active TMS combined with antidepressants dosed in the same manner for all trial participants versus sham TMS combined with the same antidepressants as in the active arm. Our meta-analysis found that active TMS combined with antidepressants dosed in the same manner for all trial participants may have greater efficacy for the treatment of MDD compared to antidepressants combined with sham TMS. Since sham TMS uses a significantly weaker electric field than active TMS (Smith and Peterchev, 2018), it would be pragmatic to state that the combination of TMS and antidepressants included in this review showed a large effect size compared to antidepressants alone. However, the findings of the meta-analysis should be interpreted with a high degree of caution given the large amount of heterogeneity between the studies, encompassing experimental procedures (antidepressant class, TMS frequency, number of sessions, and stimulation intensity) used in the studies. A key takeaway for clinicians from the limited synthesis in this review/meta-analysis is to continue the antidepressants that patients are on when starting TMS so that there would not be any need to re-initiate the antidepressant when tapering the index TMS regimen.

We stratified studies by class of antidepressants (Table 1). Although there are mechanistic differences between different classes of antidepressants (SSRI versus SNRI versus TCA), we are limited in suggesting mechanistic differences between these antidepressant classes when combined with TMS (Taylor et al., 2005). We intend for this meta-analysis to advocate for RCTs comparing the short-term and long-term effects of various antidepressant classes on TMS in MDD.

Theta burst stimulation (TBS) has shown a greater effect size than sham TMS (0.64) for the treatment of MDD(Voigt et al., 2021). TBS is more efficient than other high-frequency TMS protocols because it requires 3 to 9 minutes for administration compared to an hour for 10 Hz TMS. An accelerated intermittent TBS (iTBS) protocol called Stanford Neuromodulation Therapy (SNT) using ten daily sessions of intermittent TBS (iTBS) for five days has recently shown great promise in the treatment of MDD(Cole et al., 2022; Cole et al., 2020). Borrowing from ECT studies wherein nortriptyline and lithium combined with ECT showed less relapse compared with just nortriptyline or placebo (Prudic et al., 2013; Sackeim et al., 2001), and extrapolating from a notable study in Group 1(Rossini et al., 2005), it would be worthwhile to compare the effect of accelerated iTBS protocols with candidate molecules from one antidepressant class (SSRIs, SNRIs) versus accelerated iTBS and placebo on remission rates in MDD.

Multiple medical examples already exist where combining treatment methods results in greater efficacy. For example, multiple antidiabetic drugs with different mechanisms are often administered simultaneously to elicit a greater drop in hemoglobin A1c than drug monotherapy(American Diabetes, 2020; Petrides et al., 2015). In a seminal crossover trial, combining ECT and clozapine resulted in an initial 50% response in patients with treatment-resistant schizophrenia who received the combination and a 47% response rate in patients randomized to receive only clozapine initially but were then crossed over into the clozapine-ECT arm(Petrides et al., 2015).

We would also like to advocate for maintenance TMS sessions and future endeavors to bring consensus in the regimen for maintenance TMS based on patient/illness characteristics and current psychotropics. In the context of combining antidepressants and TMS effectively to leverage the benefits of both treatments, it is germane to discuss maintenance TMS. Antidepressants have been the mainstay to maintain remission in MDD (Borges et al., 2014; Williams et al., 2009). Maintenance TMS regimens could also offer ways to sustain remission in combination with antidepressants or their stead (d’Andrea et al., 2023; Matsuda et al., 2023; Rachid, 2018). A previous study showed significantly higher proportions of patients sustaining remission in MDD when maintenance TMS was combined with antidepressants compared to getting either treatment alone (Wang et al., 2017a). Given that the patient population getting TMS for MDD may already be treatment-resistant, relapse rates over time are high in MDD without maintenance TMS sessions (Kedzior et al., 2015; Senova et al., 2019). Therefore, it is imperative to consider individualized TMS treatment regimens in combination with suitable antidepressants to sustain remission (d’Andrea et al., 2023; Matsuda et al., 2023; Rachid, 2018). Akin to the symptom-titrated algorithm-based longitudinal ECT (STABLE) regimen in ECT, it is essential to move towards consensus regimens for maintenance TMS protocols to sustain remission in MDD (Perera et al., 2016; Rachid, 2018; Wilson et al., 2022). This needs individualizing maintenance TMS regimens based on patients’ antidepressants. Future research needs to identify differences, if any, in remission depending on antidepressants, and the only way to conclusively examine this would be blinded RCTs comparing a combination of active maintenance TMS and uniformly dosed antidepressant(s) to a combination of active TMS and a placebo. This would also help identify if remission rates differ between antidepressant classes when combined with TMS, thereby helping to tailor maintenance TMS regimens based on antidepressants, in addition to patient and illness characteristics.

Conclusions and Future Directions

This review and meta-analysis aim to fill a critical knowledge gap in clinical practice. Currently, there is no consensus on clinical decision-making with antidepressants when delivering TMS in MDD. Given that antidepressants continue to be the mainstay of treatment to prevent relapse in MDD, there is no consensus on tailoring index or maintenance TMS regimens based on antidepressant regimens. As a first step in filling this knowledge gap, we reviewed the literature on studies that compared active TMS combined with antidepressants dosed in the same manner for all trial participants versus sham TMS combined with the same antidepressants as in the active arm. Our goal with the review and meta-analysis is to advocate for RCTs comparing the short-term and long-term effects of various antidepressant classes on TMS in MDD. The results of our meta-analysis indicate the potential benefits of combining antidepressants with TMS.

Given the benefit showcased by accelerated protocols such as the SNT in expediting remission in treatment-resistant MDD, it would be worthwhile to identify optimal antidepressants to combine with SNT to maintain remission in MDD. This would also help to generate consensus guidelines on combining antidepressants and TMS. Further, these results could also provide directions to combine medications that have shown effectiveness for treating MDD, such as ketamine, and potentially effective molecules like psilocybin with SNT (Anand et al., 2023; Goodwin et al., 2022; Goodwin et al., 2023) to maintain remission.

Highlights.

• Both TMS and antidepressants can treat treatment-resistant MDD.

• There is a knowledge gap on antidepressant selection and dosing strategies when adding on TMS in treatment-resistant MDD.

• We reviewed studies that combined TMS and antidepressants rigorously in RCTs for MDD.

• The results of our meta-analysis indicate the benefit of continuing antidepressants when initiating TMS in TRD.

• We advocate for RCTs examining the short-term and long-term effects of various antidepressant classes on accelerated TMS protocols in MDD.

• This can also optimize and individualize maintenance TMS protocols to prevent relapse in MDD.